ML23163A143

PR-050 - 51FR09829 - Station Blackout

ML23163A143
Person / Time
Issue date:	03/21/1986
From:	Chilk S NRC/SECY
To:
References
PR-050, 51FR09829
Download: ML23163A143 (1)

Text

ADAMS Template: SECY-067 DOCUMENT DATE: 03/21/1986 TITLE: PR-050 - 51FR09829 - STATION BLACKOUT CASE

REFERENCE:

PR-050 51FR09829 KEYWORD: RULEMAKING COMMENTS Document Sensitivity: Non-sensitive - SUNSI Review Complete

DOCKET NUMBER PROPOSED RULE OOt [ iEr:

~1 u~~.c [7590-01)

~ D15 p3:22 Off IC:: r. '.:t t I I OOCK[TIM.1 ;, " ,nilr:

Federal Register Notice of Final Rulemaking i3R Nt,

• NUCLEAR REGULATORY COMMISSION 10 CFR Part 50 Station Blackout

'. '\~ -;_.. ,,..\I; Nuclear Regulatory Co11111ission.

ACTION: Fina 1 rule.

SUMMARY

The Nuclear Regulatory Commission is amending its regulations to require that light-water-cooled nuclear power plants be capable of withstanding

• a total loss of alternating current (ac) electric power (called "station black-out") for a specified duration and maintaining reactor core cooling during that period. This requirement is based on information developed under the Commission's study of Unresolved Safety Issue A-44, "Station Blackout." The am~ndment i ~ intended to provide further assurance that a loss of both offsite

,_,.~*'!':':.:;,~c,iisite emergency ac power systems will not adversely affect the public health and safety.

EFFECTIVE DATE: July 18, 1988 FOR FURTHER INFORMATION CONTACT: Aleck Serkiz. Division of Reactor and Plant Systems, Office of Nuclear Regulatory Research. U.S. Nuclear Regulatory Commission, Washington, DC 20555, Telephone: (301) 492-3555.

(7590-01]

SUPPLEMENTARY INFORMATION:

Background

The alternating current (ac) electric power for essential and nonessential service in a nuclear power plant is supplied primarily by offsite power.

Redundant onsite emergency ac power systems are also provided in the event that all offsite power sources are lost. These systems provide power for various safety functions, including reactor core decay heat removal and containment heat removal, which are essential for preserving the integrity of the reactor core and the containment building, respectively. The reactor core decay heat can also be removed for a limited time period by safety systems that are independent of ac power.

The term 11 station blackout 11 means the loss of offsite ac power to the essential and nonessential electrical buses concurrent with turbine trip and the unavailability of the redundant onsite emergency ac power systems (e.g., as a result of units out of service for maintenance or repair, failure to start on demand, or failure to continue to run after start). If a station blackout persists for a time beyond the capability of the ac-independent systems to remove decay heat, core melt and containment failure could result.

The Commission's existing regulations establish requirements for the design and testing of onsite and offsite electric power systems that are intended to reduce the probability of losing all ac power to an acceptable level. (See General Design Criteria 17 and 18, 10 CFR Part 50, Appendix A.) The existing regulations do not require explicitly that nuclear power plants be designed to assure that core cooling can be maintained for any specified period of loss of a11 ac power.

As operating experience has accumulated, the concern has arisen that the reliability of both the onsite and offsite emergency ac power systems might be less than originally anticipated, even for designs that meet the requirements of General Design Criteria 17 and 18. Many operating plants have experienced a total loss of offsite power, and more occurrences can be expected in the

[7590-01]

Complete and Partial losses of Offsite Power at Nuclear Power Plants, 11 published in February 1985; and NUREG/CR-4347, "Emergency Diesel Generator Operating Experience, 1981-1983, 11 published in December 1985. The major results of these studies are given below.

0 Losses of offsite power can be characterized as those resulting _from plant-centered faults, utility grid blackout, and severe-weather-induced failures of offsite power sources. Based on operating experience, the frequency of total losses of offsite power in operating nuclear power plants was found to be about one per 10 site-years. The median restoration time was about one-half hour, and 90 percent of the offsite power losses were restored within approximately 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> (NUREG/CR-3992).

0 The review of a number of representative designs of onsite emergency ac power systems has indicated a variety of potentially important failure causes. However, no single improvement was identified that could result in a significant improvement in overall diesel generator reliability.

Data obtained from operating experience in the period from 1976 to 1980 showed that the typical individual emergency diesel generator failure rate was about 2.5 x 10- 2 per demand {i.e., one chance of failure in 40 demands), and that the emergency ac power system unavailability for a plant which has two emergency diesel generators, one of which was required for decay heat removal, was about 2 x 10- 3 per demand (NUREG/CR-2989).

°Compared to the data in NUREG/CR-2989, updated estimates of emergency diesel generator failure rates indicated that diesel generator reliability has improved somewhat from 1976 to 1983. For the period 1981 to 1983, the mean failure rate for all demands was about 2.0 x 10- 2 per demand (i.e., one chance of failure in 50 demands). However, the data also indicate that the probability of diesel generator failures during actual demands (i.e., during losses of offsite power) is greater than that during surveillance tests (NUREG/CR-4347).

[7590-01]

0 Given the occurrence of a station blackout, the likelihood of resultant core damage or core melt is dependent on the reliability and capability of decay heat removal systems that are not dependent on ac power. If sufficient ac-independent capability exists, additional time will be available to restore ac power needed for long-term cooling (NUREG/CR-3226).

0 It was determined by reviewing design, operational, and site-dependent factors that the expected frequency of core damage resulting from station blackout events could be maintained near 10- 5 per reactor-year with readily achievable diesel generator reliabilities, provided that plants are designed to cope with station blackout for a specified duration. The duration for a specific plant is based on a comparison of the plant's characteristics to those factors that have been identified as the main contributors to risk from station blackout (NUREG-1032).

The staff's technical findings show that station blackout does not pose an undue risk to public health and safety. The findings summarized above show that recovery from loss of offsite power occurs for the most part in less than 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />, emergency diesel generator reliability is high (i.e.,~ 0.95), and that given a station blackout the likelihood of core damage is more dependent on decay heat removal systems that are non-ac-dependent. However, plant design and operational characteristics, plus site-dependent factors (such as anticipated weather conditions) introduce a level of variability which warrants a need for plant-specific coping analyses to provide greater assurance that core cooling can be maintained until ac power is restored. Thus the Commission believes that §50.63 of 10 CFR Part 50 will bring about a significant increase in protection to the public health and safety. As a result of station blackout coping analyses, improved guidance will be provided to licensees regarding maintaining minimum emergency diesel generator reliability to minimize the probability of losing all ac power. In addition, the Commission is amending its regulations by adding a new §50.63 to require that all nuclear power plants be capable of coping with a station blackout for some specified period of time.

The period of time for a specific plant will be determined based on a comparison

[7590-01]

of the individual plant's design with factors that have been identified as the main contributors to risk of core damage resulting from station blackout.

These factors, which vary significantly from plant to plant because of considerable differences in design of plant electric power systems as well as site-specific considerations, include: (1) redundancy of onsite emergency ac power sources (i.e., number of sources minus the number needed for decay heat removal), (2) reliability of onsite emergency ac power sources (usually diesel generators), (3) frequency of loss of offsite power, and (4) probable time to restore offsite power. The frequency of loss of, and time to restore, offsite power are related to grid and switchyard reliabilities, historical weather data for severe storms, and the availability of nearby alternate power sources (e.g., gas turbines). Experience has shown that long duration offsite power outages are caused primarily by severe storms (hurricanes, ice, snow, etc.).

The objective of the rule is to reduce the risk of severe accidents resulting from station blackout by maintaining highly reliable ac electric power systems and, as additional defense-in-depth, assuring that plants can cope with a station blackout for some period of time. The rule requires all plants to be able to cope with a station blackout for a specified acceptable duration selected on a plant-specific basis. All licensees and applicants are required to assess the capability of their plants to cope with a station blackout (i.e.,

determine that the plant can maintain core cooling with ac power unavailable for an acceptable period of time), and to have procedures and training to cope with such an event. Licensees may use an alternate ac power source if that source meets specific criteria for independence and capacity and can be shown to be available within one hour to cope with a station blackout. A coping analysis is not required for those plants that choose this alternate ac approach if the alternate ac can be demonstrated by test to be available to power the shutdown buses within 10 minutes of the onset of station blackout.

Use of an alternative ac source, one that minimizes colTll'lon mode failure, is a preferred option since this approach will also benefit other safety concerns.

On the basis of station blackout studies conducted for USI A-44 and presented in the reports referenced above, the NRC staff has developed Regulatory Guide

[7590-01]

1.155 entitled "Station Blackout," which presents guidance on (1) maintaining a high level of reliability for emergency diesel generators, (2) developing procedures and training to restore offsite and onsite emergency ac power should either one or both become unavailable, and (3) selecting a plant-specific acceptable station blackout duration which the plant would be capable of sur-viving without core damage. Application of the methods in this guide ~ould result in selection of an acceptable station blackout duration (e.g., 2, 4, 8, or 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />) which depended on the specific plant design and site-related characteristics acceptable to the staff. However, applicants and licensees could propose alternative methods to those specified in the regulatory guide in order to justify other acceptable durations for station blackout capability.

Additionally, the regulatory guide on station blackout presents guidance on quality assurance and specifications for alternate ac source(s) and non-safety-related equipment required for coping with station blackout. The equipment installed to meet the station blackout rule must be implemented so that it does not degrade the existing safety-related systems. This is to be accomplished by making the non-safety-related equipment independent to the extent practicable from existing safety-related systems. The guidance provided in the regulatory guide illustrates the specifications that the staff would find acceptable for non-safety systems and equipment. The quality assurance guidance for the non-safety-related equipment for which there are no existing NRC quality assurance requirements (e.g., Appendix B, Appendix R) embody the following elements: (1) design control and procurement document control, (2) instructions, procedures and drawings, (3) control of purchased material, equipment and services, (4) inspection, (5) test and test control, (6) inspection, test and operating status, (7) non-confonning items, (8) corrective action, (9) records, (10} audits. NRC inspections will focus on the implementation and the effectiveness of these quality controls as described in the regulatory guide.

Based on the rule and regulatory guide, those plants with an already low risk from station blackout would be required to withstand a station blackout for a relatively short period of time and probably would need few, if any, modifi-cations as a result of the rule. Plants with currently higher risk from

(7590-01]

station blackout would be required to withstand somewhat longer duration blackouts. Depending on their existing capability, these plants might need to make hardware modifications (such as increasing station battery capacity or condensate storage tank capacity) in order to cope with the longer station blackout duration. The rule requires that each light-water-cooled nuclear power plant licensed to operate must be able to withstand for a specified duration and recover from a station blackout. The rule requires each plant to perform a coping analysis and identify the coping duration, along with the basis therefor and a description of procedures established for coping and recovery. If modifications to equipment or plant procedures are necessary, these are to be identified and a schedule provided for implementing such changes.

It should be noted, based on all evidence that staff has on hand, that no undue risk exists with, or without, the promulgation of the station blackout rule.

However, station blackout may still remain an important contributor to residual risk. This station blackout rule will enhance safety by accident prevention and thereby reduce the likelihood of a core damage accident being caused by a station blackout occurrence. This does not mean however, that further enhancements in reducing the overall residual risk are not achievable by additional improvements in severe accident management, given the assumption that core damage occurs, whether from station blackout sequences or other causes (such as small or large loss-of-coolant accident sequences).

Initiatives that provide such safety enhancements (through improvements of core damage management procedures) are currently being pursued apart from the station blackout rule. Therefore, this rule should be viewed as being in the same accident prevention context as the ATWS rule (§50.62) and the fire protection rule (§50.48) in that it recognizes, as the other two rules recognize, multiple failure possibilities resulting from co11111on cause effects that should be addressed. This concern has been recognized in the Introduction to Appendix A of 10 CFR Part 50.

[7590-01]

Proposed Rule On March 21, 1986, the Commission published a proposed rule in the Federal Register (51 FR 9829) that would require (1) light-water-cooled nuclear power plants to be capable of coping with a station blackout for a specified duration, and (2) licensees to determine the maximum duration for which their plants as currently designed are able to cope with a station blackout. A 90-day comment period expired on June 19, 1986.

On April 3, 1986 (13 days after the proposed rule was published), the NRC published in the Federal Register (51 FR 11494) a notice of availability and request for corrments on a draft regulatory guide entitled "Station Blackout 11 (Task SI 501-4). This draft guide provided guidance for licensees to comply with the proposed station blackout rule. Many lette:r.s corrmenting on the proposed rule also included comments on the draft regulatory guide. Responses to these comments provided below address the public co1t111ents on the draft guide as well as on the proposed rule.

Comments on the Proposed Rule The Convnission received 53 letters commenting on the proposed rule. 2 Forty-five of these were from the nuclear industry, comprised *Of electric utilities, consortiums of electric utilities, vendors, a trade association, and an architect/engineering firm. Other letters were submitted by the Union of Concerned Scientists, the Department of Nuclear Safety of the State of Illinois, a representative of the Professional Reactor Operator Society, a citizens group, a consultant, and three individuals. Largely, the industry comments were opposed to generic rulemaking to resolve the station blackout issue. The Nuclear Management and Resources Council (NUMARC), formerly the Nuclear Utilities Management and Resources Committee, submitted, along with its 2Copies are available for public inspection and copying for a fee at the NRC Public Document Room at 1717 H Street, NW, Washington, DC.

[7590-01]

conwnents on the proposed rule, a set of four industry initiatives that it believes would resolve this issue without rulemaking. Thirty-nine of the industry letters supported NUMARC's submittal. NUMARC proposed a fifth initia-tive {see item 21) by letter dated October 5, 1987. On the other hand, the Union of Concerned Scientists, the Illinois Department of Nuclear Safety, and the citizens group supported the Conwnission's objective in the proposed rule, but did not believe the rule and guidance associated with the rule went far enough to reduce the possibility of a serious accident that could be initiated by a total loss of ac power.

Every letter was reviewed and considered by the staff in formulating the final resolution of USI A-44. Because of the large number of corrments, it was not practical to prepare formal responses to each one separately. However, since many colTIT!ents were on similar subjects, the discussion and response to the comments have been grouped into the following subjects: 3

1. Quality classification of modifications

2. Whether the backfit analysis adequately implements the Backfit Rule

3. Cost-benefit and whether §50.63 meets "substantial increase in the overall protection of the public health and safety"

4. Whether NRC should require substantial improvements in safety that go beyond those proposed in this rulemaking

5. The need for generic rulemaking

6. Applicability of the proposed §50.63 to specific plants

7. Plant-specific features and capabilities 3The first four subjects are ones on which the ColTITlissioners specifically requested public corrments when the proposed rule was published.

[7590-01]

8. The source term used to estimate consequences

9. Specificity on the extent of required coping studies

10. Acceptable duration for coping with a station blackout

11. Credit for alternate or diverse ac power sources

12. Trends on the reliability of ac power sources

13. Sharing of emergency diesel generators between units at multi-unit sites

14. Clarification of the definitions of station blackout and diesel generator failure

15. Specificity and clarification of requirements

16. Technical comments on NUREG-1032

17. Relationship of USI A-44 to other NRC Generic Issues

18. An alternative of plant-specific probabilistic assessments

19. Procedures and operator actions during station blackout

20. Schedule provisions in the proposed §50.63.

21. Industry initiatives The comments and responses to each of these subjects are presented on the following pages.

[7590-01]

1. Quality Classification of Modifications The Conmission requested conments on whether the staff should give further consideration to upgrading to safety grade the plant modifications needed (if any) to meet the proposed rule. Upgrading to safety grade would further ensure appropriate licensee attention is paid to maintaining equipment in a High state of operability and reliability.

Conments - The prevailing view by industry on this subject is represented by the following corrunents submitted by NUMARC:

Quality classification is unnecessary - Equipment used to prevent or respond to a station blackout should be sufficiently available and operable to meet its required function. To this extent, the Commission's desire that appropriate attention be paid to maintaining a sufficiently high state of operability and reliability is appropriate. The point of departure begins with the method for achieving this objective.

Specifically, by itself, a "safety grade" classification scheme does not solely equate with high states of equipment operability and reliability.

Such classification systems too often can become a documentation exercise more than a process for providing the requisite level of system functionality.

Duquesne Light agreed with this view and expressed the following conments:

Any plant modifications or additional equipment required to meet the proposed rule should not be specified safety grade. For equipment which is to be manually started and placed in service for testing or in the event of a loss of power condition there is no necessity for specifying safety grade since adequate reliability can be obtained through normal surveillance testing and the proper maintenance of conmercial power plant equipment. The cost difference in safety grade vs. conmercial grade modifications is significant and must be emphasized.

The opposite point of view was taken by the Illinois Department of Nuclear Safety.

No credit should be given for the capability of equipment to respond to a station blackout unless that equipment was originally designed, constructed, inspected, perfonnance tested, qualified, certified for the intended safety-related purpose, and the equipment is maintained to the highest industry safety standards.

[7590-01]

Gulf States Utilities conmented, The proposed rule does not provide sufficient direction on the quality classification of plant modifications that may be required to meet the rule .

• • • the quality classification of plant modifications implemented to meet the proposed rule should be conmensurate with classification of the system they support.

Response - The proposed §50.63 does not specifically address the topic of safety classification of plant modifications; however, detailed guidance is provided in Regulatory Guide 1.155 dealing with quality assurance and equipment specifications for non-safety-related equipment. Any safety-related equipment used either presently, or in modifications resulting from this rule, should meet the criteria currently applied to such equipment.

The technical analyses performed for US! A-44 (NUREG-1032) show that plant-centered events (i.e., those events in which design and operational characteristics of the plant itself play a role in the likelihood of loss of offsite power), and area- or weather-related events (e.g., grid reliability or external influences on the grid) are the dominant causes of loss of offsite power. Neither seismic events nor events related to single failure causes were found to be major contributors to loss of offsite power. Therefore, both the staff's findings and public co111T1ents received do not support an explicit need for plant modifications for coping with station blackout to be seismically qualified.

The substantial increase in protection sought by this rule can be achieved by modifications which meet criteria somewhat less stringent than generally required by safety grade criteria. Safety-related equipment modifications to meet all safety-grade-related criteria would be more burdensome and expensive and would likely achieve only a very small further reduction in risk. The major contributors to the residual risk of loss of offsite power are adequately dealt with by modifications which conform to the quality assurance and equipment specification guidance provided in Regulatory Guide 1.155.

[7590-01]

2. Whether the Backfit Analysis Adequately Implements the Backfit Rule In addition to conrnents on the merits of the proposed rule, the Convnission specifically requested conrnents on whether the backfit analysis for this rule adequately implements the Backfit Rule, §50.109 of 10 CFR Part 50.

Conrnents - The Convnission received two differing views in response to this request. On one hand, NUMARC expressed the view that the proposed rule does not meet the backfit rule standard because the analysis of the factors set forth in §50.109{c) were not adequately considered by the staff. Specifically, NUMARC stated:

1. Installation and continuing costs associated with the backfit have been underestimated.

2. Potential impacts on radiological exposure of facility employees should be further addressed.

3. The relationship to proposed and existing regulatory requirements should be considered further.

4. Potential impacts of differences in facility, type, design, or age should be considered further.

5. The reduction in risk from offsite releases to the public has been overestimated.

On the other hand, the Ohio Citizens for Responsible Energy (OCRE) and the Union of Concerned Scientists conrnented that the backfit rule should not apply to the proposed rule. OCRE took the position that "application of the backfit rule to [NRC] rulemakings *** is plainly illegal," and the Commission is not empowered to consider costs to licensees in deciding whether to impose new requirements. The Union of Concerned Scientists co11Jnented that the cost-benefit analysis should not be applied in this case because safety improvements

[7590-01]

are needed to secure compliance with existing NRC regulations, specifically General Design Criterion 17, Electric Power Systems (Appendix A to 10 CFR Part 50).

Response - NUMARC's corrments on the backfit analysis were taken into account by the staff in revising the draft version of NUREG-1109, "Regulatory Bac'kfit Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout," and a separate appendix that addresses the factors in §50.109(c) was added to that report. All but Item 2 above are on the same subjects as letters from other commenters and are discussed in more detail under subjects 3 (Item 1), 6 (Item 4), 8 (Item 5), and 17 (Item 3) in this section. NUMARC's Item 2, the potential impact on radiological exposure of facility employees, would need to be assessed in detail only if it were a major factor in the value-impact analysis. The effect of radiological exposure on facility employees, if any, would be extremely small in comparison to the reduction in radiological exposure to the public from accident avoidance. Therefore, this factor would have no impact on the overall value-impact analysis.

Contrary to OCRE's and the Union of Concerned Scientists' conments, the Commission may subject the rulemaking process to internal controls. Moreover, the Commission is empowered to consider the costs of incremental safety improvements which go beyond the level of safety necessary to ensure no undue risk to the public health and safety. See UCS, et al., v. NRC, D.C. Cir. Nos.

85-1757 and 86-1219 (August 4, 1987). The improvements embodied in §50.63 go beyond the level of safety necessary to ensure no undue risk. Finally, contrary to the Union of Concerned Scientists' corrment on GDC 17, new station blackout measures cannot be imposed on licensees as a matter of compliance with GDC 17, under the compliance exception in the backfit rule, paragraph 50.109(a)(4)(i). GDC 17 does not explicitly require that each plant be able to withstand station blackout for a specified time, or that each licensee perfonn a coping assessment and make whatever modifications may be necessary in the light of that assessment. Nor are any of these highly specific requirements logically compelled by any part of GDC 17. Moreover, GDC 17 has never been

[7590-01]

interpreted by the staff or the Commission to contain these specific requirements. Thus, to impose them under GDC 17 would amount to a backfit which resulted from a new staff and Co1t111ission interpretation of GOC 17.

The issue in this rulemaking is whether some additional protection is warranted beyond that already provided. The Co1t111ission is entitled to inquire, *and seek public conment on, whether additional safety measures should be imposed where there is a substantial increase in the overall protection of public health and safety and the cost of implementation is justified in view of this increased protection.

3. Cost-Benefit Analysis and Whether §50.63 Meets the "Substantial Increase in the Overall Protection of the Public Health and Safety" Chairman Zech and Co111T1issioner Roberts requested comments on the analysis of cost benefit, value impact, and safety improvements and the station blackout standing on the overall risk (e.g., is the reduction of risk only a small percentage of the overall risk, or is it a major component of an already small risk?). Chairman Zech and Conmissioner Roberts were particularly interested in specific comments assessing whether or not this proposal meets the "substantial increase in the overa 11 protection of the pub 1i c hea 1th and safety .*. 11

- threshold now required by the backfit rule.

Comments - (A) One of the major conments by industry on the cost-benefit analysis was that the costs of implementing the proposed requirements have been underestimated. NUMARC and the Atomic Industrial Forum (AIF) colTITlented that the cost estimates for hardware modifications reported in NUREG/CR-3840, "Cost Analysis for Potential Modifications To Enhance the Ability of a Nuclear Plant To Endure Station Blackout," were too low. Co1T1T1onwealth Edison and other utilities felt that performance of an analysis to determine the maximum duration a nuclear plant could cope with a station blackout would be substantially costlier than what is estimated in NUREG-1109. Industry also expressed concern that the interpretations associated with the proposed rule could lead to substantial costs above those addressed by the NRC staff in its

[7590-01]

backfit analysis. AIF conmented that "The estimate of 120 NRC man-hours per plant [for NRC review] *** appears inadequate to account for technical review and evaluation of the determination of maximum coping capability and of the description of station blackout procedures which the rule would require each licensee to submit."

(B) Several conmenters expressed the view that the NRC failed to consider all the risks associated with a station blackout in its value-impact assessment.

The Union of Concerned Scientists thought independent failures, in addition to failures that lead to a station blackout, should be included. One individual stated that "both NRC reports [NUREG-1109 and NUREG-1032] are completely deficient in that neither look at sabotage." OCRE commented that seismic events should also be considered.

(C) With respect to safety improvements and overall risk, different points of view were expressed. On one hand, NUMARC commented that, while the risk reduction might be large for a limited number of plants, the risk reduction associated with the majority of plants will be small. Thus, as a general matter, the reductions in risk offered by the proposed rule constitute a small percentage of the overall risk, a risk which is already small (and acceptable).

AIF stated that there is no standard by which to conclude that "substantial additional protection will be realized."

A different view was expressed by the Union of Concerned Scientists who stated that "station blackout is clearly a major component of the total risk posed by operating nuclear plants. The magnitude of the total risk is largely unknowable due to the enormous uncertainty which surrounds probabilistic assessments."

Response - (A) In order to adequately respond to industry's corrments above, the staff and NRC contractors reviewed the cost estimates associated with imple-menting the station blackout rule. Based on this review, the estimated costs for hardware modifications were reviewed and are in the range of from 20 percent to almost 140 percent greater than the estimates in NUREG/CR-3840,

[7590-01]

depending on the specific modification considered. On average, the cost estimates for hardware backfit were found to be approximately 80 percent greater than estimated in NUREG/CR-3840. However, the cost estimates in NUREG/CR-3840 were not used by the staff in the value-impact analysis in the draft version of NUREG-1109 where estimates approximately 100 percent greater than the NUREG/CR-3840 estimates were used. Therefore, the revised c~st estimates used in the final value-impact analysis are not significantly different from the estimates used in the draft version.

Industry's comments on the costs to assess a plant's capability to cope with a station blackout were based on the proposed rule that required an assessment of the maximum coping capability and the potentially unbounded nature of such an assessment. Based on public conments, the Convnission has revised the final rule to modify the requirement for licensees to determine the maximum coping capability. (See response to public conments in subject number 9.) Instead. a coping assessment is required only for a specific duration. The cost for such a study is estimated to be from 70 to 100 percent higher than the original estimates by the staff, and these revised costs are used in the final value-impact analysis.

The staff revised its estimate of the resource burden on NRC for review from 120 to 175 person-hours per reactor. This revision was based on technical review required for other comparable NRC activities.

(B) The technical analyses performed for USI A-44 indicated that the contribution to core damage frequency from independent failures, in addition to failures that must occur to get to a station blackout, is low. Likewise, results of USI A-44 studies and other probabilistic risk assessments have shown that, for station blackout sequences, the contribution to core damage frequency from seismic events is low.

Not all events can be analyzed on a probabilistic basis. Sabotage is an example. Even though sabotage was not explicitly considered in the staff's value-impact analysis, it is discussed in NUREG-1109 under other

[7590-01]

considerations. These considerations support the conclusion that a station blackout rule will provide a substantial safety benefit.

(C) The revised value-impact analysis perfonned for the resolution of USI A-44 indicates that there are substantial benefits in tenns of reduced core damage frequency and reduced risk to the public that result from the station blackout rule, and the costs are warranted in light of these benefits. The best estimate for the overall value-impact ratio is 2,400 person-rem per million dollars. Even if those plants with the highest risk (and therefore the greatest risk reduction) were not considered, the value-impact ratio for the remaining plants is still favorable (i.e., about 1,500 person-rem per million dollars).

Analyses reported in NUREG-1150, "Reactor Risk Reference Document (draft 11 issued for co1T1T1ent in February 1987), 4 indicate that station blackout is a dominant risk contributor to overall residual risk for most of the six plants analyzed. These results support the colTITlent by the Union of Concerned Scientists in response to the Commissioner's request for colTITlents on this subject.

4. Whether NRC Should Require Substantial Improvements in Safety that Go Beyond Those Proposed in this Rulemaking.

Conmissioner Asselstine requested colTITlents on whether the NRC should require substantial improvements in safety with respect to station blackout, like those being accomplished in some other countries, which can be achieved at reasonable cost and which go beyond those proposed in this rulemaking.

Co111nents - NRC received eight letters that included colTITlents on this subject.

Five of these were from the nuclear industry, none of which felt that the 4Free single copies may be obtained from the Division of Information Support Services, U.S. Nuclear Regulatory Con111ission, Washington, DC 20555.

[7590-01]

approach to station blackout taken in European countries should be used to justify safety improvements that go beyond the proposed §50.63. The main justification for industry's argument is that foreign countries may have reasons for requiring activities that differ from, or exceed, those in the U.S.

For example, Washington Public Power Supply Systems {WPPSS) conmented, "It is not apparent that the details of U.S. grid stabilities and onsite power reliabilities are substantially similar enough to those found abroad to warrant a simple adoption of these [European] measures."

In another conment from industry on this subject, NUMARC stated that there are several reasons why many of the features for coping with a station blackout in new French nuclear power plants may already exist at most U.S. plants. In fact, they said, "The French approach to station blackout does not appear to depart significantly from current regulatory approaches in the U.S. 11 Similarly, AIF stated, "The assertions of extensive station blackout coping capability at foreign {notably European) nuclear power plants are not sufficiently substantiated to serve as even part of the basis for the proposed requirements."

Three other letters {Union of Concerned Scientists, OCRE, and Illinois Department of Nuclear Safety) supported the NRC rulemaking to require all plants to be able to cope with a station blackout, but urged the Co1TUTiission to go beyond the proposed rule. The Illinois Department of Nuclear Safety stated that:

The goal of holging the expected frequency of core damage from station blackout to 10- per reactor-year is not sufficiently stringent. With _7 relatively modest modifications to the proposed rule, a frequency of 10 appears achievable at reasonable cost. Specifically, the rule should require no less than 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br /> decay heat removal capacity instead of only four or eight hours in the proposed rule, in the event of a blackout.

Reponse - The staff agrees with industry's conments that foreign countries may have valid reasons for imposing requirements that differ from or exceed those in the U.S. For example, it appears that there is a higher frequency of losses of offsite power in France than in the U.S. This experience, along with French safety objectives, led the French to design their new standard nuclear power

[7590-01]

plants to be able to cope with a very long duration station blackout (i.e., up to three days). The French safety approach and their station blackout design features are documented in NUREG-1206, "Analysis of French (Paluel) Pressurized Water Reactor Design Differences Compared to Current U.S. PWR Designs," June 1986.

The ColTITlission believes that the staff has adequately considered foreign approaches in preventing core melt from station blackout in developing the resolution of USI A-44. Although the rule requires plants to be able to cope with station blackout for a specific duration, that duration is not specified in the rule. Guidance to determine an acceptable duration is included in Regulatory Guide 1.155. This guidance should apply to most plants, but if there were adequate justification, different requirements {either more or less stringent than the regulatory guide) could be applied to specific plants. The use of alternate ac sources provides a means to achieve further incremental decreases in core melt frequency.

5. The Need for Generic Rulemaking Comments - Five letters from the nuclear industry co1T111ented that generic rulemaking is not necessary to resolve the station blackout issue. Their reasons for this issue were as follows:

A generic rulemaking is inappropriate since the historic number of sites experiencing a loss of all offsite power is small. (Texas Utilities)

The station blackout issue should be handled on a plant-specific basis and does not need to be resolved by generic rulemaking. Each plant has unique probability for a loss-of-power event based on transmission system, location of plant, and onsite power systems. (Duquesne Light)

The Commission need not pursue generic rulemaking in order to resolve a non-generic issue. In the proposed station blackout rule, the number of plants of concern is acknowledged to be limited. (NUMARC)

Station blackout has been found not to be a generic issue. Station blackout risk is plant specific and, according to the staff's own analyses, the proposal requirements are expected to result in modifications at no more than a few facilities, if at any. Requiring all

[7590-01]

licensees to undertake extensive analyses under the provisions of the proposed rules when only a small group of plants may have a need for remedial action is not appropriate. (Alf)

Response - The Co11111ission believes that a rule is appropriate to ensure that station blackout is addressed at all nuclear power plants. The plant-specific features that contribute to risk for station blackout (e.g ** diesel gene'rator configuration, probability of loss of offsite power) are considered by the staff in the station blackout regulatory guide to determine an acceptable coping duration for each plant. Even though not all sites have experienced a loss of offsite power, there is not sufficient assurance that such events would e not occur in the future. Since historic experience has shown that a total loss of offsite power occurs about once every 10 site-years, and many nuclear plants have operated for less than 10 years, it is not surprising that some plants have experienced a loss of offsite power while others have not.

Even though it is likely that many plants will not need hardware modifications to comply with the rule, the assessment of station blackout coping capability for a specific duration and implementation of associated procedures will affect a safety benefit for all plants. The "limited number of plants of concern" in NUMARC's letter refers to those plants having the highest risk from station blackout (i.e., those that would need hardware modifications). Without a

- plant-specific assessment, these plants can not be identified. Even excluding these plants from consideration, the staff's analysis has shown that the improvements in safety associated with the rule are consistent with backfit considerations set forth in §50.109.

6. Applicability of the Proposed §50.63 to Specific Plants Co11111ents - Four letters included co11111ents or questions regarding the applicability of the rule to specific plants. For example, does the rule apply to high-temperature gas-cooled reactors (HTGR) (i.e., Fort St. Vrain)? What about TMI-2 or plants that are near completion but will not have an operating license prior to the amendment's effective date? Houston Power and Lighting Company wrote:

[7590-01]

Proposed Section 50.63 provides schedular guidance for implementing station blackout-related modifications on plants that already hold oper-ating licensees or will be licensed to operate prior to the effective date of the amendment. Plants who may be NTOL's [near-term operating license]

but will not be licensed prior to the amendment's effective date should be accorded the same compliance period under parts (c} and (d} of this section. Otherwise this proposed rule could be interpreted to imply that plants not licensed prior to the effective amendment date must comply with the rule and make all necessary modifications prior to receiving 'an O.L.

[operating license]. The rule should be amended to address plants which are scheduled to receive an O.L. within a short time following imple-mentation of this rule.

Response - Rather than identifying specific plants for which the rule does not apply, §50.63(a) specifies when it does apply {i.e., "each light-water-cooled nuclear power plant licensed to operate"). Since Fort St. Vrain is an HTGR, the generic rule would not apply. Station blackout will be considered individually for that plant based on its unique design. Since TMI-2 is not licensed to operate, likewise the rule would not apply to that plant. Any plant licensed to operate after the date the rule becomes effective will comply with the same 270-day schedule for inforr.1ation submittal applied to plants previously licensed. This affords NTOLs the same compliance features as plants already licensed to operate.

7. Plant-Specific Features and Capabilities Conments - A number of utilities described plant-specific features and capabilities that reduced the risk posed by a station blackout event compared to the staff's analysis. Examples of such features are given below.

0 Availability of alternate, independent ac power sources such as diesel generators, gas turbines, or nearby "black start" ac power sources.

0 Extremely reliable offsite power supplies because of multiple right-of-ways or underground feeders to back up above ground transmission lines.

0 Dedicated shutdown systems and associated diesel generators to meet the fire protection requirements of Appendix R to 10 CFR Part 50.

[7590-01]

°Common or shared systems between two units at multi-unit sites such as direct current (de) power, auxiliary feedwater, or diesel generators.

Response - The analyses perfonned for USI A-44 clearly show that plant-specific features do affect the risk from station blackout, and the station blackout regulatory guide takes this into account in providing guidance on diffrent acceptable coping durations depending on the most significant of these features. Those plants with extremely reliable offsite and onsite ac power supplies need only have a very short (e.g., 2-hour) coping duration to be acceptable. Plants that have a dedicated shutdown system with its own inde-

- pendent power supply could take credit for this system to cope with a station blackout. The final rule and Regulatory Guide 1.155 have been clarified to give credit for alternate ac power supplies (see response to subject 11).

Therefore, the Commission believes that for almost all sites, plant-specific differences have been adequately accounted for in the resolution of USI A-44, but the door is open to licensees who believe their plants have additional capability that should be considered by the staff in demonstrating compliance with the rule.

8. The Source Term Used To Estimate Consequences Comments - NUMARC and others in the industry commented that the consequences of offsite releases that would result from a station blackout event are overestimated, and new source term infonnation would lead to the prediction of much lower consequences for this event. Several corrmenters felt that the approach taken by the staff to estimate consequences of a station blackout event was improper -- decreasing by a factor of three the estimated consequences of the siting source term (SSTl) from NUREG/CR-2723, "Estimates of the Financial Consequences of Nuclear Power Reactor Accidents" (September 1982).

AIF felt that "implementation of any requirements resulting from the resolution of USI A-44 should be deferred until the results of the source term research

[7590-01]

can be taken into account." They based this statement on the premise that if the consequences used in the staff's value-impact analysis were reduced by a factor of 10, none of the alternatives would be feasible.

The Union of Concerned Scientists expressed a different point of view in their letter which said 11 * *

• available evidence indicates that the consequences of an accident involving station blackout may be even worse than those estimated either in WASH-1400 or the NRC's more recent studies. 11 Response - NRC has had an extensive research effort underway since about 1981 to evaluate severe accident source terms. The staff has reviewed the results of this research to take into account the public co11111ents received on this subject. Since there is still a great deal of uncertainty regarding source terms and associated consequences, the staff revised its value-impact analysis for USI A-44 considering a range of estimates for consequences of a station blackout.

The NRC research on severe accident source terms has resulted in the develop-ment of significant new analytical tools by NRC contractors, as discussed in NUREG-0956, "Reassessment of the Technical Bases for Estimating Source Terms, 11 July 1986. The analytical methods developed, generally referred to as the Source Term Code Package (STCP), have been used to analyze a number of severe accident sequences for five reference plants, namely: Peach Bottom, a BWR Mark I design; Sequoyah, a PWR ice condenser; Surry, a PWR with a sub-atmospheric containment; Grand Gulf, a BWR with a Mark III containment; and Zion, a PWR with a large dry containment (NUREG-1150, "Reactor Risk Reference Document,"

Draft for Conment, February 1987).

The results of these analyses show that releases from station blackout sequences can be expected to vary significantly depending upon the plant and the specific sequence. Although generalizations are difficult, it appears that calculations using the STCP yield release fractions for most of the sequences range from about one third of an SSTl release (for the case of Surry, without

[7590-01]

condensation) to roughly one order of magnitude less than this. However, the uncertainties in our present understanding also do not preclude the possibility of a large release, approaching that of the SSTl estimate.

To determine the consequences in tenns of person-rem, given the above range of release fractions, data taken from ~UREG/CR-2723 indicate that the variations in person-rem associated with releases of magnitude SSTl, SST2 and SST3 are virtually identical to the variations in latent cancer fatalities for the same three releases. Hence, the estimated change in latent cancer fatalities with release fractions provides a reliable indication of change in person-rem as well.

Table 10 in NUREG/CR-2723 presents variations in estimated latent cancer fatalities associated with changes in SSTl release fractions (for all elements except noble gases). This table shows that a release fraction of one-third of an SSTl release would yield a value of about 50 percent of the latent cancer fatalities (and person-rem) of an SSTl release. Similarly, a release fraction of one-third of an SSTl release would yield an estimated person-rem of about 15 percent of that associated with an SSTl release. Consequently, for value-impact calculations, the staff estimated the range of consequences of station blackout, in terms of person-rem, to be from 0.15 to 0.5 of the estimated person-rem of an SSTl release. As noted, the original value-impact analysis was based on 0.3 times the estimated person-rem of an SSTl release.

With regard to a possible delay in the resolution of USI A-44 until better 11 11 source terms become available, key considerations appear to be when better source terms are likely to become available and to what degree uncertainties in phenomenology as well as diffrences between investigators will be resolved.

Although research on source terms is expected to continue well into the future, improvements in our knowledge are expected to be largely evolutionary beyond this point, in that the major phenomena appear to have been accounted for, at least in a first-order fashion, both in NRC as well as industry models.

Resolution and narrowing of the remaining uncertainties would also benefit from improved experiments and analytical models that are likely to become available

(7590-01]

gradually. For these reasons, significantly better source terms than those presently available are likely to be forthcoming only after a number of years.

Since the range of severe accident source terms and consequences suggested above from estimating station blackout sequences is sufficiently broad to cover likely improvements in source term knowledge, the resolution of USI A-44 should not be delayed.

9. Specificity on the Extent of Required Coping Studies ColTlllents - Several letters by industry expressed concern that the studies necessary to demonstrate that a plant can cope with a station blackout are not well defined and could potentially be unbounded. These co1T111ents focused on two main points. First, the proposed rule required plants to determine the maximum duration the plant could cope with a station blackout, yet the draft regulatory guide included specific guidance on acceptable coping durations (e.g., 4 or 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />). Determining the maximum duration, rather than assessing the plant's capability for a specific acceptable duration, could be an open-ended requirement. Along these lines, NUMARC stated:

Unless the required coping demonstration is specifically bounded by clearly stated definitions, assumptions, and criteria, there could conceivably be hundreds of supporting special effects analyses which licensees may have to consider as a result of the exercise of discretion by individual staff reviewers. Under the rule as proposed, licensees cannot ascertain the ultimate requirements they will be expected to meet (including the potential plant modifications they will need to make) to demonstrate compliance.

Second, industry also colTlllented on the potential open-endedness of analyses to determine the operability of equipment in environmental conditions resulting from a station blackout (e.g., without heating, ventilation, and air condition-ing). Unless these analyses were well defined, industry felt the analyses could be much more costly than estimated by the staff. However, NUMARC made the following statement relating to the need for detailed prescriptive require-ments by NRC that appears to contradict their earlier statement.

[7590-01]

The point **** is not that regulations must be prescriptive by their very nature. Prescriptive regulations, which outline in detail exactly what steps are required by licensees to satisfy a proposed regulation, are, in many instances, unnecessary and counterproductive.

Response - With regard to the proposed requirement that each plant detennine its maximum duration for coping with station blackout, the staff agrees, with the industry co11111ents. First of all, it would be difficult to adequately define "maximum duration" in this sense. Second, if licensees determine that their plants can cope with a station blackout for a specified duration and restore ac power through an acceptable coping analysis, the additional safety benefit gained from simply the knowledge that a longer, or "maximum duration,"

coping duration exists is small. Third, the costs for assessing "maximum duration" will be higher since more extensive analyses will be required to analyze a transient which would go beyond the coping analysis for a specified duration and recovery from station blackout. Therefore, the rule and regulatory guide have been revised accordingly to delete the requirement for licensees to determine a plant's maximum coping capability.

With regard to the corrments on assessments to determine equipment operability during a station blackout, the staff feels strongly that such assessments are necessary to determine a plant's response to station blackout. By deleting the requirement to detennine a plant's "maximum" coping capability, the assessment of equipment operability would not be as costly as assumed by industry.

Guidance on acceptable coping assessments is provided in the station blackout regulatory guide. Also, guidelines to evaluate the effects of loss of ventilation under station blackout conditions are provided in Appendix E of NUMARC-8700, "Guidelines and Technical Bases for NUMARC Initiatives Addressing Station Blackout at Light Water Reactors. These efforts provide additional 11 definitions, criteria, and standards for licensees' assessments of equipment operability without the need for "prescriptive regulations" by NRC.

In order to further evaluate industry's co11111ents on this subject, NRC requested Sandia National Laboratories to identify specific tasks necessary to determine operability of equipment during a station blackout and to estimate the cost to

[7590-01]

perform these tasks. Results of this study were used in the revised value-impact analysis performed for this issue {"Equipment Operability During Station Blackout Events," NUREG/CR-4942).

10. Acceptable Duration for Coping with a Station Blackout Corrments - Several corrments with differing views were directed at guidance in the draft regulatory guide on acceptable station blackout coping durations in order for plants to comply with the proposed rule.

- Washington Public Power Supply corrmented that "it should be possible for certain utilities to demonstrate [an acceptable] zero hour blackout."

One individual recorrmended "that a 30 minute period be a margin, and that no duration under 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> be accepted by the staff." NucleDyne Engineering commented that "advanced reactors should require the capability to safely withstand a station blackout of at least 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />," and the Illinois Department of Nuclear Safety wrote that "the rule should require no less than 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br /> decay heat removal capability instead of only 4 or 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />."

Response - Although diverse corr1nents were received on this subject, none provided supporting analysis or information to back up the opinions expressed.

- However, the staff did reanalyze the estimated risk from station blackout events for different plant- and site-related characteristics and revised its guidance on acceptable coping durations accordingly based on a goal of limiting the average contribution to core damage from station blackout to about 10- 5 per reactor-year. Most plants would still need a 4- or 8-hour coping capability.

Those few plants with the most redundant onsite emergency ac power system, coincident with significantly lower than average expected frequency of loss of offsite power, would need only a 2-hour capability to be acceptable. Any plant with minimum redundancy in the onsite emergency ac power system coincident with low reliability and a significantly higher than average expected frequency of loss of offsite power would need to substantially improve its ac power reliability or be able to cope with a station blackout for 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />.

[7590-01]

11. Credit for Alternate or Diverse AC Power Sources Conrnents - Ten letters from the utility industry coRJTiented that more credit should be allowed for the availability of alternate power sources such as onsite gas turbines. The coRJTients below represent the utilities' viewpoint.

The station blackout rule should be clarified to allow credit for diverse and very reliable offsite power sources or diverse and very reliable onsite electrical generation. (Public Service Company of Colorado)

The option of providing an additional alternate source of ac power is eliminated by [the proposed resolution]. The inconsistency in this approach can best be understood by considering an example at a generic nuclear power station. (Toledo Edison)

If the licensee were to provide an additional independent diesel generator capable of providing the necessary ac power to prevent station blackout, the licensee *** would still be required to withstand at least 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> without ac power. They would receive no credit for the additional diesel generator in the coping analysis. If the licensee were to use that same diesel engine to power a charging pump, even though it would be of less significance to mitigation of reactor core damage than the diesel generator, the licensee could take credit for it in coping with the blackout. (Toledo Edison)

Since a diesel-powered charging pump will not provide for equipment loading flexibility, lighting, ventilation, instrumentation, etc., it is obviously of lower value than an additional source of ac power. The fixed category approach taken in [the proposed resolution], however, will not permit taking credit for the same diesel engine when used as a generator though the actual reliability for the machine is the same. (Toledo Edison)

Response - The proposed regulation did not intend to ignore the alternative of adding additional power sources or taking credit for such sources if they already exist. For example, as specified in the regulatory guide, if a licensee added an emergency diesel generator to one of its plants that had minimum redundancy in the onsite emergency ac power sytem, the acceptable station blackout coping duration could be reduced. For some plants, however, adding a diesel generator would not result in a reduction in the acceptable coping duration, and the point made by Toledo Edison is a valid one. The rule and regulatory guide have been revised to clarify that alternate ac power sources are given credit to cope with a station blackout provided that certain

[7590-01]

criteria are met (e.g., independence, redundancy, high reliability, maintenance, and testing).

12. Trends on the Reliability of AC Power Sources Comments - Five letters included comments on the reliability of ac power sources. Four letters from industry felt that improved ac power reliability should be factored into the staff's technical analysis. Examples of these comments include the following:

" **. the frequency of loss of offsite power activities has been decreasing *** " (Washington Public Power Supply System);

" **. offsite power availability in the absence of regulation has significantly improved over the past decade." (Southern California Edi son Company) ;

"[NUREG/CR-4347] *** shows an improvement in diesel generator reliability over that shown in the earlier document [NUREG/CR-2989]." (General Electric); and "Typically the reliability of onsite power systems increases during the first few years following startup." (Gulf States Utilities)

The Illinois Department of Nuclear Safety, on the other hand, felt that potential vulnerabilities still exist in onsite emergency ac power systems, and

- licensees should demonstrate that they have taken steps to reduce the probability of loss of ac power.

Response - The staff and its contractors have extensively analyzed the industry experience and trends in ac power reliability as documented in NUREG-1032, NUREG/CR-2989, NUREG/CR-3992, and NUREG/CR-4347. Trends have shown that two aspects of ac power reliability have improved somewhat -- the reduced frequency of losses of offsite power due to plant-centered events, and a slight improvement in average diesel generator reliability from 1976 through 1983.

These factors have been taken into account in the staff's analyses and the resolution of USI A-44. However, data also demonstrate that there are practical limits on ac power reliability, and the defense-in-depth approach of being able to cope with a station blackout is warranted.

[7590-01]

13. Sharing of Emergency Diesel Generators Between Units at Multi-Unit Sites Corrments - Several letters from industry stated that some plants with two units on a site have the capability to crosstie electrical buses between units and therefore have improved flexibility in providing ac power. Since the magnitude of the electrical loads necessary to provide core cooling during a st~tion blackout is significantly less than that required for a design basis accident, it could be possible to provide ac power to both units at the site using only a single diesel generator.

Response - The proposed rule and draft regulatory guide do not prohibit the approach discussed above. If licensees can demonstrate that such crosstie capability exists, procedures are in place to accomplish the crosstie and shed nonessential loads (if necessary), and no NRC regulations are violated (such as separation, minimum redundancy, and independence), then credit would be given for this capability as shown in Regulatory Guide 1.155 (i.e., reduced acceptable station blackout coping durations for greater diesel generator redundancy).

14. Clarification of the Definitions of Station Blackout and Diesel Generator Failures Comments - (A) Three conmenters from the utility industry reconmended that the definition of station blackout in §50.2 should be clarified to exclude ac power from the station batteries through inverters. This source of ac power from the station batteries would be available in the event of a loss of both the offsite and onsite emergency ac power sources (i.e., diesel generators).

(B) Several from industry conmented that the definition of diesel generator failure should be clarified, particularly with respect to the treatment of short-term failures that can be recovered quickly. Sargent and Lundy Engineers CORITlented that:

[7590-01]

A definition of failure on demand for emergency diesel generators needs to be provided. Under the context of a station blackout, a diesel generator which fails to start automatically upon detection of an offsite power loss, but is successfully started manually from the main control room or from the local control panel, should not be considered a failure on demand.

Response - (A) The staff agrees with conment A and revised the definition of station blackout accordingly.

(B) Based on actual experience, failures of diesel generators to start due to failures in the auto-start system make up less than 20 percent of all diesel generator failures. Therefore, discounting these failures would not have a 4t significant impact on overall diesel generator reliability statistics. However, the staff agrees in principle with conment Band has clarified the station blackout regulatory guide so that auto-start failures of diesel generators need not be counted in determining the failure rate if the diesel generator is capable of being started manually immediately after it does not start automatically.

15. Specificity and Clarification of Requirements Convnents - Public comments were received regarding the specificity and clarifi-cation of the proposed rule and draft regulatory guide. These ranged from

- general to specific comments as the following two excerpts indicate:

We are concerned that, if the proposed rule is adopted, the staff will promulgate regulatory guidance criteria which will be unrealistic and excessive, i.e., compounding the event with other accidents, imposing passive failure criteria, applying seismic, environmental qualification and other qualifications to equipment that could otherwise be used in response to such an event, etc. (Maine Yankee Atomic Power Company)

Definitions of Pl and P2 [in Table 3 of the draft Regulatory Guide] use frequency of extremely severe weather and severe weather interchangeably, thus creating confusion in the definition. (Washington Public Supply System)

Response - Some of the comients on this subject relate to other subjects discussed elsewhere in this section. Some conments were quite specific while others were general in nature or expressed views that were not substantiated

[7590-01]

with backup material. The staff has taken these co11111ents into consideration and revised and clarified the rule and regulatory guide accordingly. Additional guidance is provided in NUMARC-8700 which has been reviewed by the staff and referenced in the regulatory guide as providing a method the staff finds acceptable for meeting the rule.

16. Technical Co11111ents on NUREG-1032 Comments - In addition to co11111ents on the proposed rule and draft regulatory guide, several letters contained co11111ents on the staff's draft technical report, NUREG-1032, "Evaluation of Station Blackout Accidents at Nuclear Power Plants."

Response - NUREG-1032 was issued in draft form for public co11111ent in May 1985 (50 FR 24332). The comments received were reviewed and considered by the staff and resulted in a re-evaluation of the technical analysis. Details of the specific comments and responses are not presented here. Rather, NUREG-1032 was revised extensively over the past year to address the public co11111ents. In general, the overall conclusions on the risk from station blackout events did not change significantly as a result of the reanalysis. One of the major changes resulting from the reanalysis was a revision to the definitions of plant characteristics, especially the clustering of plants into site and

- weather-related groups (Appendix A in NUREG-1032). These changes are reflected in revisions to the guidance in the station blackout regulatory guide to determine plant-specific acceptable station blackout coping durations.

17. Relationship of USI A-44 to Other NRC Generic Issues Comnents - The major public co11111ent regarding the relationship of USI A-44 to other NRC generic safety issues was that the proposed rule may not be necessary or should be postponed because of ongoing work to resolve related generic issues. Some colllllents were general in nature such as the following one from Southern California Edison Company:

[7590-01]

Promulgation of a final station blackout rulemaking at this time will unnecessarily complicate the final resolution of related generic technical issues ***. The NRC must develop and implement a program to coordinate the resolution of all power-related generic issues prior to finalizing any individual proposed rule.

AIF suggested that the implementation of any requirements for station blackout be deferred until the requirements from USI A-45, Shutdown Decay Heat ,Removal Requirements, are known and until the effect of source tenn changes can be evaluated.

NUMARC mentioned specific proposed and existing regulatory requirements that should be considered because they could reduce the need for a station blackout rule (e.g., B-56, Diesel Generator Reliability, and GI 23, Reactor Coolant Pump Seal Failures). Other related issues mentioned in the public comments were A-30, Adequacy of Safety-R~lated DC Power Supplies, and implementation of safe shutdown facilities to meet the fire protection requirements of Appendix R.

Response - The question that needs to be addressed is "should a requirement be imposed now to reduce risk, or should it be postponed until related issues are resolved sometime in the future?" Potentially, this could result in sub-stantial delays, thereby not resolving generic safety issues in a timely manner. The staff has considered the resolution of USI A-44 in light of the related issues mentioned in the comments. Although these issues are identified

- as separate tasks within NRC, they are all managed in a well established program that coordinates all related issues. A brief discussion of the most relevant issues is presented below. (Additional information is provided in NUREG-1109, "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout.")

Resolution of USI A-45 will occur at some time following issuance of the station blackout rule (§50.63) and after plant-specific station blackout coping evaluations have been perfonned by licensees per NUMARC/NUGSBO Initiative 5, utilizing guidelines provided in NUMARC-8700. Further, the resolution of USI A-45 is expected to be highly plant-specific and focused on loss of decay heat removal considerations from other causes beyond station blackout. Utilization will be made of A-44 evaluations (as applicable) and any plant equipment

[7590-01]

modification needs identified from A-45 will be carefully evaluated to maximize effective use of previously identified A-44 equipment needs.

Maintaining emergency diesel generator reliability, the purpose of B-56, is an integral part of the resolution of USI A-44. However, the CoD111ission believes that additional defense-in-depth will achieve a substantial increase i~ protec-tion to public health and safety.

The resolution of GI 23 (reactor coolant pump seal leakage) deals with loss of reactor coolant system inventory and associated degraded core conditions. USI A-44 deals with station blackout induced effects, which result in loss of ac power, thereby impacting a broader spectrum of plant equipment and safety-related functions. Although the resolution of GI 23 will contribute to establishing a higher level of assurance that seal leakage will be minimized (thereby minimizing the need for power to replace water inventory losses over the station blackout duration and recovery phase), resolution of GI 23 by itself will not address the broader scope of USI A-44 safety concerns.

Some licensees have implemented dedicated shutdown systems that are independent of normal and emergency ac power to meet Appendix R requirements. If appli-cable, these features would be credited in the resolution of USI A-44 by providing the capability to cope with a station blackout.

Thus, the resolution of LISI A-44 is coordinated with related generic issues, and implementation of a final resolution should not be delayed further.

(Response to coDlllents on the effect of source term changes is included in subject number 8.)

18. An Alternative of Plant-Specific Probabilistic Assessments CoD111ents - Several utilities suggested that, in lieu of the requirements in the rule, licensees should be permitted to submit plant-specific evaluations to demonstrate that the frequency of core damage from station blackout events is 10- 5 per reactor-year or less. In a similar vein, the suggestion was made that

[7590-01]

NRC should specify a target level of reliability for ac power systems in order to satisfy NRC's criteria for core damage frequency. A few licensees submitted limited probabilistic assessments to show that for some plants station blackout could have a very small probability of severe consequences.

Response - The ColTlllission does not preclude licensees from submitting ~lant-specific probabilistic assessments to support a detennination that station blackout would have a very small probability for causing core damage. However, the requirements of the rule must be met. The Commission would observe that the use of probabilistic assessments was important as input to the regulatory decisionmaking that culminated in the station blackout rule and related guidance. As expressed in the Commission's Safety Goal Policy statement of August 1986 (51 FR 28044), the Commission has acquired a reasonable degree of confidence about the usefulness and value of probabilistic assessments in assisting regulatory decisionmaking on complex safety issues. In short, such assessments are of value in complementing and focusing the more traditional and deterministic defense-in-depth approaches. On the other hand, any licensee must decide whether or not its plant-specific ac power configuration and other related equipment are sufficiently unique to merit the conduct and submittal of a probabilistic assessment as part of achieving compliance to §50.63. The Commission's experience also indicates that probabilistic assessments are resource intensive and can be of marginal utility if their only end result is to delay rule compliance.

19. Procedures and Operator Actions During Station Blackout Co111T1ents - (A) Several letters from industry conmented that, in response to Generic Letter 81-04, "Emergency Procedures and Training for Station Blackout Events, dated February 21, 1981, utilities already have procedures in place to 11 prepare plant operations for station blackout events. Owners' groups have established generic guidance for station blackout operating procedures for licensees to use in developing plant-specific procedures. A representative of the Professional Reactor Operator Society corrmented that:

[7590-01]

Generic procedures are used by most operating facilities. These procedures are not carried into adequate depth of specific power plant operations. The industry has relied too heavily on generic procedures and has not given a real look at what specific steps must be taken.

Extrapolation of these procedures must be required. Specific maintenance procedures must be established and followed.

(B) Other conrnents on procedures related to the timeliness of operator, actions, both inside and outside the control room. Houston Lighting and Power suggested that:

In Section 3.1 (Part 6) [of the regulatory guide], the first sentence should be revised to read, 'Consideration should be given to timely operator actions both inside and outside of the control room that *** , so that credit can be taken for existing equipment that may not have actuation and control from the control room.

Illinois Power Company recolTITlended that:

• • . Section C.3.3, Item 3.a, of the proposed regulatory guide should be modified to read:

a. The system should be capable of being actuated and controlled from the control room, or if other means of control are required (e.g.,

manual jumping of control logics or manual operation of valves), it should be demonstrated that these steps can be carried out in a timely fashion.

Response - (A) Licensees may take credit for station blackout procedures already in place to comply with the station blackout rule. However, for the most part, these procedures were developed without having the benefit of a plant-specific assessment to determine whether a plant could withstand a station blackout for a specific duration. Therefore, these procedures may need to be modified after licensees have detennined an acceptable station blackout coping duration and evaluated their plant's response to a station blackout of this duration.

(B) The staff agrees with the co1T1Tients related to operator actions outside the control room, and the regulatory guide was revhed accordingly.

[7590-01]

20. Schedule Provisions in the Proposed §50.63 Corrments - Two letters contained conments on the proposed schedule in §50.63.

OCRE felt the scheduling provisions in the proposed rule were far too generous.

One individual reconmended that the schedule be modified to require licensees to submit, within 9 months of the date of the amendment, a list of modifications along with a proposed schedule to implement those modifications.

(According to the proposed rule, licensees would not have to submit a schedule for implementing equipment modifications until after the staff received and reviewed licensees' submittals on their plant's acceptable station blackout duration.)

Response - The staff agreed in part with these corrrnents, and the schedule was revised accordingly. Section 50.63(c)(l)(iii) now requires that licensees submit within 9 months after the rule is issued a list of equipment modifications and a proposed schedule for implementing them. A final schedule would be developed after NRC has reviewed the licensees' submittal of their plant's acceptable station blackout duration.

21. Industry Initiatives Corrunents - In addition to conments on the proposed rule, NUMARC endorsed the following five initiatives 5 to address the more important contributors to station blackout:

1. Each utility will review their site(s) against the criteria specified in NUREG-1109, and if the site(s) fall into the category of an eight-hour site after utilizing all power sources available, the utility will take actions to reduce the site{s) contribution to the overall risk of station blackout. Non-hardware changes will be made within one year. Hardware changes will be made within a reasonable time thereafter.

5NUMARC initially proposed a set of four initiatives. The fifth initiative regarding the performance of a coping assessment was provided in NUMARC-8700, which was submitted by letter from J. Opeka (NUMARC) to T. Speis (RES) dated November 23, 1987. A copy is available for public inspection and copying for a fee at the NRC Public Document Room at 1717 H Street NW., Washington, DC.

[7590-01]

2. Each utility will implement procedures at each of its site(s) for:

a. coping with a station blackout event

b. restoration of ac power following a station blackout event, and

c. preparing the plant for severe weather conditions (e.g.,

hurricanes and tornados) to reduce the likelihood and consequences of a loss of offsite power and to reduce the overall risk of a station blackout event.

3. Each utility will, if applicable, reduce or eliminate cold fast-starts of emergency diesel generators for testing through changes to technical specifications or other appropriate means.

4. Each utility will monitor emergency ac power unavailability utilizing data utilities provide to INP0 on a regular basis.

5. Each utility will assess the ability of its plant(s) to cope with a station blackout. Plants utilizing alternate AC power for station blackout response which can be shown by test to be available to power the shutdown busses within 10 minutes of the onset of station blackout do not need to perform any coping assessment. Remaining alternate AC plants will assess their ability to cope for 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />.

Plants not utilizing an alternate AC source will assess their ability to cope for 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />. Factors identified which prevent demonstrating the capability to cope for the appropriate duration will be addressed through hardware and/or procedural changes so that successful demonstration is possible.

NUMARC previously opposed generic rulemaking and felt that the first four initiatives would resolve the station blackout issue.

Response - These five initiatives now include many of the elements that are included in the NRC resolution of USI A-44. The staff has followed up on the NUMARC initiatives through a series of meetings in 1986 through 1987. The result has been the development of NUMARC-8700 which provides guidelines and criteria acceptable to the staff. The procedures in NUMARC-8700 have been referenced in Regulatory Guide 1.155 as providing guidance acceptable to the staff for meeting the requirements of the rule. Table 1 in Regulatory Guide 1.155 provides a cross-reference to NUMARC-8700 and notes where the regulatory guide takes precedence. NUMARC's previous concerns have been addressed in the development of Regulatory Guide 1.155 and NUMARC-8700.

(7590-01]

Finding of No Significant Environmental Impact: Availability The Commission has detennined under the National Environmental Policy Act of 1969, as amended, and the ColTITlission's rules in Subpart A of CFR Part 51, that this rule is not a major Federal action significantly affecting the quality of the human environment, and therefore, an environmental impact statement is not required. There are not any adverse environmental impacts as a result of the rule because there is no additional radiological exposure to the general public or plant employees, and plant shutdown is not required so there are no additional environmental impacts as a result of the need for replacement power.

The environmental assessment and finding of no significant impact on which this determination is based are available for inspection and copying for a fee at the NRC Public Document Room 1717 H Street, NW, Washington, DC. Single copies of the environmental assessment and the finding of no significant impact are available from Mr. Warren Minners, Office of Nuclear Regulatory Research, U. S.

Nuclear Regulatory Commission, Washington, DC 20555, Telephone:

(301) 492-7827.

Paperwork Reduction Act Statement This final rule amends infonnation collection requirements that are subject to the Papen*wrk Reduction Act of 1980 (44 U.S.C. 3501 et seq.). These requirements were approved by the Office of Management and Budget approval number 3150-0011.

Regulatory Analysis The Conrnission has prepared a regulatory analysis on this final regulation. The analysis examines the costs and benefits of the alternatives considered by the Corrmission. A copy of the regulatory analysis, NUREG-1109, "Regulatory/

Backfit Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout," is available for inspection and copying for a fee at the NRC Public Document Room, 1717 H Street, NW, Washington, DC 20555.

(7590-01]

Regulatory Flexibility Certification As required by the Regulatory Flexibility Act (5 U.S.C. 605(b, the ColTITlission certifies that this rule does not have a significant economic impact on a substantial number of small entities. The rule requires that nuclear power plants be able to withstand a total loss of ac power for a specified time duration and maintain reactor core cooling during that period. These facilities are licensed under the provisions of §§50.2l(b) and 50.22 of 10 CFR Part 50. The companies that own these facilities do not fall within the scope of 11 small entities" as set forth in the Regulatory Flexibility Act or the small business size standards set forth in regulations issued by the Small Business Administration in 13 CFR Part 121. List of Subjects in 10 CFR Part 50 Antitrust, Classified information, Fire prevention, Incorporation by reference, Intergovernmental relations, Nuclear power plants and reactors, Penalty, Radiation protection, Reactor siting criteria, Reporting and recordkeeping requirements. For the reasons set out in the preamble and under the authority of the Atomic - Energy Act of 1954, as amended, the Energy Reorganization Act of 1974, as amended, and 5 U.S.C. 553, the NRC is adopting the following amendments to 10 CFR Part 50. Part 50 - Domestic Licensing of Production and Utilization Facilities

1. The authority citation for Part 50 is revised to read as follows:

AUTHORITY: Secs. 102, 103, 104, 105, 161, 182, 183, 186, 189, 68 Stat. 936, 937, 938, 948, 953, 954, 955, 956, as amended, sec. 234, 83 Stat. 1244, as amended (42 U.S.C. 2132, 2133, 2134, 2135, 2201, 2232, 2233, 2236, 2239, 2282); secs. 201, as amended, 202, 206, 88 Stat. 1242, as amended, 1244, 1246 (42 u.s.c. 5841, 5842, 5846).

[7590-01] Section 50.7 also issued under Pub. L. 95-601, sec. 10, 92 Stat. 2951 (42 U.S.C. 5851). Sections 50.10 also issued under secs. 101, 185, 68 Stat. 936, 955, as amended (42 U.S.C. 2131, 2235); sec. 102, Pub. L. 91-190, 83 Stat. 853 (42 U.S.C. 4332). Sections 50.23, 50.35, 50.55, and 50.56 also issued under sec. 185, 68 Stat. 955 (42 U.S.C. 2235). Sections 50.33a, 50.55a and Appendix Q also issued under sec. 102, Pub. L. 91-190, 83 Stat. 853 (42 U.S.C. 4332). Sections 50.34 and 50.54 also issued under sec. 204, 88 Stat. 1245 (42 U.S.C. 5844). Sections 50.58, 50.91, and 50.92 also issued under Pub~ L. 97-415, 96 Stat. 2073 (42 U.S.C. 2239). Section 50.78 also issued under sec. 122, 68 Stat. 939 (42 U.S.C. 2152). Sections 50.80-50.81 also issued under sec. 184, 68 Stat. 954, as amended (42 U.S.C. 2234). Section 50.103 also issued under sec. 108, 68 Stat. 939, as amended (42 U.S.C. 2138). Appendix F also issued under sec. 187, 68 Stat. 955 (42 U.S.C. 2237). - For the purposes of sec. 223, 68 Stat. 958, as amended (42 U.S.C. 2273);

 §§50.lO(a), (b), and (c), 50.44, 50.46, 50.48, 50.54, and 50.80(a) are issued under sec. 161b, 68 Stat. 948, as amended (42 U.S.C. 220l(b)); §§50.lO(b) and (c), and 50.54 are issued under sec. 16li, 68 Stat. 949, as amended (42 U.S.C.

220l(i)); and §§50.9, 50.55(e), 50.59(b), 50.70, 50.71, 50.72, 50.73, and 50.78 are issued under sec. 1610, 68 Stat. 950, as amended (42 U.S.C. 2201(0)).

2. In §50.2, definitions of "alternate ac sourc~" and "station blackout" are added in the alphabetical sequence to read as follows:

 §50.2 Definitions "Alternate ac source" means an alternating current (ac) power source that is available to and located at or nearby a nuclear power plant and meets the

- following requirements: (1) Is connectable to but not normally connected to the offsite or onsite emergency ac power systems; (2) Has minimum potential for conman mode failure with offsite power or the onsite emergency ac power sources; (3) Is available in a timely manner after the onset of station blackout; and (4) Has sufficient capacity and reliability for operation of all systems required for coping with station blackout and for the time required to bring and maintain the plant in safe shutdown (non-design basis accident).

(7590-01]

 "Safe shutdown (non-design basis accident (non-OBA))" for station blackout means bringing the plant to those shutdown conditions specified in plant technical specifications as Hot Standby or Hot Shutdown, as appropriate (plants have the option of maintaining the RCS at nonnal operating temperatures or at reduced temperatures).
 "Station blackout" means the complete loss of alternating current (ac) electric power to the essential and nonessential switchgear buses in a nuclear power plant (i.e., loss of offsite electric power system concurrent with turbine trip and unavailability of the onsite emergency ac power system). Station blackout does not include the loss of available ac power to buses fed by station batteries through inverters or by alternate ac sources as defined in this section, nor does it assume a concurrent single failure or design basis accident. At single unit sites, any emergency ac power source(s) in excess of the numbei required to meet minimum redundancy requirements (i.e., single failure) for safe shutdown (non-OBA) is assumed to be available and may be designated as an alternate power source(s) provided the applicable requirements are met. At multi-unit sites, where the combination of emergency ac power sources exceeds the minimum redundancy requirements for safe shutdown (non-OBA) of all units, the remaining emergency ac power sources may be used as alternate ac power sources provided they meet the applicable requirements. If these e criteria are not met, station blackout must be assumed on all the units.

3. A new §50.63 is added to read as follows:

 §50.63 Loss of all alternating current power.

(a) Requirements. (1) Each light-water-cooled nuclear power plant licensed to operate must be able to withstand for a specified duration and recover from a station blackout as defined in §50.2. The specified station blackout duration shall be based on the following factors:

[7590-01] (i) The redundancy of the onsite emergency ac power sources; (ii) The reliability of the onsite emergency ac power sources; (iii) The expected frequency of loss of offsite power; and (iv) The probable time needed to restore offsite power. (2) The reactor core and associated coolant, control, and protection systems, including station batteries and any other necessary support systems, must provide sufficient capacity and capability to ensure that the core is cooled and appropriate containment integrity is maintained in the event of a station blackout for the specified duration. The capability for coping with a station blackout of specified duration shall be determined by an appropriate coping analysis. Utilities are expected to have the baseline assumptions, analyses, and related information used in their coping evaluations available for NRC review. (b) Limitation of Scope. Paragraph (c) of this section does not apply to those plants licensed to operate prior to July 18, 1988 , if the capability to withstand station blackout was specifically addressed in the operating license proceeding and was explicitly approved by the NRC. (c) Implementation. (1) Information Submittal: For each light-water-cooled nuclear power plant licensed to operate on or before July 18, 1988 , the licensee shall submit the information defined below to the Director of the Office of Nuclear Reactor Regulation by April 14, 1989

• For each light-water-cooled nuclear power plant licensed to operate after the effective date of this amendment, the licensee shall submit the infonnation defined below to the Director by 270 days after the date of license issuance.

(i) A proposed station blackout duration to be used in detennining compliance with paragraph (a) of this section, including a justification for the selection based on the four factors identified in paragraph (a) of this section;

[7590-01] (ii) A description of the procedures that will be implemented for station blackout events for the duration determined in paragraph (c)(l)(i) of this section and for recovery therefrom; and (iii) A list of modifications to equipment and associated procedures, if any, necessary to meet the requirements of

• paragraph (a) of this section, for the specified station blackout duration determined in paragraph {c)(l)(i) of this section, and a proposed schedule for implementing the stated modifications.

(2) Alternate ac source: The alternate ac power source(s), as defined in §50.2, will constitute acceptable capability to withstand station blackout provided an analysis is performed which demonstrates that the plant has this capability from onset of the station blackout until the alternate ac source(s) and required shutdown equipment are started and lined up to operate. The time required for startup and alignment of the alternate ac power source(s) and this equipment shall be demonstrated by test. Alternate ac source(s) serving a multiple unit site where onsite emergency ac sources are not shared between units must have, as a minimum, the capacity and capability for coping with a station blackout in any of the units. At sites where onsite emergency ac sources are shared between units, the alternate ac source(s) must have the capacity and capability as required to ensure that all units can be brought to and maintained in safe shutdown (non-OBA) as defined in §50.2. If the alternate ac source(s) meets the above requirements and can be demonstrated by test to be available to power the shutdown buses within 10 minutes of the onset of station blackout, then no coping analysis is required. (3) Regulatory Assessment: After consideration of the infonnation submitted in accordance with paragraph (c)(l) of this section, the Director, Office of Nuclear Reactor Regulation, will notify the licensee of the Director's conclusions regarding the adequacy of the proposed specified station blackout duration, the proposed equipment modifications and procedures, and the proposed

[7590-01] schedule for implementing the procedures and modifications for compliance with paragraph (a) this section. (4) Implementation Schedule: For each light-water-cooled nuclear power plant licensed to operate on or before July 18, 1988 , the licensee shall, within 30 days of the notification provided in accordance with paragraph (c){3) of this section, submit to the Director of the Office of Nuclear Reactor Regulation a schedule co11111itment for implementing any equipment and associated procedure modifications necessary to meet the requirements of paragraph {a) of this section. This submittal must include an explanation of the schedule and a justification if the schedule does not provide for completion of the modifications within two years of the notification provided in accordance with paragraph (c)(3) of this section. A final schedule for implementing modifications necessary to comply with the requirements of paragraph (a) of this section will be established by the NRC staff in consultation and coordination with the affected licensee. Dated at Rockville, Maryland, this / 6 f! day of _J__,. . ~

                                                             -----:<,__ 1988.

Fo the Nuclear Regulatory Commission.

(7590-01] BACKFIT ANALYSIS Analysis and Detennination That the Rulemaking To Amend 10 CFR 50 Concerning Station Blackout Complies With the Backfit Rule 10 CFR 50.109 The Conmission's existing regulations establish requirements for the design and testing of onsite and offsite electrical power systems (10 CFR Part 50, Appendix A, General Design Criteria 17 and 18). However, as operating experi-ence has accumulated, the concern has arisen regarding the reliability of both the offsite and onsite emergency ac power systems. These systems provide power for various safety systems, including reactor core decay heat removal and con-tainment heat removal, which are essential for preserving the integrity of the reactor core and the containment building, respectively. In numerous instances emergency diesel generators have failed to start and run during tests conducted at operating plants. In addition, a number of operating plants have experienced a total loss of offsite electric power, and more such occurrences are expected. Existing regulations do not require explicitly that nuclear power plants be designed to withstand the loss of all ac power for any specified period. This issue has been studied by the staff as part of Unresolved Safety Issue (USI) A-44, 11 Station Blackout. 11 Both detenni ni sti c and probabilistic analyses were performed to determine the timing and consequences of various accident sequences and to identify the dominant factors affecting the likelihood of core melt accidents from station blackout. Although operational experience shows that the risk to public health and safety is not undue, these studies, which have evaluated plant design features and site-dependent factors in detail, show that blackout can be a significant contributor to the overall residual risk. Consequently, the Commission is amending its regulations to require that plants be capable of withstanding a total loss of ac power for a specified duration and to maintain reactor core cooling during that period. An analysis of the benefits and costs of implementing the station blackout rule is presented NUREG-1109, "Regulatory/Backfit Analysis for the Resolution

[7590-01] of Unresolved Safety Issue A-44, Station Blackout. The estimated benefit from implementing the station blackout rule is a reduction in the frequency of core damage per reactor-year due to station blackout and the associated risk of offsite radioactive releases. The risk reduction for 100 operating reactors is estimated to be 145,000 person-rem and supports the Co111nission's conclusion that §50.63 provides a substantial improvement in the level of public*health and safety protection. The cost for licensees to comply with the rule would vary depending on the existing capability of each plant to cope with a station blackout, as well as the specified station blackout duration for that plant. The costs would be primarily for licensees (1) to assess the plant's capability to cope with a station blackout, (2) to develop procedures, (3) to improve diesel generator reliability if the reliability falls below certain levels, and (4) to retrofit plants with additional co111µu11~nts or systems, as necessary, to meet the requirements. The estimated total cost for 100 operating reactors to comply with the resolu-tion of USI A-44 is about $60 million. The average cost per reactor would be around $600,000, ranging from $350,000, if only a station blackout assessment and procedures and training are necessary, to a maximum of about $4 million if substantial modifications are needed, including requalification of a diesel generator. The overall value-impact ratio, not including accident avoidance costs, is about 2,400 person-rem averted per million dollars. If the net cost, which includes the cost savings from accident avoidance (i.e., cleanup and repair of onsite damages and replacement power following an accident), were used, the overall value-impact ratio would improve significantly to about 6,100 person-rem averted per million dollars. These values, which exceed the $1,000/person-rem interim guidance provided by the Co111nission, support proceeding with the implementation of §50.63.

[7590-01] The preceding quantitative value-impact analysis was one of the factors considered in evaluating the rule, but other factors also played a part in the decision-making process. Probabilistic risk assessment (PRA) studies perfonned for this USI, as well as some plant-specific PRAs, have shown that station blackout can be a significant contributor to core melt frequency, and, with consideration of containment failure, station blackout events can represent an important contributor to reactor risk. In general, active systems required for containment heat removal are unavailable during station blackout. Therefore, the offsite risk is higher from a core melt resulting from a station blackout than it is from many other accident scenarios. Although there are licensing requirements and guidance directed at providing reliable offsite and onsite ac power, experience has shown that there are prac-tical limitations in ensuring the reliability of offsite and onsite emergency ac power systems. Potential vulnerabilities to coumon cause failures associated with design, operational, and environmental factors can affect ac power system reliability. For example, if potential conman cause failures of emergency diesel generators exist (e.g., in service-water or de power support systems), then the estimated core damage frequency from station blackout events can increase significantly. Also, even though recent data indicate that the average emergency diesel generator reliability has improved slightly since 1976, these data also show that diesel generator failure rates during unplanned demand (e.g., following a loss of offsite power) were higher than that during surveillance tests. The estimated frequency of core damage from station blackout events is directly proportional to the frequency of the initiating event. Estimates of station blackout frequencies for this USI were based OTI actual operational experience with credit given for trends showing a reduction in the frequency of losses of offsite power resulting from plant-centered events. This is assumed to be a realistic indicator of future perfonnance. An argument can be made that the future perfonnance will be better than the past. For example, when problems with the offsite power grid arise, they are fixed and, therefore, grid reli-

[7590-01] ability should improve. On the other hand, grid power f9ilures may become more frequent because fewer plants are being built, and more power is being trans-mitted among regions, thus placing greater stress on transmission lines. The factors discussed above support the determination ttmt additional defense-in-depth provided by the ability of a plant to cope with station blackout for a specific duration would provide substantial increase in tile overall protection of the public health and safety, and the direct and indirect costs of implemen-tation are justified in view of this increased protection. The Commission has considered how this backfit should be prioritized and scheduled in light of other regulatory activities ongoing at operating nuclear power plants. Station blackout warrants a high priority ranking based on both its status as an "unresolved safety issue" and the results and conclusions reached in resolving this issue. As noted in the implementation section of ttH! rule (§50.63(c)(4)), the Jchedule for equipment modification (if needed to mee~ the requirements of the rule) shall be established by the NRC staff in consultation and coordination with the licensee. Modifications that cannot be scheduled for completion within two years after NRC accepts the licensee's specified station blackout duration must be justified by the licensee. The NRC retains the authority to determine the schedules for modifications. In addition, some foreign countries, including France, Britain, Sweden, Germany and Belgium, have taken steps to reduce the risk from station blackout events. These steps include adding design features to enhance the capability of the plant to cope with a station blackout for a substantial period of time and/or adding redundant and diverse emergency ac power sources. Analysis of 50.109(c) Factors

1. Statement of the specific objectives that the backfit is desiyned to achieve

[7590-01] The NRC staff has completed a review and evaluation of infonnation developed since 1980 on Unresolved Safety Issue (USI) A-44, Station Blackout. As a result of these efforts, the NRC is amending 10 CFR Part 50 by adding a new§ 50.63, "Station Blackout." The objective of the station blackout rule is to reduce the risk of severe accidents associated with station blackout. Specifically, the rule requires all light-water-cooled nuclear power plants to be able to cope with a station blackout for a specified duration and to have procedures and training for such an event. A regulatory guide, to be issued along with the rule, provides an acceptable method to detennine the station blackout duration for each plant. The duration is to be detennined for each plant based on a comparison of the individual plant design with factors that have been identified as the main contributors to risk of core melt resulting from station bldckout. These factors are (1) the redundancy of onsite emergency ac power sources, (2) the reliability of onsite emergency ac power sources, (3) the frequency of loss of offsite power, and (4) the probable time needed to restore offsite power.

2. General description of the activity required by the licensee or applicant in order to complete the backfit In order to comply with the resolution of USI A-44, licensees will be required to --

0 Maintain the reliability of onsite emergency ac power sources at or above specified acceptable reliability levels. 0 Develop procedures and training to restore ac power using nearby power sources if the emergency ac power system and the nonnal offsite power sources are unavailable.

[7590-01] 0 Detennine the duration that the plant should be able to withstand a station blackout based on the factors specified in §50.63, "Station Blackout, 11 and Regulatory Guide 1.155, "Station Blackout. 11 0 If available, an alternate ac power source that meets specific criteria for independence and capacity can be used to cope with a station blackout. 0 Evaluate the plant's actual capability to withstand and recover from a station blackout. This evaluation includes:

      - Verifying the adequacy of station battery power, condensate storage tank capacity, and plant/instrument air for the station blackout duration.
      - Verifying the operability of equipment needed to operate during a station blackout and the recovery from the blackout for environmental conditions associated with total loss of ac power (i.e., loss of heating, ventilation, and air conditioning}.

0 Depending on the plant s existing capability to cope with a station 1 blackout, licensees may or may not need to backfit hardware modifica-tions (e.g., adding battery capacity) to comply with the rule. (See item 8 of this analysis for additional discussion.) Licensees will be required to have procedures and training to cope with and recover from a station blackout.

3. Potential change in the risk to the public from the accidental offsite release of radioactive material

[7590-01] Implementation of the station blackout rule will result in an estimated total risk reduction to the public ranging from 65,000 to 215,000 person-rem with a best estimate of about 145,000 person-rem.

4. Potential impact on radiological exposure of facility employees For 100 operating reactors, the estimated total reduction in occupational exposure resulting from reduced core damage frequencies and associated postaccident cleanup and repair activities is 1,500 person-rem. No significant increase in occupational exposure is expected from operation and maintenance activities associated with the rule. Equipment additions and modifications contemplated do not require work in and around the reactor coolant system and therefore are not expected to result in significant radiation exposure.

5. Installation and continuing costs associated with the backfit, including the cost of facility downtime or the cost of construction delay For 100 operating reactors, the total estimated cost associated with the station blackout rule ranges from $42 to $94 million with a best estimate of $60 million. This estimate breaks down as follows:

Estimated number of Estimated total cost (million dollars) Activity reactors Best Hi8h Low Assess plant's capability to 100 25 4 cope with station blackout Develop procedures and 100 10 15 5 training Improve diesel generator 10 2.5 4 1.5 reliability Requalify diesel generator 2 5.5 11 2.5 Install hardware to increase 27 17 24 13 plant capability to cope with station blackout Totals 60 94 42

[7590-01]

6. The potential safety impact of changes in plant or operational complexity, including the relationship to proposed and existing regulatory requirements The rule requiring plants to be able to cope with a station blackout should not add to plant or operational complexity. The station blackout rule is closely related to several NRC generic programs and proposed and ~xisting regulatory requirements as the following discussion indicates.

Generic Issue B-56, Diesel Generator Reliability The resolution of USI A-44 includes a regulatory guide on station blackout that specifies the following guidance on diesel generator reliability (Regulatory Guide 1.155, Sections Cl.1. and C.1.2): The minimum emergency diesel generator (EOG) reliability should be targeted at 0.95 per demand for each EOG for plants in emergency ac Groups A, B, and C and at 0.975 per demand for each EOG for plants in emergency ac Group O (see Table 2). These reliability levels will be considered minimum target reliabilities and each plant should have an EOG reliability program containing the principal elements, or their equivalent, outlined in Regulatory Position 1.2. Plants that select a target EOG reliability of 0.975 will use the higher level as the target in their EOG reliability programs. The reliable operation of onsite emergency ac power sources should be ensured by a reliability program designed to maintain and monitor the reliability level of each power source over time for assurance that the selected reliability levels are being achieved. An EDG reliability program would typically be composed of the following elements or activities (or their equivalent):

1. Individual EDG reliability target levels consistent with the plant category and coping duration selected from Table 2.

2. Surveillance testing and reliability monitoring programs designed to track EOG performance and to support maintenance activities.

3. A maintenance program that ensures that the target EOG reliability is being achieved and that provides a capability for failure analysis and root-cause investigations.

4. An information and data collection system that services the elements of the reliability progam and that monitors achieved EOG reliability levels against target values.

(7590-01]

5. Identified responsibilities for the major program elements and a management oversight program for reviewing reliability levels being achieved and ensuring that the program is functioning properly.

The resolution of B-56 will provide specific guidance for use by the staff or industry to review the adequacy of diesel generator reliability programs consistent with the resolution of USI A-44. Gener,c Issue 23, Reactor Coolant Pump Seal Failures Reactor coolant pump {RCP) seal integrity is necessary for maintaining primary system inventory during station blackout conditions. The estimates of core damage frequency for station blackout events for USI A-44 assumed that RCP seals would leak at a rate of 20 gallons per minute. Results of analyses perfonned for GI 23 will provide the information necessary to estimate RCP seal behavior during a station blackout. The industry coping analysis guidelines {NUMARC-8700) recognize the possibility of leakages exceeding an assumed 25 gpm per pump and incorporate the need to reevaluate the plant-specific coping analysis if the resolution of GI 23 identifies higher levels. USI A-45, Shutdown Decay Heat Removal Requirements The overall objective of USI A-45 is to evaluate the adequacy of current licensing design requirements to ensure that the nuclear power plants do not pose an unacceptable risk as a result of failure to remove shutdown decay heat. The study includes an assessment of alternative means of shutdown decay heat removal and of diverse "dedkated" systems for this purpose. Results will include proposed reco11111endations regarding the desirability of, and possible design requirements for, improvements in existing systems or an alternative dedicated decay heat removal method. The USI A-44 concern for maintaining adequate core cooling under station blackout conditions can be considered a subset of the overall A-45 issue. However, there are significant differences in scope between these two

(7590-01] issues. USI A-44 deals with the probability of loss of ac power, the capability to remove decay heat using systems that do not require ac power, and the ability to restore ac power in a timely manner. USI A-45 deals with the overall reliability of the decay heat removal function in tenns of response to transients, small-break loss-of-coolant accidents, and special emergencies such as fires, floods, seismic events, and sabotage. Although the recommendations that might result from the resolution of USI A-45 are not yet final, some could affect the station blackout capa-bility, while others would not. Recommendations that involve a new or improved decay heat removal system that is ac power dependent but that does not include its own dedicated ac power supply would have no effect on USI A-44. Recommendations that involve an additional ac-independent decay heat removal system would have a very modest effect on USI A-44. Reconmendations that involve an additional decay heat removal system with its own ac power supply would have a significant effect on USI A-44. Such a new additional system would receive the appropriate credit within the USI A-44 resolution by either changing the emergency ac power config-uration group or providing the ability to cope with a station blackout for an extended period of time. Well before plant modifications, if any, will be implemented to comply with the station blackout rule, it is anticipated that the proposed technical resolution of USI A-45 will be published for public comment. Those plants needing hardware modifications for station blackout could be reevaluated before any actual modifications are made so that any contemplated design changes resulting from the resolution of USI A-45 can be considered at the same time. Generic Issue A-30, Adequacy of Safety-Related DC Power Supply The analysis perfonned for USI A-44 assumed that a high level of de power system reliability would be maintained so that (1) de power system failures would not be a significant contributor to losses of all ac power and (2) should a station blackout occur, the probability of immediate de

[7590-01] power system failure would be low. Whereas Generic Issue A-30 focuses on enhancing battery reliability, the resolution of USI A-44 is aimed at ensuring adequate station battery capacity in the event of a station blackout of a specified duration. Therefore, these two issues are consistent and compatible. Fire Protection Program Section 50.48 of 10 CFR Part 50 states that each operating nuclear power plant must have a fire protection plan that satisfies GDC 3. The fire protection features required to satisfy GDC 3 are specified in Appendix R to 10 CFR Part 50. They include certain provisions regarding alternative and dedicated shutdown capability. To meet these provisions, some licensees have added, or plan to add, improved capability to restore power from uffsite sourc~s or onsite diesels for the shutdown system. A few plants have installed a safe shutdown facility for fire protection that includes a charging pump powered by its own independent ac power source. In the event of a station blackout, this system can provide makeup capability to the primary coolant system as well as reactor coolant pump seal cooling. This could be a significant benefit in terms of enhancing the ability of a plant to cope with a station blackout. Plants that have added equipment to achieve alternate safe shutdown in order to meet Appendix R requirements could take credit for that equipment, if available, for coping with a station blackout event.

7. The estimated resource burden on the NRC associated with the backfit and the availability of such resources The estimated total cost for NRC review of industry submittals required by the station blackout rule is $1.5 million based on submittals for 100 reactors and an estimated average of 175 person-hours per reactor.

8. The potential impact cf differences in facility type, design, or age on the relevancy and practicality of the backfit

[7590-01] The station blackout rule applies to all pressurized water reactors and boiling water reactors. However, in detennining an acceptable station blackout coping capability for each plant, differences in plant charac-teristics relating to ac power reliability (e.g., number of emergency diesel generators, the reliability of the offsite and onsite emergency ac power systems) could result in different acceptable coping capabilities. For example, plants with an already low risk from station blackout because of multiple, highly reliable ac power sources are required to withstand a station blackout for a relatively short period of time; and few, if any, hardware backfits would be required as a result of the rule. Plants with currently higher risk from station blackout are required to withstand somewhat longer duration blackouts; and, depending on their existing capability, may need some modifications to achieve the longer station blackout capability.

9. Whether the backfit is interim or final and, if interim, the justification for imposing the backfit on an interim basis The station blackout rule is the final resolution of USI A-44; it is not an interim measure.

• DOCKETU:

PACIFIC GAS usx ~ D ELECTRIC COMPANY

                              +   n s"B&E Jli.E25. 8.2 ;S.4c1sco, CA LIFORNIA 941 05 * ( 4 15 > 781- 4 2 11

• Twx 91 0-372 -6587

 .JAMES D . SHIFFER OFF ICE OF Sf. vtil T.4KY VICE PRE SIDENT NUCLE AR POWER GENER AT IO N DOCKETING & SER VI CL BRANC H June 18, 1986 PGandE Letter No.:           DCL-86-201 Secretary of the Commission U. S. Nuclear Regulatory Commission Washington, D.C. 20555 ATTN:           Docketing and Service Branch Re:         Docket No. 50-275, 0L-DPR-80 Docket No. 50-323, 0L-DPR-82 Diablo Canyon Units 1 and 2 Comments on Station Blackout Rulemaking Gentlemen:

In response to the public notice (51 F.R. 9829) requesting comments on the NRC's proposed rule on Station Blackout (SB0), PGandE submits the following comments. As a member of the Nuclear Utility Group on Station Blackout (NUGSB0) and the Nuclear Utility Management and Human Resources Committee (NUMARC), PGandE endorses the comments submitted by these organizations in response to the proposed rule. Those comments provide an in-depth critique of the proposed rule as it would affect the industry as a whole. Further, PGandE offers the following specific comments with respect to the impact of the proposed rule on its Diablo Canyon Power Plant. Based on the methods of analysis used in NUREG-1032, "Evaluation of Station Blackout Incidents at Nuclear Power Plants," the risk of SB0 at Diablo Canyon is exceedingly low. Each factor in the SB0 risk equation is low at Diablo Canyon because: (1) there is no history of severe or extremely severe weather at the California coastal location; (2) the overall grid stability is excellent, with ample hydro and pumped storage capacity, strong intersystem tielines, and sophisticated system protective schemes; (3) the switchyard arrangements are more than adequate, with five separate offsite power circuits capable of feeding the emergency safety features buses; and (4) the site emergency ac power sources require only two of five diesel generators for safe shutdown after a loss of offsite power.

