ML20245E857

Forwards Revised Commission Paper Containing Recommendations of Mark I Containment Improvement Program.Paper Revised to Reflect Comments Provided by Crgr,Acrs & Addl Input from NRR & Ofc of General Counsel

ML20245E857
Person / Time
Issue date:	01/04/1989
From:	Beckjord E NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To:	Jordan E Committee To Review Generic Requirements
Shared Package
ML19316E281	List: ML20245E852 ML20245E857 ML20245L427
References
NUDOCS 8901230048
Download: ML20245E857 (56)

Text

,_ - - -- - - _ - _ . - - _ - - - - _ .

..-. e

'. SS REQ

'o UNITED STATES 8 p( e NL9 LEAR REGULATORY COMMISSION

. { .y jf WASHINGTON D.C.20556 -

L

%..... 'JAN 4" 1989 -

MEMORANDUM FOR: Edward L. Jordan, Chairman i Committee to Review Generic Requirements FROM: Eric S. Beckjord, Director Office of Nuclear Regulatory Research

SUBJECT:

CRGR REVIEW 0F THE MARK I CONTAINMENT PERFORMANCE IMPROVEMENT PROGRAM Enclosed is a revised Connission paper containing recommendations of the Mark I Containment Improvement Program. -The CRGR reviewed this program on December 14, 1988 based on information provided by memorandums dated December 7, 1988 and December 9, 1988. The paper has been revised to reflect comments provided by the CRGR, the ACRS and additional input from NRR and OGC.

MO or' Changes to the paper since the CRGR review are:

1. The recommended method of implementation is through rulemaking.

2. A sensitivity study has been added to the regulatory analysis (section 4.1.7 of enclosure 4) evaluating the cost-benefit of the incremental addition of the ADS and backup water supply improvements. In addition, the regulatory analysis now uses 25 years remaining plant life (versus 20 years'previously) which is representative of Mark I pleits with a 40 year license.

3. The relationship of the power requirements of the proposed improvements and the power requirements of the Station Blackout Rule is explicitly stated.

An expedient review of this package is requested in order to meet a January 19, 1989 schedule to provide these recommendations to the Commission.

/ s Nh Eric S. Beckjord, Director Office of Nuclear Regulatory Research

[

l l

I-het2seeif SG PP '

L___-_______

\.'

For: . The Connissioners From: Victor Stello, Jr.

Executive Director for Operations

Subject:

MARK I CONTAINMENT PERFORMANCE IMPROVEMENT PROGRAM

Purpose:

To present staff recommendations on Mark I containment performance improvements and other safety enhancements.

Category: This paper covers a major policy question.

Sunnary: As noted in the Integration Plan for Closure of Severe.

Accident Issues (SECY 88-147) and in interim reports to the Commission (SECY 87-297 and SECY 88-206), the staff has undertaken a program to determine what actions,-if any, should be taken to reduce the vulnerability of containments to severe accident challenges. The containment performance improvement effort is one main element of the integrated approach to closure of severe accident issues. Staff efforts have focused initially on BWR plants with a Mark.I contain-ment. The staff has now completed its assessment of generic.

severe accident challenges.and' failure modes as well as potential improvements for plants with the Mark I containment.

The review of !! ark II, Mark'III, and other containment

! types are the subject of parallel but separa e program efforts, as discussed in SECY 88-147.-

ProbabilisticRiskAssessment(PRA)studieshavebeen i performed for a number of BWRs with !! ark I containments.

These studies indicate that.BWR Mark I risks are dominated by Loss of Decay Heat Removal, Station' Blackout (SBO),'and Anticipated Transient Without Scram (ATWS) sequences. Although these studies do not show the BWR Mark I plants to be risk

outliers as a class relative to other plant designs, they do suggest that the Mark I containment integrity could be challenged by a large scale core melt accident, principally due to its smaller size. However, estimates of containment failure likelihood under such conditions are based on analysis of complex accident conditions, where there remains a broad band of uncertainty.

The staff has concluded that the optimum way to reduce overall risk in BWR Mark I plants is to pursue a balanced

- approach utilizing accident prevention and mitigation.

p

Contact:

W. Beckner, RES 492-3975 L. Soffer, RES 492-3916 f

4 4'

i Based on our. assessment including the above described-balanced approach, the staff. recommends five specific

-improvements'for Mark I containment plants: 1)an. improved'

-hardened vent capability, 2)-improved automatic depressur-

-ization system (ADS) ~reliabi'lity, 3) an alternate water supply to the reactor' vessel' and drywell' sprays, 4)' extended emergency procedures and training and 5) accelerated imple-mentation of the' existing ATWS and SB0 rules. These im-provements, when fully implemented, will substantially enhance the safety of Mark I plants',' including improvement to containment performance. The staff has evaluated them and found them to be cost effective. The. staff proposes-that a rulemaking be initiated to implement these improve-ments for all licensees with Mark I containments.

Background:

The Reactor Safety Study (WASH-1400) found that, for the Peach Bottom BWR Mark I nuclear plant, even though the. core melt probability was'relatively low, the containment could-be severely challenged if a large core melt occurred.

Based-on this conclusion and reinforced by the anticipation-confirmed)inthedraft of similar Reactor findings Risk (subsequently Reference Document (NUREG-1150, February 1987) a five element program was proposed.in June 1986 to enhance the_ performance of the BWR Mark I containment, Elements'of-this proposal included 1) hydrogen control, 2) containment drywell contro.1,spray,)3) containment and 5 emergency venting, procedures 4) coreAfter and training. debris the initial proposal, the staff held two separate meetings in early 1987 with researchers representing NRC contractors and industry. There was a wide range of views expressed.

D

[

regarding accident phenomenology as'well as the efficacy of the various improvements. In view of the lack of technical i:

consensus on the effectiveness of the proposed improvements, the staff decided to undertake additional efforts. In July 1987, the staff informed the Commission of its intention to examine the Mark I issue in the context of an integrated approach to the closure of severe accident issues.

l On December 18, 1987, thestaffissuedaplan(SECY87-297) i for resolving generic severe accident containment performance issues for Mark I and other containment types. As part of the plan, a workshop was held on February 24-26, 1988 to discuss a number of issues associated with Mark I containment challenges, failure modes and potential containment improvements a

with researchers, industry representatives and interested members of the public. A major topic at the workshop was the phenomena associated with containment shell meltthrough as discussed in Enclosure 6. The Integration Plan for Closure of Severe Accident Issues, (SECY 88-147) characterizes the containment performance improvement effort as being one i of the main elements of the integrated approach to closure of severe accident issues. Other main elements include Q__--_-_

3 a)IndividualPlantExaminations(IPEs),b)improvedplant operations, c) the severe accident research program, d) examination of external events, and e) a program on accident management. The containment performance improvement program is related to the IPE effort, and is considered complementary to it, since this effort is primarily focused on the potential generic vulnerabilities of specific containment classes, whereas the IPE effort is focused on plant unique vulner-abilities. ,

ACommissionpaper(SECY88-206)datedJuly 15, 1988 provided a status report on the staff's efforts regarding the Mark I containment. This paper reaffirmed that the risk from BWR Mark Is is low. Nevertheless, the staff proposed a program intended to further reduce overall risk in BWR Mark I plants by pursuing a balanced approach involving accident prevention and mitigation. A number of safety enhancements were identified which appeared attractive in terms of their potential risk reduction capability as well as implementation costs.

Following that meeting the Commission requested additional information via a staff requirements memorandum dated August 1, 1988. Responses to these questions are included as Enclosure 1.

Discussion: Probabilistic Risk Assessment (PRA) studies for BWRs indicate that accidents initiated by transients rather than Loss-Of-Coolant-Accidents (LOCAs) dominate the total core damage frequency estimates. The principal accident sequences l for BWRs consist of Long-term Loss of Decay Heat Removal (TW), Station Blackout (SBO), and Anticipated Transient Without Scram (ATWS). WASH-1400 indicated that TW is the dominant core damage accident sequence for Peach Bottom.

Draft NUREG-1150, however, indicated that the dominant contribution to core melt frequency at Feach Bottom is due to Station Blackout, and estimated that TW has been greatly reduced at Peach Bottom by implementation of containment i venting procedures with the assumption that said venting actions can be successfully accomplished. For those plants in which TW has been eliminated as the dominant contributor, the residual risk is largely due to ATWS and SB0 sequences.

These studies also indicate that the estimated likelihood of l

core damaging accidents for existing Mark I plants is predicted to vary widely over two orders of magnitude or more. The primary containment challenges and potential c

l failure modes for BWR Mark I containments are shown in g

Enclosure 2.

I The staff has examined potential Mark I containment and plant improvements in the following six areas: (1) hydrogen control, (2) alternate water supply for reactor vessel I injection or containment drywell sprays, (3) containment pressure relief capability (venting), (4) enhanced ADS reliability,(5)coredebriscontrols,and(6) procedures i

y j 1 4

/

and training. -Each of these was evaluated to determine their potential benefits in terms of reducing the (1) core melt frequency, (2) containment failure probability, and (3) offsite consequences.

Hydrogen Control:

Although BWR Mark Is are required to be operated with an, inerted containment atmosphere, plant Technical Specifi-cations permit de-inerting to commence 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> prior to plant shutdown, and do not require inerting to be completed until 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after plant startup, in order to permit plant personnel access. In the event of a severe accident, such as a long-term station blackout, a concern was expressed that loss of control of the valves and containment leakage could eventually lead to containment de-inerting.

Two potential improvements with regard to hydrogen control were evaluated. These were: (1)eliminationofthetwo24 hour de-inerted periods and (2) providing a backup supply of nitrogen. Since the probability of a severe accident occurring during either of the two 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> de-inerted periods is small compared to the probability of accident occurrence during normal operations, eliminating this time of de-inerting would not significantly reduce risk.

During a severe accident, reactor pressure is anticipated to increase, releasing steam and non-condensable gases into the containment. This will increase containment pressure, preventing ingress of air. Therefore, the containment atmosphere would not become de-inerted for an extended period of time. Sinr.e offsite supplies of nitrogen could readily be obtained during this period, an onsite backup supply of nitrogen would not significantly reduce risk.

l Therefore, the staff concludes that additional Mark I improvements to control hydrogen beyond the existing hydrogen control rule and the procedures in Revision 4 of the Emergency Procedure Guidelines would have no significant benefit and are not warranted.

Alternate Water Supply for Drywell Spray / Vessel Injection An important proposed improvement would be to employ a backup or alternate supply of water and a pumping capability that is independent of normal and emergency AC power. By connectin residual heat removalRHR) (g this source system to the as well low as to thepressure existing drywell sprays, water could be delivered either i F

into the reactor vessel or to the drywell, by use of an

! appropriate valving arrangement. ,

hu _._.__ _ _ _ _ _ -

4 L 5 An alternate source of water injection into the reactor vessel would greatly reduce the likelihood of core melt due to station blackout or loss of long-term decay heat L

! removal, as well as provide significant accident management capability.