                                  '!i ti I I

I I 11,1, f<<IClW 8GULAT~ COMMISSIOS DOCK~TING & SERVICE SECTION OFFICE OF ll-'E ~~CRETARY 0 u C'ION

PGandE Letter No. DCL-86-201 July 18, 1986 Page 2 Diablo Canyon also has a feature not mentioned in NUREG-1032 but contributing to SBO risk reduction, namely, a net load rejection capability for both main turbine-generator units. This allows a reactor to automatically power the internal ac power needs upon loss of offsite power for either or both units. With both units operating and a 50% success possibility for either unit, the chances of successful main turbine assumption of internal loads upon loss of offsite power is roughly 75%. Hence the risk calculable using NUREG-1032 methodology should be reduced by 75% for Diablo Canyon. The total risk for SBO-caused core damage should be below 10-6 per site year at Diablo Canyon. This level is acknowledged as acceptable in various NRC documents. The method of information gathering in the proposed Rule could be improved. Specifically, Oiablo Canyon can be determined to be at low risk by inspection of docketed information available to the NRC. The analysis that would be required by the proposed rulemaking appears laborious and expensive, and the results may not be useful. Calculating the ultimate coping capability as required by the proposed Rule is expensive and would be unnecessary when the simpler checklist method under development in a joint effort of the NRC Staff and the industry is completed. Finally, a few plants may indeed have unacceptable risk of core damage due to SBO. However, as can be gathered from NUREG-1109, 80% of the total U.S. risk from SBO comes from only 20% of the sites. The NRC Staff has not identified any of these plants as being unacceptable. A plant with low SBO risk, such as Diablo Canyon, should not be required to perform the same amount of analyses as a plant with higher risk. The NRC Staff can readily identify high risk plants and implement these requirements using existing regulatory mechanisms. In summary, PGandE believes that the proposed rulemaking will place unnecessary burdens upon PGandE and, hence, will result in unnecessary costs with little or no substantive improvement in plant safety at Diablo Canyon. ,e Kindly acknowledge receipt of this material on the enclosed copy of this letter and return it in the enclosed addressed envelope. Sincerely, cc: J. B. Martin Diablo Distribution 0931S/0046K/GHM/1572

NUCLEAR UTILITY GROUP ON STAT IO N BLACKOUT SUITE 700 "86 Jl 24 AB :26 1200 SEVENTEENTH STREET, N . W. WASHING TO N , O . C . 20036 TELEP HONE ( 202 ) 857-9833 OFFICE OF S[

• KL TARY DOCKETING SERVlr.r:

BRANCH July 21, 1986 Mr. Samuel J. Chilk Secretary U.S. Nuclear Regulatory Commission Washington, D.C. 20555

Dear Mr. Chilk:

On June 19, 1986 I informed you of the Nuclear Utility Group on Station Blackout's (NUGSBO) intent to comment on NUREG/CR-3840, "Cost Analysis for Potential Modifications to Enhance the Ability of a Nuclear Plant to Endure Station Blackout." Attached are NUGSBO's comments. I trust they will be taken into consideration in the ongoing Station Blackout rulernaking. Very truly yours,

                                                                        /~W/....f, Larry Kuncl V.P. Nuclear Nebraska Public Power District

I U,t, ~t~M. ltttUt /.. ""'n ,. IIV\.KETING & OFFICE F I OF E '(

• NUGSBO 86-001 DOC. Kt rrlflLy 1986 U5NRC
• '86 Jl 24 AS :27
• COMMENTS ON COST ESTIMATES IN THE REGULATORY ANALYSIS SUPPORTING THE PROPOSED STATION BLACKOUT RULE
• NUCLEAR UTILITY GROUP ON STATION BLACKOUT 1200 Seventeenth Street, NW Suite 700 Washington, D.C. 20036
• NUGSBO 86-001 DOCKETED JULy 1986 USNRC
• *"86 JI. 24 AB :27
• COl\lThllENTS ON.

COST ESTIMATES IN THE

• REGULATORY ANALYSIS
• SUPPORTING THE PROPOSED STATION BLACKOUT RULE
• NUCLEAR UTILITY GROUP ON STATION BLACKOUT 1200 Seventeenth Street, NW Suite 700 Washington, D.C. 20036

TABLE OF CONTENTS

• 1. Summary .~*********************************************************************************************** 1

2. General Comments on the Scope, Methods, and Accuracy of the Regulatory Analysis Cost Estimate ................................................................................. 2

• 3.

4* Comments on the Underlying Assumptions of the Proposed Design Solution's Cost Estimate .. ............................................................................. 8 Comments on the Methodology and Calculations of the NUREG/CR-384Q Proposed Design Solution's Cost Estimate *.................**..........*......................... 15

5. Comparative Cost Estimate Using Alternative Nuclear Power Plant Source Data and Assumptions .......*........****.....****.*..**............................. 21

• References Appendix
• NUGSBO 86-001
• I.

SUMMARY

• This report reviews the basis for cost estimates provided in the regulatory analysis supporting the Nuclear Regulatory Commission's (NRC) proposed station blackout rule, "Regulatory Analysis of Unresolved Safety Issue A-44, Station Blackout," NUREG-1109. This review concentrates on the estimates contained in "Cost Analysis for Potential Modifications to Enhance the Ability of a
• Nuclear Plant to Endure Station Blackout," NUREG/CR-3840, published in July 1984.
• Important findings resulting from this review are:

A. The design upon which the regulatory analysis cost estimate was based is not the product of an extensive, detailed coping analysis which the proposed rule and corresponding regulatory guides require. If the design had been so based the regulatory analysis' cost estimate would have been significantly higher. B. Traditional cost estimating methodologies and nuclear construction cost data were not used in preparing the regulatory analysis. Had they been used, a

• significant increase in the cost estimate for the given designs would have resulted.

C. Significant deficiencies exist in the NUREG/CR-3840 methodology and calculations which underestimate the costs of implementing individual backfit options by factors of up to twenty. These deficiencies include arithmetic

• mistakes, use of inappropriate productivity factors, and the absence of certain indirect costs.

 ***        D.       If the regulatory analysis made use of available and realistic cost data, the total cost impact of the proposed regulation would be over 200% higher than that indicated Based on the above findings it can be concluded that the regulatory analysis cost estimates are so understated as to call into question the validity of any conclusions made using these cost estimates.

1

• NUGSBO 86-001
• 2.

GENERAL COMMENT

S ON THE SCOPE, METHODS, AND ACCURACY OF THE REGULATORY ANALYSIS COST ESTIMATE This review approached the cost analysis from the following perspectives: (1) the maturity of the design upon which the cost estimate was performed; (2) the cost estimating methodology and data source for materials and labor costs; (3) the accuracy of the cost estimate in the regulatory analysis; and,

• (4) the application of the cost estimates to the regulatory analysis.

During this review, several general comments were identified and are summarized in this section. These comments are contextual and establish a framework within which the cost analysis should be viewed.

• Overall Validity Historically, regulatory analyses underestimate the final cost of backfits. This review did not
• encounter evidence that the cost analysis for station blackout significantly departs from this pattern.

It was found, for example, that the analysis significantly underestimates both labor and material

 **  costs in assessing the impact of station blackout modifications. Further, no contingency is provided for the costs of the inevitable criteria evolution and reinterpretations which occur after backfits are imposed. Correcting the underestimate for identifiable deficiencies and including the contingencies would considerably improve the validity of the estimated backfit costs.

Proposed Modifications Do Not Meet Requirements of the New Design Basis

• The proposed station blackout rule contemplates changes to the design basis of nuclear power plants based on a coping analysis which all licensees must perform. The draft regulatory guide accompanying the notice of rulemaking defines the scope of the coping analysis and directs
• licensees to consider plant conditions which may emerge both during a station blackout and in the 2
• NUGSBO 86-001
• subsequent recovery period when AC power is restored The minimum coping duration will be determined by the NRC following a review of the licensee's coping analysis. Although the draft regulatory guide discusses a 4 and 8-hour minimum required coping duration, it is possible that shorter periods would be acceptable.
• A major segment of the new design basis coping requirement affects the instrumentation and control function supporting systems, such as heating, ventilation, and air conditioning (HV AC).

These impacts represent a significant portion of the modifications that must be made to satisfy the

• design basis coping requirement. The problem with the NUREG/CR-3840 analysis is that it does
• not recognize the potential for modifications to these systems in any of the four basic options for which cost estimates are provided. Consequently, a significant basis for the cost estimate (ie., the list of backfit modifications from which a cost estimate is derived) is missing.

Major Conclusion Conflicts With Actual Cost Data

• The NUREG/CR-3840 methodology conflicts with two reliable sources of nuclear plant construction data: (1) "Handbook for Cost Estimating", NUREG/CR-3971 (Ball[1984]), and (2) the Energy Economic Oat~ Base. NUREG/CR-3971 has a goal "... to provide a consistent methodology and constant set of assumptions to assist the NRC user in preparing absolute as well
• as comparative cost estimates of generic requirements for light-water-reactor nuclear power." (l)

Although available at the time the regulatory analysis was performed, this methodology was not

 **  used in preparing NUREG/CR-3840. Moreover, certain productivity and other major assumptions found in NUREG/CR-3840 conflict with the "rules of thumb" found in this handbook and lead to sytematic underestimates. This point is discussed further in Sections 4 and 5 of this review.

(1) page ix, NUREG/CR-3971 I 3

• NUGSBO 86;001
• In addition to NUREG/CR-3971, another source of cost information readily available at the time the cost analysis was prepared is the EEDB,published by the Department of Energy. This program has been ongoing for at least 20 years and provides reliable cost data for newly constructed power plants, notes the existence of cost trends, and provides a methodology for use in estimating new
• nuclear construction costs. Again a key NUREG/CR-3840 finding - that hardware costs tend to dominate the cost estimate for the postulated station blackout backfit - directly conflicts with the established EEDB data. This point will also be discussed at greater length in Section 4 .
• Estimate Not Based On Engineering Data Another. serious difficulty encountered in this review was the lack of important details in the cost estimate. Specifically, while the backfit options are outlined and a detailed bill of materials is provided in NUREG/CR - 3840, no engineering and construction cost information is provided.

This information is essential to developing a complete cost estimate. A partial listing of the necessary information missing from NUREG/CR-3840 includes:

• (1) DESIGN STANDARDS USED - needed to determine quality assurance and equipment qualification requirements; (2) SYSTEM SPECIFICATIONS - needed to determine quality assurance and equipment qualification requirements;
• (3) PIPING AND CONDUIT ROUTING DIAGRAMS - needed to identify construction conflicts and task sequencing;

 **              (4)

(5) DESCRIPTIONS OF OPERATING FEATURES - provides a means for checking the design adequacy; SUPPORT SYSTEM REQUIREMENTS - needed to identify other modifications necessary; (6) SURVEILLANCE FEATURES - provides a means for determining design completeness; and,

*                (7)    PLANT TECHNICAL SPECIFICATION REQUIREMENTS - essential for operability considerations.

The practical impact of this missing information becomes evident in the ~due importance placed on

• equipment costs in the NUREG/CR-3840 analysis. Without a detailed engineering design to 4
• NUGSBO 86-001
• consider, it is not possible to accurately estimate the removal or relocation costs for existing plant equipment, a significant portion of a backfit project. The focus on the bill of materials also overlooks modifications required to meet existing regulatory standards, such as qualifying plant equipment to blackout conditions as is discussed in the draft regulatory guide.
• Another example of oversights created by not considering detailed engineering design is the cost of additional instrumentation and control capability used during the postulated blackout or to properly restore AC power. Such costs can be extensive. For example, simply automating the initiation of
• auxiliary feedwater flow at Millstone 2 and providing the necessary indication, a TMI Action Plan
• item not requiring any new pumps or large valves, cost over $700,000 to install. (Counsil [1981])

This single alteration to existing control circuitry is almost twice the cost estimated in NUREG/CR-3840 for the entire set of blackout modifications. NUREG/CR-3840 Methodology is Inconsistent and Inaccurate e NUREG/CR-3840 acknowledges that the work necessary to develop the proposed backfits and prepare the cost estimates was performed in a very short period of time (approximately two months). Consequently, it should not be surprising that there are a significant number of problems due to the limited time available to do the cost study. A partial list of the problems identified in this

• review includes:

(1) use of low unit costs for nuclear construction; (2) misquotes of cost data from the report's own references; (3) arithmetic errors; (4) unrealistically low material costs assumed for nonstandard materials; (5) inconsistent treatment of labor and material rates without explanation; and;

• (6) no inclusion of the cost of capital.

The NUREG/CR-3~40 analysis also employs a significant number of productivity assumptions 5

• NUGSBO 86-001
• which are either unsubstantiated or are contradicted by actual nuclear plant construction data.

Staff Costs Associated With Implementing a Station Blackout Rule are Low

• Certain costs presented in NUREG/CR-3840 analysis which are not attributable to hardware costs appear to have been underestimated. For example, the analysis indicates that the NRC's costs for reviewing a licensee's coping analysis and proposed plant modifications is only $7,000 per plant.
• This estimate is substantially less than that required to review the design of a Technical Support
• Center (TSC) and issue a safety evaluation report (SER). For example, an estimat~ of the TSC NRC Staff review costs provided in NUREG/CR-3971 (functions 35-38) is $93,800 per site.

Assuming that ~ cost per Staff manhour has not changed, then the cost of reviewing plant-wide

• modifications required to comply with a 4 or 8-hour station blackout design basis coping requirement is only 7.5% of that required to review a TSC design. The apparent differences in Staff workscope require clarification.
• Effects of Low Cost Estimate for Performing a Coping Analysis The proposed station blackout rule may require some licensees to modify their facilities. The total
• regulatory analysis cost estimate is a function of the number of facilities requiring modifications .

However , all plants are required to perform a coping analysis. Hence, any deficiencies in the cost

 **   estimate for performing a coping analysis bears a one-to-one correspondence to a deficiency in the t~tal regulatory analysis cost estimate. The draft regulatory guide accompanying the proposed rule provides some guidance as to the scope of the coping analysis. However, it does not provide sufficient detail to be able to provide a cost estimate for performing a coping analysis.

NUREG-1109 estimates the cost of performing a "coping analysis" to be in the range of $100,000 to $200,000. In contrast, industry estimates for such an analysis range from $500,000 to

*     $2,000,000. The industry estimates are based on a scope of coping analysis described in ANS-STD-58.12 and on previous experience where station blackout was considered as a part of a license application. The importance of these costs is significant. For example, a six-fold increase in the cost of a coping analysis to $750,000 raises the regulatory analysis cost estimate by over $40 million .

6

• NUGSBO 86-001
• Additional Considerations The proposed resolution described in the report places heavy reliance on reactor coolant pump (RCP) seal cooling proposal which is based on installing a new AC-independent charging pump.
• This proposal is directed at mitigating the effects of postulated seal failure due to loss of cooling in a station blackout. Although seal failure is important to station blackout risk resolution, resolution of this concern is in the advanced stage in the Generic Issue (GI) 23 task action plan. However, the GI-23 approach concentrates on improving the seal design and does not involve a new pump.
• Further, recent tests have shown that RCP seals do not fail in a manner requiring extensive makeup capability. In any event, at the time NUREG/CR-3840 was prepared, a new charging pump was

 ,* considered "preliminary" within the Generic Issue 23 task action plan. The only conclusion that can be drawn at this point is that the proposed resolutions discussed in NUREG/CR-3840 may no

• longer be necess~ .

7

• NUGSBO 86-001
• 3. COMMENTS ON THE UNDERLYING ASSUMPTIONS OF THE PROPOSED DESIGN SOLUTION'S COST ESTIMATE
• NUREG/CR-3840 considers four types of modifications designed to mitigate the effects of a station blackout These modifications seek to enhance the operability of AC-independent equipment in a station blackout The enhancements include:

(1) increasing battery capacity; (2) providing an AC-independent RCP seal injection capability; (3) increasing condensate storage tank (CST) capacity; and,

• (4) increasing instrument air system supply.

The NUREG/CR-3840 analysis determines the backfit costs of imposing each of these four modifications on two base case reactors (Arkansas Nuclear One, Unit 1, and Quad Cities, Unit 1 ).

• Six explicit assumptions are also used in performing this analysis which are the subject of this section's review. These NUREG/CR-3840 assumptions are summarized below:

(1) Modifications Purine Shutdown - Modifications will be made during normal plant operation or during scheduled shutdowns such that no replacement power

• (2) cost will be incurred; AL AR A - Occupational radiological exposure during installation and subsequent operation and maintenance of the added equipment will be minimal or zero and are not included as an increment of cost;
• (3) Socio-Economic Costs - Socio-economic impacts are considered as being minimal and, therefore, are not included as an increment of cost; (4) Seismic Qualification - All new equipment, structures, etc., needed to implement the proposed modifications will not be designed to
• (5) meet seismic requirements; EQ.Uipment Qualification - All new equipment and components installed outside containment will not require harsh environmental equipment qualificati~n; (6) Quality Assurance - To
• ensure reliability, all electrical components and 8
• NUGSBO 86-001
• equipment will be assumed to meet Class 1E requirements (other than seismic) and quality assurance requirements normally afforded safety grade components; and, (7) No Cost of Capital - No interest for the cost of capital during the period between initiation of design studies and incorporation of
• the capital improvements in the rate base is included in this estimate.
• Additional Costs of Implementation (Modification During Shutdown)

The assumption of no replacement power costs due to these backfits does not have substantial basis and is not accurate. The apparent motivation for neglecting these costs is that it simplifies the economic analysis. However, this assumption cannot be justified if replacement power

• costs are both significant and likely to occur in the course of installing station blackout equipment.

As further evidence of the magnitude of replacement power costs it can be readily shown from experience that power plant outage frequency durations have grown significantly in recent

• years due to regulatory change. While some of the growth is certainly attributable to equipment refurbishment and replacement, activities not directly caused by regulatory interventions (e.g.,

recirculation pipe replacements in BWRs and steam generator overhauls in PWRs), it is also clear that a significant portion of this growth is the result of plant backfits. Komanoff[1981]

• makes the argument that this phenomenon is likely to continue, given the Commission's policy of backfitting new requirements in response to the identification of new licensing issues .

Using the TMI Action Plan as a model, Komanoffs argument is as follows: Most reactors have sustained only brief (one to four week) shutdowns or outage extensions since TMI for minor equipment modifications, but NRC has committed itself in its post-TMI Action Plan to weigh major plant changes involving instrumentation, containment, and heat-removal systems... . TMI also takes some credit for NRC's recent establishment of compliance schedules for equipment installation (with attendant outages) to address long standing safety issues such as environmental qualification of

• electrical equipment and fire protection. In addition, the accident has directed NRC's attention away from reactor licensing toward reactor operations ... , making it less likely that licensees will be able to operate plants with equipment problems or shorten maintenance and repair outages (Komanoff [ 1981 ], page 252).
• These replacement impacts can be translated into significant replacement power costs as analysis 9
• NUGSBO 86-001
• performed by W.A. Buehring and J.P. Peerenboom [1982] at Argonne National Laboratory demonstrates. Buehring and Peerenboom argue that three types of consequences are likely to result from unplanned outages or scheduled outage extensions caused by regulatory interventions:

(1) increased costs of system generation;

• (2) increased demand for fossil fuels; and, (3) reduced electrical system reliability.
• To illustrate this point, six cases were examined involving shutdowns of various reactor types and geographical locations(3). The results show that undiscounted production costs increase significantly with reactor shutdown, yielding normalized increases ranging from $0.125 million/Mwe - yr to $0.33 million/Mwe-yr in the first year of outage(4). By way of example, for a
• nominal 700 Mwe unit in shutdown for 4 months to implement a backfit, these costs translate to between $29 million and $77 million in additional undisclosed production costs.

(3) The plants examined were Zion, Oconee, Prairie Island, Browns Ferry, Indian Point, and Three Mile Island. The latter two analyses were based on work independently performed by the

• General Accounting Office. .

(4) In this context, production costs conform with the accounting requirements of the Fuel, Operation and Maintenance Accounts of the Federal Energy Regulatory Commission Uniform System of Accounts . 10

• NUGSBO 86-001
• Buehring and Peerenboom also noted sensitivities to fuel mix, load growth, and seasonal variations. In worst case situations affecting grid reliability, economic losses due to unserved energy were comparable to or greater than anticipated production cost increases. At the bottom line, their conclusions directl,y contradict the assumptions of NUREG/CR-3840:
• The loss of benefits that result from nuclear plant shutdowns, whether temporary or permanent, are potentially si&nificant and warrant consideration in the re~Iatory decision rnakin11 process. (Buehring and Peerenboom [1981], page 133.) (Emphasis added)

The magnitude of these costs can be easily estimated. Van Kuiken, et al, reviewed FERC data to

• determine the time for planned outages to cover refueling, maintenance, and repair. Their analysis concluded 76 days per year was the typical planned outage duration. Since straight refueling outage time (i.e., the time to cool down, shuffle fuel and heat up) is typically eight weeks (56
• days), then the typical plant is §peDclini an additional 20 days per year on maintenance and modification. A significant fraction of these 20 days are clearly associated with changing regulatory requirements.
• The above studies all point to the conclusion that any backfit will contribute some amount of time to the annual outage duration. This contribution is due to resource, craft density and operational constraints which require the work to be performed during the planned outage or as a result of work displaced into the outage from when the plant was in normal power operation. Any amount of
• time yields significant costs which are excluded from the NUREG/CR-3840 analysis .

*

• ALARA Considerations The NUREG/CR-3840 assumption that radiological exposure need not be considered in estimating backfit costs conflicts with normal regulatory practice and ignores plant experience. In fact, guidelines provided several years ago by the NRC's Executive Director of Operations specify
• radiological safety consequences as a cost to be considered in backfit decision making (Dircks

[19821). This guidance also directs the Staff to include occupational exposure during plant installation, operation, and maintenance as part of the estimated consequences of the proposed backfit. The NUREG/CR-3840 assumptions violate these guidelines. 11

• NUGSBO 86-001
• Until recently, the industry experienced a significant growth in worker exposure contemporaneous with the increasing rate of change in regulatory requirements. This growth has peaked and declined somewhat in recent years. The trends are significant to support a correlation between radiological exposure and*the imposition of new regulatory requirements. Power plant experience further
• indicates that contractor personnel dominated the upward trend in occupational doses, constituting approximately 70% of the total dose received in the past (Silver and Mays [1983]). These workers are generally employed in support of refueling, special maintenance, and modification installation.

In particular, Brooks [1983] notes a consistent pattern from 1975 to 1982 in the allocation of dose

• to maintenance and inspection, constituting approximately 75-80% of the total dose received. In
• 1982 alone, the total annual collective dose represented 52,190 man-rem for 74 reactors. The maintenance and surveillance portion in that year was 81.2% of the total or over 42,000 man-rem.

This data underscores the magnitude of the ALARA concerns and its importance in the regulatory analysis. NUREG/CR-3840 should reflect this consideration in its analysis. Seismic Qualification

• The NUREG/CR-3840 analysis e~cludes station blackout mitigation equipment from seismic qualification requirements. While designed to reduce the impact of station blackout modifications, the problem with this assumption is that it also conflicts with the apparent direction being pursued by the Staff within the seismic qualification task action plan (USI A-46). The resolution of USI
• A-46 is expected to result in new qualification guidelines affecting equipment required for safe shutdown as well as equipment whose failure could result in adverse conditions which might impair shutdown functions. At the time NUREG/CR-3840 was prepared, this new guidance was under consideration by CRGR and should have been factored into the cost estimate.
• Presently, it is not clear how station blackout can be made a design basis accident without also imposing seismic qualification requirements on coping equipment. Even if station blackout mitigation equipment can be excluded from qualification requirements, existing regulations may
• require that some portions of the blackout backfit be qualified simply to ensure that seismic failure of this new equipment does not threaten existing safety systems. Examples of blackout mitigation structures, systems and components possibly requiring seismic qualification include the proposed new building housing the blackout systems discussed in Subtasks 1 and 2, steam piping required 12
• NUGSBO 86-001
• for the turbine-driven charging pump, and new DC switchgear. Qualifying such structures, systems, and components is an extremely expensive activity and missing from the NUREG/CR-3840 analysis.
• Environmental Qualification of Blackout Mitigation Equipment The NUREG/CR-3840 analysis assumes that new blackout mitigation equipment installed outside containment will be Class lE and conform with quality assurance standards without necessarily
• being qualified for harsh environments. This assumption conflicts with the requirements of 10 CFR 50.49 (Environmental Qualification Rule) if station blackout becomes a design basis event.
• Specifically, Section 2(b) of 10 CFR 50.49 defines safety related equipment affected by the rule as

    "... that relied upon to remain functional during and following design basis events." If 10 CFR

• 50.49 requirements are imposed on station blackout, licensees would have to embark on a comprehensive analytical program to determine environmental profiles (ie., temperature, pressure, and humidity) during a station blackout. Should these profiles qualify as mild environments, then no further action would be necessary under 10 CFR 50.49. However, if thermal or humidity
• conditions exceed the severity associated with a mild environment, then utilities would be required to take the next step and qualify blackou~ mitigation equipment by further testing or analysis.

In addition to questions concerning the applicability of 10 CFR 50.49, the "no EQ" assumption

• also conflicts with Section 3.1.4 of the draft station blackout regulatory guide which states:

All AC-independent decay heat removal systems and associated equipment needed to function during a station blackout should meet design and performance standards that ensure adequate reliability and operability in extreme environments, including hazards due to severe weather, that may be associated with a station blackout.

• Reconciling these conflicts between NUREG/CR-3840 and the proposed new requirements may not be easy. The cost of analysis and subsequent qualifications are extremely significant. Indeed, typical qualification costs for Class lE equipment are often as much as if not more than the costs of
• the equipment itself.

Any savings offered by "exempting" this equipment from the full documentation requirements of 10 CFR 50.49 may not be meaningful since the station blackout regulation still establishes the need

• for an analysis to demonstrate equipment operability during blackouts for 4 or 8 hours under the 13
• NUGSBO 86-001
• loss of both HVAC and most equipment cooling systems.

No Cost of Capital *

• In establishing utility revenue requirements, economic regulatory practice defines the rate base as the net book value of the plant considered used or useful in dispensing service l2.lYs. some reasonable allowance for working capital Improvements to existing commercial facilities increase
• the net book value of the plant. The capital improvement cost includes engineering, craft and white collar labor and materials. There is a substantial amount of time between start of design,
• procurement of materials, installation, and declaration of operation. It is not until after these costs have ~een expended and the resulting modifications have been put in service that the assets are
• cons~dered useful iI) dispensing service. In addition to this consideration, there is another period between the declaration of operation and the inclusion of the capital improvement in the rate base upon which consumer rates are set. During this time, the utility must finance the capital improvement at rates established by the market for alternative investments of comparable risk. The
• period of time involved before the modifications can be credited to the plant's book value would realistically be on the order of three years. This cost of capital represents a substantial sum which

      . should be included in the cost analysis.

14

• NUGSBO 86-001
• 4. COMMENTS ON THE METHODOLOGY AND CALCULATIONS OF THE NUREG/CR-3840 PROPOSED DESIGN SOLUTION'S COST ESTIMATE This section provides detailed comments on methodology and calculations found in the cost analysis of NUREG/CR-3840. Particular attention is provided to errors which affect the overall
• estimated costs and generally involve costing problems in unit rates and productivity. The problems
• manifest themselves in the data sources, a series of procedural errors, a general lack of data substantiation, and exclusion of additional costs normally incurred in plant backfits .

Data Sources The NUREG/CR-3840 cost analysis attempts to reflect standard engineering practice for estimating

• costs. Material, engineering, and labor costs are developed based on data provided by a standard cost manual for commercial construction, the Dodge Manual [1983], supplemented by telephone surveys where necessary. Adjustments to productivity and labor costs are then made to account for factors which may alter these numbers, such as geographical differences, the impact on
• productivity of work inside containment, engineering, quality assurance, and management overhead.

 **   To begin with, many of the more serious problems with the NUREG/CR-3840 cost estimate can be traced to the decision to use commercial construction cost data in estimating the magnitude of nuclear power plant backfit costs. While useful to its own purpose, the Dodge ~anual simply does not reflect the unique attributes of nuclear construction experience, ie. it systematically underestimates the material costs, and overestimates the productivity of nuclear crafts.

• For example, since material costs are likely to be relatively independent of industry application, it might seem reasonable to use the Dodge Manual as a basic source of data. However, unit costs for safety-related equipment consist of both the actual material costs .and the documentation costs necessary to demonstrate the component's ability to perform the intended function. The 15
• NUGSBO 86-001
• documentation costs for safety-related equipment can be as much as or more than the base material costs alone, due to the cost of qualifying equipment to requisite standards and the anticipated environment.
• Actual labor costs are also greater than the Dodge Manual projects for commercial construction. The NUREG/CR-3840 analysis recognizes this potential. To compensate for this situation, the analysis develops a composite work crew for cost estimation and introduces adjustments for nuclear-related work and geographical costs. Additional multipliers are also provided to account for overhead
• (25%) and quality control (25%). However, the combined effect of these multipliers does not
• accurately reflect the actual productivity experienced in installing safety-related materials. This point is best illustrated in Figure 4-1 which compares labor rates for selected items taken from the Dodge Manual against actual nuclear construction experience reported in the NRC/ERDA PWR Capital Cost Report, NUREG-0242 (UEC [1977]). The three civil/structural related items presented in the figure were selected as representative. The materials are typically used in normal construction applications and do not appear to have unique features which might *affect the cost of their installation. Yet, the average labor hours required to install these items in a new PWR is almost~
• times mater than reported by Dodge. Further, these rates are based on constructing a new PWR and do not involve work in radiation areas. In addition, the NUREG-0242 data in the Capital Cost Report is based on ID1 construction data while the Dodge Manual is more current. The growth rate for structural craft costs is substantial, as shown by Figure 4-2, suggesting that more current
• costs should be even greater.
• The inappropriateness of the Dodge Manual is also evident in comparing material costs (Figure 4-3). This figure illustrates recent material costs for the three items discussed. Costs are shown for the Dodge Manual, NUREG-0242 costs, and NUREG-0242 costs based on an average cost escalation of 30% since the 1977 NUREG-0242 study ( ~ UEC [1984], page 3-3). The differences between these costs and those projected by the Dodge Manual are striking. In fact, some of the material costs estimated by NUREG/CR-3840 in 1984 for station blackout modifications are .!ell than previously reported in 1977 for new PWR construction. Adjusting for
• the 30% average cost escalation experienced in commodity/equipment costs since 1977 only widens this gap.

While this discussion has concentrated on unit labor and material costs, there are also problems 16

• NUGSBO 86-001
• with the use of adjustment terms to account for overheads, as is done in the NUREG/CR-3840 analysis. This is best illustrated in Figure 4-4 which presents two overhead ratios for data presented in each of three reports: NUREG/CR-3840, NUREG/CR-3971, and the EEDB (DOE/NE 0051/2).

The first ratio presents engineering, quality assurance, and project management costs to total

• equipment and material costs. The second is the ratio of craft labor costs to total equipment and material costs. These comparisons clearly demonstrate the systematic underestimates of these overheads by factors of 4-5 or more.
• These comparisons are important because a key conclusion of the Dodge Manual analysis is that
• material costs dominate the overall cost estimate for the proposed backfits. The EEDB and NUREG/CR-3971 studies arrive at the opposite conclusion: labor costs dominate the overall cost estimate. This point is underscored by the Figure 4-4 comparisons. In addition to placing undue emphasis on materials in the cost estimates, an important implication of this result is that the Dodge Manual is inaccurate in estimating the cost of nuclear construction. Clearly, it is an inappropriate data source for estimating current backfit costs. A preferable approach to the cost analysis would have been to use the more traditional nuclear construction data sources cited in the previous
• comparisons. Had these sources been used, it is more likely that the cost estimate would have been more accurate as well as yielding significantly greater backfit costs.
• Procedural Errors in the NUREG/CR-3840 Analysis In addition to underestimating unit costs, a number of errors also plague the NUREG/CR-3840 analysis. While of somewhat lesser importance than the unit rate problems previously discussed,
• the mistakes distort the results and tend to further underestimate the backfit costs beyond that previously noted These errors include arithmetic mistakes, misquotations from the Dodge Manual, and inappropriate data extrapolations. The errors were identified by independent verification of spreadsheet analysis provided in the appendices to the NUREG .
• The importance of these procedural errors should not be overlooked because they compound the systematic errors arising out of the Dodge Manual methodology. For example, on the bottom of page A-4 of NUREG/CR-3840, 4 hours of resistance measurements, priced at $40/hour, are shown to cost $120 instead of $160. This error adds another 25% to the unit rate errors previously 17
• NUGSBO 86-001
• discussed. Since other costs are based on adjustments to this value, the 25% error is propagated throughout the overhead and geographical costs.

In other cases, the Dodge Manual is simply misquoted. Case in point, the unit cost for a 3/0 bare

• wire is given in the NUREG/CR-3840 analysis as $1.10/ft. (pages A-14, 18, 34, and 39, NUREG/CR-3840). In contrast, the Dodge Manual quotes this wire at $1. 703/ft (page 235).

Other problems arise from an inconsistent cost treatment for the same item. Backfill, for example,

• has a unit labor value of 0.25 hr./YD3 (page A-13, NUREG/CR-3840) and 1.00 hr.!YD3 (page
• B-16, NUREG/CR-3840).

In other locations, the inconsistencies affect the interpolation and extrapolation procedures applied to non-standard items. For example, NUREG/CR-3840 reports the cost of an enclosed 250 Vac circuit breaker rated at 1,000 amps to be $1,875, an almost linear extrapolation of costs based solely on the electrical current rating of a standard 70-100 amp, 250 Vac breaker. Contrary to this assumption, though, the cost of non-standard items increases non-linearly due to special design

• costs, setup charges, and additional material requirements.

Other difficulties are simply the result of using the wrong adjustment productivity factors. For example, normalized productivity and labor cost adjustment terms are used for implementing the

• backfits at ANO and Quad Cities despite the fact that the Dodge Manual provides adjustment rates for Arkansas and Illinois of 0.73 and 0.94, respectively .

 **   Lack of Substantiation One of the problems encountered in reviewing this analysis is that NUREG/CR-3840 provides only the first and last steps of the cost estimation process. In the first step, general assumptions are made (e.g., pump capacity, incremental battery size, etc.), and one-line diagrams illustrate where the

• major components are located. In the last step, a bill of materials is created which forms a basis for the material cost estimates. The missing elements are the detailed engineering and construction details used in preparing the labor cost and schedule_. This information must exist since a bill of materials was created. It would have been beneficial if it was provided with the report .

18

• NUGSBO 86-001
• Another problem area concerns the lack of substantiation for the labor productivity and cost assumptions used. Many of these assumptions do not compare well with actual experience, as previously discussed Of particular interest are the considerations which led to the use of composite
• construction teams and multipliers to account for work inside containment, geographical variations, and non-construction costs (e.g., engineering, overhead, etc.). Although experience has shown that these costs are more significant than craft-related costs alone, there may be some f01J.ndation for the values used which would permit their use in spite of the experience .
• Additional Costs Not Considered In NUREG/CR-3840 In addition to the problems identified in the costs estimated in NUREG/CR-3840, there appear to be other costs which should have been considered in the estimate and were not. These costs are in addition to that provided by using more realistic unit costs discussed and include:
• Equipment Removal Costs
• Plant modifications often require the removal of existing equipment These tasks affect both the cost and schedule of a backfit project and have not been considered in the NUREG/CR-3840 analysis .
• Delay and Disruption Costs Backfits generally introduce delay and disruption effects ("ripple effects") on work planning. These effects translate into lower labor productivity within crafts and engineering disciplines on other projects.
• Engineering Costs As discussed previously, engineering costs can represent a significant fraction of backfit costs, especially in backfit issues subject to evolving requirements. Using new plant construction
• costs as an example, DOE/NE-0051/2 (UEC [1984]) suggests 50% more engineering costs for a more recent PWR than used in the NUREG/CR-3840 analysis .

19

• NUGSBO 86-001
• Field Supervision and Indirect Labor Costs DOE/NE-0051/2 (UEC [1984]) suggests almost four times more field supervision, administrative, and overhead costs than assumed by the NUREG/CR-3840 analysis .
• Equipment Conformance to Codes and Standards Material costs for materials meeting requisite codes and standards are likely to be twice those assumed by the NUREG/CR-3840
• analysis based on DOE/NE-0021/2 (UEC [1984]), Counsil [1983], and common experience.

These costs should have been included in the NUREG/CR-3840 cost estimate. Failure to do so

• may result in a substantial underestimation of the costs.

The last section of this report discusses the impact of these additional costs on the proposed backfits using the Department of Energy (DOE) database (UEC [1983], [ 1984] ) and *more accurate productivity data. The revised cost estimate is then compared with previous backfit experience to

• provide a picture of what the impact of the staffs proposals might ultimately be.

20

I COMPARISON OF LABOR HOURS/UNIT l 0.8 0.7 NUREGICA-3840 HRS/FTsq 0.6 12! NUREG-0242 0.5 0.4 0.3 0.2 0.1 0.0 1 2 3 1 = Concrete external walls 2 = Concrete blocks 3 = Personnel doors FIGURE 4-1

250 - DEPARTMENT OF ENERGY ESTIMATES OF PWR LABOR COST INCREASES 200 150 - r -... GROWTH RATE El 1978 -82 100 - 1 1982 - 83

                 \..

50 0 STRUCTURAL MECHANICAL LABOR ENGINEERING FIELD CRAFT CRAFT SUPERVISION

      - 50                                         SKILL / CRAFT FIGURE 4 - 2

16 [ MATERIAL COST COMPARISON ) 14 NUREGICR-3840 12 E! NUREG-0242 ADJUSTED TO '83 10 ll NUREG-0242

$/FTsq  8 6

4 2 0 1 2 3 1 = CONCRETE EXTERNAL WALLS 2 = CONCRETE BLOCKS 3 = PERSONNEL DOORS FIGURE 4-3

COST ESTIMATE COMPARISON I NUREGICR-3840 181 NUREGICR-3971 120 Im DOEINE 0051/2 100 80 % 60 40 20 0 (ENG.QA,QC,PROJ.MAN) /( MATS & EQUIP) (CRAFT LABOR )/( MATS & EQUIP) FIGURE 4 - 4

• NUGSBO 86-001
• 5. COMPARATIVE COST ESTIMATES USING ALTERNATIVE NUCLEAR POWER PLANT SOURCE DATA AND ASSUMPTIONS This section examines the effects of individual assumptions on the cost estimate and provides a comparative analysis of the NUREG/CR-3840 analysis by correcting arithmetic errors and modifying several basic assumptions to reflect actual plant experience. These revisions are provided in order to identify the principal sensitivities inherent in the NUREG/CR-3840 analysis. The basic design solutions are the same as in NUREG/CR-3840 as well as are the equipment costs. The results of this analysis highlight the variation in the NUREG/CR-3840 analysis to data provided by the EEDB and NUREG/CR-3971. The results also demonstrate that the cost estimate supporting
• the regulatory analysis increases by a factor of 2-3 simply by applying industry cost data to the NUREG/CR-3840 methodology. Three factors that are considered in the analysis are: (1) correcting the arithmetic errors, (2) updating labor productivity assumptions, and (3) considering the cost of capital and replacement power.

Arithmetic Errors

• The first modification to the NUREG/CR-3840 analysis corrects the arithmetic errors and data

   .misquotations from the Dodge report. The details of each option are set forth in the appendix .

Table 5-1 provides the cost for each option estimated, without the arithmetic errors. The combined effect of these errors underestimates the cost of individual NUREG/CR-3840 modifications by up

• to 69% .

Labor Productivity Assumptions

• The NUREG/CR-3840 analysis was also corrected to reflect the labor hours/unit cited in the EEDB (UEC,"Phase VI Update" [1984 ]). Table 5-2 provides the results of this calculation for each option estimated. In this case, the cost underestimation for the various options provided in the
• NUREG/CR-3840 analysis is up to 115%.

21

• NUGSBO 86-001 e Labor Productivity (NUREG/CR-3971 Rule of Thumb)

An important "real-life" consideration in estimating nuclear construction costs is the sensitivity of productivity to rework, overcrowding, and work in radiation areas. Guidance provided in

• NUREG/CR-3971, "A Handbook for Cost Estimation," includes several rules of thumb concerning productivity factors for use in preparing such estimates. These rules suggest that a correction factor be applied to the labor productivity rate when estimating work to be performed in a radiation environment. This correction factor only affects the RCP seal cooling modifications.
• Even this singular case decreases overall labor productivity (ie., weighted productivity) to as low
• as 34%, thereby increasing all labor costs by 33%.

For all areas outside radiation areas, a loss in productivity results from rework and overcrowded

• work quarters. For example, to install piping or conduit cable trays, other piping which is in the way of newly added piping or conduit may have to be removed and replaced. NUREG/CR-3971 estimates that rework can add 10-35% to the labor cost of a modification, and overcrowding can result in an estimated 10% reduction in labor productivity. In contrast, the NUREG/CR-3840
• analysis assumes no labor for rework and overcrowding. A nominal 17% increase in labor costs may be used to account for the combined effect of rework and overcrowding. Using this value results in a productivity loss of 10% and can be shown to lead to additional labor costs of up to 190% of the NUREG/CR-3840 baseline estimate (See Table 5-3) .
• Engineering and QA/QC (NUREG/CR-3971 Rule of Thumb)
• Engineering, project management, QA/QC and other clerical costs are estimated by the NUREG/CR-3840 analysis to be 25% of the total craft labor and material costs. In the NUREG/CR-3971 sample estimate provided for a Technical Support Center, the ratio ofnoncraft labor costs (ie., engineering, project management, QA/QC drafting and clerical) to craft labor and
• materials cost exceeds 54%. This cost variance translates into a significant increase in noncraft labor costs by multiples up to 250%. Table 5-3A presents the underestimate percentage for all modifications.

22

• NUGSBO 86-001
• Cost of Capital The NUREG/CR-3840 analysis assumes no interest or other cost of capital charges for money from the start of the work until the inclusion of the capital improvement in the rate base. While such
• an assumption simplifies the analysis, it is also umealistic. Approximately two years is normally required from the initial engineering through inclusion in the rate base. Throughout this period capital charges accrue to the expenditures. The cost of capital varies but even if a conservative rate of 7% is assumed, the total dollar amount of the backfit would be sufficiently large to be
• considered. Thus the cost of capital, if included in the cost estimate, would increase the cost
• estimate by 14% or more.

Replacement Power Costs . In addition to other costs discussed above, replacement power costs can be a significant increment to the overall impact of a backfit. The NUREG/CR-3840 analysis assumes zero replacement power

• costs for all modifications reflecting the assumption that all work can be done while the plant is oper~ting. However, due to resource constraints, site craft density restrictions, and operating condition.s at the location of installation, the suggested modifications may not all be completed in a non-outage situation. Experience supports that back:fits and regulatory requirements contribute to
• plant outage time. For example, in NUREG/CR-4012, Van Kuiken, et al, calculate that the typical planned operation and maintenance outages are 72-76 days in duration. Since the "shell outage"
• which includes cooldown, fuel movement, and heatup is normally on the order of eight weeks (56 days) there exists ~20 days of scheduled operations and maintenance outage attributable to other work (e.g., modifications, maintenance, surveillance,plant betterment, etc.). To determine an upper bound on the costs that could be attributed to the associated replacement power costs, a calculation was performed assuming the crew labor hours are consumed in two shifts per day, six days per week for each option analyzed in NUREG/CR-3840. This calculation yields the number of crew days required to complete the task. The number of crew days can then be multiplied by
• $500,000 per day to obtain approximate replacement power costs. These results are summarized in Table 5-4.

The replacement power costs associated with the proposed modifications are significant. Typically 23

• NUGSBO 86-001
• these costs amount to 5-10 times the estimated capital costs of the backfit. Yet, they are not normally considered in the value-impact analysis.
• Comparative Analysis While each of the factors previously discussed will increase the size of the backfit cost estimates, it is not feasible to readily present all possible combinations of factors in a sensitivity analysis.

Consequently, the factors which could easily be corrected or otherwise substantiated in the existing cost database (e.g., the use of actual cost data from EEDB or NUREG/CR-3971) were isolated for analysis. These factors were limited to: (a) correcting arithmetic errors;

• (b)

(c) EEDB craft labor costs and productivities; substituting a ratio of Engineering, QA, QC, and Project Management costs to materials and equipment costs of 50%; (d) NUREG/CR-3971 productivity factors for rework and overcrowding; and,

• (e) considering the cost of capital of 7% for two years on the total cost.

The assumptions used in the comparative analysis are summarized in Table 5 - 5. It should be noted that replacement power costs were not considered in keeping with the basic structure of the

• NUREG/CR-3840 study.
• The results of the sensitivity analysis are provided in Figures 5-1 and 5-2 for the representative "base case" examples listed in NUREG/CR-3840. Figure 5-3 provides a category comparison for Option 2 in Figure 5-1 . As is clearly evident, the modified analysis results are, for the most part, two or more times lar~er than the NUREG/CR-3840 base case estimates. These results are significant in themselves. Further, when replacement power is also added, the modified case estimate could exceed the base case by up to a factor of 500.
• Extending these results to the regulatory analysis supporting the proposed station blackout rule (i.e., NUREG-1109) demonstrates the inefficiency of the proposed rule's requirements. For example, a factor of two difference between the NUREG/CR-3840 cost estimate and the modified analysis estimate applied to the number of affected plants listed in Table 6 of NUREG-1109 24
• NUGSBO 86-001
• doubles the costs for these hardware modifications. Doubling the total cost of the backfit without a comparable improvement in safety reduces the value-impact ratio for the proposed rule by 1/2.

Similarly, combining the revised estimate with other compliance costs only reduces the

• value-impact ratio further. For example, another potential modification to the Table 6 estimate involves increasing the best estimate cost for a coping analysis to a more realistic $750,000. Even this estimate is a number of factors less than the amount spent by some utilities in demonstrating the coping capability for their plants. If this revised cost for a coping analysis is combined with the
• revised hardware costs, the best estimate cost for the proposed regulation :rises by another $46
• million, further suppressing the value-impact ratio. Figure 5-4 illustrates this revised cost in comparison with the total backfit cost presented in Table 6 ofNUREG-1109 .

Conclusion This reanalysis demonstrates the cost elasticity in the NUREG/CR-3840 estimates for station

• blackout modifications. Simply using realistic productivity and unit rate data in the NUREG/CR-3840 cost estimate can dramatically shift the value-impact ratio for the proposed rule.

The implications highlight the need to reconsider the relevance of the NUREG/CR-3840 analysis to the resolution of the station blackout issue . 25

COMPARISON OF BASE COSTS PWR-INCREASED BATTERY CAPACITY 2500 2000 NUREGICR-3840

 $1000 1500                              El  COMPARATIVE COST ESTIMATE 1000 500 0

1 2 3 4 1 = Add 4 hour capacity 2 = Add 12 hour capacity 3 = Add battery charger/ DG 4 = Add battery charger/ Gas turbine generator FIGURE 5 - 1

COMPARISON OF BASE COSTS BWR-INCREASED BATTERY CAPACITY 1600 NUREGICR-3840 1400 1200 la COMPARATIVE COST ESTIMATE 1000

$1000    800 600 400 200 0

1 2 3 4 1=ADD 4 HOUR CAPACITY 2=ADD 12 HOUR CAPACITY 3=ADD BATTERY CHARGER/ DG 4=ADD BATTERY CHARGER/ GAS TURBINE GENERATOR FIGURE 5 - 2

COST COMPARISON PWR SUBTASK-1 OPTION-2 TOTAL Interest Charges Engineering & QNQC Costs Craft Labor & Materials Materials Cost NUREGICR-3840 (base case) Rework + Crowding Cost 181 SENSITIVITY ANALYSIS Productivity Adjustment Cost Craft Labor 0 500 1000 1500 2000 2500

                                                       $1 , 000's FIGURE 5 - 3

COMPARISON OF ESTIMATED COST FOR INDUSTRY TO COMPLY WITH THE PROPOSED RESOLUTION OF USI A-44 1 2 NUREG-1109 (best estimate) 3 loi COMPARATIVECOST ESTIMATE 4 5 TOTAL 0 10 20 30 40 50 60 70 80

                                                   $1 , 000 , 000 "s 1-ASSESS PLANTS ABILITY TO COPE WITH A SBO                            5-ADD AN AC-INDEPENDENT CHARGING 2-DEVELOP PROCEDURES AND TRAINING                                      PUMP (non-seismatic) CAPABLE OF 3-IMPROVE DIESEL GENERATOR RELIABILITY                                 DELIVERING 50 TO 100 GPM TO 4-INCREASE CAPABILITY TO COPE WITH A SBO                               REACTOR COOLANT PUMP SEALS (i.e., increase capability of station batteries condensate storage tank, and instrament air)

FIGURE 5 - 4

COMPARISON OF NUREG/CR-3840 BASE CASE VERSUS CORRECTED BASE CASE ($1000)

• SUBTABLE J Reactor Type Option NUREG/CR-3840 Corrected Variance 1 290.8 4045 39.2%

PWR 2 1145.8 1164.4 1.6% 3 285.1 288.6 1.2%

• 4 486.1 492.9 1.4%
• BWR SUBTABLE2 1

2 3 4 315.1 838.8 295.3 495.8 339.3 850.0 298.3 499.1 1.1% 1.3% 1.0% 0.6% 2A 870 823.2 -5.6% PWR 2B 829.2 8295 2C NIA NIA 2D 369.2 375.2 1.6%

• BWR 2A 2B 2C NIA NIA 180.8 NIA NIA 307.2 69.9%

2D 287.9 292.8 1.7% SUBTAllLEJ

,. Reactor Type Option PWR BWR 37.4 72.85 37.6 73.1 05%

0.4%

• SUBTABLE4 PWR 1 25.8 25.8 2 56.5 56.4
• BWR 1 2

20.9 39.2 20.9 39.2 TABLE 5-1

COMPARISON OF NUREG/CR-3840 CRAFr LABOR HOURS/UNIT TO UE/C DAT ABASE FOR BASE COST NUREG/CR-3840 UEC Variance

• SUBTABLE t Reactor Type Option 1 38.1 72.7 90.8%

PWR 2 105.4 200.5 19.0% 3 12.8 26.8 109.4% 4 10.9 23.5 115.6% 1 BWR 2 90.4 175.2 93.8% 3 15.1 30.2 100.0% 4 13.1 26.5 102.3%

• SUBTABLEi PWR 2A 2B 60.4 55.2 114.9 116.2 90.2%

105.0% 2C 2D 18.8 38.0 102.1%

• BWR 2A 2B 2C 18.0 35.7 98.3%

2D 12.0 23.4 95.0% SUBTABLEJ

• Reactor Type Option PWR BWR 3.7 7.4 100.0%

SUBTABLE:\ 1 2.3 4.6 100.0% PWR 2 5.1 10.2 100.0% 1 1.5 3.1 106.6%

• BWR 2 3.6 7.2 100.0%
• TABLES-2
• COMPARISON OF REWORK/OVERCROWDING LABOR COSTS TO BASE LABOR COSTS NUREG/CR-3840 Corrected Variance SllllIAllLE 1
• Reactor Type Option
• PWR BWR 1

2 3 4 1 2 38.1 105.4 12.8 10.9 90.4 29.3 80.8 10.8 9.5 70.6 76.9% 76.6% 84.4% 87.2% 78.1% 3 15.1 12.2 80.8% 4 13.1 26.5 102.3% SllllIAllLE 2 2A 60.4 51.7 85.6%

• PWR 2B 2C 2D 55.2 18.8 71.3 20.2 129.2%

107.4% 2A BWR 2B 2C 18.0 14.4 80.0%

• 2D 12.0 35.0 191.6%

SllllIA6LE J Reactor Type Option PWR

• BWR 3.7 3.0 81.1%

SllllIAllLE ~ PWR

• 1 2 5.1 4.1 80.4%

BWR 1 1.5 1.2 80.0% 2 3.6 2.9 80.5% Table 5

• 3
• NON-CRAFf HOUR COMPARISON NUREG/CR-3971 vs. NUREG/CR-3840
• Corrected Variance NUREG/CR-3840 SlmIABLE 1 Reactor Type Option
• PWR 1

2 43.9 176.1 164.4 466.6 274.5% 164.9% 3 36.7 89.9 144.9% 4 73.3 161.5 120.3% 1

• BWR 2 127.0 355.4 179.8%

3 38.3 94.8 155.4% 4 74.9 166.0 121.6% SlmIABLE2

• PWR 2A 2B 2C 134.7 128.2 330.0 372.0 144.9%

190.2% 2D 37.1 88.8 139.4% 2A

• BWR 2B 2C 2D 20.0 37.1 61.0 88.8 205.0%

139.0% SlmIABLEJ Reactor Type Option

• PWR BWR 7.2 18.6 158.3%
• SlmIABLE4 PWR 1 2

3.1 8.0 8.8 21.9 183.9% 173.8% BWR 1 3.3 6.1 84.8% 2 5.2 14.6 180.8%

                        *includes 50% of craft labor & materials

• TABLE5-3A
• UPPER BOUND REPLACEMENT POWER COSTS FOR EACH OPTION
• Option Crew Labor Hours Crew Days_*

Cost ($1,000,000) PWRSl0l 1380 14.4 7.2 PWRS102 3802 39.6 19.8

• PWRS103 476 4.96 2.5
• PWRS104 BWRS101 BWRS102 BWRS103 404 1206 3304 561 4.2 12.6 34.4 5.8 2.1 6.3 17.2 2.9 BWRS104 481 5.0 2.5 PWRS20A 2157 22.5 11.3
• PWRS20B PWRS20C 2004 20.9 10.4 698 7.3 3.6 BWRS20D 659 6.9 3.4
• BWRS20D PWRS301 442 12 4.6 0.13 2.3 0.06
• BWRS301 PWRS401 PWRS402 134 84 186 1.4 0.88 1.9 0.7 0.44 0.97 BWRS401 56 0.58 0.3 BWRS402 131 1.36 0.68
• *assumption:Two eight-hour shifts, six days a week

   + Replacement Power Costs = 500,000 per day

• TABLES-4

• NUREGICR-3840 Cost Estimate and Comparative Cost Estimate Assumptions NUREGICR-3840 Comparative Cost Estimate
• Assumptions Assumptions Uses Materials List Based upon ** Same Proposed Desi1n Solutions
• Uses (l)Dodge Manual Labor hrs/Unit Uses (l)EEDB 83 Adjusted Labor hrs/Unit (2)Dodge Manual (2)1977 UEC Adjusted at

                  $/Labor hr                                             5% per Yr. to '83 (8 Yrs. since '77 Report Based on

• '76 Wages)
• Applies Labor Productivity Factor or:

LOW BASE HIGH 75% 67% 50%

                                                         ** Same NO Factors for Rework                                 Applies Rework & Overcrowding or Overcrowdln&                                          Factor of:

• LOW BASE IDGH Rework 10% 35%

Crowding 5% 15%

• ** Adds Material Cost to Labor Cost ** Same Adjusts Material & Labor Cost for Geographical Differences
• as Follows:

LOW BASE HIGH

                                                          ** Same 85%       0%       115%

• Uses an Engineering &

QA Factor or 25% Uses an Engineering & QA Factor or 50% Adds Contractor Mark-up of 25% ** Same to Materials and Labor

    ** NO Amount for Interest                                Adds Interest of 7% or Total for Two Years

• TABLES-5
• NUGSBO 86-001 REFERENCES
• American Nuclear Society Standard ANS-58.12 "Criteria for Evaluation of Response Capability for Loss of all AC Power (Station Blackout) at Light Water Reactor Nuclear Power Plants" .

Consolidated Draft Revi,sion 2, May 1984.

• Ball, J.R., Cohen, S. Ziegler, E.J. ,."A Handbook for Cost Estimating," NUREG/CR-3971 ,

ANLJEES-TM-265, U.S. Nuclear Regulatory Commission, Washington DC, (October 1984). Brooks, B.G. "Occupational Radiation Exposure at Commercial Nuclear Power Reactors 1982", NUREG-0713, V4, U.S. Nuclear Regulatory Commission, Washington, DC, (December 1983).

• Buehring, W.A. and Peerenboom, J.P., "Loss of Benefits Resulting from Nuclear Power Plant Outages", NUREG/CR-3045, ANL/AA-28, Argonne National Laboratory, Argonne, IL, (March 1982).

Oark, R.A., Riordan, B.J., Thomas, W.R., and Wattington, B.E., "Cost Analysis for Potential Modifications to Enhance the Ability of a Nuclear Plant to Endure Station Blackout",

• NUREG/CR-3840, U.S. Nuclear Regulatory Commission, Washington, DC, (July 1984).

Cook, D.H., Greene, S.R., Harrington, R.M., Hodge, S.A., and Yue, D.D., "Station Blackout at Browns Ferry Unit One - Accident Sequence Analysis", NUREG/CR-2182, ORNLJNUREG/IM-455/Vl, Oak Ridge National Laboratory, Oak Ridge 1N, (November 1981) .

• Letter from W.G. Counsil (Northeast Utilities) to Dr. John Ahearne (NRC),

SUBJECT:

Capital Costs for Implementation of Action Plan Requirements, dated March 2, 1981. Letter from W.G. Counsil (Northeast Utilities) to J.R. Tourtellotte (Chairman, Nuclear Regulatory Reform Task Force), dated March 21, 1983 .

• Letter from T.J. Dente (BWR Owners Group) to D.G. Eisenhut (Director, Division of Licensing),

SUBJECT:

BWR Emergency Procedure Guidelines, Revision 2 (prepublication form) , BWROG-8219, dated June 1, 1982. Memorandum from W .J. Dircks to the Commissioners,

SUBJECT:

Revised Guidelines for

• Preparing Value - Impact Analyses, SECY 82-187, U.S. Nuclear Regulatory Commission, Washington, DC, dated May 7, 1982.

Memorandum from J.J. Jackson to V.S. Noonan,

SUBJECT:

Trip Report - Review of NRC Sponsored Reactor Coolant Pump Seal Testing and Proposed Westinghouse Owners Group Seal

• Test Program at AECL, U.S. Nuclear Regulatory Commission, Washington, DC, dated April 25, 1984.

Kolaczkowski, A.M. and Payne, A.C., "Station Blackout Accident Analyses (Part of NRC Task Action Plan A-44)", NUREG/CR-3226, SAND 82-2450, Sandia Laboratories, Albuquerque, NM, (May 1983) . R-1

• NUGSBO 86-001 Komanoff, C. "Power Plant Cost Escalation", Van Nostrand Reinhold Company, New York, (198 1). .
• Pereira, P.E. (editor), "1983 Dodge Manual for Building Construction Pricing and Scheduling",

McGraw-Hill Information Systems, Parsippany, NJ (1982).

• Rubin, A. M., "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44,Station
• Blackout", NUREG-1109, U.S. Nuclear Regulatory Commission, Washington DC, (January 1986)

Sandia Laboratories, "Interim Reliability Evaluation Program:Analysis of the Arkansas Nuclear One - Unit 1 Nuclear Power Plant, NUREG/CR-2787 Vl, SAND 82-0978, Sandia Laboratories, Albuquerque, NM, (August 1982). Silver, E.G., and Mays, G.T. "Nuclear Power Plant Operating Experience - 198 1", NUREG/CR-3430, ORNLJNSIC-215, Oak Ridge National Laboratory, Oak Ridge, TN, (December 1983). United Engineers and Constructors (UEC), "Capital Cost: Pressurized Water Reactor Plant.

• Commercial Electric Power Cost Studies. Volumes 1 and 2", NUREG-0242, U.S. Nuclear Regulatory Commission, Washington, DC , (June 1977).

UEC, "Phase VI Update (1983) Report for the Energy Economic Data Base Program EEDB-VI, OOE/NE-005/1, U.S. Department of Energy, Washington, DC, (September 1984).

• UEC, "Phase VII Update (1984) Report for the Energy Economic Data Base Program EEDB-VII",

OOE/NE-0051/2, U.S. Department of Energy, Washington, DC, (August 1985). VanKuiken, J.C., Buehring, W.A., Guziel, K.A., "Replacement Energy Costs for Nuclear Electricity Generating Units in the United States", NUREG/CR-4012, U.S. Nuclear Regulatory

• Commission, Washington, DC, (October, 1984).

R-2

• APPENDIX
• NUGSBO 86-001 SUBTASK 1- INCREASE BATTERY CAPACITY NUREG/CR-3840 Assumptions
• The NUREG/CR-3840 analysis assumes that most station batteries currently have _a four-hour capacity and considers three options which are evaluated for increasing that capacity. The options include: (1 ) shedding loads, (2) adding batteries, and (3) providing an AC-independent charger.
• Additional batteries are considered in multiples of existing capacity while charg~r specifications are based on existing equipment The PWR battery option consists of two subparts providing 4 and 12 hour expansions:
• (1) adding two 125v DC 1350 amp-hr batteries; and, (2) six 1350 amp-hr battery additions
• The BWR battery analysis also provides two options for 8- and 16-hour blackout durations: the first option consisting of one 125v DC 500 amp-hr battery, and a 250v DC 900 amp-hr battery; the second option contains three such batteries at both voltage levels.
• The third option, AC-independent chargers, has two alternatives: a turbine-driven charger, and a diesel-driven charger. However, unlike the battery options, both charger options require a new
• building to house and support the equipment additions.

It should be noted that load shedding was initially considered in the NUREG/CR-3840 analysis and was later discarded because it was " ... not within the scope of the analysis" (page 44, NUREG/CR-3840). Yet, Table 1 of NUREG/CR-3840 which lists the loads considered in the analysis and concludes that an 8-hour blackout would require 923 amp-hrs at 125v DC and 190

• amp-hrs -for each additional hour. The analysis further concludes that, even conservatively assuming a 60% battery derating, most reactors could, nevertheless, maintain DC power for at least 10 hours if load shedding is attempted.

The NUREG/CR-3840 analysis also provides an assessment of the relative merits of the traditional A-1

• NUGSBO 86-001 lead-acid battery and a lithium battery. Although largely untested in commercial nuclear power plants, the benefits of the lithium battery appear to reside in its higher density, reducing the
• requirements for additional space.
• Comments on NUREG/CR-3840 Assumptions The principal comments regarding the NUREG/CR-3840 assumptions affect:

(1) load shedding as an option;

• (2) battery qualifications; and,

 *        (3)      charger reliability.

The credibility of the options presented is detracted by the failure to consider load-shedding as a viable option for providing DC power to the identified loads. In fact, a good argument is made in the NUREG that for small enough loads (such as those listed in Table 1 in the report), load shedding alone can extend battery availability out to eight-hour station blackouts and longer without the need for additional batteries. For load shedding to work, it would have to occur early in a

• blackout, which may be accomplished by procedure. In this review, the Table 1 loads are recalculated using peak loads as an indicator of DC power requirements. The total DC power requirements considered are presented below for the loads specified in the NUREG/CR-3840 analysis:
• LOAD PEAK REQUIREMENTS
• Emergency illumination DC MOY Operation Instrumentation and Display Display Lighting EDG Field Flashing and Control Switchgear and Breaker Control 12.50 0.75 31.25 2.50 1.50 kw kw kw kw kw l2...5!l kw TOTAL 81.00 kw
• Adding power requirements for the emergency lubricating oil and hydrogen seal oil pumps increases the total power requirements to nearly 100 kw .

A-2

• NUGSBO 86-001
• Analysis was performed to test the impact of these requirements on several power plants. This analysis assumed a 60% battery derating to yield anticipated availability with the load shedding option. The results are summarized below:
• STATION TYPE VOLTAGE CAPACITY (UNIT/COMBINED)

CAPACITY (HRS DURATION ) Indian Point PWR 125 1320/1584 11.3 Comanche Peak PWR 125 1950/2340 16.7 e Ginna PWR 125 1050/1260 9.0 Peach Bottom BWR 250/125 1520/1824 13.6 Brunswick BWR 250/125 1200'1440 10.7 As a brief review of this table suggests, load shedding appears t~ offer the requisite DC power for

• the blackout durations under consideration. On the basis of this cursory review, it is not clear why load shedding was not considered in NUREG/CR-3840.

While this brief analysis does present the benefits of load shedding, this option may not be viable in

• all cases. For example, there may be additional loads necessary to support plant operations during a blackout which were not considered in the NUREG/CR-3840 inventory of DC requirements.

HVAC could be one function which might have to be provided by. DC power in a blackout, especially for BWR drywells and small rooms containing important plant equipment (e.g., turbine driven auxilliary feedwater pumps or BWR core cooling equipment). Similarly, the NUREG/CR-3840 assumption of only 0. 75 kw for .all DC MOY operations appears to be low and

 '*    additional power might also be required. The nature of these loads would depend on the particular shutdown scenario envisioned.

In addition to load shedding, other battery sources already onsite should also be considered. It may be recalled that only Class lE batteries are acceptable under the draft revisions of the regulatory

• guide. However, if neither equipment or seismic qualification is necessary, taking credit for capacity present in non-Class lE batteries offers a significant enhancement to the DC power capabilities assumed to be available.
• The new batteries considered by the NUREG/CR-3840 analysis do.not credit load shedding and A-3

NUGSBO 86-001 The new batteries considered by the NUREG/CR-3840 analysis do not credit load shedding and

• require a significant amount of room. Since this feature may introduce a design constraint, an incentive was created for the lithium battery option. But introducing these batteries may not be as simple as the NUREG suggests. Lithium batteries are highly reactive and have experienced many instances of fires and explosions. Consequently, there is a need to ensure that their introduction
• into a power plant environment does not pose additional safety and design requirements which are not addressed in the cost analysis. Since no engineering details are available concerning this option, additional comments are reserved on the technical issues of this option.
• The NUREG/CR-3840 analysis of the use of AC-chargers without load shedding also lacks
• sufficient detail for meaningful comment. For example, an argument is presented for selecting a gas-turbine driver over a diesel driver, primarily on the basis of the potential for common cause failure. This goal would be appropriate if the reliability of gas-turbines were very high and comparable to diesels. However, that assumption may not be valid. Further, gas turbines are not readily available in sizes less than 500 kw and may not be a viable option. Finally, while gas turbines may not be susceptible to common cause failures associated with diesel maintenance, fuel and electrical related failures are not excluded simply by using a different driver. In short, the
• NUREG fails to make a strong case for the gas-turbine charger over the diesel on the basis of the engineering information presented.

A-4

• NUGSBO 86-001 SUBTASK 2 - RCP SEAL COOLING NUREG/CR-3840 Assumptions
• An important contribution to station blackout risk is the potential for losing primary coolant inventory due to reactor coolant pump seal failure. To mitigate this event, the NUREG/CR-3840 analysis considers an AC-independent charging pump in one of four configurations:

 *        (a)

(b) (c) Steam-driven turbine generator providing power to an existing motor-driven pump; Steam-driven turbine directly driving a charging pump; Dedicated diesel coupled directly to a charging pump; and, (d) Dedicated diesel-generator providing power to an existing motor-driven charging pump.

• The charging pump capacity desired is in the range of 50-100 gpm at full system pressure.

Since cooling to the seal injection heat exchanger would also be lost in a station blackout, two additional DC motor-driven valves would be needed to provide.the 1 gpm bleedoff from each RCP .

• This leakage would vent (flash) directly to containment For steam-driven turbine options, the additional steam line would tap inside the MSIVs with isolation provided by a normally closed
• DC-powered motor operated valve. Turbine exhaust steam would be vented to the atmosphere.
• Otherwise, existing seal injection piping would be used.

For the BWR, high pressure makeup capability is substantial with both High Pressure Coolant Injection (HPCI/HPCS) and Reactor Core Isolation Cooling Systems (RCIC) independent of AC

• power for operation. Therefore, any additional injection deemed necessary would be directed at maintaining seal integrity. For the boiler, seal injection is not considered to be as significant a concern as the PWR. Additional turbines are ruled out for the BWR due to high costs associated with new drywell pipe penetrations. For this reason, diesel driven equipment is proposed along
• with a new non-seismic building. If a new pump is added, this pump would be in parallel with the A-5
• NUGSBO 86-001 AC-driven control rod drive (CRD) pumps .

Comments on NUREG/CR-3840 Assumptions The principal comment on the NUREG/CR-3840 assumptions concerns the need for high pressure

• makeup in the light of the current understanding of RCP seal failure potential. This potential is presently viewed as having an extremely remote likelihood for large-scale leaks. Consequently, the need for a RCP seal makeup system may no longer be necessary. The importance of this issue is
• underscored by the fact that the seal injection option represents the bulk (i.e., over 70%) of the
• backfit costs estimated in NUREG/CR-3840. It should be noted,in contrast, that NUREG-1109 assumes that this modification will not be required (see Table 6, NUREG-1109).

An additional problem with the charging pump concept is that which it does not do. A high pressure seal cooling system such as the one proposed does not address the other cooling issues raised by the proposed station blackout mle. These cooling loads include:

• (1) PWR containment cooling (fans and spray) to ensure containment integrity and equipment availability; (2) PWR primary sampling for boron concentration; and, (3) Auxiliary and Reactor Building cooling to ensure equipment operability.
• Normally, charging pump cooling is another support feature required for long-term operability.

1 This cooling is generally provided by component cooling water either directly or indirectly. Moreover, injection may not be the sole means of ensuring sear integrity since, for many plants,

• injection merely complements cooling provided to the heat exchanger sU1Tounding the seals. The system providing the cooling is, again, component cooling water, a system not available in a station blackout Thus, the proposed solutions may not be complete.

A-6

• NUGSBO 86-001 SUBTASK 3 - INCREASE CONDENSATE STORAGE TANK
• CAPACITY
• NUREG/CR-3840 Assumptions The NUREG/CR-3840 proposal for expanding CST capacity is directed at maintaining an ultimate heat sink during a station blackout For PWRs, this would be provided by a portable diesel-driven
• fire pump making up to the CST. The turbine-driven auxiliary feedwater pump would provide the
• necessary feedwater for the steam generators. The water chemistry constraints of BWRs apparently preclude this approach from consideration in NUREG/CR-3840 and a more permanent CST makeup arrangement is proposed. For the BWRs, water would be transferred from the
• condenser hotwell to the CST using an AC-independent pump. Pump capacity would match usage rate at the approximate time (4 hours) when nominal CST level approaches exhaustion (calculated at 5 hours by Cook and Greene [1981])*

Comments on NUREG/CR-3840 Assumptions

• As with the previous discussion provided for Subtasks 1 and 2, the particular problems with the

_ NUREG/CR-3840 proposals concern that which is not included but necessary to completely

  ,. consider decay heat removal in a station blackout. For BWRs, this would include the impact of an
     . eight-hour station blackout on suppression pool stability and drywell thermal limits, and

• consequential effects on system operation.

Decay heat removal for a BWR in a station blackout is provided through safety relief valves (SRVs) releasing energy to the suppression pool. Initially, makeup is provided by HPCI/HPCS and RCIC.

• Later, RCIC alone may be used intermittently to control reactor pressure vessel (RPV) water level.

Since no AC power is available, the Residual Heat Removal System (RHR) is not available in the torus cooling mode. Hence, the suppression pool serves as the ultimate heat sink and, itself, has no means of being cooled. A-7

• NUGSBO 86-001 To determine suppression pool temperature in a blackout, a pool heatup analysis was performed
• for a typical BWR 4 with a Mark I containment . The model plant is assumed to have a 400-day operating history and a thermal power rating of 3293-Mwt ~ssuming an initial pool temperature of 95°F, it was found that heat rejection to the torus quickly raises the temperature to 124°F within one hour of blackout initiation based on BTP ASB 9.2 assumptions for decay heat generation rates.
• Due to the loss of normal heat sink, RPV pressure in this scenario was found to hover near the lowest SRV setpoint, which is nominally 1020 psig. Under these conditions of high pool temperature and RPV pressure, General Electric Emergency Procedure Guidelines call for emergency depressurization to cut-in RHR in order to maintain pool stability. (T. Dente [1982]).
• But, since the RHR system is not available in a station blackout, the success of this
• depressurization depends on restoring power prior to losing the RCIC turbines on low steam pressure. Analysis performed at Oak Ridge by Cook et.al. [1981] confirms the need for RPV depressurization within the first hour but for another reason. In their analysis early depressurization
• is necessary to maintain drywell air temperatures below the drywell design temperature of 281 °F.

While early depressurization limits peak drywell temperatures, it does not prevent the rapid temperature rise to 250-300°F early in the blackout nor does it substantially reduce temperatures through the event Cook found that high drywell temperature does not threaten containment structural integrity and short-term survivability of some equipment However, several problems still remain. High drywell

• temperatures can seriously affect the accuracy of RPV water level instruments by altering the differential pressure sensed between the reference leg and the vessel itself. (Dente [1982]). Should
• instrument accuracy erode to the point of yielding false high water level indications, it is possible to receive a high level trip of HPCI, precluding the ability to provide makeup. Overall equipment
• operability in this environment would also be open to question.

This brief review underscores the complexity of designing a station blackout coping capability. Emergency procedures, design limits and potential system malfunctions all need to be carefully

• considered as part of proposed modification. These key factors are apparently missing from the NUREG/CR-3840 analysis and detract from its validity .

A-8

• NUGSBO 86-001 SUBTASK 4 - INCREASE INSTRUMENT AIR SUPPLY NUREG/CR-3840 Assumptions
• The NUREG/CR-3840 analysis assumes that most compressed air systems are capable of providing sufficient air for 4 hours of operation under blackout conditions. This operability is based on the assumption that receivers constitute 10% of system volume and are normally
• maintained at 100 psig pressure. At 80 psig, air system loads are assumed in the NUREG/CR-3840
• analysis to lose their function. These assumptions are used to justify a proposed doubling of present air capacity to achieve 8-hours of operation, and tripling the air capacity for 16 hours operation.

Since the ANO-1 capacity was not available when NUREG/CR-3840 was prepared, the P\YR option is based on an average of the Ft. Calhoun and Palisades instrument and service air

• capacities. These volumes are reported at 3100 scf and 1000 scf, respectively. The average used in the analysis is 200 scf. The proposed option is based on adding a number of standard 2000 psig bottles yielding 250 scf each. At this rate, between 15 and 45 bottles are deemed necessary.
• Receiver capacity for the Quad Cities plant is calculated at 1100 scf, thereby requiring between 2200 scf and 6600 scf for 8 and 16 hour blackouts.

  ** Comments on NUREG/CR-3840 Assumptions With respect to the particular circumstances of this option, compressed air systems are normally

,* used to operate safe shutdown equipment in a blackout. For PWRs, air systems are normally required for many auxiliary feedwater pump controllers and valves, charging and letdown flow control, atmospheric dump valve operation and containment isolation. Under blackout conditions, BWR air systems are not as essential for control of safe shutdown systems . A-9

NUGSBO 86-001 It is difficult to comment on the generic analysis of air capacity requirements without having a

• detailed engineering design to review since the need for air would depend on the blackout scenario, design details, and any assumptions concerning equipment failure. A stuck-open valve or break in a line due to seismic event, environment conditions, or random failures could significantly alter air requirements. In addition, the assumption that operability is lost below 80 psig air pressure is very I
• conservative.

I The question of whether air is essential depends on the shutdown scenario. As a general rule, if air is not available to operate the valves, manual operation is always possible. Thus, the significance I ** of these options is not clear. . A-10

I - Westinghouse Water Reactor DOCKETED USNRC

                                                                 *Nuclear Technology Division Electric Corporation   Divisions       16 Jl 11 AlO :36        Box 355 Pittsburgh Pennsylvania 15230 July l, 1986 NS- NRC 3146 Secretary of the Conmission U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Attn:     Docketing and Service Branch

Subject:

Conments on Proposed Station Blackout Rule, 51 F.R. 9829, March 31, 1986 Westinghouse has reviewed the NRC 1 s Proposed Rule regarding Station Blackout and has compiled these comments into the attached document. It is our opinion that issuance of this rule should be reconsidered. The basis for this conclusion is explained in item numbers 2 and 3 of the attachment. Very truly yours, WESTINGHOUSE ELECTRIC CORPORATION

                                               ~1               *Manager Nuclear Safety Department BJJ/wh/128ln Attachment

11,1. NUCLF R ~GlJL,UnP.y COM MIS510N DOCKETlr-./G & ERVICE SECTION OFFI E OF T r SECRET ARY OF THE COMMISS ION

NS- NRC 3146 July 1, 1986 bee: E. P. Rahe, Jr., 4 R. 6. Saint- Paul, Brussels J. Cobian, Madrid J.M. Moore, PC 3 600 M. Beaumont, Bethesda J. Ga 11 agher, 4 J. Moore, 5

WESTINGHOUSE COMMENTS REGARDING PROPOSED RULE TO RESOLVE USI A- 44, STATION BLACKOUT

1. Clarification to definition of Station Blackout is needed:

The proposed definition of "Station Blackout 11 1s as follows: 111 Station Blackout* means the complete loss of alternating current

• (AC) electric power to the essential and nonessential switchgear buses in a nuclear power plant, (i.e. loss of the offsite electric power system concurrent with turbine trip and unavailability of the onside emergency AC power system)".

The following clarification should be added by revising the parenthetical statement starting with 11 i.e. 11 as follows: {i.e. loss of the offsite electric power system concurrent with turbine trip and unavailability of the onsite emergency AC power system source, but not including the energy source stored in the station batteries from which AC can be made through inverters for instrumentation use during the station blackout.)

• This clarification is needed so that credit can be taken for AC powered instrumentation and control during a station blackout scenario .

Recognition for such credit has been generally accepted and understood in prior generic and plant specific discussions with the Staff and ACRS in connection with Unresolved Safety Issue A- 44 reviews and studies. Unless this clarification is made in the rule, a literal interpretation of the proposed rule seems to preclude credit for AC powered instrumentation and control that operate from battery power inverters.

2. Steps have already been taken in the industry to improve capability to mitigate Station Blackout events:

Nuclear power plants have already addressed the need to prepare plant operations for station blackout events. Industry initiatives have been implemented to improve the capability of plant operators to mitigate the consequences of the event and recover the plant following restoration of AC power. The initiatives have addressed the eight requirements of Generic Letter 81 - 04, Emergency Procedures and Training for Station Blackout Events, dated February 25, 1981. Requirements included in this generic letter are:

a. The actions necessary and equipment available to maintain the reactor coolant inventory and heat removal with only DC power available, including consideration of the unavailability of auxiliary systems such as ventilation and component cooling.

b. The estimated time available to restore AC power and its basis.

c. The actions for restoring offsite AC power in the event of a loss of the grid.

d. The actions for restoring offsite AC power when its loss is due to postulated onsite equipment failures.

e. The actions necessary to restore emergency onsite AC power. The actions required to restart diesel generators should include consideration of loading sequence and the unavailability of AC power.

f. Consideration of the availability of emergency lighting, and any actions required to provide such lighting, in equipment areas where operator or maintenance actions may be necessary.

g. Precaut1ons to prevent equ1pment damage during the return to normal operating conditions following restoration of AC power. For example, the limitations and operating sequence requirements which must be followed to restart the reactor coolant pumps following an extended loss of seal injection water should be considered in the recovery procedures.

Generic Letter 81 - 04 also required utility annual requalification training programs to consider the emergency procedures and include simulator exercises involving the postulated loss of all AC power with decay heat removal being accomplished by natural circulation and the steam- driven auxiliary feedwater system for PWR plants. To assist ut1lities in this effort and to establish generic guidance for station blackout operat1ng procedures, the Westinghouse Owners Group included the subject event in their effort to develop improved guidance for emergency transients as required in Item I.C.l of NUREG- 0737. The resulting Emergency Response Guidelines include guidelines ECA- 0.0 Loss of all ac Power, ECA- 0.l Loss of all ac Power Recovery Without SI Required, and ECA- 0.2 Loss of all ac Power Recovery With SI Required. These guidelines address the generic aspects of Generic Letter 81 - 04. As part of their plant specific procedures development and training activities, ut111t1es have used this generic guidance as well as addressing the plant- specific aspects of Generic Letter 81 - 04 to improve their capability to mitigate station blackout events. Because of these efforts, and efforts taken in other areas such as diesel generator reliabiity, we believe that issuance of the blackout rule presents little additional benefit.

3. The proposed rule does not satisfy the Backfit Rule criteria:

Analysis was performed by the NRC Staff to determine that the proposed rulemaking complies with the criterion specified in the Backfit Rule 10CFR50.109. The majority of this analysis and conclusions are documented in NUREG/CR 1109. The Staff's favorable conclusion was based on a quantitative Benefit/Cost assessment of one person - rem eliminated per five hundred dollars expended and a qualitative assessment that industry PRAs have highlighted Blackout as a dominant r i sk contributor and the Europeans have already responded to this issue with hardware modifications . Westinghouse has reviewed this analysis and revised what we feel are numerous inappropriate assumptions. The results of the quantitative assessment lead to a conclusion that the proposed rule is highly cost ineffective and that the qualitative justifications provided are wrong. These are discussed in more detail below. Thus, the proposed Blackout Rule would not satisfy the Backfit Rule criteria as currently written. The quantitative value/impact analysis is highly conservative. The primary sources of this conservatism are due to the use of antiquated source terms and the minimal use of plant specific recovery actions following a blackout. These two factors alone would result in a change of the value- impact ratio by factors to 20 to 1000. Thus, instead of one person- rem/500$, the expected range of values is one person- rem/10000$ to 50000$. The Rule does not meet the 1 person- rem/1000$ value- impact ratio typically used for these types of assessments. The consequence values in NUREG- 1109 do not reflect current knowledge of fission product source term behavior for severe accidents. This is a general trend within the NRC due to the long time delay in implementing the conclusions and recommendations of NUREG- 0956. The NRC has advised that a plan for implementation of new source term information into the

regulatory process is underway, with an expected completion date of February, 1987. In light of the far reaching impact of the regu l atory policy developed for the resolution of USI A-44, the implementation should be achieved before further regulatory action is undertaken. This conclusion is supported by the fact that the consequences used in the value- impact analysis of NUREG- 1109 would be reduced by a factor of 10 to 100 rendering any of the alternatives infeasible. The factor of 10 to 100 is obtained by comparing the results of recent analyses of the station blackout to that given for the NUREG-1109 source term. These are prov i ded in Table 1. Table 2 provides a ratio of the relative releases compared to the 1109 value. All of these recent analyses indicate that the release fractions for the station blackout, for the fission product release categories which are dominant contributors to offsite consequences (iodines, cesiums and telluriums), are overestimated for the 1109 report by at least a factor of 10 to 100. Therefore, the consequences used in the value- impact analysis should be reduced by a factor of 10 to 100. The second area of major conservatism in the quant i tative assessment is the minimal use of recovery actions for AC/DC power and for condensate. In most PRAs, AC recovery is separated into two categories: on- site emergency AC restoration and non- emergency AC feeds to the emergency AC buses. NUREG- 1032 provides an assessment that the probability of diesel generator recovery within a two to eight hour post blackout t i meframe is low, potentially a factor of two reduction in the frequency of core melt and consequences. Recovery of numerous types of common cause failures of diesels is not a complicated maneuver. In particular for failures such as diesel trips on overspeed and compressed air start failures, operating plant data bears this conclusion out. Since the value- impact assessment results would be significantly affected by a factor of 2 change in this assumption, this effect should have been explicitly modeled. The second area of recovery is the use of non- emergency power sources. Again, NUREG- 1032 provides a comparison of the Staff assessment model and the

impact of recovery. For Indian Point (Figure A.9), the use of nearby gas - turbine generators results in a factor of 10 reduction in the frequency of offsite power exceeding durations. Although a nearby gas - turbine generator is not standard at most plants, nearly all plants have alternative power sources which could accomplish the same function. These alternatives include crosstieing emergency AC sources from sister units and linking up startup diesels from non- nuclear stations and substations nearby. The NUREG-1109 assessment is not clear as to what extent credit is given for these types of actions on a generic basis. Our view is that these effects have not been credited. The use of these types of recoveries would results in a factor of 2 to 10 reduction in the average plant's ability to cope with a station blackout. Condensate alternatives for long duration blackouts appears to be ignored as part of the 11 09 assessment. Most PWRs have alternative modes of obtaining condensate. These range from spool pieces which can link the RWST, city water or river water to the auxiliary feedwater system to merely refilling the condensate tanks via fire trucks. The procedures for these kinds of operations have been included in many plants. Therefore, these should be incorporated prior to making generic conclusions. The second part of the staff ' s assessment was qualitative. The statement was made that prior PRAs have shown that Station Blackout sequences are dominant core melt and risk contributors. This is true, however, it must be recognized that the studies quoted all represent PRA modeling techniques which are several years old. In particular, for Westinghouse NSSS designs, Blackout issues have dominated due to the assumption that reactor coolant pump seals will fail catastrophically within 30 minutes of blackout. Due to this assumption, there was little time for the operating staff to institute recoveries prior to core melt and even to containment failure. Thus, there was minimal credit in the model for recovery. current experimental and analytical work demonstrate that this is a highly

conservative assumption. Little or no degradation occurs for seals utilizing the new materials. The result is that reanalysis of these prior PRAs with the most recent data would significantly alter the risk dominance of these sequences. Thus, previous analysis is not a good indicator of industry impact when the state of knowledge is changing rapidly as in seal performance. Another qualitative argument is that European countries have taken steps to reduce the risk of station blackout through design modifications. The assumption that these changes were added only to mitigate blackout conditions is an oversimplification. The basis for these changes were two fold. The first was an over- reaction to perceived pump seal performance. As stated previously, experimental data justifies that this was an over- reaction and that the risk reduction due to these modifications is significantly less than previously expected. The second reason for these designs was political. The expectation for threats of war and/or terrorism has been much higher in these countries than in the USA. This resulted in governmental intervention by providing additional design criteria. These criteria do not directly relate to the issue of USI A44 resolution and should not be used as a basis for justifying a consensus of world nuclear opinion on blackout fixes . In conclusion, we feel that the backfit analysis significantly overestimates the risks from Station Blackout events and that the net benefits of the proposed modifications are significantly overvalued. The conservatism is such that the conclusions reached would be reversed under more realistic assumptions. In our analysis, the value- impact is highly unfavorable for implementation of this rule in its present form and therefore issuance of this rule should be reconsidered.

TABLE 1 F1ssion Product Releases for the TMLB 1 Event (all values are fraction of core inventory) Noble Study Gas Iodine Ces1um Tellurium Barium Ruthenium Lanthanum SST- 1 1.0 4. 5-1 6. 7- 1 6.4- 1 7.0-2 5.0- 2 9.0-3 RSS 9 .0- 1 7 .0-1 5.0- 1 3.0-1 6.0- 2 2.0-2 4.0- 3 ( PWR 2) NUREG- 1.0 7.8- 3 3.9-4 8.5- 2 1.8- 2 3.3- 6 8.1 - 5

• 0956 NUREG/ 1.0 CR 4540 IDC0R l .0 2.4- 2 1.7- 3 2.4- 2 1.7-3 3.0- 2 2.0- 5 2.6- 3

                                                     <l.0- 5 2.3- 3
                                                                <l.0-5 3.9- 4
                                                                           <l.0- 5 Zion IDC0R 1.0       5.1 - 4     6.4- 4    2.6-5      <l.0-5     <l.0- 5    <l.0- 5 Sequoyah j

TABLE 2 Fission Product Releases for the TMLB Event (all values are fraction SST- 1 source term) Noble Study Gas Iodine Cesium Tellurium Barium Ruthenium Lanthanum SST- 1 1.0 1.0 1.0 1. 0 1.0 1.0 1.0 RSS 0.90 1. 55 0.75 0.47 0.86 0.40 0.44 ( PWR 2) NUREG- 1.0 0.017 0.00058 0.13 0.26 0.000066 0.44

• 0956 NUREG/ 1.0 CR 4540 IOCOR 1.0 0.053 0.0038 0.036 0.0025 0.047 0.000031 0.037

                                                 <0.00014 0.046
                                                             <0.00020 0.043
                                                                        <0.00011 Zion IDCOR 1 .0     0.0011      0.00096   0.000040   <0.00014    <0.00020   <0.00011 Sequoyah

  '                                            ~lr     ~PR-~

(_5/P~ 9?~t:/ July 2, 1986 COMMENTS O OHIO I IZENS FR RESPON IBLE ENE GV, INC. <1 0 E1 ) ON THE PROPOSED STATION 8 A KOUT RULE C 1 FR 9829, MARCAb,it* 1986), THE DRAFT REGUL ORV GUIDE ON ST TON B CKOUT, ~1i'11JR~D NUREG-110 . OCR!:: supper t re~u

• remen t tt,a nuc1' acQp Tb reactor licensees consid~r t ation blac ut and upg ad r ac i l i t:ies to with tand u ~ on ev n t. Howeve ,

with ommissioner Asselst .i n _ that he propo ~ O~i:i h .,- 3=22 ree ru1tJoc;f/" G.~Cl-~t,~RY rar enoug

• Commission r A s e lstine not& , othar n g ~tRWe~

attacking the tQtion blackout threat ra mo e ogg es iv ly, equiring their plants to cope wit ~ a station blac out ror 20 ours, ereas the proposed rule would require consideration of a station blackcu duration of eith e r o 8 hou s. OC E sees no imp e diment to emulating the r e asonable practices of th French in mor edly educing the threat of static blackou t

• he commission speciFicallY re q uested comments on the implementation or the BackFit Rule in this proposed rulemaking.

It i- OCRE ' s position that app1

• cation or t e sackfit Rue t rulemakings promulgated pursuant to th Administrativ p r o edu e Act, 5 USC 553, is lainly ill eg al. Se e conn c icut Li ht and r "'o. v. NRC, 673 F."d 525 (1982):

During the rulemaking process, the NRC i force to justify the need fo regulat1on involving backfitting b Y vi t ua of the rulemaking process. A further finding on the m act or the regulations on the public safety would be otiose. 6 3 F.2d at 536. Furthermore, t e Commission i not mpowered o cons1der o>>ts I to licensees under th Atomic Energy Act and t e Ene y I Reorganization A~t; rath e r, the Commission is authorized wit le safeguarding the public h e alth and safety, This was a ticu ated as long ago a 1961 in the pow r nt co. decision of the Sup eme court, specific comments on th propo ed rule , draft regulato ry guid , on NUREG-1109 follow.

1. e sche uling provisions in th proposed rul CFR 50.63(c ) (1), (d) ( 1) and ( d) ( 2 )) are ar ~oo gene ous, f'ter-the d lays inh rent i n t e rulemaking process lice se s o~e given 9 month in which to conduct the r quired analy es ; then ,

oft e r an indet rmi ate p iod or time ror sto,r r view o f the ubmitt d analyses, th licensee are giv n 6 mor months in ic to propose a ch e dule Fo implem nti g any modi ica ions round to be n cessary, There is no date certain fo completing these modifications; final schedules need on y be agreed upon b y JUL -8\986 Acknowledged by aerd.

• __ _ _ _ _J

                    . q 0.S. NUCLEAR RFGIJI AT'1RY COM ISSI l'I DOCKETING & SEP' ICE SECTI N OFFICE OF Tr       C~ET RY OF THE COM ISSION Docume-nt Statistics

age 2 th tarr and licensee. hile th~ e pectat

• on
• s voic d that he modification s ould b* com l ted w*t in 2 years o f t e Starr's no *
• cation, th's 1s not a r quir ment.

The station blackou t i$sue has alread y been unresolv d for too long, e issue has be op n i c e th e APP al Board's 1980 St, Lucie d Ci ion (ALAB-6 3), here is no logical r ason to ndanger t e ublic safety by prolong *n g its resolution and im?lemen otion, especiall y sine much or the analytical work has alr ady b n don

• n FSAR ona yses, SEP analyses, or PRAs, OCRE would propose the ro110win9 schedule as being far more reasonabl and protective or the public safety, The analyses of 10 FR 50. 3 (c) should be submitted within 90 day after the err ctive date o the rule, Th Stoff s ould be given 90 days in ich to vi~w the licensees ' submittals ond to give the noHf'ication o 10 CFR 50.63(c) ("). S ct:ion 50, 3(d ) should be modi ied to stot~ that, for r actors licensed on or before th effec ive date of the rule, all modifications nee ssary to meet GOC 17 shal be comple ed within one y ea or th Staff's notif'icatio ** i,e., within 18 mont s or t e e rective date of the rule. It also hould e clearly stated that reactors licensed o op rate of er the errective dot of' th rule hall co ply with the rul* b*fore the issuance or an operating license at any power l!!Vel, Any licensee purporting an inobi ity to meet this sch dule should b* requir d to seek a icense a endme t ands ould be required ode onstrate (1) that complianc with the -chedule i impossible and <2> that the public health and af ty is not ndanger*d by e schedule rope ed by the licensee.

2, her gulatory guide should equire consideration o seismic events as a caus for loss of off'site power and as a actor in the availability and accessibility or offsit quiprn nt < e section 3,1, factor 5, P. 7). C ramie insulators ar gene rally assumed to rail at 0,2 9. see, ,9., NUREG-0 9, suppleme nt 3, P. 15-15 <GESSAR I SER ).

3. Factor 1 of s ction 3.1 or the regulatory guides ould, in addition to the assumption that the plant 1s op~ at1n9 o full power i diately before t e pos ulated station blackout ,

include the as umption that h& core contain the maximum ri sion product inventory, J u t as 10 CFR 50 AP endix K intends that the heat sourc s during a OCA should b_ maximiz d ro ECCS analyses, t e regulatory guide should enumerate the sources of hat to be conside ed du ing th ;tation blackout t nsient and en ure tho' t ey are maximized. Factor 2 o section 3.1 or er ulatory uide s o d

page 3 include the errec or cable v lta~a drops on DC ~~uipm nt operability wh n batte r v olt g alone c ith out the r1oat c argers i to be relied upo The abi l1 y or motors to s art under t ese conditions (inc luding cab l volta ge raps ass oc iat d w th st rting cur ents ) ShO Ul d e pec iallY be evaluated.

5. Th e term *condensate storage tan ca acity* used on P. 3 o NUREG-1109 ( and in the r egu atory gu i de ) s ould ot mean he maximum capacity or the a nk, but the minimu amount o* ater in t e ank as equir d by plant procedures or technical sp cificotions. Ir such procedu es o lY requi t a t t e tank b 1/ 3 full, i t is no reasonable to assume t a t the tonk will be Fill d to 100% or its capac ty ror s ation blackout valuations.

6. NUREG- 09 o P. 20 states t a t c edit was ive n r ri e pot ction improvements, i.e., the use of fire p otectio n equipment in maintaining core coolant 1nventory. such credit may b inappropriate since fir e protection syste s us ow water , and plant operate~ may be elucta t to int educe t i s di ty water into the reactor v essel.