Water for the drywell sprays would also provide significant mitigative capability to cool core debris, to cool the containment liner to delay or prevent failure, and to scrub-air borne particulate fission products from the atmosphere.

A review of some BWR Mark I facilities indicates that most plants have one or more diesel driven pumps which could be used to provide an alternate water supply. The flow rate using this backup water system may be significantly less than the design flow rate for the drywell sprays. The potential benefits of modifying the spray headers.to assure a spray were compared to having the water run out of the spray nozzles. Fissior. product removal in the small crowded volume in which the sprays would be effective was judged to be small compared to the benefit of having a water pool on top of the corium. Therefore, modifications to the spray nozzles are not considered warranted.

ContainmentPressureReliefCapability'(Venting): .

Venting of the containment is currently included in BWR emergency operating procedures. The vent path external to existing containment penetrations typically consists of a

- ductwork system which has a low design pressure of only a few psi. Venting under high pressure severe accident.

conditions would fail the ductwork, release the containment atmosphere into the-reactor building, and potentially contaminate or damage equipment needed for accident recovery.

In addition, with the existing hardware and procedures at some plants, it may not be possible to open or to close the vent valves for some severe accident scenarios. The staff has concluded that venting, if properly implemented, can have a significant benefit on plant risk. However, venting via a sheet metal ductwork path, as currently implemented at some Mark I plants, is likely to greatly hamper or com-plicate post-accident recovery activities, and is therefore viewed by the staff as yielding reduced improvements in safety. The capability to vent has long been recognized as important in reducing risk from operation of BWR Mark I facilities for loss of long term decay heat removal events. Controlled venting can prevent the failure cf ECCS pumps from inadequate NPSH and re-closure of th'e ADS

!- valves. The staff agrees with this view as long as the potential downsides of using the existing hardware are  ;

corrected.

l: A hard pipe vent capable of withstanding the anticipated )

severe accident pressure loadings would eliminate these

l i

• 1 i

6 I disadvantages. The vent isolation valves should also be f remotely operable from the control room and should be j i

prcvided with a power supply independent of normal or J emergency AC power. Other changes, such as raising the vent valve operability pressure and/or raising the RCIC turbine back pressure trip setpoint, may also be desirable and should be considered as part of the IPE. This capa-bility, in conjunction with proper operating procedures and other improvements discussed in this paper, would result in greatly reducing the probability of core melt due to the TW and SB0 sequences.

Given a core melt accident, venting of the wetwell would provide a scrubbed venting path to reduce releases of particulate fission products to the environment. Venting l has been estimated to reduce the likelihood of late containment over-pressure failure and to reduce offsite consequences for severe accident scenarios in which the containment shell does not fail for other reasons. Failure of the shell due to corium attack (shell meltthrough) would reduce the benefits from venting in that it would release fission products directly into the reactor building.

Inadvertent venting could result in the release of normal coolant radioactivity to the environment even when core degradation is averted or vessel integrity maintained.

fleasures to reduce the probability of inadvertent venting, such as a rupture disk, should be considered in the vent design.

Enhanced ADS Reliability:

The Automatic Depressurization System (ADS) consists of relief valves which can be operated to depressurize the reactor coolant system. Actuation of the ADS valves requires DC power. 7 an extended station blackout after station batteries hav= been depleted, the ADS would not be available and the reactor would re-pressurize. With enhanced ADS reliability, depressurization of the reactor coolant system would have a greater degree of assurance. Together with a low pressure alternate source of water injection into the reactor vessel, the major benefit of enhanced ADS reliability would be to provide an additional source of core cooling which could significantly reduce the likelihood of high pressure severe accidents, such as from the short-term station blackout.

Another important benefit is in the area of accident' mitigation. Reduced reactor pressure would greatly reduce the possibility of core debris being expelled under high pressure, given a core melt and failure of the reactor pressure vessel. Use of the ADS would also delay I containment failure and reduce the quantity and type of fission products ultimately released to the environment.

In order to increase reliability of the ADS, assurance of electrical power beyond the requirements of existing I

7

~

regulations may- be necessary as ' discussed later in this .

paper. In addition, performance of the ADS cables needs  !

- to be reviewed for temperature capability during a severe accident.

Core Debris Controls:

Core debris controls, in the form of curbs in the drywell and/or curbs or weir walls in the torus room under the wetwell have.been proposed in the past to prevent.

containment shell meltthrough and to retain sufficient water to permit fission product scrubbing. However, as noted in SECY 88-206, the technical feasibility for such controls has not been established, and the design and installation costs as well as the occupational exposure during installation could be significant. The staff intends to pursue research programs to evaluate the need for and feasibility of core debris controls. There is a growing consensus that water in the containment (from an alternate supply to the drywell sprays) may help mitigate risk either by fission product scrubbing or by ?reventing or delaying shell melt by core debris. Researca is continuing in order to confirm and help quantify these initial conclusions.

A discussion of Mark I shell melt phenomena and the current state of knowledge is included in Enclosure 6.

Emergency Procedures and Training:

A major element of the Mark I containment performance improvement evaluation involves emergency procedures and training. Current emergency operating procedures (EOPs) are symptom-based procedures that originated from require-ments of TMI Task Action Plan item I.C.1. Plant-specific E0Ps are generally implemented based on generic Emergency Procedure Guidelines (EPGs) developed by the BWR Owners Group. As part of the balanced approach to examining potential BWR Mark I plant improvements, both the generic EPGs and the plant-specific implementation of E0Ps and training have been examined.