Since el y, Susan . Hiatt OCR E Representative 82 5 Munson Road le Men~or , OH 40 0 ( 216 ) 255-3158

   ~~
  ~~                                                                                  f ,

Duquesne Ug,t Telephone (412) 393-6000 Nuclear Group P.O. Box4

• June 19, 1986 DOC:KETEO USNRC Shippingport, PA 15077-0004

                                                                                 ~ JI. -3 AU :.11 Secretary of the Commission U. S. Nuclear Regulatory Commission Washington, DC 20555 Attn: Docketing and Service Branch

Reference:

Beaver Valley Power Station, Unit No. 1 & Unit No. 2 Docket Nos. 50-334 and 50-412, License No . DPR-66 Comments on Proposed Rulemaking, Station Blackout 51FR9829 Gentlemen: On March 21, 1986 the Nuclear Regu l atory Commission (NRC) published a proposed rule on station blackout in response to Unresolved Safety Issue (US!) A-44. The proposed rule would require commercial nuc l ear power plants to withstand and recover from a station blackout for a specified duration acceptable to the Staff. Accompanying the proposed ru l e is a backfit analysis which relies primarily on the regulatory analysis found in NUREG-1109. Reference is also made in the proposed rule to a draft regulatory guide which was also published March 1986. The nuclear utility industry has been actively engaged in following the reso l ut i on of this issue. The Nuclear Utility Group on Station Blackout (NUGSBO) and the Nuclear Utility Management and Resources Committee (NUMARC) provided comments on the proposed rule which we endorse .

• In addition, we are providing separate comments on the proposed ru l e and draft regulatory guide (Attachment I).

If there are any questions or clarifications, please contact my staff. J. J. Carey

a

---

--~'

          *s "

Beaver Valley Power Station, Unit No. 1 & Unit No. 2 Docket Nos. 50-334 and 50-412 Comments on Proposed Rulemaking, Station Blackout 51FR9829 Page 2 cc: Mr. W. M. Troskoski, Resident Inspector U. S. Nuclear Regulatory Commission Beaver Valley Power Station Shippingport, PA 15077 U. S. Nuclear Regulatory Commission c/o Document Management Branch Washington, DC 20555 Director, Safety Evaluation & Control Virginia Electric & Power Company

• P. 0. Box 26666 One James River Plaza Richmond, VA 23261 Mr. Peter S. Tam, Project Manager Director of Nuclear Reactor Regulation U. S. Nuclear Regulatory Commission Division of PWR Licensing - A Washington, DC 20555

ATTACHMENT I COMMENTS I. Generic Rulemaking not Necessary The Station Blackout issue should be handled on a plant-specific basis and does not need to be resolved by generic rulemaking. Each plant has unique probabilities for a loss-of-power event based on transmission system, location of plant, and on-site power systems. II. Regulatory Focus on Coping Capability Regulatory focus on coping capability will divert licensee resources from the real issue (i.e. improve AC power reliability). Each plant can be requested to provide a coping capability analysis under 10CFR50.54(f). This application would be more expedient and efficient versus the proposed rulemaking process. III. RCP Seal Leakage Proposed rule does not take into consideration the different RCP seal designs and provide no credit for plants with back-up floating seals which limit leakage in the event of seal failure. Recent information from the WOG (Westinghouse Owner's Group) indicates that RCP seals do not represent a significant potential for large inventory losses as earlier assumed. IV. Safety Grade Equipment Any plant modifications or additional equipment required to meet the proposed rule should not be specified safety grade. For equipment which is to be manually started and placed in service for testing or in the event of a loss of power condition there is no necessity for specifying safey grade since adequate reliability can be obtained through normal surveillance testing and the proper maintenance of commerical power plant equipment. The cost difference in safety grade vs. commerical grade modifications is significant and must be emphasized. To require safety grade equipment to meet the proposed rule would significantly increase the cost estimates derived in the staff's backfit analysis.

v. Diesel Generator Reliability In the past, different approaches have been taken by licensees to calculate reliability resulting in uncomparable and inconsistent data.

Therefore, to ensure future consistency, we recommend a common method for calculating Diesel Generator reliability and the method should be highlighted in the Regulatory Guide, Table 1. One such method is the INPO/NSAC reporting instructions. This method provides a good assessment of diesel generator reliability by considering all start demands.

TENNESSEE VALLEY AUTHORITY

                                                                        *L   T l'WM8£R CHATTANOOGA, TENNESSEE 37401 5N 157B Lookout Pl ace JPC>ffD BUJ.E   PR -5
                                                                                 ~ I      FR.. 98d.9)

D0tKfr£Q JUN 2 7 1986 USNRC

                                                                   ~     JM 30 P2:2s Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Attention:   Docketing and Ser vice Br anch

Dear Sir:

On beha lf of the Tennessee Valley Authority, we wish to expre ss our agreement with, and support for , the comments which the Nuclear Utility Management and Resources Committee is filing on the proposed rule on stat ion bl ackout (USI A- 44). We wish to thank the Commission for the opportunity to comment on this issue. If our comments shou l d engender furthe r questions by the Commission or Staff, we stand ready to answer them . Very t ru ly you r s, TENN ESSEE VALLEY AUTHORI TY

                                                      ;2                irector and Licensing k.now led <J bY r 7'-H- k,.~~   ;-:-.~ --1 An Equal Opportunity Employer

Washington Public Power Supply System~P03£D RULE u0c* ET NUMBfR p * - SA

                                                                                                                 \! )

Box 1223 Elma, Washington 98541 -1223 (206)482-4428 fi:. F-re 1gJ 9 D0CK£T£D USNRC l..:!l

                                                                    ~ ~ Jo ilu            - . .

ia June 24, 1986 OFFICc OF s ,.. _

DOCKEitNG 1s1t r, R~ BRANC/.J R\ ICE. The Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, D. C. 20555 Attention: Docketing and Service Branch 1

Subject:

PROPOSED RULE STATION BLACKOUT

Reference:

a} Letter (603-85-0654), G. C. Sorensen to Karl Kniel, subject, NUREG-1032, "Eva 1uati on of Station Blackout Accidents at Nuclear Power Plants", dated November 8, 1985. b) Letter (RWW-85-071), R. W. Wells to K. Kniel, subject, NUREG-1032, "Evaluation of Station Blackout Accidents at Nuc 1ear Power Pl ants 11 , dated October 11 , 1985. On March 21, 1986, the Nuclear Regulatory Commission (NRC} published, for comment in the Federal Register, a proposed rule for resolving Unresolved Safety Issue (USI} A-44, "Station Blackout." The Supply System has reviewed the proposed rule and appreciates this opportunity to provide comments .

• In general, our findings support the results of industry initiatives to evaluate this issue. As such, the Supply System supports and endorses comments provided by NUMARC, NUGSBO and the Atomic Industrial Forum concerning the proposed rule, the technical bases for the rule (NUREG-1032), "Evaluation of Station Blackout Accidents at Nuclear Power Plants"), and the appropriateness of the staff's backfit analysis (NUREG-1109, "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout"). Specifically, we believe that the above NUREG documents do not provide an adequate technical justification for the proposed rule.

                                                                          ,.,  1~ ~ kt..

nowledged by carcl * *,/. ~ * * *

.1 *

• The Secretary of the Commission June 24, 1986 Attention: Docketing and Service Branch Page Two We would also like to direct the staff's attention to earlier comments submitted on NUREG-1032 by the Supply System [Reference a)] and by the Combustion Engineering Owners Group (CEOG) [Reference b)], and re-emphasize the detailed technical comments provided in those transmittals.

Attachment 1 represents the Supply System's comments on the proposed rule and associated Draft Regulatory Guide. Attachment 2 provides our comments on NUREG-1109 "Regulatory Analysis for the Resolution of USI A-44, Stat ion Blackout." Attachment 3 provides a copy of our comments on NUREG-1032 "Evaluation of Station Blackout Accidents at Nuclear Power Plants", previously transmitted via Reference a). Again, the Supply System appreciates this opportunity to review and prov i de comments on this subject. Very truly yours, ,!.c~kn:;er Regulatory Programs AJM/cae Attachments

Attachment l COMMENTS FEDERAL REGISTER PUBLICATION OF PROPOSED RULE CONCERNING STATION BLACKOUT

1) Consideration of additional single failures, as implied (but not included) by the Commi ssioners in the proposed Rule, would add considerably to the costs of implementation, and should be explicitly removed from consideration. Examp 1es would include a second steam driven train of auxiliary feedwater for PWRs and a backup to RCIC for BWRs. Substantial justification should be required prior to making any add on features safety grade, since utilization of available non-safety grade equipment will provide the desired result. This is acknowledged in the draft Regulatory Guide, Section C3. l .5.

Availability can be maintained at a high level by appropriate preventative maintenance and surveillance testing programs.

2) We strongly disagree with Cammi ssi oner Assel sti ne that additional measures as pursued in 11 Countries abroad" should be considered for domestic applciation. It is not apparent that the details of U. S.

Grid stabilities and on site power reliabilities are substantially similar enough to those found "abroad 11 to warrant a simple adoption of these measures.

3) The proposed definition of Station Blackout to be added to Section 50.2 should exclude AC power provided by battery backed inverters.

4) Since this rulemaking will alter the Licensing basis for the entire industry, it should be made very clear that the issues of fuel design 1imi ts, pressure boundary integrity, core coo 1i ng and containment integrity are to be addressed using best estimate codes and assumptions.

5) The implementation of this rule would require a demonstration of the ability to maintain core cooling and containment integrity for the duration of the Station Blackout (SBO) condition. This raises a substantial issue of qualification of equipment which has not been addressed and which will add significantly to the cost of implementation. In general, the equipment to be relied upon during the SBO may not have been qualified for the elevated temperatures (no HVAC) and durations (up to 8 hours). Previously most of this gear has been considered to be in a mi l d environment. Also, not addressed are the potential Human Factors implications associated with the Blackout Scenario, such as Emergency Lighting. We be 1i eve this could have a major impact on the Cost/Benefit Analysis for this rule.

Attachment l (Continued}

6) Table l of the draft Regulatory Guide identifies the criteria to be used for determining acceptable SBO duration capability. This table provides 4 hour and 8 hour criteria based on off-site power system design, diesel generator reliability, and emergency power system configuration. Using the same bases, it should be possible for certain utilities to demonstrate a zero hour blackout. This should be left as an option.

7) A definition of initial plant parameters to be assumed for analysis purposes is not provided. An example of such parameter would be Reactor Coolant Pump seal leakage. The wide variation in time dependence, due to the RC pump seal leakage issue alone, militates against a rulemaking, at least until the issue is resolved .

8) "Ability to Cope" with a Station Blackout needs to be defined.

9) With regard to diesel generator reliability, we would like to draw your attention to the recently published NUREG/CR-4557 (4/86} A 11 review of Issues Related to Improving Nuclear Power Plant Diesel Generator Reliability." This document analyzes and summarizes data and recommendations of the utility responses to Generic Letter 84-15.

The Supply System finds the recommendations in this document regarding reduction/minimization of Cold Fast Starts, prelube systems, maintenance, root cause failure analysis and training, in general to be sound, common sense approaches. We would take under advisement any recommendations for additional record keeping. Of particular interest is the finding that the average DG reliability at Nuclear Power Plants is 98%. This is a remarkable finding, given the Commission's concerns with reliability. It suggests that with a

• bit more attention to those plants with lower than acceptable availabilities, and industry acceptance of the above noted recommen-dations, the average DG reliability could be brought much closer to 100%. Thus reducing even further the probabilistic contribution of this factor to the overall core melt risk.

l 0) Clarification regarding required responses to a DG failure rate of

        >0.05 failure per demand is needed.

11} Clarification is needed as to whether failure rate is to be based upon all test and valid emergency start signals or simply on valid emergency start signals only.

Attachment 2 COMMENTS ON PROPOSED NUREG 1109 "Regulatory Analysis For The Resolution of USI A-44, Station Blackout" l) In Table 3, Page 9 - Definitions of Pl and P2 use frequency of extremely severe weather and severe weather interchangeably, thus creating confusion in the definition. I. 2) The offsite power design characteristic is dominated by weather related failures. This is not representative of the northwest BPA grid with its large hydro capacity and mild weather. The regulations should allow "no action" for grids with a demonstrable reliability above a "cut off" value, (e.g., 10- 4/yr).

3) In Table 5, Page 17 - The estimated reduction in core damage frequency, except for the worst case, is not as large as uncertainty in the calculated median values. If this is carried through the value impact statement, the logical conclusion is that the regulation is not cost beneficial except in the cases of poor grid and / or diesel generator reliability. However, these cases would be required to be remedied under current regulations.

4) Without publication of a safety goal, the probabilistic arguments presented are irrelevant. That is, for loss of all AC power to be significant, it is not enough that it is potentially a large contributor to core melt frequency. If the total core melt frequency is less than an establ ished safety goal, no change to the plant should be imposed irregardless if blackout is the largest contributor or not.

5) Pages 30 through 37 discuss the impact of this issue on eight other regulatory issues. Rather than enumerate opposing arguments in each case, we urge the NRC to consider the AIF Working Group on Station Blackout's original arguments for not separating the Station Blackout from the resolution of these issues. It is more reason ab 1e to coa 1esce the issues under the IDCOR - degraded core and source term work now on going. To separate Station Blackout frequency out as an independent variable when many sequences are dependent on it is a misapplication of the probabilistic technique (also see Item 4 above).

Attachment 3 COMMENTS REGARDING NUREG 1032 "Evaluation of Station Blackout Accidents at Nuclear Power Plants"

1) The NUREG utilizes overly conservative frequencies for loss of offsite power events. The Nuclear Safety Analysis Center ( NSAC) has published two reports summarizing loss of offsite power (LOOP) events in the U. S., NSAC/80 ( a11 events through 1983) and NSAC/85 ( event s up to 1984).

As noted by NSAC, the frequency of LOOP activities has been decreasing due to improved grid configurations, larger utility systems and improved switchyard designs. Combining NSAC frequencies for th e last 3 years gives a LOOP frequency of O. 045 per site year compared to the NUREG-1032 frequency of 0.088 per site year. The NUREG-1032 estimates seem to be based on data that is not as current as NSAC 1 s.

2) The NUREG utilizes overly conservative estimates of the time to restore AC power. Again, the NUREG-1032 data is not current and thus does not reflect the substantially upgraded grid reliabilities and switchyard designs.

3) The NUREG uses the Regulatory Guide 1.108 definition of a fa ilure to start for the Emergency Diesel Generators (EOG). This definit ion does not credit degraded mode operation or valid starts that exceed l O seconds. The ongoing effort to reduce challenges and wear on t he EDG s 1 (see for example Generic Letter 83-30) due to overly prescriptive testing requirements, will likely lead to a future relaxation of this definition in the Regulatory Guide.

In any case, NUREG-1032 is misapplying a regulatory criterion in place of more realistic success criteria based on actual experience data. The actual operating experience used by NUREG-1032 is out of date (NUREG/CR-2989-7/83). More recent data (NUREG/CR-3831-1/85) suggests that another over conservatism may have been introduced in terms of EOG failure rates and time to repair.*

4) The NUREG-1032 Accident Sequence Analysis assumes that the Loss of Core Cooling leads directly to core melt due to the short (1-2 hour) time period assumed from onset of core damage to onset of core melt, and the low probability assigned to the restoration of AC power dur ing this interval.

As noted above, in comment 2 & 3, this assumption may not be reflective of reality in terms of the actual ability to restore AC power in some form.

• Also, see Attachment 1, comment 9, regarding NUREG/CR- 4557.

Attachment 3 (Continued)

5) NUREG-1032 uses overly conservative values for the fraction of fission products released to the containment. This leads to a significantly over conservative result for time to containment failure. It is suggested that attention be given to the results obtained by the Source Term Reassessment (NUREG-0956) and the IDCOR Program. The IDCOR result of 32 hours from start of SBO to containment failure (for large dry containments; the shortest time calculated is 18 hours for Mark I and II BWR s) is based on more realistic physical and chemical models, 1

which are in good agreement with the NUREG-0956 results. However, depending on accident sequence, a difference of from 10 to 22 hours exists between NUREG-1032 and IDCOR . In view of the actual experience in recovering AC power ( no LOOP has ever lasted longer than 10 hours), this difference in time to containment failure becomes critical to the ultimate outcome of the analysis. Whether containment failure occurs at all is another issue which ought to be considered (see NUREG-0956). Thus, the probability of recovering AC power and arresting or substantially changing the accident sequence, prior to containment failure, becomes much greater. As noted on Page 7-15 of NUREG-1032, "The time to containment failure after the onset of core damage and the containment failure mode is an important factor in determining fission product release and ultimately public risk. " Core damage and vessel melt through, while undoubtedly involving

• substantial financial risk to the uti l ities, do not translate directly into public safety and health risks. A realistic evaluation of the true risk to the general public is likely to show that it is not substantially increased by Station Blackout sequences.

GULF ST.ATES POSTOFFICEBOX2951 UTILITIES

• BEAUMONT.TEXAS CO.JWP.ANV

                                                                     ~

777antxnr , AREACODF 4~6631 USNRC [

                                           '(!_y                 ~    kllliJO A1~ ;,{

I), KtT NU ~BfR PR 50 June RBG- 23923 )V--:-4l

                  'OOPO ED RULE     -                     File (#f tcJ;or2s1 / '

let- FR.. ('

• DO'

                                        -1  (>
                                               "9 )

Ol DOCKETING l ~it. BRANCH RV/Cf. IAR i' Secretary of the Commission U. S. Nuclear Regulatory Commission Washington, D. C. 20555 Attention: Docketing and Service Branch Gentlemen: Gulf States Utilities Company (GSU) is pleased to submit its comments on the proposed revisions to 10CFR Part 50, "Station Blackout". GSU is not convinced that the proposed rule to address the station blackout concern has been properly justified. It is our understanding that the current station blackout (SB0) contribution to core melt is due in part to low emergency diesel generator reliability and a 24 hour estimated SB0 duration. The estimated 24 hour SB0 duration conflicts with a proposed four (4) to eight (8) hour SB0 capability. Use of a more reasonable SBO duration and diesel generator reliability in the PRA, coupled with current industry knowledge on the source term, reduces the benefits derived from a SB0 backfit by orders of magnitude over what was stated in the Valve/Impact statement. Notwithstanding, it is felt

• that there are areas in the proposed rule that should be clarified.

Specifically;

1. The proposed rule does not provide sufficient direction on the quality classification of plant modifications that may be required to meet the rule. The rule should address the method to be used to assure the operability and reliability of these modifications. In lieu of this, it is suggested that the quality classification of plant modifications implemented to meet the proposed rule should be commiserate with the classification of the system they support.

2. The phrase important-to-safety is used throughout the proposed rule.

It is our understanding that the definition of this phrase has not been completely agreed upon by the nuclear industry and the NRC. Until this has cleared up, it is recommended that this new regulation avoid the use of this term. Acknowledged by cord 1/4J.t -~

d s n I

3. According to the proposed 10CFR50.63 rule, the station blackout duration is to be based, in part, on the reliability of onsite AC power sources. This may unneccessarily penalize new plants which have not had sufficient time to establish reasonable onsite AC power reliability data. Typically the reliability of onsite power systems increases during the first few years following startup. It does not appear that adequate consideration for this has been allowed.

We appreciate the opportunity to comment on the subject proposed rule and hope that the above comments will assist you in its finalization . Sincerely, f ~.:~ke~ Manager-Engineering, Nuclear Fuels & Licensing River Bend Nuclear Group JEB/LAE/DWL/lp I

UNION OF CONCERNED SCIENTISTS 1616 P Street, NW S. 310 Washington, DC 20036 * (202) 332-0900 DOCKETED June 27, 198'>5 RC 06CKET NU P.ROPOSED RULE RPR'- 5

                                                          'B6 JJ1 21     P4 :35             @'I-FR.?~~)

Secretary of the Comnission U.S. Nuclear Regulatory Corrmission Washington, DC 20555 Attn: Docketing and Service Branch RE: Union of Concerned Scientists Comments on Proposed Station Blackout Rule

Dear Mr. Secretary:

Enclosed are the Union of Concerned Scientists comments on the proposed Station Blackout Rule. I spoke with Alan Rubin last week about sul:xnitting these a week after the June 19th due date and was told that there was unlikely to be any problem. The conversation was subsequently confinned in writing. In any case, conflicting committments made an earlier submission impossible and I therefore request an extension that these corrments be accepted and considered although they are submitted one week after the published due date. Very truly yours,

                                                ~~Ellyn R. Weiss General Counsel Union of Concerned Scientists Acknowledsed by Cambridge Office: 26 Church Street

• Cambridge, Massachusetts 02238 * (617) 547-5552

     -\

UNION OF CONCERNED SCIENTISTS 1616 P Street, NW S. 310

• Washington, DC 20036 * (202) 332-0900 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION UNION OF CONCERNED SCIENTISTS' COMMENTS ON PROPOSED STATION BLACKOUT RULE Introduction
• The Union of Concerned Scientists supports the NRC's decision to require that all nuclear plants be capable of coping with a station blackout. These c01I111ents take no position with regard to the specific solutions outlined in the proposed Station Blackout Rule; they are intended to support the general conclusion that action is required to provide reasonable assurance that an accident involving total loss of onsite and offsite power will not result in harm to the public health and safety.

Station blackout -- total loss of onsite and offsite power -- is one of

• the longest-standing and most serious of the NRC's list of generic unresolved safety issues. The problem is so serious because 1) such total losses of power have occurred in the past are likely to occur in the future, 2) the margin of time available to restore power before the onset of core damage is quite short (1-16 hours depending on plant design),11 and 3) since a total loss of power ensures the unavailability of vital safety systems, accidents involving station blackout lead to the greatest postulated releases of radio-11 NUREG-1032, p. C-3 Cambridge Office: 26 Church Street
• Cambridge, Massachusetts 02238 * (617) 547-55l>2

activity and the largest postulated numbers of deaths and injuries. This has been known and remained unchallenged since the publication of the Reactor Safety Study (WASH-1400) in 1975. 21 In that study, station blackout is the most important component of the accident sequences which constitute 40% of the total risk from

              . all PWR dominant accident sequences.    -Id. Main Report at 99.

Thus, station blackout constitutes both a relatively high probability event and an event of extraordinarily high potential consequences to public health and safety. Action to preven~ station blackout is required to prevent undue risk to public health and safety *

• 1. A Cost-Benefit Analysis Approach is Inappropriate in this Case The Commission's new backfit rule contains the so-called "compliance exception" which establishes that the cost-benefit analysis otherwise mandated by the rule shall not be applied in cases where "a modification is necessary to bring a facility into compliance with a license or the rules or orders of the Commission, or into conformance with written cornmittments by the licensee." 10 CFR 50.109 (a)(4)(i). In this case, action is required to
• secure compliance with a pre-existing rule; namely, GDC 17 of Appendix A to CFR Part 50.

GDC 17 provides in pertinent part: Criterion 17--Electric power systems. An onsite electric power system and an offsite electric power system shall be provided to permit functioning of structures, systems, and components important Y Of the nine PWR release categories postulated in the Reactor Safety Study, the two most serious involve station blackout.

to safety. The safety function for each systans (assuming the other system is not functioning) shall be to provided sufficient acpacity and capability to assure that (1) specified acceptable fuel design limits and design conditions of the reactor coolant pressure boundary are not exceeded as a result of anticipated operational occurrences and (2) the core is cooled and containment integrity and other vital functions are maintained in the event of postulated accidents. Provisions shall be included to minimize the probability of losing electric power from any of the ranaining supplies as a result of, or coincident with, the loss of power generated by the nuclear power unit, the loss of power from the transmission network, or the loss of power from the onsite electric power supplies. The facts gathered by the NRC which form the basis for the proposed rulanaking establish that, in fact, the electrical systans of current plants do not meet the fundamental requirement of providing assurance that the plant's safety systems will function, that the core will remain cooled and containment integrity will be maintained, nor do they "minimize" the probability of losing the remaining power supplies as a result of or coincident with the loss of power generated by the nuclear plant, the loss of power from the grid or the loss of power from the onsite power supplies. In short, the evidence demonstrates that since the historical incidence and future probability of total loss of onsite and offsite power (station blackout) is relatively high, the measures proposed by the NRC are necessary to bring licensed plants into compliance with current rules. This case is analogous to the situation presented to the Commission in the areas of fire protection (GDC 3) and environmental qualification of safety equipment (GDC 4), where the Coomission concluded that test data and other information demonstrated that the actual equipment installed in plants did not

provide the assurance required by the applicable rules. The Corrmission therefore ordered all operating plants to take a series of remedial actions in order to bring them up to the level of safety called for by the rules, which had been assumed to be present at the time of their original licensing,~ Petition for Emergency and Remedial Action, CLI 80-21, 11 NRC 705, 711-712. In similar fashion, the facts here clearly show a widespread failure to meet the fundamental purpose of GDC 17: providing a high degree of assurance that there will be electrical power available when needed to run the panoply of vital safety systems necessary to mitigate an accident, maintain containment

• integrity and safely shut down a nuclear plant and remove decay heat after the shutdown.

As the NRC data show, the historical frequency of loss of offsite power is once per ten (10) site-years. That is, each plant can expect about four losses of offsite power during its 40 year operating life. If 100 plants are operating, one can project roughly 300-400 such events. Moreover, onsite emergency power sources - diesel generators -- are notoriously unreliable. NRC data from operating experience show a typical failure rate of an individual diesel of one in 40 demands. While total emergency power availability for a typical facility with two diesel generators is estimated by NRC at 2 X 10-3 per demand, it should be noted that plants with two diesels are not infrequently operating with one unavailable because of maintenance or testing, as permitted by the license technical specifications. Thus, historical operating data, not hypothetical future probability analyses, establish that the requirement embodied in GDC 17 for highly reliable sources of power to "permit functioning of structures, systems, and components important to safety" is not currently being met and further action

is necessary to secure compliance with the fundamental intent of the rule and the standards set by the Atomic Energy Act. Under these circumstances, the backfit rule itself makes cost-benefit analysis inapplicable.

2. Even Applying Cost-Benefit Principles, Action To Require Plants to Pursue Means of Coping with Station Blackout is Clearly Required Even assuming that cost-benefit analysis is appropriate in this case, the costs of taking the steps proposed by the staff to cope with station blackout are clearly outweighed by the benefits. Indeed, the costs are so trivial - an average of $600,000 per reactor -- compared with the human and financial cost of an accident as to make this conclusion beyond serious

• question. The cost of this action is comparable to the cost of replacement power for a one or two .£§Y outage for a typical nuclear plant.

illustration, in a sworn statement recently filed by the Pacific Gas and As Electric Company, it was estimated that purchase power to replace that generated by San Onofre Unit 1 would increase fuel costs to that utility's customers by $1 million per day of outage *.11 Moreover, the benefits of substantially reducing the potential for core damage from station blackout are great. As noted above, the Reactor Safety Study estimated that station blackout is a major component of those accident sequences which constitute about 40% of total risk from all PWR dominant accident sequences. WASH-1400, Main Report, p. 99. Moreover, there is good reason to believe that WASH-1400 underestimated the potential consequences of these accidents. Affidavit of James D. Shiffer in support of Pacific Gas and Electric Co.'s Response to Emergency Motion for Stay Pending Review, San Luis Obispo Mothers for Peace. et al. v. NRC, U.S. Court of Appeals, 9th Cir., No. 86-7297, June, 1986.

First, NRC's most recent source term research results, contained in NUREG-0956, indicate that the consequences of an accident could be greater than calculated in WASH-1400 due to containment failure. One of the scenarios most likely to lead to containment failure involves station blackout, since the safety features designed to mitigate an accident, such as containment spray, cannot operate without power. The ice condenser containments pose special risks in this regard. NUREG-0956 states that a station blackout event in an ice condenser plant "most like~y" will cause early containment failure. Second, the fallout from the Chernobyl accident suggests that one of the

• key assumptions used in calculating source tenns may be wrong. Data published from Sweden shows that 75-85% of the iodine isotopes released from Chernobyl were in gaseous form, in sharp contrast to the assumption used in NRC's source term estimates that most of the iodine isotopes would be in particulate form and would plate out and remain in the containment even should the containment be breached. Nucleonics Week, June 5, 1986, p. 2. Changing this assumption to conform with the Chernobyl evidence could substantially raise consequence projections.

In addition, it is UCS's understanding that a new advanced computer code

• developed at the national laboratories estimates accident consequences to be significantly greater than the CRAC codes currently used.

In sunmary, available evidence indicates that the consequences of an accident involving station blackout may be even worse than those estimated* either in WASH-1400 or in the NRC's more recent studies. In addition, the benefits of the proposed action have been understated. As NRC acknowledges, the action proposed to reduce station blackout risks will also have benefits in addressing several of the other generic unresolved safety issues, such as USI A-45, shutdown decay heat removal; Generic Issue

19 A-30, DC power supply adequacy; Generic Issue B-23, reactor coolant punp seal failure, and a number of other safety issues discussed at NUREG-1109, pp. 17-21 . Thus, calculation of the overall benefit of the proposed action should properly include consideration of the added safety benefit associated with reducing the risk related to those generic issues as well. Indeed, were the Commission to properly consider the benefits achievable by addressing all of these related safety issues together, it would probably conclude that the most cost-effective way of achieving a su~stantial overall safety improvement would be to follow the approach used in German plants of a dedicated, bunkered,

• totally independent decay heat removal system or the French approach delineated in Conmissioner Asselstine's separate views accompanying this proposed rulemaking. By unjustifiably narrowing the scope of its consideration, the Conmission has effectively precluded even a rational analysis of these alternatives. UCS believes this is a serious error in the Corrmission's approach.

Another serious error in the analysis is the NRC's failure to consider at all the risks associated with a station blackout combined with another single failure. The NRC's analysis essentially assumes that blackout is the

• single causitive event i nvolved. However, there are many other possible scenarios involving station blackout. For example, one can postulate a stea:n generator tube rupture leading to a plant shutdown which causes a loss of offsite power. The plant is operating with one emergency diesel shut down for maintenance (as allowed by typical technical specifications) and the remaining diesel fails, as historical data indicate that it will for 40% of demands.

This sequence of events is entirely possible and, we believe, not less likely than many others considered by NRC. It is certainly not technically justified to simply assume it away, as NRC has done. UCS believes that the Conmission's

-- analysis should properly include the risk associated with accidents where station blackout occurs as a result of another event, not simply alone, and where the station blackout accident sequence includes one random single failure. The separate remarks of Conmissioners Roberts and Zech which accompany the proposed rule ask for cooments on whether the action proposed "[i]s the reduction of risk only a small percentage of the overall risk or is it a major component of an already small risk?". Neither formulation of the issue is a useful or comprehensive statement of the situation presented. As shown above,

• station blackout is clearly a major component of the total risk posed by operating nuclear plants. The magnitude of the total risk is largely unknowable due to the enormous uncertainty which surrounds probabilistic assessments. Indeed, it is beyond question that the overall risk calculations, particularly the arbitrary point estimates, are the least reliable of all numbers generated by these probabilistic risk estimates. UCS incorporates herein "The Source Tenn Debate," A Report by the Union of Concerned Scientists, January, 1986, sent to NRC as UCS's technical comments on NUREG-0956, which details the technical reasons for UCS's conclusion that
• NRC's overall risk estimates are nonconservative and not reliable.

However, even NRC's calculations are that, at the 95% confidence level, the estimated core damage frequency for the four reference plants for station blackout sequences alone straddles NRC's safety goal (See NUREG-1032, Figure 8.5, p. 8-9, copy attached.), that the proposed action would in two of three designs given approximately halve that risk (NUREG-1109, Table 5, p. 10) and that the action is necessary to keep expected core damage frequency from station blackout alone at or lower than 10-5 per reactor year for all plants. NUREG-1109, p. 25. Considering that station blackout, should it occur for

-- even relatively short time periods, disables safety systems and therefore presents high containment failure probabilities, this is clearly a risk that should be immediately addressed. The low cost of the action proposed provides added support for this conclusion. Certainly, if the backfit rule is ever going to permit any b3ckfit, this is such a case. If the Commission votes otherwise, it will confirm that the backfit rule is a mechanism for preventing any safety improvements at operating nuclear plants. Conclusion

• For the reasons stated above, the Union of Concerned Scientists sutmits that action is required to substantially reduce the risk associated with station blackout at operating nuclear plants. That action is necessary to meet current NRC rules, namely GOC 17 of Appendix A to 10 CFR Part 50, and to achieve the level of safety mandated by the Atomic Energy Act. Cost-benefit anlysis is not appropriate in this case either under the new benefit rule or the Atomic Energy ~ct.

Moreover, even using a cost-benefit approach, action is clearly

• required. The costs of the step; proposed by the staff are trivial and are greatly outweighed by the benefits of addressing one of the single largest contributors to risk posed by operating plants.

Respectfully Submitted, June 27, 1986 Q~pUJeAr-Genera1 ounsel Union of Concerned Scientists

PWR with 1 steam BWR with 1 BWR with BWR w ith 10*3 Driven AFW Train Isolation Condenser r-------+----...;...;.__;..__;..~+~~.;....;;.-+-;.;...,,;...,;.;. HPCI/RCIC HPCS/RCIC _ ___

          -.CII CD 0

10-4

                    ~-.-. _

(J ca a: CD ll. CD u 10-5 .. 2 w - *-*- ... d w a: LL w

        . (!I
           <                                                    .,~

. :E C

10-6 *- -  :  ::*"":--*~ . ~

                                                                                        -=-:. i.

w a: 0 0 C w .. t- .. . ..

E *.

j:: 10*1 Upper 95% Confidence Limit en w Mean ,. 10--8 Median Point Valve Lower 5% Confidence Limit al N al m N al cc

                                                                            -              m N

a:i al N N Cl

                                                                                                                           ... al N

N

                                      ~          0     d                :::::i             0            :::::i  :::)    :::)      :::>
                               ~      ~                                 :E                              ~       :E      ~         ~

t- t- ~ t- ~ t-

E t- t- t- t- ~

ACCIDENT SEQUENCES Figure 8.5 Estimated core damage frequency showing uncertainty range for four reference plants Source: NUREG/CR-3226 8-9 .

1 )J Mcllll Str, t

     \Vh,r r' ,ns I\J,wYnk Ul 1(

9 14 GR I r l4(1 oUCKH NUWIBtR \> R- 50 \ l!llllPllst:0 RULE@!- FR q'l ~ Cf )

 , . NewVorkPower                                                                      John C. Brons S"n1or V c' L' Pre, rlent
 . , Authori1y                                                    OOCKETEO US NRC N,irlem li, nc'r non June 19, 1986
                                                           ~=~~ mo :13 OFF ICE Of St LH .i AH'f The Secretary of the Commission                        OOCKETING & SERV ICf.

BRANCl-1 U. S. Nuclear Regulatory Commission Washington, D.C. 20555 Attention: Docketing and Service Branch

Subject:

Indian Point 3 Nuclear Power Plant Docket No. 50-286 James A. FitzPatrick Nuclear Power Plant Docket No. 50-333 Proposed Rule on Station Blackout

Dear Sir:

The Nuclear Regulatory Commission has proposed a rule on Station Blackout. This rule has been reviewed in detail by the New York Power Authority as an active participant in the Nuclear Utility Group on Station Blackout (NUGSBO). The NUGSBO comments on the proposed rule have been endorsed by the Nuclear Utility Management and Resources Executive Committee (NUMARC). The New York Power Authority is in agreement with and supports the comments which are being filed on the proposed rule by NUMARC. Should you or your staff have any questions regarding this matter, please contact Mr. P. Kokolakis or Mr. J. A. Gray, Jr. of my staff . Very truly yours,

                                                             ,LCl/L cJ ohn C. Brans enior Vice President N clear Generation cc List Attached Acknowledged by card.J!i?!t./e....~

s l\

cc: Office of Resident Inspector Indian Point Unit 3 U. s. Nuclear Regulatory Commission P. o. Box 66 Buchanan, N. Y. 10511 Office of Resident Inspector James A. FitzPatrick Nuclear Power Plant

u. s. Nuclear Regulatory Commission P. o. Box 136 Lycoming, New York 13093 Mr. Steven A. Varga, Director PWR Project Directorate No. 3 Division of PWR Licensing - A U. S. Nuclear Regulatory Commission 7920 Norfolk Avenue Bethesda, Md. 20014 Mr. Daniel R. Muller, Director BWR Project Directorate No. 2 Division of BWR Licensing U. s. Nuclear Regulatory Commission 7920 Norfolk Avenue Bethesda, Md. 20014

 .A ARKANSAS POWER & LIGHT COMPANY POST OFFICE BOX 551 LITTLE ROCK. ARKANSAS 72203 (501) 371-40CO Junr;rn-986 clOCKrf NUutfR oROPOSfo RU~E   PR~ l)

(jt f:'R. 9Fc,)9) 0CAN068606 Secretary of the Commission Office of the Secretary U.S. Nuclear Regulatory Commission Washington, D.C. 20555 ATTENTION: Docketing and Service Branch

SUBJECT:

Arkansas Nuclear One - Units 1 & 2 Docket Nos. 50-313 and 50-368 License Nos. DPR-51 and NPF-6 Comments on Proposed Rule for Resolving USI A-44, "Station Blackout" Gentlemen: On March 21, 1986, the Nuclear Regulatory Commission (NRC) published for comment in the Federal Register a proposed rule for resolving Unresolved Safety Issue (USI) A-44, "Station Blackout. 11 Arkansas Power and Light Company has reviewed the proposed rule and appreciates the opportunity to provide comments .

• In general, our findings support the results of industry initiatives to evaluate this issue. As such, Arkansas Power and Light Company supports and endorses comments provided by NUMARC/NUGSBO and the Atomic Industrial Forum concerning the proposed rule, the technical bases for the rule (NUREG-1032, "Evaluation of Station Blackout Accidents at Nuclear Power Plants"), and the appropriateness of the staff's backfit analysis (NUREG-1109, "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout 11 ) . Specifically, we believe that the above NUREG documents do not provide an adequate technical justification for the proposed rule.

We would also like to direct the staff's attention to earlier comments submitted on NUREG-1032 by the Combustion Engineering Owners Group (CEOG) [Mr. R. W. Wells to Mr. K. Kniel, RWW-85-71, dated October 11, 1985] and reemphasize our advocacy of the detailed technical comments provided in that transmittal. Ac now1 MEMBER MIDDLE SOUTH UTILITIES SYSTEM

I I I I I

                               'O, I u s. t l,

L I L

                      ~/:23Jr0        I I

I I I I I I I I I I I

June 16, 1986 In conclusion, based on the NUMARC/NUGSBO and Atomic Industrial Forum comments on the proposed rule, and the CEOG comments provided, Arkansas Power and Light Company does not consider a Station Blackout event to be a significant contributor to the overall risk to the public health and safety. This conclusion is also expected to be supported by the Industry Degraded Core studies. Station Blackout would be expected to be even less of a contributor if the revised source term data were to be utilized by the staff. It does not seem credible, therefore, that the proposed rule will provide a substantial increase in the overall protection of the public health and safety and that the costs of implementation are justified for Arkansas Nuclear One. Again, we appreciate the opportunity to review and provide comments on this subject. Very truly yours,

                                   ~~

• JTE:LVP:ji Ll;_Ted Enos, Manager Nuclear Engineering and Licensing

COMBUSTION~ ENGINEERING June 1 9, 1 986 LD-86-028 Secretary of the Commission Office of the Secretary u.s. Nuclear Regulatory Commission ttU 8£1\PR -

• O RULE \

Washington, DC. 20555 Attention: Docketing and Service Branch @'I Ft:. C/tJ.9 J

Subject:

Comments on Proposed Rule for Resolving USI A-44, "Station Blackout".

References:

(A) R. W. Wells to K. Kniel, "NUREG-1032, 'Evaluation of Station Blackout Accidents at Nuclear Power Plants', Combustion Engineering Owners Group Comments (50 FR 24332)", RWW 71, October 11, 1985. (B) A. E. Scherer to K. Kniel, "Comments on Draft NUREG-1032,

                            'Evaluation of Station Blackout Accidents at Nuclear Power Plants' (50FR24332)", LD-85-048, October 14, 1985.

(C) E. J. Butcher to A. E. Lundvall, Jr., "NUREG-0737 Item II.K.3.25, Reactor Coolant Pump (RCP) Seal Integrity Following Loss of Offsite Power", October 7, 1985.

Dear Sir:

On March 21, 1986, the Nuclear Regulatory Commission (NRC) published for comment in the Federal Register a proposed rule for resolving Unresolved Safety Issue ( USI) A-44, "Station Blackout". Combustion Engineering (C-E) has reviewed the proposed rule and appreciates the opportunity to provide comments *

• In general, our findings support the results of industry initiatives to evaluate this issue. As such, C-E supports and endorses comments provided by NUMARC/NUGSBO and the Atomic Industrial Forum on the proposed rule, the technical bases for the rule (NUREG-1032, "Evaluation of Station Blackout Accidents at Nuclear Power Plants"), and the inappropriateness of the staff's backfit analysis (NUREG-1109, "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout"). Specifically, we believe that the above NUREG documents do not provide an adequate technical justification for the proposed rule.

C-E would also like to direct the staff's attention to earlier comments submitted on NUREG-1032 by the Combustion Engineering Owners Group (CEOG) [Reference (A)] and re-emphasize our advocacy [Reference (B)] of the detailed technical comments provided in that transmittal. A copy of Reference (A ) is provided as an attachment to this letter for your convenience. Power Systems 1000 Prospect Hill Road (203) 688-1911 Combustion Engineering, Inc. Post Office Box 500 Telex: 99297

                                          'Ii/_, J,   Windsor, Connecticut 06095-0500 Ackr,ow edg d by
                                            /~.~b... 94

0 s n 1

Secretary of the Commission June 19, 1986 Page 2 LD-86-028 With respect to NUREG-1032, C-E believes that the issue of reactor coolant pump seal integrity during a station blackout is worthy of additional comment. Combustion Engineering and the CEOG have long endorsed the position that reactor coolant pump seal cooling is not necessary to assure integrity of the RCP seals in an idle pump and, furthermore, complete loss of seal function will not occur due to station blackout in a C-E designed Nuclear Steam Supply System (NSSS). The staff appears to acknowledge this position in Table 7.1 on page 7-7 of NUREG-1032 where core uncovery time due to a 100 gpm total leak rate from RCP seals is not calculated nor apparently applicable to C-E PWRs. In addition, on October 7, 1985, the NRC issued a Safety Evaluation Report [Reference (C)] which we feel further supports our position that seal cooling is not necessary to assure seal integrity. The SER states, "*** the staff no longer requires automatic reinitiation of coolant to the reactor coolant pump seals following a loss-of-offsite power event." Since seal cooling is not considered to be necessary to protect seal integrity following a loss of offsite power event, it follows that seal cooling should not be necessary to maintain seal integrity during a station blackout event. This position is supported by the successful 50 hour station blackout test performed on a production seal cartridge by a licensee having a C-E NSSS. In view of the above discussion, C-E believes that the staff should re-examine the technical bases in NUREG-1032 as we feel that the generic conclusions in that document are inappropriate and unjustified. In conclusion, based on the NUMARC/NUGSBO and Atomic Industrial Forum comments on the proposed rule, the C-E and CEOG comments provided in References (A) and (B), and our position with respect to seal integrity, C-E does not consider a Station Blackout event to be a significant contributor to the overall risk to the public health and safety. This conclusion is also expected to be supported by the Industry Degraded Core Rulemaking studies. Station Blackout would be expected to be even less of a contributor if the revised source term data were to be utilized by the staff. It does not seem credible, therefore, that the proposed rule will provide a substantial increase in the overall protection of the public health and safety nor does it seem likely that the costs of implementation can be justified. If the staff should have any questions concerning our comments, they may feel free to contact me or Mrs. R. o. Hoogewerff of my staff at (203) 285-5217. Very truly yours, COMBUSTION ENGINEERING, INC. Director Nuclear Licensing AES: bkm Attachment

Attachment NGRTIIEAST UTILfflES TN( C.0-.CTIQIT LG-IT ANO IIIOWl!IIII C~AHY General Offices

• Selden Street, Berlin, Connecticut

(([fJ WIST(- ~ T T S ELECT-.C C,OWlt#Y HQ.'t()q WAT(IIII ,owfllll ~ 1 , # Y HQR1'to<<AIT Ul ...Tl(t M ~ CJ:/ll!l#llff NQIIIT.,._Ml NUQ.I_. l,_IIIO'f' CC:.-Ml' PO. BOX 270 HARTFORD, CONNECTICUT 06141-0270 (203) 665-5000 October 11, 1985 RWW-85-71 Mr. Karl Kniel Division of Safety Technology U.S. Nuclear Regulatory Commission Washington, DC 20555

Subject:

NUREG-1032, "Evaluation of Station Blackout Accidents at Nuclear Power Plants", Combustion Engineering Owners Group

• Comments (SO FR 24332)

The Combustion Engineering Owners Group (CEOG) appreciates the opportunity to review and provide comments on the subject draft NUREG. completed its review of NUREG-1032 and is herewith submitting comments which we hope will be both timely and useful. The CEOG has Since Station Blackout (SBO) was designated an Unresolved Safety Issue in 1980, the NRC and Industry have conducted a number of studies which have produced a wealth of data on the issue. The CEOG believes that, although the availability of current data is acknowledged in NUREG-1032, it has not been properly utilized in establishing a technical basis for resolving this issue. The CEOG strongly recommends that the NRC take advantage of these recent studies which, for example, reflect the many upgrades the industry has implemented resulting in more reliable grids and switchyards. The CEOG therefore submits that, with the current and more representative data, the NRC will find that SBO does not contribute significantly to the risk to public health and safety from operation of nuclear power plants .

• The CEOG additionally recommends that the NRC take advantage of other related industry efforts, such as IDCOR, which are referenced in NUREG-1032 but not fully utilized. This has led, for example, to a large disparity between containment failure times as determined in NUREG-1032 and IDCOR.

Utilizing unnecessarily conservative fa~lure times, as in NUREG-1032, only leads to over estimates of the SBO consequences. Although the CEOG does not consider the SBO event a significant contributor to the overall risk to public health and safety, we are incorporating the loss of offsite power and station blackout events into our emergency procedure guidelines. In doing so, the CEOG believes that the SBO concern will be satisfactorily resolved for utilities with CE designed Nuclear Steam Supply Systems.

Mr. Karl Kniel October 11, 1985 Additionally, the CEOG's findings are consistent with those of the Nuclear Utility Group on Station Bl ackout in that SBO does not significantly contribute to the risk from operating nuclear power plants. Detailed comments on NUREG-1032 are attached in support of the above discussions. Again, the CEOG appreciates the opportunity to review and provide comments on this subject. If you have any questions or would like to discuss our comments, please feel free to contact me at ( 20~) 665-3614. Very truly yours, 20/4)lltt-

                                     ~J:'    Wells, Chairman CE Owners Group RWW/drg cc: F. Maraglia (NRC)

CEOG Detailed ColTITlents on NUREG-1O32

1. LOSS OF OFFSITE POWER FREQUENCY Loss of Offsite Power (LOOP) frequency has continued to decline over the years due to major changes that have occurred in both the onsite and offsite electrical distribution systems.

As part of the ongoing Nuclear Safety Analysis Center's (NSAC) program, NSAC has published two reports sulTITlarizing all LOOP events in the U.S .*

• NSAC/8O discusses all events through 1983 and NSAC/85 revises the data to include 1984. These reports document the decreasing LOOP frequency and discuss the phenomena of those events.

The frequency of LOOP, as noted above, has continually decreased due to improved grid designs, larger interties and improved site switchyards. These upgrades have contributed significantly to the reduction in NRC estimated LOOP of .088/site year to NSAC frequency of .O45/site year. Some specific changes that have occurred in an effort to improve grid and site reliability with respect to LOOP are discussed below.

1. Utilities have installed load shedding relays to prevent grid collapses such as the one that occurred in the Northeast in 1965 .

2. Emphasis has been put on wheeling power between utilities or import-ing power from major generation centers. This has led to larger, higher voltage interties that are less susceptible to load changes and/or weather effects.

3. Three of the LOOP events longer than one hour have occurred in the Southeast during the oil embargo of 1977 when local generation was minimized and power importation was maximized. Since then, two non-oil power plants and a major 345 kV lie line has been added.

4. Only a small percentage of sites have had LOOP events of any duration and frequency. Sites that have experienced LOOP events have taken corrective actions to prevent reoccurrence.

a) Only 3 sites have had 2 losses of off-site power longer than 1 hour (none have had more than 2 losses longer than 1 hour). These losses all occurred a number of years ago: 4 in 1977, 1 in 1978, and 1 in 1980. These three sites have taken or are taking major correc-tive actions. b) 1/4 of all losses have occurred at 2 sites . At one of these sites, the last loss of off-site power was 8 years ago. The problem was corrected by expanding the 345 kV transmission system to become a multiple supply source for the 115 kV system that supplies a source of off-site power to the site. The site is also connected to the 345 kV transmission grid. The other site is in the process of expanding its switch-yard. Current plans call for adding five 240 kV breakers to reduce interactions between the four generating units at the site *

• Based on the results of the review of the NUREG-1032 LOOP frequencies, the CEOG believes that NRC utilizes overly conservative frequencies for the S80 accident analysis. This is due to utilizing outdated data and not taking into consideration more reliable and updated grids and switchyards.

2

2. DIESEL GENERATOR RELIABILITY NUREG-1032 assumes that diesel generators (DG) fail to start in 10 seconds. Additionally, it does not give credit for operation in a degraded mode. These assumptions have led to higher failure rates for both the single and double DG failure events analyzed in NUREG-1032 for the S80 event.

More recent work on DG reliability(!) by Oak Ridge National Laboratory has shown that for a single DG failure, excluding failures to start in 10 seconds and degraded mode operation, the failure rate to be 2.9 x 10- 3;

• demand. Additionally, for the common mode failure of both DGs, where a geometric mean assumption was used, the failure rate was estimated to be 1.6 x 10- 4 per demand.

Utilizing more recent DG reliability information, as in the above mentioned ORNL report, will reduce unnecessary uncertainty and contribute to the fact that S80 poses less risk than NUREG-1032 postulates.

3. RCP SEAL INTEGRITY It is the CEOG's position that RCP seal cooling is not necessary to assure the integrity of the RCP seals in an idle pump and that complete

• loss of seal function will not occur due to a station blackout in a CE designed NSSS. This position is supported by two tests on prototype seal cartridges and operating history.

(1) Kaul, W. K., Borkowski, R. J., "The In-Plant Reliability Data Base for Nuclear Plant Components: Interim Report-Diesel Generators, Batteries, Chargers and Inverters," NUREG/CR-3831 (ORNC/TM-9216), January, 1985. 3

The first test was a simulated station blackout test on a prototype seal cartridge performed at SBO conditions, i.e., no shaft rotation, plant operating temperature and pressure. The four seals were mounted in a cartridge which has piped up to the Byron Jackson test loop in order to simulate-actual plant conditions. The test was run for in excess of 50 hours without loss of seal function. Seal c~ntrolled leakage remained within normal limits {approximately 1.0 gpm) and vapor seal leakage did not exceed 0.25 gpm. Vapor seal leakage is that which leaks past the fourth seal and potentially into the containment, however with such a low leakage rate it is normally piped off by gravity drain to the reactor drain tank. Controlled leakage exits the seal cartridge between the third seal and fourth (vapor) seal and is piped off to the volume cont:ol tank. Seal pressure breakdowns remained within acceptable limits and the vapor seal temperature reached approximately 400°F. Post test inspection showed a cracked vapor seal rotating ring and permanent compression of the "O" rings and hardening of the "U" cups. After the test, the seal cartridge was rebuilt to ope~ational standards by replacing the broken rotating ring and all elastomers. This test confirmed the capability of the seals to withstand a SBO for a time period in excess of any of the times considered in NUREG-1032. The second test was a 30 minute loss of component cooling water test with the pump running. This test was potentially more severe than the SBO test because the rate of heat generation in the seals is greater for an operating pump than an idle pump without CCW. The test was run on a prototype RCP in the Byron Jackson test loop at plant operating conditions. During the 30 minute test, seal function was maintained. Maximum controlled leakage was 2.0 gpm (1.0 gpm normal value) and vapor seal leakage did not exceed 0.5 gpm. Again pressure breakdowns remained within acceptable limits and the vapor seal temperature peaked at 450°F. After CCW was restored the pump ran 2-1/2 hours and the seal leakage rates returned to normal. Post test inspection showed a cracked vapor seal rotating ring and some deterioration of the elastomers. This test confirmed the capability of the seals to withstand an abnormal event equal to or more severe than a SBO. 4

At various CEOG plants there have been a number of loss of CCW events with the pumps running or idle or a combination of both which add further evidence to the CEOG position that seal function will not be lost as a result of a S80. The events are su1T1T1arized below: EVENT A: 1974 &1975 Seals changed in all four pumps after loss of CCW incidents. Time periods were not available but seal function was maintained. 1981 - Pumps remained on hot standby for one hour after loss of CCW. Pumps restarted without seal problems. (Hot standby refers to idle pumps in a hot. pressurized reactor coolant system.) EVENT B: 1977 & 1980 Replaced all four seals on two occasions where CCW was lost.

             -  1980 incident involved a natural circulation cooldown because CCW could not be restored.
             -  No complete loss of seal function occurred during either

• EVENT C:

of these events. 1984 - 2 of 4 pumps remained on hot standby for approximate ly six hours without CCW.

             -  Seal temperature reached in excess of 350°F prior to restoration of CCW.

5

             - Upon pump restart seals in one pump operated normally.
             - Lower seal in other pump did not pressure stage properly but recovered completely later.
             - No evident increase in normal_ operating seal lea kage.
             - Seals in both pumps operated for approximately 2 more months then replaced during 1985 refueling outage.

Seals in both pumps were in operation for 10 and 12 months prior to event.

             - Event confirmed capability of seals to withstand conditions similar to S80.

EVENT D: 1984 - Shaft/impeller separation occurred on one pump.

             - While at 60% power, lower and middle seals failed simultaneously.
             - Plant shutdown commenced and after two hours upper seal failed .
             - After three hours affected pump secured.
             - During entire event vapor seal functioned to prevent gross leakage to containment.
             - Event did not involve loss of CCW but confirmed capability of vapor seal to maintain seal system integrity.

6

EVENT E: 1985 - CCW to RCP seal coolers isolated after inadvertent containment spray activation.

             - Attempts to restore CCW to 3 of 4 pumps were unsuccessful so these pumps were tripped within 3 minutes.
             - After approximately 1 hour and 15 minutes seal leakage to containment from one pump was estimated to be approximately 3 gpm.
             - Plant instrumentation showed that upper seal had failed but lower, middle and vapor seals functioned.
             - Seal leakage from other three pumps remained normal and plant was cooled down.
             - During entire event seals functioned to prevent gross leakage to containment.
             - Event confirmed capability of sea 1 package to ma i ntain seal system integrity.

EVENT F: 1985 - CCW isolation valves for RCP seal coolers closed due to inadvertent safety injection actuation.

             - RCP's secured in less than 10 minutes due to loss of CCW.
             - CCW restored to 2 pumps after 20 minutes and after approximately 4 hours to other 2 pumps.

7

             -  Controlled leakage from all four pumps remained normal (approximately 1.0 9pm). Vapor seal leakage did not increase.
             - After 4-1/2 hours from start of event, 2 pumps which had lost CCW for 20 minutes were restarted and plant cooldown.
             -  Seals for all four pumps were replaced as a precautionary measure.
             -  Event confirmed capability of seals to withstand conditions similar to SBC *

• Evaluation of the effects of Station Blackout on RCP seal integrity, with respect to C-E designed NSSSs, leads to the following conclusions:

a. Seal cooling is not necessary to assure seal integrity during a SBO.

b. S80 will not result in a loss of seal function nor lead to an unacceptable loss of reactor coolant system inventory.

c. No operating plant with C-E des igned NSSSs has ever experienced a significant degradation of the seals due to SBO, loss of CCW or other incidents. Seal leakage past the vapor seal has never come near the leakage rate of concern (100 gpm) cited in NUREG-1032 .

• d. RCP seal capabilities and integrity confirmed by spec i al tests on prototype pumps and operating experiences.

e. Loss of RCP seal integrity due to SBO is not a credible event for C-E designed NSSSs.

8

4. Accident Sequence Analysis The CEOG believes that utilization of the data available (discussed in detail items 1, 2, 3, and 5) will result in the overall conclusion that the risk due to the Station Blackout Event is low and does not constitute additional requirements. To support this conclusion, the CEOG had Combustion Engineering provide ,a table (Table 1) utilizing NUREG-1032 methodology and CE designed NSSS probabilities, comparing the dominant sequences for core damage from SBO. It is clear that SBO does not make significant contribution to the risk from nuclear power p1ants.

5. Time to Core Uncover/Vessel Failure/Containment Failure Following Loss of AC Power NUREG-1032 predicted times for the subject events appear reasonable with exception of containment failure times. The detennination of containment pressure vs. time should be based on realistic treatment of the physical phenomena. This will likely give longer times to predicted containment failure (or even no failure) with attendant improvements in time available to restore AC power and reduction in source term. NUREG-1032 acknowledges the disparity with IDCOR results but does not justify utilizing a lower value

• 9

TABLE 1 DOMINANT SEQUENCES FOR CORE DAMAGE FROM STATION BLACKOUT C-E NUREG-1032 Analysis Table *7. 2 Table CS TMS 1B1 (No. 14* )/RY 1 E-8 5 E-6 3 E-6 TM/RY 0.045 0.088 .088 Li/0 2.4E-2 2E-2 2E-2 B1/D OG 1. 2E-4 8.2E-3 lE-2 AC .11 .35 .17 TML 2B2 (No. 8* )/ RY 4 E-9 1 E-5 2 E-6 TM/RY .045 0.088 .088 L2/D 0.06 .5 .5 B2/0 DG 0.8 E-4 1.1 E-3 5. 7 E-4

• AC 0. 02 . 21 . 08 TMQ2B2 (No. 6* )/RY 0 1 E-5 2 E-6 TM/RY .045 .088 .088 Q2/D e .5 .5 B2/D DG 0.8 E-4 1.1 E-3" 5.7 E-4 AC 0.02 . 21 . 08
• No. refers to sequence number in the attached Figure (extracted from NUREG CR/ 3226) 10

4 Plgure 1 GENERIC PWR EVENT TREE FOR STATION BLACKOUT 2 hr11.- 2-12hra. >12->24 hrs. I

                             .        .                                   t..J 0

w,..:a

                                                                                                                            .~:i
                                  .. J                  0           OQ II,<

f4J ". w O,<

                                  >,1,-4 II:> VI
                                                .zM Q
                                              .. JW Id 0

M a: w

                                                                  *w 0 C,OM z a:

f-4 W Id O,c

                                                                          >o!::

CII > v, 0:

                                                                                            ~ E i;       .,

u Iii Ulil

r: < z  !;! 0 ~

c.cao. zM .. Po w r. 0111:~

                                                                                                         ..,::,:z:   :z: lC
                 ~~
                   *O
                     >     M u

01-41-< u .,, Wu>o

                                                  >z
                                                *OWW
                                              ~:i~~             UZM
                                                                    )(
                                                                .,, WW E
                                                                          >ow 0 .....

u .,, It.I u >-

                                                                                            ~~a:

UUf4

                                                                                                .,, . 0 w

WO

,u Of-4 w:::,

s~

                 .,, a:    a:     a:   ,c .,, Ill a: Id t-,     c,i:Mf,t  er:<.,,           ,c a: :r:    u,          Ill 0 CLt >   (01)           (B1)      CL2)              (02)        (82)           (83)

TMBo OK 1 TMBt OK 2 II I TNB2 OK l {Succe1111 I TMB3 CD 4 ...._. Pellure S111all LOCA

                                                                                     --- - - - - -    TM02B1          OK      5 ffl02" 2        CD      6 TNL2B 1         OK     7
                                         -                                  I I                         TML2B2          CD      8 TMB~-

S111all LOCA I I TML20l91 TML202B2 OK CD

                                                                                                                             '10 S111all LOCA      TN01Bo          OK      11 TM01B1          CD      12 fflL B          OK      lJ I                                                                          1 0 I                                                               TNL1B1          CD      14 S111al l LOCI\    TNL101Bo        011:    15 I

TML10181 CD 16

H Nebraska Public Power District GENERAL OFFICE P.O. BOX 499, COLUMBUS, NEBRASKA 68601-0499 TELEPHONE (402) 564-8561 June 18, 1986 Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, DC 20555

Reference:

Comments on the Proposed NRC Rule and Draft Regulatory Guide on Station Blackout (1) 51 Federa l Register 9829

Dear Mr. Secretary:

As requested in Reference 1 , please find enclosed comments on the proposed NRC rule and draft regulatory guide on Station Bl ackout. The District concurs with these commen ts and asks your consideration of these prior to taking any action regarding issuance of a final NRC rule and regulatory guide on Station Blackout. If you have any questions concerning these comments please contact me. Sincerely, - J~u~

 ~ Vice President - Nuclear LGK:GMC:rg Enclosure

~ U s. r Dl

TABLE OF CONTENTS

• 0 Abstract 0 Execu tive Summa ry I. The Station Blackout I s sue Need No t Be Res o lved By Generic Rul e mak ing .*..*........*.* 1 II. The Te c hn ical Record Does Not Suppo rt This Generic Rulemaking . . . . . . . . . . . . . . . . . . . . . . . . . 7 II .A. Technical Reports Relied Upon in the Proposed Rule Do Not Lend Support to the Proposed Rule *.................*....*.....*.. 9 II.A.1 The Probability of a Station Blackout Is Not Clearly Established .*.*...*...**.*.....*..... 9 II .A. 2 Station Blackout Consequences Are Overstated ..... 13 II.A.3 NUREG-1032 Contains Errors and Omissions .*....... 23 II.B Additional Matters Included in the Technical Record Do Not Support the Proposed Rule .*...*.... 27 II.B.1 Clarifications of European Approach Are Necessary .......*......*......*.....*.*....*. 28 II.B.2 Clarifications of RCP Seal Integrity Are Necessary ....*.............*.......*.*....... 36 II.B.3 Clarifications of the Significance of Hurricane Gloria Are Necessary ....*.*....*...*... 38
• II.C Responses to the Additional Comments and Views of the Commission Must Be Considered ....... 42

III. The Proposed Rule Itself Should Be Reevaluated ... 48 I I I .A The Ultimate Requirements of the Proposed Rule Are Indefinite and Depend Upon the Future and Uncertain Exercise of I

• Discretion ....................................... 48 III.B The Proposed Rule Will Not Achieve a Consistent or Efficient Resolution of the Station Blackout Issues .......................... 52 IV. The Proposed Rule Cannot Meet the Backfit Rule Standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 IV.A Introduction ..................................... 56 IV.B Installation And Continuing Costs Associated With The Backfit Have Been Underestimated ........ 58 IV.C Potential Impacts on Radiological Exposure of Facility Employees Should Be Further Addressed ........................................ 64 IV.D The Relationship Between Proposed and Existing R0gulatory Requirements Should Be Considered Fu 1the r .......................................... 67 IV.E Po t e ntial Impacts of Differences in Facility Typ e , Design or Age Should Be Considered F11r t he r .......................................... 68 IV.F The Reduction In Risk From Offsite Releases To The Public Has Been Overestimated ............. 69
• IV.F.l IV.F.2 Factors Reducing the Probability of Station Blackout Require Further Attention ............... 70 Fa c tors Reducing The Consequences of Offsite Releases Require Further Attention ............... 74 IV.G Conclusion ...................................... 82

v. Objections to the Proposed P.ule .. . . .............. 84

EXECUTIVE

SUMMARY

On March 21, 1986 the Nuclear Regulatory Commission (NRC) published a proposed rule on station blackout. 51 Fed. Reg. 9829 et ~g_. Station blackout is defined by the NRC Staff in proposed 10 C.F.R. §50.2, Definitions, to mean: the complete loss of alternating current (AC) electric power to the essential and

• nonessential switchgear buses in a nuclear power plant (i.e., loss of the offsite electric power system concurrent with turbine trip and unavailability of the onsite emergency AC power system).

The proposed rule is in response to Unresolved Safety Issue (USI) A-44, which was designated as a usr in 1979, partly as a result. of the findings of the Reactor Safety Study (WASH-1400, 1975) that station blackout could be an "important contributor to the total risk of nuclear power plant accidents." 51 Fed. Reg. at 9830, col. 1. A concern was also

• expressed "that the reliability of both the onsite and offsite emergency AC power systems might be less than originally anticipated." Id. (See also SECY 85-163 at 1-2).
• To address this concern, the proposed rule creates a new section, §50.63, which would require commer~ial nuclear power plants to withstand and recover from a station blackout for a specified duration in accordance with Ge0er2l Design Criterion

E-2

• (GDC) 17. The "specified duration" is defined only in that four factors set forth in GDC 17(e) should be considered. A draft regulatory guide published in March 1986 ("Draft Regulatory Guide, Station Blackout") could be used to determine an acceptable method of determining a specific duration(~,

four hours, eight hours) for each licensed plant. Amendments to GDC 17 are proposed to make station blackout a design basis event, because it is felt that existing regulations do not

• require plants to be designed to assure core cooling and containment integrity for any specified period of loss of all AC power.

The proposed rule sets forth implementation schedules for determination of the specific durations mentioned above (270 days from the effective date of the rule) and the schedule for implementing necessary plant modifications (within two years from the NRC's notification of its findings regarding the

• acceptability of a licensee's specific station blackout duration, unless a longer schedule is justified and mutually agreed to by the NRC and a licensee).

Accompanying the proposed rule is a backfit analysis which relies in large measure on the Staff"s regulatory

          . l ana 1 ys1s. 51 ~ e d * !3_e g . at 9 8 3 3 - 3 5 . Reference is made to a 1/ "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout." NUREG-1109.

E-3

• draft regulatory guide (temporarily identified by its task 2

number, SI 501-4) which was also published in March 1986. 51 I 6 Fed. Reg. 11494. Finally, the proposed rule makes reference to 3 various suppor~ing documents. The nuclear utility industry has been actively engaged in seeking a resolution of the station blackout issue. In the spring of 1984, a number of utilities formed the Nuclear

• Utility Group on Station Blackout (NUGSBO).

utilities, and with the help of technical and legal Through its member consultants, NUGSBO reviewed existing literature and made 4 several submittals to the NRC. Subsequently, on August 28, 2/ "Draft Re gulatory Guide, Station Blackout," Task SI 501-4 (March 1986).

     }/ i.e., NUREG/ CR-3226 (May 1983), NUREG/ CR-2989 (July 1983),

NUREG/ CR-3992 (February 1985), and WASH-1400 (1975).

     !/ On May 8, 1985, NUGSBO submitted its "Proposal for Resolution of USI A-44 (Station Blackout)." The proposal contained a seven point AC reliability program, proposed policy statement and discussion of integration of related issues. on June 5, 1985, NUGSBO submitted a report entitled "Estimation of Site-Specific Station Blackout Core Damage Frequency Using NRC Staff Methodology. On July 17, 1985, NUGSBO submitted a letter to the NRC Staff further clarifying NUGSBO's proposal.      This letter addressed those points raised by the Office of Nuclear Reactor Regulation (NRR) related to:    (1) utility support of the NUGSBO proposal and for the concept of integr a ti o n; (2) NUGSBO preparAdness to e laborate on the seven p~int program; (3) the basis for existing blackout coping capability.          Each of these documents is incorporated by refe r ence.        NUGSBO also made a brief presentation to ACRS on ~0 ~ r uary 26, 1985.

NUGSBO stated that the trend i n l o ss- of -o f fsite-power (LOOP) data and di e sel generator reli a bility together indicate station blackout is not a generic issue. As a result of the (Footnote 4 Continued on Next Page

E-4 1985, the Nuclear Utility Management and Resources Committee (NUMARC) voted to establish a Technical Subcommittee to focus executive attention and resources upon hardware issues such as station blackout. A NUMARC Station Blackout Working Group was formed. NUGSBO was named technical consultant to the NUMARC Station Blackout Working Group. Based upon a review of the rulemaking material it is

• concluded that rulemaking is not warranted for the following reasons:

o First, the issue is not generic; rather, station blackout is a concern at a limited number of plants and, thus, should be resolved in a manner other than rulemaking. 0 Second, the technical record does not support the proposed rule: the risk of a station blackout (both probability and consequence) is overstated; the calculations of risk contain errors and omissions; and the reliance upon the European approach, reactor coolant pump

• a seal degradation issue and severe weather events such as Hurricane Gloria is of concern.

Third, the proposed rule itself should be reevaluated: the coping analysis requirement is not well-defined and, based on experience with other rules, has the very real potential for (Footnote 4 Continued from Previous Page) NUGSBO presentation, the ACRS indicated in a letter to the Executive Director for Operations. Mr. Dircks, dated March 12, 1985, that "if a better altc:rnati_*.**:* t 11an rulemaking is advanced, we recommend that it be given serious consideration."

E-5 confusion and misplaced efforts without commensurate benefits; the resort to a coping analysis diverts attention from the proper goal of onsite and offsite AC power reliability. 0 Fourth, the proposed rule does not meet the backfit rule requirements: the costs are underestimated and the benefits are overestimated. The basic positions encompassing the above reasons are highlighted below (parenthetical references correspond to

• sections in the comments):

The Station Blackout Issue Need Not Be Resolved __ by Generic Rulemaking (SI) The r e cord supporting the rule reflects that the number of plants o f c on ce rn is limited. The regulatory analysis (NUREG-11 09) , a ssumes that only 15 reactors need to increase di e sel ge nerator r e liabilities and 10 reactors need to increase their capa bility to cope. If the analysis supporting the rule gave adequ a te c redit for a number of site-specific risk reduction measures currently in place or available ( ~

   "blacksta rt diesels) even fewer plants would be of concern in this rulem a k i ng .

. An analysis of prior technical findin g s (NUREG/ CR-3992 ) shows th at ap p rox imately 40 ~ of th e 52 si t e s conside r ed in the r egul a t ory ana l ysis neve r experie n c e~ a I oss - of - of fs i te-p owe r event. Further , ap pr ox imat e l y 1/ 3 of t he eve nts oc c urred at 4 of the r ema ining 30 sites whi c h had experienced an event.

E-6 Since the number of plants which are of concern appears to be small, it would be more appropriate to proceed on a case-by-case basis -- focusing only on those plants at which some action may ~e appropriate to re du c e risk. The Comm ission, in the exercise of its discretion s h ould use means other than rulemaking to accomplish this. The Techn ical Reco rd Does Not Suppo rt This Generic Rulemak i ng (§ I I)

• A review of the technical reports relied upon in the proposed rule and the draft regulatory guide accompanying the proposed rule shows that the proposed rule does not have the necessary supporting technical record.

First, the magnitude of the probability of occurrence of a station blackout at any site or group of sites has not been established. The principal technical report, NUREG-1032, does not identify the actual ~requency distribution for station blackout. Instead, a set of "bins" is created in which each bin is assigned a core damage frequency defined by plant features considered important for blackout. Thus, it is conceivable that the distribution of plants could concentrate . at any point on the spectrum -- and, in fact, could concentrate at a point on the spectrum where the overal l risk is very low. However, no definite statement is made c o~c~r ning how many plants, or which plants, are in each of the bins.

E-7 Second, the technical analysis of the consequences of a station blackout event makes certain erroneous assumptions regarding the absence of containment integrity and the amount of fission product released. The assumption that core damage is synonymous with near-term breach of containment for station blackout accident sequences is inconsistent with other technical analyses. Further, the use of the Siting Source Term (SSTl) assumptions (i.e., direct containment breach and early release of fission products caused by over-pressurization in only 1.5 hours following onset of core damage) to support this proposition in lieu of SST2-5 source term assumptions (describing long-term containment failure and late release) is inappropriate. This is because the frequency of a large early release such as SSTl for station blackout sequences is so small (on the order of 10 -7 or less per reactor year) that such release is not risk-significant compared to a late release. As a result of the SSTl source term assumption, the offsite

• release consequences may be.overestimated by two or more orders of magnitude. This point is underscored by recent analyses which question the short time intervals assumed in the technical analysis to precede containment failure and offsite release. Longer time intervals would not only provide utilities with additional opportunities to ?revent releases but, at a minimum, would make the SST2-5 release categories more ~ppropriate than SSTl for a s t0ti o~ *::~ckou t. The fission

E-8 product releases associated with these accident categories are well-below SSTl values. Third, the methodologies used in the technical analysis contain errors and the plant-specific interpretation of data in the scientific literature is inconsistent. These errors and inconsistencies affect the validity of the postulated categories of loss-of-offsite-power events (LOOP), which are

• distinguished by offsite power design characteristics and emergency AC power system reliability. There is an error in the hypothesis that a relationship exists between offsite power system features and the potential for a loss-of-offsite power.

This can be ilttributed to improper assignment of durations to certain events, double counting events and adding in several non-LOOP events to the data. Further, upon reproducing the analysis it appears that there is an important statistical error in failing to account for the large sum-of-squares error

• associated with the results and, thus, the analysis cannot draw correlations with the normal statistical confidence levels.

Also, the analysis of the correlation between grid stability and weather does not entail a rigorous statistical review of plant experience. A key technical paper cited in the analysis was improperly interpreted and, if used correctly, concludes that there is no correlation between the duration of loss-of-offsite-power events and precipi tet i o n a ~o u 0t s.

E-9 Fourth, in presentations before the Commission and in response to written questions, several additional matters were I J raised in support of the rule. These matters included the European approach to station blackout, integrity of reactor coolant pump seals and the Hurricane Gloria experience. The European approach is purported to be that of coping for a longer duration than would be required for U.S. plants in the proposed rule. However, no information or technical analysis has been put into the record as a basis for this proposition. Thus, detailed comments cannot be provided on this proposition. It appears that central to the European approach is the capability to provide sufficient water inventories for makeup purposes. Under this logic, a large number of U.S. plants could cope for the long durations attributed to the French design. It is questionable as to whether any of the European plants could meet the requirements

   ** of the proposed rule and the technical guidance in the regulatory guide, particularly the equipment qualification aspects. Failure to demonstrate that French designs could comport with the terms of the proposed rule calls into question any reliance upon the European approach as a benchmark for U.S.

plants.

E-10 The proposed rule also links the ability of a plant to cope with a station blackout to the issue of reactor coolant I j pump (RCP) seal integrity. With respect to station blackout, however, only leakages of 100 gpm per pump or more are of concern because of their impact on limiting decay heat removal. This criteria eliminates from consideration all boiling water reactors and plants operating pumps with hydrodynamic seals. Thus, reactor coolant pump seal integrity is only relevant to

• the Westinghouse pumps with hydrostatic seals (due to the larger seal leak potential). The concern over the Westinghouse pumps is approaching resolution and may be resolved as a result of a May 8, 19A6 submittal of a revised Westinghouse report (WCAP-10541, Rev. 1). Thus, it does not appear that high pressure AC-independent injection pumps are needed to provide seal injection and cooling. The impact of this discussion highlights the fact that concerns over RCP seal integrity do not contribute to the overall station blackout concern .
• Hurricane Gloria, the September 1985 Hurricane which moved up the East Coast from Florida to New England, has been cited as an example of potential weather-related causes of loss of offsite power which supports rulemaking. However, the actions by utilities in anticipation of the hurricane to place plants in safe condition during the event demonstrates that effective measures could reduce th~ se ver~ ~e ather concern.

E-11 The Proposed Rule Itself Should Be Reevaluated (§III) The ultimate requirements of the proposed rule are indefinite and depend upon the future and uncertain exercise of discretion. Neither the codified rule nor the draft regulatory guide establishes bounds for the required coping demonstrations. Under the rule as proposed, licensees cannot ascertain the ultimate requirements they will be expected to

• meet or the potential plant modifications they will need to make to satisfy the Staff of their compliance.

The proposed rule will not achieve a consistent or efficien t r e solution of the station blackout issue. The rule does not fo c us on reducing the likelihood of a station blackout but, inst e ad, focuses on coping duration. The rule allows for a range of c oping from zero to eight hours. A more practical approach to meeting this objective of reducing the likelihood

• of station bl a ckout consequences is to focus on an AC power reliability program.

The Proposed Rule Does Not Meet the Backfit Rule Standard (§IV) 10 C.F.R. §50.109, the Commission's backfitting standard, provides that a proposed backfit must resul t in. . a substant i al in c r e ase in overall pr ot~c ti 0~ ~f th e public heal t h and safet y . . and that the . . c osts of implementation for

E-12 that facility are justified in view of this increased protection . . . . " 10 C.F.R. §50.109(c) lists the nine (9) I , specific factors which are to be addressed as appropriate. The backfit analysis provided in the record to support the proposed rule falls short of demonstrating the justification for this proposed rule for the following reasons .

• First, the direct and indirect costs of the proposed backfits are underestimated. The number of plants that would be significantly affected by the proposed rule is understated in the regulatory analysis. This alone would significantly increase thP costs. Further, the costs associated with activities at individual plants are based upon standard commercial labor productivities rather than information available for the nuclear industry. Handbooks for cost estimating developed under NRC auspices were not used. Also,
• the projected costs for a coping analysis, based on both actual figures incurred by utilities and reasonable utility estimates, are low by a factor of 5 to 20. This underestimation is consistent with Staff projections in other rulemakings.

Second, the record does not appear to c onsider the

• potential impact of differences in facility type, design or age, despite the site-specific natu r 0 o f th e issue.

E-13 Third, the potential impacts on radiological exposure of facility employees should be reassessed. The regulatory analysis asserts that no significant increase in occupational exposure is expected. This appears to be drawn from the belief that there will be no worker exposure if the backfit does not specifically involve the "reactor coolant system." This is not the case. The types of hardware fixes that could be contemplated include replacement of valve operators, new

• instrumentation and surveillance. A recent study indicates that up to forty percent of the total occupational exposure at light water reactors was attributable to NRC initiated multi-plant actions.

Fourth, the regulatory analysis on this record should further consider the relationship between proposed and existing regulatory requirements. This consideration is mandated by 10 C.F.R. §109. Although the regulatory analysis acknowledges the relationship between station blackout and the two issues of diesel generator reliability and reactor coolant pump seal integrity, it assumes resolution of these items would have little impact. However, to the extent there is improvement in the diesel generator reliability, the benefits claimed from the proposed rule are dissipated. Also, and despite the fact the regulatory analysis assumes small leakage, the RCP seal issue has been a motivating force in the propo~~d sta tion blackout

E-14 rulemaking. Recent information makes it apparent that the risk of catastrophic seal failure is significantly less than thought. Thus, resolution of this issue should de-emphasize the need for a rule. Fifth, the reduction in risk from offsite releases to the public has been overestimated. This overestimation is attributable to a failure to account for flexibility in

• emergency AC power system designs, emergency procedures and additional power sources when determining the probability of occurrence of a station blackout event. Further, use of the SSTl fission product release assumption rather than SST2-5 release assumptions is inappropriate. The regulatory analysis makes no s pe cific statement as to time of containment failure except to assume the SSTl scenario of 1.5 hours after onset of core damage. However, the regulatory analysis contains a recognition that this assumption is erroneous since SSTl
• releases are reduced by 1/3.

public. The time to containment failure, if any, is a significant parameter in calculating doses to the Because containment integrity is not properly credited, the dose to the public has been overestimated. Also, evolving source term information is ignored which calls into question tho ussumptions contained in the r?gulatory analysis.

E-15 Q~j~ctions to the Proposed Rule (§V) The effectiveness of the rulemaking tool in addressing so isolated an issue as station blackout is questioned. However, it is recognized that the Commission may believe that rulemaking satisfies other imperatives which transcend the ability to demonstrate public health and safety benefits across all of industry. Should such imperatives exist within the

• Commission, then it is believed that the proposed rule should focus on the concerns raised. Accordingly, instead of the proposed rule with its focus on coping, guidelines should be prepared concerning what constitutes acceptable emergency and non-emergency AC power systems for preventing or mitigating station blackout events. Without such guidance, a rule which requires AC power independent coping is subjective, difficult to implement and of questionable value with respect to providing real safety margins .
• A shift in focus from compensatory measures to problem resolution would entail modifying several aspects of the proposed rule. To this end, the kinds of changes suggested include:

(1) expanding the focus of the rule and draft regulatory guide to include e x plicit and equally weighted provision for c r editing onsite and offsite bac~ u ..o AC no~~ r sour c es or improvements to AC \:)O**r'= r ::; :1* s ~ *2 rn availability in responding to the loss of normal and emergency AC power;

E-16 (2) deleting the addition to the General Design Criterion of an AC-independent coping requirement and the associated coping analyses; (3) eliminating the open-ended coping demonstration aspects of the regulatory guide and substituting a simpler closed-form coping capability checklist; (4) considering the potential ripple effects of new station blackout requirements, particularly those involving AC-independent coping, on other areas of nuclear regulation, such as potential equipment

• (5) qualification of station blackout coping equipment by the methods called for in 10 C.F.R. §50.49; and eliminating the need to consider severe weather as a dominant factor in establishing whether a plant is in a 4 or 8 hour coping duration category (NUREG-1109, Tables 1 & 3) .

I. The Station Blackout Issue Need Not Be Resolved By Generic Rulemaking The Commission need not pursue generic rulemaking in order to resolve a non-generic issue. In the proposed station blackout rule, the number of plants of concern is acknowledged to be limited. First, in preliminary screening to determine the existence of "any plants with especially high risk that might require further analysis or action on an urgent basis",

• none were identified.!/ Second, the regulatory analysis (NUREG-1109, January 1986) states (at 9):

It was also assumed that all plants, as currently designed, can cope with a station blackout for 2 hours, and, with proper procedures and training, plants could cope with a 4-hour station blackout without having to make major modifications. (Emphasis added.) (See also 11.)l/ Third, concerns over station blackout risks are focused on the 14 reactors listed in Group C in Table 4 of NUREG-1109. This

• is evident from a review of Table 2 which defines Group C plants, and Table 5, which lists the estimated core damage frequency per reactor-year of these plants as 15 x 10 -5 . The next group of plants in Table 5 receive only 10% of the benefits attributed to implementation of the proposed rule at Group C plants. Further, the risks associated with all other

 ! / NUREG-1032 (May 1985) at 2-3.

ll The initiatives advanced by the industry will provide for station blackout coping procedures.

plants set forth in Table 5 are at or near the goal sought by the proposed rule. Fourth, in Table 6 of NUREG-1109, only 15 reactors are identified as needing to improve diesel generator reliability and only 10 as needing to increase their capability to cope.}/ In the proposed rule, the highly plant-specific nature of station blackout is further acknowledged in connection with the ability of power plants to cope. In listing several "factors" as the "main contributors to risk of core melt resulting from station blackout," the Notice of Proposed Rulemaking states (51 Fed. Reg. at 9830, col. 3 - 9831, col. 1): These factors . . . vary significantly from plant to plant because of considerable differences in design of plant electric power systems as well as site specific considerations . . . . Other points support a position that station blackout is not a generic issue which should be resolved by rulemaking .

• For example, a major contributor to the overall risk of station blackout is diesel generator reliability. Station blackout ceases to become a generic issue if diesel generator reliability is high at many or most plants. Such is the case, in that the individual diesel generator reliability for most plants currently exceeds the minimum levels called for in this proposed rulemaking. See 51 Fed. ~e9 . at 9830, col. 3 and

 ~/ The initiatives advanced by industry include a program for increased diesel generator reliability.

NUREG-1109 at 6-7 and at Tables 1 and 6. Therefore, as to the majority of plants, no action need be taken under the proposed rule to assure greater diesel generator reliability. Similarly, the contribution of loss-of-offsite-power to station blackout risk is not spread evenly among the operating sites. A review of NUREG/CR-3992 (1985) shows that approximately 40% of the 52 sites representing 67 reactors . (i.e., 22 sites) considered in the regulatory analysis (NUREG-

• 1109) never experienced a loss-of-offsite-power event.

Moreover, approximately 1/3 of the ~oss-of-offsite-power events occurred at 4 of the 30 remaining sites. See~-, NUREG/CR-3992 at §5 and Appendix. Therefore, rather than being roughly the same at all plants, or even a majority of plants, the contribution of loss-of-offsite-power potential to station blackout risk is concentrated at only a handful of sites and is not significant elsewhere.ii

• Beyond this, the proposed rule does not reflect adequate consideration for a number of site-specific factors which reduce the probability of a station blackout event. For example, no credit is given for the availability of additional Recent information developed by the Nuc lg ar Safety Analysis Center indicates that, for all years th r ough 1985, fewer losses-of-offsite-power may have occurred than is indicated in NUREG/ CR-3992. NSAC-103, "Losses of Off-Site Power at U.S. Nuclear Power Plants," Apr il 198 5 . a t 2-13 (indicating that 51 of 65 sites have not had a total loss-of-offsite-power longer than 30 minutes; 38 of 65 sites have never had a loss-of-offsite-power).

sources of AC power.~/ The proposed rule also does not credit the ability of some utilities to reconfigure their diesel generators and to shed non-essential loads so that they are effectively in a different category of plant diesel generator configuration.~/ Finally, the ability of utilities to manually As an example, the Indian Point site has gas turbines available in the unlikely event that all offsite and onsite emergency power is lost. The H.B. Robinson plant can rely upon its fire protection diesel to provide emergency AC power. In the case of Florida Power and Light Company's

• Turkey Point plant, by design the units are configured so that there are a total of two safety related emergency diesel generators supported by five non-safety related black-start dies~ls of similar capacity. The Turkey Point plants (Units #3 & #4) are located on a two-unit nuclear site adjacent to two conventional fossil (oil and/or gas) fired plants (Units #1 & #2). Each unit requires one out of two diesel generators in order to meet all essential loads during a design basis LOCA event. Also by original plant design, both diesel generators are aligned such that either unit's safety related emergency buses can be powered by either diesel generator from their respective trains.

Additionally, any two of the five black-start diesels can be used to power each unit's emergency buses (via dedicated cable, independent of station switchyard) during a blackout condition within 30 minutes of the initiating event. The Turkey Point site, by original design, also has the capability to provide both steam and electric power (via the switchyard) from the fossil-fired units to the nuclear units. This power can be used to

• power safety related equipment.

 ~/ Each diesel at Dresden is capable of either parallel or
    . independent operation. Likewise, the emergency 4 kV bus system is designed to operate either sectionalized or as a ring bus between both units. Procedures currently exist to reduce the AC loads to only those absolutely necessary for safe shutdown of the plant. In the event of loss-of-offsite-power, with the use of the isolation condenser, condensate storage, service water, and control-rod-drive hydraulic systems, both units can achieve and sustain hot shutdown conditions with one diesel generator. Quad Cities has a similar capability using the re?ctor core isolation cooling (RCIC) and RHR systems in lieu of an isolation condenser. As a result, the Dr esden and Quad Cities plants, which are characterized as a 2 out of 3 diesel generator configuration, would actually serve during a station blackout event as a 1 out of 3 diesel generator configuration.

start diesel generators which fail to start automatically is not fully taken into account.I/ Since the number of plants of concern is limited, the issue is not generic and rulemaking shoul d not result.~/ It is recognized that a basic premise of administrative law is that agencies such as the NRC may impose new requirements either by promulgating generic rules or on an individual plant basis. DAVIS, ADMINISTRATIVE LAW TEXT chs. 6 and 8 (1972). Further,

• it is well-recognized "[that] the choice between rulemaking and adjudication lies in the first instance with the [agency's]

discretion." NLRB v. Bell Aerospace Co., 416 U.S. 267, 294 (1974). However, the line dividing the choices may not always be a "bright one." united States v. Florida East Coast Railway, 410 U.S. 224, 245 (1973).~/ In this instance, the 11 Procedures exist at a number of plants which permit manual diesel starts in a matter of minutes, should these sources fail to start automatically. It would be more efficient to proceed on a case-by-case basis focusing only on those relatively few plants where actions are judged to be necessary to minimize the risks posed by station blackout.

   ~/ For example, one court recognized that:

This simple proposition, however, is incapable of being reduced to mathematical terms. One cannot say, for example, that an issue which affects 15 % of the nuclear plants in the country should be resolved generically, while one which affects only 10% of the nation's plants is inappropriate for generic resolution. Deukmej ~an v. NRC, 751 F.2d 1287 (D.C . Cir. 1984).

Commission should exercise its discretion and refrain from promulgating the proposed station blackout rule. The Commission recognized in its 1986 Policy and Planning Guidance (PPG) that: NRC must be sensitive to the large number of requirements imposed on licensees. Requirements imposed on the regulated industry by NRC are to provide a positive contribution to the public health and safety . . . There should be no unnecessary regulatory burdens._!Q/

• This guidance is important in the context of station blackout.

First, new requirements should not be imposed on licensees unless they are necessary. Second, orders to individual plants are a viable means of addressing issues of concern at those few plants. In short, the Policy and Planning Guidance makes clear that when the focus of a perceived issue is confined to only a few plants, the resources of the NRC should be selectively applied .

• In sum, a rule is not required to resolve issues at those few plants where station blackout is of concern. The need for any generic rule which would change a General Design Criterion to address highly site-specific issues is questionable.

10/ PPG at 23.

II. The Technical Record Does Not Support This Generic Rulemaking The proposed rule is "intended to provide further assurance that a station blackout . . . will not adversely affect the public health and safety."l.!_/ The rule would provide this assurance by requiring that plants cope with a station blackout for some period of time. This duration would not be specified in the rule itself. Rather, the selection of

• l.!_/ 51 Fed. Reg. at 9829, col. 1. The Notice of Proposed Rulemaking contains the assertion that in a "few cases" there have been complete losses of both onsite and offsite AC power systems, and, in these cases, "AC power was restored in a short time without any serious consequences."

Id. at 9830, col. 1 (emphasis supplied). While station bTackouts have occurred at several plants, they either have involved momentary loss of all AC power or have occurred while the plant was not engaged in normal operations. There have apparently been four events involving brief simultaneous losses of offsite and onsite AC power. In 1968, there was a loss of offsite power event at Connecticut Yankee. The diesel generators started and loaded. Due to a switching error an erroneous signal caused a "stuck breaker protection scheme," tripping the diesel generators. There was no AC power for a period of 4 minutes. In 1976, Millstone 2 suffered a loss of offsite power. The diesel generators started, but when a large 1* pump was being started, an undervoltage trip occurred. AC power was lost for 5 minutes. In 1983, Fort St. Vrain experienced a loss of offsite power while in cold shutdown. One diesel generator was out of service. The other diesel had been started and loaded, but tripped due to overloading when the loss of offsite power occurred. AC power was restored after 25 minutes. Finally, in 1984, during a test simulating a loss of offsite power event at Susquehanna, Unit 1, there was a switching error by an operator, resulting in the diesel generators failing to start automatically and having to be started ~2nually. It took 11 minutes to restore AC power. As all of these incidents indicote. t~ere has been no event rluplicative of a sustainld station blackout, the concern of the proposed rule. Rather, there have been events which resulted from testing and switching errors under circumstances for which there is little concern about sustained losses of AC power.

a 4-hour or 8-hour duration would be the result of application of methods provided in a regulatory guide. 50 Fed. Reg. at 9830, col. 3. In addition to this requirement, improved guidance will be provided to licensees regarding maintaining minimum emergency diesel generator reliability to minimize the probability of losing all AC power. Id. The technical bases for the proposed requirement are the result of information developed by the Staff and by NRC contractors in the course of their studies of the station blackout issue, including oral statements of the Staff made at Commission meetings and Staff responses to specific Commissioner's questions. A review of this material, and the below provided responses to the additional comments and views of the Commission accompanying the proposed rule, shows that the technical record does not support the rule. Technical comments fall into three categories: Technical Reports Relied Upon In The 0 Proposed Rule Do Not Lend Support to the Proposed Rule (§II.A below);l2/ o Additional Matters Included in the Technical Record Do No t Support the Proposed Rule (§II.B below); and o Responses to the Additional Comments and Views of the Comm is sion Must Be Considered (§II.C below). The Supp le menta ry information portion of the Notice of P roposed Rulemak i ng highlights the major results of the Staff's tec hnical studie s on the stat io n blackou t iss ue. 51 Fed. ~1:3

• a t 9830, cols. 2 and 3.

II .A. Technical Reports Relied Upon In The Proposed Rule Do Not Lend Support to the Proposed Rule A review of the technical reports relied upon in the proposed rule (i.e., WASH-1400 (1975), NUREG-1032, NUREG/CR-3226 (1983), NUREG/CR-2989 (1983) and NUREG/CR-3992 (1985)), and the draft regulatory guide accompanying the proposed rule (March 1986),_!_l/ reveals that station blackout risks have been overestimated. This position is based upon the following conclusions regarding the technical bases for the

• proposed rule:

(1) The Probability of A Station Blackout Is Not Clearly Established (§II.A.l below); (2) Station Blackout Consequences Are Overstated (§II.A.2 below); and (3) NUREG-1032 Contains Errors and Omissions (§II.A.3 below). II.A.1. The Probability of a Station Blackout Is Not Clearly Established The principal document in support of the proposed rule is

• NUREG-1032. It identifies the various factors and plant features which affect the estimated frequency of core damage resulting from station blackout events. The principal result from NUREG-1032 upon which the proposed rule is based is that the potential variability of estimated station blackout likelihood and core damage frequency is large, depending on the precise combination of features which mav ~xist at a given site .