NRC has recently reviewed and approved Revision 4 of the BWR Owners Group EPGs (General Electric Topical Report NED0-31331, BWR Owner's Group " Emergency Procedure Guidelines, Revision 4," March 1987). Revision 4 to the BWR Owners Group EPG is a significant improvement over earlier versions in that they continue to be based on symptoms, they have been simplified, and all open it' ems from previous versions have been closed. The BWR EPGs extend well beyond the design bases and include many actions appropriate for severe accident management.

k The improvement to EPGs is only as good as the plant-I specific E0P implementation and the training that operators Ci___-_

1 receive on use of the improved procedures. A recent staff safety evaluation report (Ltr. Thadani to Grace, " Safety Evaluation of 'BWR Owners' Group - Emergency Procedure Guidelines, Revision 4,' NE00-31331, March 1987," dated September 12,1988) encouraged licensees to implement Revision 4 of the EPGs and reiterated the need for proper implementation and training of operators. Implementation of the guidelines has been voluntary, but is strongly recommended in the SER.

Impact of Existing Requirements:

As part of the balanced approach, for completeness, and to provide a more accurate picture of Mark I plant risk, the staff has also evaluated the impact on Mark I risk of several recent rules that have been imposed on light water reactors - the Station Blackout Rule and the ATWS Rule. As discussed earlier, PRAs typically indicate that Mark I reactor risks are dominated by TW, SB0 and ATWS sequences.

Upon implementation of these two rules at all Mark I plants, risk from SB0 and ATWS sequences would be expected to be reduced to a low level. The response to Question #2 in Enclosure 1 provides a discussion cf expected risk reductions from changes to Mark I plants as a result of these rules.

Assuring the operability of the proposed improvements under severe accident conditions, including an extended period of station blackout, may require assurance of electrical power beyond the requirements of the recent StationBlackout(SBO) rule,10CFR50.63. The proposed improvements have been coordinated with the requirements of this rule in order not'to cause an undue proliferation of power supplies, which could be counter-productive to safety. The staff proposes that licensees intending to implement the SB0 rule by use of an alternate AC (AAC) source, need not provide additional electric power supplies for the proposed Mark I improvements, provided that the capacity of the AAC is sufficient for the requirements of '

both the SB0 rule and the improvements proposed here.

Further details are given in Enclosure 7.

Benefit of Improvements:

The improvements (1) an that the improved hardened staff is venting recommending) capability, include:

(2 improved ADS reliability, (3) an alternative water supply to the reactor vessel and drywell sprays, and (4) emergency' proce-dures and training. Accelerated implementation of the existing station blackout and ATWS rules is also planned.

These improvements are unchanged from those indicated in the interim report (SECY 88-206) to the Commission.

A major benefit of these improvements is that they can ,

l provide a reduction in core melt frequency of about a {

factor of five to ten.

l

~

L Withtheproposedenhancements,thecoremeltfrequgncy would be expected to be reduced to about I to 2x10 per reactor-year. .It should be noted that these estimates apply to internal. events only.

For plants with a high.TW probability, a large fraction of the reduction in core snelt frequency is attributable to improved venting which, by allowing the removal of long-term decay heat from the containment, greatly reduces the likeli-hood of core melt from the TW sequence. Another reduction in core melt frequency from station blackout is attributable to the enhancements taken together. In the event of station  ;

blackout, enhanced ADS reliability would permit depressurization of the reactor, availability of a low pressure backup source of water injection into the vessel would permit core cooling, while venting would allow decay heat removal from the con-tainment. It is important to note that under these circum-stances, venting would prevent core damage and not result in releases of fission products of any significance.

Accident mitigation benefits are also considered to be significant. Mitigation of fission product releases would be realized for all accident sequences, including ATWS.

Venting would be effective in preventing containment failure arising from slow over-pressurization. Venting via the suppression pool would provide significant scrubbing of non-noble gas fission products by about a factor of 10 to 100 if no containment shell failure occurs. Water in the drywell may be effective in preventing or at least delaying failure of the shell by molten core debris. Finally, even~

if shell failure should occur, the presence of a water layer atop the core debris combined with the drywell spray would reduce any source term releases to the environment by a factor judged to range from 2 to 10.

Because of the combination of reduced core melt likelihood, reduced fission product releases due to mitigation, and possible reduction or elimination of a significant contain-ment failure mode, the staff concludes that the overall risk reduction of the proposed improvements is in excess of one order of magnitude.

The benefits of the proposed enhancements in terms of their l

reduction in offsite risk can be calculated in terms of person-rem. Depending upon the probability of core melt due to the TW sequence the estimated reduction in risk, expressed in person-rem, for the proposed enhancements j.

ranged from about 145 person-rem per reactor-year to about i

1330 person-remperreactor-year,forplantghavinga probability of core melt due to TW of 1x10' per reactor-year and 1x10~4 per reactor-year, respectively. Of this ,

I f

C-_ _

10 total value, the risk reduction produced by lowering the likelihood of core melt due to station blackout and mitigation of ATUS accounts for a reduction of about 33 to 210 person-rem per reactor-year. For plants whose robabilit p(about 10~y of core melt due to the TW sequence is highp can be attributed to the large reduction in the TW sequence brought about by improved venting. Additional details are provided in Enclosure 4 Finally, as noted earlier, the recommended improvements i

form a package in the sense that they complement one another in prevention or mitigation. This results in the maximura risk reduction when all are taken together.

Summary of Costs of Improvements:

Cost estimates were made of the proposed improvements.

These are given in Enclosure 3 which provides a summary for all improvements that includes high and low estimates ranging from $3.1 to $1.6 million dollars. For purposes of the regulatory analysis included in Enclosure 4, a best estimate cost of $2.0M has been used. Estimates of cost as high as $7.3M were obtained based on actual costs of similbr improvements at an existing Mark I plant. Actual costs at many plants may be less since, as shown in Enclosure 5, l

some plants already have many features of the proposed improvements.

l 1

Conclusions:

Many of the proposed enhancements would require plant backfits. The staff has examined these in light of the backfit rule, 10 CFR 50.109. Section (a) 3 of that regulation indicates that the Commission shall require backfitting only when "there is a substantial increase in the overall protection of the public health and safety" and "that the direct and indirect costs ... are justified in view of this increased protection".

In reaching a conclusion with respect to the first test indicated above, the staff considered the effect of the l proposed enhancements upon reductions in core melt frequency and improved containment performance. A major benefit of these enhancements is in their ability to reduce the likelihood of core melt. Core melt frequencies for BWR I

Mark I plants prior to any of the enhancements considgred would be expected to range from about 1x10~ to 2x10- per t'eactor year. With the combined enhancements, core melt i

i frequency would be reduced by about a factor of five to ten. Thus, the proposed enhancements clearly offer a L

substantial reduction in core melt frequency. The core melt frequency reductions do not give credit for existing p

venting capability assumed in NUREG-1150 since the current venting capability at plants has significant uncertainty regarding its overall effectiveness.

11 The increased ability to cool core debris and to remove excess heat from the containment by venting, given the occurrence of an accident, is also expected to reduce the likelihood of containment failure, although this is not as readily quantifiable because of the uncertainty in core melt progression and shell meltthrough phenomenology which is discussed in Enclosure 6. In addition, the ability to scrub particulate fission products by use of venting through the suppression pool and by the use of a water layer atop any core debris also adds significant mitigative capability.

Since the proposed enhancements would be expected to reduce the likelihood of core melt by about a factor of five to ten, and provide significant additional accident mitigation capability as well, the staff concludes that the proposed 4 enhancements do provide a substantial increase in the  ;

overall protection of the public health and safety.

With regard to the second or cost-benefit test required by the backfit rule, the discussion given earlier has shown that the costs of the enhancements are estimated to range from 1.6 to 3.1 million dollars per plant, although similar improvements at an existing Mark I plant may have cost about 7.3 million dollars. Both the estimated cost and the cost associated with an existing Mark I plant were used in the cost-benefit analysis. Based on the survey results for nine Mark I plants, the staff believes that many plants have some of these improvements already in place. Since the estimated benefits ranged from 3.6 toJ9'73 million dollars per reactor based upon 1000 dellars per person-rem and an average remaining plant life of 25 years for Mark I plants, the staff concludes that the proposed enhancements are generally cost beneficial.

For the' reasons stated above, the steff concludes that backfit of these proposed enhancements is warranted for all Mark I plants.

Options: 1. Ta'ke no action. This option would result in a situation where a number of enhancements to safety that the staff ,

believes to be cost effective would not be implemented and closure of severe accident issues would not be obtained for l Mark I plants.

2. Issue a oeneric letter. This option would be the quickest and require the least resources. The generic letter can inform industry of the staff's finding and can request i licensees to make changes to their facilities.

3. Issue an order. This option could be accomplished quickly and provide a regulatory requirement to implement the I improvements. However, this type of regulatory action is i not viewed as the appropriate vehicle for generic  ;

requirements such as the proposed improvements.

I 12-4 Initiate Rulemakinc. This option would require some staf1 resources anci cause a delay in implementing the proposed improvements. However, it would provide a firm regulatory basis for requiring the improvements.

Recommendations: The staff recommends that it initiate a rulemaking to require the improvements. As part of the rulemaking..the staff will prepare an Environmental Assessment of venting of the containment using the improved hardware and procedures.

A draft proposed rule is attached as Enclosure 7.

Coordination: OGC has no legal objections. The ACRS has reviewed these recommendations and will provide their comments separately.

Victor Stello, Jr.

Executive Director for Operations

Enclosures:

1. Response to Commission Questions

2. Mark I Challenges and Relative Likelihood of Failure Modes

3. Summary of Costs

4. Regulatory Analysis

5. Results of Survey of Mark I Plants

6. Mark;l Liner Melt Status-

7. Draft Proposed Rule-l

-l 1

I

/

ENCLOSURE 4 i

l REGULATORY ANALYSIS l- 1.0 STATEMENT OF THE PROBLEM 1 Accidents which exceed those evaluated during the licensing of facilities (design basis accidents) have a low probability of occurrence. These accidents, known as " severe" accidents, could result in core damage or core melt. The General Electric Company has designed and constructed several BoilingWaterReactor(BWR)configurationswiththreebasiccontainmentdesigns designated as Mark I, Mark II, and Mark III. The BWRs with the Mark I containment design have the smallest free volume and have been considered to be most susceptible to severe accidents which could challenge containment integrity. The potential challenges to containment integrity were reviewed and potential enhancements were proposed to improv the probability of co.itainment survival or to reduce the possiblity of a severe accident.