 .!l_l The proposed tule provides the title of each of these documents. 51 Fed. RE:_3. at 9830, cols. 1 and 2.

However, NUREG-1032 does not identify the actual frequency distribution for station blackout. Instead, the analysis presented in NUREG-1032 creates what can be characterized for discussion purposes as a set of "bins." Each bin is assigned a core damage frequency defined by the factors and plant features considered important to station blackout ( ~ , emergency diesel generator reliability and offsite power design characteristics). Bins at the high end of the core damage frequency spectrum are characterized by the worst

• possible set of plant features (which maximizes frequency).

Similarly, the low end of the spectrum is defined by the best combination of factors (which minimizes frequency). Thus, the core damage frequency range reported in NUREG-1032 actually constitutes the universe of potential values which could exist. The problem with this approach is that the analysis is never completed. That is, after having conceptualized the factors important to the probability of a station blackout and creating the core damage frequency spectrum, NUREG-1032 does not identify how many plants, or which plants, are in each of the bins. It is conceivable that the distribution of plants among the various bins could concentrate at any point on the spectrum. Thus, overall risk could be very high, very low, or somewhere in between. However, without making that 0~sig11ment based on plant features, the current probability of occurrence of a station

blackout posed by U.S. nuclear plants is simply not established beyond a mere generalization of what it could be._!_i/ Thus, the premise that this generic rulemaking is necessary to reduce a current unacceptable risk to the public health and safety is not adequately supported by the analysis in NUREG-1032. Put another way, a specific resolution for an issue has been proposed before determining whether the issue really exists. This is unlike equipment qualification, which arose in

• response to pre-1980 test results at Sandia National Laboratories. This is unlike Appendix R, which was premised upon another set of Sandia tests, and subsequent Appendix A (i.e., Branch Technical Position 9.5-1) fire reviews at the majority of plants. This is unlike emergency planning, which arose in response to the experience of TMI. No information is presented to cast doubt on the size of the overall risk associated with nuclear plant operation or that such risk is now unacceptable .
• The information necessary to assess the probability includes data on emergency AC power system configuration, emergency diesel generator reliability, switchyard design features, weather experience, utility procedures currently in place for responding to a loss-of-offsite-power and station blackout event, and the ability of various plant systems to operate without any AC power lonsite and offsite). This information is currentlf available for estimating plant-specific station bl~c 1;out risks. Because this information was not processed in NUREG-1032, the risk of station blackout cannot be established on an overall basis or with respect to any specific plant.

Absent specific information, the proposed rule appears simply to be a response to the early WASH-1400 study and other PRA's, which show station blackout to be an "important" contributor to risk -- a risk that is "small" (51 Fed. Reg. at 9830, col. 1) and which is already acceptable.15/ Emergency diesel generator reliabilities also may have been a contributing factor. 51 Fed. Reg. at 9830, col. 3. However, Electric Power Research Institute (EPRI) studies show the industry average diesel reliability to be approximately 0.975,

• which exceeds the 0.95 goal in the regulatory guide accompanying the rule (see draft regulatory guide at §C.1.2).

Further, the proposed rule raises the specter of growing grid instability in the future. However, no evidence is presented in support of such speculation. To the contrary, since General Design Criteria 17 and 18 went into effect in the early 1970s, the overall site frequency of loss-of-offsite power in any calendar year has declined significantly. NUREG-

• 1032 acknowledges (at A-12) that such improvements are evident in the "plant-centered" category after 1978. Further, the Staff stated at the November 14, 1985 Commission meeting (Transcript at 19) that:

NUREG-1032 estimates that the frequency of station blackout events in the future will be better than the past experience has shown . . . [W]e recognize that there have been some trends showing improvements in loss of offsite power experien c e. tha t we don't 15/ See discussion below (§II.A.2) of WASH-1400 results in contrast to other more recent Staff technical information.

expect as many losses as have occurred in the past. This statement confirms the conclusion which found that, taking the recent loss-of-offsite power data, such events were decreasing (even as the number of plants has increased) due to the knowledge and experience gained by utilities since the inception of commercial nuclear power operation.16/ In sum, when taken as a whole, the above facts suggest

• that the motivation for proceeding toward a station blackout final rule cannot be linked to evidence of degrading plant experience over the years.17/ Thus, since it has not been clearly established that the probability of station blackout events is unacceptably high, the Commission should not promulgate the proposed rule.

II.A.2 Station Blackout Co~_s_~_guences Are Overstated~/

• 16/ See NUGSB-85-005, "NUGSBO Comments on NUREG-1032," October 1:11 1985 and Revision 1, November 1985 at Appendix B, which is incorporated by reference into these comments.

The initiatives advanced by the industry will provide reasonable assurance that AC power availability will not degrade in the future.

 ~ / The comments on the proposed rule contai_ned in this section also form the basis for various positions regarding the published backfit analysis (§IV).      In §IV below, the comments focus on the overestime~ed h nefit or risk-0 reduction value of the proposed rule 2s e xpressed in terms of averted person-rems. In this section, however, the comments focus on the underlying technical basis for the conclusion that station blackout risks are substantial enough to be adrlressed via the current proposed generic rulem.:ik i.ng.

The technical findings in support of the rule tie the consequences of a station blackout event to certain assumptions regarding the absence of containment integrity and fission product releases, (i.e., core damage is synonymous with near-term breach of containment as discussed below). Namely, the assumption is made that one of the so-called "Siting Source Term" fission product release categories, the SSTl category, is appropriate for purposes of modeling the

• generic characteristics of the consequences of all station blackout events for all types of containments.19/

1109 at 8. See NUREG-In contrast to the regulatory assumptions for "a maximum credible accident" postulated for siting studies under 10 C.F.R. §100 -- in which containment integrity is assumed to exist throughout the accident (see NUREG- 0771 at 7), the SST! release category assumes a loss of all installed safety features and a severe direct breach of containment at 1.5 hours following onset of core damage -- as a result of a "core melt"

• accident and containment failure due to "overpressure." See NUREG/CR-2239 at Table 2.3.1-2 (Novemper 1982) which describes the characteristics of the SSTl-SSTS release categories.

The SSTl fission product release category, and companion categories SST2-SST5, are described in variou s documents, including NUREG-0771 (at 6-8) (June 1981). In sum, the categories reflect the conce ptual spectrum of potential consequences from severe accidents resulting from some form of core damage and fission product release, ranging from limited consequences (SST5) to maximum consequences (SST!).

The above-mentioned assumptions are inappropriate for purposes of estimating the consequences of station blackout. First, the assumptions underestimate the value of containments I

• in mitigating reactor accidents. They do not address the remedial effects of timely restoration of AC power (i.e., to restore to service those systems which will help to assure that containment integrity is preserved). The technical findings fail to reconcile the large disparity between estimates of time to containment failure in NUREG-1032 (at Table 7.3) and those
• of IDCOR -- despite the recognition that the more recent IDCOR time estimates "may be cause for revision" of the results in NUREG- 1032.20/ Had the disparity been reconciled, the opposite conclusion might have been reached regarding current plant coping abilities and the need for the proposed rule.

Second, the effect of choosing this early containment failure mode (as opposed to a late failure mode) is to place inappropriate regulatory focus on a non-dominant risk scenario I -- with an extremely low probability of occurrence. While an early release scenario could be postulated to result from mechanisms such as a steam explosion or direct heating, the frequency of such a release is so small (on the order of 10- 7 or less per reactor year) that it does not contribute significantly to the overall _risk from station blackout (frequency x consequences).21/ In addition, in its Technical 20/ See NUREG-1 032 at 7-15, 16. 21/ The technical findings maintain the "typical" estimated (Footnote 21 Continued on Next Page

Summary Report (November 1984), IDCOR states that "[t]he necessary conditions for a steam explosion sufficiently energetic to cause primary system or containment failure cannot occur" in a commercial U.S. nuclear power plant. Instead, the risk-dominant sequence is long-term overpressurization of containment long after core melt (!:...:._9_., one to three days). Third, the implication of improper selection of SSTl is significant. Each successive category SST2-5 has progressively

• lower offsite consequences. In fact, the range of offsite consequences spans about 7 orders of magnitude.22/ According to NUREG/CR-2723 (at 2-11, 12), the SST2-5 categories would not be expected to produce "substantial numbers of offsite consequences compared to the SSTl source term." Thus, the (Footnote 21 Continued from Previous Page) frequ~~cy of core damage resulting from a station blackout is 10 per reactor year (NUREG-1032 at 1-3). NUREG-1032 al~o states the assumption (at 7-19, Table 7.4) that the containment failure mode associated with the SSTl release ijategory is a steam explosion," with a probability of 10 22/ According to NUREG-0771 (at 15):

The difference in released source term between a Group 1, 'worst case' release, and a Group 2,

         'spray functional' release, is about~ factor of 100. The source term difference between a Group 2 release and a Group 3, 'melt through' accident, is about a factor of 10. The difference between Group 3 and Group 4 is about 2 factor of 1000, and between Group 4 and Group 5 of another factor of 10.

choice of SSTl rather than SST2-5 may have resulted in a significant overstatement of the risks of station blackout.23/ Fourth, it is significant to note that the weakness in relying on the SSTl release category is recognized, since a reduction factor of one-third is applied to the SSTl release (at a 50 mile radius). See NUREG-1109 at 8. This modification is made in order to account for the fact that "[i]f a core melt resulted from station blackout [i.e., as opposed to the types

• of severe accidents for which the prompt containment failure assumption of which SSTl is appropriate], containment failure would be delayed for a number of hours." Id. (emphasis added).

No rationale is provided as to why the application of a simple reduction factor of one-third to the offsite release characteristics of the SSTl category is appropriate in lieu of choosing from categories SST2-5 -- a choice which, significantly, would lower the person-rem offsite consequences, not by a factor of three, but from one to four orders of magnitude. Further, it is unclear how the straight- forward application of a multiplier on the person-rem offsite consequences associated with the SSTl release category is a technically sound approach for purposes of simultaneously correcting the (1) incorrect containment failure mode and timing characteristics and (2) the incorrect fission product inventory characteristics of SSTl. 23/ If the use of SSTl was supplanted by the use of SST2 or SST3, the person-rem calculation would be lower by 1 to 4 orders of magnitude. NUREG/CR-2239 at Table 2.3.1-3.

A better approach to risk estimation would have been to continue the analysis of accident sequences in §7 of NUREG-1032 using a model which would correctly match containment failure modes on a mechanistic basis with the various accident sequences. Instead, (1) the accident sequence analysis in NUREG-1032 was simply performed to the point where core damage is predicted (see NUREG 1032 at 7-1), (2) core damage was assumed synonymous with core melt and a precursor certain to result in near-term containment failure (Id. at 7-5) and, (3)

• the SSTl release category and its attendant containment failure mode characteristics (i.e., direct breach of containment at 1.5 hours after the onset of core damage) was assumed appropriate for representing offsite consequences (see NUREG-1109 at 8).~/

The failure to more fully and carefully substantiate the selection of the SSTl release category -- which is the most severe release category which can be identified for purposes of predicting both large offsite releases (as consequences) and large averted person-rems (as benefits) of the proposed rule-- serves to demonstrate that the technical record does not support the proposed rule. 24/ The ACRS has previously commented that the equating of core melt with fission product release is ill-advised. As recently as March 28, 1986, at the ACRS meeting on Safety Goals, Professor Okrent stated that ACRS has "[n]oted the need to be alert to the fact that core ~elt frequency may be a poor measure of the frequency of release of a significant amount of radioactive material from cont a i nm en t. " Mee t i n g Tran s c r i i::i t ,3 t 1_ n . r u r the r , the Staff has elsewhere stated that ** [ s] '::'****:'re core damage sequences would not necessarily involve complete core meltdown." NUREG-0772 at 3.16 (June 1981) (citing Three Mile Island as an example of a beyond-design-basis event which did not lead to core melt-through and containment failure).

Fifth, the SSTl release category may have been chosen to correspond to the dominant station blackout accident sequences which require AC recovery in 1 to 2 hours to avoid core damage. These accident sequences are summarized on Tables 7.2 and C.1 of NUREG-1032. They include the "TML B " sequence for all PWRs 1 1 and the "TMu 1 s " sequence for all BWRs.25/ Thermal hydraulic 1 analyses are cited for these early core cooling failure types of accidents which estimate that, following these station blackout sequences, core uncovery (i.e., onset of core damage)

• would occur in 1-2 hours and reactor vessel melt-through would occur in another 1-2 hours. See NUREG-1032 at C-3. The selection of SSTl, which assumes containment failure in 1.5 hours after the onset of core damage, could be attributed to these thermal hydraulic studies. The citation to the studies and explanations of timing tend to expose an underlying technical premise of the proposed rule, namely, that it is appropriate to equate the onset of core damage with vessel melt through and containment release .
• In this regard, the above-mentioned PWR and BWR dominant accident sequences and their specific characteristics (~~.,

thermal hydraulic analysis) are inappropriate to support the choice of SSTl as the release category which is generically reflective of station blackout risks. More recent analysis of 25/ See NUREG-0771 at 8-10 and NUR~G.*CR-2-:.~.3 ut 2 for other examples and explanations of these sequences and the alphanumeric terminology. See n.26 below for an explanation of a specific secTuence, the "TMLB' II sequence.)

these same station blackout sequences indicate that the technical findings and assumptions in NUREG-1032 and NUREG-1109 may unrealistically predict the occurrence of rapid containment failure in connection with station blackout. For example, EPRI has performed a reanalysis of the TMLB' accident sequence26/ for the Surry Plant (subatmospheric containment) and compared it to specific WASH-1400 analysis of this accident sequence (from which the findings regarding the timing of containment failure are apparently derived27/ ). EPRI NP-4096, "Surry

• Source Term and Consequence analysis," June 1985. The EPRI calculations of a Surry TMLB', which utilized the post-WASH-1400 MARCH-2 computer code, predicted a much lower rate of The "TMLB'" sequence is characterized by loss of all AC power and reactor coolant heat removal including failure of emergency core cooling and AC-powered containment heat removal systems. The general features are described in Appendix I to WASH 1400 to involve a transient event (T) in combination with loss of both main feedwater (M) and auxiliary feedwater flow (L) to the steam generators, and the loss of all AC power (B) (onsite and offsite) for a period long enough to cause core meltdown. In this preconceived scenario the likelihood of early containment failure by overpressure would be "potentially high" and the consequences "potentially severe." NUREG-0772 at 3.20.

It is useful to compare the different release category conventions to demonstrate that the timing of containment failures is consistently represented in NRC documents. The SST! release category utilized in technical documents on station blackout(~., NUREG-1032, NUREG/CR-3226, NUREG-0772) is comparable to WASH-1400 (NUREG-75/014) category PWR-2. See NUREG-2723 at 2. In WASH-1400 (at Appendix VI, Table vr --r=-1), the PWR-2 release category assumes a time of containment failure of 2.5 hours from the time the accident se~~nce is initiated, (i~ -, c o~pa~ able to SSTl which uses 1.5 hours f rom onset of core damage coupled with the Staff assumpt ions that core damage would begin between 1-2 hours after accident initiation).

pressure buildup during the early i;tages of the accident.28/ In fact, while the simplified calculations of WASH-1400 predicted a containment pressure of 77 psia at 1.5 hours and containment failure by overpressure at 3.67 hours, the updated EPRI analysis predicted that the containment pressure at 1.5 hours would not exceed normal atmospheric pressure, and, further, that there would not be overpressure failure for "at least the first 12 hours of the accident." Id. at 3-1, 3-2. The EPRI analysis suggests that the technical findings

• regarding accident sequences may be incorrect, and, at a minimum, do not support the selection of the SSTl release category._?_~ /

Sixth, the proposed rule's association with WASH-1400 (see 51 Fed. Reg. at 9830, col. 1) technical findings regarding the amounts of fission products released and containment performance (i.e., failure modes and timing of release) is itself cause for concern that station blackout risks have been

• overestimated. For example, NUREG-0772 states (at F.3) that fission product deposition in the primary system, the occurrence of which tends to reduce fission product releases in 28/ EPRI NP-4096 at 3-1. The application of the MARCH-2 computer code utilized Surry plant desi1n parameters listed in WASH-1400.

   ~~ / In NUREG-0956, a SJ.milar c onch1 s i on '*r,} 1.- eached in a reanalysis of Surry containme nt re rf *J~m a nce during severe accidents. The NUREG states (a t C.1-9) that (c]ontainment failu[e within the first!~."~ .b ~~~r_s is quite unlikely" (emphasis supplied).

L__

the event of containment failure, ""[c]ould not be evaluated in WASH-1400 because of' lack of applicable models" and that "significant extensions" beyond WASH 1400 methodologies are now I

• possible.iQ_/

Such improved methods have been used for risk calculations related to containment performance in other areas, but not for station blackout.31/ In NUREG- 0956, Appendix C.2 provides the Executive Summary of the Containment Performance

• Working Group (CPWG) Report (i.e., NUREG-1037). The CPWG studied containment failure modes as part of the overall NRC severe accident research program. Specifically, the CPWG considered containment leakage as a function of time and the impact of containment pressure relief on the mode and timing of containment failure. In contrast to the WASH-1400 results, the CPWG found that, in most cases, containment integrity would not be challenged "until several hours after vessel failure."

NUREG-0956 at C.2-11. Thus, a characteristic feature of the

• SSTl release scenario as applied to station blackout events 30/ NUREG-0956 describes (at 8-1) the "two major deficiencies" in WASH-1400 which are "often cited as causes for overestimation of radioactivity released from a plant during a severe accident [ ~ . , station blackout]" . .

One was WASH-1400's treatment - of iodine as elemental iodine rather than the less volatile, more soluble salt, cesium iodide. The other was the omission of natural processes which would retain radionuclides, particularly in the reactor coolant system. l!/ NUREG-0956 also explains (at 8-1 ) that the BMI-2104" suite of co mput er codes "represents a mDjor advancement in technology and can be used to replace" the WASH-1400 methods (which yielded the SSTl release category used in NUREG-1109).

(i.e., timing of containment failure and offsite release) has been seriously questioned. In conclusion, the use of the SSTl release category in the technical documents underlying the station blackout rulemaking, NUREG-1032 and NUREG-1109, has not been justified nor reconciled with the technical findings in NUREG-0956 and other technical reports. II.A.3 NUREG 1032 Contains Errors and Omissions The provisions of the proposed rule and regulatory guide most affected by errors and omissions pertain to loss of off-site power assumptions. In this regard, the Notice of Proposed Rulemaking states that the likelihood of a station blackout is "directly proportional" to the likelihood of a loss-of-offsite power event. 51 Fed. Res. at 9834, col. 3. The proposed rule links (Id. at 9830, col. 2) the potential and duration for a loss-of-offsite power event to three factors: plant offsite-

• power design features, grid stability, and severe weather.

Each of these factors is discussed below.32/ Offsite Power System Design. In postulating a relationship between offsite power system features and the In any discussion of loss- o f-off~site power events it is important to keep in mind tha t . i ndu stry-wide, there are reasons to believe that t he fr~aue ncv of loss-of-off-site powe r e vents is declining, the r eby r e ducing the emphasis it should play in station blackout. See NSAC-103, "Losses of Off- Si t e Powe r at U.S. Nuclear PowerPlants," April 1985 at i i i .

potential for a loss-of-offsite power, durations were assigned in NUREG-1032 to certain events of less than one minute duration "[t]o facilitate the statistical analysis."1_i/ Also, one event was double-counted and several non-LOOP events were introduced in the analysis._li/ While none of these items, standing alone, is dispositive, the net effect of these adjustments is to result in questioning the database and add unnecessary conservatisms. At a minimum, these adjustments prevent others from reproducing the analysis in an objective

• manner.

Further, it appears that a significant statistical error exists in the failure to account for the large sum-of-squares error associated with the results.}_~/ While some correlation was found, the overall quality of the statistical model, as measured in "R-Square" terms_36 / , is less than 25%. This type of error undercuts the validity of the correlations obtained between offsite power design features and the likelihood of a

• loss-of-offsite-power.fl/

ll_/ NUREG-1032 at A-10. In view of the above, a sound basis _lil See NUGSB at 85-005, supr~ n.16, at 4-3 to 4-6 and 4-7 to 4-8. 35/ Id. at A-9 (Appendix A, Item 3). i§_/ "R-Square" refers to a term in the statistical model which measures the cap~bility of the model t o a ccount for variation in the dependent v ari ab le. J f R-Square is small, errors in the model are l 0 rg c ~ ~d t h: ~o de l i s unreliable. _3_?__/ NUGSBO hc1d previously perf o rme d a difference of means test (based upon t h e "t" st0tisti c ) to determine if the offsite power d e sign groupings were distinguishable. This test (Footnote 37 Continued on Next Page

has not been established for associating certain offsite power design features with the potential for a loss-of-offsite power in the manner done in NUREG-1032. The statistical errors also tend to question the distinction between "offsite power design characteristic groups" underlying the 8-hour coping requirements of Tabie 1 of the draft regulatory guide (i.e., the "Pl" and "P2") groups. Grid Stability and Weather. NUREG-1032 also postulated grid-related and weather categories. Unlike the category of plant-centered events, however, these two categories are not based on a rigorous statistical review of plant experience. Instead, NUREG-1032 arbitrarily assigns event frequencies and power restoration times. No citation to other data sources or supporting analyses is provided to show that actual experience, or even the basis for the values assumed, confirms the appropriateness of the grid-related and weather-related categories of events. Thus, it is not possible to determine the reasonableness of the assumptions made.38/ In addition, the methodology used to determine the effects of weather on the frequency and duration of loss-of-(Footnote 37 Continued from Previous Page) only established statisti cally distinguishable mean durations of offsite power loss amon9 adj acent groupings of 60 and 80 % confidence, well below th~ levels of confidence ordinarily necessary to conclude that a distinction exists. Id. at A-9 through A-11. 38/ 1~- at 3-7 to 3-22 (§§3.2.2-3.2.5).

offsite power events lacks a sound technical basis. NUREG-1032 cites (at A-22) one paper (i.e., by Lauby, et al.) in the technical findings as yielding the correlation between weather hazards and loss-of-offsite power frequency.22./ The Lauby paper is central to the technical findings concerning this asp~ct of the proposed rule. However, in contrast to the use of the Lauby paper in NUREG-1032, Lauby never actually assumed a linear relationship between power loss frequency and weather-hazard rate. The paper did not even examine the potential for

• such correlations. Also, contrary to the implications in NUREG-1032, Lauby did not evaluate hurricane or tornado hazards.

In fact, with respect to the conclusions in NUREG-1032 regarding grid stability and weather, the Lauby paper reaches opposite conclusions. For example, the Lauby paper found (at 2349, col. 2) that: [t]he 345 kV Line-Related and the 230 kV

• and 345 kV Terminal-Related weather associated Forced outage rates were not significantly different between the winter sub-divisions of mean annual snowfall which was used to simulate the frequency of snowstorms. (Emphasis added.)

 ~ / M. G . Lauby , e t a 1. , " E f f e ct s o f Po o l i n g We a the r Assoc i a t e d MAPP Bulk Trar1smTssion Outage Octa r:,n Cc"l_culated Forced Outage R<1tes," (January 198tl) :?~;1er 1:~c1 1*YMO<ll0) Presented at IEEE Winter Power Meeting, Dallas, Texas.                     IEEE Transactions On Power Apparatus and Systems, Vol. PAS-103, No. 8, 2345-51.

The implication of this conclusion is that the paper was not intended to support the hypothesized relationship between weather hazards (cumulative amount of snowfall in this instance) and the potential for loss-of-offsite power events at a given plant. In addition to the above described errors in NUREG-1032, similar errors were found in the reliability equations apparently used in calculating the unavailability of the

• emergency AC power system.40/ However, since the actual risk equations used in calculating core damage frequency are not presented in NUREG-1032, it is unclear whether the various derivation errors in the treatment of the emergency AC power system reliability are merely typographical or, in fact, are carried forward in a substantive manner into the analysis.

Such substantive errors in the analysis of core damage frequency due to station blackout, together with the above-described errors in the loss-of-offsite power analysis, raise further concerns as to the sufficiency of the underlying technical bases in supporting the generic rulemaking. II.B Additional Matters Included in the Technical Record Do Not Support the Proposed Rule During the Commission's conside r ation of the recommendation to publish the proposed rule, several matters 40/ Corrections to the Staff's derivations are provided in NUGSBO's comments of November 1985. See NUGSB 85-005, su~ n. 16 at 3-25 (§3.3.2).

were raised as additional support for the rule. These matters are as follows and will be further detailed below. I . The European Approach to Station Blackout Reactor Coolant Pump Seals Hurricane Gloria II.B.l Clarifications of European Approach are Necessary Various oral and written representations regarding the European approach (particularly that of the French) to station blackout risks have been made and are part of the technical record.41/ In short, the European approach to station blackout should be viewed separate from the European approach to protecting against unique external circumstances which are unrelated to station blackout. In this regard, it should be clearly recognized that the French found the likelihood of a station blackout to be on the order of 10 -5 per reactor-year.42/ This value compares well with point estimates provided by the Staff for design features present at most U.S. plants. The

• French considered this risk to be acceptable, with two conditions: (1) the implementation of special procedures43/

and (2) providing the means for diminishing the threat posed by 41/ See Transcript of November 14, 1985 Commission meeting on station blackout ( ~ , at 28, 39, 40, 47 , etc.). 42/ See Tanguy, infra n. 49, at 595, col. 2. These procedures are directed toward ~eas onable operator actions which could be taken to rest ore power and provide core cooling. The initiatives advanced by industry provide comparable procedures.

loss of reactor coolant pump (RCP) seal integrity. Recent information indicates that RCP seals do not represent I - significant potential for large inventory loss (see §II.B.2 below). At the time the French examined this issue, however, the integrity of reactor coolant pump seals in a station blackout was viewed with great uncertainty. This concern prompted the installation of a turbine-driven pump to provide makeup. More recent information (discussed in Section II. B.2) indicates that seal integrity negates the need for a pump of

• this kind.

The French approach to station blackout does not appear to depart significantly from current regulatory approaches in the U.S._!!/ Other countries are not much different. In Belgium, for example, current U.S. practices concerning onsite and offsite AC power reliability and redundancy, as well as the lack of a design basis coping capability, remain the norm. The current position is that the diversity in AC power sources makes the event highly improbable.45/ Italy and Switzerland provide similar layers of AC power system redundancy and reliability and, thus, are relatively unconcerned about station blackout. 46/ _!!/ One area of departure may be coping ca pab ility. The coping capab i lity of Fr ench plants wi l l be ois~u ssed below. 45/ See IAEA-TECDOC-332, "Safety Aspects of Station Blackout at Nuclea r Power Plants," at §8.l, International Atomic Energy Agency, Vienna, 1985 (hereafter "IAEA"). 46/ Id. at §§8.7 and B.12.

                             -  30 -

Perhaps the best explanation for this perspective on the part of the Europeans is provided in the German design philosophy which is summarized in the following statement:47/ In German nuclear power plant design the station blackout (which is loss of on-site and off-site normal and emergency AC power), is expected to be an event of extremely low probability of occurrence due to the design of emergency AC power supply systems. (emphasis added) This reliability can be measured by the summary results of a recent German reliability study which reports 8 loss-of-offsite-power events in 33 years of operation of six plants.48/ This experience translates to a frequency of approximately 0.04 events per reactor per year, a frequency which is close to the annual U.S. event frequency experienced since 1979. That there are strong similarities should not be surprising since U.S. designs formed the basis for most plants in operation throughout non-Soviet bloc nations.49/ Similarly,

• 47/ Id. at §B.6.

48/ Id. at §B.6 (summarizes an article published in Atomwirtschaft, February 1984).

 ~9/ In a paper by P. Tanguy, Director of the Institute of Safety and Protection at the French Atomic Energy Commission (CEA), the French approach to nuclear safety was addressed:

[t]here are strong similarities between French safety philosophy and the safety philosophies of the other major nuclea r 9ower c ountries; this is hardly surp ri s i ng b~ca u se fo r mo r e than 20 yr there has been an extensive and continuous exchange of information between nearly all organizations around the world, either by means of bilateral agreements or (Footnote 49 Continued on Next Page

U.S. regulatory practices are referenced extensively in regulating power plant design and operation.2.9_/ From this common base, local regulations have evolved which recognize the differences in hazards faced by U.S. and foreign plants. Such hazards in Europe are unrelated to station blackout events, and generally result from greater population density in the vicinity of power plant sites, differing seismic criteria, and the greater likelihood of external events, such as military

• aircraft impact and gas explosions. Protection against these hazards is provided in some European countries through a number of local design features which have the effect of providing greater reliability of the AC power systems required for plant safety. It has been the practice in some countries to provide an additional level of redundancy in the number of diesel generators available to the plant. Discounting the common cause contribution, the overall emergency AC power system reliability sought is approximately 10 -4 per demand.51/ A
• (Footnote 49 Continued from Previous Page) under the sponsorship of such international bodies as the International Atomic Energy Agency in Vienna and the Nuclear Energy Agency in Paris.

P. Tanguy, The French Approach to Nuclear Power Safety, Nu c 1 ear safety, Vo 1 . 2 4 , No . 5 , - a t 5 8 9 , co 1 s . 1- 2 , - September-October 1983 (hereino.fter 1'2n_guy"). 11

 ?_QI See gene r o.11 y: ~<:! . , at 5 9 2 , col . "

21-.I J. A. Richardson, "Summary Comparison of West European and U.S. Licensing Regulations for LWRs, Nuclear Engineering International, February 1976.

similar concern for the potential loss of the decay heat removal function due to external events motivates the use of 4-50% capacity cooling systems or 3-100% capacity cooling systems. In an accident, two of the four (2/4) or one of the three (1 / 3) systems need to operate to provide core cooling. The picture which emerges from this brief review of the European approach to station blackout is clear: AC power reliability over and above U.S. levels results from European

• unique events and not from a particular concern with a station blackout event.?_~/

Another issue related to the features of European design raised question in the Notice of Proposed Rulemaking is whether the proposed rule goes far enough in comparison. See 51 Fed. Reg. at 9831-2. Two features cited in the Notice of Proposed Rulemaking are believed to be erroneous assumptions: (1) that the French plants possess a three-day coping capability; and

  * (2) that the French "have a goal of achieving a probability of one in ten million (10- 7 ) per reactor-year for a major event such as station blackout." Id. The claimed three-day French coping capability is deemed to support the proposed rule in that it would provide time to bring a mobile gas turbine onsite and terminate a station blackout event .         It is also considered 52 / Core cooling co nsiderations asso :.: iat :_, 1 -,1 it h ex t e rnal e v ents 1

(i.e., similar to the Eur o pean-unique hazards) is a subject for th e ongoing Task Action Plan for resolution of USIA-45, decay heat removal, not station blackout. See NUREG-1109 at 18.

( COMMENTS ON THE PROPOSED

• NRC RULE AND DRAFT REGULATORY GUIDE ON STATION BLACKOUT (51 I~d. Reg. 9829 March 21, 1986)

June 19, 1986

ABSTRACT On March 21, 1986 the Nuclear Regulatory Commission (NRC) published a proposed rule on station blackout. The proposed rule is in response to Unresolved Safety Issue (USI) A-44. The term "Station Blackout is defined in the proposed rule which would require commercial nuclear power plants to withstand and recover from a station blackout for a specified duration acceptable to the St~ff. The proposed rule includes an amendment to General Design Criterion (GDC) 17. Accompanying the proposed rule is a backfit analysis which relies in large measure on the regulatory analysis found in NUREG-1109. Reference is made in the proposed rule to a draft regulatory guide which was also published March 1986. The nuclear utility industry has been actively engaged in following the resolution of this issue. The Nuclear Utility Group on Station Blackout (NUGSBO) was formed by a number of utilities in the spring of 1984 and has made various presentations to the NRC. In the fall of 1985, the Nuclear Utility Management and Resources Committee (NUMARC) formed a Working Group on Station Blackout and took the lead on utility industry efforts to resolve this issue. NUGSBO remained as technical consultant to NUMARC. A review of the rulemaking material leads to the conclusion that rulemaking is not warranted for the following reasons: (1 ) The issue is not generic and, thus,

 *           (2)

(3) need not be resolved by a generic rulemaking. The technical record does not support the proposed rule. The proposed rule itself should be reevaluated.

• (4) The proposed rule does not meet the backfit rule standard.
• In the event the Commi_ssion decides to promulgate a rule, despite these reasons to the contrary, there are several factors which are of concern and warrant special consideration.

achievable at reasonable costs. Similarly, the Notice of Proposed Rulemaking states that the French 10-? objective represents a higher safety objective than the 10-S objective associated with the proposed rule.53/ With regard to coping, it is important to point out that there are several reasons why many of the features attributed to the new French plants may already exist at most U.S. plants. The apparent focus of the described plant features at the new

• French plants is to provide sufficient m~keup water to primary systems to provide for reactor coolant pump (RCP) seal cooling and preclude seal failure. (Recent full-scale RCP seal tests under stat i on blackout conditions indicate that seal injection and cooling are necessary to maintain seal integrity. (See

   §II.B.2.). These features are similar to the current capabili tie s of many U.S. pl ants which utilize emergency pumps, such as f ir e pumps and portable equipment, to provide makeup to the condensate and feedwater systems.                 Such capabilities offer

• broad flex i bility to PfOVide decay heat removal should the main auxiliary feedwater and condensate systems be disabled by a station blackout. Indeed, in the context of providing makeup water, it i s likely that a l arg~ number of U.S. plants could claim a three-day cop i ng capability compara ble t o the French. 54/

. 5 3/ See Tan guy , s u :e. r a , n . 49 , a t 5 9 5 . 54/ I n the separate views accompaP.ying t h-2 p r opose d rule , it i s rep resente d that new Fre nch 1300 MWe plant s are desi gned with a goal o f "co ping" wi th a stati on blackout for "a t lea st 20 hours" and, fu rther , tha t the plants can "wi th sE~ nd" a s t a ti on b l ackout f o r "three days." 51 Fed. (Footnote 54 Continued on Nex t Page

                             -   34 -

If the French definition of coping capability is limited to simply having sufficient cooling water inventory, then the burdens of coping without AC power would be significantly reduced. The test applied to U.S. plants by the proposed rule and regulatory guide is significantly more rigorous: in addition to requiring that power plants have access to sufficient water for decay heat removal, the coping features of the proposed rule also include the requirement for a demonstration of equipment operability in environmental

• conditions associated with a station blackout.

9835, col. 1. Draft regulatory guide at 7. 51 Fed. Reg. at The current capability of French plants to "cope" with station blackout may be inconsistent with the concept of coping as contained in the regulatory guide published with the proposed rule. No information is provided that French plants can meet the terms of the regulatory guidance on coping capability, nor has it been demonstrated that in terms of the

• regulatory guide French plants can cope with station blackout for as long as the Notice of Proposed Rulemaking suggests they can.

7 With regard to the 10- figure, it was noted above that the actual frequency of station blackout i0. Europe is (Footnote 54 Continued from Previou s Pag ~ 1

      ~~- at 9831, col. 3. It is not clear what distinction, if any, exists between the asserted ability of the French design to "cope" with a station blackout for 20 hours and to "withstand" a station blackout for three days.

comparable to most U.S. plants, i.e., 10 -5 . More importantly, in contrast to the separate views accompanying the Notice of Proposed Rulemaking (51 Fed. Reg. at 9831-2), the number 10- 7 per reactor per year is not a goal but a screening limit for excluding families of events from consideration(~, low probability events).55/ To this end, the screening process used for PWR designs is aimed at maintaining the probability of accidents l eading to individual offsite exposures between 5 and 50 person- r ems to be less than 10- 6 per reactor per year .

• As has been discussed, NUREG- 1032 assumes a core damage frequency of 10- 4 per reactor per *year arising from a station blackout event. Core damage, in and of itself, does not give rise to i ndividual offsite exposur e s, rather, core melt, vessel failure and breach of containment integrity are necessary.56/

The frequ e nc ies associated with these additional events drive the frequen cy i nto the range of the French screening limit (i.e., 10 - 7 ). For example, as noted (supra n. 21), even the

• use of the SST! offsite release frequency of 10- 4 results in a core damag e/conta i nment failure / offsite release frequency of 10- 7 or l ess. See NUREG 1032 at Table 7.4 at 7-19.

According l y, there i s l i ttle dispar i ty between the U.S. and ... French in th i s regard . 55/ See Tanguy, s~.E..£_~ , n. 49 , a t 594-596 . As noted i n §II .B. 2 above, cor~ damage is imp roperly equated (wit h the ela pse of 1 . 5 hour s ) wit h co ntai nment bre ach and pub lic exposu re . It shoul d be po i nted out, the Fr ench do no t e quate co re damag e with conta i nment failure on such a non-mec hanist i c basis. See Id. at 597.

II.B.2 Clarifications of RCP Seal Integrity are Necessary The regulatory analysis, NUREG-1109 (at 19) links reactor coolant pump (RCP) seal integrity to a plant's ability to cope with a station blackout. The issue of RCP seals was also discussed at the November 14, 1985 Commission meeting. Transcript at 27. The relationship between reactor coolant pump seal

• integrity and station blackout requires clarification.

issue concerns the ability of the seals to serve as a reactor coolant p~essure boundary when deprived of cooling water in a station blackout. Assumptions regarding the This timing of seal failure and the rate of coolant loss are important to determining a plant's ability to cope with a station blackout.57/ To begin with, it needs to be recognized that the reactor coolant pump seal integrity issue is a generic issue having its own Task Action Plan (Generic Issue B-23). The principal features of that plan are prototype tests of various seal designs to determine the most viable approach to extending seal integrity should cooling or injection be

• lost. The current state of knowledge sepa r ates the seal leakage potential of reactor coolant pumps i nto two groups:

Westinghouse pumps with a hydrosta tic se cl. a nd pumps with a 57/ See NUREG-1032 at §6.

hydrodynamic seal. Pumps with hydrostatic seals are believed to pose a greater risk due to seal failure. With respect to station blackout, however, only leakages of 100 gpm per pump or more are of concern because of their impact on limiting decay heat removal.58/ Smaller leakages do not have significant impacts on coping. This criteria eliminates from consideration all boiling water reactors and plants operating pumps with hydrodynamic seals. Thus, reactor coolant pump seal integrity is only relevant to the Westinghouse pumps with hydrostatic seals due to the larger seal leak potential. The concern over the Westinghouse pumps is also approaching resolution. Recent tests of this design have introduced substantial doubt regarding the previously assumed large leakage potential. At this time, it does not appear that high pressure AC-independent injection pumps are

• needed to compensate for seal leakage. In fact, since 1983, the Staff has recognized this improvement in RCP seal integrity and uses a 20 gpm per pump primary inventory loss rate in their analysis of station blackout.59/ This value is below the 100 gpm criterion cited in NUREG-1032 and is supported by the revised Westinghouse repo rt (WCAP-10541, Rev. 1).

58/ NUREG-1032 at 6-3. _59/ See NUREG/ CR-3226 at Appendix G at 248.

The impact of this discussion highlights the fact that concerns over RCP seal integrity do not contribute to the overall station blackout concern. This issue was originally addressed to Westinghouse plants, but with the impending resolution of the RCP seal integrity issue, further discussion of the risk of potentially large inventory losses in a station blackout does not appear to be relevant. II.B.3 Clarifications of the Significance of Hurricane Gloria are Necessary

• At the November 14, 1985 Commission briefing on station blackout, a general concern was raised (Transcript at 8) that nuclear utilities were ill-prepared for severe weather events believed to be major contributors to station blackout risks.60/

In particular, reference was made to the late-September 1985 Hurricane Gloria which moved up the East Coast from Florida to New England. At the briefing, Hurricane Gloria was represented as illustrative of the basis for the concern that utilities cannot adequately respond to rapidly moving severe weather in a timely fashion, i.e., in time to take proper precautionary measures. These representations seriously understated the present capability of utilities to take timely and effective risk-reduction measures in response to such seve r e weather events as Hurricane Gloria. 60/ See also, ~ - , 51 Fed. R~. at 9830, col. 2 and at 9831, coI. -r:-

For example, at Millstone unit Nos. 1 and 2 in Connecticu t, the licensee had advance notification of the storm's approach, easily tracked Hurricane Gloria and routinely implemented its onsite hurricane action plan. This plan includes, among other things, a checkout of the Emergency Response Facilities, the selection of two Station Emergency Organization shifts and verification tests of all emergency onsite AC power sources.61/ Despite the storm's rapid movement toward the plants, the licensee had ample time to take all necessary precautions for station blackout risk-reduction purposes, including orderly plant shutdowns, and preheating/starting the emergency onsite AC power sources. Due to these precautions, the Millstone units were in stable shutdown operations during the weather-related loss-of-offsite power at the site. In conjunction with bringing the units off-line, all Millstone emergency on-site AC power sources successfully

• started and loaded and ran until prudent plant actions were completed to allow for restoration of normal off-site power.

If necessary, Millstone Unit No. 1 could have had off-site power restored within 3 1/2 hours and Millstone Unit No. 2 could have had off-site power restored within 5 1/2 hours. Since more rapid restoration of off-site power was not vital, the utility elected to pursue a mo r ~ delibe r ate and thorough cleaning and checking restorat i on pro cess.

 ?1/ Northeast Utilities letter to NRC (J.F. Opeka to H.R.

Denton), December 27, 1985, at 1.

The advance notification associated with severe weather events of this kind permits advance precautionary actions not usually credited by the Staff or plant probabilistic safety studies. As noted by members of the Advisory Committee on Reactor Safeguards during the November 19, 1985 Subcommittee meeting in Waterford, Connecticut (Transcript at 207 through 209), the failure to credit utility capabilities for actions in response to advance warning prior to a severe storm arrival led to conservatisms in a probabilistic risk assessment, and accordingly, these events should be categorized in a fashion different from other loss-of-offsite-power events. In this regard, the Atomic Industrial Forum's Public Affairs and Information Program has assembled information from affected utilities which is in contrast to the assertion mentioned above that the present capability for severe weather precaution at nuclear plants is inadequate.62/ In spite of

• §2/ See INFO Report, Number 204, October 1985, which, with respect to other affected utilities, states as follows (at 2-3) :

A number of nuclear plants on the East Coast from the Carolinas to New England voluntarily shut down or reduced power in precautionary actions in late September as Hurricane Gloria came calling - or threatened to call - with her blustery winds. No significant damage was reported at any nuclear units. Carolina Power & Light Co. brought Brunswick 2 in North Carolina down because of t~ e oossibility the storm would move towards th~ site. Unit 1 was down for scheduled maintenance a t th e time. CP&L's Wayne Ennis said the utility was not concerned about the plant withstanding the hurricane winds, but about the possible loss of transmission lines (Footnote 62 Continued on Next Page

suggestions at the November 14, 1985 briefing, the utilities (Footnote 62 Continued from Previous Page) into and out of the plant. The plant was down less than 24 hours due to the storm, but CP&L took the opportunity to do some maintenance work while it was off line. In.New Jersey, Jersey Central Power & Light Co. [sic, GPU Nuclear Corp.] reduced power at Oyster Creek to 35 percent on the evening of September 26. JCP&L's [sic, GPU Nuclear's] John Fidler said the action was taken to avoid a potential situation in which the plant lost offsite power and then tripped at full capacity. However, Gloria caused no such problems and the plant was returned to 100 percent power the next day. Public Service Electric & Gas Co. reported that Salem 1 & 2 escaped Gloria unscathed, although power was reduced slightly after the storm so intake screens could be cleaned. On Long Island, where the brunt of the storm was fP.lt, the Long Island Lighting Co. began shutting down the Shoreham plant September 26 and declared an unusual event and an alert the next day. "One of the triggering mechanisms for an usual event condition is wind speeds or anticipated winds in the vicinity of 80 mph or above," said LILCO spokeswoman Carol Clawson. "For an ale rt, it is wind speeds of 100 mph in the vicinity of the plant," she said. Shoreham, which currently has only a low-power license, was back up by October 1 [sic, 3). In Massachusetts, Boston Edison Co. reduced power at its Pilgrim 1 plant to 25 percent late in the afternoon on September 27 and k e pt power at that level until 1 a.m. Saturday morning because winds of about 70 mph were reported in the area. The a c tion was taken in acc or dc1nce ,.*i. -. h th e unit's 1 t Pc h n i ca 1 s p e c i f i ca t i on s . T lw 1J t. i__l i t y then sh u t the plant down voluntarily to wash salt off the insulators. Pilgrim 1 was back at 100 percent by Sunday. (Footnote 62 Continued on Next Page

threatened by Hurricane Gloria successfully averted the risks of station blackout. II.C Respo n ses to the Additional Commen ts and Views of the Commi ssion Must be Considered Among the considerations of the rulemaking are part'icular concerns raised by the Commissioners in their additional c ommen ts and separate views accompanying the proposed rule (51 Fed. Reg. at 9831-2). Since much of the information that would

• (Footnote 62 Continued from Previous Page)

Other utilities reported that their nuclear plants ran smoothly, despite the storm. Baltimore Gas & Electric Co., for example, said Calvert Cliffs 1 and 2 operated at full power throughou t the storm. Had Gloria hit the area, however, the utility was ready for her: the utility had called in extra workers to be on hand. In addition to these events described above, other utilities were able to take appropriate measures. In accordance with the Indian Point Unit No. 2 (IP-2) Technical Specifications, Consolidated Edison commenced hot shutdown procedures when the center of Hurricane Gloria was within 320 nautical miles of Indian Point with sustained winds exceeding 100 mph (Approximately 0230 EDT, September 27, 1985). The Technical Specifications also required that appropriate actions be taken to ensure that the plant be in the cold shutdown condition prior to arrival on site of hurricane winds exceeding 100 mph. This cold shutdown action was not required in the case of Gloria since winds of 100 mph or greater were not experienced in the IP-2 proximate area. As an additional precautionary measure, the site declared a "Notice of Unusual Event" (NUE) in accordance with its Emergency Plan. (An NUE is the first and least significant of the four action levels in the Emergency Plan.) On September 27, 1985 at approximately 1600 EDT the National Weather Service had lifted the hurricane warning for the area and IP-2 began start-up procedures. Virginia Power took Surry 1 offline and reduced Unit 2 to 25 percent power for a few hours as a precaution for employees that were working outside to assist in shutting down the plant.

be contained in response to Commissioner questions is presented throughout this document, the appropriate sections of the comments are cited. There are five issues raised by the Commissioners as comments or separate views: (1) The need for "quality classification" of station blackout modifications; (2) Whether the backfit analysis for the proposed rule "adequately implements" the Backfit Rule, 10 C.F.R. §50.109; (3) Whether the reduction in risk offered by the proposed rule constitutes a "small percentage ' of the overall risk" or "a major component of an already small risk"; (4) Whether the proposed rule meets the "substantial increase in the overall protection of the public health and safety. threshold required by the backfit rule"; and, (5) Whether the NRC "should require substantial improvements in safety with respect to station blackout, like those being accomplished in other countries, which can be achieved at reasonable cost and which go beyond those proposed in this rulemaking." (1 ) Quality Classification is Unnecessary. Equipment used to prevent or respond to a station blackout should be sufficiently available and operable to meet its required function. To this extent, the Commission's desire that appropriate attention be paid to maintaining a sufficiently high state of operability and reliability i s appropriate. Th8 oo i n t o f departure begins with the me t hod for achieving this ob jective. Specifically, by itself, a "safety grade" classification scheme does not solely

equate with high states of equipment operability and reliability. Such classification systems too often can become a documentation exercise more than a process for providing the requisite level of system functionality. In lieu of a classification system, the goal of system availability should be pursued programmatically.63/ (2) Implementation of the Bac kfit Rule Is Inadequate. The backfit analysis for this rule does not a dequately implement the Backfit Rule, 10 C.F.R. §50.109. Additional information concerning this point is provided in §IV. (3) Reduction in Risk of Station Blackout Is Small. An important aspect of this rulemaking is the estimated freque n cy of core damage due to station blackout. For most plants, this value can be calculated using the methods in NUREG-1032 to be currently near or below 10 -5 per reactor year. If the first

• half of the industry's loss-of-offsite-power experience is excluded in recognition of industry improve ments,64/ the estimated frequency of all station blackouts greater than 2 hours in duration is near or below 10- 5 per reactor per year for all plants.

This goal may best be achieved through the initiatives advanced by the industry. For the re a s ons discussed therein, that program can make si gni fic ant progress towards that goal without incurring significant burdens in the area of documentation. 64/ NUGSBO previously presented information to the NRC supporting this position. _S ee sup~~ n. 16.

The Notice of Proposed Rulemaking acknowledged that the "total risk from nuclear power plant accidents . was found to be small." 51 Fed. Reg. at 9830, col. 1.65/ This "small" risk is clearly acceptable and does not pose an undue risk to the public health and safety.66 / Thus, any reduction resulting from the proposed rule will be in the category of diminishing returns. This point is underscored in NUREG-1109, table 5. In two of the three categories addres*sed, the reduction in risk is 5 5 small (i.~~' 1.3 x 10- and 0.1 and 10- ). In the other category, wherein the risk reduction is comparatively large (14 5 x 10- ) two points are in order. First, as noted, several of the plants falling into this category are able to take actions which reduce the initial risk and, thus, render small any reduction associated with the rule.67 / Second, even assuming the risk reduction advanced by the proposed rule, the number of plants falling into this category is small (i.e., 14 plants, 7 sites) and, thus, while the risk reduction might be large as to

• §2/ The proposed rule's acknowledgment was made on the basis of WASH-1400 results (i.e., one of the primary technical sources of reliance in the proposed rule). WASH-1400 was published in 1975. Clearly, given the numerous rulemakings and requirements that have been issued since then (i.e.,

NUREG-0737 responses, fire protection, equipment -- qualification, emergency planning, etc.), the risk is even smaller now. 66/ It was pointed out in §I that no plants pose an unacceptably high risk r e sulting fr o m s~ation blackout as to require immediate action (i.e., the r isk was not a threat to public health and safety). Se e NUREG-1032 at 2-3. 6_}._/ See §I, supra n. 6, regarding the ability of plants with 2/ 3 eme i: genc:y diesel g e nerators to cross connect and thus have 0n p [fective c onfiguration of 1/3.

these limited number of plants, the risk reduction associated with the majority of plants will be small. Thus, as a general matter, the reductions in risk offered by the proposed rule constitute a small percentage of the overall risk, a risk which is already small (and acceptable). The reduction in risk for a limited number of plants could be larger, however, as some of these plants are able to take actions which reduce the risk and, thus, the proposed rule's risk reduction benefits .

• Placing station blackout risk in the context of overall risk requires knowledge of several other parameters.

elements of that state of knowledge are the estimated frequency Important of other accident scenarios, containment performance in a severe accident, and source term assumptions. Updated information in this regard is expected in the next severe accident research report. This report, NUREG-1150, is currently scheduled to be released late in the summer of 1986. Additional comments may be appropriate at that time which will

• further address the relative risk question.

(4) Rule Does Not Provide Substantial Additional Protections. The proposed rule does not me et the threshold provided by_ the Backfit Rule. See §IV for comments concerning this issue. (5) Additional Modificati on s Are _Not Necessary. The European approach t o station blac ko ut is no t i ncon s i s te nt with the present U.S. approach. Accordingly, additional modifications

premised upon the European approa~h are unnecessary. See discussion at §II.B.1 .

III. The Proposed Rule Itself Should Be Reevaluated III .A The Ultimate Requirements of the Proposed Rule Are Indefinite and Depend Upon the Future and Uncertain Exercise of Discretion The proposed rule would create indefinite requirements. The Supplementary Information accompanying the proposed rule explains that the rule would require an open-ended determination of "[t)he amount of time the plant can maintain

• core cooling and containment integrity with AC power unavailable." 51 Fed. Reg. at 9831, col. 1.68/ The problem is further illustrated by the fact that the regulatory guide and proposed station blackout rule do not provide limits on utility efforts to demonstrate requirP.d coping. See the regulatory guide (March 1986) at §3.1. In fact, the text proposed for codification by the rule does not specify a numerical value for required coping time.

Unless the required coping demonstration is specifically bounded by clearly stated definitions, assumptions, and criteria, there could conceivably be hundreds of supporting special effects analyses which licensees may have to consider as a result of the exercise of discretion by individual Staff reviewers. Under the rule as proposed, lic~nsees cannot ascertain the ultimate requirements they will be expected to 68/ This demonstration would be a meaningless exercise if it results in coping periods in excess of the 4 or 8 hour categories referenced in the Draft Regulatory Guide.

meet (including the potential plant modifications they will need to make) to demonstrate compliance. For example, in the draft regulatory guide it is stated (at 7) that equipment necessary to withstand a station blackout must "[m]eet design and performance standards that ensure adequate reliability and operability in extreme environments, that may be associated with a station blackout including hazards due to severe weather. II It is not clear how this guidance will be interpreted. If licensees are required to consider loss of AC power in conjunction with 10 C.F.R. §50.49 reviews, the scope of §50.49 will be significantly expanded and this will have a substantial impact on industry. In this regard, the potentially significant activities necessary to demonstrate adequate reliability and operability in environments that may be associated with a station blackout do not appear to have been addressed. It should be clearly stated that the consideration of total loss of AC power is unnecessary

• for §50.49 dP.monstrations of equipment qualification.

proceed otherwise would subject utilities to substantial costs associated with changeout and replacement of much equipment To which is*functionally sound in all respects but which may lack documentation sufficient to satisfy the rule. Accordingly, the technical guidance accompanying any rule sh~uld be flexible enough to permit suitable engineering analysis and prudent compensatory measures ( ~__:_9~, open i '7? eqti j_ prn<? 11 t cabinet doors to provide cooling) to assure that the necessary design and performc1.nu~ standards of concern have been satisfied.

The point above is not that regulations must be prescriptive by their very nature. Prescriptive regulations, which outline in detail exactly what steps are required by licensees to satisfy a proposed regulation, are, in many instances, unnecessary and counterproductive. Rather, the point is that demonstrating conformance with new design requirements is time-consuming and resource intensive, especially when the criteria applied to judge the adequacy of licensee verifications are not identified from the outset (and, in the case of the proposed rule, will be indirectly developed as a result of licensee submittals rather than directly in the rulemaking).§2/ Licensee activity needed to implement the draft rule has not been accurately described. The regulatory analysis, NUREG-1109, assumes that all plants will derive benefits from implementing the rule. In the value/impact assessment (discussed in Section IV) the rule is calculated to result in a

• total dose reduction of 80,000 person-rems across all of industry over the remaining life of the 67 plants it considered.70/ Yet it is also claimed that "almost all plants

 §9/ As addressed below in connection with the discussion of cost estimates (§IV.B), licensee activities cannot be determinPd for purposes of the NRC sati;fying its obligations under the backfitting rule, 10 C.F.R. §50.109, unless the proposed rule first identifies clearly what it proposes to require and rlPscribes wit~ [ asonable detail 0

how licensees are expected to ~c~ie *2 compliance. 0

 ]_J_/ In NUREG-1109 (at 8) it w0s explained that the 80,000 person-rems was derived by utilizing information from NUREG/CR-2723 which provides estimates of offsite (Footnote 70 Continued on Next Page

                              - Sl -

should be ahle to meet [the requirements of the proposed rule] without major modifications."71/ It is difficult to understand how such significant benefits across all of industry can be assumed to accrue from a backfit that the regulatory analyses claims will not require any hardware at "most plants." If, in fact, there will be a total dose reduction of 80,000 person-rems over the industry, it would appear to follow that some licensee activities will be required "over the industry." These inconsistent statements point to the vagueness in the scope of the proposed rule. Industry concerns are long-standing over insufficient precision in regulations. On October 27, 1980, in the fire protection rulemaking, the basis for Section III.G of 10 C.F.R. 50, Appendix R to the Commission was presented. The purpose of the section was to achieve a consistent level of fire protection safety at those plants where safe shutdown system separation remained an open item. Approximately 20 plants

• II

       . would be significantly affected by the backfit of

[Section III .G]. "]_}I plants, While backfitting this requirement on all the Commission recognized that the majority of plants which had completed safe shutdown system modifications should (Footnote 70 Continued from Previous Page) consequences for each of 91 sites. It ~ould appear that, for each of the 67 plants considered in NUREG-1109, the corresponding offsite conseguenc~ dat~ in NUREG/CR-2723 was used. _71/ NUREG-1109 at 11.

 ~~/  October 27,  lq80 Commission Meeting Transcript at 20.

not be affected by this feature of the rule. Consequently, the Commission took the unusual step of adding an exemption process to its fire protection regulations at 10 C.F.R. §50.48 to allow utilities to receive credit for the fire protection safety added to their plants.2l_/ With Appendix R, a rule was sought to resolve issues that may otherwise be closed out in individual dockets. In so doing, most plants were expected to do little since their fire protection features were already reviewed and accepted by the Staff. It was believed that no additional compliance costs would be incurred by these plants._2i/ But the impact of new regulatory requirements cannot always be isolated to the weaker plants. In the case of Appendix R, contrary to this assumption, Section III.G disrupted licensee activities not at 20 plants, but at almost 90t of all operating plants with most plants expending millions of dollars to comply.75/ III.B The Proposed Rule Will Not Achieve a Consistent or Efficient Resolution of the Station Blackout Issues A major factor in the current proposal is the central importance of coping as a means of reducing the potential core damage frequency due to station blackout. The Notice of Proposed Rulemaking states that the expected frequency of core 73/ See 45 Fed. I\eg. 76602 (Noverril~or 1.9, 1<:iq()). 2il Enclosure B to SECY 80-88, February 13, 1980.

 ]5_1 See §IV.B below.

damage from station blackout could be maintained near or below 10- 5 per reactor-year provided the plant is designed to cope for a specified duration. 51 Fed. Reg. 9830, at col. 3. This goal apparently is a major reason for the proposal that coping become a design feature for nuclear power plants. Yet, the analysis does not support the need for the rule for several reasons:76/ (1) More than half the generic combinations of loss-of-offsite power susceptibility and diesel generator reliabilities do not require 4 ~ more hours of plant coping to meet the 10 5 goal; (2) four combinations of generic features do not require plant coping at all; and (3) The expected frequency of station blackouts lasting 4 hours or longer is an order of magnitude less than the total expected blackout frequency (i.e., plant coping for these durations represents only a limited amount of risk reduction compared to the tot a 1 r i s k ) . r}_ / If the proposed rule is implemented, the equipment qualification issue will potentially be reopened at many plants by extending the number of plant areas which may be deemed to be subject to a "harsh" environment. See §III.A above. In order for plants to comply with the proposed station blackout rule, a new program of qualifications and, possibly, 76/ See NUGSB 85-005, supra n. 16. 77 / In this regard, §II.C abovi? acldr<?ssecl thc. additional comments of Commissioners Roh r-*:::, 2 !1 r 1 -= =*: h, pertaining to arguments that the proposed rul e would at best result in a small reduction to an already small risk.

requalification of existing equipment to the station blackout environment could be necessary. While some equipment may be able to take credit for existing qualification, equipment in the new areas will be introduced to the program. The effect of this action would be to extend the schedule for compliance with both the station blackout and equipment qualification rules beyond current expectations. The cost of such compliance activities could also be significant and was not considered in the regulatory analysis of the rule. See §IV.B below . The most practical approach to this issue is to reduce the likelihood of a station blackout. The dominant factors which contribute to this potential have been identified: loss-of-offsite-power and the early availability of emergency AC power. As noted in NUREG-1032, improvements in both areas are effective in reducing the potential for core damage to 10 -5 per reactor-yPar or less .

• Additional redundancy in the emergency AC power system may be provided by reducing the number of diesels deemed necessary for station blackout or by crediting the availability of another backup power source. The impact of such improvements is a substantial reduction in the likelihood of a station blackout. As the analysis in NUREG-l032 demonstrates, such measures could easily achieve the factor of 2 improvement in core dam a ge frequency sciught b y ~~e ~~l0~~k ing with a design basis coping capability. In fact, as previously discussed, it

is more likely that greater improvements could be realized in this fashion. Another concern with the proposed rule is its potential to divert licensee resources away from improved reliability and towards the more limited improvements achievable through compensatory measures, i.e, coping. Moreover, with a regulatory focus on coping, licensees receive effectively no credit under the proposed rule for actions taken to improve AC power system availability. Such improvements offer substantially more safety benefits than the proposed rule, and with earlier results .