Draft NUREG-1150 I evaluated the dominant accident sequences for five plants, one of which was a BWR Mark I. The dominant accident sequences were identified as station blackout (TB), which includes the loss of all AC and DC power; anticipated transient without scram (TC); and would have included the loss of long term decay heat removal (TW) except that for the particular plant being reviewed this sequence was considered to be non-dominant due to assumed successful venting of the containment. For severe accidents initiated by a station blackout,' a11 existing systems are assumed to fail due to a lack of electri-city. The short term station blackout fails ai, AC and DC power sources immediately while the long term station blackout has immediate failure of all AC power sources and failure of all DC power sources after 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />.

4 INUREG-1150, " Reactor Risk Reference Document", Draft, February 1987.

1

(

l ri _ _ _ . _ _ _ _ _ _ _ _ _ . _ _ . _ _ . _ _ _ _ . _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _

y-  ; .; .

= x w, .==w = ---------------==:-----===---

, Probabilistic Risk Assessment (PRA) studies have been performed for a number of BURS with Mark I containments. Although these PRA studies do not show the BWR Mark I plants to be risk outliers as a class relative to other plant designs, they do suggest that the Mark I containment could be challenged by a large scale core melt' accident, principally due to its smaller' size. However, estimates of containment failure likelihood under.such conditions are based on calculations of complex accident conditions, which contain significant uncertainty.

'2.0 OBJECTIVES The staff objective is to reduce overall risk in BWR Mark I plants by pursuing a balanced approach utilizing accident prevention and accident mitigation.

Most recent PRA studies indicate that BWR Mark I risk is dominated by loss of decay hedt removal' (TW), station' blackout (TB), and anticipated transient with outscram(TC) sequences. The balanced approach includes: (1) accident prevention - those features or measures that are expected to reduce the likelihood of an accident occurring or measures that the operating staff can use to control the course of a accident and return the plant to a controlled, safestate,-and(2)accidentmitigation-thosefeaturesormeasuresthatcan reduce the magnitude of radioactive releases to the environment in the event of an accident. The containment performance ic:provement program would provide enhanced plant capabilities and procedures with regard to accident prevention and mitigation.

3.0 ALTERNATIVE RESOLUTIONS Plant modifications are being proposed to reduce the probability of or to mitigate the consequences of a severe core melt accident which consists of modifications to three existing plant systems. The modifications considered are (1) venting of the wetwell, (2) a backup water supply for the residual heat q removal system and the containment sprays, and (3) assuring the operability of theautomaticdepressurizationsystem(ADS). Other modifications were considered, such as additional hydrogen controls, but were not considered to

! significantly reduce either the probability of a severe accident or

i. 2

4 consequences given the occurrence of a severe accident. The details of each j

proposed enhancement is described in this section and estimates of the enhancements' benefits in Section 4.

For all prcposed modifications, the new components need not be safety-grade or safety-related. However, no failure of a modified system or non-safety-related component for design basis accidents is to adversely affect any safety-related structure, systrin, or component.

The effects of the proposed enhancements were evaluated by using a simplified containment event tree (S-CET) for station blackout events with 15 to 20 of the 107 top events used in draft NUREG-1150. Tne development of the S-CET and corresponding branch point split-fractions relied heavily on the data and insights generated by the draft NUREG-1150 effort. However, instead of trying to consider the entire range of possibilities and their uncertainties, the S-CET assigned best-estimate branch point probabilities. While this approach produced a point estiraate of the risk and does not identify the range of uncertainty in the calculations, it provides a concise and flexible model which was easily used to perform sensitivity studies. The Nsults of the S-CET identified each specific event tree end-state and its associated probability.

These end-states were compared with the similar accident progressions from the list of Peach Bottom accident progression bins.2 The end-states from the S-CET )

were characterized according to the draft NUREG-1150 accident progression bin format and then compared and assigned to the best-match accident progression bin. This process reduced the number of source terms that needed to be evaluated. Once the S-CET and states were related to those identified in draft NUREG-1150, the consequences were taken directly from draft NUREG-1150 or scaling factor.; were applied to the draft NUREG-1150 results and interpolated to generate the consequences. The risks were then calculated by multiplying i

i f

I 2 Draft HUREG/CR-4551, " Evaluation of Severe Accident Risks and the Potential

{ for Risk Reduction: Peach Bottom, Unit 2", Volu;ne 3, Draft, May 1987.

3 >

L___-______

4 the plant damage state frequency, the bin probability, and the consequences of that bin together.

l To evaluate the approximate accuracy of the S-CET, the draft HUREG-1150 information related to Peach Bottom was input into the S-CET and the results compared with those of draft NUREG-1150. In all categories, the results of the l S-CET compared with those in draft NUREG-1150 within about 25% accuracy, well within the uncertainty band of draft NUREG-1150. Once verified, the advanced information related to final NUREG-1150 was used to form a'new base case and to evaluate the benefits of the proposed enhancements. Details of this methodology is documented in NUREG/CR-5225, "An Overview of Boiling Water Reactor Mark I Containment Venting Risk Implications".

3.1 Alternative (1)

Under this alternative no action would be taken.

3.2 Alternative (iijf This alternative w d.o eccelerate the implementation of the existing rules (station blackoJt and anticipated transient without scram), without any other modifications.

'3 Alternative (iii)

This alternative would involve alternative (ii) plus a hardened venting capability from the containment wetwell to the plant stack.

The proposed venting improvement would provide a wetwell vent path to the plant stack capable of withstanding the anticipated environmental conditions of a severe accident. This modification would include installation of a hard pipe from the outlet of an existing wetwell vent outboard containment isolation valve to the base of the plant stack. This pipe would be routed through a new DC operated isolation valve which would bypass the existing ductwork and SGTS.

The hard pipe to the stack could contain a rupture disk to prevent inadvertent A

l

= _ _ _ __ _ - _ _-__ _ __ _ _-__ ___ __ _ - _ .__ ___

operation and release of radioactivity. In order to vent the wetwell, all isolation devices, except the rupture disk, need to be capable of being operated without reliance on AC power. The emergency procedures would need to' be modified to provide appropriate instructions for the operator.

3.a Alternative (iv)

This alternative would involve alternative (ii) plus enhanced operability Of the ADS.

The proposed improvement'to the ADS would consist of a portable generator (and related cabling and controls) to supply DC power for actuation of the ADS valves, ensuring that ADS cables within the containment could survive the severe accident environment, and additional nitrogen gas bottles for valyn operation, where necessary. Emergency procedures would need to be modified to provide crpropriate instructions for the operator. The use of the ADS would reduce the probability of early containment failure from high pressure melt ejection. In addition, the corium would exit the reactor sooner, but would be cooler than when the reactor is pressurized, thus delaying the potential containment meltthrough due to corium attack. The cooler vessel alsc promotes plateout of non-gaseous fission products within the reactor vessel.

3.5 Alternative (v)

This alternative would involve alternative (ii) plus a backup water supply to the containment drywell sprays and as a low pressure water injection source to the reactor vessel.

The proposed improvement of the containment sprays would use an alternative water supply and pump to the residual heat removal (RHR) pump discharge line outside of the outboard containment isolation valve. At some plants, this alternate system could use an existing 1000 gpm diesel driven fire water pump f or a portable generator to power an existing water pump, such as a service water pump. In addition, spool pieces, piping, and isolation valves would be necessary to cross cnnnect the alternate water system to the RHR system.

! 5 W _ _ _ - _ -

'4 s i

Modifications may be required to some of the RHR valves to permit remote manual operation without reliance upon AC power and to bypass interlocks. The emergency procedures would need to be modified to provide appropriate instructions for the operator.

Cool water.through.the sprays will condense steam and thereby reduce the

< potential failure of the containment by condensible gas over-pressure. The water will tend to scrub the non-not.le gas fission products from the drywell atmosphere. The water which runs down the inside of the steel containment shell will tend to wash and cool the shell and may prevent re-volitalization of fission products. This shell coolirg may reduce the potential for containment failure due to over-temperature. With water on and around the corium on the drywell floor, the water will cool the corium and may assist in the formation of a crust. With a crust formed between the molten corium and the containment shell, the likelihood of containment f ailure by corium attack may be reduced.

The pool of water over the corium is also expected to reduce the fission products that could be released to the environment.

The use of a diesel powered puc.p into the RHR system provides an additional low pressure water injection system for the reactor as a preventive featura. Thus, if the reactor is at low pressure and the alternate water system is initiated in a timely manner, the alternate teater system could prevent core degradation and arrest core melt within the reactor vesse13 ,

3.6 Alternative (vi) ,

This anernative would reduce the overall risk in BWR Mark I plants by a combination of accelerated implementation of existing rules, extended emergency operating procedures and training, and potential implementation of the following hardware modifications:

3Even for an alternate water supply system that does not provide an adequate I amount of water to prevent core degradation, the alternate system would delay severe core aamage and thereby increase the likelihood of recovery of a system to arrest core failure and prevent vessel failure.

i 6

\

u=__________________

4 i

'_(1) Containment drywell. spray: assurance of a backup water supply to the residual heat removal (RHR) system and.drywell sprays with AC -

. independent pumping capability, (2)'Containmentventing: a hardened containment wetwell venting L

capability with the ability to_open and reclose the_ isolation valves independently of AC power, and (3) Improved reliability of the automatic depressurization system (ADS):

providing additional DC-power for the solenoids, upgrade of cables, and additional nitrogen gas supply.

- 4.0 CONSEQUENCES u

4e1 Costs and Benefits of Alternative Resolutions PRAs that the. staff has available have been performed on 9 BWR Mark I plants.

These' assessments are of varying quality and, in some cases, have considered both external and internal event core melt initiators. The range of core melt -

frequencies for these' units range from 3 x 10-4/ Reactor-Year (RY)4 to 8.2 x 10-6/RY5 . We have recently' received the Brunswick PRA which has a core melt frequency of 2.5 x 10-6/RY. However, the staff has not completed the evalua-tion of this PRA.

WASH-1400 identified the dominant accident sequence to be the loss of long term decay heat removal (TW). Peach Bottom core melt frequency due to TW is 6

estimated to be 1x10-5/RY or less without assured venting ,

Core melt frequency from A-45 studies for internal events only.

S ibid, #1.

6 I The existing hardware consists of low pressure ductwork from the outboard f

containment isolation valve, through the standby gas treatment system (SGTS) to I i

the plant stack. The ductwork is designed for less than one psig interr.a1 pressure while the venting procedures identify venting containment when the containment pressure is 60 psig.

7 I

i I

For purposes of this regulatory analysis, it is assuntd that all 24 BWR Mark I plants which would use low pressure rated ductwork as part of the containment vent path would have a core melt frequency associated with TW between 1 x l 10~4/RY and 1 x 10-5/RY. NUREG-1150 has estimated that venting may reduce the probability of the TW sequence for peach Bottcin by three orders of magnitude.