IV. The Proposed Rule Does Not Meet the Backfit Rule Standard IV.A Introduction Included in the Notice of Proposed Rulemaking on station blackout is the backfit analysis performed pursuant to the NRC's backfitting rule, 10 C.F.R. §50.109. 51 Fed.

 ~~~-  9829, at 9833-35.         The backfitting rule mandates that the NRC Staff demonstrate by analysis, and the Commission find prior to the imposition of a backfit, that the proposed backfit will result in:

a substantial increase in the overall protection of the public health and safety or the common defense and security . . and that the direct and indirect costs of implementation for that facility are justified in view of this increased protection. [10 C.F.R. 50.109(a)(3)] In order to make this overall finding, Commission regulations require that an analysis be performed consistent with 10 C.f.R. §50.109(c). 10 C.F.R. §50.109(c) lists nine (9)

• specific factors which are to be addressed as appropriate.

In order to address the requirements of 10 C.F.R.

 §50.109(c)(l-9), NUREG-1109, "Regulatory Analysis For The Resolution of Unresolved Safety Issue A-44, Station Blackout" was prepared.'?_~/       This document estimates t 1,'= risk reduction to 78/ The h,c1ckfit ano.lysis olso refe 1:*:; to 1.~l ~Jrobabli.stic risk asses~,rnr~nt studies, (.2) romrnon cause failures, (3) estimated frequency of core damage and (4) European exp e r i enc e . I terns ( 1 ) , ( 3 ) and ( 4 ) a re add re s s ed i n § I I above. With respect to common cause failures, the (Footnote 78 Continued on Next Page

be 80,000 person-rems and asserts that the total cost of compliance would be about $40 million. On this basis, an overall value-impact ratio of about 2,000 person-rems averted per million dollars is reached.79/ A review of the backfit analysis and documents referenced therein, particularly NUREG-1109, shows that these backfit (Footnote 78 Continued from Previous Page) following is presented . The potential for station blackout is linked to the likelihood of coincidental failure of all emergency diesel generators (EOG). Such EOG failures could occur independent from each other, or as a result of a common cause. The potential for common cause failures provides an upper limit to EDG reliability and the availability of the emerg(mcy AC power system. See generally NUREG/CR-2099 (1982). The effects of common cause failures need to be considered in viewing station blackout risk and the value of improved AC power reliability. The state of knowledge today is sufficiently advanced to give explicit consideration to these effects in availability estimates. The kinds of remedies which are effective in reducing the impacts of common cause failure are also known and include greater diversity in design and maintenance. It should be emphasized that the contribution of common cause failure to overall system unavailability is low. Typical values for dependent failures cited in the literature are on the order of 1-2% of all components with some estimates ranging up to 5%. Estimates of the value of improving EOG reliability relative to station blackout risk, which are discussed in this document, consider the common cause contribution as a limiting parameter. It is believed that the common cause failure issue is both limited in scope and manageable. 79/ 2,000 person-rems per million dollars equates to $500 per person-rem. The Commission regulations state that additional measures shall be ta.ken i. F 11:1 favorable cost-benefit ratio [can] effect reductions in dose to the population." 10 C.F.R. Part 50, Appendix I, Sec. II.D. The Commission has determined that the value $1,000 per person-rem shall be used in this cost-benefit analysis." Id.

analysis materials fall short with regard to the consideration of the specific matters set forth in 10 C.F.R. §50.109(c). Specifically, the backfit analysis can be characterized as follows:

1. Installation and Continuing Costs Associated With the Backfit Have Been Underestimated (§IV.B below);

2. Potential Impacts on Radiological Exposure of Facility Employees Should Be Further Addressed (§IV.C below);

3. The Relationship to Proposed and Existing Regulatory Requirements Should Be Considered Further (§IV.D below);

4. Potential Impacts of Differences in Facility, Type, Design or Age Should Be Considered Further (§IV.E below); and

5. The Reduction in Risk from Offsite Releases to the Public Has Been Overestimated (§IV.F below).

There is insufficient basis for the Commission to find that the backfitting standard has been satisfied. For this reason, the proposed rule should not be promulgated .

• IV.B Installation And Continuing Costs Associated With The Backfit Has Been Underestimated Section 50.109(c)(5) requires the consideration of installation and continuing costs associated with its proposed backfit, including the cost of facility downtime or the cost of construction delay. The anal ysis in connection with this point is incorrect; the direct and indirect costs of the proposed backfit are underest i mat 9d . When the underestimated costs of the proposed backfit are considered

in light of the overestimated benefits cited as justification for the proposed blackout rule, the findings required under §5O.1O9(c)(5) to support the proposed rule cannot be made. The direct and indirect costs of the proposed backfit are underestimated in two significant respects. First, the number of plants which will be required to make modifications as a result of the rule is underestimated .

• Second, with respect to modifications and coping, the costs that would be incurred to satisfy the proposed rule are not accurately identified. These errors become even more significant when viewed against the past experience with regulatory projections of the costs to licensees of backfits. These items are addressed in turn below.

Number of Plants Affected rs Larger Than Assumed. The technical record has consistently claimed that most plants

• already meet the requirements of the proposed rule. For example, NUREG/CR-3226 suggests that most plants can readily provide the plant f ea tu res (fl:~. , condensate, battery depletion, reactor coolant system isolation, loss of HVAC effects, ~~£- )~p; necessary to cope with blackouts of 4-hours. This theme is developed further in ~UREG-1O32 where
• it is noted that all plants have the ability to remove decay heat for some period. NUREG-11O9, rable 0, 3tates that only 10-15 reactors will need to make modifications.

80/ Se~ gen~-,rally NUREG/CR-3226, Appendices D and E.

                             -  60 -

As discussed in §III.A above, the regulatory analysis should take into consideration the potential that subsequent regulatory interpretations of the rule will evolve as a result of review of the coping analyses (which will be required for all plants, regardless of station blackout risk for each plant). The various interpretations could lead to a significant increase in the number of plants ultimately affected by the proposed station blackout rule. This is best demonstrated by the impact that regulatory interpretations of licensee compliance has had on the number of plants required to make modifications pursuant to the safe shutdown requirements for Appendix R, the TMI Action Plan, and equipment qualification guidelines. In fact, it is difficult to identify any backfits involving licensee analysis that did not eventually lead to additional facility modifications. This is not to imply that additional modifications are never necessary. The

• point simrly is that a sizeable number of unanticipated facility modifications inevitably occur, and, therefore, should be considered in the analysis of the backfit costs.

Activities at Individual Plants Reflect Higher Costs Than Projected. The above mentioned cost e3timate does not take into account all of the costs associat~d with implementing tl1e draft rule at par ~ic ul2r pla nts. There are two issues of concern: (1) underestimation of the cost

of preparing a coping analysis; and (2) errors in the NRC contractor report. First, NUREG-1109 projects that a coping analysis should cost each licensee approximately $150,000. Based on actual cases where such analyses have been completed, or sufficiently scoped, this projection is low by factors of 5-

20. Second, the cost analysis involves a contractor report, NUREG/CR-3840, which is inadequate in the following

• respects:

(1) The use of labor productivity figures are higher than those provided in "Handbook for Cost Estimating," NUREG/CR-3971; (2) The use of unit rates for certain materials that are lower than those rates historically experienced in nuclear construction; (3) The failure to consider design criteria imposed by other Commission requirements (e.g., quality assurance, separation, equipment qualification, etc.); (4) The underestimation by a factor of two or more of project management, engineering, and QA/QC costs using historical data; and, (5) Typographical and mathematical errors in the source document. Based on the above, it is reasonable to conclude that the cost of in1plementing any single hardware-reloted coping feature discussed in NUREG / CR-3840 will cost bet wee~ 2 and 5 times more than estimated. To be sure, this oredi c t io n is not intended to be exact. Rather, it is intended as a more reasonable estimate regarding implementation costs of the station blackout rule.

Costs for Demonstrations of Equipment Qualification and/or Equipment Replacement Have Not Been Considered. The cost analysis should have considered the costs associated with demonstrations that equipment used to cope with a station blackout is qualified to function in the environment associated with the event.Bl/ First, should the analysis require certain equipment to function in a "harsh environment", 10 C.F.R.

 §50.49 would provide the requirements for demonstration of capability of the equipment to function in such environment .

As mentioned above (§III.A), if the rule is interpreted to require full compliance with 10 C.F.R. §50.49 regarding the demonstration of equipment environmental qualification for substantial amounts of equipment not previously subject to

 §50.49, significant costs and resource consumption will result just for this demonstration.        Based upon industry's experience, such demonstrations entail significant costs.

Second, should the coping analysis require the equipment

• to be demonstrated to function in a "mild environment," the lack of such "demonstration" coui'd result in changeout of significant amounts of equipment which is otherwise functionally sound. The costs of additional tests and analysis for equipment which may lack sufficient documentation to demonstrate operability in mild environment;, and changeout of See the draft regulatoly *Juid e 1;:,* )IJJ.j_ . . !i*:*:l with the proposed rule ( at 7), :1hich s e ts f o rth guidance regarding evaluation 1

of equipment operability and reliability during station blackout events.

mild environment equipment should have been considered in the backfit analysis. Therefore, the costs of the proposed rule are underestimated. Concern with the underestimation of costs is deep-rooted. The value-impact analysis supporting the fire protection rulemaking stated: Most licensees already comply with the requirements of the proposed rule and they will incur no additional costs associated with this rule. Those licensees who have not yet complied with the NRC fire protection guidelines will incur some additional capital and operating costs . . . . Enclosure B to SECY-80-88, "Fire Protection Actions," February 13, 1980, at 5. Notwithstanding those optimistic assurances, the actual costs incurred by individual utilities have ranged from approximately $5,300,000 to $20,000,000 per plant site.82/ The value-impact analysis accompanying the proposed emergency planning rule stated as follows regarding the typical costs for State and local government programs to achieve* adequacy in radiological emergency response plans for a 10-mile Emergency Planning Zone: For a State, the initial costs of planning, exercises, training and resources (communication and radiation monitoring instrumentation) will typically total ahout $2d0,000 with associated annual updatlno cost~ r~ about

           $ ,1 4 , 0 0 0 . Po c* lo ca 1 gov c 1: " " '::' n t .'° 1 :~ h, ~

82/ Based on informal contacts with nuclear utility members of NUGSBO.

initial costs typically total about

            $120,000 (four jurisdictions) with annual updating costs of about $30,000.        The typical total costs to State and local governments to obtain an NRC finding of adequacy in their emergency response plans would be about $360,000 initial costs plus $74,000 in annual updating costs.

NUREG-0685, "Environmental Assessment for Effective Changes to 10 C.F.R. Part 50 and Appendix E to 10 C.F.R. Part 50; Emergency Planning Requirements for Nuclear Power Plants," (August 1980) at 5, 7. In addition, the one-time cost of

 $500,000 per facility was estimated for the public notification system.          In actuality, initial State and local government costs have ranged from $170,000 to $1,714,500.

Annual state and local updating costs have ranged from

 $135,000 to $740,000.          Initial utility costs for the public notification system ranged from $400,000 to $2,400,000.           In addition, other one-time utility expenses have in certain instances exceeded $2,000,000.          Annual utility costs range from $125,000 to $667,000.~~/

• IV.C Potential Impacts on Radiologi~al Exposure of faci_li _ty__E~plo_yees Should be Further Addressed Section 50.109(c)(4) requires the regulatory analysis to address the potential impact of proposed backfits on the radiological exposure of facility employees. NUREG-1109 states (at Table 8, n. 2):

 ?J/  Cost fi gures a re based on informal contacts with NUGSBO members and are based on costs per site.

No significant increase in occupational exposure is expected from operation and maintenance or implementing the recommendations proposed in this resolution. Equipment additions and modifications contemplated do not require significant work in or around the reactor coolant system and therefore would not be expected to result in significant radiation exposure. This statement is questioned in two respects. First, licensee actions required to.implement a design basis coping capability of 4- to 8-hours will probably encompass facility

• modifications in and around the reactor coolant system (or other contaminated systems). For example, valve operators may require replacement, new instrumentation may need to be installed and surveillance of station blackout systems will have to be implemented. All of these activities are normally associated with hardware backfits of the type that have contributed to the rapid growth in incidental onsite exposure experienced since the early 1970s .
• According to a recently published industry report, prepared by the Nationai Environment Studies Project of the Atomic Industrial Forum,84/ during the period 1979 through 1983 forty percent of the total occupational exposure at U.S. light water reactors was attributable to NRC-initiated multi-plant 84/ AIF/NF:SP*-033, "Occupational Radiation Exposure Implications of NRC-lnitiated Multi-Plant Actions," March 1986.

actions.~~/ This percentage represents approximately 99,000 person-rem of collective exposure.86 / Second, one of the estimated financial costs for licensees to comply with the draft rule is the resolution of the reactor coolant pump seal integrity issue.87/ Notwithstandjng the above discussion (§II.B.2), if licensees have to replace or modify their existing seals in order to resolve the reactor coolant pump seal integrity issue, workers will be exposed to limited occupational doses. Thus, if the the financial impacts of seal resolution are to be considered, the radiologi c al impacts of this backfit should also be considered . 85 / Id. at vi . See NUREG-07 48 for a listing of so-called "r1ul ti--Plant--Ac tions. II 86 / A major contributor (8.8 i of the total, o r 8712 person-rem) was licensee actions to ~chie ve co mpli a~ce with NRC fire protection requi r ements, yet the regul ator y an a l y sis for those requirements did not ac c ount f or a ny oc cupational exposure amount s. Se e En c l o s u r e B t0 S~CY 88 ,

       "Valu e / I mpact Assessment of P L"')l_)U S P.(1 r> i_ r*~ Pr o tecti o n Rul e ,"

February 13, 1980. _87/ NUREG-1109 at Table 6, n. 5.

IV.D The Relationship to Proposed and Existing Regulatory Requirements Should Be Considered Further The backfitting rule requires consideration of the relationship of a proposed backfit with other proposed and existing regulatory requirements. 10 C.F.R. §50.109(c)(6f. Such a relationship is acknowledged between station blackout and the two issues of diesel generator reliability and reactor coolant pump seal integrity. However, these issues are being

• addressed through other regulatory initiatives. Even if concerns on these issues were substantiated(~., see §II.B.2 above regarding clarifications of the RCP seal issue), when these separate initiatives are completed, the need for the proposed station blackout rule would appear to be further diminished.

DiesPl generator reliability is central to any resolution of the station blackout issue. While the proposed rule reflects minimum standards in this regard, there is no account for the fact that, to the extent that there is an improvement in the diesel generator reliability, the benefits claimed from the proposed rule (and, thus, the need for a rule at all) are dissipated.88_/ As mentioned above, the benefits of the proposed rule are also strongly dependent on the resolution of the reactor

 ?_8/ Initiativi=>s advanc0rJ by industry are designed to improve diesel generator performance.

coolant pump seal integrity issue. When the proposed rule was drafted, current data suggested that certain reactor coolant pump seals may be susceptihle to catastrophic failure should cooling be lost. Thus, the proposed rule places strong emphasis on coping and the value of seal injection. However, as discussed in §II.B.2 above, significant new information has recently become available. Based on this information, it is now apparent that the risk of catastrophic seal failure is significantly less than previously thought . In short, the backfitting rule was intended to ensure that proposed modifications be related to other ongoing regulatory activities in order to avoid unnecessary expenditures of NRC and licensee resources. The failure in the backfit analysis to consider realistically both diesel generator reliability improvements and recent developments regarding seal integrity, results in failure to take into account these ongoing regulatory activities .

• IV.E Potential Impacts of Differences in Facility Type, Design or Ag~_Should Be Considered Further Section 50.109(c)(8) requires the consideration of the potential impact of differences in facility type, design or age on the relevancy and practicality of the proposed backfit.

Table 6 in NUREG 1109 indicates that betwe~~ 42 and 57 plants need not do anything to improve diesel gener a tor reliability or coping capability in order to c omp.l }" with :: h,'° rule. Thus, the benefits of the proposed rule will go to the 15 plants needing

to upgrade diesel generator reliability or to the 10 plants needing to increase coping capabilities. It therefore follows that 42-57 plants will enjoy little or no benefits from this rule but each will spend at least $150,000 89/ to document their compliance. Clearly, for this population of plants, there will be significant costs with minimal corresponding benefits. Under these circumstances, where the risks associated with particular plants are not, in fact, spread evenly throughout the industry, it is believed that unique plant-specific factors were artificially and unrealistically ignored in order to justify a favorable cost-benefit ratio. Thus, compliance with Section 50.109(c)(6) has not been achieved. IV.F The Reduction In Risk From Offsite Releases To The Public Has Been Overestimated The backfitting rule requires the "potential change in the risk to the public from the accidental offsite release of

• radioactive material" be addressed. 10 C.F.R. §50.109(c)(3).

As described above in §II.A, the representation made in NUREG-1109 was that the estimated total risk reduction achievable from the proposed resolution of UST A-44 is 80,000 person-rem, assuming an average remaining plant life of 25 years. This estimate represents the benefit for the pro 9o sed rule. 89 / $150,U00 represents the hest estimate for a coping study. NUREG-1109, T~hle 6. As stated ahove, this figure could hav~ h0on underestimated hy a factor of 5-20.

It is not readily apparent how this figure was derived. NUREG-1109, Table 5 provides three examples of the reduction in frequency of core melt per reactor year which is attributed to the proposed rule. Table 4 identifies 3 groups of plants, totalling 67, which form the basis for the calculation of a total dose reduction of 80,000 person-rems attributed to the proposed rule. However, there is no disclosure whether a relationship exists between the three frequencies in Table 5 and three groups of plants in Table 4 . Assuming some rational relationship exists between Tables 4 and 5, i t appears nevertheless that the risk of a station blackout event is overestimated and, therefore, the expected benefits o f the proposed rule (in terms of person-rem averted) are also overestimated. This is apparent because a number of factors were not considered which (1) reduce the probability of a station blackout (see §IV.F.1 below) and (2) reduce the consequences of station blackout (see §IV.F.2 below) .

• IV.F.1 Factors Reducing the Probability of Station Blackout Require Further Attention More full account in the backfit analysis should be made for a number of factors that, as a practical matter, reduce the risk of a station bla c kout. These factors ~ re (1) emergency diesel generator configurations; (2) emergency procedures; and (3) additional power sources. Wh e n t he s 0 foctors are taken into account, the probability of an off-site release in connection with a station blackout is much less significant than that postulated by the Staff.

Diesel _Generator Conf~guration. The first factor is present in the category of plants which pose the greatest risk based on the conclusion in Table 5: two unit plants sharing 2 of 3 (2 / 3) emergency diesel generators to respond to a loss of coolant accident. It appears from Tables 2 and 4 in NUREG-1109, that 14 plants at 7 sites fall in this category (i.e., Group C). Assuming that the proposed rule is implemented and that all 14 plants now have a 2 hour station blackout coping capability, NUREG-1109 indicates that an estimated risk reduction in core damage frequency of 14 x 10- 5 per reactor-year (RY)- 1 , will be achieved at each plant. By comparison, the other two plant categories will achieve reductions in core damage frequency of 1.3 x 10 -5 (RY) -1 and 0.1 x 10 -5 (RY) -1 , respectively. While these 14 "higher risk" plants represent only 21% of the facilities considered in NUREG-1109 Table 5, by implication approximately 78 ~ of the expected offsite consequences of station blackout which can be averted by the rule is attributed to these sites .

• Apparently this concern is based ultimately on the observation that these two-unit plants do not have as much redundancy in the emergency AC power system as do most other plants . In a station blackout event the magnitude of electrical loads necessary to provide core ~ooling is significantly less than that required under design basis accident conditions. Inde ed , su~h 0~ci~~nts are specifically excluded from coping consideration in the proposed rule. As

previously discussed, in several of these 2/3, two-unit plants, it is possible to provide AC power to both units in a station blackout while using only a single emergency diesel generator, effectively making the emergency diesel configuration 1/3. This can be done by shedding non-essential loads and/or by cross-connecting the diesel generators servicing each unit in a two-unit plant. The impact of this effective 1/3 configuration on the regulatory analysis is dramatic. The reliability of a 1/3 system is approximately 4.5 to as much as 13 times greater than a 2/3 system, assuming individual component reliabilities of

 .95-.99 per demand. Most of these plants are operating their diesels at reliabilities of above 0.97 per demand.              Crediting these improvements in AC power reliability significantly reduces the station blackout risk at the very sites the regulatory analysis suggest pose the greatest risk.              When the number of effective 1/3 configurations is taken into account,

• the risk of station blackout is not as great as that suggested in the regulatory analysis.

Emergency Procedures. In assessing the risk of off-site power loss it was concluded that severe weather was a prime contributor, particularly with respect to '=-,-ent durc1tion. (Indeed, sevcrA wea t her is cr i tical to deter~ining minimum coping duratil)nS and underlies the 1_1 s e r:- f: *; 1Jµing to resol v e station blackout.) This conclusion does not give full credit

for the potential benefits of onsite emergency procedures (i.e., as requested by Generic Letter 81-04, "Emergency Procedures and Training for Station Blackout Events," February 25, 1981) which are designed to address extremely severe weather conditions.90/ The backfit analysis does not reflect the procedures licensees have in place to reduce power, verify emergency equipment operability, and shutdown their units in the event of severe and extremely severe weather, such as hurricanes. The effectiveness of these procedures was most recently demonstrated by Hurricane Gloria. In every instance where the storm posed a risk, licensees took appropriate action to safely respond to the severe weather.91/ Additional Power Sources. The assessment of loss of on-site power is not realistic due to its undue focus on emergency diesel generator reliability. Other alternate sources are present either at or nearby nuclear power plants, such as fire protection diesel generators, gas turbines, black-start diesel

• generators and steam driven pumps.

relied upon in an emergency, These power sources can be thereby enhancing onsite power and decreasing the frequency and consequences of loss of onsite power. Information submitted by licensees in response to a request based on 10 C.F.R. §50.54(f) for example, could be used 90/ In addition to such procedures, the initiative advanced by industry provides for severe w~?ther rr~c~dures. 21_1 See discussion of Hurricane Gloria in §II.B.3, supra. See

      §IV.F.2 below for cl discussion of emergency planning factors designed to reduce the consequences of offsite releases.

to demonstrate the availability of these additional power sources.92 / IV.F.2 Factors Reducing The Consequences of Offsite Releases Require Further Attention Notwithstanding the above discussion of the failure to account fully for flexibility in emergency AC power system designs and other emergency procedures, the consequences of offsite releases that would result from a station blackout event are overestimated. First, inappropriate assumptions were made which overestimated the expected impact of early containment failure on station blackout risk. Second, information to the effect that the source term can be reduced by at least two orders of magnitude should have been considered more explicitly. Third, for the reason mentioned above (§II.A.2), the SSTl release category is inappropriate as used by the Staff to represent offsite consequences of accident sequences involving a station blackout . Fourth, the

• contribution that implementation of the Commission's emergency planning regulations makes to the reduction of population exposure has been ignored. These items are considered in turn below .

   ~1/  A utility demonstration wo uld i'.J *?nti.*~ :~* the power sour c e and provide assurance of its availability. Such assurance could entail description of maintenance, testing, reliability, and procedures to avoid common cause failure.

• Inappropriateness of Containment Performance Assumption.

The expected impact of early containment failure on station blackout risk has been overestimated. As explained fully in

 §II.A.2, the siting source term, SSTl, has been used improperly to model the offsite consequences of a station blackout accident. As explaiQed above, SSTl postulates a containment failure in 1.5 hours after the onset of core damage.            NUREG/CR-2239 at 2-13. To be sure, the expected SSTl release is discounted by a factor of 3 to account for the fact that the SSTl was never intended for use with station blackout events, and the importance of accounting for the effect of containment integrity in delaying and mitigating the release is acknowledged. _~ /  The effect of such delay and mitigation is to reduce both the energy of the assumed release and its contents.

However, NUREG-1109 provides no specific information as to how long containments would be assumed to remain intact in a station blackout event. Yet this is a critical factor because it is directly related to person-rem exposure.2i/ The

• subjective use of a reduction factor of 3 without any relation to containment integrity is not justified. See §II.A.2 above.

NUREG-1032 does provide what is characterized as some "insights" on this question. In §7.0 of NUREG-1032, IDCOR containment failure times are compared with NRC Staff times in most instances . It i s stated ( at 7 -1 5 , 1 7 ) that recent 2.11 NUREG-1109 at 8. 94/ NUREG-1032 at 7-15.

estimates of containment performance such as those produced by IDCOR "may be cause for revision" of prior published NRC results. However, no attempt is made in NUREG-1109 to reconcile the large disparity between prior NRC estimates for containment failure times and those of IDCOR. These "insights" show that the value of containment has been underestimated in mitigating reactor accidents. For example, it is stated that large dry containments can fail in as little as 10 hours. The IDCOR study suggests that the failure time for these containments can be at least 18 hours. NUREG-1032 indicates (at Figure A.1) that the likelihood of offsite AC power being restored in 10 hours or less approaches

1. Restoring AC power will return to service those containment cooling systems which will assure containment integrity.

ThereforP, the NUREG-1032 analysis demonstrates that the potential tor containment failure associated with failure to restore AC power should be very small, regardless of whether

• containment integrity is maintained for 10 hours, as the Staff suggests, or 18 hours, as IDCOR found. Accordingly, the expected impact of containment failure is overestimated in the regulatory analysis.

New Source Term _Data_~hould Be Reli~d ~pon. It was improper to base the value-impact ratio sol ~l y on the SSTl release C <:tC-i CJCHy. DuL* ing a Mu rc h - ,~ 1,:,*> *=* )m mi ssi on meeting the Commis si on was advised (Transcript at 78) of the latest

                                   -  77 -

schedule for completion of source term related activities. In addition, a draft of NUREG-1150 is scheduled to be issued in the late summer of 1986. This NUREG will, among other things, provide the results of accident analyses using the new source term assumptions. This information will provide the basis for the Commission to issue new direction concerning the risk of severe reactor accidents. Most of the analyses in this regard have been completed and are currently available .

• In fact, during the March 26 Commission meeting, Dr.

David Torgenson, from the OECD Committee on the Safety of Nuclear Installation's Special Task Force on Source Terms, told the Commission: The thing I want to point out is the reduction in the source term. This is the old technology here for Surry in Wash 1400. This is the new technology that is coming out of RMI 2104. You can see that the source term has been reduced by more than two orders of magnitude.2_~/ Zoltan Rosztoczy also stated: The programs in the Office of Nuclear Reactor Regulation. .have to be closely coordinated and coupled with the unresolved and generic safety issues such . . as station blackout . .96 / We agree. Under these circumstances, this potential new direction should have been accounted fo~ in the

• regulatory analysis. Section 7.0 of NUREG-1032, states (at 95/ Id. at 10.

_9-_~/ Id. at 71.

7-18) that the "reader is cautioned that ongoing research could cause substantial revision" in fission product release fractions utilized in the analysis. However, an explanation should have been provided in NUREG-1109 as to the potential effect such substantial revisions will have on the validity of the regulatory analysis accompanying the proposed rule. The Commiss i on should not act on any final rule until this information is available.97/

• Inappropriateness of Fission Product Inventory Assumption. As discussed above in §II.A.2, the use of SSTl as a generic assumption applicable for all plant containment types when d e termining the avoided person-rems is inappropr ia te. A better approach would be to select r elease categorie s on the basis of the different containment types, failure moctes an d station bla c kout accident sequences.98 /

For e xamp l e , t he c haracteristics of an SSTl are containment failure and r e lease time at 1.5 hours, with release duration

• of 2 hours. 99/ If the assumed energy release rate is high, wider dispersion of radioactive materials and a higher curie 97 / NUREG-1150, when published, will be reviewed and comments will be provided as appropriate. Indeed, an argument could be made that the st a tion bl a c ko ut rulemaking should be deferred until a t l ea st su c h time as comme nts can be made on NUREG-1 150 .

 ~-~/ NUREG --1032 odd re ss Ps thes e factrns ( ,:1t 7- 19 ) . Howev e r ,

NUREG-J. 10 9 u ti l iz es t he SS Tl r e l ,-,asP r: ,:i 1~<::gory ( red u ce d by 1/ 3 ) 0 n o ge neric basis f or 2! 1 ~l 2~~s to d i sco unt SSTl rel0as~s (at 50 mi les) f o r s t at i on blackou t ev ents. 9~/ NURE G/ CR-2239 ot 2-13.

content of that material which is released will occur.100/ These characteristics are in direct contrast to the risk-dominant accident sequences associated with station blackout.\01/ Specifically, these accident sequences can all be characterized -- not by the SSTl containment failure mode, but by a containment failure later than 1-1/2 hours, a lower energy release, and a significantly reduced curie content in the released materials.102/ Moreover, if SSTl is used when SST2 assumptions are more appropriate, an error of more than 2 orders of magnitude can be injected into the averted person-rem calculation and, thus, the risk-reduction benefits of implementing the proposed rule would be overestimated. For example, if an SST2 or SST3 release had been assumed at the outset instead of SSTl, the estimated offsite release associated with station blackout could have been lower by as much as 4 orders of magnitude. The practical

• impact of doing so would have been to reduce averted person-l_OO/ Id. at 2-73.

101/ As e~plained in §II.A.2 above, while early release could be postulated with a mechanism such as steam explosion or direct heatt9g, the frequency of such a release is so small (<10 per RY) thct it does not contribute to the risk from station blackout, (i.e., frequency x consequence). The risk do~inant sequence is the long term overpressuri z ation of containment, 24 hours to 72 hours after core melt. 102/ In this regard, it is not clear from the analysis in NUREG-1109 what energy release rates were used in conjunction with the SSTl assumption.

rems -- not merely by the value of one-third cited in NUREG-1032, but by a factor of 100 or more -- as a result of the highly non-linear relationship among the various categories of releases, SST1-SST5. See §II.A.2 above. This is not to imply that the SST2 category, or a modified SST2 category, should have been relied upon in all cases instead of an appropriately modified SSTl. However, a regulatory analysis should provide a sound, articulated technical basis for its selection of key assumptions. In this case, instead of attempting to fit the station blackout issue into a preconceived fission product release category, such as SSTl, the most appropriate technique would have been to develop a mechanistic model for calculating the potential for offsite consequences due to station blackout accident sequence, including expected containment failure mode and failure times, and to use that model in the regulatory analysis supporting the proposed blackout rule .

• Alternatively, a more thorough offsite consequences analysis could have been provided, which would reconcile the predicted offsite release with the various SST categories selected and match the release characterist i cs ( ~ , plant effects) of a station blackout accident with o ne of the pre-defined source terms, such as SST2. Ins tea d the analysis in NUREGs-1032 and -1109 appar e ntl y h a ve 1 *3~J the most conservati ve s o urc e t erm yet identified with o ut presenting an adequ ~ tP t e chni c al b a sis for the choice.

Accordingly, the averted person-rems to be derived from the proposed rule are overestimated and the underlying basis for key assumptions in identifying these averted person-rems is not established. Contributions of Emergency Plans to Dose Reduction. The estimation of 80,000 averted person-rems should have accounted for the dose reductions resulting from compliance with the Commission's emergency planning regulations. 10 C.F.R. §50.47 and 10 C.F.R. Part 50, Appendix E. The Statement of Considerations accompanying the emergency planning rule states that such rule was promulgated to protect the public health and safety. 45 Fe~. Reg. 55402 (1980). Furthermore, NUREG-0396,103/ an authoritative source document in emergency planning, specifically states (at 5) that: [T]he objective of emergency response plans should be to provide dose savings for a spectrum of accidents that . could produce offsite doses . Accordingly, it can be seen, that in complying with the emergency planning regulations, a "dose savings results. This "dose savings" should have been deducted from any estimate of person-rem avArted as a result of implementation of the proposed rule. 103/ "Planning Basis For The Development of State and Local Government Radiological Emergency Response Plans In Supp o rt of Light Water Nuclear Power Plants." NUREG-0396 (1978).

IV.G Conclusion A number of problems have been identified in the backfit analysis developed in support of the proposed rule. Resolving only two of the problems significantly alters the value/impact ratio, raising a question as to whether the proposed rule meets the backfitting standard. For example, Table 5 in NUREG-1109 implies that 78% of the 80,000 person-rem are at the 14 plants with a 2/3 emergency diesel

• generator configuration. That risk (62,400 person-rem) diminishes by a factor of from about 4 to 13 times if all plants in this category can be converted to a 1/3 configuration. Using a 4.5 factor as representative of the risk-reduction value of these conversions, the 80,000 man-rem projected savings by the Staff could be reduced by as much as 31,500 person-rem.

In addition, the total projected costs of implementing

• the proposed backfit amounts to in excess of $80 million.

Thus, the value/impact ratio of the draft rule, even without discounting for containment performance and source terms, could be closer to $2,500 per person-rem averted. Proper treatment of containment perfo~mance and the use of more realistic source terms, coupled with consideration of more realistic cc:: t~5 , s 1' .) t'J.d result in a value / impact ratio of at least $50,000 per person-rem

averted and could well exceed this amount.

• On these bases,

 )

it is clear that the backfitting rule has not been satisfied and thus the proposed rule should not be implemented. In view of these errors, adequate justification has not been provided in support of the position that the backfitting standard has been satisfied. Thus, the Commission should not proceed with this rulemaking .

v. Objections to the Proposed Rule These comments can be summarized in four key arguments:

(1) The bulk of perceived station blackout risk is concentrated at a very small number of sites having distinctly identifiable design features -- thus, generic rulernaking is not the best mechanism to address an issue that involves a small percentage of the industry;

 * (2 )   Even if the Commission concludes a generic rule is appropriate, the technical basis does not support the rule; (3)    Even if the underlying technical basis did support the rule, the proposed rule itself should be reevaluated as a regulatory tool for it is likely to fall short of the goal of consistent and efficient resolution of the
*  (4) station blackout concern; and, Notwithstanding the above, the proposed rule does not meet the standards of the Commission's backfit rule --

either in the net value of the anticioated modifications or in the completeness of the analysis. Any one of the above conclusi o ns i ~ sufficient to preclude enactment of the rule currently before the Commission.

                             -' 85 -

If there is particular concern, then any regulatory action considered must be directed at the concern. In this instance, an issue has been raised as to the availability of AC power sources and the potential that station blackout poses an unacceptable risk at an unknown number of plants. Yet, the proposed rule seeks to identify the coping capability of operating plants and impose minimum durations which plants must provide as part of their design basis. It appears that the proposed solution does not address the stated problem in an

• effective manner.

A rule is not needed to identify whether or not station blackout risk is unacceptable at a particular plant. Other mechanisms are available for obtaining this information without the need fnr the Commission to utilize the rulemaking process. Similarly, concern about the availability of AC power sources suggests that the best solution to the problem is to enhance that reliability.104/

• The effectiveness of a generic rulemaking in addressing so isolated an issue as station blackout is a concern.

However, it is recognized that the Commission may believe that rulemaking satisfies other imperatives which transcend the ability to demonstrate public health and s2fety benefits across all of industry. Should such imperatives exist within the 10~/ Initiatives advanced by industry are designed to enhance reliability.

Commission, then the proposed rule should focus on the concerns raised. If the Commission believes a rule is necessary, guidelines should be prepared concerning what constitutes acceptable.emergency and non-emergency AC power systems for preventing or mitigating station blackout events. Without such guidance, a rule which requires AC power-independent coping is subjective, difficult to implement and of questionable value

• with respect to providing real safety margins.

Such a shift in focus from compensatory measures to problem resolution would entail modifying several aspects of the proposed rule. To this end, the kinds of changes we would suggest include: (1) expanding the focus of the rule and draft regulatory guide to include explicit and equally weighted provision for crediting onsite and offsite backup AC powP. r sources or improvements to AC power system availability in responding to the loss of normal and emergency AC power; (2) deleting the addition to the Gener a l Design Criterion of an AC-independent coping requi r em e n t and the associat e d c oping ana ly ses):0 5_.

• 105/ The propos e d rule would make GDC-17 unique among other (Footnote 105 Continued on Next Page

(3) eliminating the open-ended coping demonstration aspects of the regulatory guide and substituting a simpler closed-form coping capability checklist; (4) considering the potential ripple effects of new station blackout requirements, particularly those involving AC-independent coping, on other areas of nuclear regulation, such as the potential need for equipment qualification of station blackout coping

• equipment by the methods called for in 10 C.F.R.

       §50.49; and (5)   eliminating the need to consider severe weather as a dominant factor in establishing whether a plant is in a 4-hour or 8-hour coping duration category (i.e.,

NUREG-1109 at Tables 1 and 3) . (Footnote 105 Continued from Previous Page) General Design Criteria, in that no other GDC explicitly negates the availability of AC power.

VIROINIA ELECTRIC AND POWER COMPA RICHMOND, VIRGINIA 23261

w. L . STHWABT

                                         ~::019,t198R~W \

VICE PRBS I DB?fT N u cLBA.R 0PBRAT101"8 The Secretary of the Commission 57 Fl< q S~~t) No. 86-233 Attn: Docketing and Service Branch NO / ALM/ acm U.S. Nuclear Regulatory Commission Docket Nos. 50-280 Washington, D.C. 20555 50-281 50-338 50-339 Gentlemen: License Nos. DPR-32 DPR-37 NPF-4 NPF-7 VIRGINIA ELECTRIC AND POWER COMPANY SURRY AND NORTH ANNA POWER STATIONS PROPOSED RULE ON STATION BLACKOUT Virginia Electric and Power Company has reviewed the proposed rule on station blackout. Our comments on this issue were incorporated into the Nuclear Utility Management and Resources Committee (NUMARC) comment document and we endorse this document as being consistent with our views. Additionally, we support the industry initiatives that NUMARC has proposed to address the more important contributors to station blackout. Finally, we believe rulemaking is not necess ary and the previously referenced industry initiatives, when fully implemented, should resolve your concerns with respect to station blackout. Very truly yours , W. L. Stewart

lJ s.

VIBOINIA ELECTRIC AND PoWEB COMPANY TO The Secretary of the Commission cc: NRC Senior Resident Inspector Surry Power Station NRC Senior Resident Inspector North Anna Power Station Mr. Chandu P. Patel NRC Surry Project Manager PWR Project Directorate No. 2 Division of PWR Licensing-A Mr. Leon B. Engle NRC North Anna Project Manager PWR Project Directorate No.

• 2 Division of PWR Licensing-A

.. u- 600609 130- 86 (06 -23 )-1 lA-120 lll/NOIS POWER COMPANY Docket No. 50-461 Secretary of the Commission Attention: Samuel J. Chilk Docketing and Service Branch U.S. Nuclear Regulatory Commission Washington, D.C. 20555

Subject:

Clinton Power Station Comments on Proposed Rule to 10CFR Part 50 - Station Blackout

Dear Mr. Chilk:

Illinois Power (IP) has reviewed the proposed rule "Station Blackout" as published for comment in the Federal Register dated March 21, 1986, and the draft regulatory guide "Station Blackout" published for comment in March 1986. IP has the following comments:

1) The draft regulatory guide does not specify whether the equipment needed for station blackout (SBO) mitigation must meet single failure criteria. Must failures, in addition to the failures that initiated the station blackout, be considered when implementing the rule?

2) Section C.3.3, Item 3.a, of the proposed regulatory guide states that if a system is required for primary coolant charging and make up, reactor coolant pump seal cooling or injection, or decay heat removal specifically to meet the recommended station blackout duration, then the system should be capable of being actuated and controlled from the control room. Implementation of this requirement could be expensive and may not be cost effective. The actuation and control of these systems can in certain instances be accomplished through the use of effective emergency procedures that specify control actions to be taken such as manual jumpering of control logics or manual operation of valves. Therefore, IP believes that Section C.3.3, Item 3.a, of the proposed regulatory guide should be modified to read:

Acknawledgedbycar.i .JJ~~ ~

        ' u s.

Po Ad I

          ~IJt

u- 600609 L30-86(06-23)-L lA-120 "a. The system should be capable of being actuated and controlled from the Control Room, or if other means of control are required (e.g., manual jumpering of control logics or manual operation of valves), it should be demonstrated that these steps can be carried out in a timely fashion." Please contact me if there are any questions. Sincerely yours,

                                     ~-~-~

Manager - Licensing and Safety TLR/pjr cc: B. L. Siegel, NRC Clinton Licensing Project Manager NRC Resident Office Regional Administrator, Region III, USNRC Illinois Department of Nuclear Safety

Georgia Power Company 333 Piedmont Avenue Atlanta, Georgia 303Q8 Telephone 404 526-6526 Mailing Address: Post Office Box 4545 Atlanta, Georgia 30302 J0CK£T NU BER PROPOSED RULE PR- 50 ,\ Georgia Power L. T. Gucwa Manager Nuclear Safety and Licensing

                                     @I {'1C.C/f~

June 19, 198 Mr. Samual J. Chilk Secretary of the Co11111ission U. S. Nuclear Regulatory Commission Washington, D. c. 20555 Attention: Docketing and Service Branch Georgia Power Company Co11111ents On Proposed Rulemaking Regarding Station Blackout

Dear Mr. Chilk:

On March 21, 1986, the Nuclear Regulatory Commission (NRC) published in the Federal Register proposed station blackout rulemaking for public co1T111ent. The proposed rulemaking would define the tenn "station blackout" and would require commercial nuclear power plants to withstand and recover from a total loss of alternating current (AC) electric power (called "station blackout") for a specified duration. Georgia Power Company (GPC) has reviewed the proposed rulemaking and has the following comments. General Comments The proposed rulemaking concentrates on the duration to cope with a station blackout and not on the reduction in the probability of the incidents of station blackouts. GPC believes a practical approach to reducing the risk of the consequences of a station blackout would be through an AC power reliability program. The improvement in the reliability and perfonnance of existing emergency AC power sources could provide a more practical and economical benefit without the new costs and obligations of a restrictive rule. If station blackout rulemaking is detennined by the NRC to be necessary, GPC strongly recommends that such rul emaking provide for an AC reliability program as an alternative to coping with a station blackout for a specified duration. Industry Initiatives Georgia Power Company is currently a participant in the Nuclear Utility Group on Station Blackout (NUGSBO) and the Nuclear Utility Management and ACkn I dged by car<l7./~~/a**"a.at

' u* $. NU CI N DOCf 0 Por Co, Adel Spc

Georgia Power , \. U. S. Nuclear Regulatory Commission June 19, 1986 Page 2 Human Resources Committee (NUMARC). The NUMARC Station Blackout Working Group, with NUGSBO serving as its technical consultant, has made numerous comments to the NRG staff regarding the proposed rulemaking, backfit analysis and regulatory analysis. Based upon the NRC's analysis methodology, NUMARC has determined that the risk of a station blackout appears to be dominated by the redundancy and reliability of emergency AC power systems. The review of the analysis indicates that almost 80% of the anticipated benefits of the proposed station blackout rulemaking would be derived from modifications to plants with less redundant emergency AC power systems. In addition, there may be sites that have low emergency diesel generator reliability, are vulnerable to severe weather, and/or have less than desirable off site AC power system configurations. Consequently, NUMARC has approved certain industry initiatives in the belief that the station blackout issue is not generic to the nuclear industry and, thus, need not be resolved by rulemaking. The industry i ni ti atives approved by NUMARC are as follows:

1. Each utility will review their site(s) against the criteria specified in NUREG-1109, and if the site(s) falls into the category of an eight-hour site, the utility will take actions to reduce the site's contribution to the overall risk of station blackout to the extent possible.

2. Each utility will implement procedures at each of its site(s) for:

• a.

b. c. coping with a station blackout event, restoration of AC power following a station blackout event, and preparing the plant for severe weather conditions, such as hurricanes and tornadoes, to reduce the likelihood of a loss of offsite power and to reduce the overall risk of a station blackout event.

3. Each utility will, if applicable, reduce or eliminate cold fast-starts of emergency diesel generators for testing through changes to technical specifications or other appropriate means.

4. Each utility will monitor emergency AC power unavailability utilizing data utilities provide to INPO on a regular basis.

4297A 700775

Georgia Power , \ U.S. Nuclear Regulatory Commission June 19, 1986 Page 3 Georgia Power Company supports these industry initiatives approved by NUMARC and the NUMARC conclusion that the station blackout issue is a concern to a limited number of plants and should be resolved in a manner other than rulemaking. However, if the NRC determines . that station blackout rulemaking is necessary, GPC strongly recommends that the proposed rule, 10 CFR 50.63, and the proposed station blackout definition in 10 CFR 50.2 be revised. Comments on the Proposed Rule 10 CFR 50.63, "Loss of all Alternating Current Power" The proposed revision to 10 CFR 50, Appendix A, General Design Criteria 17 (GDC 17) would require that nuclear power plants cope with a station blackout for a specified duration and that this duration would be determined by a consideration of four factors. The proposed rule, however, allows the NRC to notify the licensee of the duration it must cope with in order to satisfy GDC 17 based on plant-specific analyses and does not establish the bounds for coping with a station blackout. The ultimate requirements of the proposed rule are uncertain and are based on the discretion of the NRC. As written, the proposed rule will not achieve a consistent or efficient resolution of the station blackout issue. GPC recommends that the proposed rule itself be revised to specify additional guidance in determining the minimum duration to cope with a station blackout. Addi ti anally , the proposed rule requires each 1i censee to inform the NRC of the maximum duration that the plant, as currently designed, could cope with a station blackout. The plant-specific analysis to determine this maximum duration beyond the minimum specified duration (e.g., 4 or 8 hours) would be costly and would not serve to reduce the risk from a station blackout. Rather, GPC recommends that the proposed rule require an analysis only to the minimum specified duration required to satisfy GDC 17. The proposed rule also does not specify the design requirements for equipment to cope with a station blackout. The assumptions for a station blackout go beyond the normal single failure criterion and involve not only a loss of all offsite AC power but the loss of both trains of safety-related onsite AC power. Because of these unusual assumptions, GPC believes that the 4297A 700775

Georgia Power ,.\ U. S. Nuclear Regulatory Commission June 19, 1986 Page 4 equipment necessary to cope with a station blackout need not be safety-grade or environmentally qualified in accordance with 10 CFR 50.49, but rather this equipment should be commercial grade. Additionally, operator action outside the control room should be acceptable, since a radiological release would not be involved. GPC strongly recommends the proposed rule be so revised. Comments on the Definition of Station Blackout

• The proposed amendment to 10 CFR 50.2 would add the following definition:

              "Station blackout" means the complete loss of alternating current (AC) electric power to the essential and nonessential switchgear buses in a nuclear power plant (i.e., loss of the offsite electric power system concurrent with turbine trip and unavailability of the onsite emergency AC power system).

Georgia Power Company believes that the phrase 11 unavailability of the onsi te emergency AC power system 11 could be interpreted as being too restrictive and might not allow the AC power available from the station batteries through inverters in coping with a station blackout. This AC power from the station batteries could be used in the event of a station blackout to power instrumentation, room cooling fans and other plant equipment. Exclusion of this reliable power source would have a significant impact on the new costs and obligations of the proposed station blackout rulemaking and would not represent a realistic approach in coping with a station blackout event .

• In order to clarify a potential misinterpretation, GPC recommends this phrase of the station blackout definition be revised to "unavailability of the onsite emergency AC power system, but not a loss of the available AC power from the station batteries through inverters."

Conclusion Georgi a Power Company concurs with the NUMARC position that the station blackout issue is a concern for a 1imi ted number of pl ants and should be resolved in a manner other than rulemaking. A more practical and economical approach would be the industry initiatives approved by NUMARC. This would involve an approach that balances both preventing station blackouts and coping with their consequences. 4297A 700775

Georgia Power , \ U. S. Nuclear Regulatory Commission June 19, 1986 Page 5 If industry initiatives are not acceptable to the NRC, then the proposed rulemaking should be revised to:

1) provide for an AC reliability program as an alternative to the capability to cope with the station blackout for a specified duration;

• 2) 3)

specify additional guidance in determining the minimum duration to cope with a station blackout; require a review of the pl ant, as currently designed, to cope with the minimum specified duration only, not the maximum duration;

4) specify that the equipment used to cope with a station blackout needs only to be commercial grade and not safety related or environmentally qualified to 10CFR50.49;

5) allow operator action outside the control room; and

6) acknowledge the availability of AC power from the station batteries through the inverters.

If you have questions regarding this matter, pl ease feel free to contact

• this office.

Sincerely, L. T. Gucwa MAL/blm c: Georgia Power Company Nuclear Re~ulatory Co11111ission Mr. J. H. Miller, Jr. Dr. J. N.race, Regional Administrator Mr. J. P. 0 1 Rei 11 y Senior Resident Inspector Mr. J. T. Beckham, Jr. Mr. G. Bockhold, Jr. Mr. H. C. Nix, Jr. GO-NORMS 4297A

ROCHESTER GAS AND ELECTRIC N. Y. 14649*0001 ROGER W . KOBER VICE PRESIDEN T ELECTRIC & STEAM PRODUCTION June 18, The Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Attention: Docketing and Service Branch

Dear Sir:

This letter is in response to the Proposed Rulemaking package concerning "Station Blackout - Unresolved Safety Issue A-44," published in the Federal Register on March 21, 1986. As a member of the Nuclear Utility Group on Station Blackout, Rochester Gas and Electric has actively pursued resolution of this issue in a timely, cost effective manner, with due consideration for its perceived risk. To this end, we endorse the comments filed via the letter of June 17, 1986 by the Nuclear Utility Management and Resources Committee (NUMARC) on this issue. In addition to agreement with those comments, RG&E wishes to document our perspectives on the potential risk of Station Blackout, specifically as related to Ginna Station.

1. We consider the likelihood of a Station Blackout event at Ginna Station to be very low, in part because of the high reliability of the emergency diesel generators.

The diesel generator reliability figure is 0.990. This reliability figure has been calculated using RG&E 1 s

• response to Generic Letter 84-15 excluding those reported events which would not have impaired diesel generator operability and safe shutdown capability. A recent study by EPRI, the "EPRI Data Base on Diesel Generator Reliability," confirms this figure.

1983-1985 time period, the reliability calculated by For the EPRI is 0.991.

2. There appears to be an overemphasis in the Rulemaking package, and the related NUREG-1109, on severe weather being a cause of Station Blackout. Although severe weather certainly increases the probability of a loss of offsite power, it should have only a slight effect on the risk from Station Blackout. The emergency power systems at Ginna were thoroughly reviewed for operability in the instances of severe and extreme natural phenomena such as floods, tornadoes, and snowstorms as part of the Systematic Evaluation Program.

The Ginna Station design basis thus already includes the system design features and procedures to ensure that no unacceptable loss of emergency onsite power will occur during severe weather events. -,/"~-n_nP Acknowledged by card .* . ~- ~t- ~ "1 1

lON I u s. Nl I Du t Pco;tr l.O I

                    <ti{~q ~(p A lei I

        --~--

ROCRESTE R GAS AND ELECTRIC CORP. SHEET NO . 2 DATE June 18, 1986 TO The Secretary of the Commission

3. Additional safety features independent of the emergency ac power distribution system are available at Ginna Station. In addition to the 200% capacity turbine-driven AFW system, Ginna Station has a) a TSC diesel generator, which can be used to power a charging pump to provide reactor coolant pump seal injection, b) a diesel driven fire pump, taking suction from Lake Ontario, which can provide an inexhaustable source of secondary cooling water to the steam generators, and c) a TSC battery system double the capacity of each station emergency battery, and which can be cross-connected to the station batteries to supply vital loads for much longer than 8 hours *

• Thus, RG&E believes that the Station Blackout Rulemaking Package should be withdrawn, and the NUMARC proposals, which emphasize the most cost-effective methods for reducing risk from Station Blackout, should be considered by the Commission
• l

                                                              '-i:" \  lt HE ELECT RIC COMPAN SAS GAS ANO ELECTRI C COMPANY DOCKETED NU 81,1\  0
                                                                      ~I' ...-

• 5 ) USNRC
• June 20, 1986 G L E N N L. K O E S T E R VIC E PR ES ID E NT
• N UCLEAR RUlf. ct[89

{jl Ff?- I Secretary of the Coomission U.S. Nuclear Regulatory Coomission Washington, o. c. 20555 KMLNRC 86-114 Re: Docket No. STN 50-482

Dear Sirs:

On Subj : Ccmnents on the Proposed March 21, 1986, a proposed rule on station blackout was published for carment (51FR2829). The proposed rule is in response to Unresolved Safety Issue A-44, "Station Blackout". The proposed rule would define the term "station blackout" and require comnercial nuclear power plants to withstand and recover from a station blackout of a specified duration acceptable to the Staff. The proposed rule would also include an amendment to General Design Criteria 17. Accanpanying the proposed rule is the "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout" (NUREG-1109) that sunmarizes the value of the proposed modifications and the anticipated cost for compliance. Kansas Gas and Electric Canpany (KG&E) concurs with the NUMARC position that rulemaking is not warranted for the reasons outlined in their letter dated June 17, 1986 from J. H. Miller, Jr., NUMARC, to Chairman Palladino, NRC. KG&E also endorses the NUMARC cooments on the draft regulatory guide on station blackout and cooments provided in letter dated June 19, 1986 by KM:, Inc. on behalf of the Utility Safety Classification Group. If you have any questions concerning this matter please contact me or Mr. o. L. Maynard of my staff. Very truly yours,

                                                                                          ~~~ Glenn L. Koester Vice President - Nuclear GLK:see cc: ro *connor (2)

JCunmins 201 N. Market - Wichita, Kansas - Mail Address: P.O. Box 208 I Wichita, Kansas 67201 - Telephone: Area Code (316) 261 -6451

DOCKETED USNRC Florida

                                                                              .,,..    ~-**

11P Power CORPORATION June 19, 1986 3F0686-09 Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, DC 20555 Attention: Docketing and Service Branch

Subject:

Crystal River Unit #3 Docket No. 50-302 Operating License No. DPR-72 Comments: NRC Station Blackout Proposed Rule

Dear Sir:

Florida Power Corporation (FPC) has reviewed the proposed rule, its technical basis, and other support information. We have been actively participating in resolution of the Station Blackout issue, indpendently and through the Nuclear Ut i lity Group on Station Blackout (NUGSBO). Our proactive effort has been directed at Crystal River Unit #3 (CR-3) fact finding and subsequent reasonable resolution of the issue. A preliminary coping analysis has been conducted and a comprehensive coping study was scoped for cost estimating purposes. The comprehensive study would be required to assess compliance with the proposed rule. The CR-3 PRA model has been significantly enhanced and carefully analyzed

• At this time, FPC extends its total support to the comments forwarded to the NRC by NUMARC. This extensive review and comment effort utilized significant resources of NUGSBO as techni ca 1 consultant. In summary, the issue is not generic and should not be resolved by generic rulemaking. The technical record does not support the proposed rule, and requirements to justify a backfit do not appear to be satisfied.

Many points of the industry* s comments can be reinforced and supported by FPC, with Crystal River Unit #3 specific information. However, in the interest of brevity and to avoid redundancy, we forward our comments, correlating with the industry's comments, in two areas only.

1. The Draft Regulatory Guide on SBO, Task SI 501-4, Division I, March 1986, Section C.3.1.5. indicates "Consideration should be given to using available non-safety-related equipment, as well as safety-related equipment, to cope with a station blackout provided such C.21 equipment meets the recommendations of Item 4 in Section GENERAL OFFICE 3201 WI :dged L_ *~"" '°~o.f Thirty-fourth Street South* P.O. Box 14042, St. Petersburg, FIC>rida 33733

• 813-866-5151

June 19, 1986 3F0686-09 Page 2 this Regulatory Guide." Item 4 of Section C.3.1 requires such equipment to meet design and performance standards that ensure adequate reliability and operability in extreme environments that may be associated with a station blackout, including hazards due to severe weather. Notwithstanding our opposition to the proposed rule, together with the draft Regulatory Guide, this is an unreasonable requirement; it eliminates the Li censee 1 s ability to take credit for existing equipment that could be used to demonstrate coping capability for a very unlikely event (extended SBO) simply because the existing equipment may not be qualified to the undefined "extremes" of environmental conditions resulting from an SBO. Guidance should provide for utilization of existing equipment for a period of time during which extreme conditions are developing. Additionally, in plant areas where, during an SBO event, the environment remains within the limits of the equipment design, no environmental qualification testing requirements should apply.

2. The preliminary, high level SBO coping study conducted by FPC identified a number of areas requiring further investigation and review to determine if any coping problems exist. The scope of a comprehensive study has been defined ( the areas requiring further investigation constitute a significant portion of this scope), and an estimate of the cost of the coping study has been determined to be

       $1,175,000. This estimate significantly exceeds the NRC estimate of
       $150,000 for a coping study.          The FPC estimated cost for a comprehensive coping study, of course, does not provide for any hardware or software corrections or enhancements. This supports the industry comments to the point that the NRC has understated the cost burden on the utilities, for this portion, by nearly an order of magnitude. FPC does not intend to proceed with the comprehensive study at this time.

From a somewhat separate perspective, the attached "Crystal River Unit 3 Risk-Based Perspective of the Proposed Station Blackout Rule" stands on its own merit asserting that the Station Blackout issue should not be resolved by generic rulemaking. The CR-3 Station Blackout core-melt frequency satisfies the goal of 1.0E-5/y, the basis for acceptable risk in the proposed rule. In addition, a very important point is drawn from the comparison of the cost estimate for a comprehensive coping study ($1,000,000+) and the cost that could reasonably be justified to increase coping capability from 2 hours (if our capability was as little as 2 hours) to 8 hours ($249,000 - see page 3 of attachment). In the unlikely event that the cost of our coping study could be reduced to the NRC estimate of

 $150,000 by scope reductions, owners group cost sharing and other potential

June 19, 1986 3F0686-09 Page 3 streamlining, less than $100,000 would be available to effect any enhancement of coping time. Furthermore, if the conservatisms were reduced or eliminated (i.e., battery life of 3 or 4 hours rather than 2 hours, EOG reliability upgraded by data through end of 1985, a more reasonable release category selected rather than SST-1), the cost that could reasonably be justified to increase coping capability becomes significantly less than the

 $249,000 above.

Expenditures to resolve the Station Blackout issue can be better spent, securing quicker and more appropriate results, by implementing the AC Reliability Program proposed by the industry comment package and described briefly by NUMARC's transmittal of the industry's comments on the proposed rule

• Sincerely,

 ~~

Manager, Nuclear Operations Licensing and Fuel Management DEP/feb Attachment

June 18, 1986 3F0686-09 CRYSTAL RIVER UNIT 3 RISK-BASED PERSPECTIVE OF THE PROPOSED STATION BLACKOUT RULE The proposed station blackout (SBO) rule is based on two key assumptions: (1) core-melt accidents due to SBO events are a dominant contribution to overall plant risk, and (2) a significant reduction in risk can be achieved by improving a plant's capability to cope with an SBO event. Results of a probabilistic risk assessment (PRA) for Crystal River Unit 3 (CR-3) dispute both of these assumptions. The following sections discuss the development of SBO core-melt accident sequences and reliability data in the CR-3 PRA. Core-melt frequency results for the current CR-3 configuration are presented along with estimates of the decrease in core-melt frequency due to SBO coping capability improvement. Station Blackout Core-Melt Sequence Development SBO core-melt sequences for CR-3 are unusual si nee the time when core melt commences depends on which diesel generator (DG) fails last. (This distinction was not made in Appendix B of NUREG-10321.) Seven time lines, shown in Figure 1, have been developed to encompass all possible combinations of DG failures. These time lines are based on the following relations:

1. The motor-driven emergency feedwater pump ( MDEFWP) fails when the "A" DG fails.

2. The turbine-driven emergency feedwater pump (TDEFWP) fails two hours after the 11 8 11 DG fails. It is conservatively assumed that the train "B" batteries are exhausted in two hours after loss of the "B" DG.

Both control valves fed from the TDEFWP wi 11 fail open when the "B" batteries are exhausted, resulting in steam generator overfill. The steam generator overfill will continue until either (1) operators outside the control room manually close the block valves between the TDEFWP and the control valves, (2) operators outside the control room manually trip the TDEFWP, or (3) the overspeed device trips the TDEFWP when water fills the steam piping. (Note that the block valves cannot be shut and the TDEFWP cannot be tripped from the control room due to the loss of the "B" battery.) It is assumed that the operators wi 11 not attempt to use the TDEFWP after the "B" batteries fail (by manually restarting the TDEFWP and controlling its flow through manual positioning of the block valves) if the "A" DG is operating. If the "A" DG has failed prior to "B" battery exhaustion, then manual control of the TDEFWP is not possible since all steam generator level indication will be failed due to the total loss of de power.

3. All high pressure injection cooling (primary feed-and-bleed) fails when both DG's fail.

4. Core uncovery commences in 50 minutes following a total loss of decay heat removal2.

June 18, 1986 3F0686-09 TT Case 1: Both D/G's fail ~,_------+-0-+----*..._ T at T=O I 2:50 Case 2: 0/G A fails at T=O y and 0/G B fails at O T t t+2:50 T=t Case 3: D/G B fails at T=O and D/G A. fails at T=t' where t' < 2h r \I t' e 2:50 T 2

sz Case 4: 0/G B fails at T=O and D/G A fails at T=t' where t' > 2h T 2 e

t' t'+:50 T Case 5: 0/G A fails before sz~ 0 T D/G B (t'<t) r;. t t+2 t+2:50 , Case 6: D/G A fails after 0/G B such that

                                            ~      sz             e      T t      t'          t+2:50 t<t' <t+2h                                  t+2 Case 7:   0/G A fails after            :T:                sz     911 T 0/G B such that t' > t+2h                     t               I t'  t'+:50 t+2
    '] = A D/G fails        t'= time when A D/G fails     T = time after LOSP T  = B D/G   fails     t  = time when B D/G fails   O = core melt FIGURE 1. Time Lines Developed to Quantify Station Blackout Sequences

June 18, 1986 3F0686-09 Data Collection The frequency of loss of off-site power ( LOSP) events is based on the Electric Power Research Institute/Nuclear Safety Analysis Center (EPRI/NSAC) database3 for all years through 1985. Events in each EPRI/NSAC category have been reviewed for applicability to the CR-3 site. The criteria used in this review are summarized below:

1. All cold weather (snow or ice) events have been removed.

2. All total LOSP events which would not result in a total LOSP at CR-3 due to swi tchyard equipment and configuration differences have been removed.

3. All partial LOSP events which would have resulted in a total LOSP event at CR-3 have been added.

This review shows a total of 18 events in 649.7 site-years which are applicable to the CR-3 site. Two events concern hurricane-induced LOSP events; the exposure time of these events is 190.8 site-years since only coastal plant sites are subject to hurricanes. Hence, the LOSP frequency for CR-3 is 0.035/y. The time-dependent off-site power restoration probability is taken from Figure 3.2 of NUREG-10321. Use of this information does not imply that FPC agrees with the methods and conclusions of NUREG-1032 regarding the duration of LOSP events. This information provides a conservative upper bound for the true off-site power restoration probability at the CR-3 site, and is used for convenience. There are four significant failure modes of the DG's: (1) the DG 1 s can fail to start due to malfunction of the engine, generator, or associated support systems (e.g., fuel oil), (2) the DG's can fail to start due to maintenance activities in progress at the occurrence of a LOSP, (3) the DG's can fail to load due to battery unavailability, and (4) the DG's can fail to run until off-site power has been restored. The likelihood of each failure mode is discussed below:

1. The probability that a DG fails to start on demand due to engine, generator, or support system malfunction is based on plant experience (computerized maintenance history, NPRDS, and LER) for the period from September 18, 1978 until April 18, 1985. This data yields a failure probability of 0.022/d.

2. The probability that a DG is unavailable when needed due to maintenance is based on estimates by the plant staff. These estimates, based on a formal interview technique4, have been aggregated using a method which includes a statistical means for accounting for biases due to expert judgement5. The DG unavailability due to maintenance is 0.0056.

June 18, 1986 3F0686-09

3. The probability that a DG fails to load due to battery unavailability (de power is required to close the DG output circuit breaker) is based on a battery failure rate of 1.2E-6/h (obtained through analysis of plant experience as described in note 1 above) and an exposure time of 3 months (the interval between battery load tests). This data yields a battery unavailability of 0.0026.

4. The DG failure rate, based on plant experience as described in note 1 above, is 0.0057/h. A convolution of this data and the off-site power restoration information shows the probability that a DG fails to run for the duration of a LOSP event is 0.0073.

In summary, the unreliability of a DG at CR-3 is estimated as about 0.0375/d. Note that, unlike NUREG-10321, no credit has been taken for repair of a failed DG during a LOSP event. Core-Melt Frequency Due to Station Blackouts The results of the PRA show that the frequency of SBO core-melt accidents is about 1.16E-5/y (28.5% of the total level I PRA core-melt frequency of 4.07E-5/y). DG failure modes which prevent successful DG start comprise about 88% of the total SBO core-melt frequency. When one considers the uncertainties in data, the SBO core-melt frequency for CR-3 satisfies the goal of 1.0E-5/y, which is the basis for acceptable risk in the proposed SBO rul e6. A limited sensitivity study has been conducted to assess the influence of battery lifetime on core-melt frequency. Table I shows the effect of increasing battery 1 ifetime to four and eight hours, assuming that other effects of an SBO event (notably, loss of auxiliary building ventilation) are less limiting than battery lifetime. The greatest decrease in total core-melt frequency (due to increasing the battery lifetime to eight hours) is 4.30E-6/y; this value is about an order of magnitude below the NRC Staff 1 s estimate? of core-melt frequency reduction following implementation of the proposed SBO rule (3E-5/y). Assuming (1) the off-site consequences of an SBO core-melt accident are about two million person-rems per event* and (2) the remaining life of CR-3 is 29 years (based on an operating license extension to 2016}, this reduction in total core-melt frequency equates to a risk reduction of about 249 person-rems. Such a reduction in risk would be justified if the cost of plant modifications is less than

 $249,000 (based on $1,000 per person-rem averted8).

In summary, alt hough the cont ri but ion of SBO events to over a 11 core- me 1t risk is significant, CR-3 currently satisfies the goal for acceptable risk. In addition, increasing the ability to cope with SBO does not lower the overall core-melt risk for CR-3 as significantly as the NRC has projected. For CR-3, the most significant benefit would be derived from improving AC reliability.

• This value is the NRC staff's estimate of the consequences, based on rel ease category SSTl for an average pl ant7; the consequences for an accident at CR-3 would be significantly lower since there are no large population areas within 50 miles of the site, nearly half of the area is in the Gulf of Mexico, and SBO core-melt accidents should be expected to result in releases less severe than those associated with SSTl.

 .,                                                           June 18, 1986 3F0686-09 TABLE I SENSITIVITY OF CORE-MELT FREQUENCY TO BATTERY LIFETIME Battery Lifetime 2              4              8 SBO core-melt frequency (/y)          1.16E-5         9.37E-6       7.25E-6 Total core-melt frequency (/y)        4.07E-5         3.85E-5       3.64E-5 Percentage of total core-melt frequency due to SBO events              28.5           24.3          19.9 Percent reduction in SBO

• core-melt frequency Percent reduction in total core-melt frequency 0

0 19.2 5.4 37.5 10.6 June 18, 1986 3F0686-09 References

1. United States Nuclear Regulatory Commission, Evaluation of Station Blackout Accidents at Nuclear Power Pl ants, NUREG-1032, draft report, May 1985.

2. Nuclear Safety Analysis Center and Duke Power Company, Oconee PRA: A Probabilistic Risk Assessment of Oconee Unit 3, NSAC-60, June 1984.

3. Nuclear Safety Analysis Center, Losses of Off-Site Power at U.S.

Nuclear Power Plants: All Years Through 1985, NSAC-103, May 1986.

4. Fragola, Joseph R. and Collins, Erin P., 11 Interview Process for Obtaining Component Unavailability Data from Plant Personnel Experience, 11 SAIC/NY Draft Report No. 85-9-1, September 3, 1985.

5. Martz, H. F. and Bryson, M. C., 11 A Statistical Model for Combining Biased Expert Opinions, 11 IEEE Transactions on Reliability, Vol. R-33, No. 3, August 1984, pp. 227-232.

6. Federal Register, Vol. 51, No. 55, March 21, 1986, p. 9830.

7. United States Nuclear Regulatory Commission, Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout, NUREG-1109, draft report, January 1986.

8. 10 CFR 50, Appendix I, section II.D

• 0ut,; fT NU et.:11 DUKE POWER COMPANY P.O. BOX 33189 iiafo o Bu P - So (st Ff.

HAL B. TUGKER CHARLOTTE, N.C. 28242 DOCK[ O USN~PHONE C/f..2!,) ( VIGE PRESIDENT (704) 373-4331 NUCLEAR PRODUCTION June 18, 1986 The Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, D.C. 20555 ATIENTION: Docketing and Service Branch

Dear Sir:

.Duke Power Company submits the following comments on the Nuclear Regulatory Commission Proposed Rule on Station Blackout. This item was published for public comment on March 21, 1986 in Federal Register Volume 51, Number 55. The nuclear utility industry has closely followed the resolution of the station blackout issue. In this regard two industry groups were formed to address station blackout: The Nuclear Utility Group on Station Blackout (NUGSBO) and the Nuclear Utility Management and Resources Committee (NUMARC) Working Group on Station Blackout. Duke Power has been an active participant in both these groups and hereby provides its endorsement of the consolidated comments which NUMARC is filing on the proposed rule on station blackout(USI A-44). Duke Power considers the NUMARC comment document a detailed and thorough technical analysis of the rule. It also provides reasons why the station blackout issue is not generic, points out a lack of adequate backfit analysis, and provides an aggressive industry AC power reliability improvement program. Additionally, Duke would like to emphasize its

• agreement with the NUMARC comment which recommends that this issue not be resolved by rulemaking.

In the event the Commission decides to proceed with the rulemaking process on station blackout, Duke Power submits the following specific comment on the proposed rule. Our review of Part 50.63(b), Limitation of Scope, indicates this paragraph should be broadened. These provisions should also cover plants where safe shutdown facilities have been installed in accordance with earlier Appendix R requirements and/or other considerations. The proposed rule should fully recognize these installations and not require these licensees to provide further analyses, since the capability for withstanding extended station blackout already exists. Acilnowled by r

Secretary June 18, 1986 Page 2 Duke Power appreciates the opportunity to connnent on this issue. Very truly yours, H.B. Tucker .JSW/jgm

Commonwealth Edison One First National Plaza, Chicago, Illinois Address Reply to: Post Office Box 767 Chicago, Illinois 60690 June 17, 1986 Secretary U.S. Nuclear Regulatory Commission Docketing and Service Branch 1717 H Street N.W. Room 1121 Washington, DC. 20555

Subject:

Comments on Proposed Rule Regarding Station Blackout

Dear Mr. Chilk:

Commonwealth Edison has reviewed the draft documents regarding Unresolved Safety Issue USI A-44; Station Blackout. The documents reviewed included Draft Rule 10CFR50.63, Draft Reg. Guide and NUREG 1109. We agree and support the comments the Nuclear Utility Management and Resources Committee has filed on the proposed station blackout rule (USI A-44). In addition, we have the following supplemental comments. The factors identified as the main contributors to risk from station blackout (SB) are: a) redundancy of onsite emergency power sources (number of sources available minus the number needed for decay heat removal), b) reliability of onsite emergency AC (EAC) power sources c) frequency of loss of offsite power (LOOP), and d) probable time to restore offsite power. The frequency and duration of LOOP being related to grid and switchyard reliability and weather dependency. With respect to a) redundancy of onsite EACs, both LaSalle and Zion Stations fall into Group A (page 5 NUREG 1109) or 2 out of 5 (2/5) category. Dresden and Quad Cities Stations are 2/3 stations based on LOCA considerations. Based on the definition in Table 2 of NUREG 1109 we would be

• looking at " ... number of EAC power sources required to operate AC powered decay heat removal systems ... This number is based on all the AC loads required to remove decay heat to achieve and maintain hot shutdown at all units at the site with offsite power unavailable".

Since Dresden and Quad Cities Stations have the capability to crosstie 4KV buses, and due to the limited emergency AC power requirements of the two units, both units can achieve and sustain hot shutdown conditions with only one of the three available EDG's (Emergency Diesel Generators). ~hus all of our nuclear power plants fall in EAC power configuration Group A. This translates to a requirement that we show a plant coping capability of 4 hours for a loss of offsite power event. We believe performance of a rigorous, auditable coping analysis as recommended in NUREG 1109, would be substantially costlier than what is stipulated in the NUREG, without making our plants any safer. The item which

CO MISSION U $. NUCLE CH oo ... K y 0 r L. A'

    ~p  ti i

                                         -2 we feel could be the costiiest 'is the requirement to verify that equipment needed to operate during a station blackout will be able to operate under the environmental conditions associated with a total loss of AC power. Due to the vintage of a number of our plants not all monitoring equipment was environmentally qualified. The requirement to go back and document the acceptability of this equipment could mean its replacement due to unavailability of documentation.

We believe that any funds spent on this rule should go toward equipment modifications rather that paper exercises. Item iii. of page 3 of NUREG 1109 would defeat that intent and preference. Second main factor contributing to station blackout risk is reliability of onsite EACs. Based on the definition in NUREG 1109, p 5, EAC reliability is measured by counting EDG failures in the last 100 valid demands. This method does not truly reflect the site EDG reliability especially if one EDG is started and fails more then the remaining ones .

• Furthermore, credit is not given to a start that may be a delayed start that for LOCA purposes is considered a failure. A better defined, station blackout related accounting is the minimum required if this draft rule is approved.

The final main factors contributing to station blackout risk are related to weather and switchyard. Based on the EPRI published report NSAC/103 dated May, 1986, titled Losses of Offsite Power at US Power Plants all years through 1985, the following important statistics emerge:

1) Only four losses of all off-site power longer than 30 minutes occurred in the last three years.

2) In all years through 1985, only seven sites experienced losses of off-site power longer than two hours.

3) Only two substantial weather events occurred since 1980.

Of Commonwealth Edison's five plants only Dresden experienced a LOOP event longer than 2 hours and that was in 1965. (One event in 139.8 reactor years). Since then, additional transmission lines on separate right of ways, have improved substantially the grid tie-in. The above statistics appear to indicate that the risk of Station Blackout has been overstated, and having the entire industry perform expensive coping analyses does not seem to be justified. The Supplemental Information published in 51FR9829 and NUREG-1032 implies that the objective of the proposed rule is to reduce the probability of blackout-caused core melt to 10-5 per reactor year. The proposed rule requires that licensees demonstrate the capability to cope with a blackout of specified duration as the means of demonstrating compliance with the rule's objective. We suggest that the rule should permit other means of demonstrating compliance with this objective. Specifically, a licensee should be permitted to submit, in lieu of items required to be submitted under the proposed 50.63(c)(l), a plant-specific evaluation that directly demonstrates compliance with the rule's objective of limiting the probability of blackout-caused core melt to 10-5 per reactor year. Similarly, direct compliance with this objective should be permitted as an alternative to paragraph (e) of GDC 17.

The *Following are Deficiencies of NUREG 1032 Analysis of LOOP Events Due to Severe Weather:

1. Only six loss of offsite power (LOOP) events due to severe weather are in the data base. Only five of 52 sites experienced such an event. The NRC staff recognized that this experience was inadequate for use in a regression model correlating failure rate to design factors and/or weather hazards. In an attempt to improve the data base the staff gave consideration to "partial" LOOP events in which some but not all sources of offsite power were lost. However there is no reason to believe that such partial LOOP events are consistently reported by utilities, nor to believe that they are distributed in the same manner as complete LOOP events.

Furthermore, such partial LOOP events pose much less risk to the public than complete LOOP events. 2* Notwithstanding their inability to develop a regression model for

• weather-caused LOOP events by conventional techniques, the staff attempted to accomplish the same result by assuming proportionality of weather-caused LOOP events to the associated weather variable, weighted by years of exposure. Such estimates were biased upward by a) inclusion of "partial" LOOP events in the numerator and b) exclusion of experience at sites for which the weather variable was unavailable from the denominator. The staff claimed to use methods described by M.G. Lauby et. al. (1) to correlate outage data to weather hazards. However Lauby's work employed a Chi-squared test of the equality of two or more Poison distributions to assess the statistical significance of the differences observed. The staff claims that "the statistical validity could not be ascertained." As the calculations are elementary, we suggest this amounts to failure to use results not supporting the staff's preconceptions. Furthermore, the weather hazard data used for each site has also been omitted from NUREG 1032, precluding independent review of its accuracy, replication of the calculations, or computation of statistical significance .

3. While claiming to utilize Lauby's methodology to demonstrate a correlation between LOOP events and weather hazard, the NRC staff apparently disregarded Lauby's conclusions which are as follows.

a) At the 345kV and 230kV voltage levels studied, the frequency of weather caused outages is higher in summer than in winter. The difference is statistically highly significant for line related outages and not significant for terminal related outages. b) In the winter, only for 230kV line related outages is there a statistically significant relationship with snowfall. c) In the sununer, only for 230kV terminal related outages is there a statistically significant relationship with isokeraunic level. (1) M. G. Lauby et. al., "Effect of Pooling Weather Associated MAPP Bulk Transmission Data on Calculated Forced Outage Rates," IEEE Trans. (PAS), vol. PAS-103, August 1984, pp. 2345-2351.

Mr. Lauby indicates that the apparent susceptibility of 230kV lines to weather conditions is actually confined to a particular group of 230kV lines among those studied. These are of an old design which is being upgraded to reduce outage rates. In any event there is no statistical justification to extending these very limited results to other voltage levels or geographic regions.

4. Of the six valid LOOP events due to weather, only one, at Dresden, was caused by a tornado. At that time all transmission lines to the site were on a single right of way. The proposed requirement for coping capability gives no recognition of the importance of multiple rights of way in reducing the frequency of LOOP due to tornadoes.

on page A-28 of NUREG 1032 it is noted that "For sites in relatively high tornado frequency locations, the results may be more appropriately treated as a high, rather than a best, estimate." On page A-32 the tornado frequency coefficient was annotated "(for single right of way)," further emphasizing the importance of this factor in relating the frequency of tornadoes to LOOP events. As a minimum improvement, licensees should be permitted to use the LOOP frequency predicted by a more rigorous site specific tornado model in lieu of the term in the staff's equation dealing with tornadoes. This would reduce the frequency by a factor of ten in the case of Edison sites, using a model published by Teles, Landgren and Anderson.

5. The only LOOP event due to high winds other than tornadoes was due to a a hurricane. The staff attempted to correlate this event with the frequency of winds exceeding 75 miles per hour, predicted by fitting a statistical distribution to the annual peak wind speed observed at city and airport reporting locations. It seems implausible that such extrapolation gives valid estimates of hurricane frequencies as hurricanes seem to be a separate phenomenon from other occurrences of high winds .

• 6. NUREG/CR-2890, the principal source of wind speed predictions for midwestern locations, was based on "fastest mile" observations of wind. Such measurements imply that the faster the speed the shorter the time needed at that speed to be recorded. For instance, a 60 mile wind would be recorded only if sustained for 1 minute; a 90 mile wind, if sustained 40 seconds, etc. Transmission lines are designed to withstand wind loadings imposed by such brief gusts, but are more likely to be damaged by blowing debris in a period of sustained high winds such as a hurricane. As in the case of tornadoes, the number of lines and diversity of their routing should be considered when relating the frequency of hurricanes to the frequency of LOOP events.

Michael S. Turbak Operating Plant Licensing Director 1792K

P.O. Box 724928 Atlanta,Ga. 30339 P.O. Box 33189 Charlotte, N.C. 28242 NUCLEAR UTILITY MANAGEMENT AND RESOURCES COMMITTEE June 23, 1986 The Honorable Samuel J. Chilk Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, DC 20555

Dear Secretary Chilk:

• The June 17, 1986 NUMARC letter to Chairman Pal l adino , forwarding NUMARC comments on the proposed rulemaking for station blackout, contained a minor error in the cover letter.

Enclosed i s a corrected copy of the June 17, 1986 cover letter. Please replace the initial letter with the corrected copy and docket the corrected copy. The correction is noted by a sidebar in the right-hand margin. Thank you for your assistance. You rs truly,

• H. Mil er, Jr.

Chairman Steering Committee

• JHM: las Enclosure

l

         * *s n \

CORRECTED COPY P 0. Box 724928 Atlanta, Ga. 30339 P 0 . Box 33189 Charlotte, N.C. 28242 NUCLEAR UTILITY MANAGEME T AND HUMAN RESOURCES COMMITTEE June 17, 1986 The Honorable Nunzio J. Palladino Chairman U.S. Nuclear Regulatory Commission Washington. DC 20555

Dear Chairman Palladino:

In August 1985, NUMARC established a Technical Subcommittee to address some selected generic technical (hardware-related) issues and use the NUMARC process to bring those issues to closure. Station blackout was identified as the first technical issue to be addressed and the NUMARC Station Blackout Working Group was formed. The working group received both technical and administrative support from the Electric Power Research Institute (EPRI} and utilized the Nuclear Utility Group on Station Blackout (NUGSBO) as its technical consultant. The various aspects of the station blackout issue were evaluated and numerous interactions took place between the working group and the NRC staff regarding the station blackout issue. the proposed rule, and its regulatory analysis and associated backfit analysis. On March 21. 1986. the Nuclear Regulatory Commission (NRC) published a proposed rule on station blackout. The proposed rule, intended to resolve the Unresolved Safety Issue A-44, defines the term 11 station blackout 11 and requires commercial nuclear power plants to withstand and recover from a station blackout of a specified duration acceptable to the NRC. The proposed rule also includes an amendment to General Design Criterion 17. Accompanying the proposed rule is the "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout" (NUREG-1109) that summarizes the value of the proposed modifications and the anticipated cost for compliance. NUMARC's review of the station blackout issue and the proposed rule, with

• its regulatory and backfit analysis, indicates that many of the concerns related to station blackout can be alleviated through industry initiatives to reduce the individual sites' contribution to the overall risk of station blackout through the enhancement of the reliability of on-site and off-site AC power systems. We believe such initiatives are warranted since NUMARC agrees with the NRC staff that most of the anticipated benefits from the proposed rule would be derived from improvements at a limited number of plants.

On June 10, 1986, the NUMARC Executive Group overwhelmingly endorsed industry initiatives to address the more important contributors to station blackout. These initiatives consist of the following:

1. Each utility will review their site(s} against the criteria specified in NUREG-1109, and if the site(s) fall into the category of an eight-hour site after utilizing all power sources available, the utility will take actions to reduce the site(s) contribution to the overall risk of station blackout. Non-hardware changes will be made within one year. Hardware changes will be made within a reasonable time thereafter.

The Honorable Nunzio J. Palladino June 17, 1986 Page Two

2. Each utility will implement procedures at each of its site(s) for:

a. coping with a station blackout event,

b. restoration of AC power following a station blackout event, and

c. preparing the plant for severe weather conditions, such as hurri-canes and tornados to reduce the likelihood and consequences of a loss of off-site power and to reduce the overall risk of a station blackout event.

3. Each utility will, if applicable, reduce or eliminate cold fast-starts of emergency diesel generators for testing through changes to techni-cal specifications or other appropriate means.

4. Each utility will monitor emergency AC power unavailability utilizing data utilities provide to INPO on a regular basis

• NUMARC has endorsed the above industry initiatives in the belief that the station blackout issue is not generic and thus need not be resolved by rulemaking; rather station blackout is a concern at a limited number of plants and should be resolved in a manner other than rulemaking. We believe that the proposed rule would cause the unnecessary expenditure of resources at a majority of the plants in the United States even though they have been identified by the NRC as having an acceptable station blackout risk. In addition, the proposed rule focuses on coping with a station blackout event, whereas, we believe that improvement in AC power reliability is a more appropriate means to reduce the risk of station blackout. Detailed technical comments on the proposed rule, regulatory analysis, and backfit analysis are enclosed.

We are ready to work with you and your staff to bring the issue of station blackout to a timely closure *

• Yours very truly, Chairman

, Jr.

Steering Committee 0 JHM:las Enclosure cc/w: Commissioner Asselstine Commissioner Bernthal Commissioner Roberts Commissioner Zech Victor Stello, Jr. Secretary of the Commission

                    .ARGE                    u FOUND ED 18QI DOCKETED 55 EAST  MONROE    S T REET USNRC CHICAGO, ILLINOIS eoe03
                   ~ c 31 2> 2eg-2000
                   ~ T W X QI0 - 221-2807                      2:51 June 17, The Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Attention:  Docketing and Service Branch

• RE: Comments on Proposed Rule on Station Blaokout (51FR9829; March 21, 1984) and Draft Regul tory Guide and Value/Impact Statement on Station Blackout (dated March 1986)

Dear Sir:

Thank you for the opportunity to comment on the subject proposed rule and draft regulatory guide. Sargent & Lundy's specific comments are provided in the enclosure. Yours very truly, yJµ~~

• DWK:tss Danie l w. Kane Assistant Head, Nuclear Safeguards Licensing Division Enclosure

COMMENTS ON NUCLEAR REGULATORY COMMISSION PROPOSED 10CFR, PART 50 RULE ON STATION BLACKOUT REF. FEDERAL REGISTER , VOL. 51, NO. 55 DATED MARCH 21, 1986 A. SECTION TITLED "SUPPLEMENTARY INFORMATION" The proposed rule states " *** that the expected frequency of core damage resulting ;rom station blackout events could be maintained near or below 10- per reactor-year for any nuclear plant ** ". S&L COMMENT Since a frequency of occurrence of 10 -S is the apparent acceptance criteria for SBO analysis, the rule should provide a basis and justification for using this value. B. SECTION TITLED " ADDITIONAL COMMENTS BY THE COMMISSION" The commission states that the rule does not require a single failure to be assumed concurrent with a station blackout because SBO already exceeds this criteria by assuming failure of four power supplies, (i.e., two offsite and two onsite). It is thus concluded by the commission that based on the probabilities of losing the four power supplies, any modification work resulting from SBO should be upgraded to safety grade. S&L COMMENT We recommend reconsideration of the conclusion reached by the commission regarding upgrading of equipment to safety grade for the following two reasons:

• 1. The SBO evaluation performed by the NRC and summarized in NUREG-1032 assumes far more than the loss of four power supplies. Equation No . 1 to Appendix C of NUREG-1032 derives the probability of core damage due to a station blackout.

The first set of terms in Equation No. 1 considers con-current with an SBO, (i.e., loss of the four ac supplies) failure of all the ac independent decay heat removal systems or a small LOCA resulting from a stuck open relief valve. Thus , the SBO evaluation goes far beyond the loss of four power supplies.

2. Components designated as safety grade are required to respond to an accident condition. SBO is not an accident, but an event.

2. C. SECTION TITLED "COMMENTS BY COMMISSIONERS ROBERTS AND ZECH" Commissioners Roberts and Zech have requested additional comments regarding the impact of SBO on overall risk . S&L RESPONSE This type of assessment is difficult to address due to the lack of detail in NUREG-1032 regarding risk assessments and the apparent discrepancy in appropriate source term values. Specific details on risk and resolution of the source term need to be presented before the comments of commissioners Roberts a n d Zech can be responded to accurately. The NRC evaluation should include a quantitative %nd qualitative discussion on the risk impact of meeting a 10- core melt frequency.

COMMENTS ON DRAFT REGULATORY GUIDE AND VALUE/ IMPACT STATEMENT ON STATION BLACKOUT DATED MARCH 1986 "REGULATORY POSITION Cl. 2 - MAXIMUM FAILURE RATE" S&L COMMENT A definition of failure on demand for emergency diesel generators needs to be provided. Under the context of a station blackout, a diesel generator which fails to start automatically upon detection of an offsite power loss, but is successfully started manually from the main control room or from the local control pane~ should not be considered a failure on demand.

 "REGULATORY POSITION C3 .1 - EVALUATION OF PLANT-SPECIFIC STATION BLACKOUT S&L COMMENT The requirements specified for equipment performance during an SBO without HVAC should apply only to those pieces of equipment determined necessary for coping with an SBO. If the expected temperature excursions are within the temperature ranges specified as the design range for the equipment, no further evaluations are to be required.

It should be clarified that non-safety related equipment required for a station blackout need not be qualified or upgraded to safety-related status. The requirement that all equipment necessary to cope with an SBO during the first eight hours be located on site is excessive. Provided it can be demonstrated that equipment can be made available within the time period deemed necessary, the equipment need not be located onsite. This would allow sharing of portable generators, for example *

• "REGULATORY GUIDE POSI TION 3.2 - MINIMUM ACCEPTABLE STATION BLACKOUT DURATION CAPABI LITY" S&L COMMENTS "Table 1 - Acceptable Station Bl ackout Duration Capability" Table 1 outlines acceptable SBO duration capabilities. Variations from the tabulated values are allowed if justification, including cost-benefit analysis, is provided by the licensee. The tabulated durations are derived based on onsite power system reliability, offsite power system reliability and reactor dynamics. The overall criteria is to meet a core damage frequency of 10-5. In order to adequately assess ac power system reliability, one must account for reactor dynamics. Since the primary focus of SBO is on the ac power system reliability, a target level of reliability for ac power systems should be spec i fied. That is, a frequency of station blackout. This

2. would allow for a more concentrated evaluation of ac power systems, which would result in utilities being able to identify areas within the system for improvement which may preclude modifications for coping and allow a target to judge and monitor ac power system performance. Thus, in addition to this table, the NRC should specify what levels of ac power system reliability should be met in order to satisfy the NRC's criteria for core damage frequency .

          "Table 2 - Emergency ac Power Configuration Groups" The following should be added to the end of footnote "b" " *** with offsite power unavailable and no accident".
          "Table 3 - Offsite Power Design Characteristic Groups" Transmission lines are designed to account for expected regional weather conditions. These include snowfall, icing, and wind speeds.

The equation for estimated frequency of loss of offsite power due to severe weather (footnote b) should be based on weather conditions which exceed the design levels

• BALT IMORE GAS AND ELECTRIC DOCKETED USNRC 2.fi.b 3>

CHARLES CENTER* P. 0 . BOX 1475

• BALTIMORE, l@ f i 1*1 j§j}. :54 JO S EPH A . Tl ER NAN VICE PRES I DEN T NUCLEAR ENERGY uOCKEI NUM8£R P.R0£0$f0 BUJ.E PR- s (51 H:.<ttn)

The Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, DC 20555 ATTENTION: Docketing and Service Branch

SUBJECT:

Comments on NRC Proposed Rule on Station Blackout

REFERENCES:

(A) Federal Register, Vol. 51, No. 55, March 21, 1986, "Proposed Station Blackout Rule" (B) NUREG-1109, "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout" (C) Draft Regulatory Guide on Station Blackout, dated March 1986 (Task SI 501-4) (D) NUREG-1032, "Evaluation of Station Blackout Accidents at Nuclear Power Plants, Technical Findings Related to Unresolved Safety Issue A-44," Draft, May 1985 (E) NRC letter from E. J. Butcher, to A. E. Lundvall, Jr. (BG&E),

                                "NUREG-0737 Item 11.K.J.25, Reactor Coolant Pump Seal Integrity Following Loss-of-Off site Power," October 7, 1985 Gentlemen:

The Baltimore Gas and Electric Company is in agreement with and supports the comments which the Nuclear Utility Management and Resources Committee (NUMARC} is filing on the proposed rule on station blackout. We would also like to provide additional comments concerning features and capabilities unique to Calvert Cliffs which have not been considered by the NRC Staff in the proposed rule (Reference A) and the associated regulatory analysis (Reference B). Recognition of these factors reduces the risk of station blackout significantly below that calculated by the staff and further disputes the need for a rule.

June 19, 1986 The staff's objective through this rulemaking appears to be to reduce the risk of station blackout at a relatively small number of plant sites. These plant sites are characterized by either low redundancy in the emergency AC power supply or low reliability of the emergency diesel generators. The staff is also concerned with sites which appear to be susceptible to a loss of normal offsite power due to severe weather. Our review of the draft regulatory guide issued with the proposed rule (Reference C) indicates that Calvert Cliffs appears to be one of the plants with which the staff is especially concerned. This is mainly because the Calvert Cliffs diesel generator configuration falls into the regulatory guide's Category "C"; that is, a two-out-of-three configuration. The following information is provided for consideration by the staff and demonstrates that the risk of a station blackout involving core damage at Calvert Cliffs is extremely remote. In fact, the level of protection against station blackout-related core damage at Calvert Cliffs is among the highest in the industry. Specifically:

1. We have added a third independent offsite power source at Calvert Cliffs consisting of a 69kv line from the Southern Maryland Electric Cooperative (SMECO) system. This power source is capable of handling

• all of the safe shutdown loads at the site (it has the load-carrying capacity of two diesel generators) and is electrically and geographically separated from the two normal 500kv offsite sources. In a recent safety evaluation report supporting license amendments for Units 1 and 2, the NRC approved the SMECO tie as a fully qualified GDC-17 power source.

2. The reliability of the Calvert Cliffs emergency diesel generators is outstanding, based on the results of an industry survey conducted by the EPRI Nuclear Safety Analysis Center in 1985. This survey evaluated the previous 100 consecutive diesel generator tests. All three Calvert Cliffs diesels were found to be essentially 100 percent reliable from the perspective of preventing a station blackout.

3. Even in the unlikely event of a loss of offsite power followed by a simultaneous failure of two of the three diesel generators at Calvert Cliffs (resulting in a station blackout at one of the two units), two full-capacity steam-driven auxiliary feedwater pumps would be available to maintain the "blacked-out" unit in a safe shutdown condition until either offsite or onsite power was restored. This redundancy in the AC-independent feedwater system exceeds that which is typically required by the NRC {a typical auxiliary feedwater system for a PWR with a C-E or Westinghouse NSSS is one motor-driven and one steam-driven pump, or two motor-driven and one steam-driven pump.)

4. In the event of a station blackout at one of the two units at Calvert Cliffs, there are a number of important plant features and services which are shared between the two units. This commonality would allow the unit with an operable diesel generator to support the more important safe shutdown functions of the other unit. Among the most notable of the features is the recently installed cross-connect between the Unit 1 and Unit 2 motor-driven auxiliary feedwater pumps. In the event of a station blackout at one unit and a subsequent failure of both steam driven auxiliary feedwater pumps serving that unit, this cross-"

connect would provide a source of water from the unaffected unit's motor-driven pump.

June 19, 1986

5. The Calvert Cliffs DC electrical power system (including the batteries, the battery chargers, and the inverters) is common to Units 1 and 2. In the event that only a single diesel generator remained operable following a loss of offsite power (placing one unit in a station blackout condition), it would provide battery charging to a battery that serves both units. In this situation, neither unit would suffer a lack of battery capacity.

6. Calvert Cliffs Units 1 and 2 share a common control room. Even in the event that two of the three diesel generators failed following a loss-of-offsite-power, the third diesel would power the control room emergency ventilation system. Therefore, the control room would remain a mild environment from both an instrumentation and habitability standpoint during a station blackout at one of the two units.

From the above discussion, it can be seen that Calvert Cliffs is more accurately characterized by a one-out-of-three diesel generator configuration for the safety parameters most important to station blackout coping (i.e., decay heat removal, DC power supply and instrumentation survivability). The impact of these additional plant features and capabilities on the station blackout core damage risk calculated by the NRC is considerable. In the case of Calvert Cliffs, the NRC's regulatory analysis would tend to assign the plant a higher-than-average risk by virtue of the two-out-of-three diesel generator configuration. This is because the staff's regulatory analysis does not appear to take into consideration the degree of commonality in safety systems which is often attendant to those twin-unit sites utilizing such a configuration. Proper consideration of these Calvert Cliffs design features would cause the staff to place the plant in the risk category assigned to plants with a one-out-of-three diesel generator configuration, the category which is assigned the lowest risk of those evaluated by the NRC. With regard to Reference D, the issue of reactor coolant pump seal integrity during a station blackout is worthy of additional comment. The C-E Owners Group has long endorsed the position that reactor coolant pump seal cooling is not necessary to assure integrity of the RCP seals in an idle pump and, furthermore, complete loss of seal function will not occur due to station blackout in a C-E designed NSSS. The staff appears to acknowledge this position in Table 7 .1 on page 7-7 of Reference D where the core uncovery time due to a 100 gpm total leak rate from the RCP seals is not calculated nor apparently applicable to C-E PW Rs. In addition, on October 7, 1985, the NRC issued a Safety Evaluation Report to BG&:E (Reference E) which further supports our position that seal cooling is not necessary to assure seal integrity. The SER states," **. the staff no longer requires automatic reinitiation of coolant to the reactor coolant pump seals following a loss of offsite power event." Since seal cooling is not required to protect seal integrity following a loss-of-offsite-power event, it follows that seal cooling is not necessary to maintain seal integrity during a station blackout event. Despite this recognition of seal integrity under station blackout conditions, the staff continues to characterize the seal integrity issue as one of the largest contributors to core damage risk from station blackout. This is highly inconsistent with the staff's own findings. We also note with interest that the NRC's Interim Reliability Evaluation Report for Calvert Cliffs concludes that station blackout sequences only contribute three percent of the total core melt probability.

June 19, 1986 In summary , the current level of protection against station blackout-related core damage at Calvert Cliffs is adequate. We believe that these comments, when evaluated in concert with the detailed generic comments filed by NUMARC, demonstrate that the proposed rule is not warranted and should not be issued. As an alternative to the rule, we support the industry initiat ives being proposed by NUMARC as an effective means of maintaining the present level of safety regarding station blackout at Calvert Cliffs. Should you have additional questions regarding this matter, please do not hesitate to contact us. Very truly yours,

• JA T /BSM/dmk cc: D. A. Brune, Esquire J . E. Silberg, Esquire D. H. Jaffe, NRC T. Foley, NRC

The Light COIDpany Houston Lighting & Power P.O . Box llJQtKftfe9ston, Texas 77001

                                                                   ,,.. ~ C:-"

(713 ) 228-9211

                                                                                    -. 5 June 18, 1986 PR                                             ST-HL-AE-1690 00' tROPO  O RUl£    -

The Secretary of the Commission {st PR ffiJ..7) U.S. NUCLEAR REGULATORY COMMISSION Washington, D. c. 20555 Attention: Docketing and Service Gentlemen: DOCKETS 50-266 AND 50-301 COMMENTS TO PROPOSED RULE ON 10 CFR 50.63 LOSS OF ALL ALTERNATING CURRENT POWER On March 21, 1986, the Commission released for comment the proposed rule (10 CFR 50.63) on Unresol ved Safety Issue A-44, "Station Blackout", in the Federal Register. The purpose of this letter is to provide comments on this proposed ru l e as it appl i es to the Point Beach Nuclear Plant, a two-unit, two-loop Westinghouse pressurized water reactor plant started up in 1970 .

• Wisconsin Electric endorses the NUMARC submittal, "Comments on the Proposed NRC Rule on Station Blackout", dated June 19, 1986.

We feel that these comments provide a thorough review of the proposed rule and address those areas that are of concern to us. Of additional concern is the manner in which the emergency AC power configuration is determined on Table 2 of the draf t Regul atory Guide. Each of the two Point Beach Nuclear Plant emergency diesel generators is capable of suppl ying the required l oads of both uni ts of the station t o safely shut down, remove decay heat, and remove sufficient heat from the containment to maintain containment integrity. Assuming the conditi ons of a hypothesized station blackout, based on the capability of the emergency diesel generators, there is, therefore, no difference between a singl e-unit station with two emergency diese l generators and a two-unit station wi th two emergency diesel generators. I n either case sufficient capability is avai lable t o supply the r equired loads of the entire station, even i f one assumes the unavailabi l ity of one of the two emergency diesel by

The Secretary of the Commission June 19, 1986 Page 2 generators. As an additional comment, more credit should be allowed for the availability of on-site, non-emergency AC power sources, such as gas turbines, in determining the station blackout duration capability of a specific station. If you have any questions, please contact us *

• C. W. Fay Vice President Nuclear Power Copy to Resident Inspector

NUCLEAR UTILITY GROUP ON STATION BLACKOUT DO<:Knrn USNRC

                                                  *12 June 19, 1986 Secretary                                        uOC<<tl NUltliR PR."-

U.S. Nuclear Regulatory Commission eROPOSED ,21:E - ${) Washington, D. c. 20555 e_SI F~C/ CX1

Dear Mr. Secretary:

The Nuclear Utility Group on Station Blackout (NUGSBO) has been actively engaged in the station blackout issue since early 1984. NUGSBO has had numerous meetings with the NRC Staff and made various submittals to the Commission. Presently, NUGSBO serves as technical consultant to the Nuclear Utilities Management and Resources Committee (NUMARC) on this issue. NUGSBO has reviewed the Proposed Station Blackout Rule and Draft Regulatory Guide and supports the comments filed by NUMARC. NUGSBO also has reviewed NUREG/CR 3840 entitled "Cost Analysis for Potential Modifications to Enhance the Ability of a Nuclear Power Plant to Endure Station Blackout." This document provides the basis for the backfit cost estimates set forth i n the regulatory analysis supporting the proposed station blackout rule (see Table 6, NUREG 1109). NUGSBO anticipates submitting comments on this document in the near

• future.

Sincerely,

                                              ~u~

Executive Chairman Nuclear Utility Group on Station Blackout

            - n l

 **o Ps~G                       Public Service Electric and Gas Company 80 Park Plaza, Newark, NJ 07101 / 201 430-7000 June 19, Mr. Samuel J. Chilk Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, DC 20555                      -* -   * -* '

• Attention : Docketing and Service Branch

Dear Mr. Chilk:

STATION BLACKOUT 10CFR50 PROPOSED RULE 51FR9829 DRAFT REGULATORY GUIDE 51FR11494 We have reviewed the above listed documents along with appropriate existing material, and offer the following for your consideration. We have concluded that the present information and analysis does not provide an adequate technical basis for rulemaking. The risk analysis distribution indicates that station black-out is not generic, but is an issue to be reviewed by only a small number of plants. The source term assumptions used by the staff, although consistent with current NRC policy, should have reflected the IDCOR work which resulted in lower consequences. The presumption that core damage is synony-mous with near-term breech of containment for the station blackout accident sequence, is inconsistent with other tech-nical analyses. The staff treatment of severe weather is technically unsupported. Regarding diesel generator availability and capability, the ability to manually restart, after auto-start failure, should be considered in the analysis. In addition, station blackout loads are less than experienced during LOCA, and therefore can be handled by fewer diesels through proper load management procedures. With respect to the backfit analysis, the risk of station blackout is overstated and the cost to comply with the rule is understated leaving the analysis incomplete and the The Energy People 95-0942

Mr. Samuel J. Chilk 2 6/19/86 proposed rule not cost beneficial. As an example, the coping analysis requirement is not well defined and has the real potential to develop into a costly, detailed analysis. Furthermore, we agree with AIF that the backfit analysis is difficult to follow and, in the future, should be more self-contained and correlate to the required nine factors listed in 10CFR50.109C. We have been actively involved in the Nuclear Utility Group on Station Blackout (NUGSBO) efforts from its inception and have participated in the technical comments that will be . submitted by NUMARC by June 19, 1986. We endorse the NUMARC comments and initiatives and support the comments to be sub-

• mitted by AIF.

Very truly yours, R. L. Mittl General Manager - Regulatory and Environmental Support - 22A CWV:srd

• FA06 1/2

YANKEE ATOMIC ELECTRIC COMPANY Telephone (61 7) 8 72-8100 TWX 710-380-7619 19', 1986

                                                                     , 6-007 6-111 Secretary of the Commission U.S. Nuclear Regulatory Commission                             JOCIITT   Nllll8Elf ., @

Washington, DC 20555

                                                                       , .,,- ~ Pr?-

Attention: Docketing and Service Branch {!7 Fl(

Subject:

    "Comments pertaining to Proposed Rule (10CFR50) and Regulatory

Dear Sir:

Guide (Task Sl501-4) Regarding Station Blackout (FR Doc. 86-6284) and FR Doc. 86-7428) Yankee Atomic Electric Company (YAEC) appreciates this opportunity to comment on the proposed rule and regulatory guide on Station Blackout. YAEC owns and operates a nuclear power plant in Rowe, Massachusetts. Our Nuclear Services Division also provides engineering and licensing services for other nuclear power plants in the Northeast, including Vermont Yankee, Maine Yankee, and Seabrook. The Nuclear Utility Management and Resource Committee (NUMARC) is filing detailed comments resulting from their analysis of the subject proposal. We endorse these comments and feel strongly that the initiatives by NUMARC to improve AC power source reliability standards, much of which has already been implemented at Yankee facilities, are sufficient. Beyond that, a plant-specific analysis can be requested if the specific situation warrants it *

• Station blackout is very plant specific. It is not an issue that can be solved with a generic rule. NUREG-1109 highlights this plant-specific and non-uniform characteristic. The potential risk improvement and the associated costs under the generic analysis presented in NUREG-1109 are invalid because an average risk improvement at an average cost is assumed.

We believe that this issue is not appropriate to a rule change, and this rulemaking should be terminated. Very truly yours,

                                           ,12a/~

D. w. Edwards DWE:mtd Director of Industry Affairs

CP&L Carolina Power & Light Company JUN 19 JSe6 Mr. Samuel J. Ch ilk The Secretary of the Commission U.S. Nuclear Regulatory Comm ission JUN i* OOCKET NUMMR p* 'R'

• _ 50 Washington, DC 20555 IBOfOlf_D RULE _ _ _ .

Attention: Docketing and Service Branch {!I FR na-1J

Dear Mr. Ch ilk:

Carolina Power & Light Company (CP&L) wishes to express agreement with, and support for, the comments which the Nuclear Utility Management and Resources Committee (NU MARC) is filing on the proposed rule on Station Blackout (USI A-44)

• We would also like to add the following comments concerning features and capabilities unique to our H. B. Robinson and Brunswick Steam Electric Plants which have not been taken into consideration by the Staff in the rule and supporting regulatory analyses.

Recognition of these factors results in a reduction of the risk and further calls into questions the need for a rule. Our H. B. Robinson Plant has installed an Appendix R diesel and associated dedicated shutdown system which is fully capable of achieving and maintaining cold shutdown conditions independent of off-site power and emergency diesel generator power. This capability is in excess of the proposed rule requirement to achieve hot shutdown subsequent to a station blackout. However, it is not apparent that the Staff considered this capability in their station blackout risk assessment and determination of the need for a rule. Our Brunswick Steam Electric Plant has previously recognized the potential for loss of off-site power during severe storms such as hurricanes. As a result, we have a

• comprehensive weather procedure in place which calls for unit shutdown in advance of a hurricane as well as other precautionary measures. These actions serve to minimize the risk of a station blackout by placing the plant in a safe shutdown condition in advance of a possible blackout. This reduction in risk was not factored into the Staff's analysis, thereby resulting in an overstatement of the risks associated with a station blackout and the conclusion that rulemaking is necessary.

We wish to thank the Commission for the opportunity to comment on this issue. If our comments should engender further questions by the Commission or Staff, please do not hestitate to contact us. Yoursveri,

                                                                @~e:1~ent Nuclear Engineering & Licensing ABC/SDF /pgp      (3981 SDF) 411 Fayetteville Street

• P. 0 . Box 1551
• Raleigh , N. C . 27602 Acknowledged by cara

N 0

                                 @)              DOCKETEllTune 17, 1986 USNRC 1)0CID NUUBfR    PR - 57)
            <'IIOPJJSfD RU)-;- _ _       ~

Secretary of the c':I2Lf:R- q)? ~q U.S . Nuclear Regulatory Commission Washington, DC 2055 Attention: Mr. Alan Rubin Docketing and Service Branch

Dear Sir:

I have reviewed the proposed rule, regulatory guide,

• the referenced NUREG reports and the following is a li technical comments on the rule and regulatory guide. I am pro-viding these comments as a private individual since I feel that companies (architect/engineers, equipment suppliers, and utilities) tend to dilute any valid comments from the engineers that work in the nuclear energy field.

My technical background is electrical engineering and I have 15 years nuclear experience in design, construction, startup, and support of nuclear plants. My comments are divided into two parts. The first part is related to the proposed rule and the second part is directed at the draft regulatory guideline. These comments are on the attachments. I have listed various questions and request a response (either verbal or written). If you have any questions on my comments I can be contacted at (415) 768-2316 or 768-1180. iJl~~~ Kenneth M. Cooke PE19/8

COMMENTS ON PROPOSED RULE 50 . 69(a) The requirements section states it applies to 'light-water-cooled nuclear power plant'. The Fort St . Vrain nuclear plant is a HTGR and utilizes gas and air for cooling, therefore, does the new ruling apply to a HTGR type nuclear plant? Also , there are many research nuclear reactors operated by universities that are light water cooled plants. Does the new ruling apply to these small reactors in research facilities? 50.69(b)

• No comments on this paragraph.

50.63 Sect i on (c) . l This paragraph states that within 270 days of the date of the amendment the licensee shall submit the items in 50.63(c)(l) i, ii , iii, iv, and v which essential is a station blackout evalua-tion. However, in 50 . 63 Section c(2)d it states that once this information is submitted the commission will notify the licensee, and upon notification the licensee shall submit a schedule within 180 days of all modifications. It is suggested that the commis-sion request the licensee to submit a list of modifications and proposed schedule during the initial 270 day deadline. This will enable the commission to evaluate whether or not the utility is serious about complying with the 2-year completion schedule and also provide a basis for determining whether the proposed modifications support the station blackout evaluation . A large portion of the utilities will not offer any modifications if they are not requested, but wait for the Commission to notify them. Also, this partially removes the Commission from being in the critical path for the implementation schedule of the modifications . 50 . 63(c)i As usual, the rulemaking does not explicitly identify exact design basis criteria . As a result , each NRC inspector at the region or NRR engineer will interpret and judge whether or not the licensee has complied with the rule. This occurred on Appendix R and caused additional reviews and generic letters to be developed. I strongly recommend that the rule identify station blackout be a design basis event and list exact criteria. As a minimum, COMMENTS ON PROPOSED RULE the criteria should include all safety-related equipment used during the station blackout must be environmentally and seismically qualified (by test), specific exact time limit (8 hours minimum), identify whether or not single failure criteria applies, and state whether or not it is necessary to evaluate a station blackout event during various operating conditions of the plant. 50.63(c)ii This section requires a description of the procedures to cope with a station blackout. These is no guidelines or criteria identified as to exactly what "description of procedures" encompasses or what is needed. It is recommended that the commission request a brief overview of the procedure in conjunction with a detailed step-by-step itemized procedure to support the blackout analysis. This is necessary in order to surface key problem areas that will come as a result of the step-by-step procedure. For example, a brief procedure will not identify if sufficient plant operators are available to open doors, install portable HVAC, trip circuit breakers, etc. Also, the order of the procedure may be extremely important (a door to an inverter room may be necessary to open within 15 minutes time period to ensure the heat buildup in the inverter room will not damage the inverter). S0.63(c)iii This section should be more definitive in stating the identifica-tion of the factors and whether or not this identification should include an overall process system, individual electrical compon-ents or exceeding design parameters for containment temperature, equipment temperature, and/or battery capacity

• S0.63(c)iv It is unclear on how item Dis utilized in determining a station blackout duration.

50.63(c)v The comments on 50.63(c)iii are applicable to this section. S0.63(c)2 This section could be relocated into the 50.63(d) section. COMMENTS ON PROPOSED RULE 50.63(d)(l) and (2) The NRC agency must learn from the past issues like environmental qualification (EQ) and fire protection. The Brown's Ferry plant fire caused the existing Appendix R law to be developed and the Brown's Ferry plant, after eleven years, still does not comply with the Federal law. A large number of nuclear utilities did not initiate a serious effort to comply with the EQ and fire protection laws, but only waited until the NRR office threatens penalty before performing any actions. As a result, the NRC office received crictism about the nuclear plants not complying with their new requirements in a timely manner. I recommend that the NRC state clearly to the licensee's that

• schedule slippage is unacceptable, and 180 days for a submittal and 2 years for an implementation is mandatory. If the commission does not develop a hard line, the same group of utilities will not comply with your schedule and will not take the station blackout issue as a serious issue
• PE19/8 COMMENTS ON DRAFT REGULATORY GUIDE B. Discussion This section states that the staff assumes licensee's have adhered to standard RG 1.6, IEEE 387-1984, and RG 1.75 in order to minimize single-point vulnerabilities in design and operation. This is not a good assumption because the older plants have not complied to RG 1.75 or IEEE 387-1984. The majority of the plants were designed, constructed, and operating prior to IEEE 387-1984 or RG 1.75 were issued. (I.e., Trojan, SONGS, Rancho Seco, Arkansas 1.)

This section states that 4 hours or 8 hours is the minimum accept-able duration. However, the last paragraph states the licensees may propose durations different than the minimum. It appears to contradict the minimum requirement *

• What happens if a licensee submits a blackout analysis with a 3-1/2 hour availability? Does the rulemaking require a margin in the duration and if the answer is yes, how much? I recommend that a 30 minute period be a margin and that no time duration under four hours be accepted by the staff.

1.1 Reliability Program The reliability program is specified for the emergency AC power only. If the station blackout design event is relying upon battery capacity, then why is there no emphasis on the reliability of the DC system (battery, charger, inverter). The DC system is needed to start the emergency AC system and also to connect the AC power source to the electrical switchgear. Please read the study by Nucler Safety Analysis Center (NSAC) 48/1982 on DC systems and the failures that have occurred in the industry. A reliability program with numerical goals should be specified for the DC system

• Are the proposed staff actions in generic letter 84-15 mandatory as a result of the proposed rule?

Is the proposed reliability program a requirement similar to EQ and require full audit documentation file? Is a reliability program utilizing IEEE 577 and 352 standards acceptable? 1.2 Maximum Failure Ratio If the maximum failure rate is determined to be deficient over a certain period of time, then is this reportable and will it change the minimum duration from 4 to 8 as a result of maximum failure rate not being achieved? COMMENTS ON DRAFT REGULATORY GUIDE

2. Offsite Power The footnote states that lightning induced faults should be con-sidered as a potential cause and address in the procedures for restoring offsite power. The issue of a lightning strike on plants with full solid state modular circuitry utilizing floating ground system has never been completely reviewed for impact nor from a procedural viewpoint. The Information Notice 85-86 demonstrates that lightning strikes have caused a LOP and damaged equipment. The area of lightning strikes on plants having solid state logic modular systems should be addressed in the station blackout issue.

I doubt whether or not all the plants have improved their ground-ing systems to minimize lightning strikes (reference NSAC/41-1981

• study and NFPA 780 3.1 Evaluation of Plant Specific Station Blackout In 3.1.1, is it assumed that none of the DC systems are under test or maintenance at the time of the postulated station blackout?

In 3.1.4, is full environmental qualification (test or just analysis) data require for mild environment equipment? Is it necessary to requalify diesel generator and associated com-ponents or qualify control room equipment for extreme temperatures during loss of HVAC? In 3.1.5, is it acceptable to assume that a non class lE security system diesel generator is available but the emergency diesel generators are not available during the postulated station blackout? An onsite security system diesel generator could be utilized to recharge batteries over an extended period. In 3.1.6, what does 'consideration should be given to timely operator actions' actually mean? For example, on B&W type PWR reactors a postulated station blackout would cause loss of feed-water, no auxiliary feedwater flow (assuming no emergency AC power), and the steam generators would empty in 30 to 60 seconds, requiring the operator to manually trip local breakers to increase battery capacity appears to be unacceptable. Is it acceptable to provide a single switch or set of control switches which will trip the local appropriate breakers at the control room instead of sending operators to local stations? COMMENTS ON DRAFT REGULATORY GUIDE 3.2 Minimum Acceptable Station Blackout Duration Capability The methods to obtain minimum station blackout duration from the Tables 1, 2, and 3 is confusing and appears by the footnotes on Table 1 that the minimum duration is negotiable by showing cost-benefit analysis. It is unclear on how the probable time to restore offsite power is being applied in the Tables 1, 2, and 3. 3.3 Modifications to Cope with Blackout This section appears to be oriented towards PWR rather than all types of reactors. Also, I believe there should be some technical guidance on the method for evaluating battery capacity. For example, can I assume that all the capacity can be utilized and ignore hydrogen build up. I believe utilities will assume techni-cal items to their advantage in order to retain existing battery sizes. 3.3.1 If an additional battery charging system is necessary to achieve the appropriate blackout duration, why does an onsite power source have to be dedicated solely for station blackout? This eliminates flexibility that a licensee might have available. For example, many plants have an onsite gas turbine, adjacent solar power facility, or security diesel generator that could be manually tied to the class lE battery charging systems. Please provide more direction and definition on onsite diesel generator (does this diesel generator require reliability goals; what about technical specification for this new DG; what about the interface between existing charging system and new DG, etc.). 3.3.2 This section appears to be related strictly to PWRs and not BWRs. Provide a section on BWR's condensate storage tank and switch over to the water inventory in the Torus

• 3.3.3a&b Is the automatic capability within 10 minutes necesary?

Where does 10 minutes come from and what is the basis? Do these electrical circuits that are for the station blackout system have to comply with Appendix R, single failure criteria, environ-mental and seismic qualification, and GDC 17 (separation between onsite and offsite sources)? The staff should take a definite position on RCP seal cooling and demand an independent power source to protect these seals. I am sure every PWR licensee will show that the loss of the seals and the resulting water inventory will not create a problem for cooling down the reactor system. However, the utilities position will be motivated because of cost rather than the overall function of the RCP seals. PE19/8 Nuclear y ey 07054 umber Mr. Samuel J. Chilk Secretary of the Commission U.S. Regulatory Commission Washington, DC 20555

Dear Mr. Chilk:

Subject:

Request for Comments on Station Blackout Proposed Rule (10 CFR50)

• The staff of GPU Nuclear Corporation herewith submits comments on the subject proposed rule. Comments were requested in a March 21, 1986 Federal Register notice (51 FR 9829).

As a general comment, we wish to express our agreement with, and support for, the comments which the Nuclear Utility Management and Resources Committee (NUMARC) is filing on this proposed rule. As with NUMARC, a review of the proposed rule, and its supporting material, leads us to conclude that rulemaking is not warranted for the following reasons: (1) The issue is not generic and, thus, need not be resolved by a generic rulemaking. (2) The technical basis underlying the proposed rule is lacking. (3) The proposed rule itself is deficient .

• (4) The proposed rule does not meet the backfit rule standard.

Not withstanding the above, in the event the Commission feels compelled to promulgate a rule, GPU Nuclear wants to call to the Commission's attention the particular case of TMI-2. We believe that this rule is not applicable to TMI-2 and that TMI-2 should be exempted by specific language from any rule the Commission might issue. The present mode of core cooling at TMI-2 is "loss-to-ambient" which the TMI-2 Technical Specifications define as 11 a passive cooling mode by which decay heat, generated by the reactor core, is removed and transferred to the surrounding environment by air and passive components (i.e., reactor vessel) inside the Reactor Building. 11 Thus, even if TMI-2 were to suffer a station blackout, core cooling would continue to be maintained by loss-to-ambient cooling since no active components are required. Sincerely,

                                                      ~-~~

JRT:RJ:nk:3564f J. R. Thorpe ~ Director, Licensing & Reg. Affairs GPU Nuclear is a part of the General Pu~%Jitilities System

                                    .~n wledgea u     u. 7;.~"t:~ .... ~

1 1' :i~dc;

                    )    I.PP\/

I O IC. OJ 1Jf)o/ J

            ' /)

1J eW~SOd

                '~0
               >iJOO J-1::inN *s *n        f t
                                'J

Atomic Industrial Forum, Inc. 7101 Wisconsin Avenue Bethesda, MO 20814-4891 Telephone: (301) 654-9260 TWX 7108249602 ATOMIC FOR DC June 20, 1986 Secretary of the Commission U. S. Nuclear Regulatory Commission Washington, D. C. 20555 Attn: Docketing and Service Branch

Subject:

Federal Register Notices of Proposed Change to 10CFR Regarding Station Blackout (51FR9829) and Regulatory Guide Task No. SI 501-4 (51FR11494)

• The subject Notices and the documents to which they refer have been reviewed by several Subcommittees of the AIF Committee on Reactor Licensing and Safety. We appreciate the opportunity to review these proposals and provide below general comments. Addi-tional detailed comments are included in the enclosure to this letter.

Our review of the proposed requirements and their technical basis results in the conclusion that they are not justified. The bases for this conclusion include:

• The public health and safety benefits are significantly exaggerated by the analysis. Inappropriate use of siting source term SST-1 overestimates by at least an order of magnitude the potential radiation exposure to the public which would be prevented by the proposed requirements (see Enclosure, attachment 2);
• The costs of implementing the proposed requirements have been underestimated. In the particular case of the proposed coping studies, the estimated costs appear to reflect only the conduct of a battery capacity evaluation. It is not clear, however, that such an analysis would be sufficient to demonstrate coping, and special effects analyses which could be required could increase coping study costs significantly.

Other costs are underpredicted by factors of two to three (see Enclosure, attachment 3);

• Station Blackout has been found not to be a generic issue.

Station blackout risk is plant specific and, according to the staff's own analyses, the proposal requirements are expected to result in modifications at no more than a few facilities, if at any. Requiring all licensees to undertake extensive analyses under the provisions of the proposed rules when only a small group of plants may have a need for remedial action is not appropriate; l'cknowJedg by

-- O::) Od 1*rT

                '1
                   ')Q 1-1 n..i  *s * \

• The assertions of extensive station blackout coping capability at foreign (notably European) nuclear power plants are not sufficiently substantiated to serve as even part of the basis for the proposed requirements. The allusion to changes made at these plants to increase blackout coping is not substan-tiated in the record of this rulemaking. Based on information in the record, it is not clear that European plants are substantially more "blackout proof" than the current population of US Reactors (See Enclosure, attachment 4); and
• Other considerations presently under review by the NRC make it premature to impose the proposed requirements at this time. Specifically, the results of NRC and industry reevalua-tions of accident source terms would so significantly reduce the calculated benefit from the proposed changes as to make them unfeasible. Revised source term assumptions are expected to be ready for regulatory use in the near future, perhaps before the proposed requirements would be ready to promulgate in final form (see Enclosure, attachment 2).

We note that the Nuclear Utility Management and Resource Committee (NUMARC) has also performed an assessment of the subject proposal and intends separately to submit detailed comments. We endorse these comments. Further, we note that NUMARC intends to implement an initiative to improve the reliability of emergency AC power sources and to identify and address perceived station blackout vulnerabilities. This initiative encompasses those elements which can be implemented at all facilities to reduce any risk from

• station blackout and further obviates the need for the proposed requirements.

Additional detailed comments are provided in the enclosure. The proposed station blackout rules are the first technical rules relating to nuclear power plant design and operation to be proposed since the revised Backfit Rule, 10CFRS0.109, became effective. The Backfit Rule provides a framework for considering such proposals and establishes certain standards which must be met in NRC's supporting analyses. A "Backfit Analysis" was published concurrently with the subject proposal. We have reviewed the subject proposal in the context of the Backfit Rule.

e fi rmly believe that the Backfit Rule provides a long-needed framework for assessing the need for and the justification of new requirements. With this conviction, we were somewhat dismayed by the analysis accompanying these rules which, because of its structure, is somewhat inscrutable. The analysis is principally a loose fabric of references to other documents (e.g., NUREG-1109), which in turn often direct attention to further references to other documents (e.g., NUREG-1032). Thus it is often difficult to determine the precise basis for the conclusion presented. It is our opinion that future backfit analyses should be significantly more self-contained, and should avoid the use of non-specific

• references entirely. Further comments on the station blackout backfit analysis are contained in the enclosure.

Based on the above comments, and the detailed comments in the enclosure, we believe the proposed regulatory requirements to resolve USI A-44 have not been adequately justified under the Backfit Rule. We further believe that any generic rule would be inappropriate as a resolution of this issue. Therefore, the proposed requirements should not be imposed. We encourage NRC's consideration of these comments in developing its final position. Sincerely,

                                   ~           t1- ~

Murray Edelman, Chairman Committee on Reactor Licensing and Safety

• ME/jlc cc: Chairman Palladino Commissioner Roberts Commissioner Asselstine Commissioner Bernthal Commissioner Zech Victor Stello, Jr.

Rules and Procedures Branch, DRR, ADM, USNRC

ENCLOSURE Detailed Comments Regarding Proposed Requirements to Resolve USI A-44, Station Blackout Our review of the proposed regulatory requirements to resolve USI A-44, and the backfit analysis supporting them, results in a conclusion that the proposed requirements have not been adequately justified. We further conclude, based on the nature of this issue, that it is unlikely that any generic requirements can be justified. Our conclusions are based on:

• Deficiencies in the backfit analysis and its technical basis which, taken together, indicate that the required "substantial increase in overall protection" has not been demonstrated; Deficiencies in the estimates of benefits and costs associated with the proposed requirements; Evaluation of the technical basis for the proposed requirements which demonstrates that the issue is non-generic; and
• The unbounded nature of the proposed coping study requirements.

Comments related to each of these areas are enumerated below. Questions raised by the Commission and individual Commissioners are also addressed in the comments below. In addition, since the station blackout backfit analysis is the first such analysis related to rulemaking, we provide comments intended to improve its usefulness as a precedent for future such actions. The attachments to this enclosure provide more detailed information in support of specific comments as referenced below.

1. It has not been demonstrated that a "substantial increase in overall protection" would be realized by the proposed requirements:

a. The NRC Staff has proposed no standard by which to conclude that substantial additional protection will be realized. The backfit analysis concludes that radio-logical benefits would be realized at a value-impact ratio of 2000 man-rem per $1,000,000. The backfit analysis further concludes that imposition of the proposed requirements is justified. Although no

standard is explicitly stated, the form of the overall value-impact ratio calculated implies a comparison to the standard of $1,000 per person rem in the provisional safety goals policy statement. As discussed elsewhere in these comments, we conclude that the staff has overstated the radiological benefit and underestimated the costs. As a result, we conclude that the provisional safety goal standard has not been met, and that the requirements have not been demonstrated to be justified.

b. Commissioners Roberts and Zech, in their additional comments question whether station blackout is a small percentage of the overall risk, or perhaps a major component of an already small risk. As noted above, the staff's analysis shows that the risk is highly non-uniform. At most plants, it is our conclusion that the overall risk is low and that station blackout is a small contributor. At some plants, station blackout may be a more dominant sequence although the total risk could still be low. The Backfit Rule appropriately

• c.

places the burden on the Staff to demonstrate that the risk in such instances is such that it must be reduced by specific actions. That demonstration has not yet been made. By letter dated October 15, 1985 the Nuclear Utility Group on Station Blackout (NUGSBO) submitted to NRC technical comments regarding NUREG 1032, "Evaluation of Station Blackout Accidents at Nuclear Power Plants, Technical Findings Related to Unresolved Safety Issue A-44". These comments questioned apparent overstate-ments of risk. Although the Statements of Consideration accompanying the proposed rules refer to NUREG-1032 as the technical findings of the staff's study of the station blackout issue, and thus as their basis for concluding that a substantial increase in protection will be realized, there has been no public response by the staff to the NUGSBO critique. (The evaluation of NUGSBO work included in SECY-85-163A addresses an earlier proposal to resolve USI A-44, and does not address the group's comments on NUREG-1032). The Staff should be required to publicly respond to detailed criticisms of its technical work before that work is relied upon as the basis for new requirements. We recommend that the NUGSBO comments be addressed formally, e.g., as a supplement to NUREG-1032.

d. The analysis does not define the AC loads which are important in a station blackout and thus apparently misses the point that they are significantly less than the loads which must be accommodated in the event of a design basis loss of coolant accident (LOCA). This is particularly important in the case of two unit sites having three diesel generators, apparently representing some of the plants perceived to be most at risk. The Staff's analysis assumes that 2 of these 3 diesels would be required to accept load to assure safety in a blackout situation. In fact, proper load management procedures could so significantly reduce required ac power that one diesel could handle the loads or, at a minimum, greatly extend the time available to recover other sources of ac power. Proper consideration of this factor would decrease the number of plants perceived to be at risk and thus the level of increase in protection to be realized.

e. It is not clear to what extent credit is granted for the ability to manually start a diesel generator. Most diesel generator "failure" data reflects the ability of diesels to accept load in very short periods, on the order of seconds, following an automatic start. These requirements, again, are related to LOCAs. In the case

• of station blackout, much more time would be available to manually recover a diesel. Thermal inertia within the reactor systems can accomodate decay heat for some period. The Staff's own analysis acknowledges a minimum period of two hours coping capability at any nuclear power plant. This provides a significant period in which operator actions to restore AC power could be taken, including manual recovery of a "failed" diesel, and these actions should be credited. So doing would also reduce the perceived level of risk and the degree of increased protection which could be gained.

2. The regulatory analysis referenced in the backfit analysis is flawed:

a. We conclude that the benefits which would result from implementation of the proposed requirements are overestimated in the Staff's backfit analysis. The basis for this conclusion is discussed in Attachment 2.

b. Attachment 3 discusses the Staff's cost estimates and concludes that these are underestimated. This conclusion is reached despite the conservative assumption that costs for a coping study have been accurately estimated.

If the costs for potential special effects analyses are considered, the total cost estimate would exceed that presented in this analysis by an even greater margin.

c. Paragraph 3.1.4 of the proposed Regulatory Guide states, in part, "[t]he design adequacy and capability of equipment needed to function in environmental conditions associated with a station blackout should be evaluated."

This should not be difficult for equipment qualified for harsh environments in accordance with 10CFR50.49. Much equipment, however, is located such that pipe break accidents (the concern underlying 50.49) do not subject it to a harsh environment. Equipment located in such mild environments was not encompassed under 50.49, but could be affected to varying degrees by the loss of ventilation which would be a consequence of a station blackout. Experience gained in implementing 10CFR50.49 has shown that the costs of demonstrating equipment qualification is very high. These costs include analyses and tests. In many cases, in fact, costs of demonstrating qualification were found to be prohibitive and equipment was instead replaced. These costs are not considered in NUREG-1109.

d. The regulatory analysis indicates that the value-impact ratio would be improved by consideration of the on-site costs which would be predicted to be saved by averting core-damaging accidents through implementing the

• proposed requirements (averted on-site costs). We note that the Commissioners addressed the question of whether to base actions on consideration of averted on-site costs in their decision to publish provisional safety goals in 1983 (see NUREG-0880 Rev.l, Section IX).

Absent further action by the Commission, present NRC policy appears to be to place no reliance on averted on-site costs in justifying new requirements.

3. The Station Blackout Issue is not Generic:

a. NUREG-1109, "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout," reports that only 10 to 15 of 67 reactors considered in the analysis could be expected to require improvements in diesel generator reliability or modifications to increase coping capability. The remaining plants, a significant majority, are expected to require only a

• coping analysis and improved procedures. Thus, the staff's analysis acknowledges that all but a few operating plants are believed to already incorporate the capability to cope with a blackout for the durations which would be required by the proposed rule. Thus, it is inappropriate to require all licensees to take specific actions in response to this issue. The staff should, instead, utilize other means to further assess those plants having perceived vulnerabilities and/or to require actions at these facilities. The requirements of the Backfit Rule would apply to these plant-specific actions.

b. A generic analysis can be expected to significantly overestimate the cost/benefit for some plants and similarly underestimate the result for other plants when vulnerabilities are not uniformly distributed, as is the case here. In other words, the cost-benefit analysis is inaccurate for most of the plants to which it is generically applied, and use of a generic analysis is thus inappropriate. It follows that a generic cost-benefit analysis should only be used when there is a relatively small deviation in probable costs and benefits among the plants included in the analysis.

This test should be applied before a single generic cost-benefit analysis is used to justify a backfit. If the application fails the test, as it would in this case, plant-specific cost-benefit analyses or separate analyses of groups of plants which are sufficiently similar should be considered. This should not be construed to be questioning the use of cost-benefit analyses in relation to backfits. It is merely intended to recognize the limitations of such analyses and that their applicability should be examined on a case basis.

4. The Required Coping Study is Not Bounded

a. There is no specification in the rule or its accompanying Regulatory Guide which defines what must be demonstrated to show that a plant can "cope" with a station blackout.

It is conceivable that this demonstration could range from simply showing that a means of providing cooling water to remove decay heat is available, to assuring that instrumentation utilized under routine conditions to monitor plant status continues to be available, or to proving that no conditions in the plant will exceed nominal values. Our evaluation of the cost estimates in NUREG-11O9 (See Attachment 3) is premised on a coping study which is principally a battery load evaluation. Our cost estimate for performing such an evaluation is consistent with NUREG-11O9's estimates for a coping study. There are references elsewhere, however, to possible special effects analyses regarding issues such as environmental qualification of equipment (see Comment 2.c above). This open-ended possibility of a need for special effects analyses raises parallels to the situation which resulted from promulgation of 10CFRSO.48 and Appendix R, which also originally contemplated relatively modest analyses. Special effects analyses could significantly increase the resource burden for conducting coping studies beyond the levels predicted in NUREG-11O9.

b. It is not appropriate to require the majority of plants with a low risk of injury to the public from station blackout to spend resources on a coping study. The study itself provides no benefit. NUREG-1109 predicts a significant reduction in core melt probability at all plants, but any such improvement results solely from revised procedures and training, and not from the study. Revisions to procedures to maximize the extent to which DC power supplies can be conserved under station blackout conditions, and to assure that other actions which might be appropriate under such conditions will be taken, can be accomplished without a coping study. This is precisely the type of improvement which will result from the NUMARC initiative regarding this issue irrespective of the outcome of this rulemaking.

s. The additional comments by the Commission, included in the Statements of Consideration accompanying the proposed rule, question the need to make any modifications required "safety grade". As noted above, it is our conclusion that the proposed requirements have not been justified, and should not be imposed. Thus, questions of necessary safety classifica-

• 6.

tion are moot in this case. The separate views of Commissioner Asselstine suggest that the proposed requirements do not go far enough. The basis for this conclusion is a comparsion with perceived practice in European nuclear power plants. European practices have not, however, been explained sufficiently within the record of this rulemaking to conclude that they are better than or even significantly different from those in the U.S. In particular, it is not apparent that even the French plants, cited specifically by Commissioner Asselstine, can "cope" with a station blackout for extended periods because of the lack of definition for "coping" (see comment 4.a above). This subject is discussed further in Attachment 4.

7. The backfit analysis prepared for the station blackout rulemaking requires improvements to serve as a precedent for future rulemakings:

a. While we recognize that there will be disagreements between the Staff and industry on specific technical issues, we believe that the Staff has an obligation to address comments concerning the technical bases for new requirements. In the present instance, the Staff has not addressed the technical comments regarding NUREG-1032 submitted by NUGSBO (see comment l.c above).

We, and the Commission through enactment of the backfit rule, recognize that poorly justified rules may not improve safety and could result in unnecessary backfits. For future rules, the Commission should insist that all legitimate comments regarding technical basis documents be addressed before the documents are relied upon in backfit analyses.

b. The station blackout backfit analysis presents a discussion of the nine factors in 10CFRS0.109.(c) in a straightforward manner. However, the discussion of individual factors refers to other documents (e.g.,

NUREG-1109 or NUREG-1032), which often refer to additional documents (e.g., various NUREG/CR documents relied upon in NUREG-1032). In many cases, these references are not specific; they refer to documents by title only, and not to specific sections or pages within the documents. This makes it difficult to follow the logic of the analysis, and thus tends to make it inscrutable. We suggest that future backfit analyses be organized to more completely address, in a self-contained manner, the nine factors. Where references must be used, they should be keyed to specific sections, chapters, or pages in the supporting documents. I. c. The backfit analysis addresses several of the nine factors in 10CFRS0.109.(c) in a cursory manner. In particular, potential occupational radiation exposure and the expected burden on NRC have not been seriously assessed. It is recognized that the importance of these factors to an ultimate decision to impose new requirements will vary depending on the issue under consideration, and that they may not be of vital importance in this instance. Failure to adequately assess these impacts, however, does not provide the decisionmaker with an accurate understanding of their relative importance. Cursory treatment also provides an appearance of lack of diligence in conducting and reporting the backfit analysis. The treatment of each of the nine factors is critiqued individually in Attachment 1. Additional information is provided in the attachments which discuss: Attachment 1 - Critique of the Backfit Analysis. Attachment 2 - Review of the Stated Benefits of the Proposed Requirements. Attachment 3 - Review of the Cost Estimates of the Proposed Requirements. Attachment 4 - Comments on European Practices Regarding Blackout. ATTACHMENT 1 Critique of the Backfit Analysis In September 1985, the revised Backfit Rule, 10CFRS0.109(c) was published. The final rule specified that rulemaking should be subject to the backfitting requirements. The Commission majority concluded that "the Commission should fully understand the effects of a proposed backfit before its imposition". Section S0.109(c), ennumerates nine factors which must be considered by the Staff in its analysis. The backfit analysis supporting the proposed station blackout requirements was published along with the proposed rules. We conclude that the analysis is flawed as discussed below. Paragraph S0.109(a)(3) states, "The Commission shall require the backfitting of a facility only when it determines, based on the analysis described in paragraph (c) of this section, that there is

• a substantial increase in the overall protection of the public health and safety". Based on the Staff's statements regarding expected actions at individual plants and their overall cost estimate of $40 million, it is apparent that they do not expect to find many sites which can not presently cope with a station blackout. NUREG-1109 estimates the total costs for most reactors, on an individual basis, to be $225,000, most of which is the cost of a coping analysis. We interpret this to mean that installed equipment is expected to remain the same, having been proven by the coping analysis to be adequate. Given this interpretation, it is difficult to support a conclusion that a "substantial increase in overall protection" would be realized at most facilities by imposing the proposed requirements.

The backfit analysis is addressed point-by-point in the following paragraphs. Headings correspond to the nine factors required by 10CFRS0.109(c). This critique is intended to identify discrepan-cies which should be resolved before a decision to impose the requirements under consideration and, by example, to identify the degree of completeness we believe should be the norm for future backfit analyses. Item cl - Statement of the Specific Objective The Staff has described the perceived problem, but has not adequately described the objective. This relates to the need to demonstrate a "substantial increase" in protection as discussed above. Additionally, a cost-benefit ratio was calculated but no standard was specified. Item c2 - General Description of Activity Required by Licensees The Staff has outlined much of the information that they would require for a coping study. The required study is not precisely

ATTACHMENT 1 bounded, however (See comment 2 of the Enclosure to this letter). Furthermore, the Supplementary Information strongly suggests that safety grade equipment must be used for modifications and the Draft Regulatory Guide implies that equipment must be environmen-tally qualified. These types of related issues made addressing Appendix Rand the constantly "moving target" of requirements, reinterpretations, and compliance inspections difficult at best and frustrating to licensees, the Staff, and the Commissioners. The vague manner in which these proposed actions are described does not reflect the discipline that the revised Backfit rule was to produce. Item c3 - Potential Change in Risk to the Public We believe the analysis fails to establish a sufficiently grounded basis to determine public risk reduction. Benefits are acknow-ledged to be non-uniform and very plant-specific. The use of

• SST-1 source terms significantly overstates risk and thus potential risk reduction. This is discussed further in Attachment 2.

Item c4 - Potential Impact on Radiological Exposure of Facility Employees The analysis indicates no increase in occupational exposure is expected because equipment additions and modifications contemplated do not require work in and around the reactor coolant system. This argument is not well founded. Radiation exposure is received in work at nuclear power stations which does not involve work "in and around the reactor coolant system". It is not expected that occupational exposure resulting from the proposed rule would be high enough to be a dominant factor in decisionmaking, but it would not be zero. Item c5 - Installation and Continuing Costs Associated with the Backfit The costs of complying with the rule appear to be underestimated. The estimated cost for a coping study is perhaps sufficient to accomplish a battery load evaluation. That estimate is, however, likely well below the amount that would be expended if additional special effects analyses or demonstration of environmental qualification of all equipment is involved (See item cZ above). Even assuming that a battery load analysis would be adequate to fulfill the requirement for a coping study, total costs are still underestimated as discussed in Attachment 3. The analysis does not address potential delays in construction for Near Term Operating License plants. In fact, the manner in which the proposed rule would be implemented at NTOLs is not discussed ATTACHMENT 1 in the rule itself. Obviously, if any rule were to be adopted, NTOL's should have the same period to comply as other plants. If, somehow, it were concluded that compliance with the new require-ments must be demonstrated before licensing, delays could result for some facilities. The costs of such a delay at one NTOL could easily exceed the total estimated costs for implementing the rule at 67 reactors. Item c6 - Potential Safety Im~act of Changes in Plant or Operational Complexity Including the Relationship to Proposed and Existing Regulatory Requirements The analysis states that the proposed rule "should not add to plant or operational complexity". This statement would apply only to the plants at which no modifications are expected. Changes to a plant to increase its ability to cope would necessarily increase complexity, albeit only slightly. The conclusion in the backfit

• analysis reflects not that such an impact is known to be small, but rather that it has not been considered.

The discussion of the relationship to other requirements addresses other generic issues which are under ongoing review. We believe, however, that this element was included in the backfit analysis requirements, in large part, so that a logical priority could be established for implementing requirements. As such, it should not be limited to issues specifically affecting the same plant systems, but should include related regulatory changes. In this instance, the most important shortcoming of the analysis discussion is the omission of the effect of source term changes, including estimated time of containment failure. If the resolution to the source term question is near at hand, as we believe, a logical approach would have been to defer the station blackout evaluation until its effects were known. In fact, as discussed in Attachment 2, new source term information obviates the need for the proposed station blackout requirements .

• Item c7 - Estimated Resource Burden on NRC The estimate of 120 NRC man-hours per plant is questionable. This level of effort is perhaps sufficient to account for project man-ager attention to assuring that licensees meet their obligations and to preparing correspondence (the rule would require NRC to pre-pare at least one letter to each licensee, informing of concurrence or disagreement with the determination of required blackout dura-tion). It appears inadequate to account for technical review and evaluation of the determination of maximun coping capability and of the description of station procedures which the rule would require each licensee to submit. Comparison to the resources required for review of fire hazards analyses required by 10CFRS0.48 and Appendix R would be informative. The estimate would also appear to include ATTACHMENT 1 no allowance for inspection efforts to verify compliance, particu-larly if they are to include team efforts such as used for other issues (again fire protection provides a relevant example).

Item c8 - Potential Im act of Differences in Facilit Desifn, or Age on t e Relevancy and Practicality oft Back As noted elsewhere in these comments, the risk from station blackout is expected to be highly non-uniform as a result of precisely the kind of design differences which should be considered under this factor. The backfit analysis essentially states that the rule has been constructed such that licensees will be able to account for the differences. This appears to be an unwarranted shifting of the Staff's burden onto the licensees. Additionally, the effect of plant age is not addressed other than to assume, generically, that 25 years remain in an individual

• plant lifetime (for purposes of estimating benefits and costs in NUREG-1109). In fact, some facilities have little more than a decade remaining before expiration of their licenses. Performing detailed analyses and implementing any hardware modifications which may be found to be necessary could well take several years, based on past experience. For some facilities, therefore, only a limited amount of time would remain to actually realize any benefit from the changes. If an issue is such that age is immaterial as to whether a backfit should be imposed, that fact should be explicitly stated.

Item c9 - Whether the Backfit is Final or Interim To the extent that station blackout is a separate issue, the proposal is a final resolution. On the other hand, USI A-45, Shutdown Decay Heat Removal Requirements, is addressing all potential causes for inability to remove decay heat, of which lack of ac power is inherently a subset. Any new requirements resulting

• from A-45, therefore, have a potential for affecting the need for the proposed A-44 requirements, and thus makes them interim in nature. NUREG-1109 discusses this interrelationship and indicates that the resolution of the two issues is being coordinated.

Considering that resolution of A-45 is expected in the relatively near term, we conclude it would be more appropriate to defer implementation of any requirements for station blackout, assuming any could be justified, until the requirements which result from the A-45 program, if any, are known. We note that the discussion in NUREG-1109, although germane, is not referred to in the backfit analysis; this is another example of why we believe that future backfit analyses should be more self-contained. ATTACHMENT 2 Review of the Stated Benefits of the Proposed Requirements STATEMENT OF PROBLEM The consequences for a station blackout incident used in the value-impact analysis in NUREG-1109 are based on the following assumptions: The siting source term SST-1 fission product release was used to represent this event. The SST-1 source term was reduced by a factor of three to account for the differences between the station blackout

• event sequence and that reflected by the SST-1 source term.

In particular, containment failure is not predicted to occur for several hours for this event, barring recovery, whereas the SST-1 source term is representative of a prompt containment failure. The consequences of the event were taken from NUREG/CR-2723, with a reduction by a factor of five to account for the differences in the distance to which consequences are calculated. The NUREG/CR-2723 consequences were calculated for a distance of 350 miles whereas Enclosure 1 of NRR Office Letter No. 16 specifies the use of a 50 mile distance for regulatory analyses of safety issues. EVALUATION OF SOURCE TERM METHODOLOGY The methodology used in the calculation of the consequences of station blackout, while consistent with current NRC policy and procedures, is quite outdated. New source term information, which has been available for several years, would lead to the prediction of much lower consequences for this event. The NRC is presently in the process of updating their policy and procedures to specify the use of this new source term information. The NRC Regulatory Improvements Branch, Division of Safety Review and Oversight, has prepared a detailed implementation plan for the regulatory use of new source term information. This was discussed at the February 24, 1986 meeting of the ACRS Subcommittee on Class 9 Accidents. Of particular relevance, a revision to NRR Office Letter No. 16 with respect to the use of source terms in safety issue evaluation is scheduled for February, 1987. The major issue preventing the immediate revision of this document was stated to be the completion of NUREG-1150, which is presently scheduled for issuance in the summer of 1986.

ATTACHMENT 2 The NRC has published a major draft report, NUREG-0956 "Reassess-ment of the Technical Basis for Estimating Source Terms", in which several major conclusions are stated, including: Conclusion 1 "The BMI-2104 suite of computer codes represents a major advance in technology and can be used to replace the Reactor Safety Study methods." Conclusion 8 "Source terms were found to be depend strongly on plant design and construction details, thus making development of generic source terms difficult." Conclusion 10 "A comparative appraisal for the Surry plant using the Reactor Safety Study accident

• frequencies, source terms based on BMI-2104 results, and a preliminary reevaluation of the containment shows a reduction in estimated risk compared to the Reactor Safety Study."

The conclusions stated in draft NUREG-0956 are also accompanied by several recommendations, including: Recommendation 1 "The new source term analytical methods should be used to reevaluate regulatory practices that are based on Reactor Safety Study methods .*.* Improvements are so significant that utilization of the new methods is warranted while additional confirmator research is bein completed." emphasis added SOURCE TERM ANALYSIS The source terms presented in draft NUREG-0956 for the Surry plant for the TMLB' sequence (station blackout) with late containment failure due to overpressure are significantly lower than those in the Reactor Safety Study, as shown in Table 1. (Late failure due to overpressure is the most likely containment failure mode for such a sequence). Also included in Table 1 are release estimates for the TMLB' sequence for the Seabrook plant which were developed during the review of the Seabrook Probabilistic Safety Study, and reported in NUREG/CR-4540, February, 1986. These latter estimates are not based on detailed analyses using the NRC recommended code suite; they are based on extrapolation of previous results to the Seabrook evaluation. Additionally, the results of the IDCOR reference plant analyses for the Zion, Sequoyah, and Peach Bottom plants are included in Table 1. These analyses are based on a MAAP code analysis and are reported in IDCOR Reports 23.lZ, 23.1S, and 23.lPB respectively. Table 2 presents the same information in terms of the fraction of SST-1 values for each species.

                      ---- ----~-

ATTACHMENT 2 EVALUATION OF CONSEQUENCE METHODOLOGY The consequence methodology is based on the CRAC-2 computer code, which is the industry standard for calculation of consequences. The methodology used in this particular application interpolates consequence analysis results for 350 miles, as reported in NUREG/CR-2723 to a distance of 50 miles. The estimates were reduced by a factor of five to account for the reduction in distance for consequence calculations. There is no apparent reason to believe that this is not a fair estimate of a generic site consequence evaluation for an average site with a 50 mile consequence radius. However, the use of an average site consequence is questioned since there are a few highly populated sites which would heavily weight the average, and the risk from station blackout has been acknowledged to be non-uniformly distributed across the population of nuclear power plants *

• CONCLUSIONS The consequence values in NUREG-1109 do not reflect current knowledge of fission product source term behavior for severe accidents. The NRC has advised that a plan for implementation of new source term information into the regulatory process is underway, with an expected completion date of February, 1987.

Assuming they could otherwise be justified, implementation of any requirements resulting from resolution of US! A-44 should be deferred until the results of the source term research can be taken into account. This conclusion is supported by the fact that the consequences used in the value-impact analysis of NUREG-1109 would be reduced by an additional factor of approximately 10 or more, thereby rendering any of the alternatives UNFEASIBLE. The factor of 10 is obtained by comparing the results of the recent analyses of station blackout to the NUREG-1109 source term as shown in tables 1 and 2. These recent analyses indicate that the release fractions for station blackout, for the fission product

• release categories which are dominant contributors to offsite consequences (iodine, cesiums, and telluriums), are overestimated in the NUREG-1109 report.

The Staff's analyses should be reperformed, untilizing the best information presently at hand, before being relied upon as the basis for justifying new requirements. Table 1 Fission Product Releases for the TMLB' Event (all values are fraction of core inventory) STUDY Noble Iodine Cesium Tellurium Barium Ruthenium Lanthanum Gas SST-1 1.0 4.5-1 6.7-1 6.4-1 7.0-2 5.0-2 9.0-3 9.0-1 7.0-1 I 5.0-1 3.0-1 6.0-2 2.0-2 4.0-3 2) G 0956 1.0 7.8-3 3.9-4 8.5-2 1.8-2 3.3-6 8.1-5 NUREG/CR 1.0 2.4-2 2.4-2 3.0-2 2.6-3 2.3-3 3.9-4 4540 IDCOR 1.0 1. 7-3 1. 7-3 2.0-5 1. 0-5 1. 0-5 1. 0-5

• Zion IDCOR 1.0 5.1-4 6.4-4 2.6-5 1. 0-5 1. 0-5 1.0-5*

Sequoyah IDCOR 1.0 5.0-2 5.0-2 6.0-2 8.0-5 1. 0-4 1.0-s* Peach Bottom (TQVW)

• *Based on independent calculations performed in support of IDCOR task 23 rather than integrated MAAP analysis.

Table 2 Fission Product releases for the TMLB' Event (all values are fraction SST-1 source term) I STUDY Noble Iodine Cesium Tellurium Barium Ruthenium Lanthanum Gas

 .      1  1.0     1.0      1.0      1.0          1.0       1.0       1.0 RSS      0.90    1. 55    0.75     0.47         0.86      0.40      0.44 (PWR 2)

NUREG- 1.0 0.017 0.00058 0.13 0.26 0.000066 0.44 0956 NUREG- 1.0 0.053 0.036 0.047 0.037 0.046 0.043 4540 IDCOR 1.0 0.0038 0.0025 0.000031 0.00014 0.00020 0.00011 Zion IDCOR 1.0 0.0011 0.00096 0.000040 0.00014 0.00020 0.00011 Sequoyah

 -~

Bottom (TQVW) 1.0 0.11 0.075 0.094 0.0011 0.00200 0.00011

                --------~----------*-- - -------------

ATTACHMENT 3 Review of the Cost Estimates of the Proposed Requirements The Subcommittee on Cost-Impact generated cost estimates indepen-dently to compare to those presented in Table 6 of NUREG-1109. To accomplish this, certain assumptions had to be made regarding work-scope or the nature of modifications which might be required. Since specific changes are not incorporated in the proposed requirements, the conclusions drawn from this effort are neces-sarily quite general and subject to large uncertainties. Specific assumptions and/or conclusions regarding individual elements of this evaluation are summarized below. The overall conclusion is presented in Table 3, which contrasts our estimates to those of the NUREG. We conclude that the cost of complying with the proposed requirements have been underestimated by nearly a factor of 2 assuming, conservatively, that coping

• study costs have been correctly estimated by NRC.

As discussed elsewhere in these comments, we are concerned that the effort involved in a coping study could expand greatly as a result of special effects analyses which may be interpreted to be a required part of the study's scope. Experience with previous regulations involving unbounded analyses subject to NRC review (e.g., Appendix R) leads us to conclude that coping study costs could be significantly more than estimated. This would not only increase the total industry impact, but more importantly would increase the burden on every nuclear power plant licensee regard-less of the present ability of their facility to accommodate a station blackout event. This would further decrease the plant-specific cost-benefit justification for the majority of facilities which appear to present little risk from station blackout. DISCUSSION

• 1* Coping Study o The evaluation of the coping study costs was based on a definition of scope for conduct of a battery load analysis provided within NUREG/CR-3840 and the proposed regulatory guide. Excluded from consideration were component performance analyses or proof tests for operating conditions beyond the equipment's original design criteria, (e.g., performance proof test at degraded voltage levels). Also excluded from considera-tion were possible special effects analyses. Should such tests and analyses be required to meet coping criteria, the costs could be expected to escalate significantly; perhaps by as much as 5 to 20 times the best estimated values described in this evaluation.

ATTACHMENT 3 o Our evaluation indicates that the costs for conducting a coping study could range from as little as $40,000 to as much as $250,000 . Our best estimate of the cost per reactor is $140,000 which closely approximates the NRC value of $150,000 o We conclude that the NRC best estimate figure of

         $150,000 per reactor is reasonable given the assumed limitation on scope described above. It is not clear, however, that this limitation accurately reflects the desired coping analysis.

2. Procedures/Training 0 Our evaluation indicated a likely cost ranging from

         $50,000 to as much $200,000 to complete the required

• procedure development and training. Our expected best estimate value of $90,000 includes both training and procedure development and assumes no credit for procedures which may currently be in place. From our investigations, we conclude that many utilities currently have procedures describing load management practices during loss of AC scenarios. It would not be unreasonable, therefore, to assume that the average cost across the plant population may actually be lower than the expected value provided here. These costs do not include, however, extensive technical defense of the adequacy of procedures in support of a technical NRC review.

3. Improved Diesel Generator Reliability 0 Our estimate of the cost to conduct a "reliability investigation" closely approximates the NRC value of

• 0

         $100,000 per reactor.

The cost estimates for equipment modifications which may alter fuel systems, electronic, sequencing, and / or other equipment, are based on some general assumptions regard-ing these modifications. Equipment modifications are expected to be complex, and there is a great deal of uncertainty as to the benefit in improved reliability for any of the assumed changes. o Our evaluation indicates that some modifications might be incorporated for as little as $150,000 as suggested by the NRC. However, it is expected that these costs are more likely to fall in the range $1 - 1.5 million for the majority of the reactors requiring these improvements. ATTACHMENT 3 o Our expected value of $1 million per reactor for diesel generator reliability improvements includes both the cost of modifications, as well as the initial reliability investigation. However, specifically excluded are replacement power costs which may be incurred to effect requalification of the diesel generator.

4. Increased Blackout Coping Capability with Plant Modifications 0 Due to the lack of design detail associated with the potential modification to station batteries, condensate storage, and/or instrument air systems, we chose to revisit NUREG/CR-3840, "Cost Analysis for Potential Modifications to Enhance the Ability of a Nuclear Plant to Endure Station Blackout". This NUREG provides some detail of the NRC's initial estimates in this regard and

• 0 is purported to result in expected costs similar to those provided within the NUREG-1109 analysis.

Our earlier evaluation of NUREG/CR-3840 provided evidence that the proposed modification costs were significantly underestimated (refer to letter from Murray Edelman to the Executive Director for Operations dated November 30, 1984). At that time, we expressed concern that the Dodge Manual for building construction pricing and the R.S. Means Mechanical and Electrical Cost Data Handbook, both used by NRC, were inappropriate sources/references for estimating costs for nuclear power plant construction. The unit rates provided therein for installation are significantly understated as compared to actual nuclear industry experience. The industry provided evidence that the unit prices for installation of mechanical/electrical components, in general, were two to five times lower than what could be

• expected in nuclear plant construction/modification. In addition, material cost data from these sources has historically been 25% to as much as 125% lower than material pricing experiences for nuclear applications.

In a handbook for cost estimating published in October 1984 (NUREG/CR-3971), the NRC adopts the Energy Economic Data Base as a reference source for unit pricing. We believe the methods described within this handbook (NUREG/CR-3971), provide a much more reasonable basis for estimating the cost impact of potential modifications. o Our evaluation indicates that as a result of the factors discussed above, the NRC best estimate value of $1 million per reactor underestimates by a factor of 2 to 3 the cost which a utility can expect to incur in effecting proposed modifications.

• Table 3 COMPARISON OF COST ESTIMATES FOR COMPLIANCE WITH PROPOSED RESOLUTION TO US! A-44 NRC AIF

               # of Reactors       Best Estimate ($000)         Best Estimate ($000)

Modification Considered Per Reactor Population Per Reactor Population

1. Coping Study 67 150 10,000 150 10,000

2. Procedures/ 67 75 5,000 90 6,000 Training

3. Improve D/G 15 250 11,000 1,000 20,000 Reliability

4. Inc. Blackout 10 1,000 10,000 2,500 25,000 Coping Capability.

w/Plant Mods. TOTAL INDUSTRY IMPACT 36,000 61,000

ATTACHMENT 4 Comments on European Practices Regarding Blackout We support the concept of considering the experience and practices of foreign nuclear power plants in deciding on appropriate actions for U.S. plants. However, it is absolutely necessary that we accurately understand the totality of the resulting comparison. Besides explicit differences, we must know what is being done, how and why it is being accomplished, and how differing regulatory requirements (e.g., single failure, etc.) influence the perceived need for certain requirements. In the present case, it does not appear that any of this informa-tion is available. References to European practices consist of comments within the Statements of Consideration and backfit analysis, and overview summary statements during a November 14,

• 1985 Commission briefing, all of which are lacking in technical detail. Before using these statements as a basis for regulatory action which will affect each U.S. nuclear power plant, the Commissioners should obtain more complete information from which conclusions can be drawn.

It goes without saying that foreign regulatory agencies occasional-ly require activities which differ from or exceed those required in the U.S. It is equally true, however, that the basic design of most nuclear power plants regulated by European agencies are fund-amentally similar to U.S. designs. They include both PWRs and BWRs, many of which have been constructed under cooperative arrangements involving U.S. reactor vendors. Commissioner Asselstine, in particular, cites actions reportedly being taken at French nuclear power plants to provide additional protection from core damage resulting from station blackout events. These actions include a turbine-driven pump to provide

• cooling to reactor coolant pump seals. The motivation for this addition was apparently concern regarding the potential for gross leakage, on the order of several hundred gpm, which might begin soon after loss of seal cooling (as would occur in a station blackout). Subsequent to the decision to install this pump, extensive testing was conducted of the reactor coolant pump seals. This testing demonstrated that the feared gross failures are not likely to occur, and that leakage following a loss of seal cooling is only a small fraction of that expected.

Commissioner Asselstine further refers to a reported ability of French plants to cope with a station blackout for up to three days. Again, the details of this capability have not been provided, although it appears to relate to the availability of water supply for decay heat removal. (There is also some confusion as to whether capabilities are for three days or for 20 hours, both of

ATTACHMENT 4 which have been referred to in NRC statements). With respect to the availability of water, this capability does not appear to be significantly greater than that which is available at most U.S. plants which could utilize such sources as self-driven fire pumps to provide makeup to decay heat removal systems. Fire trucks could also be utilized in many cases because of the extensive period in which remedial actions could be taken. It is not apparent that the French have addressed other issues wh ich have been referred to in the context of "coping", particularly the issue of qualification of equipment for the conditions which could be hypothesized following a loss of ventilation in a blackout event. In summary, we are concerned that more is not known than is known about the actual capabilities of European plants and the reasons behind those capabilities. The references to foreign experience which are available in the rulemaking record are too brief. Until more detailed information is provided, we believe that no conclu-sions can be drawn regarding the effect that European practices should have on station blackout requirements in the U.S . John D. O'Toole Vice President Consolidated Edison Company of New York, Inc. 4 Irving Place, New York, NY 10003 Telephone (212) 460-2533 June 1 Re: Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, D. C. 20555

• ATTN:

SUBJECT:

Dear Sir:

Docketing and Service Branch Proposed Rule - 10 CFR Part 50.63 Station Blackout On March 21, 1986 the Nuclear Regulatory Commission ( the "Commission" or the NRC") published in the Federal Register a proposed revision to 10 CFR Part SO to include 10 CFR Part SO. 63 "Station Blackout" requiring that light-water cooled nuclear power plants be capable of withstanding a total loss of alternating (AC) electric power for a specified duration and maintaining reactor core cooling and containment integrity during that period. Consolidated Edison Company of New York, Inc. ("Con Edison"), as holder of Operating License DPR-26 for Indian Point Unit No. 2 (IP2), is pleased to provide the following comments on that proposal

• We take this opportunity to express our agreement with, and support for, the comments submitted by the Nuclear Utility Management and Resources Committee (NUMARC) on the proposed rule. We also offer the following comments concerning features and capabilities unique to IPZ which evidently have not been taken into consideration by the staff in its proposed rule and the generic backfit analysis. These may also be applicable to other plants.

At IP2 normal off-site AC power supply consists of two (2) 138 Kv feeders with back-up from three underground 13. 8 Kv feeders. This arrangement, which has proven to be extremely reliable, along with a continuing enhancement of grid stability has significantly minimized the possibility of a loss-of-offsite event at IP2. Nevertheless, off-site AC power restoration procedures are in place which require high-priority restoration of power to the Indian Point site in the event of loss-of-offsite power. Acknowledged by o.r,/).t.J.re. .~

I

             .,I I

I I I I

  -            I I

I (\ \

By original plant design, IP2 has three (3) onsite emergency diesel generators. A minimtnn of two generators is required for maintaining FSAR design basis safeguard system load for hypothetical accidents, and any one of the three diesel generators can be used to provide sufficient emergency power for the load associated with a station blackout condition. The IP2 diesel generators have a proven track record for starting and loading on demand. Our response to Generic Letter 84-15 indicated a reliability of 100% for the IP2 diesel generators. Additionally, in the unlikely event that all offsite and onsite emergency power is lost, three (3) gas turbine generators are directly available to IP2. One is located onsite and dedicated to IP2 and two additional units are located at the adjacent Buchanan Substation connected to IP2 via underground feeders. The limiting conditions for operation and surveillance requirements for these gas turbines are included in our Technical Specifications. One gas turbine generator is more than adequate to provide electric power for mitigating a station blackout event and for bringing the plant to a safe shutdown condition .

• Taking these features into consideration for IP2 results in a significant reduction in the risk posed by a station blackout event for this Unit, and brings into question the practicality of a generic rulemaking addressing such a highly site-specific issue.

Nonetheless, we have reviewed the proposed rule's requirements that apply to the needs for per£ orming an IP2 "coping analysis". The engineering and design resources and outside contractor resources necessary to perform such an analysis are similar to those required for analyses done for 10 CFR SO, Appendix R in order to ascertain the need for and justify exemptions and alternative shutdown configurations and develop and implement plant modifications. The engineering costs associated with these types of analyses, asstnning no modifications result, are underestimated in the backfit analysis accompanying the draft rule. For IP2, these costs would be approximately four times the amount estimated by the staff. For example, assessing environmental conditions in rooms containing equipment needed during a station blackout without HVAC requires substantial time-dependent computer analysis and room modeling .

• Such dynamic environmental analyses without HVAC are expensive to perform and are currently not available for IP2 except for some components relied on for alternate safe shutdown pursuant to the requirements of Appendix R. In lieu of per£ orming such analyses for equipment, such as that in the Central Control Room (CCR), containment, etc, the rule should permit use of the Appendix R minimtnn set of components and alternate shutdown from outside the CCR. In our judgement, when the costs of potential modifications and the engineering resources to develop them, are added to cost of the study, the total could double the maximtnn amount given in the backfit analysis. While we believe IP2 can appropriately accorrnnodate the station blackout condition in its present configuration by using the alternate safe shutdown system, we would be required by the proposed rule to demonstrate this through an extensive study and doctnnentation effort.

We believe that the time, resources and costs of compliance with this proposed rule are not commensurate with any incremental safety improvement for plants such as IP2; therefore, the NRC could better direct its efforts toward maintaining the reliability of on-site and off-site power sources. Accordingly, for the site specific characteristics described above, the rule, as it applies at least to IP2, should be revised to refocus the required efforts and resources on achieving and maintaining a high reliability of the AC power systems rather than require an expensive coping study and mitigating features. Additionally, we believe that the rule should allow a licensee to identify its limiting capability to fulfill a required safety function for coping with a station blackout event and proposed mitigating measures without the need for expensive coping studies. Determining such a limiting capability can be derived from earlier conservative station blackout studies in response to Generic Letter 81-04, and in work done for Appendix R, and Reactor Coolant Pump seal performance work being done

• by the Westinghouse Owners Group. We point out that certain Appendix R scenarios can lead to a station blackout event, and mitigating features have already been provided to shutdown from outside the CCR, as well as revisions made to emergency operating procedures for Generic Letter 81-04 for shutdown from inside the CCR. Using conservative engineering judgements and a qualitative assessment of the earlier work, limiting capability can be determined and any necessary mitigating measures proposed, without the need for burdensome and time-consuming studies.

We hope that the areas of concern noted above will be of some assistance in revising Part 50 into standards which both the Commission and industry find acceptable. We appreciate the opportunity to comment on this important rulemaking proceeding. J ruly yours '.

• Jo~ ~ ~- - -

Vice President cc: Senior Resident Inspector U.S. Nuclear Regulatory Commission P. O. Box 38 Buchanan, New York 10511

NRC 88 WPSC[414)433-1234 TELEX 51010 12698 WPSC GAB TELECOPIER [414] 433-1 297 EASYUNK 62891 993 WISCONSIN PUBLIC SERVICE CORPORATION 600 North Adams

• P.O. Box 19002
• Green Bay, WI 54307-90 oocn JUN 2 09R1i p

June 18, 1986 (!j) Secretary of the Commission

• Attention Docketing and Service Branch U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Gentlemen:

Docket 50-305 Operating License DPR-43 Kewaunee Nuclear Power Plant Comments on Proposed Rule 10 CFR 50.63 Station Blackout Reference~: 1) 51 FR 9829 Publication of the Proposed Station Blackout Rule We have reviewed the above Federal Register Notice (reference 1) and find that we do not agree that rulemaking is required to resolve this issue. WPSC is not a member of the Nuclear Utility Group on Station Blackout (NUGSBO) although we have formally agreed to support their effort on this issue. WPSC is a member of the Nuclear Utilities Management and Human Resources Committee (NUMARC}, which has undertaken to resolve this technical issue through direct NRC/utility interaction and thereby preclude the need for rulemaking. This letter provides WPSC concurrence in support of the NUMARC comments on the proposed rule. We take this opportunity to encourage the Commission to pursue resolution of this issue with NUMARC to our mutual satisfaction. J;~ D. C. Hintz Manager - Nuclear Power DJM/jms cc - Mr. George E. Lear, US NRC Mr. Robert Nelson, US NRC Mr. James H. Miller, Jr., Georgia Power Co. Mr. John Opeka, Northeast Utilities Service Co.

                                                                -.:knowlerlged by ~rrJ. 1.//.~-

• t t

NORTHEAST UTILITIES General Offices

• Selden Street, Berlin, Connecticut THE CONNECTICUT LIGHT ANO POWER COMPANY

((ill WESTERN MASSACHUSETTS ELECTRIC COMPANY HOLYOKE WATER POWER COMPANY NORTHEAST UTILITIF.S SERVICE COMPANY NORTHEAST NUO..EAR ENERGY COMPANY P.O. BOX 270 HARTFORD, CONNECTICUT 06141 -0270 (203) 665-5000 June 18, 1986 50-213 50- 245 5 0-336 50-423

                                                                                      - - ~~      l,J" 21 4 Q Secretary of the Commission Attn:  Docketing and Service Branch U.S. Nuclear Regulatory Commission Washington, D.C. 20555

• Gentlemen:

Haddam Neck Plant Millstone Nuclear Power Station, Unit Nos. Comments on Proposed Station Blackout Northeast Nuclear Energy Company (NNECO) and Connecticut Yankee Atomic Power Company (CYAPCO) offer the following comments regarding the proposed Station Blackout Rule which appeared in the Federal Register on March 21, 1986 (51FR9829): NNECO and CYAPCO agree with and support the comments whic h the Nuclear Utility Management and Resources Committee (NUMARC) is filing on the proposed Station Blackout Rule: NNECO and CYAPCO have reviewed the station blackout issue and the proposed rule, with its regulatory and backfit analysis, and have concluded that many of the concerns related to station blackout can be alleviated through the industry initiatives being pursued by NUMARC. We believe the

 - Industry initiatives will reduce the ind i vidual sites' contribution to the overall risk of station blackout through the enhancement of

• the reliability of on- site and off-site AC power systems. We believe
• this attention to AC power reliability is a more appropriate means to address the risk from station blackout than the proposed rule which focuses on coping from a station blackout event.

We concur with NUMARC's detailed technical commen t s on the proposed rule, regulatory guide and backfit analysis. Further, we offer the following specific comments. Millstone Nuclear Power Station, Un i t No. 2 util i zes a Combustion Engineering (CE) designed nuclear steam supply system. The CE owners Group has maintained the position that reactor coolant pump seal cooling is not necessary to ass~re integrity of the reactor coolant pump seals in an idle pump. Furthermore, complete loss of seal function will not occur due to station blackout in a CE designed nuclear steam supply system. It follows, then, that seal cooling should not be necessary to mainta i n seal integrity during a station blackout event. This position, when f actored into the Staff's

technical justification for the proposed rule, further supports NUMARC's position that a generic station blackout rule is not necessary. The Staff indicated during their November 14, 1985 Commission briefing on the resolution to the station blackout issue that Hurricane Gloria provided further justification for the need of a rule to resolve this issue. The Staff cited Millstone Nuclear Power Station as having had a 20 hour loss-of-offsite-power event. Our December 31, 1985 letter discussed in detail the actions taken by NNECO and CYAPCO prior to, during, and after Hurricane Gloria affected the Millstone Station. As the storm reached its peak, it became evident that, because of a lack of any effective rainfall, a heavy buildup of salt spray was taking place as evidenced by an increased frequency of arcing on outside transformers, switchyard transmission lines and circuit breakers. Steps were taken to bring the units offline. All the Millstone emergency on-site AC power

• sources successfully started, were loaded and run until prudent plant actions were completed to allow for restoration of normal off-site power. (On September 27, 1985 Millstone Unit No. 3 had not yet loaded fuel and accordingly did not have core cooling requirements.

However, upon loss of normal power, the emergency diesel generators automatically started and loaded). If necessary, Millstone Unit No. 1 could have had off-site power restored within 3 1/2 hours and Millstone Unit No. 2 could have had off-site power restored within 5 1/2 hours. Since more rapid restoration of off-site power was not vital, NNECO elected to pursue a more deliberate and thorough cleaning and checking restoration process. The advance notification associated with severe weather events of this kind permits advance precautionary actions not usually credited by the Staff or in plant probabilistic safety studies. In conclusion, NNECO and CYAPCO support NUMARC's initiatives and comments regarding the Staff's proposed station blackout rule. Further, a review of the station blackout issue and the proposed rule, with its regulatory guide and backfit analysis, indicates that most of the anticipated benefits from the proposed rule would be derived from improvements at a limited number of plants and therefore the station blackout issue should be resolved in a manner other than rulemaking. Very truly yours, CONNECTICUT YANKEE ATOMIC POWER COMPANY NORTHEAST NUCLEAR ENERGY COMPANY J.F.eka Senidr Vice President

IQ :f19 Pennsylvania Power & Light Company Two North Ninth Street

• Allentown, PA 18101
• 215 I 770-5151 Harold W. Keiser Vice President-Nuclear Operations 21s1no-1so2 I

JUN 2 o 19Afi lo June 18, 1986 SER~~ 8ECY-lUiC The Secretary of the Commission U.S. Nuclear Regulatory Commission Washington DC 20555 Attention: Docketing and Service Branch SUSQUEHANNA STEAM ELECTRIC STATION COMMENTS ON PROPOSED STATION BLACKOUT RULE FILE Al7-20F DOCKETS 50-378 PLA-2667 50-388

Dear Sir:

On behalf on Pennsylvania Power & Light Co., we wish to express our agreement with, and support for, the comments which the Nuclear Utility Management and Resource Committee is filing on the proposed station blackout rule (USI-A-44). PP&L would like to add the comments contained in Attachment 1 to this letter concerning features and capabilities unique to Susquehanna Units 1 and 2 which have not been taken into consideration by the Staff in the rulemaking. Recognition of these factors results in a reduction of the risk and further calls into question the need for a rule. We wish to thank the Commission for the opportunity to comment in this issue. If our comments should engender further questions by the Commission we stand ready to answer them. Very truly yours, Vice President - Nuclear Operations Attachment(s) cc: M. J. Campagnone - USNRC L. R. Plisco - USNRC

I ~~ I s n '

Page 3 SSES PLA-2667 File Al7-20F ATTACHMENT 1 TO PLA-2667 Draft Regulatory Guide and Value Impact Statement Page 9, Item 32 - Credit should be given for action outside of the control room after a periods of time. For example, the hanging of jumpers to prevent automatic transfer of HPCI from the condensate storage tank to the suppression pool should be allowed, provided procedures exist and the operators receive adequate training *

• Draft Report NUREG - 1109 Page 1: Asterisk Susquehanna's Individual Plant Evaluation (IPE) shows that for station blackouts which end in core damage, 58 percent terminate with the vessel intact.

Page 2: Objectives The objective of making station blackout a relatively small contributor to core melt is open-ended. A more appropriate objective would be to establish a measurable standard for evaluating contribution to core melt. Page 3: Item iii Susquehanna utilizes and takes credit for load shedding, use of fire protection system pumps and use of bottled air.

EDISON DRIVE ATom,c POWER comPAllY . AUGUSTA, MAINE 04336 (207) 623-3521 uocn-,.

                ~l u

IJ'Ult8£,r

             ~POSED RULF PR - so June 17, 1986 MN-86-82 GDW-86-147 4{st F£. 9fc99)

Secretary of the Commission United States Nulcear Regulatory Commission Washington, D.C. 20555

• Attention:

Subject:

Gentlemen: Docketing and Service Branch Comments on Proposed Station Blackout Rule Maine Yankee Atomic Power Company offers the following comments regarding the proposed Station Blackout Rule which appeared in the Federal Register on March 21, 1986 [51FR9829]. We agree that Unresolved Safety Issue A-44, Station Blackout", should be 11 addressed to provide a high level of assurance that public health and safety will not be affected by multiple failures of AC power systems. However, we believe that the proposed rule is flawed, inappropriate, and not the best way of achieving the intended objective. We concur and fully support the comments on the proposed rule submitted by NUMARC and offer the following additional specific comments:

• 1. The reliability of AC power systems and the ability to cope with a failure of such systems is very much site and unit specific. The proposed rule would require remedial measures which would be inappropriate, ineffective, and costly in many plants. Such generically imposed changes may serve to make plant systems more complex, less reliable, and perhaps less safe.

A requirement that each nuclear unit have a very high level of AC power system reliability or be able to cope with a station blackout would be more appropriate and effective than the proposed rule. Where coping time is required, it should be proportional to reliability. Licensees could then select the best of alternative remedies for their particular site specific conditions, i.e., improving reliability and/or improving coping ability. 7710L-SDE

N

MAINE YANKEE ATOMIC POWER COMPANY United States Nuclear Regulatory Commission Page Two Attention: Docketing and Service Branch MN-86-82

2. The proposed rule should be more specific regarding its intended safety objective. In our view the objective should be to provide a high level of assurance that the reactor can be placed in a hot shutdown (or safer) condition, with the ability to remove decay heat, maintain reactor water inventory, and the reactor subcritical for an indefinite period of time following the loss of all offsite power and a loss of two redundant on-site emergency power sources.

Licensees should be able to take credit for other on-site power sources and directly connected highly reliable nearby power sources such as hydro and gas turbine units.

3. The proposed rule is inconsistent. The summary parenthetically defines a station blackout as a 11 toss of both offsite power and onsite emergency AC power systems ... 11

• The text defines it as a 11

              *** loss of offsite AC power to the essential and nonessential electrical buses ..... and ..... onsite emergency AC power systems ... 11 (emphasis added).

4. We are concerned that, if the proposed rule is adopted, the staff will promulgate regulatory guidance criteria which will be unrealistic and excessive, i.e., compounding the event with other accidents, imposing passive failure criteria, applying seismic, environmental qualification and other qualifications to equipment that could otherwise be used in response to such an event, etc.

The cost impact of such a rule could be much higher than than anticipated.

5. We believe the assumptions used in determining the benefits associated with such a rule are unrealistic. These assumptions lead to an unrealistically high public exposure which is used to justify

• the cost.

In summary, we believe Unresolved Safety Issue A-44 should be addressed, however, the proposed rule will not achieve the desired objective and may, in fact, have a negative impact on safety. It will be costly to implement and its benefits are perhaps exaggerated. Very truly yours, MAINE YANKEE ATOMIC POWER COMPANY

                                           ~~--

J. B. Randazza Executive Vice President cc: Mr. Ashok C. Thadani Mr. Pat Sears Dr. Thomas E. Murley Mr. Cornelius F. Holden 771OL-SDE

0 Public Serviceru DOC:KfTEo USNRc Public Service Company of Colorado 2420 W. 26th Avenue, Suite 100D,1ifnver, June 18,. Fort St. Unit No.

                                                               -86433
                                                                             /~

Secretary o e Corrmission U.S. Nuclear Regu atory Commission oocm rw**"pff ~(J

• Washington, D.C. 20555

                                                'IWlPO D RULE     ...               )

Attn: Docketing and Service Branch (SI r:;-t:_ 9 fd9 Docket No. 50-267

SUBJECT:

Proposed Rule on Station Blackout (USI A-44) Gentlemen: On March 20, 1986, the Co111T1ission published the Proposed Rule on Station Blackout to resolve unresolved safety issue (USI) A-44. This proposed rule is to amend NRC regulations to require light-water-cooled nuclear power plants to be capable of withstanding a total loss of alternating current (AC) electric power from normal and first line back-up sources for a specified duration and maintain reactor core cooling during that period. Applicability: The rule is ambiguous as to its applicability to the Fort St. Vrain HTGR. The proposed section 50.63(a) refers to "each light-water-cooled nuclear power plant", 50.63 (c) refers to "each l i ght-water-cooled nuclear power plant", 50.63 (d) refers to "For each light-water-cooled nuclear power plant". Draft guide SI 501-4 indicates on page 1 "This guide applies to all corrrnercial light-water-cooled nuc l ear power plants, and on page 3 "The issue is of concern for both PWRs and BWRs. The proposed section 50.2 refers to "essential and nonessential switchgear buses in a nuclear slant", Appendix A, Criterion 17 refers to "the loss of power generate by the nuclear power unit, and section 3.1 on page 6 of draft guide SI 501-4 refers to "each nuclear power plant". The Supplementary Information to the Proposed Rule on Station Blackout states: "In addition, the Commission is proposing to amend its regulations by adding a new 50.63 and by adding a new final paragraph to General Design Criterion 17, Appendix A of 10 CFR Part 50, to require that all nuclear power plants be capable of coping with a station blackout for some specified period of time."

vl ON

 ' U, $ . r UCL D

0 (

• Post*

Co leS I k ,f;/r, ----- I .- Add I

    !:,p  I"' I

P-86433 Page 2 June 18, 1986 As written the applicability of the regulation to the Fort St. Vrain HTGR is at best questionable for the following reasons:

1. Electric power is normally supplied to Fort St. Vrain by five 230 KV transmission lines through a common substation. These normal sources are backed up by two emergency electric generators driven by four diesel engines.

2. A fire protection electric generator (Alternate Cooling Method generator) is available and completely independent of all other electric sources and cables and is capable of supplying power to essential cooling equipment to protect the health and safety of the public.

3. The background information for the Proposed Regulation and Regulatory Guide says the purpose is to reduce the probability and frequency of core melt due to station blackout. The Fort St .

• Vrain HTGR core consists primarily of graphite which does not melt under accident conditions and, therefore, is not a contributor to core melt frequency statistics.

Site Specific Analysis: The severity of station blackout should be determined on a site specific basis. Light-water reactor sites vary in availability of offsite power supplies, emergency generators, supplemental generators, engine driven pumps or cooling water with natural head. The rule should be made clear as to the difference between emergency generators and other generators that are onsite for supplemental purposes. Otherwise, statements dictating a loss of all onsite emergency generation have different implications for different plants. Diversity of Solutions:

• The station blackout rule should be clarified to allow credit for stations with diverse and very reliable off-site power sources or diverse and very reliable onsite electrical generation. The rule should also permit reliability improvements in these areas to reduce the time period required for alternate cooling. In this fashion reliability improvement programs, diverse onsite generators, or transmission line additions may be included as solutions to the station blackout problem. The rule should be clear as to the requirements for accident scenario boundaries with regard to additional onsite generators, AC power from battery source inverters, and engine driven pumps during the station blackout scenario.

i

P-86433 Page 3 June 18, 1986

Conclusion:

The new section 50.63 (b) 11 Limitation of Scope" should state specifically that the regulation is not applicable to HTGRs and that the HTGR will be considered individually based on specific plant features. Very truly yours,

                                 ~ry~ryrfl~
                                ..l- H. L. Brey, Manager
                               ~- Nuclear Licensing and Fuels HLB/JW:jmt

Southern California Edison Company P. 0. BOX 800 2244 WALNUT GROVE AVENUE ROSEMEAD, CALIFORNIA 91770 M.O. MEDFORD MANAGER, NUCLEAR LICENSING June 19, 1986 Secretary of the Conm1ss1on U.S. Nuclear Regulatory Conm1ss1on Wash1ngton, D.C. 20555 Attent1on: Docketing and Serv1ce Branch u

Dear S1r:

Subject:

Reference:

Federal Reg1ster, Volume 51, Number 55, Fr1day March 21, 1986, "Proposed Rules" In the referenced ed1t1on of the Federal Reg1ster, the Nuclear Regulatory Conm1ss1on proposed a rev1s1on to 1ts regulat1ons in 10 CFR Part SO to address Stat1on Blackout. Publ1c conments were 1nv1ted on the proposed rule through June 19, 1986. Southern Cal1forn1a Ed1son apprec1ates th1s opportun1ty to conment on the proposed rule. We have been act1vely work1ng with the Nuclear Ut111ty Management and Resources Conm1ttee (NUMARC) and the Nuclear Ut111ty Group on Stat1on Blackout (NUGSBO). Southern Cal1fornia Ed1son supports and endorses the conments of these organ1zat1ons and the conments developed by the Atom1c Industr1al Forum concern1ng the proposed rule, the techn1cal bases for the rule (NUREG-1032) and the appropr1ateness of NRC Staff's backf1t analys1s (NUREG-1109). Spec1f1cally, we do not bel1eve that these NUREG documents provide an adequate technical or regulatory justificat1on for the proposed rule. Finally, Southern California Edison's specific conments on the proposed rule are enclosed. Very truly yours, Enclosure

Enclosure SOUTHERN CALIFORNIA EDISON S COMMENTS 1 ON PROPOSED STATION BLACKOUT RULEMAKING Southern Cal1forn1a Ed1son conments on the proposed stat1on blackout rule address the follow1ng: (1) SCE bel1eves that current regulat1ons are suff1c1ent and add1t1onal requ1rements are not needed, (2) the proposed rule 1s not the result of an 1ntegrated resolut1on of related gener1c 1ssues, (3) the proposed rule does not address the two key 1ssues of stat1on blackout and (4) SCE be11eves that NRC has not adequately taken 1nto account 1nformat1on prev1ously subm1tted by 11censees. Add1t1onally, our comments on three other areas of the proposed rule 1nclude (1) the rule 1 s fa11ure to address determ1n1st1c causes (1ssues). (2) the fact that "cop1ng" 1s not well def1ned and (3) the fact that the requ1rements are open- ended 1n nature. Based upon these conments, SCE recommends that the proposed rule be w1thdrawn .

• The Nuclear Regulatory Conm1ss1on states 1n the Federal Reg1ster. Volume 51, No. 55, that:

       "The ex1st1ng regulat1ons do not requ1re expl1citly that nuclear power plants be des1gned to assure that the core can be cooled and the 1ntegr1ty of the reactor coolant pressure boundary can be ma1ntained for any spec1fied per1od of loss of all AC power."

It is Southern California Ed1son's (SCE) pos1tion that such explic1t regulat1ons, beyond those currently 1n effect, are not requ1red. Contrary to the NRC s central prem1se as stated above, exp11c1t regulations address1ng 1 stat1on blackout are not requ1red based upon the ex1sting regulations and report1ng requ1rements currently in effect. Plant- spec1fic Techn1cal Spec1f1cat1ons already address both offs1te power ava11abil1ty and d1esel generator re11ab111ty, both of wh1ch are the key 1ssues 1n stat1on blackout. Add1t1onally, regulat1ons have been issued wh1ch address reporting requ1rements for power source 1noperab111ty. For example, the plant Techn1cal Spec1f1cat1ons conta1n conservat1ve requ1rements for ma1nta1n1ng the operab111ty of both offs1te and ons1te AC power sources. G1ven 1noperabil1ty of a power source or sources, act1on statements requir1ng restorat1on of operab111ty or plant shutdown are entered. Addit1onally, an array of surve11lance requ1rements 1nclud1ng preventive maintenance and testing is included 1n the Technical Spec1f1cat1ons to demonstrate and mainta1n power source operab111ty. F1nally, the plant Techn1cal Spec1f1cations and 10 CFR 50 Sect1ons 72 and 73 would require prompt NRC not1f1cat1on w1th wr1tten followup by any 11censee exper1enc1ng cond1t1ons wh1ch could lead to stat1on blackout. These report1ng requirements allow the NRC opportun1ty to rema1n aware of trends 1n d1esel generator and offs1te power unava1labil1ty. Such reporting requ1rements also 1nclude the requirement to address correct1ve act1on taken to preclude further occurrence. NRC currently has the author1ty to so rev1se plant 11cense cond1tions and Techn1cal Spec1f1cat1ons for any ut111ty as well as to enforce necessary correct1ve act1ons ident1f1ed subsequent to occurrence of such operat1onal events. SCE bel1eves that the ex1st1ng operab111ty and reporting requirements are suffic1ent to ensure that the publ1c health and

safety w111 be ma1nta1ned for losses of AC power sources which can be reasonably ant1c1pated to occur. Therefore, add1t1onal regulat1ons are unnecessary. Promulgat1on of a f1nal stat1on blackout rulemak1ng at this t1me will unnecessarily compl1cate the f1nal resolut1on of related generic techn1cal issues. On October 12, 1984, the Nuclear Utility Group on Stat1on Blackout (NUGSBO) forwarded to the NRC an Integration Plan for Station Blackout (USI A-44) and Power- Related Generic Issues. The plan proposed that the NRC integrate the process for resolv1ng stat1on blackout and the power- related aspects of other gener1c issues. NUGSBO bel1eved that power- related technical cons1derations and correspond1ng analyses and solut1ons, addressed 1n the context of several other gener1c 1ssues, could have a s1gn1f1cant 1mpact upon the disposit1on of USI A- 44. L1kew1se, the d1sposit1on of USI A- 44 may be an important cons1derat1on toward resolut1on of the power- related aspects of the other gener1c 1ssues. In the plan, NUGSBO 11sted the follow1ng 1ssues as be1ng related to station blackout: (1) shutdown decay heat removal requirements (USI A- 45), (2) d1esel generator reliab1lity (GI B-56), (3) reactor coolant pump seal failures (GI - 23), (4) component cool1ng water system fa1lures (GI - 65), (5) adequacy of safety- related DC power suppl1es (GI A- 3O) and (6) safety 1mpl1cat1ons of control systems (USI A-47). SCE believes that NRC has not properly 1ntegrated resolut1on of these generic 1ssues into the proposed stat1on blackout rule. For example, stat1on blackout cop1ng capab111ty, which is the sole focus of the proposed rule, 1s d1rectly dependent upon reactor coolant pump seal performance and the adequacy of DC power sources, yet the proposed regulatory gu1de does not address the resolut1on of these gener1c 1ssues. D1esel generator reliab111ty, which 1s a key 1ssue to the station blackout concern, is mistreated in the proposed station blackout rulemaking. The resolut1on of generic 1ssue B- 56 1s also not addressed 1n the proposed rulemaking. Finally, any perceived resolution of the stat1on blackout concern 1s 11kely to complicate resolution of shutdown decay heat removal requ1rements (A- 45). The proposed rulemaking only cursorily addresses adequacy and reliability of both AC-dependent and independent decay heat removal systems. The absence of a coordinated resolution of related technical issues could contribute to a less reliable or less safe final plant configurat1on. The NRC must develop and 1mplement a program to coord1nate the resolution of all power- related generic issues prior to finaliz1ng any ind1vidual proposed rule. The real issues involved in the stat1on blackout concern are offs1te power availability and diesel generator reliabil1ty. There is nothing within the proposed regulations wh1ch will enhance either offsite power or diesel generator reliability. Thus, any efforts expended by NRC and industry towards compliance with the new rule detracts resources which would be better directed toward improvements in these two key areas. For example, offsite power availability in the absence of regulations has significantly improved over the past decade. The proposed regulation not only was formulated without taking the results of these efforts into account, but further fails to encourage future improvements in any way. Indeed, the proposed rule is based on a perce1ved national average loss of offsite power. This approach ignores basic

facts 1n terms of the actual d1str1but1on of losses of offs1te power. for example, the character1st1cs of certa1n east- coast ut111ty systems or the effects of severe weather (such as snow or hurr1canes} are not appl1cable to San Onofre. Reference 1s made to EPRI 1 s report 11 Losses of Offs1te Power at U.S. Nuclear Power Plants 11 (May 1984} and NSAC- 103, Losses of Offs1te Power at U.S. Nuclear Power Plants, (June 1986). Add1t1onally, the proposed rule not only fa1ls to address methods for 1mprov1ng d1esel generator re11ab111ty, 1t also re11es upon a method for assess1ng ex1st1ng rel1ab111ty wh1ch 1s not appl1cable to the stat1on blackout 1ssue; 1.e., use of Regulatory Gu1de 1.108. Thus, the proposed regulat1on unnecessar11y doubly pena11zes ut111t1es. Reference is made to the d1scuss1on of stat1on blackout 1n NSAC -95, Gener1c Safety Issue Track1ng and Evaluat1on Sunvnary Descr1pt1on (May 1986). SCE reconvnends that the proposed rule be w1thdrawn based upon the fact that 1t will missallocate both NRC and 1ndustry resources away from future 1mprovements 1n these two v1tal areas .

• By promulgat1ng the proposed rule, the NRC has apparently overlooked or d1scounted prev1ously subm1tted plant- spec1f1c 1nformat1on wh1ch prov1des a basis for assess1ng ex1sting defense- 1n- depth (cop1ng} capab111t1es. Th1s 1nformat1on was prov1ded 1n response to USNRC Generic Letter 81 - 04, Emergency Procedures and Tra1n1ng for Stat1on Blackout Events. Th1s letter requ1red that 11censees rev1ew current plant operat1ons to determ1ne capab111ty to m1tigate a station blackout event and promptly 1mplement, as necessary, emergency procedures and a tra1n1ng program for stat1on blackout events. In response to th1s requ1rement, SCE evaluated the ab111ty of our San Onofre units to cope w1th a two hour stat1on blackout event. SCE bel1eves that our response and the response of the other 11censees to th1s requ1rement has not been appropr1ately cons1dered by NRC 1n formulat1on of the subject rule . In add1t1on, the regulatory backf1t analys1s also appears to 1gnore this work.

for example, any perce1ved benef1t of the proposed rule would have to resu l t from an extens1on of our demonstrated cop1ng capab111ty from two to four hours. Indeed, the NRC staff 1tself has concluded that the major1ty of plants w1ll not be s1gn1ficantly affected by the rulemaking. SCE believes that th1s was concluded on the bas1s of ut1lity responses to Gener1c Letter 81 - 04

• Thus, the proposed rule 1nadequately addresses the current status of the 1ndustry relative to genu1ne stat1on blackout r1sks. It 1s 1n th1s manner that the backf1t analys1s 1s 1ncomplete, thus lead1ng to potent1ally 1nappropr1ate conclus1ons regard1ng the usefulness of the proposed regulat1on.

The proposed stat1on blackout rulemak1ng 1s not rea11st1c 1n that 1t 1s based upon perce1ved consequences w1thout an 1nit1at1ng cause. The proposed regulat1on assumes loss of all AC power and complete unava1lab111ty of ons1te emergency AC power sources. Th1s could 1nclude s1multaneous loss of mult1ple offs1te power 11nes. It also assumes s1multaneous unava1lab111ty of at least two ons1te emergency AC sources and essent1ally 1gnores the 1ssue of other non- Class lE AC sources or other un1ts at mult1ple un1t s1tes. In our case, th1s would requ1re assuming s1multaneous loss of mult1ple 1ndependent offs1te power 11nes (powered from two separate util1ty networks), loss of three un1ts (reactor and turb1ne tr1p) coupled w1th loss of two or more d1esel generators

at each unit. The proposed rulemaking should be based upon the plant- spec1f1c probab111ty of such an event w1th progression all the way to radioactive mater1al release. The proposed rule thus formulates a solut1on to a problem with a probability of occurrence several orders of magnitude below the NRC Staff's published goals. Such an event should never be considered as a nuclear power plant des1gn basis. In SCE's case, the rule, as proposed, would in effect merely extend our demonstrated coping capab11ity from two to four or more hours under these cond1tions. This demonstrat1on of 1tself could not seriously reduce our blackout contribution to core melt risk. Th1s lack of realistic basis 1s further borne out 1n the confus1on which could result following attempted compliance with the new rule. For example, the proposed rule would require that the additional onsite AC power source which could be added to reduce the frequency of blackout would itself have to be assumed lost. Both the lack of realist1c bas1s for the rule and 1ts amb1guous and confused methods of app11cation are suffic1ent grounds to argue for w1thdrawal of the proposed regulat1on . The proposed stat1on blackout rule 1s focused s1ngly on the concept of cop1ng capab111ty. However, the regulat1ons are 1nadequate 1n the1r definit1on of cop1ng and the bases upon wh1ch that capab1lity 1s def1ned. The proposed rule 1s apparently directed towards achieving continued core cooling and ma1ntain1ng 1ntact the reactor coolant pressure boundary. Neither of these are def1ned plant condit1ons. In fact, the NRC has not def1ned acceptable plant cond1t1ons for cop1ng nor have the requirements for stat1on blackout recovery been addressed. For example, is core uncovery acceptable? If so, under what conditions? Would fuel damage be acceptable prov1ded reactor coolant system leakage did not exceed Technical Spec1f1cat1on limits? W111 the rule require maintaining hot standby or shutdown conditions specifically? Does successful cop1ng require that 11censees have the procedures 1n place to successfully reach cold shutdown condit1ons before (or after) AC power 1s restored? Thus, the concept of cop1ng as expressed 1n the regulat1on 1s not adequately defined. The open- endedness of the proposed rule 1s reflected 1n the unnecessary requirement that util1ties perform max1mum durat1on cop1ng stud1es beyond the1r establ1shed max1mum durat1on categories. It has been po1nted out by NUGSBO and others that these duration categories (4 or 8 hours) are themselves arbitrary numbers, without basis. SCE therefore reconmends that the proposed rule be withdrawn at least until the issue of coping is thoroughly evaluated and def1ned 1n the proposed regulatory gu1de. Additional examples of the open- endedness of criteria appl1cat1on include single failure requirements, equ1pment qualificat1on issues, alternat1ves to coping and safety related versus non - safety related modificat1ons with the1r attendant Technical Specification, procedures and training requirements. For example, the NRC's discussion of the unresolved safety issue which precedes the proposed rule change states, "The proposed rule does not require that a single failure be assumed concurrent w1th station blackout." However, Cr1terion 17, Electrical Power Systems, as it is proposed to be amended, adds a paragraph (e) which spec1fies, in part, that station batteries prov1de sufficient capacity and capab111ty to assure that the core 1s cooled and

conta1nment 1ntegr1ty 1s ma1nta1ned in the event of a stat1on blackout. Paragraph (b) of that cr1ter1on requires that ons1te electr1cal d1str1but1on systems 1nclud1ng batter1es, have suff1c1ent 1ndependence, redundancy and testab111ty to perform the1r safety funct1on assum1ng a s1ngle fa1lure. It would seem that although 1t 1s not 1ntended that the backf1t result1ng from the proposed rule requ1re that a s1ngle fa1lure be assumed, 1n fact, Cr1ter1on 17 as proposed would require s1ngle fa1lure assumpt1ons. Add1t1onal clar1ficat1on is requ1red s1mply to implement backf1ts 1n accordance w1th NRC's 1ntent. SCE therefore reconmends that the proposed rule be w1thdrawn at least unt11 these cr1ter1a are 1dent1f1ed and the1r resolut1on 1s addressed 1n the proposed regulatory gu1de. Premature promulgat1on of requ1rements such as th1s rule have a h1story of creating years of delay 1n f1nal 1mplementat1on and requ1r1ng unnecessary expend1ture of m1111ons of dollars. The net result of promulgat1ng the proposed rule 1n 1ts current form w1ll be 1ssuance of further gu1dance and add1t1onal clar1f1cat1ons in a manner s1m1lar to that which resulted in the wake of the 10 CFR 50 Appendix R rulemaking. The proposed regulat1on should therefore be w1thdrawn at least unt11 the concepts of coping and cr1teria for successful 1mplementation are thoroughly evaluated and clearly defined. More 1mportantly, SCE strongly reconmends complete withdrawal of the proposed rule because the rule itself 1s not required based upon ex1sting regulat1ons. It 1s h1ghly unl1kely the rule as proposed w111 achieve the stated goals. BRD:6885F

consumers Power Kenneth W Berry Director l'OWERINli Nuclear Licensing MICHlliAN"S l'ROliRESS General Offices: 1945 West Parnell Road, Jackson, Ml 49201 * (517) 788-1636 June 18, 1986 Samuel J Chilk Secretary of the Commission US Nuclear Regulatory Commission Washington, DC 20555 ATTN: Docketing and Service Branch RE: Consumers Power Company Comments on Proposed Rule Regarding Station Blackout, Task SI 501-4 Consumers Power Company appreciates the opportunity to comment on the proposed rule regarding station blackout (51CFR9829). After careful review by Consum-ers Power Company personnel, we have the following comments. Consumers Power Company fully endorses the comments submitted by the Nuclear i r Utility Management and Human Resources Committee (NUMARC) on the proposed rule. Furthermore, as a member of NUMARC, Consumers Power Company i s commit-ted to completing the NUMARC initiative on station blackout regardless of the outcome of rulemaking on this issue. Speci fically, the NUMARC init i ative calls for Consumers Power to evaluate our Plants using the criteria in NUREG-1109 to determine the duration of station blackout capability that currently exists after utilizing all power sources available. If a Plant falls into the eight-hour category, Consumers Power Company will take actions to reduce the Plant's contribution to the overall risk of station blackout. The NUMARC initiative also includes provisions for i mproving AC power reli-ability at our Plants. Kenneth W Berry Director, Nuclear Licensing KWB 86-31

Log# TXX-4858 File# 10185 TEXAS UTILITIES GENERATING COMPANY SKYWAY TOWER

• 400 NORTH OLIVE STREET, L . B . 81
• DALLAS, TEXAS 73201 June 18, 1986 WILLIAM G . COUNSIL E X ECUTIVE VICE PRESIDENT

                                                                @)

UC ,.. uwt-U..RPR *

                                       '~OPOM:D RULE_-          SQ )

Mr. Samuel J. Ch ilk (!"I F~ qf;l9 Secretary of the Commission U. S. Nuclear Regulatory Commission Washington, D.C. 20555 Attn: Docketing and Service Branch

SUBJECT:

COMMENTS ON PROPOSED RULE ON STATION BLACKOUT

Dear Mr. Chilk:

(51 FEDERAL REGISTER 9829-9835 (3/21/86)) On Friday, March 21, 1986, the Nuclear Regulatory Commission (NRC) published a notice inviting public comments on a proposed rule on "station blackout" resulting from the Commission's study of Unresolved Safety Issue A-44, "Station Blackout." 51 Federal Register 9829-9835. The proposed rule would add a definition of "station blackout" to 10CFR Part 50 Section 50.2, add a new Section 50.63 "Loss of all alternating current power", and add to 10CFR Part 50 Appendix A General Design Criterion 17 "Electric power systems" a new requirement (e) which will require light-water cooled nuclear power plants be capable of withstanding station blackout for a specified duration and maintaining reactor core cooling during that period .

• Texas Utilities Generating Company (TUGCo) endorses the comments made by the Nuclear Utility Management and Resources Committee (NUMARC) on the subject proposed rule and incorporates them herein by reference. TUGCo respectfully submits the attached supplemental comments on the proposed rule.

TUGCo advocates cooperation between the nuclear industry and the NRC to achieve resolution of the Commission's concerns on station blackout. The AC power reliability program delineated in NUMARC's submission on this proposed kknow dgcd by r IIt~ A DIVISION OF TEXAS UTILITIES ELECTRIC COMPANY

' U. s. NUCLEAR REGULATORY COMMIS ION DO "'J'1 & StRVICE BRANCH

rule describes a consistent industry approach for addressing station blackout, placing appropriate emphasis on preventive action (versus coping after the event occurs). As brought out in NUMARC's supporting comments, pursuit of a generic rulemaking would lead to a very costly backfit for a very rare and low risk concern. Very truly yours, W$~ W. G. crJsi} By: ~ J(/' Jt!

                                                                     ~*....-t,.,~,,

G. S. Keeley Manager, Nuclear WJH/arh

ATTACHMENT (page 1 of 3) TUGCo Supplemental Comments on Proposed Station Blackout Rulemaking I. A Generic Rulemaking is Inappropriate Since the Historic Number of Sites Experiencing a Loss of All Off-site Power is Small. The number of sites experiencing a loss of all off-site power event is small. The recently published (May 1986) NSAC-103, "Losses of Off-Site Power at U. S. Nuclear Power Plants - all years through 1985", notes (at iv) that for all years of U. S. nuclear experience through 1985, 38 of 65 sites have never had a loss of all off-site power and 51 of the 65 sites have never had a loss longer than 30 minutes. An earlier analysis, NUREG/CR-3992, had found that for 30 (of a total 52) sites which had experienced a loss-of-offsite-power event, approximately one-third of the events occurred at 4 sites. The regulatory analysis (NUREG-1109, at table 6), assumes that only 15 reactors need to increase diesel generator reliabilities and only 10 reactors need to increase their capability to cope. Since the number of plants which are of concern appears to be small, it would be more appropriate to proceed on a case-by-case basis, using means other than a generic rulemaking to accomplish the Commission's goal. I I. Factors Which Further Reduce the Possibility of a Loss-of-offsite-power Event at Comanche Peak Steam Electric Station (CPSES). Four 345KV transmission lines 1 and one 138KV transmission line connect CPSES with the Texas Utilities Electric Company (TUEC) electrical network "grid". These transmission lines exit the site property on three separate right-of-ways in different directions (east, south-east and north). This construction inherently minimizes the possibility of a localized weather phenomenon or other postulated catastrophe from removing all sources of off-site power from the plant. In fact, since the CPSES switchyard was energized by both the 138KV and 345KV transformers, no loss-of-offsite-power event (i.e., loss of all off-site AC power to the essential and nonessential switchgear buses) has occurred at CPSES .

• An additional capability not taken into consideration by the NRC staff in the rulemaking is the availability of nearby "black start" capable AC power sources. For example, the Public Utilities Commission of Texas has approved a TUGCo plan to install approximately 260MW at it's DeCor2ova Steam Electric Station (nearby CPSES) in the form of gas turbines.

Tentatively four gas turbines of 64 to 67MW each are planned, depending on the manufacturer selected. These turbines will have "black start" capability and will be available to provide AC electric power to the TUEC electrical network "grid" in 30 minutes under "black start" conditions. Any one of these turbines has the capacity to provide power for the safe shutdown loads for both CPSES nuclear plants and, due to the nature 1Two more 345KV transmission lines are planned: one to be completed prior to Unit 1 fuel load (Comanche Switch Transmission Line, on an additional right-of-way south) and one to be completed prior to Unit 2 fuel load (a second circuit on the Benbrook Transmission Line). 2Installation is currently scheduled for completion in 1990.

ATTACHMENT (page 2 of 3) of the TUEC electrical network "grid", would have multiple separate paths for providing AC electrical power to CPSES. Such capabilities significantly reduce the probability that CPSES could experience an extended loss-of-offsite-power event. I I I. The Probability of a Station Blackout Event is Small. The proposed rule defines "station blackout" as "the complete loss of alternating current (AC) electrical power to the essential and nonessential switchgear buses in a nuclear power plant (i.e., loss of the off-site electrical power system concurrent with turbine trip and unavailability of the on-site emergency AC power system)." 51 Federal Register 9832 (1986). For this condition to exist, not only must the off-site power be lost but both separate and redundant trains of backup emergency AC power must be lost. The probability for a station blackout to occur would therefore be the probability of a loss-of-offsite-power event times the probability of the Train A diesel generator failing to start and load times the p3obability of the Train B diesel generator

• failing to start and load.

As an illustration of the small probability associated with such an event, the probability of a four hour station blackout can be calculated based on industry-wide experience. For all years through 1985 the occurrence of a loss-of-offsite-power event longer than 4 hours was less than 0.01 losses per site year. NSAC-103 at 2-16. This will be conservatively taken as a probability of 0.01 losses per site year. An EPRI survey for the U. S. nuclear industry for the years 1983, 1984, and 1985 showed an industry averaije diesel generator "unreliability" for unplanned "demands" of 0.022. Therefore, a generic probability of a four-hour station blackout event wo~ 6d be less than: (0.01)(0.022)(0.022) = 4.84 x 10 per site year. (As discussed below, the probability of an off-site release occurring from a station blackout is even lower.) IV. The Risk to the Public from Station Blackout is Less Than That Used to Justify the Rulemaking. The occurrence of station blackout is not intrinsically a public health and safety risk. The full scenario of concern is a station blackout with a duration which goes beyond the capability of the plant to provide core cooling, resulting in core damage and release of radiation to containment coupled with a containment failure, allowing the release of radiation to the environment. Results from IDCOR {Technical Summary Report (November 1984)) indicate that the frequency of release for an early containment 3cPSES has two separate, redundant 100% design capacity diesel generators per plant and the essential switchgear buses automatically transfer to the energized source. 4If tests (i.e., planned starts) are included, the "unreliability" improves to 0.014. No data is available to assess the probability of recovering the diesel generator within four hours. It is not uncommon to take operator action to restore diesel generator operability within a short period of time. Such operator action would decrease the probability of a station blackout of this duration.

 **                                 ATTACHMENT (page 3 of 3) failure as postulated by NUREG-1032 are on the order of 10- 7 or less per reactor year and that the risk dominant sequence would be long-term overpressurization of containment long after core melt (e.g., one to three days). NSAC-103 (May 1986) states (at iv) that the median duration of all losses of off-site power for all U. S. nuclear power plants for all years through 1985 is less than 1/2 hour and that the longest duration has been 8 hours and 54 minutes. Both of these time frames are significantly less than those associated with the dominant risk sequence.

Overall risk to the public is measured by the aforementioned "accident" frequency multiplied by the consequences (i.e., the "source term"). Both industry and NRC sponsored studies have shown that consequences based on WASH-1400, which was used by the NRC staff in its analyses, are "high" or overestimated. A better estimate of the consequences can be made when NUREG-1150 becomes available later this year. However, preliminary estimates reported to the Commission in it's quarterly briefing on source term on March 26, 1986, showed the total core melt frequencies for both plants being analyzed as reduced from previous estimates (based on WASH-1400). Taken in total, it can be seen that actions taken in response to a generic rulemaking based on improved coping will only be a small reduction in an already small risk at significant cost to the nuclear industry and their ratepayers. V. The Rule and Associated Regulatory Guide Could Produce Substantial Costs to All Utilities in Addition to Those Addressed by the Backfit Analysis. TUGCo has a significant concern that the interpretations associated with this rule could lead to substantial costs above those addressed by the NRC staff in its Backfit Analysis. The proposed coping analysis is open-ended and lacks realistic guidelines. The NRC staff has indicated that the coping analyses should be realistic, yet a detailed thermal-hydraulic analysis was required in

• the NRC's consideration of the station blackout issue on the St. Lucie docket. Such an analysis is significantly more expensive than the

       $150,000 used by the Staff in NUREG-1190.

The draft Regulatory Guide on Station Blackout (at 7) provides that equipment used to cope with a station blackout be qualified in the environment associated with the event. If the rule is interpreted to require full compliance with 10CFR Section 50.49 regarding the demonstration of equipment environmental qualification or even the documentation required to demonstrate the ability to function in a "mild environment", when not previously required to do so, significant additional costs will be incurred to provide the necessary tests and analyses to demonstrate existing equipment operability or to replace equipment with items which have been properly documented in their manufacture. The significant costs associated with the qualification or purchase of equipment which is "safety grade" were not considered by the NRC staff in its Backfit Analysis.

P. 0. Box 724928 Atlanta, Ga. 30339 P. 0. Box 33189 Charlotte, N.C. 28242 NUCLEAR UTILIT IES MANAGEMENT AND HUMAN RESOURCES CO June 17, 1986 0i

• l)()Cl!'l'ED JUN 191986*

The Honorable Nunzio J. Palladino DOCDTINCII Chairman RV CE BRAN U.S. Nuclear Regulatory Commission lluGY-NRC Washington, DC 20555 NUew. N11u11>ra

Dear Chairman Palladino:

i~~

                                              *~ ~ /
                                                     ~" PR - 5 In August 1985, NUMARC establishedfTfc~nical su2ommittee to address some selected generic technical (hardware-related) issues and use the NUMARC process to bring those issues to closure. Station blackout was identified as the first technical issue to be addressed and the NUMARC Station Blackout Working Group was formed. The working group received both technical and administrative support from the Electric Power Research Institute (EPRI) and utilized the Nuclear Utility Group on Station Blackout (NUGSBO) as its technical consultant. The various aspects of the station blackout issue were evaluated and numerous interactions took place between the working group and the NRC staff regarding the station blackout

• issue, the proposed rule, and its regulatory analysis and associated backfit analysis.

On March 21, 1986, the Nuclear Regulatory Commission (NRC) published a proposed rule on station blackout. The proposed rule, intended to resolve the Unresolved Safety Issue A-44, defines the term "station blackout" and requires commercial nuclear power plants to withstand and recover from a station blackout of a specified duration acceptable to the NRC. The proposed rule also includes an amendment to General Design Criterion 17. Accompanying the proposed rule is the "Regulatory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout" (NUREG-1109) that summarizes the value of the proposed modifications and the anticipated cost for compliance. NUMARC's review of the station blackout issue and the proposed rule, with its regulatory and backfit analysis, indicates that many of the concerns related to station blackout can be alleviated through industry initiatives to reduce the individual sites' contribution to the overall risk of station blackout through the enhancement of the reliability of on-site and off-site AC power systems. We believe such initiatives are warranted since NUMARC agrees with the NRC staff that most of the anticipated benefits from the proposed rule would be derived from improvements at a limited number of plants. On June 10, 1986, the NUMARC Executive Group overwhelmingly endorsed industry initiatives to address the more important contributors to station blackout. These initiatives consist of the following:

1. Each utility will review their site(s ) against the criteria specified in NUREG-1109, and if the site(s) fal l into the category of an eight-hour site after utilizing all power sources available, the utility will take actions to reduce the site(s) contribution to the overall risk of station blackout to the extent possible. Non-hardware changes will be made within one year. Hardware changes will be made within a reasonable time thereafter.

I u s. NlJ(..!.E 1\I 1~:..(,U, *'\TO,, 1 C A'*'l'>*;i ON DO I YI,;._ b I -.':rl

The Honorable Nunzio J. Palladino June 17, 1986 Page Two

2. Each utility will implement procedures at each of its site(s) for:

a. coping with a station blackout event,

b. restoration of AC power following a station blackout event, and

c. preparing the plant for severe weather conditions, such as hurri -

canes and tornados to reduce the likelihood and consequences of a loss of off-site power and to reduce the overall risk of a station blackout event.

3. Each utility will, if applicable, reduce or eliminate cold fast-starts of emergency diesel generators for testing through changes to techni-cal specifications or other appropriate means.

4. Each utility will monitor emergency AC power unavailability utilizing data utilities provide to INPO on a regular basis.

NUMARC has endorsed the above industry initiatives in the belief that the station blackout issue is not generic and thus need not be resolved by rulemaking; rather station blackout is a concern at a limited number of plants and should be resolved in a manner other than rulemaking. We believe that the proposed rule would cause the unnecessary expenditure of resources at a majority of the plants in the United States even though they have been identified by the NRC as having an acceptable station blackout risk. In addition, the proposed rule focuses on coping with a station blackout event, whereas, we believe that improvement in AC power reliability is a more appropriate means to reduce the risk of station blackout. Detailed technical comments on the proposed rule, regulatory analysis, and backfit analysis are enclosed. We are ready to work with you and your staff to bring the issue of station blackout to a timely closure *

• - t Yours very truly, ler, Jr.

Committee JHM:las Enclosure cc/w: Commissioner Asselstine Commissioner Bernthal Commissioner Roberts Commissioner Zech Victor Stello, Jr. Secretary of the Commission

DOCIETED UN 1919Rn DOCDnNQ& VIGEIMNQI

                                        ~Y*lRC COMMENTS ON THE PROPOSED

• NRC RULE AND DRAFT REGULATORY GUIDE ON STATION BLACKOUT (51 Fed. ~eg. 9829 March 2 , 1986)

June 19, 1986

TABLE OF CONTENTS 0 Abstract a o Executive Summary I. The Station Blackout Issue Need Not Be Resolved By Generic Rulemaking ***..**......*.. 1 II. The Technical Record Does Not Support This Generic Rulemaking . . * . * . * * . . * * * . . * * . * . * * * . . 7 I I .A. Technical Reports Relied Upon in the Proposed Rule Do Not Lend Support . to the Proposed Rule.............. . . . . . . . . . . . . . . . . . . 9 II.A.l The Probability of a Station Blackout Is Not Clearly Established .*.......**.**...*.**..... 9 II.A.2 Station Blackout Consequences Are Overstated ...*. 13 II.A.3 NUREG-1032 Contains Errors and Omissions ......*.* 23 II.B Additional Matters Included in the Technical Record Do Not Support the Proposed Rule *......... 27

• II.B.l II.B.2 Clarifications of European Approach Are Necessary *.*......**.........*....*..*......
• 28 Clarifications of RCP Seal Integrity Are Necessary *.*.* -....**.*......*..*.*.*......... 36 II.B.3 Clarifications of the Significance of Hurricane Gloria Are Necessary .... . *.......*..... 38

f. II.C Responsei to the Additional Comments and Views of the Commission Must Be Considered ....... 42

III. The Proposed Rule Itself Should Be Reevaluated ..* 48 III.A The Ultimate Requirements of the Proposed Rule Are Indefinite and Depend Upon the Future and uncertain Exercise of Discretion ....................................... 48 I III.B The Proposed Rule Will Not Achieve a Consistent or Efficient Resolution of the Station Blackout Issues .*.....*...**......*...... 52 IV. The Proposed Rule Does Not Meet the Backfit Rule Standard .................................. . 56 IV.A Introduction ..................................... 56 IV.B Installation And Continuing Costs Associated With The Backfit Have Been underestimated *...*..* 58 IV.C Potential Impacts on Radiological Exposure of Facility Employees Should Be Further Addressed **..*.. ****.*...*.*....*.**.***.*....... 64 IV.D The Relationship to Proposed and Existing Regulatory Requirements Should Be Considered Further . ..*.**.**.*.....*.*.**..***.*.*..*...**** 6 7 IV.E Potential Impacts of Differences in Facility Type, Design or Age Should Be Considered Further .......................................... 68 IV.F The Reduction In Risk From Offsite Releases To The Public Has Been overestimated *..***..*.... 69

• IV.F.1 IV.F.2 Factors Reducing the Probability of Station Blackout Require Further Attention *.....*.... .* ... 70 Factors Reducing The Consequences of Offsite Releases Require Further Attention .*..**.*.*..... 74 IV. G Conclusion ... ~ .......... ,. ....................... 82 V. Objections to the Proposed Rule ****.*..***.*.***. 84

ABSTRACT On March 21, 1986 the Nuclear Regulatory Commission

• (NRC) published a proposed rule on station blackout. The proposed rule is in response to Unresolved Safety Issue (USI) A-44. The term "Station Blackout" is defined in the proposed rule which would require commercial nuclear power plants to withstand and recover from a station blackout for a specified duration acceptable to the Staff. The proposed rule includes an amendment to General Design Criterion (GDC) 17.

Accompanying the proposed rule is a backfit analysis which relies in large measure _on the regulatory analysis

• found in NUREG-1109. Reference is made in the proposed rule to a draft regulatory guide which was also published March 1986.

The nuclear utility industry has been actively engaged in following the resolution of this issue. The Nuclear Utility Group on Station Blackout (NUGSBO) was formed by a number of utilities in the spring of 1984 and has made various presentations to the NRC. In the fall of 1985, the Nuclear Utility Management and Resources Committee (NUMARC) formed a Working Group on Station Blackout and took the lead on utility industry efforts to resolve this issue. NUGSBO remained as technical consultant to NUMARC. A review of the rulemaking material leads to the conclusion that rulemaking is not warranted for the following reasons: (1 ) The issue is not generic and, thus, need not be resolved by a generic rulemaking. (2) The technical record does not support the proposed rule. (3) The proposed rule itself should be reevaluated. (4) The proposed rule does not meet the backfit rule standard. In the event the Commission decides to promulgate a rule, despite these reasons to the contrary, there are several factors which are of concern and warrant special consideration.

• EXECUTIVE

SUMMARY

On March 21, 1986 the Nuclear Regulatory Commission (NRC) published a proposed rule on station blackout. 51 Fed. Reg. 9829 e t ~ - Station blackout is defined by the NRC Staff in proposed 10 C.F.R. §50.2, Definitions, to mean: the complete loss of alternating current (AC) electric power to the essential and nonessential switchgear buses in a nuclear power plant (i.e., loss of the offsite electric power system concurrent with turbine trip and unavailability of the onsite emergency AC power system). The proposed rule is in response to Unresolved Safety Issue (US I ) A-44, which was designated as a USI in 1979, partly as a resul t of the findings of the Reactor Safety Study (WASH-1400, 1975) that station blackout could be an "important contributor to the total risk of nuclear power plant accidents." 51 Fed. Reg. at 9830, col. 1. A concern was also expressed "that the reliability of both the onsite and offsite emergency AC power systems might be less than originally anticipated." Id. ( See also SECY 85-163 at 1-2). J To address this concern, the proposed rule creates a new section, §50.63, which would require commer ci al nuclear power plants to withstand and recover from a sta ti on blackout for a specified duration i n accordance with Ge neral Design Criterion

E-2 (GDC) 17. The "specified duration" is defined only in that four factors set forth in GDC 17(e) should be considered. A draft regulatory guide published in March 1986 ("Draft Regulatory Guide, Station Blackout") could be used to determine an acceptable method of determining a specific duration(~, four hours, eight hours) for each licensed plant. Amendments to GDC 17 are proposed to make station blackout a design basis

• event, because it is felt that existing regulations do not require plants to be designed to assure core cooling and containment integrity for any specified period of loss of all AC power.

The proposed rule sets forth implementation schedules for determination of the specific durations mentioned above (270 days from the effective date of the rule) and the schedule for implementing necessary plant modifications (within two years from the NRC's notification of its findings regarding the acceptability of a licensee's specific station blackout duration, unless a longer schedule is justified and mutually agreed to by the NRC and a licensee). J Accompanying the proposed rule is a backfit analysis which relies in large measure on the Staff's regulatory analysis. 1 51 Fed. Reg. at 9833-35. Reference is made to a

   !I  "Regula tory Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout." NUREG-1109.

E-3

  . draft regulatory guide (temporarily identified by its task 2

number, SI 501-4) which was also published in March 1986. 51 Fed. Reg. 11494. Finally, the proposed rule makes reference to various supporting documents. 3 The nuclear utility industry has been actively engaged in seeking a resolution of the station blackout issue. In the

• spring of 1984, a number of utilities formed the Nuclear Utility Group on Station Bla~kout (NUGSBO). Through its member utilities, and with the help of technical and legal consultants, NUGSBO reviewed existing literature and made several submittals to the NRC. 4 Subsequently, on August 28,

      ~/ "Draft Regulatory Guide, Station Blackout," Task SI 501-4 (March 1986) .
      }/ i.e., NUREG,/CR-3226 (May 1983), NUREG/CR- 2989 (July 1983),

NUREG/CR-3992 (February 1985), and WASH-1400 (1975).

      !/ On May 8, 1985, NUGSBO submitted its "Proposal for Resolution of USI A-44 (Station Blackout)." The proposal contained a seven point AC reliability program, proposed policy statement and discussion of integration of related issues. On June 5, 1985, NUGSBO submitted a report entitled "Estimation of Site-Specific Station Blackout Core Damage Frequency Using NRC Staff Methodology." On July 17, 1985, NUGSBO submitted a letter to the NRC Staff further clarifying NUGSBO's proposal. This letter addressed those points raised by the Office of Nuclear Reactor Regulation (NRR) related to:     (1) utility support of the NUGSBO proposal and for the concept of integration ; (2) NUGSBO prepar edness to elaborate on the seven p0int program; (3)

• • the basis for existing blackout coping capability. Each of these documents is incorporated by reference . NUGSBO also made a brief presentation to ACRS on February 26, 1985.

NUGSBO stated that the trend in loss-o f -offsite-power (LOOP) data and diesel generator reliability together indicate stat ion blackout is not a generic issue. As a result of the (Footnote 4 Continued on Next Page

E-4 1985, the Nuclear Utility Management and Resources Committee (NUMARC) voted to establish a Technical Subcommittee to focus executive attention and resources upon hardware issues such as station blackout. A NUMARC Station Blackout Working Group was formed. NUGSBO was named technical consultant to the NUMARC Station Blackout Working Group .

• Based upon a review of the rulemaking material it is concluded that rulemaking is not warranted for the following reasons:

o First, the issue is not generic; rather, station blackout is a concern at a limited number of plants and, thus, should be resolved in a manner other than rulemaking. 0 Second, the technical record does not support the proposed rule: the risk of a station blackout (both probability and consequence) is overstated; the calculat ions of risk contain errors and omissions; and the reliance upon the European approach, reactor coolant pump seal degradation issue and severe weather events such as Hurricane Gloria is of concern. o Th i rd, the proposed rule itself should be reevaluated: the coping analysis requirement is not well-defined and, based on experience with other rules, has the very real potential for (Footnote 4 Continued from Previous Page) NUGSBO presentation, the ACRS indicated i n a letter to the Executive Director for Ope r ations. Mr . Dirc ks, dated March 12, 1 985, that " i f a bette r al t~rnati ~~ tha n rulemaking is advanced, we recommend that it be g i ven serious consideration."

E-5 confusion and misplaced efforts without commensurate benefits; the resort to a coping analysis diverts attention from the proper goal of onsite and offsite AC power reliability. 0 Fourth, the proposed rule does not meet the backfit rule requirements: the costs are underestimated and the benefits are overestimated. The basic positions encompassing the above reasons are

• highlighted below (parenthetical references correspond to sections in the comments):

The Station Blackout Issue Need Not Be Resolved by Generic Rulemaking (SI) The record supporting the rule reflects that the number of plants of concern is limited. The regulatory analysis (NUREG-1109), assumes that only 15 reactors need to increase diesel generator reliabilities and 10 reactors need to increase their capability to cope. If the analysis supporting the rule

• gave adequate credit for a number of site-specific risk reduction measures currently in place or available ( ~

 "blackstart" diesels) even fewer plants would be of concern in this rulemaking.

An analysis of prior technical findings (NUREG/CR-3992) shows that approximately 40% of the 52 si tes considered in the regulatory ~nalysis never experiencect a loss-of-offsite-power event. Further, approximately 1/3 of the events occurred at 4 of the remaining 30 sites which had experienced an event.

E-6 Since the number of plants which are of concern appears to be small, it would be more appropriate to proceed on a case-by-case basis -- focusing only on those plants at which some action may be appropriate to reduce risk. The Commission, in the exercise of its discretion should use means other than rulemaking to accomplish this. The Technical Record Does Not Support

• This Generic Rulemaking (SII)

A review of the technical reports relied upon in the - proposed rule and the draft regulatory guide accompanying the proposed rule shows that the proposed rule does not have the necessary supporting technical record. First, the magnitude of the probability of occurrence of a station blackout at any site or group of sites has not been established. The principal technical report, NUREG-1032, does

• not identify the actual frequency distribution for station blackout. Instead, a set of "bins" is cr*eated in which each bin is assigned a core damage frequency defined by plant features considPred important for blackout. Thus, it is conceivable that the distribution of plants could concentrate at any point on the spectrum -- and, in fact, could concentrate at a point on the spectrum where the overall risk is very low.

However, no definite statement is made concerning how many plants, or which plants, are in each of the bins.

E-7 Second, the technical analysis of the consequences of a station blackout event makes certain erroneous assumptions regarding the absence of containment integrity and the amount of fission product released. The assumption that core damage is synonymous with near-term breach of containment for station blackout acc i dent sequences is inconsistent with other technical analyses. Further, the use of the Siting Source Term

    * (SSTl) assumptions (i.e., direct containment breach and early release of fission products caused by over-pressurization in only 1.5 hours following onset of core damage) to support this proposition       in lieu of SST2-5 source term assumptions (describing long-term containment failure and late release) is inappr opri ate. This is because the frequency of a large I -    early re lease such as SSTl for station blackout sequences is so small (on the orde r of 10- 7 or less per reactor year) that such release is no t ris k-significant compared to a late release.            As a result of the SSTl source term assumption, the offsite release consequences may be overestimated by two or more orders of magnit ude . Th i s point is underscored by recent analyses which ques ti on the short time intervals assumed in the technica l ana l ys i s to precede conta i nment failure and offsite release. Longer t i me intervals wou l d not only prov i de utilities wi th additional opportun iti es to pr event releases but, at a min imum, would ma ke the SST2 - 5 relea s e c a te go ri es more app ropri a te than SST l for a s tation ~lack ou t. The f i ss i on

E-8 product releases associated with these accident categories are well-below SSTl values. Third, the methodologies used in the technical analysis contain errors and the plant-specific interpretation of data in the scientific literature is inconsistent. These errors and inconsistencies affect the validity of the postulated

• categories of loss-of-offsite-power events (LOOP), which are distinguished by offsite power design characteristics and emergency AC power system reliability. There is an error in the hypothesis that a relationship exists between offsite power system features and the potential for a loss-of-offsite power.

This can be attributed to improper assignment of durations to certain events, double counting events and adding in several non-LOOP events to the data. Further, upon reproducing the analysis it appears that there is an important statistical error in failing to account for the large sum-of-squares error associated with the results and, thus, the analysis cannot draw correlations with the normal statistical confidence levels. Also, the analysis of the correlation between grid stability and weather does not entail a rigorous statistical review of plant experience. A key technical paper cited in the analysis was improperly interpreted and, if used cor r ectly, concludes that there is no correlation between the duration of loss-of-offsite-power events and precipita tio n amount s.

E-9 Fourth, in presentations before the Commission and in

    !esponse to written questions, several additional matters were raised in support of the rule. These matters included the European approach to station blackout, integrity of reactor coolant pump seals and the Hurricane Gloria experience.

The European approach is purported to be that of coping

• for a longer duration than would be required for U.S. plants in the proposed rule.

proposition. Howe ver, no information or technical analysis has been put into the record as a basis for this Thus, detailed comments cannot be provided on this proposition. It appears that central to the European approach is the capability to provide sufficient water inventories for makeup purposes. Under this logic, a large number of U.S. plants could cope for the long durations attributed to the French design. It is questionable as to whether any of the European plants could meet the requirements of the proposed rule and the technical guidance in the regulatory guide, particularly the equipment qualification aspects. Failure to demonstrate that French designs could comport with the terms of the proposed rule calls into question J' any reliance upon the European approach as a benchmark for U.S. plants.

E-10 The proposed rule also links the ability of a plant to cope with a station blackout to the issue of reactor coolant pump (RCP) seal integrity. With respect to station blackout, however, only leakages of 100 gpm per pump or more are of concern because of their impact on limiting decay heat removal. This criteria eliminates from consideration all boiling water reactors and plants operating pumps with hydrodynamic seals .

• Thus, reactor coolant pump seal integrity is only relevant to the Westinghouse pumps with hydrostatic seals (due to the larger seal leak potential). The concern over the Westinghouse pumps is approaching resolution and may be resolved as a result of a May 8, 1986 submittal of a revised Westinghouse report (WCAP-10541, Rev. 1). Thus, it does not appear that high pressure AC-independent injection pumps are needed to provide seal' injection and cooling. The impact of this discussion highlights the fact that concerns over RCP seal integrity do not contribute to* the overall station blackout concern .

Hurricane Gloria, the September 1985 Hurricane which moved up the East Coast from Florida to New England, has been cited *as an example of potential weather-related causes of loss of offsite power which supports rulemaking. However, the actions by utilities in anticipation of the hurricane to place plants in sa f e condition during the event demonstrates that effective measures could reduce th ~ seve re we ather concern.

E-11 The Proposed Rule Itself Should Be Reevaluated (§III) The ultimate requirements of the proposed rule are indefinite and depend upon the future and uncertain exercise of discretion. Neither the codified rule nor the draft regulatory guide establishes bounds for the required coping demonstrations. Under the rule as proposed, licensees cannot

• ascertain the ultimate requirements they will be expected to meet or the potential plant modifications they will need to make to satisfy the Staff of their compliance.

The proposed rule will not achieve a consistent or efficien t resolution of the station blackout issue. The rule does not focus on reducing the likelihood of a station blackout but, instead, focuses on coping duration. The rule allows for a range of coping from zero to eight hours. A more practical approach to meeting this objective of reducing the likelihood of station blackout consequ~nces is to focus on an AC power reliability program. The Proposed Rule Does Not Meet the Backfit Rule Standard (§IV) 10 C.F.R. §50.109, the Commission's backfitting standard, provides that a proposed _backfit must result in . . . a substantial increase in overall prot ecti0~ 0f the public health and safety . . . and that the . . . costs of implementation for

E-12 that facility are justified in view of this increased protection . . . " 10 C.F.R. §50.109(c) lists the nine (9) specific factors which are to be addressed as appropriate. The backfit analysis provided in the record to support the proposed rule falls short of demonstrating the justification for this proposed rule for the following reasons .

• First, the direct and indirect costs of the proposed backfits are underestimated. The number of plants that would be significantly affected by the proposed rule is understated in the regulatory analysis. This alone would significantly increase the costs. Further, the costs associated with activities at individual plants are based upon standard commercial labor p~oductivities rather than information available for the nuclear industry. Handbooks for cost estimating developed under NRC auspices were not used. Also, the projected costs for a coping analysis, based on both actual figures incurred by utilities and reasonable utility estimates, are low by a factor of 5 to 20. This underestimation is consistent with Staff projections in other rulemakings .

J Second, the record does not appear to c onsider the potential impact of differences in facilit y t ype, design or age, desp i te the site-specific nature o f t he i ssue.

E-13 Third, the potential impacts on radiological exposure of facility employees should be reassessed. The regulatory analysis asserts that no significant increase in occupational exposure is expected. This appears to be drawn from the belief that there will be no worker exposure if the backfit does not specifically involve the "reactor coolant system." This is not the case. The types of hardware fixes that could be

• contemplated include replacement of valve operators, new instrumentation and surveillance. A recent study indicates that up to forty percent of the total occupational exposure at light water reactors was attributable to NRC initi ated multi-plant actions.

Fourth, the regulatory analysis on this record should further consider the relationship between proposed and existing regulatory requirements. This consideration _is mandated by 10 C.F.R. §109. Although the regulatory analysis acknowledges the relationship between station blackout and the two issues of diesel generator reliability and reactor coolant pump seal integrity, it assumes resolution of these items would have little impact. However, to the extent there is improvement in the diesel generator reliability, the benefits claimed from the proposed rule are dissipated. Also, and despite the fact the regulatory analysis assumes small lBakage, the RCP seal issue has been a motivating force in the p r opo s~d s tation blackout

E-14 rulemaking. Recent information makes it apparent that the risk of catastrophic seal failure is significantly less than thought. Thus, resolution of this issue should de-emphasize the need for a rule. Fifth, the reduction in risk from offsite releases to the public has been overestimated. This overestimation is

• attributable to a failure to account for flexibility in emergency AC power system designs, emergency procedures and additional power sources when determining the probability of occurrence of a station blackout event. Further, use of the SST! fission product release assumption rather than SST2-5 release assumptions is inappropriate. The regulatory analysis makes no spe c ific statement as to time of containment failure except to assume the SST! scenario of 1.5 hours after onset of core damage. However, the regulatory analysis contains a recognition that this assumption is erroneous since SST!

releases are reduced by 1/3. The time to containment failure, if any, is a significant parameter in calculating doses to the public. Because containment integrity is not properly credited, the dose to the public has been overestimated. Also, evolving source term information is ignored which calls into question the ussumptions contained in the r~gulatory analysis.

E-15 Objections to the Proposed Rule (§V) The effectiveness of the rulemaking tool in addressing ~o isolated an issue as station blackout is questioned. However, it is recognized that the Commission may believe that rulemaking satisfies other imperatives which transcend the ability to demonstrate public health and safety benefits across all of industry. Should such imperatives exist within the

• Commission, then it is believed that the proposed rule should focus on the concerns raised. Accordingly, instead of the proposed rule with its focus on coping, guidelines should be prepared concerning what constitutes acceptable emergency and non-emergency AC power systems for preventing or mitigating station blackout events. Without such guidance, a rule which requires AC power independent coping is subjective, difficult to implement and of questionable value with respect to providing real safety margins .
• A shift in focus from compensatory measures to problem resolution would entail modifying several aspects of the proposed rule. To this end, the kinds of changes suggested J

include: (1 ) expanding the focus of the rule and draft regulatory guide to include e x plicit and equally weighted provision for crediting onsite and offsite bac ku p AC power sources o r improvements to AC powe r s y ste m availability in responding to the loss of normal and emergency AC power;

E-16 (2) deleting the addition to the General Design Criterion of an AC-independent coping requirement and the associated coping analyses; (3) eliminating the open-ended coping demonstration aspects of the regulatory guide and substituting a simpler closed-form coping capability checklist; (4) considering the potential ripple effects of new station blackout requirements, particularly those involving AC-independent coping, on other areas of nuclear

 * (5) regulation, such as potential equipment qualification of station blackout coping equipment by the methods called for in 10 C.F.R. §50.49; and eliminating the need to consider severe weather as a dominant factor in establishing whether a plant is in a 4 or 8 hour coping duration category (NUREG-1109, Tables 1 & 3) .

J

I* The Station Blackout Issue Need Not Be Resolved By Generic Rulemaking The Commission need not pursue generic rulemaking in order to resolve a non-generic issue. In the proposed station blackout rule, the number of plants of concern is acknowledged to be limited. First, in preliminary screening to determine the existence of "any plants with especially high risk that might require further analysis or action on an urgent basis",

• none were identified.!/ Second, the regulatory analysis (NUREG-1109, January 1986) states (at 9):

It was also assumed that all plants, as currently designed, can cope with a station blackout for 2 hours, and, with proper procedures and training, plants could cope with a 4-hour station blackout without having to make major modifications. (Emphasis added.) (See also 11.)~/ Third, concerns over station blackout risks are focused on the 14 reactors listed in Group C in Table 4 of NUREG-1109. This is evident from a review of Table 2 which defines Group C plants, and Table 5, which lists the estimated core damage frequency per reactor-year of these plants as 15 x 10 -5 . The next group of plants in Table 5 receive only 10% of the benefits attributed to implementation of the proposed rule at Group C plants. Further, the risks associated with all other

 ! / NUREG-1032 (May 1985) at 2-3.
 ~/ The initiatives advanced by the industry will provide for station blackout coping procedures.

plants set forth in Table 5 are at or near the goal sought by the proposed rule. Fourth, in Table 6 of NUREG-1109, only 15 reactors are identified as needing to improve diesel generator reliability and only 10 as needing to increase their capability to cope.]/ In the proposed rule, the highly plant-specific nature of station blackout is further acknowledged in connection with the

• ability of power plants to cope. In listing several "factors" as the "main contributors to risk of core melt resulting from station blackout," the Notice of Proposed Rulemaking states (51 Fed. Reg. at 9830, col. 3 - 9831, col. 1):

These factors . . vary significantly from plant to plant because of considerable differences in design of plant electric power systems as well as site specific considerations. Other points support a position that station blackout is not a generic issue which should be resolved by rulemaking . For example, a major contributor to the overall risk of station blackout is rliesel generator reliability. Station blackout ceases to become a generic issue if diesel generator reliability is high at many or most plants. Such is the case, in that the individual diesel generator reliability for most plants currently exceeds the minimum levels called for in this proposed rulemaking. See 51 Fed. Re_g. at 9830, col. 3 and

 ~/ The initiatives advanced by industry include a program for increased diesel generator reliability.

NUREG-1109 at 6-7 and at Tables 1 and 6. Therefore, as to the majority of plants, no action need be taken under the proposed rule to assure greater diesel generator reliability. Similarly, the contribution of loss-of-offsite-power to station blackout risk is not spread evenly among the operating sites. A review of NUREG/CR-3992 (1985) shows that approximately 40% of the 52 sites representing 67 reactors le (i.e., 22 sites) considered in the regulatory analysis (NUREG-1109) never experienced a loss-of-offsite-power event. Moreover, approximately 1/3 of the loss-of-offsite-power events occurred at 4 of the 30 remaining sites. See~-, NUREG/CR-3992 at §5 and Appendix. Therefore, rather than being roughly the same at ull plants, or even a majority of plants, the contribution of loss-of-offsite-power potential to station blackout risk is concentrated at only a handful of sites and is not significant elsewhere.4/

• Beyond this, the proposed rule does not reflect adequate consideration for a number of site-specific factors which reduce the probability of a station blackout event. For example, no credit is given for the availability of additional 4/ Recent information developed by the Nucl~cr Safety Analysis Center indicates that, for all years through 1985, fewer losses-of-offsite-power may have occurred than is indicated in NUREG/CR-3992. NSAC-103, "Losses 0-f: Off-Site Power at U.S. Nuclear Power Plants," Apri_J_ 1985 - at 2-13 ( indicating that 51 of 65 sites have not had a total loss-of-offsite-power longer than 30 minutes; 38 of 65 sites have never had a loss-of-offsite-power).

sources of AC power.~/ The proposed rule also does not credit the ability of some utilities to reconfigure their diesel generators and to shed non-essential loads so that they are effectively in a different category of plant diesel generator configuration.~/ Finally, the ability of utilities to manually

 ~/ As an example, the Indian Point site has gas turbines available in the unlikely event that all offsite and onsite emergency power is lost. The H.B. Robinson plant can rely upon its fire protection diesel to provide emergency AC

• power. In the case of Florida Power and Light Company's Turkey Point plant, by design the units are configured so that there are a total of two safety related emergency diesel generators supported by five non-safety related black-start diesels of similar capacity. The Turkey Point plants (Units #3 & #4) are located on a two-unit nuclear site adjacent to two conventional fossil {oil and/or gas) fired plants (Units #1 & #2). Each unit requires one out of two diesel generators in order to meet all essential loads during a design basis LOCA event. Also by original plant design, both diesel generators are aligned such that either unit's safety related emergency buses can be powered by either diesel generator from their respective trains.

Additionally, any two of the five black-start diesels can be used to power each unit's emergency buses {via dedicated cable, independent of station switchyard) during a blackout condition within 30 minutes of the initiating event. The Turkey Point site, by original design, also has the capability to provide both steam and electric power (via the switchyard) from the fossil-fired units to the nuclear units. This power can be used to power safety related equipment.

 ~/ Each diesel at Dresden is capable of either parallel or independent operation. Likewise, the emergency 4 kV bus system is designed to operate either sectionalized or as a ring bus between both units. Procedures currently exist to reduce the AC loads to only those absolutely necessary for safe shutdown of the plant. In the event of loss-of-offsite-power, with the use of the isolation condenser, condensate storage, service water, and control-rod-drive hydraulic systems, both units can achieve and sustain hot shutdown conditions with one diesel generator. Quad Cities has a similar capability using the re0rtor core isolation cooling (RCIC) and RHR systems in lieu of an isolation condenset. As a Lesult, the DresJen and Quad Cities plants, which are characterized as a 2 out of 3 diesel generator configuration, would actually serve during a station blackout event as a 1 out of 3 rliesel generator configuration.

start diesel generators which fail to start automatically is not fully taken into account.2/ Since the number of plants of concern is limited, the issue is not generic and rulemaking should not result.~/ It is recognized that a basic premise of administrative law is that agencies such as the NRC may impose new requirements either by promulgating generic rules or on an individual plant basis . DAVIS, ADMINISTRATIVE LAW TEXT chs. 6 and 8 (1972). Further, it is well-recognized "[that] the choice between rulemaking and adjudication lies in the first instance with the [agency's] discretion." NLRB v. Bell Aerospace Co., 416 U.S. 267, 294 (1974). However, the line dividing the choices may not always be a "bright one. united States v. Florida East Coast Railway, 410 U.S. 224, 245 (1973).9/ In this instance, the 21 Procedures exist at a number of plants which permit manual diesel starts in a matter of minutes, should these sources fail to start automatically .

 ~/ It would be more efficient to proceed on a case-by-case basis focusing only on those relatively few plants where actions are judged to be necessary to minimize the risks posed by station blackout.

21 For example, one court recognized that: This simple proposition, however, is incapable of being reduced to mathematical terms. One cannot say, for example, that an issue which affects 15~ of the nuclear plants in the country should be resolved generically, while one which affects only 10% of the nation's plants is inappropriate for generic resolution. Deukmejian~v~ NRC, 751 f.2d 1287 (D.C. Cir. 1984).

Commission should exercise its discretion and refrain from promulgating the proposed station blackout rule. The Commission recognized in its 1986 Policy and Planning Guidance (PPG) that: NRC must be sensitive to the large number of requirements imposed on licensees. Requirements imposed on the regulated industry by NRC are to provide a positive contribution to the public health and safety. . There

• should be no unnecessary regulatory burdens.!.Q_/

This guidance is important in the context of station blackout. First, new requirements should not be imposed on licensees unless they are necessary. Second, orders to individual plants are a viable means of addressing issues of concern at those few plants. In short, the Policy and Planning Guidance makes clear that when the focus of a perceived issue is confined to only a few plants, the resources of the NRC should be selectively applied .

• In sum, a rule is not required to resolve issues at those few plants where station blackout is of concern. The need for any generic rule which would change a General Design Criterion to address highly site~specific issues is questionable.

10/ PPG at 23.

I I. The Technical Record Does Not Support This Generic Rulemaking The proposed rule is "intended to provide further assurance that a station blackout . will not adversely affect the public heal th and safety. "l.!/ The rule would provide this assurance by requiring that plants cope with a station blackout for some period of time. This duration would not be specified in the rule itself. Rather, the selection of

     * ..!,!/ 51 Fed. Reg. at 9829, col. 1. The Notice of Proposed Rulemaking contains the assertion that in a "few cases" there have been complete losses of both onsite and offsite AC power sys terns, and, in these cases, "AC power was restored in a short time without any serious consequences."

Id. at 9830, col. 1 (emphasis supplied). While station blackouts have occurred at several plants, they either have involved momentary loss of all AC power or have occurred while the plant was not engaged in normal operations. There have apparently been four events involving brief simultaneous losses of offsite and onsite AC power. In 1968, there was a loss of offsite power event at Connecticut Yankee. The diesel generators started and loaded. Due to a switching error an erroneous signal caused a stuck breaker protection scheme," tripping the diesel generators. There was no AC power for a period of 4 minutes. In 1976, Millstone 2 suffered a loss of offsite power. The diesel generators started, but when a large pump was being started, an undervoltage trip occurred. AC power was lost for 5 minutes. In 1983, Fort St. Vrain experienced a loss of offsite power while in cold shutdown. One diesel generator was out of service. The other diesel had been started and loaded, but tripped due to overloading when the loss of offsite power occurred. AC power was restored after 25 minutes. Finally, in 1984, during a test simulating a loss of offsite power event at Susquehanna,

 ,..          Unit 1, there was a switching error by an operator, resulting in the diesel generators failing to start automatically and having to be started ~2nually.        It took 11 minutes to restore AC power.

As all a£ these incidents indic ~t e . there has been no event rluplirative o f a sustain ~d stati on blackout, the concern of the proposed rule. Rather, there have been events which resulted from testing and switching errors under circumstances for which there is little concern about sustained losses of AC power. l

                               -   8 -

a 4-hour or 8-hour duration would be the result of application of methods provided in a regulatory guide. 50 Fed. Reg. at 9830, col. 3. In addition to this requirement, improved guidance will be provided to licensees regarding maintaining minimum emergency diesel generator reliability to minimize the probability of losing all AC power. Id. The technical bases for the proposed requirement are the

• result of information developed by the Staff and by NRC contractors in the course of their studies of the station blackout issue, including oral statements of the Staff made at Commission meetings and Staff responses to specific Commissioner's questions. A review of this material, and the below provided responses to the additional comments and views of the Commission accompanying the proposed rule, shows that the technical record does not support the rule. Technical comments fall into three categories:
• 0 o

Technical Reports Relied Upon In The Proposed Rule Do Not Lend Support to the Proposed Rule (§II.A below);.!l/ Additional Matters Included in the Technical Record Do Not Support the Proposed Rule (§II.B below); and o Responses to the Additional Comments and Views of the Commission Must Be Considered (§II.C below). 12/ The Supplementary information ~Joi:tion of the Notice of Proposed Rulemaking highlights the major results of the Staff's technical studies on the station blackout issue. 51 Fed. Reg. at 9830, cols. 2 and 3* ---- -

II .A. Technical Reports Relied Upon In The Proposed Rule Do Not Lend Support to the Proposed Rule A review of the technical reports relied upon in the proposed rule (i.e., WASH-1400 (1975), NUREG-1032, NUREG/CR-3226 (1983), NUREG/CR-2989 (1983) and NUREG/CR-3992 (1985)), and the draft regulatory guide accompanying the proposed rule (March 1986),ll/ reveals that station blackout risks have been overestimated. This position is based upon the

• following conclusions regarding the technical bases for the proposed rule:

(1) The Probability of A Station Blackout Is Not Clearly Established (§II.A.1 below); (2) Station Blackout Consequences Are Overstated (§II.A.2 below); and (3) NUREG-1032 Contains Errors and Omissions ( §II . A. 3 be 1 ow) . II.A.1. The Probability of a Station Blackout Is Not Clearly Established

• The principal document in support of the proposed rule is NUREG-1032. It identifies the various factors and plant features which affect the estimated frequency of core damage resulting from station blackout events. The principal result from NUREG-1032 upon which the proposed rule is based is that the potential variability of estimated station blackout likelihood and core damage frequency is large, depending on the precise combination of features which mav ~xist at a given site.

    -------**----**-----*-~----*-

lll The proposed Lule provides the title of each of these documents. 51 Fed. RE:..9_. at 9830, cols. 1 and 2.

However, NUREG-1032 does not identify the actual frequency distribution for station blackout. Instead, the analysis presented in NUREG-1032 creates what can be characterized for discussion purposes as a set of "bins." Each bin is assigned a core damage frequency defined by the factors and plant features considered important to station blackout ( ~ , emergency diesel. generator reliability and offsite power design characteristics). Bins at the high end of the core

• damage frequency spectrum are characterized by the worst possible set of plant features (which maximizes frequency).

Similarly, ~he low end of the spectrum is defined by the best combination of factors (which minimizes frequency). Thus, the core damage frequency range reported in NUREG-1032 actually constitutes the universe of potential values which could exist. The problem with this approach is that the analysis is never completed. That is, after having conceptualized the factors important to the probability of a station blackout and creating the core damage frequency spectrum, NUREG-1032 does not identify how many plants, or which plants, are in each of the bins. It is conceivable that the distribution of plants among the various bins could concentrate at any point on the spectrum. Thus, overall risk could be very high, very low, or somewhere in between. However, without making that 0ssignment based on plant features, the current probability of occurrence of a station

                            -  11 -

blackout posed by U.S. nuclear plants is simply not established beyond a mere generalization of what it could be.1:_i/ Thus, the premise that this generic rulemaking is necessary to reduce a current unacceptable risk to the public health and safety is not adequately supported by the analysis in NUREG-1032. Put another way, a specific resolution for an issue has been proposed before determining whether the issue really

• exists. This is unlike equipment qualification, which arose in response to pre-1980 test results at Sandia National Laboratories. This is unlike Appendix R, which was premised upon another set of Sandia tests, and subsequent Appendix A (i.e., Branch Technical Position 9.5-1) fire reviews at the majority of plants. This is unlike emergency planning, which arose in response to the experience of TMI. No information is presented to cast doubt on the size of the overall risk associated with nuclear plant operation or that such risk is now unacceptable .

 -~/ The information necessary to assess the probability includes data on emergency AC power system configuration, emergency diesel generator reliability, switchyard design features, weather experience, utility procedures currently in place for responding to a loss-of-offsite-power and station blackout event, and the ability of various plant systems to operate without any AC power fonsite and offsite). This information is currently available for estimating plant-specific station bl2c 1t out risks. Because this information was not processed in NUREG-1032, the risk of station blackout cannot be established on an overall basis or with respect to any specific plant.

Absent specific information, the proposed rule appears simply to be a response to the early WASH-1400 study and other PRA's, which show station blackout to be an "important" contributor to risk -- a risk that is "small" (51 Fed. Reg. at 9830, col. 1) and which is already acceptable.15/ Emergency diesel generator reliabilities also may have been a contributing factor. 51 Fed. Reg. at 9830, col. 3. However, Electric Power Research Institute (EPRI) studies show the

• industry average diesel reliability to be approximately 0.975, which exceeds the 0.95 goal in the regulatory guide accompanying the rule (see draft regulatory guide at §C.1.2).

Further, the proposed rule raises the specter of growing grid instability in the future. However, no evidence is presented in support of such speculation. To the contrary, since General Design Criteria 17 and 18 went into effect in the early 1970s, the overall site frequency of loss-of-offsite power in any calendar year has declined significantly. NUREG-1032 acknowledges (at A-12) that such improvements are evident in the "plant-centered" category after 1978. Further, the Staff stated at the November 14, 1985 Commission meeting (Transcript at 19) that: NUREG-1032 estimates that the frequency of station blackout events in the future will be better than the past experience has shown . . . [W]e recognize that there have been some trends sh owing impro v em~nts in loss of offsite power experien c e. that we don't

 ~5/ See discussion below (§II.A.2) of WASH-1400 results in contrast to other more recent Staff technical information.

expect as many losses as have occurred in the past. This statement confirms the conclusion which found that, taking the recent loss-of-offsite power data, such events were decreasing (even as the number of plants has increased) due to the knowledge and experience gained by utilities since the inception of commercial nuclear power operation.16/

• In sum, when taken as a whole, the above facts suggest that the motivation for proceeding toward a station blackout final rule cannot be linked to evidence of degrading plant experience over the years._!2/ Thus, since it has not been clearly established that the probability of station blackout events is unacceptably high, the Commission should not promulgate the proposed rule.

II.A.2 Station Blackout Consequences Are Overstated 18/

• 16 / See NUGSB-85-005, "NUGSBO Comments on NUREG-1032," October 1985 and Revision 1, November 1985 at Appendix B, which is incorporated by reference into these comments.

17/ The initiatives advanced by the industry will provide reasonable assurance that AC power availability will not degrade in the future. The comments on the proposed rule contatned in this section also form the basis for various positions regarding the published backfit analysis (§IV). In §I V below, the comments focus on the o v er es timeterl h n e f i t o r r i sk-0 reduction v alu e of the prop o s e d rul e ~s ex pressed in terms of averted person-rems. In this se c tion, however, the comments focus on the underlying technical basis for the conclusion that station blackout risks are substantial enough to be addressed via the current proposed generic rulemaki.ng.

The technical findings in support of the rule tie the consequences of a station blackout event to certain assumptions regarding the absence of containment integrity and fission product releases, (i.e., core damage is synonymous with near-term breach of containment as discussed below). Namely, the assumption is made that one of the so-called "Siting Source Term fission product release categories, the

• SST! category, is appropriate for purposes of modeling the generic characteristics of the consequences of all station blackout events for all types of containments.19/

1109 at 8. See NUREG-In contrast to the regulatory assumptions for "a maximum credible accident" postulated for siting studies under 10 C.F.R. §100 -- in which containment integrity is assumed to exist throughout the accident (see NUREG-0771 at 7), the SST! release category assumes a loss of all installed safety features and a severe direct breach of containment at 1.5 hours le following onset of core damage -- as a result of a "core melt" accident and containment failure due to "overpressure." See NUREG/CR-2239 at Table 2.3.1-2 (November 1982) which describes the characteristics of the SST1-SST5 release categories. 19/ The SSTl fission product release category, and companion categories SST2-SST5, are described in v arious documents, including NUREG-0771 (at 6-8) ( .J u ne 1 98 1_). In sum, the categories reflect the c onceptu a l spectrum of potential consequences from severe accidents resulting from some form of core damage and fission product release, ranging from limited consequences (SST5) to maximum consequences (SSTl).

The above-mentioned assumptions are inappropriate for purposes of estimating the consequences of station blackout. First, the assumptions underestimate the value of containments in mitigating reactor accidents. They do not address the remedial effects of timely restoration of AC power (i.e., to restore to service those systems which will help to assure that containment integrity is preserved). The technical findings fail to reconcile the large disparity between estimates of time

• to containment failure in NUREG-1032 (at Table 7.3) and those of IDCOR -- despite the recognition that the more recent IDCOR time estimates "may be cause for revision" of the results in NUREG-1032.;~; Had the disparity been reconciled, the opposite conclusion might have been reached regarding current plant coping abilities and the need for the proposed rule.

Second, the effect of choosing this early containment failure mode (as opposed to a late failure mode) is to place inappropriate regulatory focus on a non-dominant risk scenario with an extremely low probability of occurrence. While an early release scenario could be postulated to result from mechanisms such as a steam explosion or direct heating, the 7 frequency of such a release is so small (on the order of 10-or less per reactor year) that it does not contribute significantly to the overall i:-j._sk from station blackout (frequency x consequences) .21; In addition, j_n its Technical

 ~ / See NUREG-1032 at 7-15, 16.

21/ The technical findings maintain the "typical" estimated (Footnote 21 Continued on Next Page

Summary Report (November 1984), IDCOR states that "[t]he necessary conditions for a steam explosion sufficiently energetic to cause primary system or containment failure cannot occur" in a commercial U.S. nuclear power plant. Instead, the risk-dominant sequence is long-term overpressurization of containment long after core melt (!:..:....9.., one to three days). Third, the implication of improper selection of SSTl is significant. Each successive category SST2-5 has progressively lower offsite consequences. In fact, the range of offsite consequences spans about 7 orders of magnitude.22/ According to NUREG/CR-2723 (at 2-11, 12), the SST2-5 categories would not be expected to produce "substantial numbers of offsite consequences compared to the SSTl source term." Thus, the (Footnote 21 Continued from Previous Page) frequ~~cy of core damage resulting from a station blackout is 10 per reactor year (NUREG-1032 at 1-3). NUREG-1032 also states the assumption (at 7-19, T-0ble 7.4) that the containment failure mode associated with the SSTl release ijategory is a "steam explosion," with a probability of 10 22/ According to NUREG-0771 (at 15): The difference in released source term between a Group 1, 'worst case' release, and a Group 2,

         'spray functional' release, is about~ factor of 100. The source term difference between a Group 2 release and a Group 3, 'melt through' accident, is about a factor of 10. The diff~ren~e between Group 3 and Group 4 is about 2 factor of 1000, and between Group 4 and Group 5 of another factor of 10.

choice of SSTl rather than SST2-5 may have resulted in a significant overstatement of the risks of station blackout.~/ Fourth, it is significant to note that the weakness in relying on the SSTl release category is recognized, since a reduction factor of one-third is applied to the SSTl release (at a 50 mile radius). See NUREG-1109 at 8. This modification is made in order to account for the fact that "[i]f a core melt resulted from station blackout [i~-, as opposed to the types of severe accidents for which the prompt containment failure assumption of which SSTl is appropriate], containment failure would be delayed for a number of hours." Id. (emphasis added). No rationale is provided as to why the application of a simple reduction factor of one-third to the offsite release characteristics of the SSTl category is appropriate in lieu of choosing from categories SST2-5 -- a choice which, significantly, would lower the person-rem offsite consequences, not by a factor of three, but from one to four orders of magnitude. Further, it is unclear how the straight- forward application of a multiplier on the person-rem offsite consequences associated with the SSTl release category is a technically sound approach for purposes of simultaneously correcting the (1) incorrect containment failure mode and timing characteristics and (2) the incorrect fission product inventory characteristics of SSTl. 23/ If the use of SSTl was supplanted by the use of SST2 or SST3, the person-rem calculation would be lower by 1 to 4 orders of magnitude. NUREG/CR-2239 at Table 2.3.1-3.

A better approach to risk estimation would have been to continue the analysis of accident sequences in §7 of NUREG-1032 using a model which would correctly match containment failure modes on a mechanistic basis with the various accident sequences. Instead, (1) the accident sequence analysis in NUREG-1032 was simply performed to the point where core damage is predicted (see NUREG 1032 at 7-1), (2) core damage was assumed synonymous with core melt and a precursor certain to

• result in near-term containment failure (Id. at 7-5) and, the SSTl release category and its attendant containment failure mode characteristics (i.e., direct breach of containment at 1.5 hours after the onset of core damage) was assumed appropriate (3) for representing offsite consequences (see NUREG-1109 at 8).~/

The failure to more fully and carefully substantiate the selection of the SSTl release category -- which is the most severe release category which can be identified for purposes of predicting both large offsite releases (as consequences) and large averted person-rems (as benefits) of the proposed rule-- serves to demonstrate that the technical record does not support the proposed rule.

 ~ / The ACRS has previously commented that the equating of core melt with fission product release is ill-advised. As recently as March 28, 1986, at the ACRS meeting on Safety Goals, Professor Okrent stated that ACRS has "[n]oted the need to be alert to the fact that core ~elt frequency may be a poor measure of the frequency of release of a significant amount of radioactive material from containment." Meeting Transcri.pt ,it 1. 0. further, the Staff has elsewhere stated that **rsJ -=*,*,=re core damage sequences would not necessarily involve complete core meltdown." NUREG-0772 at 3.16 (June 1981) (citing Three Mile Island as an example of a beyond-design-basis event which did not lead to core melt-through and containment failure).

Fifth, the SSTl release category may have been chosen to correspond to the dominant station blackout accident sequences which require AC recovery in 1 to 2 hours to avoid core damage. These accident sequences are summarized on Tables 7.2 and C.1 of NUREG-1032. They include the "TML B " sequence for all PWRs 1 1 and the "TMU B " sequence for all BWRs.25/ Thermal hydraulic 1 1 analyses are cited for these early core cooling failure types of accidents which estimate that, following these station

• blackout sequences, core uncovery (i.e., onset of core damage) would occur in 1-2 hours and reactor vessel melt-through would occur in another 1-2 hours. See NUREG-1032 at C-3. The selection of SSTl, which assumes containment failure in 1.5 hours after the onset of core damage, could be attributed to these thermal hydraulic studies. The citation to the studies and explanations of timing tend to expose an underlying technical premise of the proposed rule, namely, that it is appropriate to equate the onset of core damage with vessel melt through and containment release .

In this regard, the above-mentioned PWR and BWR dominant accident sequences and their specific characteristics (~~-, thermal hydraulic analysis) are inappropriate to support the choice of SSTl as the release category which is generically reflective of station blackout risks. More recent analysis of See NUREG-0771 at 8-10 and NUREG, CR-2723 at 2 for other examples and explanations of these sequences and the alphanumeric terminology. See n.26 below for an explanation of a specif i C secTuence, the "TMLB, II sequence. )

these same station blackout sequences indicate that the technical findings and assumptions in NUREG-1032 and NUREG-1109 may unrealistically predict the occurrence of rapid containment failure in connection with station blackout. For example, EPRI has performed a reanalysis of the TMLB' accident sequence26/ for the Surry Plant (subatmospheric containment) and compared it to specific WASH-1400 analysis of this accident sequence (from which the findings regarding the timing of containment

• failure are apparently derived27/ ) . EPRI NP-4096, "Surry Source Term and Consequence analysis," June 1985.

calculations of a Surry TMLB', which utilized the post-WASH-1400 MARCH-2 computer code, predicted a much lower rate of The EPRI 26/ The "TMLB'" sequence is characterized by loss of all AC power and reactor coolant heat removal including failure of emergency core cooling and AC-powered containment heat removal systems. The general features are described in Appendix I to WASH 1400 to involve a transient event (T) in combination with loss of both main feedwater (M) and auxiliary feedwater flow (L) to the steam generators, and the loss of all AC power (B) (onsite and offsite) for a period long enough to cause core meltdown. In this preconceived scenario the likelihood of early containment failure by overpressure would be "potentially high" and the consequences "potentially severe." NUREG-0772 at 3.20. 27/ It is useful to compare the different release category conventions to demonstrate that the timing of containment failures is consistently represented in NRC documents. The SSTl release category utilized in technical documents on station blackout (e.g., NUREG-1032, NURSG/CR-3226, NUREG-0772) is comparableto WASH-1400 (NURSG-75/014) category PWR-2. See NUREG-2723 at 2. In WASH~1400 (at Appendix VI, Table vr-2--=1), the PWR-2 release catego,y assumes a time of cont~inment failure of 2.S hot1rs from th 0 time the accident sequence is initiated, (i.e., *:::0 1'1p2,_-= 1Jl'::' to SSTl which uses 1.-5--h-ours frOm -on.se-Cof cr.Ji:e damage coupled with the Staff assumptions that core damage would begin between 1-2 hours after accident initiation).

pressure buildup during the early !;tages of the accident.28/

• In fact, while the simplified calculations of WASH-1400 predicted a containment pressure of 77 psia at 1.5 hours and containment failure by overpressure at 3.67 hours, the updated EPRI analysis predicted that the containment pressure at 1.5 hours would not exceed normal atmo~pheric pressure, and, further, that there would not be overpressure failure for "at least the first 12 hours of the accident." Id. at 3-1, 3-2 .
• The EPRI analysis suggests that the technical findings regarding accident sequences may be incorrect, and, at a minimum, do not support the selection of the SSTl release category._?_~/

Sixth, the proposed rule's association with WASH-1400 (see 51 Fed. Reg. at 9830, col. 1) technical findings regarding the amounts of fission products released and containment performance (i.e., failure modes and timing of release) is itself cause for concern that station blackout risks have been overestimated. For example, NUREG-0772 states (at F.3) that fission product deposition in the primary system, the occurrence of which tends to reduce fission product releases in

   ~ / EPRI NP-4096 at 3-1. The application of the MARCH-2 computer code utilized Surry plant design parameters listed

. in WASH-1400 .

   ~~/   In NUREG-0956, a sj_mi lar concJ.11s ion '* 2,.: :eached in a   1   1 reanalysis of Surry containment perf~rmance during severe accidents. The NUREG states (at C.1-9) that "(c]ontainment
        -~~~_!l1_1::_<:_ 1,vJ_t_lJ_i~1 tl1_~ f.!:_r_st -~~-"~ l]ou_r_s is _<ll!ite unlikely" (emphasis supplied).

the event of containment failure, "[c]ould not be evaluated in WASH-1400 because of lack of applicable models" and that "significant extensions" beyond WASH 1400 methodologies are now possible.l__Q_/ Such improved methods have been used for risk calculations related to containment performance in other areas, but not for station blackout.l!_/ In NUREG-0956, Appendix C.2

• provides the Executive Summary of the Containment Performance Working Group (CPWG) Report (i.e., NUREG-1037).

severe accident research program. The CPWG studied containment failure modes as part of the overall NRC Specifically, the CPWG considered containment leakage as a function of time and the impact of containment pressure relief on the mode and timing of containment failure. In contrast to the WASH-1400 results, the CPWG found that, in most cases, containment integrity would not be challenged "until several hours after vessel failure." le NUREG-0956 at C.2-11. Thus, a characteristic feature of the SSTl release scenario as applied to station blackout events l.QI NUREG-0956 describes (at 8-1) the "two major deficiencies" in WASH-1400 which are "often cited as causes for overestimation of radioactivity released from a plant during a severe accident [e.g., station blackout]" . One was WASH-1400's treatment of iodine as elemental iodine rather than the less volatile, more soluble salt, cesium iodide. The other was the omission of ~atural processes which would retain radionuclides, particularly in the reactor coolant system. 31/ NUREG-0956 also explains (at 8-1) th,::,i~ the "BrH-2104" suite of computPr codes "rPprPsents a rnnjor advancement in technology ;_rnd can be used to replace" the WASH-1400 methods (which yielded the SSTl release category used in NUREG-1109).

(i.e., timing of containment failure and offsite release) has ... been seriously questioned . In conclusion, the use of the SSTl release category in the technical documents underlying the station blackout rulemaking, NUREG-1032 and NUREG-1109, has not been justified nor reconciled with the technical findings in NUREG-0956 and other technical reports. II.A.3 NUREG 1032 Contains Errors and Omissions The provisions of the proposed rule and regulatory guide most affected by errors and omissions pertain to loss of off-site power assumptions. In this regard, the Notice of Proposed Rulemaking states that the likelihood of a station blackout is "directly proportional" to the likelihood of a loss-of-offsite power event. 51 Fed. Reg. at 9834, col. 3. The proposed rule links (Id. at 9830, col. 2) the potential and duration for a loss-of-offsite power event to three factors: plant offsite-power design features, grid stability, and severe weather. Each of these factors is discussed below.32/ Offsite Power System Design. In postulating a relationship between offsite power system features and the In any discussion of loss-of-off-site power events it is important to keep in mind that, industry-wide, there are reasons to believe that the fr~auenc v of loss-of-off-site power events is declining, thereby reducing the emphasis it should play in station blackout. See NSAC-103, "Losses of Off-Site Power at U.S. Nuclear PowerPlants," April 1985 at iii.

potential for a loss-of-offsite power, durations were assi gned . in NUREG-1032 to certain events of less than one minute duration "[t]o facilitate the statistical analysis."ll/ Also, one event was double-counted and several non-LOOP events were introduced in the analysis.34/ While none of these items, standing alone, is dispositive, the net effect of these adjustments is to result in questioning the database and add unnecessary conservatisms. At a minimum, these adjustments

• prevent others from reproducing the analysis in an objective manner.

Further, it appears that a significant statistical error exists in the failure to account for the large sum-of-squares error associated with the results.35/ While some correlation was found, the overall quality of the statistical model, as measured in "R-Square" terms36/, is less than 25%. This type of error undercuts the validity . of the correlations obtained between offsite power design features and the likelihood of a loss-of-offsite-power.37/ In view of the above, a sound basis ll/ NUREG-1032 at A-10. 34/ See NUGSB at 85-005, supra n.16, at 4-3 to 4-6 and 4-7 to 4-8. 35/ Id. at A-9 (Appendix A, Item 3). 36/ "R-Square" refers to a term in the statistical model which measures the capability of the model to account for variation in the dependent v a riable. If R-Square is small, errors in the model are large a nd th ~ ~odel is unreliable. 37/ NUGSBO had previously performed a difference of means test (based upon the "t" statistic) to determine if the offsite power design groupings were distinguishable. This test (Footnote 37 Continued on Next Page

has not been established for associating certain offsite power design features with the potential for a loss-of-offsite power in the manner done in NUREG-1032. The statistical errors also tend to question the distinction between "offsite power design characteristic groups" underlying the 8-hour coping requirements of Table 1 of the draft regulatory guide (i.e., the "Pl" and "P2") groups .

• Grid Stability and Weather.

grid-related and weather categories. NUREG-1032 also postulated Unlike the category of plant-centered events, however, these two categories are not based on a rigorous statistical review of plant experience. Instead, NUREG-1032 arbitrarily assigns event frequencies and power restoration times. No citation to other data sources or supporting analyses is provided to show that actual experience, or even the basis for the values assumed, confirms the appropriateness of the grid-related and weather-related categories of events. Thus, it is not possible to determine the reasonableness of the assumptions made.38/ In addition, the methodology used to determine the effects of weather on the frequency and duration of loss-of-(Footnote 37 Continued from Previous Page) only established statistically distinguishable mean durations of offsite power loss amon9 odjacent groupings of 60 and 80 % confidence, well below th~ levels of confidence ordinarily necessary to conclude that a distinction exists. Id. at A-9 through A-11. 38/ Id. at 3-7 to 3-22 (§§3.2.2-3.2.5).

offsite power events lacks a sound technical basis. NUREG-1032 cites (at A-22) one paper (i.e., by Lauby, et al.) in the technical findings as yielding the correlation between weather hazards and loss-of-offsite power frequency.39/ The Lauby paper is central to the te~hnical findings concerning this aspect of the proposed rule. However, in contrast to the use of the Lauby paper in NUREG-1032, Lauby never actually assumed a linear relationship between power loss frequency and weather-

• hazard rate. The paper did not even examine the potential for such correlations. Also, contrary to the implications in NUREG-1032, Lauby did not evaluate hurricane or tornado hazards.

In fact, with respect to the conclusions in NUREG-1032 regarding grid stability and weather, the Lauby paper reaches opposite conclusions. For example, the Lauby paper found (at 2349, col. 2) that:

 *              [t]he 345 kV Line-Related and the 230 kV and 345 kV Terminal-Related weather associated Forced outage rates were not
                ~ignificantly different between the winter su -divisions of mean annual snowfall which was used to simulate the frequency of snowstorms. (Emphasis added.)

I

   ~ / M. G. Lauby, et al., "Effects of Pooling Weather Associated MAPP Bulk Transmission Outage Data ori C2J. culated Forced Outage Rates," ( January 1984) P.:t per ( 84 i-JMO410) Presented at IEEE Winter Power Meeting, Dallas, Texas. IEEE Transactions On Power Apparatus and Systems, Vol. PAS-103, No. 8, 2345-51.

The implication of this conclusion is that the paper was not intende d to support the hypothesized relationship between weather hazards (cumulative amount of snowfall in this instance) and the potential for loss-of-offsite power events at a giyen plant. In addition to the above described errors in NUREG-1032, similar errors were found in the reliability equations apparently used in calculating the unavailability of the emer gency AC power system.40/ However, since the actu~l risk equations used in calculating core damage frequency are not presented in NUREG- 1032, it is unclear whether the various derivation errors in the treatment of the emergency AC power system reliability are merely typographical or, in fact, are carried forward in a substantive manner into the analysis. Such substantive errors in the analysis of core damage frequency due to station blackout, together with the above-described errors in the loss-of-offsite power analysis, raise further concerns as to the sufficiency of the underlying technical bases in supporting the generic rulemaking. II.B Additional Matters Included in the Technical Record Do Not Support the Proposed Rule During the Commission's consideration of the

 '        recommendation to publish the proposed rule, several matters 40/ Corrections to the Staff's derivations are provided in NUGSBO's comments of November 1985. See NUGSB 85-005, supra n. 16 at 3-25 (§3.3.2).

were raised as additional support for the rule. These matters are as follows and will be further detailed below. The European Approach to Station Blackout Reactor Coolant Pump Seals Hurricane Gloria II.B.1 Clarifications of European Approach are Necessary Various oral and written representations regarding the European approach (particularly that of the French) to station blackout risks have been made and are part of the technical record.41/ In short, the European approach to station blackout should be viewed separate from the European approach to protecting against unique external circumstances which are unrelated to station blackout. In this regard, it should be clearly recognized that the French found the likelihood of a station blackout to be on the order of 10 -5 per reactor-year.42/ This value compares well with point estimates provided by the Staff for design features present at most U.S. plants. The French considered this risk to be acceptable, with two conditions: (1) the implementation of special procedures43/ and (2) providing the means for diminishing the threat posed by 41/ See Transcript of November 14, 1985 Commission meeting on station blackout ( ~ , at 28, 39, 40, 47 , etc.). 42/ See Tanguy, infra n. 49, at 595, co l . 2 . These procedures are directed toward l easonable operator actions which could be taken to restore power and provide core cooling. The init i atives advanced by industry provide comparable procedures.

loss of reactor coolant pump (RCP) seal integrity. Recent information indicates that RCP seals do not represent significant potential for large inventory loss (see §II.B.2 below). At the time the French examined this issue, however, the integrity of reactor coolant pump seals in a station blackout was viewed with great uncertainty. This concern prompted the installation of a turbine-driven pump to provide makeup. More recent information (discussed in Section II.B.2)

• indicates that seal integrity negates the need for a pump of this kind.

The French approach to station blackout does not appear to depart significantly from current regulatory approaches in the U.S.44/ Other countries are not much different. In Belgium, for example, current U.S. practices concerning onsite and offsite AC power reliability and redundancy, as well as the lack of a design basis coping capability, remain the norm. The current position is that the diversity in AC power sources makes the event highly improbable.45/ Italy and Switzerland provide similar layers of AC power system redundancy and reliability and, thus, are relatively unconcerned about station blackout.46/

 ~ / One area of departure may be coping capability. The coping capability of French plants will be discus sed below.

45/ See IAEA-TECDOC-332, "Safety Aspects of Station Blackout at Nuclear Power Plants," at §B.1, International Atomic Energy Agency, Vienna, 1985 (hereafter "IAEA"). 46/ Id. at §§B.7 and B.12.

Perhaps the best explanation for this perspective on the part of the Europeans is provided in the German design philosophy which is summarized in the following statement:47/ In German nuclear power plant design the station blackout (which is loss of on-site and off-site normal and emergency AC power), is expected to be an event of extremely low probability of occurrence due to the design of emergency AC power supply systems. (emphasis added) This reliability can be measured by the summary results

• of a recent German reliability study which reports 8 loss-of-offsite-power events in 33 years of operation of six plants.48/

This experience translates to a frequency of approximately 0.04 events per reactor per year, a frequency which is close to the annual U.S. event frequency experienced since 1979. That there are strong similarities should not be surprising since U.S. designs formed the basis for most plants in operation throughout non-Soviet bloc nations.49/ Similarly,

• 47/ Id. at §B.6.

48/ Id. at §B.6 (summarizes an article published in Atomwirtschaft, February 1984). j9/ In a paper by P. Tanguy, Director of the Institute of Safety and Protection at the French Atomic Energy Commiss ion (CEA), the French approach to nuclear safety was addressed: [t]here are strong similarities between French safety philosophy and the safety philosophies of the other major nuclear 9ower countri es; this is hardly surpr isi ng because for more than 20 yr there has been an extensive and continuous exchange of information between nearly all organizations around the world, either by means of bilateral agreements or (Footnote 49 Continued on Next Page

U.S. regulatory practices are referenced extensively in regulating power plant design and operation.SO/ From this common base, local regulations have evolved which recognize the differences in hazards faced by U.S. and foreign plants. Such hazards in Europe are unrelated to station blackout events, and generally result from greater population density in the vicinity of power plant sites, differing seismic criteria,

• and the greater likelihood of external events, such as military aircraft impact and gas explosions. Protection against these hazards is provided in some European countries through~ number of local design features which have the effect of providing greater reliability . of the AC power systems required for plant safety. It has been the practice in some countries to provide an additional level of redundancy in the number of diesel generators available to the plant. Discounting the common cause contribution, the overall emergency AC power system reliability sought is approximately 10- 4 per demand.51/ A
• (Footnote 49 Continued from Previous Page) under the sponsorship of such international bodies as the International Atomic Energy Agency in Vienna and the Nuclear Energy Agency in Paris.

P. Tanguy, The French Approach to Nuclear Power Safety, _ Nuclear Safety, vol. 24, No. 5, at 589, cols. 1-2, September-October 1983 (hereinafter "Ta!1guy"). 50/ See generally id., at 592, col. 2 . 51/ J. A. Richardson, "Summary Comparison of West European and U.S. Licensing Regulations for LWRs", Nuclear Engineering International, February 1976.

similar concern for the potential loss of the decay heat removal function due to external events motivates the use of 4-50% capacity cooling systems or 3-100% capacity cooling systems. In an accident, two of the four (2/4) or one of the three (1/3) systems need to operate to provide core cooling. The picture which emer9es from this brief review of the European approach to station blackout is clear: AC power

• reliability over and above U.S. levels results from European unique events and not from a particular concern with a station blackout event.5~/

Another issue related to the features of European design raised question in the Notice of Proposed Rulemaking is whether the proposed rule goes far enough in comparison. See 51 Fed. Reg. at 9831-2. Two features cited in the Notice of Proposed Rulemaking are believed to be erroneous assumptions: (1) that the French plants possess a three-day coping capability; and (2) that the French "have a goal of achieving a probability of one in ten million (10- 7 ) per reactor-year for a major event such as station blackout." Id. The claimed three-day French coping capability is deemed to support the proposed rule in that it would provide time to bring a mobile gas turbine onsite and terminate a station blackout event. It is also considered Core cooling considerations associat_c 1 with external events (i.e., similar to the European-unique hazards) is a subject forthe ongoing Task Action Plan for resolution of USI A-45, decay heat removal, not station blackout. See NUREG-1109 at 18.

achievable at reasonable costs. Similarly, the Notice of Proposed Rulemaking states that the French 10- 7 objective represents a higher safet~ objective than the 10- 5 objective associated with the proposed rule.53/ With regard to coping, it is important to point out that there are several reasons why many of the features attributed to the new French plants may already exist at most U.S. plants .

• The apparent focus of the described plant features at the new French plants is to provide sufficient makeup water to primary systems to provide for reactor coolant pump (RCP) seal cooling and preclude seal failure. (Recent full-scale RCP seal tests under station blackout conditions indicate that seal injection and cooling are necessary to maintain seal integrity. (See

 §II.B.2.). These features are similar to the current capabilit ies of many U.S. plants which utilize emergency pumps, such as fire pumps and portable equipment, to provide makeup to the condensate and feedwater systems.         Such capabilities offer broad flexibility to provide decay heat removal should the main auxiliary feedwater and condensate systems be disabled by a station blackout.       Indeed, in the context of providing makeup water, it is likely that a large number of U.S. plants could claim a three-day coping capability comparable to the French.54/
 ~~/  See Tanguy, ~up_ra, n. 49, at 595.

54/ In the separate views accompanying t he pr oposed rule , it is represented that new French 1300 MWe plants are designed with a goal of "coping" with a station blackout for "at least 20 hours" and, further, that the plants can "wi thsfanc:fl'il stat ion blackout for "three days." 51 Fed. (Footnote 54 Continued on Next Page

                               -   34 -

If the French definition of coping capability is limited to simply having sufficient cooling water inventory, then the burdens of coping without AC power would be significantly reduced. The test applied to U.S. plants by the proposed rule and regulatory guide is significantly more rigorous: in addition to requiring that power plants have access to sufficient water for decay heat removal, the coping features of the proposed rule also include the requirement for a

• demonstration of equipment operability in environmental conditions associated with a station blackout.

9835, col. 1. Draft regulatory guide at 7. 51 Fed. Reg. at The current capability of French plants to "cope" with station blackout may be inconsistent with the concept of coping as contained in the regulatory guide published with the proposed rule. No information is provided that French plants can meet the terms of the regulatory guidance on coping capability, nor has it been demonstrated that in terms of the regulatory guide French plants can cope with station blackout for as long as the Notice of Proposed Rulemaking suggests they can. 7 With regard to the 10- figure, it was noted above that the actual frequency of station blackout in Europe is (Footnote 54 Continued from Previous Pag ~) Reg. at 9831, col . 3. It is not clear what distinction, if any, exists between the asserted ability of the French design to " cope " with a station blackout for 20 hours and to "withstand" a station blackout for three days.

comparable to most U.S. plants, i.e., 10 -5 . More importantly, in contrast to the separate views accompanying the Notice of Proposed Rulemaking (51 Fed. Reg. at 9831-2), the numb er 10- 7 per reactor per year is not a goal but a screening limit for excluding families of events from consideration(~, low probability events).55/ To this end, the screening process used for PWR designs is aimed at maintaining the probability of accidents leading to individual offsite exposures between 5 and

• 50 person-rems to be less than 10- 6 per reactor per year .

As has been discussed, NUREG-1032 assumes~ core damage frequency of 10- 4 per reactor per year arising from a station blackout event. Core damage, in and of itself, does not give rise to individual offsite exposures, rather, core melt, vessel failure and breach of containment integrity are necessary.56/ The frequencies associated with these additional events drive the frequency into the range of the French screening limit (i.e., 10- 7 ). For example, as noted (supra n. 21), even the use of the SST! offsite release frequency of 10 -4 results in a core damage/ containment failure/offsite release freq u ency of 10- 7 or less. See NUREG 1032 at Table 7.4 at 7-19. Accordingly, there is little disparity between the U.S. and French in this regard. 55/ See Tanguy, supra, n. 49, at 594-596. As noted in §II.B.2 above, cor~ damage is improperly equated (with the elapse of 1.5 hours) with containment breach and public exposure. It should be pointed out, the French do not equate core damage with containment failure on such a non-mechanistic basis. See Id. at 597.

II.B.2 Clarifications of RCP Seal Integrity are Necessary The regulatory analysis, NUREG-1109 (at 19) links J

. reactor coolant pump (RCP) seal integrity to a plant's ability to cope with a station blackout. The issue of RCP seals was also discussed at the November 14, 1985 Commission meeting. Transcript at 27.

The relationship between reactor coolant pump seal

• integrity and station blackout requires clarification.

issue concerns the ability of the seals to serve as a reactor coolant pressure boundary when deprived of cooling This water in a station blackout. Assumptions regarding the timing of seal failure and the rate of coolant loss are important to determining a plant's ability to cope with a station blackout.57/ To begin with, it needs to be recognized that the

• reactor coolant pump seal integrity issue is a generic issue having its own Task Action Plan (Generic Issue B-23).

principal features of that plan are prototype tests of The various seal designs to determine the most viable approach to extending seal integrity should cooling or injection be lost. The current state of knowledge sepa r ates the seal leakage potential of reactor coolant pumps i nto two groups: Westinghouse pumps with a hydrosta tic se cl, and pumps with a 57/ See NUREG-1032 at §6.

hydrodynamic seal. Pumps with hydrostatic seals are believed to pose a greater risk due to seal failure. With respect to station blackout, however, only leakages of 100 gpm per pump or more are of concern because of their impact on limiting decay heat removal.58/ Smaller leakages do not have significant impacts on coping. This criteria eliminates from consideration all boiling water

• reactors and plants operating pumps with hydrodynamic seals.

Thus, reactor coolant pump seal integrity is only relevant to the Westinghouse pumps with hydrostatic seals due to the larger seal leak potential. The concern over the Westinghouse pumps is also approaching resolution. Recent tests of this design have introduced substantial doubt regarding the previously assumed large leakage potential. At this time, it does not appear that high pressure AC-independent injection pumps are needed to compensate for seal leakage. In fact, since 1983, the Staff has recognized this improvement in RCP seal integrity and uses a 20 gpm per pump primary inventory loss rate in their analysis of station blackout.59/ This value is below the 100 gpm criterion cited in NOREG-1032 and is ' supported by the revised Westinghouse reoor~ (WCAP-10541, Rev. 1) . 58/ NUREG-1032 at 6-3. 59/ See NUFrnG/CR-3226 at Appendix G at 248.

The impact of this discussion highlights the fact that

,   concerns over RCP seal integrity do not contribute to the overall station blackout concern. This issue was originally addressed to Westinghouse plants, but with the impending resolution of the RCP seal integrity issue, further discussion of the risk of potentially large inventory losses in a station blackout does not appear to be relevant.

II.B.3 Clarifications of the Significance of Hurricane Gloria

• are Necessary At the November 14, 1985 Commission briefing on station blackout, a general concern was raised (Transcript at 8) that nuclear utilities were ill-prepared for severe weather events believed to be major contributors to station blackout risks.60/

In particular, reference was made to the late-September 1985 Hurricane Gloria which moved up the East Coast from Florida to New England .

• At the briefing, Hurricane Gloria was represented as illustrative of the basis for the concern that . utilities cannot adequately respond to rapidly moving severe weather in a timely fashion, j.e., in time to take proper precautionary measures.

These representations seriously understated the present capability of utilities to take timely and effective risk- . reduction measures in response to such seve r e weather events as Hurricane Gloria. 60/ See also,~-, 51 Fed. R~. at 9830, col. 2 and at 9831, col .1 .-

For example, at Millstone Unit Nos. 1 and 2 in Connecticut, the licensee had advance notification of the storm's approach, easily tracked Hurricane Gloria and routinely implemented its onsite hurricane action plan. This plan includes, among other things, a checkout of the Emergency Response Facilities, the selection of two Station Emergency Organization shifts and verification tests of all emergency onsite AC power sources.61/ Despite the storm's rapid movement

• toward the plants, the licensee had ample time to take all necessary precautions for station blackout risk-reduction purposes, including orderly plant shutdowns, and preheating/starting the emergency onsite AC power sources. Due to these precautions, the Millstone units were in stable shutdown operations during the weather-related loss-of-offsite power at the site.

In conjunction with bringing the units off-line, all Millstone emergency on-site AC power sources successfully started and loaded and ran until prudent plant actions were completed to allow for restoration of normal off-site power. If necessary, Millstone Unit No. 1 could have had off-site power restored within 3 1/2 hours and Millstone Unit No. 2 could have had off-site power restored within 5 1/2 hours. Since more rapid restoration of off-site power was not vital, the util ity elected to pursue a mo r e deliberate and thorough cleaning and checking restoration proc ess. 61/ Northeast Utilities letter to NRC (J.F. Opeka to H.R. Denton), December 27, 1985, at 1.

The advance notification associated with severe weather events of this kind permits advance precautionary actions not usually credited by the Staff or plant probabilistic safety studies. As noted by members of the Advisory Committee on Reactor Safeguards during the November 19, 1985 Subcommittee meeting in Waterford, Connect icut (Transcript at 207 through 209), the failure to credit utility capabilities for actions in response to advance warning prior to a severe storm arrival led

• to conservatisms in a probabilistic risk assessment, and accordingly, these events should be categorized in a fashion different from other loss-of-offsite-power events.

In this regard, the Atomic Industrial Forum's Public Affairs and Information Program has assembled information from affected utilities which is in contrast to the assertion mentioned above that the present capability for severe weather precaution at nuclear plants is i nadequate.62/ In spite of

• 62/ See INFO Report, Number 204, October 1985, which, with respect to other affected utilities, states as follows (at 2-3):

A number of nuclear plants on the East Coast from the Carolinas to New England voluntarily shut down or reduced power in precautionary actions in late September as Hurricane Gloria came calling - or threatene d to call - with her blustery winds. No significant damage was reported at any nuclear units . Carolina Power & Light Co. brought Brunswick 2 in North Carolina down because of the possibility the storm would move towards the sit e . Unit 1 was down for scheduled maintenance at the time. CP&L's Wayne Ennis said the utility was not concerned about the plant withstanding the hurricane winds, but about the possible loss of transmission lines (Footnote 62 Continued on Next Page

suggestions at the November 14, 1985 briefing, the utilities (Footnote 62 Continued from Previous Page) into and out of the plant. The plant was down less than 24 hours due to the storm, but CP&L took the opportunity to do some maintenance work while it was off line. In New Jersey, Jersey Central Power & Light Co. [sic, GPU Nuclear Corp.] reduced power at Oyster Creek to 35 percent on the evening of September 26. JCP&L's [sic, GPU Nuclear's] John Fidler said the

• action was taken to avoid a potential situation in which the plant lost offsite power and then tripped at full capacity. However, Gloria caused no such problems and the plant was returned to 100 percent power the next day.

Public Service Electric & Gas Co. reported that Salem 1 & 2 escaped Gloria unscathed, although power was reduced slightly after the storm so intake screens could be cleaned. On Long Island, where the brunt of the storm was felt, the Long Island Lighting Co. began shutting down the Shoreham plant September 26 and declared an unusual event and an alert the next day. "One of the triggering mechanisms for an usual event condition is wind speeds or anticipated winds in the vicinity of 80 mph or above," said LILCO spokeswoman Carol Clawson. "For an alert, it is wind speeds of 100 mph in the vicinity of the plant, " she said. Shoreham, which currently has only a low-power license, was back up by October 1 [sic, 3). In Massachusetts, Boston Edison Co. reduced power at its Pilgrim 1 plant to 25 percent late in the afternoon on September 27 and kept power at that level until 1 a.m. Saturday morning because winds of about 70 mph were reported in the area. The action was taken in accordance wi~h the unit's technical specifications. The utility then shut the plant down voluntarily to wash salt off the insulators. Pilgrim 1 was back at 100 percent by Sunday. (Footnote 62 Continued on Next Page

threatened by Hurricane Gloria successfully averted the risks of station blackout. II.C Responses to the Additional Comments and Views of the Commission Must be Considered Among the considerations of the rulemaking are particular concerns raised by the Commissioners in their additional comments and separate views accompanying the proposed rule (51 Fed. Reg. at 9831-2). Since much of the information that would

• (Footnote 62 Continued from Previous Page)

Other utilities reported that their nuclear plants ran smoothly, despite the storm. Baltimore Gas & Electric Co., for example, said Calvert Cliffs 1 and 2 operated at full power throughout the storm. Had Gloria hit the area, however, the utility was ready for her: the utility had called in extra workers to be on hand. In addition to these events described above, other utilities were able to take appropriate measures. In accordance with the Indian Point Unit No. 2 (IP-2) Technical Specifications, Consolidated Edison commenced hot shutdown procedures when the center of Hurricane Gloria was within 320 nautical miles of Indian Point with sustained winds exceeding 100 mph (Approximately 0230 EDT, September 27, 1985). The Technical Specifications also required that appropriate actions be taken to ensure that the plant be in the cold shutdown condition prior to arrival on site of hurricane winds exceeding 100 mph. This cold shutdown action was not required in the case of Gloria since winds of 100 mph or greater were not experienced in the IP-2 proximate area. As an additional precautionary measure, the site declared a "Notice of Unusual Event" (NUE) in accordance with its Emergency Plan. (An NUE is the first and least significant of the four action levels in the Emergency Plan.) On September 27, 1985 at approxim~tely 1600 EDT the National Weather Service had lifted the hurricane warning for the area and IP-2 began start-up procedures. Virginia Power took Surry 1 offline and reduced Unit 2 to 25 percent power for a few hours as a precaution for employees that were working outside to assist in shutt ing down the plant.

be contained in response to Commissioner questions is presented throughout this document, the appropriati sections of the comments are cited. There are five issues raised by the Commissioners as comments or separate views: (1) The need for "quality classification" of station blackout modifications;

• (2 )

(3) Whether the backfit analysis for the proposed rule "adequately i mplemen ts" the Backfit Rule, 10 C.F.R. §50.109; Whether the reduction in risk offered by the proposed rule constitutes a "small percentage of the overall risk" or "a major component of an already small risk"; (4) Whether the proposed rule meets the "substantial increase in the overall protection of the public health and safety. threshold required by the backfit rule"; and, ( 5) Whether the NRC "should require substantial improvements in safety with respect to station blackout, like those being accomplished in other countries, which can be achieved at reasonable cost and which go beyond those proposed in this rulemaking." (1) Qual ity Classification is Unnecessary. Equipment used to prevent or respond to a station blackout should be sufficiently available and operable to meet its required function. To this extent, the Commission's desire that approp ri ate attention be paid to maintaining a sufficiently high state of operability and reliability is appropriate. The poi~t of departure begins with the method for achieving this objective. Specifically, by itself, a "safety grade" classification scheme does not solely

equate with high states of equipment operability and reliability. Such classification systems ~oo often can become a documentation exercise more than i process for providing the requisite level of system functionality. In lieu of a classification system, the goal of system availability should be pursued programmatically.63/

• ( 2) Implementation of the Backfit Rule Is Inadequate.

backfit analysis for this rule does not adequately implement the Backfit Rule, 10 C.F.R. §50.109. concerning this point is provided in §IV. Additional information The (3) Reduct ion in Risk of Station Blackout Is Small. An important aspect of this rulemaking is the estimated frequency of core damage due to station blackout. For most plants, this value can be calculated using the methods in NUREG-1032 to be curren~ly near or below 10- 5 per reactor year. If the first half of the industry's loss-of-offsite-power experience is excluded in recognition of industry improvements,64/ the estimated frequency of all station blackouts greater than 2 hours in duration is near or below 10- 5 per reactor per year for all p lants. 63/ This goal may best be achieved through the initiatives advanced by the industry. For the r e ~sons discussed therein, that program can make 3i gn ific ant progress towards that goal without incurring significant burdens in the area of documentation. 64/ NUGSBO previously presented information to the NRC supporting this position. See supr~ n. 16.

The Notice of Proposed Rulemaking acknowledged that the "total risk from nuclear power plant accidents . was found to be small." 51 Fed. Reg. at 9830, col. 1.65/ This "small" risk is clearly acceptable and does not pose an undue risk to the public health and safety.66/ Thus, any reduction resulting from the proposed rule will be in the category of diminishing returns. This point is underscored in NUREG-1109, table 5. In two of the three categorie"s addressed, the reduction in risk is

• small (i.e., 1.3 x 10- 5 and 0.1 and 10- 5 ).

two points

• or d er.

are in

• First, In the other category, wherein the risk reduction is comparatively large (14 x 10 -

5)

• as note d , severa 1 o f the plants falling into this category are able to take actions which reduce the initial risk and, thus, render small any reduction associated with the rule.67/ Second, even assuming the risk reduction advanced by the proposed rule, the number of plants falling into this category is small (i.e., 14 plants, 7 sites) and, thus, while the risk reduction might be large as to
• The proposed rule's acknowledgment was made on the basis of WASH-1400 results (i.e., one of the primary technical sources of reliance in the proposed rule). WASH- 1400 was published in 1975. Clearly, given the numerous rulemakings and requirements that have been issued since then (i.e.,

NUREG-0737 responses, fire protection, equipment -- qualification, emergency planning, etc.), the risk is even smaller now. It was pointed out in §I that no plants pose an unacceptably high risk resulting from station blackout as to require immediate action (i.e., the risk was not a threat to public health and safety). See NUREG-1032 at 2-3* 67/ See § I, s upra n. 6, regarding the ability of plants with 1""/3 emergency diesel generators to cross connect and thus have an effective configuration of 1/3.

these limited number of plants, the risk reduction associated with the majority of plants will be small. Thus, as a general matter, the reductions in risk offered by the proposed rule constitute a small percentage of the overall risk, a risk which is already small (and acceptable). The reduction in risk for a limited number of plants could be larger, however, as some of these plants are able to take actions which reduce the risk and, thus, the proposed rule's risk reduction benefits .

• Placing station blackout risk in the context of overall risk requires knowledge of several other parameters. Important elements of that state of knowledge are the estimated frequency of other accident scenarios, containment performance in a severe accident, and source term assumptions. Updated information in this regard is expected in the next severe accident research report. This report, NUREG-1150, is currently scheduled to be released late in the summer of 1986.

Additional comments may be appropriate at that time which will further address the relative risk question. (4) Rule Does Not Provide Substantial Additional Protections~ The proposed rule does not meet the threshold provided by the Backfit Rule. See §IV for comments concerning this issue. (5) Additional Modifications Are Not Necessary. The European approach t o station blackout i s no t i n c on sist ent with the present U.S. approach. Accordingly, additional modifications

premised upon the Europea n approach are unnecessary. See discussion at §II.B.1 .

III. The Proposed Rule Itself Should Be Reevaluated III.A The Ultimate Requirements of the Proposed Rule Are Indefinite and Depend Upon the Future and Uncertain Exercise of Discretion The proposed rule would create indefinite requirements. The Supplementary Information accompanying the proposed rule explains that the rule would require an open-ended determination of "[t]he amount of time the plant can maintain

• core cooling and containment integrity with AC power unavailable." 51 Fed. Reg. at 9831, col. 1.68/ The problem is further illustrated by the fact that the regulatory guide and proposed station blackout rule do not provide limits on utility efforts to demonstrate required coping. See the regulatory guide (March 1986) at §3.1. In fact, the text proposed for codification by the rule does not specify a numerical value for required coping time .
• Unless the required coping demonstration is specifically bounded by clearly stated definitions, assumptions, and criteria, there could conceivably be hundreds of supporting special effects analyses which licensees may have to consider as a result of the exercise of discretion by individual Staff reviewers. Under the rule as proposed, lic~nsees cannot ascertain the ultimate requirements they will be expected to 68/ This demonstration would be a meaningless exercise if it results in coping periods in excess of the 4 or 8 hour categories referenced in the Draft Regulatory Guide.

meet (including the potential plant modifications they will

• need to make) to demonstrate compliance.

For example, in the draft regulatory guide it is stated (at 7) th~t equipment necessary to withstand a station blackout must "[m]eet design and performance standards that ensure adequate reliability and operability in extreme environments, that may be associated with a station blackout including

• hazards due to severe weather. It is not clear how this II guidance will be interpreted. If licensees are required to _

consider loss of AC power in conjunction with 10 C.F.R. §50.49 reviews, the scope of §50.49 will be significantly expanded and this will have a substantial impact on industry. In this regard, the potentially significant activities necessary to demonstrate adequate reliability and operability in environments that may be associated with a station blackout do not appear to have been addressed. It should be clearly stated that the consideration of total loss of AC power is unnecessary for §50.49 demonstrations of equipment qu~lification. To proceed o therwi se would subject utilities to substantial costs associated with changeout and replacement of much equipment which is functionally sound in all respects but which may lack documentation sufficient to satisfy the rule. Accordingly, the technical guidance accompanying any rule should be flexible enough to permit suitable engineering anal ysis and prudent compensatory measures ( ~ , openi~g equ ipmen t cabinet doors to provide cooling) to assure that the necessary design and performan ce standards of concern have been satisfied.

The point above is not that regulations must be prescriptive by their very nature. Prescriptive regulations, which outline in detail exactly what steps are required by licensees to satisfy a proposed regulation, are, in many instances, unnecessary and counterproductive. Rather, the point is that demonstrating conformance with new design requirements is time-consuming and resource intensive, especially when the criteria applied to judge the adequacy of

• licensee verifications are not identified from the outset (and, in the case of the proposed rule, will be indirectly developed as a result of licensee submittals rather than directly in the rulemaking) .§21 Licensee activity needed to implement the draft rule has not been accurately described. The regulatory analysis, NUREG-1109, assumes that all plants will derive benefits from implementing the rule. In the value/impact assessment (discussed in Section IV) the rule is calculated to result in a total dose reduction of 80,000 person-rems across all of industry over the remaining life of the 67 plants it considered.70 / Yet it is also claimed that "almost all plants 69/ As addressed below in connection with the discussion of cost estimates (§IV.B), licensee activities cannot be determinPd for purposes of the NRC satisfying its obligations under the backfitting rule, 10 C.F.R. §50.109, unle ss the proposed rule first identifies clearly what it propos e s to require and d~scribes wit~ rea sonabl e detail how licensees are expected to oc~ie 00 ? comp liance.

I!}_/ In NUREG-1109 (at 8) it WuS explained that the 80,000 person-rems was derived by utilizing information from NUREG/ CR-2723 which provides estimates of offsite (Footnote 70 Continued on Next Page

should be able to meet [the requirement s of the proposed rule] without major modifications."71/ It is difficult to understand how such significant benefits across all of industry can be

*. assumed to accrue from a backfit that the regulatory analyses claims will not require any hardware at "most plants."     If, in fact, there will be a total dose reduction of 80,000 person-re.ms over the industry, it would appear to follow that some licensee activities will be required "over the industry."

• These inconsistent statements point to the vagueness in the scope of the proposed rule.

Industry concerns are long- standing over insufficient precision in regulations. On October 27, 1980, in the fire protection rulemaking, the basis for Section III.G of 10 C.F.R. SO, Append ix R to the Commiss ion was presented. The purpose of the section was to achieve a consistent level of fire protection safety at those plants where safe shutdown system separation remained an open item . Approximately 20 plants

     " . . . would be significantly affected by the backfit of

[Section III.G]."72/ While backfitting this requirement on all plants, the Commission recognized that the majority of plants which had completed safe shutdown system modifications should (Footnote 70 Continued from Previous Page ) consequences for each of 91 sites. It woul d appear that, for each of the 6 7 plants considered in NUREG-1109, the corr esponding offsite consequence dat~ in NUREG/CR- 2723 was used. 71/ NUREG- 1109 at 11. 72/ Octobe r 27, 1980 Commission Meeting Transcript at 20.

not be affected by this feature of the rule. Consequently, the Commission took the unusual step of adding an exemption process _to its fire protection regulations at 10 C.F.R. §50.48 to allow utilities to receive credit for the fire protection safety added to their plants.73/ With Appendix R, a rule was sought to resolve issues that may otherwise be closed out in individual dockets. In so

• doing, most plants were expected to do little since their fire protection features were already reviewed and accepted by the Staff. It was believed that no additional compliance costs would be incurred by these plants.74/ But the impact of new regulatory requirements cannot always be isolated to the weaker plants. In the case of Appendix R, contrary to. this assumpti on , Section III.G disrupted licensee activities not at 20 plants, but at almost 90 % of all operating plants with most plants expend i ng millions of dollars to comply.75/

III.B The Proposed Rule Will Not Achieve a Consistent or Efficient Resolution of the Station Blackout Issues A ma j or factor in the current proposal is the central importance of coping as a means of reducing the potential core damage frequency due to station blackout. The Notice of Proposed Rulemaking states that the expected freque ncy of core 2._~I See 45 Fed . Reg. 766 02 (Novembt:?r 19 , 1980) . 74/ Enclosu re B to SECY 80-88, February 13 , 19 80. J5/ See §IV.B below.

damage from station blackout could be maintained near or below 10 -5 per reactor-year provided the plant is designed to cope for a specified duration. 51 Fed. Reg. 9830, at col. 3. This goal apparently is a major reason for the proposal that coping become a design feature for nuclear power plants. Yet, the analysis does not support the need for the rule for several reasons:76/ (1) More than half the generic combinations of

• (2) loss-of-offsite power susceptibility and diesel generator reliabili ties do not require 4

10 2s more hours of plant coping to meet the goal; . Four combinations of generic features do not require plant coping at all; and (3) The expected frequency of station blackouts lasting 4 hotirs or longer is an order of magnitude less than the total expected blackout fre quency (i.e., plant coping for these durations represents only a limited amount of risk reduction compared to the total risk).77/ If the proposed rule is implemented, the equipment

• qualification issue will potentially be reopened at many plants by extending the number of plant areas which may be deemed to be subject to a "harsh" environment. See §III.A above. In order for plants to comply with the proposed station blackout rule, a new program of qualifications and, possibly, 76/ See NUGSB 85-005, supra n. 16.

77/ In this regard, §II.C above addressed th e additional comments of Commissioners Robe ~ts a n~ ~~~ h, pertaining to arguments that the proposed rule would at best result in a small reduction to an already small risk.

requalification of existing equipment to the station blackout environment could be necessary . While some equipment may be able to take credit for existing qualification, equipment in .. the new areas will be introduc ed to the program. The effect of this action would be to extend the schedule for compliance with both the station blackout and equipment qualification rules beyond current expectations. The cost of such compliance activities could also be significant and was not considered in the regulatory analysis of the rule. See §IV.B below. The most practical approach to this issue is to reduce the likelihood of a station blackout. The dominant factors which contribute to this potential have been identified: loss-of-offsite-power and the early availability of emergency AC power. As noted in NUREG-1032, improvements in both areas are effective in reducing the potential for core damage to 10-S per reactor-year or less .

• Additional redundancy in the emergency AC power system may be provided by reducing the number of diesels deemed necessary for station blackout or by crediting the availability of another backup power source. The impact of such improvements is a substantial reduction in the likelihood of a station blackout. As the analysis in NUREG-1032 demonstrates, such measures could easily achieve the factor of 2 improvement in core damage frequency sought by the rulema king with a design basis coping capability. In fact, as previously discussed, it

                              -  55 -

is more likely that greater improvements could be realized in this fashion . Another concern with the proposed rule is its potential to divert licensee resources away from improved reliability and towards the more limited improvements achievable through compensatory measures, i.e, coping. Moreover, with a regulatory focus on coping, licensees receive effectively no

• credit under the proposed rule for actions taken to improve AC power system availability. Such improvements offer substantially more safety benefits than the proposed rule, and with earlier results .

J

IV. The Proposed Rule Does Not Meet the Back fit Rule Standard I IV.A Introduction Inc luded in the Notice of Proposed Rulemaking on station blackout is the backfit analysis performed pursuant to the NRC's backfitting rule, 10 C.F.R. §50.109. 51 Fed. Reg. 9829, ~t 9833-35. The backfitting rule mandates that the NRC Staff demonstrate by analysis, and the Commission

• find prior to the imposition of a backfit, that the proposed backfit will result in :

a substantial increase in the overall protection of the public health and safety or the common defense and security . . . and that the direct and indirect costs of implementation for that facility are justified

• in view of this increased protection. [10 C.F.R. 50.109(a)(3)]

In order to make this overall finding, Commission regulations require that an analysis be performed consistent with 10 C.F.R. §50.109(c). 10 C.F.R. §50.109(c) lists nine (9) specific factors which are to be addressed as appropriate. In order to address the requirements of 10 C.F.R.

   §50.109(c)(l-9), NUREG-1109, "Regulatory Analysis For The

... Resolution of Unresolved Safety Issue A-44, Station Blackout" was prepared.78/ This document est imate s the risk reduction to 78/ The backfit analysis also refers to (1) proba bl istic risk assessment studies, (2) common cause failures, (3) estimated frequency of core damage and (4) European expe rience . Items (1), (3) and (4) are addresse.d in §II above. With respect to common cause failures, the (Footnot e 78 Continued on Next Page

be 80,000 person-rems and asserts that the total cost of compliance would be about $40 million. On this basis, an overall value-impact ratio of about 2,000 person-rems averted per million dollars is reached.79/ A review of the backfit analysis and docume n ts reference d therein, particularly NUREG-1109, shows that these backfit (Footnote 78 Continued from Previous Page)

• following is presented.

The potential for station blackout is linked to the likelihood of coincidental failure of all emergency diesel generators (EDG). Such EDG failures could occur independent from each other, or as a result of a *common cause. The potential for common cause failures provides an upper limit to EOG reliability and the availability of the emergency AC power system. See generally NUREG/CR-2099 (1982). The effects of common cause failures need to be considered in viewing station blackout risk and the value of improved AC power reliability. The state of knowl edge today is sufficiently advanced to give explicit consi d eration to these effects in availability estimates. The kinds of remedies which are effective in reducing the impacts of common cause failure are also known and include greater diversity in design and maintenance. It should be e mphasized that the contribution of common cause failure to overall system unavailability is low. Typical values fqr dependent failures cited in the literature are on the order of 1-2 % of all components with some estimates ranging up to 5%. Estimates of the value of improving EOG reliability relative to station blackout risk, which are discussed in this document, consider the common cause contribution as a limiting parameter. It is believed that the common cause failure issue is both limited in scope and manageable. 2,000 person-rems per million dollars equates to $50Q per person- rem . The Commission regulations state that additional measures shall be taken if a favorable cost-benefit r atio [can] effect reductions in dose to the population." 10 C.F.R. Part 50, Appendix I, Sec. II.O. The Commission has determined that the value $1,000 per person-rem "shall be used in this cost-benefit analysis." Id.

analysis materials fall short with regard to the consideration of the specific matters set forth in 10 C.F.R. §50.109(c). Specifically, the backfit analysis can be characterized as follows:

1. Installation and Continuing Costs Associated With the Backfit Have Been Underestimated (§IV.B below);

2. Potential Impacts on Radiological Exposure of Facility Employees Should Be Further Addressed (§IV.C below);

• 3.

4. The Relationship to Proposed and Existing Regulatory Requirements Should Be Considered Further (§IV.D below); Potential Impacts of Differences in Facility, Type, Design or Age Should Be Considered Further (§IV.E below); and

5. The Reduction in Risk from Offsite Releases to the Public Has Been Overestimated (§IV.F below).

There is insufficient basis for the Commission to find that the backfitting standard has been satisfied. For this reason, the proposed rule should not be promulgated .

• IV.B Installation And Continuing Costs Associated With The Backfit Have Been Underestimated Section 50.109(c)(5) requires the consideration of installation ~nd continuing costs associated with its proposed backfit, including the cost of facility downtime or the cost of construction delay. The analysis in connection with this point is incorrect; the direct and indirect costs of the proposed backfit are underestimated. When the underestimated costs of the proposed backfit are considered

in light of the overestimated benefits cited as justification for the proposed blackout rule, the findings i required under §50.109(c)(5) to support the proposed rule cannot be made. The direct and indirect costs of the proposed backfit are underestimated in two significant respects. First, the number of plants which will be required to make

• modifications as a result of the rule is underestimated.

Second, with respect to modifications and coping, the costs that would be incurred to satisfy the proposed rule are not accurately identified. These errors become even more significant when viewed against the past experience with regulatory projections of the costs to licensees of backfits. These items are addressed in turn below. Number of Plants Affected Is Larger Than Assumed. The technical record has consistently claimed that most plants already meet the requirements of the proposed rule. For example, NUREG/CR-3226 suggests that most plants can readily provide the plant features ( ~ . , condensate, battery depletion, reactor coolant system isolation, loss of HVAC effects, ~tc.)~~/ necessary to cope with blackouts of 4-hours. This theme is developed further in NUREG-1032 where it is noted that all plants have the ability to remove decay heat for some period. NUREG-1109, Table 6, states that only 10-15 reactors will need to make modifications. 80/ See generalJ:.y NUREG/CR-3226, Appendices D and E.

As discussed in §III.A above, the regulatory analysis should take into consideration the potential that subsequent regulatory interpretations of the rule will evolve as a result of review of the coping analyses {which will be required for all plants, regardless of station blackout risk for each plant). The various interpretations could lead to a significant increase in the nu mber of plants ultimately affected by the proposed station blackout rule. This is

• best demonstrated by the impact that regulatory interpretations of licensee compliance has had on the number of plants required to make modifications pursuant to the safe shutdown requirements for Appendix R, the TMI Action Plan, and equipment qualification guidelines.

In fact, it is difficult to identify any backfits involving licensee analysis that did not eventually lead to additional facility modifications. This is not to imply that additional modifications are never necessary. The point simply is that a sizeable number of unanticipated facility modifications inevitably occur, and, therefore, should be considered in the analysis of the backfit costs. Activities at Individual Plants Reflect Higher Costs Than Projected. The above mentioned cost estimate does not take into account all of the costs associated with implementing the draft rule at par ti cular plants . There are two issues of concern: (1) underestimation of the cost

of preparing a coping analysis; and (2) errors in the NRC contractor report . First, NUREG- 1109 projects that a coping analysis should cost each licensee approximately $150,000. Based on actual cases where such analyses have been completed, or sufficiently scoped, this projection is low by factors of 5-

20. Second, the cost analysis involves a contractor report, NUREG/CR-3840, which is inade q u ate in the following respects:

(1) The use of labor productivity figures are higher than those provided in "Handb o o k for Cost Estimating," NUREG/CR-3971; (2) The us e of unit rates for certain materials that are lower than those rates historically experienced in nuclear con struction; (3) The failure to consider design criteria impo sed by other Commission requirements ( ~ , quality assurance, separation, equipment qualification, etc.);

• (4)

(5) The underestimation by a factor of two or more of project management, engineering, and QA/QC costs using historical data; and, Typographical and mathematical errors in the source document. Based on the above, it is reasonable to conclude that the cost of implementing any single hardware-related coping feature discussed in NUREG/9R-3840 will cost between 2 and 5 times more than estima ted . To be sure, this prediction i s not intended to be exact. Rather, it is intended as a more re asonable estimate regarding i mplementation co sts of the station blackout rule.

Costs for Demonstrations of Equipment Qualification and/or Equipment Replacement Have Not Been Considered. The cost analysis should have considered the costs associated with demonstrations that equipment used to cope with ' a station blackout is qualified to function in the environment associated with the event.Bl/ First, should the analysis require certain equipment to function in a "harsh environment", 10 C.F.R.

   §50.49 would provide the requirements for demonstration of capability of the equipment to function in such environment .

As mentioned above (§III.A), if the rule is interpreted to require full compliance with 10 C.F.R. §50.49 regarding the demonstration of equipment environmental qualification for substantial amounts of equipment not previously subject to

   §50.49, significant costs and resource consumption will result just for this demonstration.      Based upon industry's experience, such demonstrations entail significant costs.

Second, should the coping analysis require the equipment to be demonstrated to function in a "mild environment," the lack of such demonstration" could result in changeout of significant amounts of equipment which is otherwise functionally sound. The costs of additional tests and analysis } for equipment which may lack sufficient documentation to demonstrate operability in mild environments, and changeout of See the d r aft r egulato ry guide pub l ished with the proposed rule (at 7), which sets forth guidance regarding evaluation of equipment operability and reliability during station blackout events.

mild environment equipment should have been considered in the backfit analysis. Therefore, the costs of the proposed rule are underest i mated. Concern with the underestimation of costs is deep-rooted. The value-impact analysis supporting the fire protection rulemaking stated: Most licensees already comply with the requirements of the proposed rule and they

• will incur no additional costs associated with this rule. Those licensees who have not yet complied with the NRC fire protection guidelines will incur some additional capital and operating costs . . . .

Enclosure B to SECY-80-88, "Fire Protection Actions," February 13, 1980, at 5. Notwithstanding those optimistic assurances, the actua l costs incurred by individual utilities have ranged from appr oxi mately $5 ,300,000 to $20,000,000 per plant site.82/ The value-impact analysis accompanying the proposed

• emergency planning rule stated as follows regarding the typical costs for State and local government programs to achieve adequacy in radiological emergency response plans for a 10-mile Emergency Planning Zone:

J For a State, the initi al costs of planning, exercises, training and resources (communication and radiation .. monitoring instrumentation) will typically total about $240,000 with associated annual updating costs of about

              $44 ,000. For local govern~~nts, th~

82/ Based on i n formal contacts with nuclear utility members of NUGSBO.

initial costs typically

• total about

           $120,000 (four jurisdictions) with annual updating costs of about $30,000. The typical total costs to State and local governments to obtain an NRC finding of adequacy in their emergency response plans would be about $360,000 initial costs plus $74,000 in annual updating costs.

NUREG-0685, "Environmental Assessment for Effective Changes to 10 C.F.R. Part 50 and Appendix E to 10 C.F.R. Part 50; Emergency Planning Requirements for Nuclear Power Plants,"

• (Augus t 1980) at 5, 7. In addition, the one-time cost of

 $500,000 per facility was estimated for the public notification system. In actuality, initial State and local government costs have ranged from $170,000 to $1,714,500.

Annual state and local updating costs have ranged from

 $135,000 to $740,000. Initial utility costs for the public notification system ranged from $400,000 to $2,400,000.      In addition, other one-time utility expenses have in certain instances exceeded $2,000,000. Annual utility costs range from $125,000 to $667,000.83/

Iv.c Potential Impacts on Radiological Exposure of y~cjlity Employees Should be Further Addressed Section 50.109(c)(4) requires the regulatory analysis to address the potential impact of proposed backfits on the radiological exposure of facility employees. NUREG- 1109 states (at Table 8, n. 2): 83/ Cost figures are based on informal contacts with NUGSBO members and are based on costs per site.

No significant increase in occupational exposure is expected from operation and

• maintenance or implementing the
• recommendations proposed in this resolution.

Equipment additions and modifications contemplated do not require significant work in or around the reactor coolant system and therefore would not be expected to result in significant radiation exposure. This statement is questioned in two respects. First, licensee actions required to implement a design basis coping capability of 4- to 8-hours will probably encompass facility

• modifications in and around the reactor coolant system (or other contaminated systems). For example, valve operators may require replacement, new instrumentation may need to be installed and surveillance of station blackout systems will have to be implemented. All of these activities are normally associated with hardware backfits of the type that have contributed to the rapid growth in incidental onsite exposure experienced since the early 1970s .
• According to a recently published industry report, prepared by the National Environment Studies Project of the Atomic Industrial Forum,84/ during t~e period 1979 through 1983 forty percent of the total occupational exposure at U.S. light water reactors was attributable to NRC-initiated multi-plant

   ~~/ AIF/NESP-033, "Occupa tional Radiation Exposure Implications of NRC-Initiated Multi-Plant Actions," March 1986.

actions.85/ This percentage represents approximately 99,000 person-rem of collective exposure.86/ Second, one of the estimated financial costs for licensees to comply with the draft rule is the resolution of the reactor coolant pump seal integrity issue.87/ Notwithstanding the above discussion (§II.B.2), if licensees have to replace or modify their existing seals in order to

• resolve the reactor coolant pump seal integrity issue, workers will be exposed to limited occupational doses. Thus, if the the financial impacts of seal resolution are to be considered, the radiological impacts of this backfit should also be considered .

85/ Id. at v i . See NUREG-0748 for a listing of so-called

    -   ~ulti-Plant Actions."

A major contributor (8.8 % of the total, or 8712 person-rem) was licensee actions to achieve complia ~c e with NRC fire protection requirements, yet the re~ula t ory analysis for those requirements did not account for any occupational exposure amounts. See Enclosu re B to SBCY-80-88, "Value/Impact Assessment of P ropo se d P'ire P r otection Rule," February 13, 1980. 87/ NUREG-1109 at Table 6, n. 5.

IV.D The Relationship to Proposed and Existing Regulatory Requirements Should Be Considered Further The backfitting rule requires consideration of the relationship of a proposed backfit with other proposed and existi~g regulatory requirements. 10 C.F.R. §50.109(c)(6). Such a relationship is acknowledged between station blackout and the two issues of diesel ge~erator reliability and reactor coolant pump seal integrity. However, these issues are being

• addressed through other regulatory initiatives. Even if concerns on these issues were substantiated(~., see §II.B.2 above regarding clarifications of the RCP seal issue), when these separate initiatives are completed, the need for the proposed station blackout rule would appear to be further diminished.

Die sel generator reliability is central to any resolution of the station blackout issue. While the proposed rule

• reflects minimum standards in this regard, there is no account for the fact that, to the extent that there is an improvement in the diesel generator reliability, the benefits claimed from the proposed rule (and, thus, the need for a rule at all) are dissipated.88/

As mentioned above, the benefits of the proposed rule are also strongly dependent on the resolutioP of the reactor 88/ Initiativ~s advanced by industry are designed to improve diesel generator performance.

coolant pump seal integrity issue . . When the proposed rule was drafted, current data s uggested that certain reactor coolant pump seals may be susceptible to catastrophic failure should cooling be lost. Thus, the proposed rule places strong emphasis on coping and the value of seal injection. Howeve r, as discussed in §II.B.2 a bove, signific ant new information has recently become available. Based on this information, it is now apparent that the risk of catastrophic seal failure is significantly less than previously thought

• In short, the backfitting rule was intended to ensure that proposed modifications be related to other ongoing regulatory activities in order to avoid unnecessary expenditures of NRC and licensee resources. The failure in the backfit analysis to consider realistically both diesel generator reliability improvements and recent developments regarding seal integrity, results in failure to take into account these ongoing regulatory activities .

IV.E Potential Impacts of Differences in Facility Type, Design or Age Should Be Considered Further Section 50.109(c)(8) requires the consideration of the potential impact of differences in facility type, design or age on the relevancy and practicality of the proposed backfit. Table 6 in NUREG 1109 indicates that betwee n 42 and 57 plants need not do anything to improve diesel generator r eliability or coping capability in order to c ompl y wi th the rule. Thus, the benefits of the proposed rul e will go to the 15 plants needing

to upgrade diesel generator reliability or to the 10 plants needing to increase coping capabilities. It therefore follows that 42-57 plants will enjoy little or no benefits from this rule but each will spend at least $150,000 89/ to document their compliance. Clearly, for this population of plants, there will be significant costs with minimal corresponding benefits. Under these circumstances, where the risks associated with particular plants are not, in fact, spread evenly throughout the industry, it is believed that unique plant-specific factors were artificially and unrealistically ignored in order to justify a favorable cost-benefit ratio. Thus, compliance with Section 50.109(c)(6) has not been achieved. IV.F The Reduction In Risk From Offsite Releases To The Public Has Been Overestimated The backfitting rule requires the "potential change in the risk to the public from the accidental offsite release of radioactive material" be addressed. 10 C.F.R. §50.109(c)(3). As described above in §II.A, the representation made in NUREG-1109 was that the estimated total risk reduction achievable from the proposed resolution of USI A-44 is 80,000 person-rem, assuming an average remaining plant life of 25 years. This estimate represents the benefit for the p ropo sed rule. 89/ $150,UOO represents th~ best est i mate for a coping study. NUREG-1109, T~hle 6. As stated ahove, this figu re could have brcn underestimated hy a factor of 5-20.

It is not readily apparent how this figure was derived. NUREG-1109, Table 5 provides three examples of the reduction in frequency of core melt per reactor year which is attributed to the proposed rule. Table 4 identifies 3 groups of plants, totalling 67, which form the basis for the calculation of a total dose reduction of 80,000 person-rems attributed to the proposed rule. However, there is no disclosure whether a relationship exists between the three frequencies in Table 5

• and three groups of plants in Table 4.

Assuming some rational relationship exists between Tables 4 and 5, it appears nevertheless that the risk of a station blackout event is overestimated and, therefore, the expected benefits of the proposed rule (in terms of person-rem averted) are also overestimated. This is apparent because a number of factors were not considered which (1) reduce the probability of a station blackout(~ §IV.F.l below) and (2) reduce the consequences of station blackout (see §IV.F.2 below) . IV.F.1 Factors Reducing the Probability of Station Blackout Require Further Attention More full account in the backfit analysis should be made for a number of factors that, as a practical matter, reduce the risk of a station blackout. These factors are (1) emergency diesel generator configurations; (2) emergency procedures; and (3) additional power sources. When these factors are taken into account, the probability of an off-site release in connection with a station blackout is much less significant than that postulated by the staff.

Diesel _Generator Confjguration. The first factor is present in the category of plants which pose the greatest risk based on the conclusion in Table 5: two unit plants sharing 2 of 3 (2 / 3) emergency diesel generators to respond to a loss of coolant accident. It appears from Tables 2 and 4 in NUREG-1109, that 14 plants at 7 sites fall in this category (i.e., Group C). Assuming that the proposed rule is implemented and that all 14 plants now have a 2 hour station blackout coping capability, NUREG-1109 indicates that an estimated risk 5 reduction in core damage frequency of 14 x 10- per reactor-

                -1 year (RY)      , will be achieved at each plant.             By comparison ,

the other two plant categories will achieve reductions in core damage frequency of 1.3 x 10-S (RY)-l and 0.1 x 10-S (RY)- 1 , respectiv e ly. While these 14 "higher risk" plants represent only 21 % of the facilities considered in NUREG-1109 Table 5, by implication approximately 78 ~ of the expected offsite consequenc e s of station blackout which can be averted by the rule is attributed to these sites . App ar ently this concern is based ultimately on the observation that these two-unit plants do not have as much redundancy in the emergency AC power system as do most other plants. In a station blackout event the magnitude o f electrical loads necessary t o p rov ide cor e ~oo ling is

• . significantly less than tha t r e q uir e d und e r des i gn b asis accident co ndi tio n s . Indeed, such ~~c i rl ~ ~ts are s p ecif ica lly e x cluded fro m c op i ng co nsid era tion in the proposed rule. As

previously discussed, in several of these 2/3, two-unit plants, it is possible to provide AC power to both units in a station blackout while using only a single emergency diesel generator, effectively making the emergency diesel configuration 1/3. This can be done by shedding non-essential loads and/or by cross-co~necting the diesel generators servicing each unit in a two-unit plant .

• The impact of this effective 1/3 configuration on the regulatory analysis is dramatic. The reliability of a 1/3 system is approximately 4.5 to as much as 13 times greater than a 2/3 system, assuming individual component reliabilities of

    .95-.99 per demand. Most of these plants are operating their diesels nt reliabilities of above 0.97 per demand. Crediting these improvements in AC power reliability significantly reduces the station blackout risk at the very sites the regulatory analysis suggest pose the greatest risk. When the number of effective 1/3 configurations is taken into account,

• the risk of station blackout is not as great as that suggested in the regulatory analysis.

Emergency Procedures. In assessing the risk of off-site . power loss it was concluded that severe weather was a prime t contributor, particularly with respect to event duration. (Indeed, severe weather is critical to determining minimum coping durations and underlies the use of c oping to resolve station blackout.) This conclusion does not give full credit

for the potential benefits of onsite emergency procedures

   . (i.e., as requested by Generic Letter 81-04, "Emergency Procedures and Training for Station Blackout Events," February 25, 1981) which are designed to address extremely severe weather conditions.90/     The backfit analysis does not reflect the procedures licensees have in place to reduce power, verify emergency equipment operability, and shutdown their units in the event of severe and extremely severe weather, such as hurricanes. The effectiveness of these procedures was most recently demonstrated by Hurricane Gloria.      In every instance where the storm posed a risk, lice nsees took appropriate action to safely respond to the severe weather.91/

Additional Power Sources. The assessment of loss of on-site power is not realistic due to its undue focus on emergency diesel generator reliability. Other alternate sources are present either at dr nearby nuclear power plants, such as fire protection diesel generators, gas turbines, black-start diesel

• generators and steam driven pumps. These power sources can be relied upon in an emergency, thereby enhancing onsite power and decreasing the frequency and consequences of loss of onsite power. Information submitted by licensees in response to a

. request based on 10 C.F.R. §50.54(f) for example, could be used l 90/ In addition to such procedures, the initiative advanced by industry provides for severe weather ?rocedures . 91/ See discussion of Hurricane Gloria in §II.B.3, supr~. See

         §IV.F.2 below for a discussion of emergency planning factors designed to reduce the consequences of offsite releases.

to demonstrate the availability of these additional power sources.92/ IV.F.2 Factors Reducing The Consequences of Offsite Releases Require Further Attention Notwithstanding the above discussion of the failure to account fully for flexibility in emergency AC power system designs and other emergency procedures, the consequences of offsite releases that would result from a station blackout event are overestimated. First, inappropriate assumptions were made which overestimated the expected impact of early containment failure on station blackout risk. Second, information to the effect that the source term can be reduced by at least two orders of magnitude should have been considered more explicitly. Third, for the reason mentioned above (§II .A. 2), the SSTl release category is inappropriate as used by the Staff to represent offsite consequences of accident sequences involving a station blackout. Fourth, the

• contribution that implementation of the Commission's emergency planning regulations makes to the reduction of population exposure has been ignored. These items are considered in turn below .

.I A utility demonstration would identify the power source and provide assurance of its availability. Such assurance could entail description of maintenance, testing, reliability, and procedures to avoid common cause failure.

                                         *rnappropriateness of Containment Performance Assumption.
. The expected impact of early containment failure on station blackout risk has been overestimated. As explained fully in l
   §II.A.2, the siting source term, SSTl, has been used improperly to model the offsite consequences of a station blackout accident. As explained above, SSTl postulates a containment failure in 1.5 hours after the onset of core damage. NUREG/CR-2239 at 2-13. To be sure, the expected SSTl release is discounted by a factor of 3 to account for the fact that the SSTl was never intended for use with st~tion blackout events, and the importance of accounting for the effect of containment integrity in delaying and mitigating the release is acknowledged.J3/   The effect of such delay and mitigation is to reduce both the energy of the assumed release and its contents.

However, NUREG-1109 provides no specific information as to how long containments would be assumed to remain intact in a station blackout event. Yet this is a critical factor because it is directly related to person-rem exposure.94/ The

• subjective use of a reduction factor of 3 without any relation to containment integrity is not justified. See §II.A.2 above.

NUREG-1032 does provide what is characterized as some "insights" on this question. In §7.0 of NUREG-1032, IDCOR i containment failure times are compared with NRC Staff times in most instances. It is stated (at 7-15, 17) that recent 93/ NUREG-1109 at 8. 94/ NUREG-1032 at 7-15.

estimates of containment performance such as those produced by ... IDCOR "may be cause for revision" of prior published NRC results. However, no attempt is made in NUREG-1109 to l reconcile the large disparity between prior NRC estimates for containment failure - times and those of IDCOR. These "insights" show that the value of containment has been underestimated in mitigating reactor accidents. For example, *it is stated that large dry containments can fail in as little as 10 hours. The IDCOR study suggests that the failure time for these containments can be at least 18 hours. NUREG-1032 indicates (at Figure A.1) that the likelihood of offsite AC power being restored in 10 hours or less approaches

1. Restoring AC power will return to service those containment cooling syste ms which will assure containment integrity.

Therefore, the NUR EG-1 032 ana lysi s demonstrates that the potentia l for containment failure associated with failure to restore AC power should be very small, regardless of whether

• containment integrity is maintained for 10 hours, as the Staff suggests, or 18 hours, as IDCOR found. Accordingly, the expected impact of containment failure is overestimated in the regulatory analysis.

t New Source Term Data Should Be Relied Upon . It was improper to base the value-impact ratio sol~ly on the SST! release category. During a March 2S, lSJS Commission meeting the Commission was advised (Tran sc ript at 78 ) of the late st

schedule for completion of source term related activities. In

\

• addition, a draft of ~UREG-1150 is scheduled to be issued in the late summer of 1986. This NUREG will, among other things, provide the results of accident analyses using the new source term assumptions. This information will provide the basis for the Commiss ion to issue new direction concerning the risk of severe reactor accidents. Most of the analyses in this regard have been completed and are currently available .
• In fact, during the March 26 Commission meeting, Dr.

David Torgenson, from the OECD Committee on the Safety of Nuclear Installation's Special Task Force on Source Terms, told the Commis sion: The thing I want to point out is the reduction in the source term. This is the old technology here for Surry in Wash 1400. This is the new technology that is coming out of BMI 2104. You can see that the source term has been reduced by more than two orders of magnitude.95/

• Zoltan Rosztoczy also stated:

The programs in the Office of Nuclear Reactor Regulation . . . have to be closely coordinated and coupled with the unresolved and generic safety issues such . . . as station blackout . . . -~/ We agree. Under these circumstances, this potential I new direction should have been accounted for in the regulatory analysis. Section 7.0 of NUREG-1032, states (at 95/ Id. at 10. 96/ Id. at 71.

7-18) that the "reader is cautioned that ongoing research could cause substantial revision" in fission product release fractions utilized in the analysis. However, an explanation

• should have been provided in NUREG-1109 as to the potential effect such substantial revisions will have on the validity of the regulatory analysis accompanying the proposed rule.

The Commission should not act on any final rule until this information is available.97/

• Inappropriateness of Fission Product Inventory Assumption. As discussed above in §II.A.2, the use of SSTl as a generic assumption applicable for all plant containment types when determining the avoided person-rems is inappropriate. A better approach would be to select release categories on the basis of the different containment types, failure modes and station blackout accident sequences.98/

For exampl e , the characteristics of an SSTl are containment failure and release time at 1.5 hours, with release duration

• of 2 hours.22/ If the assumed energy release rate is high, wider dispersion of radioactive materials and a higher curie 97/ NUREG-1150, when published, will be reviewed and comments will be provided as appropriate. Indeed, an argument could be made that the station blackout rulemaking should be deferred until at least such time

' as comments can be made on NUREG-1150. 98 / NUREG-1032 addressPs these factors (at 7- 19). However, NUREG-1109 utili z es the SSTl rel ~asP r: a~ ~g ory (reduced by 1/ 3) on a gener ic basis for ~ ~l ~l 2~ ts to discount SSTl rel0ases (at SO miles) f or station blackout events. 99 / NUREG/C R-22 39 at 2-13.

content *of that material which is released will occur.100/ These characteristics are in direct contrast to the

• risk-dominant accident sequences associated with station
• blackout.101/ Specifically, these accident sequences can all be characterized -- not by the SST! containment failure mode, but by a containment failure later than 1-1/2 hours, a lower energy release, and a significantly reduced curie content in the released materials.102/ Moreover, if SST! is used when SST2 assumptions are more appropriate, an error of more than 2 orders of magnitude can be injected into the averted person-rem calculation and, thus, the risk-reduction benefits of implementing the proposed rule would be overestimated.

For example, if an SST2 or SST3 release had been assumed at the outset instead of SSTl, the estimated offsite release associated with station blackout could have been lower by as much as 4 orders of magnitude. The practical

• impact of doing so would have been to reduce averted person-100/ Id. at 2-73.

101/ As explained in §II.A.2 above, while early release could be postulated with a mechanism such as steam explosion or direct heat!9g, the frequency of such a I release is so small (<10 per RY) that it does not contribute to the risk from station blackout, (i.e., frequency x consequence). The risk dominant sequence is the long term overpressurization of containment, 24 hours to 72 hours after core melt. 102/ In this regard, it is not clear from the analysis in NUREG-1109 what energy release rates were used in conjunction with the SSTl assumption.

rems -- not merely by the value of one-third cited in NUREG- . 1032, but by a factor of 100 or more -- as a result of the highly non-linear relationship among the various categories

• of releases, SSTl-SSTS. See §II.A.2 above.

This is not to imply that the SST2 category, or a modified SST2 category, should have been relied upon in all

 . cases instead of an appropriately modified SSTl. However, a

• regulatory analysis should provide a sound, articulated technical basis for its selection of key assumptions. In this case, instead of attempting to fit the station blackout issue into a preconceived fission product release category, such as SSTl, the most appropriate technique would have been to develop a mechanistic model for calculating the potential for offsite consequences due to station blackout accident sequence, including expected containment failure mode and failure times, and to use that model in the regulatory analysis supporting the proposed blackout rule .
• Alternatively, a more thorough offsite consequences analysis could have been provided, which would reconcile the predicted offsite release with the various SST categories selected and match the release characteristics (i.e., plant effects) of a station blackout accident with one of the pre-defined source terms, such as SST2. Instead , the analysis in NUREGs- 1032 and -1109 apparentl y have vsed the most conservative source term yet identified without presenting an adequate technical basis for the choice.

Accordingly, the averted person-rems to be derived from the proposed rule are overestimated and the underlying basis for key assumptions in identifying these averted person-rems is not established. Contributions of Emergency Plans to Dose Reduction. The estimation of 80,000 averted person-rems should have accounted for the dose reductions re~ulting from compliance

• with the Commission's emergency planning regulations.

C.F.R. §50.47 and 10 C.F.R. Part 50, Appendix E. The Statement of Considerations accompanying the emergency planning rule states that such rule was promulgated to 10 protect the public health and safety. 45 Fed. Reg. 55402 (1980). Furthermore, NUREG-0396,103/ an authoritative source document in emergency planning, specifically states (at 5) that: [T]he objective of emergency response plans should be to provide dose savings for a spectrum of accidents that could produce offsite doses . . . . Accordingly, it can be seen, that in complying with the emergency planning regulations, a "dose savings" results. This "dose savings" should have been deducted from any estimate of person-rem averted as a result of implementation of the proposed rule.

      "Planning Basis For The Development of state and Local Government Radiological Emergency Response Plans In Support of Light water Nuclear Power Plants." NUREG-0396 (1978).

IV.G Conclusion A number of problems have been identified in the backfit analysis developed in support of the proposed rule. Resolving only two of the problems significantly alters the value/ impact ratio, raising a question as to whether the proposed rule meets the backfitting standard. For example, Table 5 in NUREG-1109 implies that 78% of the 80,000 person-

• rem are at the 14 plants with a 2/3 emergency diesel generator configuration. That risk (6?,400 person-rem) diminishes by a factor of from about 4 to 13 times if all plants in this ca t egory can be converted to a 1/3 configur ati on. Us i ng a 4.5 factor as representative of the risk-reduction value of these conversions, the 80,000 man-rem proje c ted savings by the Staff could be reduced by as much as 31 , 500 pe r son-rem.

In add i tion, the total projected costs of implementing the proposed backfi ~ amounts to in excess of $80 million. Thus, the va l ue/ impact rat i o of the draft rule, even without discount i ng f or containment performance and source terms, could be c l oser to $2,500 per person-rem averted. Proper treatmen t of containme nt perfo~mance and the .V use of mo r e re al i st ic sou r ce terms, c oupled wi th considera tion of mo re r ea li s tic co sts , should re su lt i n a value/ impact rat i o of at least $ 50,000 pe r pe r son- r em

averted and could well exceed this amount. On these bases, it is clear that the backfitting rule has not been satisfied and thus the proposed rule should not be implemented. In view of these errors, adequate justification has not been p r ov ided i n support of the position that the backfitting standard has been satisfied. Thus, the Commission should not proceed with this rulemaking . v

V. Objections to the Proposed Rule

These comments can be summarized in four key arguments:

(1 ) The bulk of perceived station blackout risk is concentrated at a very small number of sites having distinctly identifiable design features -- thus, generic rulemaking is not the best mechanism to address an issue that involves a small percentage of

• (2) the industry; Even if the Commission concludes a generic rule is appropriate, the technical basis does not support the rule; (3) Even if the underlying technical basis did support the rule, the proposed rule itself should be reevaluated as a regulatory tool for it is likely to fall short of the goal of consistent and efficient resolution of the
• (4) station blackout concern; and, Notwithstanding the above, the proposed rule does not meet the standards of the Commission's backfit rule --

either in the net value of the anticipated modifications or in the completeness o f the analysis. Any one of the above conclus ion s i s suf ficient to preclude enactment of the rule currently before the Commission.

If there is particular concern, then any regulatory action considered must be directed at the concern. In this instance, an issue has been raised as to the availability of AC power sources and the potential that station blackout poses an unacceptable risk at an unknown number of plants. Yet, the proposed rule seeks to identify the coping capability of operating plants and impose minimum durations which plants must provide as part of their design basis. It appears that the proposed solution does not address the stated problem in an effective manner. A rule is not needed to identify whether or not station blackout risk is unacceptable at a particular plant. Other mechanisms are available ~or obtaining this information without the need f or the Commission to utilize the rulemaking process. Similarly, concern about the availability of AC power sources suggests that the best solution to the problem is to enhance that reliability.104/

• The effectiveness of a generic rulemaking in addressing so isolated an issue as station blackout is a concern.

However, it is recognized that the Commission may believe that rulemaking satisfies other imperatives which transcend the ability to demonstrate public health and s~fety benefits across all of industry. Should such imperatives exist within the 104/ Initiatives advanced by industry are designed to enhance reliability.

Commission, then the proposed rule should focus on the concerns raised.

   "       If the Commission believes a rule is necessary, guidelines should be prepared concerning what constitutes acceptable emergency and non-emergency AC power systems for preventing or mitigating station blackout events. Without such guidance, a rule which requires AC power-independent coping is subjective, difficult to implement and of questionable value with respect to providing real safety margins.

Such u shift in focus from compensatory measures to problem resolution would entail modifying several aspects of the proposed rule. To this end, the kinds of changes we would suggest include: (1) expanding the focus of the rule and draft regulatory guide to include explicit and equally weighted

• provision for crediting onsite and offsite backup AC power sources or improvements to AC power system avail~bility in responding to the loss of normal and emergency AC power;

( 2) deleting the addition to the General 9esign Criterion \

 \,        of an AC-independent coping requirement and the associated coping analyses~_9-~

105/ The proposed rule would make GDC-17 unique among other (Footnote 105 Continued on Next Page

(3) eliminating the open-ended coping demonstration aspects of the regulatory guide and substituting a simpler closed-form coping capability checklist; ( 4) considering the potential ripple effects of new station blackout requirements, particularly those involving AC-independent coping, on other areas of nuclear regulation, such as the potential need for

• equipment qualification of station blackout coping equipment by the methods called for in 10 C.F.R.

         §50.49; and (5)   eliminating the need to consider severe weather as a dominant factor in establishing whether a plant is in a 4-hour or 8-hour coping duration category (i.e.,

NUREG-1109 at Tables 1 and 3) . L (Footnote 105 Continued from Previous Page) General Design Criteria, in that no other GDC explicitly negates the availability of AC power.

P. 0 . BOX 14000, JUNO BEACH, FL 33408 FLORIDA POWER & LIGHT COMPANY JUN 1 71986 L-86-261 The Secretary of the Commission Attention: Docketing and Service Branch U.S. Nuclear Regulatory Commission

• Washington, D.C. 20555

Dear Sir:

Attached are Florida Power & Light Company's (FPL) specific comments on the Nuclear Regulatory Commission's proposed rulemaking to resolve the station blackout issue (TAP-A-44). In addition to these specific comments, FPL supports and endorses comments from the Nuclear Utility Management and Resources Committee (NUMARC) and the Atomic Industrial Forum (AIF) on this rulemaking. In providing these comments, I want to stress FPL's continuing interest in NRC and industry efforts to achieve sound technical and regulatory resolution of this and other issues. Very truly yours,

• ~~~

Group Vice President Nuclear Energy COW /RWG/cvb Attachment RWGl/011/1 Acknowledge by r PEOPLE . . . SERVING PEOPLE

I us

r ATTACHMENT Current regulation in IO CFR SO, Appendix A, establishes requirements for the design and testing of onsite and offsite electric power systems. This regulation is also intended to reduce the probability of losing all AC power to an acceptable level, see General Design Criteria (GDC) IO, 17, 18. Existing regulations also explicitly require nuclear power plant systems, including emergency AC power systems, be designed with appropriate margin to assure that specified acceptable fuel design limits are not exceeded and containment integrity is maintained. In addition, GDC 17 specifically requires the Licensee "to minimize the probability of losing electric power from any remaining supplies as a result of or coincident with, the loss of power generated by the nuclear power unit, the loss of power from the transmission network, or loss of power from onsite electric power supplies" *

• We believe the above stated GDC provide the NRC with the regulatory vehicle to resolve unresolved issue A-44, "Station Blackout" without further rulemaking. The NRC staff, while evaluating station blackout in NUREG-1032 has apparently determined that: I) existing margins as defined in IO CFR 50, Appendix A, are no longer acceptable, and 2) there is no safety margin for emergency AC power systems that is acceptable to preventer mitigate a complete loss of all AC power at nuclear facilities (FPL disagrees with both assumptions). Subsequently, in developing the proposed rule on station blackout, the staff has dictated a "design fix" (i.e., AC-independent coping) to resolve the station blackout issue.

The NRC reliance on AC-independent coping should not be used as the "measuring stick" to define acceptable safety margin (or goals) to meet the intent of GDC I0, 17 and 18. Should the NRC staff believe that existing margins are not acceptable or that failure probabilities are too high, then the Staff should provide guidance on appropriate safety margins and methodologies which utilities can use as a basis to evaluate emergency AC systems. The Staff has not provided such guidance nor have they allowed appropriate credit for backup AC emergency systems to mitigate or prevent a complete loss of all AC power

• Historically, the staff has provided guidance on safety margins and methods to be used by utilities to determine if safety systems meet regulations. It would seem that the adequacy of normal and emergency AC power systems could best be addressed in th is same manner.

In summary, Florida Power & Light believes that current regulation as presented in IO CFR SO, Appendix A, already provides the regulatory vehicle to resolve the station blackout issue, thus, no new rule is necessary. In order to resolve this issue the NRC Staff should provide utilities with the following: I. Sufficient guidance on appropriate safety margins (or safety goals) to be used by utilities as an acceptable threshold to evaluate emergency AC systems; RWGl/011/2

2. stipulation of acceptable margins required in AC emergency systems (including backup sources) to mitigate or prevent a complete loss of all AC power; and

3. a methodology to evaluate emergency AC power systems with new safety margins, if new margins are necessary.

Florida Power & Light objects to using AC-independent power systems (a recommended NRC design fix) as the method to resolve the station blackout issue. FPL believes AC-independent coping is not an appropriate sole consideration to determine an acceptable safety margin:-Finally, FPL subscribes to a program which will increase AC emergency system reliability and establishment of safety margins for AC power systems which will reduce the probability of complete loss of all station AC power as provided in GDC 17

• RWGl/011/3

               ~i JJ,JI .

.0~PR-5o ~

(Sir/!_ $'? 29) ...,

               ~

MISSISSIPPI POWER & LIGHT COMPANY Helping Build Mississippi P . 0 . BOX 1640 . JACKSON. MISSISSIPPI 39215-1.64-0 June 16, 1986

0. D. KINGSLEY, JR.

VICE PRESIDENT

• NUCLEAR OPERATIONS U. S. Nuclear Regulatory Commission Office of Nuclear Reactor Regulation Washington, D. C. 20555 Attention: Docketing and Service Branch

Dear Sir:

SUBJECT:

Grand Gulf Nuclear Station Unit 1 Docket No. 50-416

• License No. NPF-29 Comment to NUREG-1109 AECM-86/0187 Mississippi Power & Light wishes to endorse the position of the Nuclear Utility Management and Resources Committee expressed in their comments to the proposed rule on station blackout (USI A-44). It is expected that Grand Gulf would have appreciable cost associated with analysis which would result in no required physical modifications and therefore no change in risk associated with a blackout. Additionally, Grand Gulf Nuclear Station has capabilities which have not been taken into consideration by the staff in its generic approach to station blackout. We believe that a more realistic individual approach to station blackout would have more meaningful results.

We wish to thank the commission for the opportunity to comment on this issue .

• ODK:bms cc: Mr. T. H. Cloninger (w/a) Mr. R. B. McGehee (w/a)

Mr. N. S. Reynolds (w/a) Mr. H. L. Thomas (w/o) Mr. R. C. Butcher (w/a) Mr. James M. Taylor, Director (w/a) Office of Inspection & Enforcement U. S. Nuclear Regulatory Commission Washington, D. C. 20555 Dr. J. Nelson Grace, Regional Administrator (w/a) U. S. Nuclear Regulatory Commission Region II 101 Marietta St., N. W., Suite 2900 Atlanta, Georgia 30323 JUN 17 1986 Acknowledged by cerd . .......... ...... . - J 16AECM86061203 - 1 Member Middle South Utilities System

       ~~ , t~lrl

( C.( ~q~ s~

       ?J1~1,

liail PWIUli lMIIIIIPR -5() ~ {_5/ F~ -;,tq LV f."*:. MARVIN LEWIS h~8~J ROOSEVELT BLVD. #61, c:"::: PHILA., PA 19152  :,

                                                                *.:,1.

Mr. A. M. Rubin U* S

• N* R* C*

Washington , D. C. 20555

Dear Mr . Rubin:

Please accept this letter on Station Blackout as my comments upan the NUREG 1109 andl032 . Both reports are completely deficient in that neither look at sabotage . The pylons whihc bring the wires carrying off site electr i c i ty into the plant are most vulnerable and they could easily all be brought down at the same time with one man laying time delay charges on each

• Then say you have 4 lines bringing off site power into the site . They could all be blown at one time by one man with time delay dev i ces and a few pounds of explosives .

This kind of damage would take an awful long time to repair . We are talking days to get power back into the site . This is only one area that has not been evaluated in the so called "resolution of Statio n Blackout . " Very truly yours ,

• MARVIN LEWIS 7801 ROOSEVELT BLVD. # 62 PHILA., PA 19152

a u.~ Mid.EAR Re°GULATORY c6MMI ~ KETING t' c;~QVICE '.;EC:T!Ol"f f Pc-

IIDD' NNlill WPNPIIU - NucLED ENGINEERING CORPORATION i28 West Mil'higan Awnue Jackson, Michigan 49201 Represented by May 23,

0. B. Falls, Jr.

Consultant The Secretary of the Commission U.S. Nuclear Regulatory Commission 1717 H Street, N.W. Washington, D. C. 20555 Attn: Docketing and Service Branch

Dear Secretary:

NucleDyne wishes to comment on the proposed rule on station blackout appearing in the Federal Register in Vol. 51, No. 55, dated March 21, 1986. Our comments also apply, particularly, to the NRC's Proposed Policy for Regulation of Advanced Nuclear Power Plants, issued on March 26, 1985. The Proposed Policy for Advanced Reactors should require the capability to safely withstand a station blackout of, at least, 8 hours. In our letter to NRC's Chairman, Dr. Nunzio Palladino of May 22, 1986, we request consideration of NucleDyne's Passive Contain-ment System (PCS) in keeping with the NRC's Proposed Policy on Advanced Reactors. In our letter we stated that it is hoped the proposed posture being taken by the NRC on Advanced Reactors applies to the PCS. Such consideration extending to confirmatory research on steam jet injectors as applied in the PCS will provide assurance of capability to withstand station blackout for 8 hours. Thank you for the opportunity to comment on the proposed rule. Sa?~ -

0. B. F~

Consultant OBF/mr cc: Dr. Nunzio Palladino, Chairman Admiral Lando W. Zech, Jr., NRC Chairman-Elect Raymond F. Fraley, ACRS

U.S. NUClEAR k!CUtA10RV COMMISS KET ING & SERVI CE SECT ION OFFICE nF TW SVRE:f A.RY o- - :: r o* ,c-c;* "'I N

May 19 - 86 from the pen of EMERY NEMETHY l~RC Re : ~reposed Rule - TT : :00 CKEHIJ G & St Htion BJackout SER I CE BRA1' Ch Fed Reg lViar 21-86 , p 98~9

• Gentlemen=

With 67 reactors now in operation , it ' s good to know you 're tackling the Unresolved Saf ety Issue of Sta tion Blackout . As an aaded precaution , we sugge st incl uding a clause that if a reactor eYceeds 4 or 8 hours of blackout, it be put into cold s hut-down (based on 1r John F Doherty ' s proposal in Fed Register Feb 11- 86 , p 5086) *

• Sec ' y AlllffcMtadRld bJ

l.uAt fMNIIIIPR -5'lJ DQPPIUI (1) (tfl Fil, 9132.q) Hay 10, Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, DC 20555 Attention: Docketing and Service Branch

Subject:

Comment for, "Proposed Station Blackout Rule" Gentlemen: The promulgation of the proposed rule on Unresolved Safety Issue

• (USI) A-44 Station Blackout must be deferred, and/or (i.e.,

accomplishing B-23, B- 48 and B-56). pending resolution integration of this USI with other related generic issues A-30, First a means for this integration in an efficient and cost-benefi-cial way must be established. Starting with the most basic systems, addressed in Generic Issue A-30, Adequacy of Safety Related DC Power Supply, and working through the more nebulous Generic Issue B-23, 0 Reactor Coolant Pump <RCP) Seal Failures and Generic Issue B-56, Enhancing Reliability of Diesel Generator and ending with the most important issues of USI A-44 Station Blackout and USI A-45 Shut-down Delay Heat Removal Requirements. Standardized Nuclear Operating Facilities would be much easier to address the issues, so the next best thing to do is - uniform systematic approach to incorporating meaningful databases and using this to promulgate beneficial regulation to the industry. The referenced NUREG of USI A-44 indicated that not enough useful information was recorded at Operating Facilities and often questionnaries had to be used. Much of the data was gathered based on experience 5 to 10 years ago. The Operating plants, we must believe, have moved along the learning curve. Steps to be taken prior to rulemaking concerning USI A-44:

1. DC-2.owe~-5uP+Ll~- S~siems must have time 1 imits imposed in Operating Stations Technical Specifications <T.S.)

concerning grounds and the magnitude thereof.

2. Establishing meaningful guide1 ines for maintaining component and system daiabasas (i.e., component IDs not used consistently in maintenance records and plant diagrams not consistent in showing IDs-NUREG/CR-3831).

                                           ~ledged by cetllJUN
                                                                      , ,-rr......,....,1/Wii'......

1

' rk D

• I

3. Before incorporating na~_+/-achnical=s~aci£ica+/-icn on diesel generator frequency and testing, a determination is needed to see if this testing will only lead to the

           *testing to death of components and systems* syndrome.

4. Banac~c-Eccceduces are used by most Operating facilities. These procedures are not carried into adequate depth of specific power plant operations. The industry has relied too heavily on generic procedures and has not given a ceai looK at what specific steps must be taKen. Extrapolation of these procedures must be required. Specific maintenance pr*ocedues must be established and followed.

5. Ica~nin.g_an.cLcpeca+/-icn treats the diesel generators as a

• pacKage component one pacKage.

instead of a system. When diesel generators are installed the sKid is usually one color - Most major systems are color coded in an Operating Facility but seldom - the subsystems of the diesel generator pacKages. This is a simple suggestion that will enhance trouble shooting technics of diesel generator components. Training must go in-depth to cover electrical equipment and distribution. Reducing the human error probability and thereby enhancing diesel generator reliability is the most cost effective pre-1 iminary step in addressing USI A-44.

6. The technical manuals which are used for references in maintenance, training and procedure preparation are not QA frozen documents. Thereby many errors have been made which are contributed to this fact.

7. The most intimately involved personnel, the 1 i censed operators must be incorporated in resolving the Safety issues through an open and independent forum - The Professional Reactor Operator Society.

Individual facility studies will bring us closer to standardization. This must be implemented with regulatory "teeth" to ensure uniformity of data and thereby streaml i ne required cost-effective rulemaKing. NUREG/CR-2989 "Reliability of AC Power Systems at Nuclear Power Plants" and NuReg/CR-3831,"The In-Plant Reliability Data Base for Nuclear Plant Components: Interim Report - Diesel Generators, Batteries, Chargers and Inverters" clearly indicated the studies were inhibited by the 1 imited amount of data. Required formats for data bases can no longer be delayed. 2

r I As an example I inquired into a generic bearing design problem for a major pump used throughout the Industry. The finding as follows wer e not fruitful: A. INPO networK - question considered inappropriate. B. Nuclear Power Reliability Data System (NPRDS) 1 ists data only for safety related components.

c. North American Electric Reliability Council - Generating Availability Data <GAD) System - tracKs only generic failures such as pump or motor but not necessar i ly the causes; plus the data base is 1 imited to last three r years *

• This is a Improvements must clear demonstration of be made viability of future studies.

If in these the areas you would require additional information I will upon your requests at the addresses below: present to increase be system. the available Sincerely yours,

                                        ~-~ ROBERT N. MEYER International         esentative of The Professional Reactor Operator Society President of Systems Management And Reliability Technology (Philippines>

1) Consultant - PNPP- 1 Philippine Project Ebasco Services Incorporated 88th Floor SE Two World Trade Center New YorK, N.Y. 10048- 0752

2) 8-2 NPV/NPC Bagac, Bataan Philippines

3) Box 70-A Hagar City, WI 54014 3

.~ IIJIHIUIPR** i:",

-        -         - J   '4    (J)
'(51 FR, 9B29 NUCLEAR ENERGY BUSINESS OPERATIONS GENERAL ELECTRIC COMPAN'/

• 175 CURTNER AVENUE
• SAN JOSE, MFN #038- 86 The Secretary of the Commission U.S. Nuclear Regulatory Commission Washington, DC 20555 Attention:

Docketing and Service Branch Gentlemen:

Subject:

Proposed 10CFR50.63 (Loss of All Alternating Current Power) On March 21, 1986, the NRC in Volume 51 , No. 55, Page 9829 of the Federal Register requested public comments on an amendment to its regulation which would require light-water-cooled nuclear power plants to be capable of with-standing a total loss of alternating current electric power (10CFR50.63). General Electric Company herein (Attachment A to this letter) provides com-ments both on the proposed rule and on the accompanying draft Regulatory Guide. Should you have any questions about our comments please contact Mr. Noel Shirley (408-925-1192) of my staff at your convenience. Very truly yours,

• R. Villa, Manager Product Licensing RV/wf Attachment

• *oMM\SSt<II

Ci\ot' R'<

ATTACHMENT A The following comments pertain to the proposed "Station Blackout Rule" pub-lished in the Federal Register, March 21, 1986.

1. The rule cites NUREG/CR 2989, "Reliability of Emergency AC Power System at Nuclear Power Plants" as the basis for the rule. This document has been superseded by NUREG/CR 4347, "Emergency Diesel Generator Operating Experience, 1981-1983". This latter document should be used as the basis for the rule. It also shows an improvement in diesel generator relia-bility over that shown in the earlier document. This increase in reliability needs to be reflected in the final rule.

2. The Commission requests comments on the question of quality classifica-tion of modifications. General Electric does not believe that any modifications needed to satisfy the Station Blackout Rule should be required to satisfy single failure criteria or to be otherwise considered safety related. Any upgrading of existing systems to meet the Rule should be consistent with the original design quality for the system.

The purpose for adding any new system would be to lower the core damage frequency. Since the evaluation takes credit for non-safety systems, the major benefit is derived from the availability of the system rather than its quality. To make additions safety related would greatly increase the cost of such modifications beyond the point where they might be consis-tent with the backfit rule. It would also greatly complicate the interface between Class IE and non-IE systems and components.

3. As we pointed out in our comments on NUREG 1109, the justification for this rule overestimates by several orders of magnitude the benefit for BWRs in terms of offsite exposure. Accordingly, it is not clear that the backfit rule is satisfied for BWRs. For example, the risk reduction cited in the Federal Register and NUREG 1109 is 80,000 person-rems for 67 reactors. At a frequency of lE-5 core damage events per year over 25 years, this is equivalent to about 5E+6 person-rem per event. The source term required to achieve this level of exposure has been described by the Staff as "upper bound". GE does not believe that such upper bound source terms are appropriate for use in application of the backfit rule.

The following comments pertain to the "Draft Regulatory Guide and Value / Impact Statement" on Station Blackout, dated March 1986.

1. The distinction between EXTREMELY severe weather frequencies of 1/ 100 years and 1/350 years in the definition of the P2 group (reference Table 3) is unnecessary since the Pl group must already have a frequency less than 1/350 years.

2. The requirements in Table 3 for achieving the Pl group when severe weather conditions exceed a 1/100 year frequency are to have procedures for restoration of power which can be implemented within two hours.

There does not seem to be any basis for this value. If a plant were able to restore power before the ability to cool the core is lost then the same benefit would be obtained. It should be sufficient to demonstrate that the procedure can be implemented prior to loosing the ability to continue core decay heat removal.

3. Section C.3.3 Item 3 specifies that if a system is required to be added to meet the station blackout requirements, that it be operable within 10 minutes. There is no need for a time restriction provided that it is shown that the system can be brought on line within sufficient time to assure core cooling. For station blackout events in a BWR this is typically greater than 45 minutes.

4. Section C.3.3 Item 4 should be deleted since the added systems should not be considered safety related (see Comment 2 on the proposed rule).

5. The apparent purpose of Section C.3.3 Item 5 is to assure that any additional systems are reviewed for adverse system interactions with safety systems. This should be considered in the resolution of Unre-solved Safety Issue A-17 and not be included in this regulatory guide *

• 6. Since the thrust of the proposed rule is to assure adequate core cooling under station blackout conditions for a prescribed period of time, we interpret the need for decay heat removal to be related to "core cooling" rather than "containment heat removal". Based upon this interpretation, Table 2 in the draft Regulatory Guide should be clarified to recognize the availability and positive impact of the BWR High Pressure Core Spray system
• Federal Register / Vol. 51, No. 55 / Friday, March Zl., 1986 / Proposed Rules 91129 ow lesa each way than the arrangement and that a two pound subsample of~

the top layer: the arrangement of the smalleat plums in each eight pound die layer may be the same aa the top sample containa not more than th lay , or may be ODe row less one way number of plums listed for the v ety in than e aITangernent ol the top l~yer. In Column C of said table. the 3 x5 and 3 1/4-4x4 packs the face of ch half of the crate shall be TABLE I packed a a unit, with no shim between Column the two b kets. Cot A. varie1y C,l)k-. pa, (2) The d meter of the smallest and largest plum in *any individual pack or pie container sha not vary more than one* Amazon .................. - .................... ....... .. 64 17 fourth(1/4) inch except that plums which Ambfa ......... _.............................. ....._..... 67 18 are placed in vo me*fill or tight.fill type Andya Pride ............._............. .............. 69 18 Angeleno ....... - ..................... ............... .. 87 19 Station Blackout containers and h we a diameter of two Angee................................. ....... -...*-- 87 18 and one*fourth (2 inches or larger Autumn Rosa ................... ..................... 72 18 AGENCY: Nuclear Regulatory Bee Gee .............. - .. - ........... - ....... .. 66 17 shall not vary more ban three*eights Blaclolmbar.................. ....................... .. 58 15 Commission. (3/s) inch. A total of t more than five Black Beaut- ............. ......................... . 6ll UI ACTION: Proposed rule. Black Olamond. -.. _ ................ _ .. .. 59 18 (5) percent. by count, f the plums in any Bliek Jewel ............ ............................. .. 54 M package or container y fail to meet !Mack Knight ........................................ .. 58 ti

SUMMARY

The Nuclear Regulatory this requirement. Carolyn t-tams .... ................................... 61 17 Casselman ............................................ . 83 17 Commission is proposing to amend its

               *

• Catalina .......... ............. - ..................- 59 te regulatiom1 to require that light*water*

Du,ado ............................................. .. 74 20 (d) When used herein" *ameter" shall Early Hawa" n Ann ................................ eo 18 cooled nuclear power plants be capable have the same meaning as et forth in

• the U.S. Sandards for Grad of Fresh EbOny ................................................... ..

El 66 88 ,. 11 of withstanding a total loss of alternating current (AC) electric power Empr 57 15 Plums and Prunes (7 CFR 51. to 58 15 {called "station blackout") for a 51.1538) and all other terms s II have Friar ............... _ .. - ....._ ......................... 56 1$ specified duration and maintaining Fron . .......... - ...............................- ...... 81 17 the same meaning as when us in the Gar osa ....................... - ..................... .. 71 19 reactor core cooling during that period. amended marketing agreement nd . Gr Rosa ................... - .............-. 54 14 This proposed requirement is based on order. "No. 12B standard fruit bo " Red ..*.. _ ...... - ........... - .. * ..*- * - 64 17 ly Santa Rosa ............... - ... - ............ 69 18 information developed under the measures 23/s to 71/4x11 1/2x161/4 in hes, elaey ................ - ......... - -............ - ... 47 ti Commission's study of Unresolved "No. 22D standard lug box" measu es King Oawl ........................... _ ............... . 50 14 King Richard .. _ ....... _ ............................. 54 14 Safety Issue A-44, "Station Blackoul" 27Ai to 71/4xt31/2x161/4 inches, "No. 2 G King*s Blade .......................................... .. 56 18 The proposed change is intended to standard lug box" measures 73/s to Laroda ..................................................... 58 18 provide further assurance that a station Late Santa Rosa (incWng improved 71/2x131/4x15 74 inches. All dimensions Lale Senta Rosa and SwaH Rosa) ... 64 17 blackout (loss of both offsitd power and are given in depth (inside dimensions Linda Rosa ........................................... 83 17 onsite emergency AC power systems) Maripoaa ................._ ........................ .. 81 17 by width by length (outside dimensi ns Midsummer ............. _ ............................ .. 83 17 will not adversely affect the public

3. Section 917.400 would be revi d to Nubi-- ................. _ ............................ 58 '5 health and safety.

read as follows: President. .... _ ...................................... 57 15 Pr1nla Blaci< ............... _ ......................... .. 69 13 DATE: The comment period expires on Ann ............................................. . 50 14 Rosa ... _ ................ _ .................. S3 14 June 19, 1986. Comments received after Beaut .................._ ......................... 74 20 this date will be considered if it is

 § 917.460    Plum Aegulatlon 19.                  Glow ............................................... .. eo        18

• practical to do so, but assurance of Rosa .. _ ...............- ... - ................ 84 17 (a) No handler shall ship y lot of 58 18 consideration cannot be given except as packages or containers of y plums ........................................... 74 20 to cqmments received before this date.

unless such plums grade least U.S. *****-**** ..***--***-****-******..****-**** 89 19 50 14 ADDRESSES: Send comments to: The No. 1, except that matu

• y shall be 60 18 Secretary of the Commission, U.S.

R 74 20 determined by the app

• ation of color Roy __ ......... __ ............... 74 20 Nuclear Regulatory Commission, standards by variety such other tests Santa ............................... .. 69 19 Washington, DC 20555, Attention:

as determined to be roper by the Simka. Yorker ................. .. so 14 Spring Bea .............................. 74 20 Docketing and Service Branch. Copies of Federal*State lnsp tion Service. Standard ...... ....................... .. 83 21 comments received may be examined Wickson ........ ..................... .. 51 14 Internal discolora on not considered and copied for a fee at the NRC Public serious damage d healed growth Document Room, 1717 H Street, NW., craclu emanat' g from the step end (c) No hand r shall ship any package Washington, DC. which do not ause serious damage or container of ny variety of plums not shall be pe *ued. In addition to the FOR FURTHER INFORMATION CONTACT: specifically nam in paragraph (b) of above, any ot of Tragedy or Kelsey this section, unle such plums are of a Alan Rubin, Division of Safety Review plume aha be permitted and additional size that an eight p nd sample and Oversight, Office of Nuclear 10 perc tolerance for defects not representative of the izes of the plums Reactor Regulation, U.S. Nuclear conside d serious damage. in the package or cont *ner contains not Regulatory Commission, Washington. (b) handler shall ship any package more than 139 plums, an that a two DC 20555, Telephone: (301) 492-8303. container of any variety of pound subsample of the s sliest plums SUPPLEMENTARY INFORMATION: The listed in Column A of the in each eight*pound sampl contains not alternating cun-ent (AC) electric power foll ing Table I unless such plums are more than 38 plums. for essential and nonessential service in of size that an eight.pound sample, , (d) As used herein, "U.S. N 1" and a nuclear power plant is supplied re esentative of the size& of the plums "serious damage" mean the sa e as primarily by offsite power. Redundant in he package or container, contains not defined in the United States Sta dards onsite emergency AC power systems are re than the number of plums listed for variety in Column B of said table, for Grades of Fresh Plums and CFR 51.1520 through 51.1538). es (7 power ,ou,o,. an, losl n,,.. *Y';~l \l also provided in the event that all offsited _ lO Q~'

9830 Federal Register / Vol. 51, Not 55 / Friday, March 21, 198!? / Proposed Rules 5 1 provide power for various safety as an Unresolved Safety Issue (USI); a indicated a variety of potentially systems including reactor core d cay* Task Action Plan (TAP A-44) was important failure causes. However, no heat removal and containment heat issued in July 1980, and work was single improvement was identified that removal which are essential for initiated to determine whether could result in a significant preserving the integrity of the reactor additional safety requirements were improvement in overall diesel generator core and the containment building, needed. Factors considered in the reliability. Data obtained from operating respectively. The reactor core decay analysis of risk from station blackout experience show that the typical heat can also be removed for a limited included: (1) The likelihood and duration individual emergency diesel generator time period by safety systems that are of the loss of offsite power; (2) the failure rate is about 2.5 x 10- 2 per independent of AC power. reliability of the onsite AC power demand (i.e., one chance of failure in 40 The term "station blackout" means system; and (3) the potential for severe demands), and that the emergency AC the loss ofoffsite AC power to the accident sequences after a ioss of all AC power system unavailability for a plant essential and nonessential electrical power, including consideration of the which has two emergency diesel buses concurrent with turbine trip and capability to remove core decay heat generators, one of which is required for the unavailability of the redundant without AC power for a limited time decay heat removal. is about 2 x 10- 3 onsite emergency AC power systems period. per demand (NUREG/CR-2989). (e.g., as a result of units out of service The technical findings of the stafrs

• Given the occurrence of a station for maintenance or repair, failure to studies of the station blackout issue are blackout, the likelihood of resultant core start on demand, or failure to continue presented in NUREG-1032, "Evaluation damage or core melt is dependent on the to run after start). ff a station blackout of Station Blackout Accidents at Nuclear reliability and capability of decay heat persists for a sufficient time during Power Plants, Technical Findings removal systems that are not dependent Related to Unresolved Safety Issue A-which Jhe capability of the AC- on AC power. U sufficient AC-independent systems to remove decay 44." 1 Additional information is independent capability exists, heat is exceeded, core melt and provided in supporting contractor additional time will be available to containment failure could result. reports: NUREG/CR-3226, "Station restore AC power needed for long-term The Commission's existing regulationa Blackout Accident Analyses" published cooling (NUREG/CR-3226).

establish requirements for the design in May 1983; NUREG/CR-2989, .

• It was determined by reviewing and testing of onsite and offsite electric "Reliability of Emersency AC Power. design, operational. and site-dependent power systems that are intended lo System at Nuclear Power factors that the expe~ted frequency of reduce the probability of losing all AC Plants"published in July 1983; and core damage resulting from station
• power to an acceptable level. (See NUREG/CR-3992, "Collection and blackout events could be maintaine.d General Design Criteria 17 and 18, 10 Evaluation of Complete and Partial near or below 10- 11 per reactor-year for CFR Part 50, Appendix A.) The existing Losses of Offsite Power at Nuclear any nuclear plant with readily regulations do not require explicitly that Power Plants" published in February achievable diesel generator reliabilities, nuclear power plants be designed to 1 1985. The major results of these studies provided that the plant is designed to assure that the core can be cooled and are given below. cope with station blackout for a the integrity of the reactor coolant
• Losses of offsite power can be specified duration. The duration for a pressure boundary can be maintained characterized as those resulting from specific plant is based on a comparison for any specified period of loss of all AC plant-centered faults, utility grid of the plant's characteristics to those power. blackout, and severe weather-induced factors that have been identified as the As operating experience has failures of offsite power sources. Based main contributors to risk from station accumulated, the concern has arisen on operating experience, the frequency blackout (NUREG-1032}.

that the reliability of both the onsite and of total losses of offsite power in offsite emergency AC power systems operating nuclear power plants was As a result of the station blackout might be less than originally anticipated, found to be about one per 10 site-years. studies, improved guidance will be even for designs that meet the The median restoration time was about provided to licensees regarding requirements of General Design Criteria one-half hour, and 90 percent of the maintaining minimum emergency diesel 17 and 18. Many operating plants have offsite power losses were restored in generator reliability to minimize the experienced a total loss of offsit1i power, approximately 3 hours (NUREG/CR - probability of losing all AC power. In and more occurrences can be expected 3992). , addition, the Commission is proposing to in the future. Also, operating experience

• The review of a number of amend its regula lions by adding a new with onsite*emergency power systems representative designs of onsite § 50.63 and by adding a new final has included many instances when emergency AC power systems has paragraph to General Design Criterion diesel generators failed to start. In a few 17, Appendix A of 10 CFR Part 50, to cases, there has been a complete loss of 1 Draft NUREG-1032 was issued for public require that all nuclear power plants be both the offsite and the onsite AC power comment on June 15, 1985. Copies of this report are capable of coping with a station systems. During these events, AC power available for public inspection and copyill8 for a fee blackout for some specified period of was restored in a short time without any al the NRC Public Document Room at 1717 H Street. time. The period of time for specific t

serious consequences. In 1975, the results of the Reactor NW., Weshill8ton, DC 20555. Free single copies of Draft NUREG-1032 may be requested by writing to the Publication Services Section. Room P-130A,

                                                                                                     . plant   would be determined based on the existing capability of the plant as well Safety Study (WASH-1400) showed that Division of Technical Information and Document                  as a comparison of the individual plant station blackout could be an important       Control. U.S. Nuclear Regulatory Commission,            design with factors that have been Washington, DC 20555.                                  identified as the main contributors to contributor to the total risk from nuclear

• Copies of these documents are available for power plant accidents. Although this public inspection end copying for a fee et the NRC risk of core melt resulting from station total risk was found to be small, the Public Document Room at 1717 H Street. NW., blackout.

relative importance of the station Washington, DC 20555. Copies may also be These factors, which vary purchased by celling (202) 275-2171 or (202) 275- significantly from plant to plant because blackout accident was established. 2060 or by writing to the Superintendent of Subsequently, the Commission Documents, U.S. Government Printing Office, P.O. of considerable differences in design of deaignated the issue of station blackout Box 37082. Washington, DC 2001~7082. plant electric power systems as well as

• Federal Register / Vol. 51, No. 55 / Friday, l,la~- 21, 1986 / Proposed Rules 9831 aite:-spei;ific considerations, include: (1) in thls guide would result.In selection of on station blackout. We will.be Redundancy of onsite eo,ergency AC a 4-hour or-8-hour station blackout. interested in comments received and
• pow~r sourc~s (i.e., number of sources duration, depending*on the specific plant staff responses associated with analysis minus the number needed (or decay heat design and site-related characteristics. of cost benefit, value impact. and safety removal (2) reliability of onsite However, applicants and licensees improvements and the station blackout emergency AC power sources (usually could propose alternative methods to standing on the overall risk (e.g.. Is the diesel generators), (3) frequency of loss that specified in the regulatory guide in reduction of risk only a small percentage of offsite power, and (4) probable time order to justify other minimum durations of the overall risk or is it a major to restore offsite power. The frequency for station blackout capability; component of an already small risk?).

of loss of, and time to restore offsite If the proposed rule and regulatory This will be one of the first-proposed power are related to grid and gulde are issued, those plants with an ~les to be evaluated by the NRC under switchyard reliabilities, historical already low risk from station blackout its new backfitting requirements. We weather data for severe storms, and the would be required to withstand a station would be particularly interested in availability of nearby alternate power blackout for a relatively short period of specific comments assessing whether or sources (e.g., gas turbines). Experience time and probably would need few, if not this proposal meets the "substantial has shown that long duration offsite any, modifications as a result of the rule. increase in the overall protection of the power outages are caused primarily by Plants with.currently higher risk from public health and safety *. ." threshold severe storms (hurricanes, ice, snow, station blackout would be required to now required by the backfit rule. etc.). withstand somewhat longer duration The objective of the proposed rule is blackouts. Depending on their existing Separate Views of Commissioner to reduce the risk of severe accidents capability, these plants might also need Asselstine resulting from station blackout by to make -modifications (such as I support the proposed rulemaking but maintaining highly reliable AC electric increasing station battery capacity or believe substantial additional safety power systems and, as additional condensate storage tank capacity) in improvements beyond those called for ln defense-in-depth, assuring that plants order to cope with the longer station this rulemaking are achievable and can cope with a station blackout for blackout duration. The proposed rule practicable. How to prevent and some period of time. If the proposed rule would require licensees to develop, in mitigate a station blackout event is one is adopted, all licensees and applicants

• consultation with the Office of Nuclear of the most.significant unresolved safety would be required to assess the Reactor Regulation, proposed plant- issues associated with nuclear power capability of their plants to cope with a specific schedules for implementation of plants. Extended station blackout,can station blackout (i.e., detennine the any needed modifications. result in,g)re meltdown and loss of amount of time the plant can maintain conta~t integrity. Since existing core cooling and containment integrity Additional Comments by the Commi1111ion mitigation features such as containment with AC power w:iavaitable), and to spray would be inoperable, a station have procedures and training to cope The proposed rule does not require blackout could result in a large release with such an event. Plants would be that a single failure.be assumed of radioactive material to the required to be able to cope with a concurrent with a-station blackout
• environment.

specified minimum duration station because station blackout goes beyond the normal single failure criterion. That Countries abroad that have made a blackout selected on a plant-specific serious commitment to nuclear power basis. is, for a station blackout to occur, four AC power supplies must fail (two offsite and to nuclear safety have, or are On the basis of station blackout planning, backfit features which studies conducted for USI A-44, and sources and two safety-related onsite emergency AC sources). The staffs markedly reduce station blackout risks. presented in the reports referenced For example, the new French 1300 MWe above, the NRC staff has developed a estimated probability of the concurrent nuclear power plants are designed with draft regulatory guide entitled "Station - failure of all four power supplies leads a goal of coping with a station blackout Blackout," 1 which presents guidance on. us to believe that the staff should give for at least 20 hours. According to the (1] maintaining a high level of reliability further consideration to upgrading to safety grade the plant modifications NRC staff, the design features that for emergency diesel generators, (2) provide this capability (listed below) developing procedures and training to needed (if any) to meet the proposed rule. Upgrading to safety grade will permit the plant to withstand a station restore offsite and onsite emergency-AC blackout for three days. power should either one or both become further ensure appropriate licensee attention is paid to maintaining a high

• A steam-driven generator provides unavailable, and (3) selecting a plant- power for a small positive displacement specific minimum duration for station state of operability and reliability. The Commission believes that the question pump that supplies cooling for reactor blackout capability to comply with the of quality classification of modifications coolant pump (RCP) seals and also proposed amendment to General Design should be addressed by interested provides power for instrumentation and Criterion 17. Application of the methods parties In comments on the proposed controls and control room lighting rule. necessary to withstand a station
• A notice of availability and requ"t for blackout. This design feature, which is comments on the draft regulatory guide will be In addition to comments on the merits published within a few days of this Notice of of the proposed rule, the Commission also being backfitted onto all operating Proposed Rulemakilll!. Copies of the draft regulatory specifically requests comments on 900 MWe nuclear plants in France, guide are available for public lnapection and addresses two factors that impact the copying for a fee at the NRC Public Document Room wh11ther the backfitanalysis for this rule adequately implemJmts the Backfit Rule, ability to cope with a station blackout-at 1717 H Street. NW., Wathington, DC 20555, and will be diatributed to thoae on the automatic 10 CFR 50.109. RCP seal cooling with AC power distribution liat for draft regulatory guldea. Pree µnavailable and battery depletion.

1tnsle cople of the draft regulatory guide may be Additional Comments by

• Two turbine-driven auxiliary obtained by writing to the U.S. Nut;:lear Regulatory Commisaionen Roberts and Zeeb feedwater (AFW) pumps included in the Comml1&ion, Washington, DC 20565, Attention:

Director; Diviaion of Technical Information and We agree with soliciting public 1300 MWE Frencn design in addition to DQCumenl-Control. comments on the proposed rulemaking two motor-driven AFW pump11. Moat

• ~-two........... . . .

U.S. 'prelllmiiled water nactan haft one ........... ~ , - . p

• n-aE..... tNt'tach . . . . Pl'IMMI additional redunducy iD tile AC-

                                                   *111r        *~iiaelladbwefno
                                                   *igrdflmml '...-it . . arai IM rl!Cla Mr.

Wana Nbwwl, Ofiim af Nw:lear Radslteplatim, U.S. Nuclear

                                                                                             *  ~ I nlatiou. Nudeat poWIII' ,lanta ucl reactol'I, Paaalty, Radiation protection, Reactor tl<<ns c:rtterJa, Reporting and recordkeeping Regulatory Commisaion. Walhiagton.           requirements.

igdeperident trams of Ille AFW *JNm. DC 20555. Telephone: (301) 492-7827.

• Gravity feed beck-up wuer ftl)ply For the reason* set out in the from oll8ite AOUJ1CU to the oomleuate Paperwelk.Reductlon Act Statement preamble and under the authority of the stomp tank proYidea additional water This praposed rwe amends Atomic Energy Act of 1954, as amended, for llacar heat nnaval via the AFW infamtatioa collection requiremenbl that the Energy Reorganization Act of 1974, lfatala ior Iona-duration ata&ion are *ubiect to tile Paperwork Reduction as amended, and 5 U.S.C. 553, the NRC blaclrout eventa, Le.. qp to tbNe dap. Ad ef 18IIO (44 U.S.C. 3501 et seq.). Thi* is proposing to adopt the followina Thi* three-day *tatiml blackout nde baa been submitted to the Office of amendments to W CFR Part 50.

capability woald permit suffw:ient.time Manatement and Budget for review and ID connect* mobile gas turbme approval of the papa work PART so-ooMESTIC UCENStNG OF 8Ufflltal' ~ provide power if AC pewer .tieqUirementa. PRODUCTION AND UTILIZATION cmuJd DQt be reatoiwdboa Mhar, FAC&UTIES preferred aourcea. A aobile sa* tubiae Replataey ~

• genertor is located at, or in the vtcinity 'l1le Commieaioa has prepared a 1. The authority citation for Part 50 of, eve17 Melear power plot 11ite in replaloly analyai1 for lhia regulation. contilµl* to read as fallows:

France. These improvements ill safety The analrsu examines the costs and Aldbmity: Sect. tt,S, 10I. 181, 18Z 183, 188, are being achinad at not uaraaaooabie benefits flltlle nde ** OOD8idered by the 189, 68 91al 138. 937, 9C8, El, 954. 955, 956,

• cos ta .and are .beios .dri11eD br 1he French Coamiaeion. A cepr of the replatory a ~ sec 234. 13 Stat. 1244. a1 amended aoal of achievinl
• probabiit, qf ona ill an. .,ia, NUREG-1109, For Comment. (42 u.s.c. 2133. Z134.2201, 2232. 2233. Z238, ten million cw-71 per reactal'--year for a "Re,u}at<<y Analysit for the Reaollltion 2231, 22112); .ca. 201. zm. 206. 88 Stat. 1242.

major eunt auch aa aatiaa blackmt of Umeselved S..fety lane A-44. Station 1244, 1241. u mmnded (4Z U.S.C. 5841. 5142.

'Out C.mrnjaaion'* me.propeaaa aw.ch               Blackout" {Publiabed ill Januuy 1988), ia    IM8). mdea adsarwiBe 110ted.

Jea It prq,oaea aa objectiw.e Gf QDe in nailable for inapectio11 nd copying for Sectiap rio.7 also issued under Aib. L one hundred thouaand (1'T) per reactOl'- a fee &t rbe NRC Pllblic Document llooaa. 1717 H Sheet. NW, Wumagtoa. 95-all. aec. 10, 92 Stal 2951 (42 U..S.C. yew .for .atatioa black.out mwaetlcore meltmwn aad an objective of ol3lf DC 20555. Free single copia of Dnft 5851). SectioDa 50.57{d), 50.58. 50,111, and SOJJ2 also isauell Ulldet Pub. L 97-415, 96 about four hours coping c:apabllitJ. NUREG-1109 may be obtai.aed by

  . I would a,ppre<iate cornrnsta on               writing to the Pa.bllcation Services         Stat. 2ml. 2073 (U U.S.C.2133. 2239}.

wheJb'1l' the NR.C sbouW requin, Section, Room P-130A, Division of Section 50.78 al,e ileued under sec. 122. substantial improvement. In .aafet;r IW.itb Tecbnic:al Information ud Document 68 Stat. 939 (42 U.S.C. .2152). Sectigm respect to station blackout.. lib these ControL U.8. Nuclear lteplatory 50.IIG-G0.81 alu i11Ud under sec. 184, 68 tieing accomplished in ofher countries, Comnlluima, Waabinp,n, DC 20555. Stat. 954. aa amended (42 U..S.C. 2234}. which can be achieved at reasonable The Commilaion n,qv.ea public Sectiana 50.100-50.102 alao iuued 1IDder coat and which go beyond those COIDID811tOll the regulator, analylie. sec.186. 68 Stat. 955 (42 U.S.C. 2238). propoaetl in this rulemaking. Comaeata on the draft analyais may be For the purposes of sec. 223, 68 Stat.

                                                   .ubmittetl to the NRC as indicated 'llllder  958, aa amended (42 U.S.C. 2213).

Finding af. Na Slpifloaat IIMawma..tal H 5'1.lO(aJ, (b), and (cl, SOM. 50.48, latpact: Availability the Arl"II Ill heading. 50.48, 50.54, and 50.BO(a) are issued

    'T1te Commission hn detennine under             Raa,*latarJ .Plnlaillty CeJtiftcatioa        under aec. 16th, 68 Stat. 918. u the NatioDal &mnmmental PoHcr Act                     llu,..a.d111ce wttb the &88ulatory        amended E42 U.S.C. 2201(btt; t f 50.tO{b)
 ~ 1989, M amended. and the

• Fl*dl!d~ Mt of 1980. & U.S.C. 805{b), and (c) aad 50.54 are ias- under sec.

CollllDleaicft'* regu1atic,m in Su~ ft. the O Phi ieeioil llentby certifie* that 1011, 68 Stat. IH9. WI amended (42 U.S.C. of 10 CFR Part 51, that this fflle. if * * ' this ,eap111 d nda, if pro,N)sated, will 2201(1>>: and H 50.55(e), 50.59(b), SA.70, adapted, woald not be a major Federal not llave a etgnlficant ecaaoaic impact 50.71, 50.72. 50.73, and 50.78 are iuaed action significantly affecting lhe qualltJ: on e. am, t tial number of small under sec. 1610, 68 Stat. 950, as of the human en.tronment. 8lld i entitiN. '11m pNpOl8d rule tpeeifiea amended (4% U.S.C. Z201(o)J, tllaafon

• eovilauneotal Impact tMt auai8ar1J0wer planta be able to 2. In I 50.Z, a definition of ..etation lta1ielaent ia 11ot requirecl. -n.re would not be any adYene envinmmemal impaela **
• 1'881llt ol the IJ'DPOled Me "pedied t*

wtdatnll a .....i &o.s of. AC ,ewer for a clurldiort and ,naintain mactor caie ceollng durins that period. blackour' is added in the alphebetical. sequence to read ae follows: liar the following l'N80ae: (tt 1lme ThesefaeiiMee RN licenaed 1Ulder the f 50.2 DeflnitlonL would be no aclclltwnal radiolop:al prouoas d 10 O'R 111.Zl(b) and 10 * * * *

• exposure to the general pllhlc or plant CFR 50.22. The companies that own "Station blackout" means the employeea, ad tz) plant sbatdawn le lileN fadUtiea de not fall within the complete Ion of alternating current (AC)

Jllt ,equireli so daere would be no scope of "8ffl&R ntittes" as set forth in electric power to the esseatial and additi8Dal e11ftl'OIUD8lltal illlpllc* as

• the R.-tGrY Plexlbltity Act or th.e noneuential awitchaear buea in a re.at of tbe need b replaoell&lt ...u ........ .a9tandards set forth nucleaT power plant (i.e., lo* of the power. Tbe em,i[Wlal ueeeement in resu}aUGIU tu.ed.by the SmaH of&ite electric power system concurrent and findlns of no aignificant illlpac,t OR Business Admini1tration in 18 CFR Part with turbine trip and unavailability of which tibia determinatioa la bned are 121. the onsite emersency AC power availaW. for inepection ud copytng for a,etem).

a fee at the NRC Public Dbclunent List of..,._ la 18 CPR Put 18 Raom..1111 H.stJiaet. NW. Wahbllltea. Antilrut. aa..Ified infonaation, Plre :1. A new I 50.63 is aaded to read as DC.8'11&le copies of fhe envtroam.1111

• pleYelltloa, h..xw,,otation by reference, . follows:

Federal Repter / Vol. 51, No.,.55 / *Pritlay, Maroh 21, 1986 / Proposed Rule1 9833 t ..... :.......... ..........QlrNIII Cemmtsnon wiH notify 188 ltceneee of likelihood of lhelr limultilneoufailllre,under poww.' its determination of tht apecified 1tation operating and poetulated accident and {a) ~ t a 8-li&lit-water- blackout duration tt> be uect-in environmental condition. A awttchyard - cooled nuclear power plant licensed ti> determinins compliance with General common to both circuit* 11 acceptable. Each operate mu1t be able to with8tand and Design Criterion 17 of Appendix A of of theae circuit shall be deaigned to be available in sufficient time following a 1011 of recover from a station blackout as this part. all onsite alternating current power 1uppliea defmed in I 50.2 for a specified duration (d) Implementation-Schedule for and the other offsite electric power circuit, to in accordance with the requirement, in Implementing Equipment Modifications. assure that 1pecified acceptable fuel deaign paragraph (e) of General Design (1) For each light-water-cooled nuclear limits and desip conditions of the reactor Criterion 17 of Appendix A of this part. power plant licensed to operate on or coolant presaure boundary are not exceeded. (b) Limitation of Scope. Paragraphs (c) before [insert the effective date of this One of theae circuit ,hall be designed to be and (d) of this section do not apply to amendment], the licensee shall, within available within a few second* following a those plants licensed to operate prior to 180 days of the notification provided in 1011-of-coolant accident to aBBure that core [insert the effective date of this accordance with paragraph (c)(2} of this cooling, containment integrity, and other vital amendment] if the capability to 1&fety function are maintained. section, submit to the Director of the (d) Provisions hall be included to withstand station blackout was Office of Nuclear Reactor Regulation a minimize the probability of lo1ing electric considered in the operating license schedule for implementing any power from any of the remaining supplies a proceeding and a specified duration was equipment and procedure modificationa a result of, or coincident with, the 1011 of accepted as the licensing basis for the necessary to meet the requirements of _ power generated by the nuclear power unit, facility. General Design Criterion 17 of Appendix the 1011 of power from the tranamiHion (c) lmplementation-Detennination of A of this part. This submittal must network, or the 1011 of power from the on1ite Station Blackout Duration. (1) For each include an explanation of the schedule electric power supplies, light-water-cooled.nuclear power plant and a justification if the schedule does (e) The reactor core and a11ociated licensed to operate on or before [insert coolant, control. and protection 1ystems. not provide for completion of the including the station batterie1, hall provide the effective date of this amendment), _ modifications within two years of the the licensee shall submit to the Director 1ufficient capacity and capability to aBBure notification provided in accordance with that the core i1 cooled and containment of the Office of Nuclear Reactor paragraph (c)(2} of this section. integrity i1 maintained in the event of a Regulation by [insert a date 270 days (2) The licensee and the NRC staff station blackout (as defined in.I 50.2} for a after the effective date of this shall mutually agree upon a final specified duration. The following facton amendment]: schedule for implementing modifications hall be conaidered in specifying the station (i) A determination of the maximum necessary to comply with the blackout duration: (t) the redundancy ,f the duration for which the plant as currently requirements of Criterion 17. onsite emersency AC power sources. (2) the designed is able to maintain core reliability of the onaite emersency AC power cooling and containment integrity in the 4. In Appendix A. General Design sources, (3) the expected frequency of 1011 of event of a station blackout as defined in Criterion 17 is revised to read as offsite power, and (4) the probable time I 50.2: follows. needed to restore off1ite power. (ii} A description of the procedures Appendix A-General Design Criteria Dated at Washingtoa, DC, this 17th day that have been established for station for Nuclear Power Plants of March 1986. blackout events for the duration For the Nuclear Regulatory Commiuion, determined in paragraph (c}(l}(i} of this * * * *

• Samuel J. Chilk.

section end for recovery therefrom: JI. Protection by Multiple Fission Product Secretary of the Commission. (iii) An identification of the factor{s) Bon-iers that limit the capability of the plant to * * *

• Backfit Analysis cope with a station blackout for a longer Criterion 11-Electric power systems. (a) Analysis and Detennination That the time than that determined in paragraph An onsite electric power system and an Proposed Rulemaking To Amend 10 CFR (c)(l)(I) of this section; offsite electric power 1ystem shall be 50 Concerning Station Blackout (iv) A proposed station blackout provided to permit functioning of 1tructures, Complies With the Back/it Rule 10 CFR duration to be used in determining systems, and components Important to 1&fety, The safety function for each system 50.109 compliance with paragraph (e} of (assuming the other system is not General Design Criterion 17 of Appendix The Commission's existing regulations functioning) shall be to provide sufficient A of this part, including a justification capacity and capability to assure that (1) establish requirements for the design for the selection based on- specified acceptable fuel design limits and and testing of onsite and offsite electric (A) The redundancy of the onsite desip conditions of the reactor coolant power systems (10 CFR Part 50, emergency AC power sources; pressure boundary are not exceeded as a Appendix A, General Design Criteria 17 (B} The reliability of the onsiteo result of anticipated operational occurrences and 18). However, as operating emergency AC power sources; and (2) the core is cooled and containment experience has accumulated, the (C) The expected frequency of loss of integrity and other vital function are concern has arisen regarding the maintained in the event of postulated offsite power; and accidents. reliability of both the offsite and onsite (DJ The probable time needed to (b) The onsite electric power supplies, emergency AC power systems. These restore offsite power; and including the batteries, and the onsite electric systems provide power for various (v) An identification of the factors, if distribution system, shall have sufficient safety systems including reactor core any, that limit the capability of the plant Independence, redundancy, and teatability to decay heat removal and containment to meet the requirements of Criterion 17 perfonn their safety functions a11uming a heat removal which are essential for for the specified station blackout single failure. preserving the integrity of the reactor duration proposed in the response to . (c) Electric power from the transmission core and the containment building, network to the onsite electric distribution paragraph (c)(1)(iv) of this section. 1yatem hall be 1upplied by two phy1ically respectively. In numerous instances (2) After consideration of the. independent circuit1 (not necesaarily on emergency diesel generators have failed information s11bmitted in ~ccordance aeparate rights of way} deisned and located to start and run during tests conducted with paragraph (c)(l) of this section. the 10 aa to minimize to tbe extent practical the at operating plants. In addition, a

Feclaal ......r\/rVol.. m., No,'15. / Frklay,. Mardl.21, 1S86' / Prapoleil iwJee . . .,. ,., . ......... . oa.ta Tb..*t.iamtacl tota!coat f* fJ7

                                                     . . . .ti .. reac:forttD~widadle syst811:18t dial dte.e&ttma.~

dall\age frequency from station blackout eveat.acea lncreue aigniftcantly. eledllls---...a . . . . . . propo~11fU9lA-44Ja. ooom...,._.,,. a peele!l. Wftl ab* t40 million. The 8ftnP coet per The estimated eqoency of core resul-"om- do not require explicltly that reactor would be amud S600.om damage from station blackout events i* nuclear power plants be deslgned to ranging from $200,000 if only a station directly proportional to the frequency of withstand the foas of all AC power for blackout aaseB1ment and procedures the initiating event. Estimates of station any fpecifaed period. and training are neceaaary, to a blackout frequencies for this USI were

   'l'hia.i uue laaa been etudied by the             maximum of about $4 million if               baaed on actual operational experience.

staff u 118ft of Unreaolved Safety haue subaiantial modificatiooa are needed. Thia is a1t1umed to be a ree.listic (USI) A-44, "Station BlackOllL" Both includiae requalification of a dieael indicator of future performance. An detanniniflic and probebiliatic analyaes gaerator. . argument can be made that the future were-performed to determine "1he timing The overall valu.e-impact ratio, not performance will be better than the past. and conaequences of various accident including accident avoidance coat.I. ta

• For example, when problems with the sequences and to ideQtify the dominant about Z,000 peNOD-rema averted per offsite power grid arise, th:ey are fixed.

factara affecting the likelihood of core million dollan. If coat savings to and therefore, grid reliability should mall accidenta '&om atatwn blackouL industry f?om accident avoidance (1.e.. improve. On the other hand, grid power Theae atadi* indicate that atatlion cleanup and repair of onsite damages failures may become more frequent blaokolat .can he a 9ignificald oontributor and replacement power) were included. because fewer plants are being built. to theonrall plant ri:ak. Coneequently, the overall value-impact ratio would and more power is being transmitted die CoamiuioR is proposill8 to amend improft significantly to about 8.000 between regions. thus placing greater ill reguJationa to require that plants be p.,.on-rema averted per million dollars. streu on transmission lines. capable of witbatanding a total loH of Thia analysis support a A number of foreign countries. AC power fur a apeclled duratiOJl and determination that a aubatantial inchidlna Prance, Britain, Sweden. to .maintain reactor core.coollna during increase in the protection ol the publh: "Germany and Belgiv.m, have taken atepa that period. laealth and safety will be derived from to reduce the riak from station blacko1&t An aelyaia of the bmefita a.ad coats the backfit ta the proposed station events. Theae steps include adding ofiln,lwmt:,adaepropowiution . bla~out rule, and that the bacldit ia design feature. to enhance the blackout Nie ia p1'88811ted in NURBG- justified in view ol the direct and capability of the plant to cope with a 1109, Draft Report For Comment, Indirect ooetB of implementing the station blacko1ll for a substantial periDd "lt.f1'8'1latorJ Analysis for the Rnolution , proposed l'llle. of time, arul/ or adding redundant and of Unraolved Safety Iasue A-44, Station The quantitative value-impact diverse emergency AC power sources. Blackout.** *'The benefit from anal,.ia disc:uued above wu one of the The factors diacuaaed above aupport implerneotina thaiJm,oNdn&le la a factors conaidered in evaluating the the determination that additional reduction in die ln,qwmcy of core melt proposed rule, but other factors also defense in depth provided by the ability per reactor-yiu due to statiOll blackout played a part in the decision-making of a plant t o ~ with station blackout and 1he auoci.a ted risk of offsite process, Probabilistic risk asseBSment for a specific duration is warranted. The radioactive releuea. Tbe risk redw:tiqn (PRA) studies performed for this USI, as Commisaiaa has coaaidered how tlria forffT-operating reactors la estimaJed to well aa some plant-specific PRAs, have backfit should be prioritized*and be 80,000 person-rema. 1 shown that station blackout can be a scheduled in light of other regu)atory The coat for ticenaen to comply with significant contributor to core melt* activitie11 ongoing at operating nuclear the proposed backfit wmdd vary frequency, and, with cooaide.ration of power plants. Station blackout warrant depending on the existina capability of containment failure, station-blackout a high priority ranking based on both its each plant to cope with a mtton events can represent an important status as an "unresolved safety issue" blaclfflut, as weft ae the specified station contributw to reactor risk. In general, and the results and conclusions reached blackout duration for that plant. The active canninment *y1tem1 required for in resolving this issue. As noted in the costs would be primarily for licensees to lteat ~ a l , preaaure mpprenic:m, and implementation section of the proposed develop procedures, to improve diesel radioactivity remoYal from the rule(§ S0.63(dl), the schedule for ,....ratGr Ntiability if the reliability containment atmosphere following an equipment modification (if needed to fella below aertaia lev-, and to n:tmfit accident are unavailable during a meet the requirements of the proposed plant.a with adottional t:OIIJPOIUtDt8 or

• station blackou.L Therefore, the offsite rule) shall be mutually agreed upon by system,, *MCIIIAl)',.to..,. tlae risk i higher from a core melt reaulting the licensee and NRC. Modificatioa.

propoaed ffi111M'819-* from atatioo blackout than it is from that cannot be scheduled for completloo Dl&DY other accident scenarios. withirt two years after NRC accepts the

• Draft ~11111 wu 1-ad.for1Nli>lic Although there are licensing licensee's specified station blackout comment la ta-rJ t-.C.,.flllliillrepad an requirements for guidance directed at duration muat be ruatified by the available IDr laapec:Kan and ~Yllll for a fee *t tba providing reliable offsite and onaite AC licensee.

NllC Public Document Room. 1717 H StmL NW., powe.r; l!xperience baa shown that there Analysis of 50.109(c) Factors Walhliltton, DC -

• Pree fflllle copvia of Draft NUREG*tt* **Y b abtind br writiD8 to the are practical limitations in ensunns the Pub~*Srvicaiille:tiea. 1i1oom P...1030A. reliability of offaite and onaite t . Statement of the specific obfectives Dlvilioo of ndmlcal lllformatiou and Docment emergency AC power systems. Potential that the propoaed bacldit is designed to

_Control. U.S. NuclerResulalOfJ Cemmlton. actueve. WHhill8ton,0C20555. . vulnerabilities to common cause failures

    'flle-A. . . .aat -i,.1 In NlJltBG-110IWH         aBBociated with design, operational and         The NRC staff has completed a review hued aa p i n t ~ illfonaatiCIJI for a tDtI of ft1   environmental factors can affect AC          and evaluation of information developed reactoa. .Al&IIGuali liHln u c:wratlf about 1Q)      power system reliability. FOl' example. JI   over the past 5 years on Unresolved cipsallna reaaton, the overall val~lmpact ra.llo In   potential com.moo cause failuret, of         Safety Issue (USO A-44, Station NUR!JC..n*-wuutd not chanp ,1gnmcantly
 ~ of-4hemlll'ftMJII ._ 11ulllber .f e,erettnf         emergency diesel aeneratora exist (c..s,. blackout. As a retrult of these efforts, the plant,.             .,                                in service-water or OC poww suppod           NRC la proposms to amend 10 CFR Part

  /

I I - Federal &egialer so.by t h e . i n ~ of new I SB.es, V~L -Sl. No. 55 / Friday, March 21, 1986

                                                   ~iniq to icupe-with and reco*ar fnJlll a Propoeed. Rules For 81 operalift8;reaciorl, tAe 9835 "Station mitckout." and an addltianal             station blackout.                           estimated total cost for NRC review of p......... tr>.GeaeraJ U...Crttericm                 3. Potential change in the risk to the   industry submittals required by the

11. "KleGtric Power: 8ytt81U" kl public from the .accidental off-site proposed rule is $500,000 [based on an AppeadixA. release of radioactift material. estimated average of 120 person-hours The objective of the proposed rule ii ' Based on an analysis of potential per reactor; see Table 8 in NUREG-to reduce the risk al aevere accidents consequences presented in Section 4 of 1109).

aslOCiated with 1tation blackout by NUREG-1109, if the proposed rule were 8. The potential impact of differences muinfl .station blackout a relatively implemented, the estimated total risk in facility type, design or age on the small contributor to total core melt reduction to the public from 67 operating relevancy and practicality of the frequency. Specifically. the proposed reactors is 80,000 person-rem. proposed back.fit rule would require all lisht-water-cooled 4. Potential impact on radiological exposure of facility employees. The proposed rule applies to all aaclear power plants to be able to cope pressW"ized water reactors and boiling wita a ution blackout for a specified For 67 operating reactors, the estimated total reduction in water reactora. However, in determining duratioa, and to have procedures and the specific minimim station blackout trailwta for 111eh an event A draft occ\lJ)ational expoaure resultill8 from reduced core melt frequencies and coping capability for each plant. Rewwatory Guide, to be iuued alq differences in plant design [e.g.. number wida lhs pmposed rule. would provide associated post-accident cleanup and repair activities ia 2.000 person-rem of emergency generators) and the an acceptable method to determine the reliability of the offsite and onsite station blackout duration for each plant. (Table 8 in NUREG-1109). No increase In occupational expoaure is expected emergency AC power systems could The dw-atioa would be determined for result in different coping capabilities. each plant baaed on a comparison of the from operation and maintenance or implementing the proposed rule. For example, plants with an already low ind.i.vidul plant design with factor that risk from station blackout would be Equipment additions and modifications ha11e been identified aa the main cootemplated do not require work in required to withstand a station blackout cODtributora to rilk of COl'e melt for a relatively short period of time: and and around the reactor coolant system re1ultill8 ma station blackout These and therefore would not be expected to few, if any, hardware baclcfits would be facton are: (1) The redundancy of onaite re,ult in significant radiation exposure required as a reault of the propoeed rule. emergency AC power aourcea. (Z) the (Table 8 in NUREG-1109). Plant,.with cummtly higher risk from reliability of onaite emergency AC 5 . Installation and continuins costs 1 station blackout would be required to power eources, (S) the frequency of lou auociated with the backfit. includins withstand somewhat longer duration of off.U. power and (4) the probable the cost of facility downtime or the coat blackouts; and. depending on their time needed to restore olfsite power. of conatruction delay. existing capability, may need !Olfte

2. General detcription of the activity For 67 operating reactors, the total modifications to achieve the longer that would be required by the licensee estimated cost for assessill8 the station station blackout capability.

or applicant in order to complete the blackout coping capability, procedures 9. Whether the proposed backfit is backfiL and training. installation of hardware interim or final and, if interim, the In order to assure that each nuclear - backflts (if necessary). plant downtime, justification for imposing the proposed power plant ia able to wtthatand and and operation and maintenance ia $40 backfit on an interim basis. recover frmn a station blackout for a million. [See Tables 6 and 8 in NUREG- The proposed rule is a final resolution specified minimum duration, licenaees 1109). of USI A--44; it is not an interim would be required to auess their plant' 8, The potential safety impact of measure. capabill.1:y to withstand and recover changes in plant or operational from a station blackout This evaluation lFR Doc. 86-6284 Filed 3-20-86; 8::45 am) complexity, including the relationship to BIWNG CODE 7590-41-11 would inchide: proposed and existing regulatory

• Verifying the adequacy of station requinmients.

battery power, condensate storase tank The p r ~ rule for plants to be capacity, and plant/instrument air for able to cope with a station blackout the atatlon blackout duration. should not add to plant or operational

• Verifying adequate reactor coolantpump complexity. The relationship between seal Integrity for the atation blackout the proposed station blackout rule and duration ao that seal leakage due to lack of propoeed and existing regulatory seal coolins would not result in a aufficient requiremfflts is discussed in Section 4.2 primary system coolant inventory reduction of NUREG--1109. This discussion [Airspace to lo11e the ability to cool the core. includes the following NRC generic programs: Proposed A,m11u1r1m
• Verifying operability of equipment needed to operate during a station
• Generic Issue B-56 "Proposed Area; DaUaa/F blackout for environmental conditions Actions for Enhancing Reliability of Correction associated with total loss of AC power Diesel Generators at Operating Plants,"

(i.e., loss of heating, ventilation and air

• Generic Issue 23, "Reactor Coolant
• on page conditioningJ. Pump Seal Failures," 7950, in th h 7, Depending on the plant's existing *
• USI A-45, "Shutdown Decay Heat 1986, ma
• s:

capability to cope with a station Removal Requirements," 1. 0 irty-blackout, licensees may or may not need

• Generic Issue A-30, "Adequacy of firs
• end insert de to backfit hardware modifications (e.g.. Safety-Related DC Power Supply." 33*

adding battery capacity) to comply with 7. The estimated reaource burden on me page, third colu the proposed rule. {5ee item afor the NRC assQciated with the proposed e, before "thence" .inll additional discunion.) Llcenaees would backflt and the availability of auch de 97*39'30" W.,". be reqtlired to have procedures and resources. - - 1IIOH1-M}}