The other two dominant accident sequences are station blackout (TB) and anticipatedtransientwithoutscram(TC). Proper implementation and compliance with the existing station blackout and ATWS rules is assumed to reduce the probability of the TC and TB sequences to less than 1 x 10-5 /RY. Thus, for the purposes of this regulatory analysis, we assume that the plant core melt frequency for all Mark I plants would be less than 1 x 10~4/RY and probably greater than 2 x 10-5jgy, 4.1.1 Alternative (i)

This alternative would be to take no action. At least one licensee has seen the need to provide improved accident management capabilities and, thus, defense-in-depth. Because the value-impact analysis has shown that it would be beneficial to implement the recommendations identified in alternative (vi) which provides for defense-in-depth for accident management, the no-action alternative is not recommended.

4.1.2 Alternative (ii) 4.1.2.1 Value: Risk Reduction Estimates For the BWR Mark I plants, the acceleration of compliance by one year represents a risk savings of approximately 1392 man-rem as identified in NUREG-1109. This could reduce the core melt frequency associated with TB sequences by an estimated 2.6 x 10-5, however, there is no effect on the dominant accident sequence of TW and thus there wo91d be no overall benefit when compared to alternative (vi).

l 4.1.2.2 Impacts: Cost Estimates L

b l

l l

l

'4.

. Implementation of the modifications required by the. ATWS rule has been accelerated and are. scheduled to be complete by the end of 1989.,.The station blackout rule could be accelerated to reduce the time required until compliance is achieved by possibly one year. This could be. achieved'

.by expeditious and timely staff review and approval of licensee submittals.

Thus, this could require expert staff to review the submittals. Accelera' tion

~

of compliance by one year represents an estimated cost to licensees of $1.4 million.

4.1.2.3 Value-Impact Ratio The overall value-impact ratio of this alternative is about 990 man-rem averted per million dollars. Because this alternative does not reduce the consequences of the dominant accident sequence (TW) nor the probability of the occurrence of the dominant accident sequence, this' alternative is not recommended.

4.1.3. Alternative (iii) 4.1.3.1 Value: Risk Reduction Estimates The alternative of installing a hardened vent capability from the contain-ment wetwell to the plant stack, in addition to accelerating the implemen-tation of the TB and TC rules, would result in a reduction in core melt frequency for TW sequences. With an independent source of power for remote operation of the valves, it would result in a reduction in core melt frequency in the range from 1x10~4 to 1x10-5 per reactor-year by reducing the contribution of the TW sequence to the total core melt frequency to be an insignificant contributor. The corresponding reduction in risk is approximately 1120 to 112 man-rem per reactor year.

i-g i

o

-I L '4.1.3.2 Impacts: Cost Est hates p

The estimated costs.for installation of the hardened vent system ranges 7 8 frcm $690,000 to $2,909,000 per plant for an estimated industry costs b from $16.6 million to $f9.8 million.

L -

i

[ '4.1.3.3 Value-Impact Ratio L The overall value-impact ratio of this alternative is from about 4060 to 9630 man-rems averted per million dollars. While the value-impact ratio indicates that this is a cost effective alternative, it is not recommended l because it does not provide any defense-in-depth for TB or TC events.

t-l

-4.1.4 Alternative (iv) 4.1.4.1 .Value

Risk Reduction Estimates i

This alternative.would provide enhanced operability of,the ADS, in f

l' additiontoalternative(ii). It would reduce the risk to the public by t

an estimated 5.7 man-rem per reactor year, but would not reduce the core I melt frequency beyond that provided by alternative (ii). The availability of the ADS would eliminate early containment over-pressure /over-temperature failure due' to direct containment-heating by changing the high pressure station blackout sequence (high pressure melt ejection) to a low pressure station blackout sequence.

i

\

i '4.1.4.2 Impacts: Cost Estimates I'

\;

[

7 Costs estimated by Science and Engineering Associates and documented in' SEA

[

Report 87-253-07-A:1, dated November 1988.

8 Cost derived from information provided by Boston Edison Company (DPU 88-28, f Request No. AG 13-6) and does not include costs related to Technical L Specification changes revising procedures or training manual, training, or NRC

$ costs.

10 u

1 4

The installation cost of enhancing the ADS has been estimated to range 10 from $500,0009 to $1,993,000 per plant for an estimated industry cost of

$12 to $47.8 million.

4.1.4.3 Value-Impact Ratio The overall value-impact ratio of this alternative is from about 290 to 72 man-rem averted per million dollars. Because this alternative is not cost effective, does not reduce the probability or consequences of the duninant accident sequences, and does not provide defense-in-depth, this alternative is not recommended.

4.1.5 Alternative'(v) 4.1.5.1 Value: Risk Reduction Estimates This alternative would provide a backup water supply system for the containment sprays and as an alternate low pressure water injection system for the reactor vessel, in addition to alternative (ii). It would provide no reduction in the probability of severe accident sequences where the reactor remains at high pressure, such as the short term station blackout scenario. However, it would delay core heatup for the long term station blackout scenarios, i.e. where the ADS has been operating, until the safety-relief valves (SRVs) are reclosed due to high containment pressure.

The reduction in core melt frequency related to TB due to the backup water supply is estimated to be approximately 9.5x10~7 per reactor-year with a ,

corresponding reduction in risk to the public of approximately 5.4 man-rem per reactor-year. Using the backup water supply to spray inside the containment drywell will not affect core melt frequency but could reduce the risk to the public by approximately 6 man-rem per reactor year.

4.1.5.2 Impacts: Cost Estimates i

9 ibid, #7 10 ibid, #8 11 l

i r

11 12 The cost of this alternative is in the range of $810,000 to $2,438,000 with the resulting industry costs in the range of $19.4 to $53.5 million, f

l

.4.1.5.3 Value-Impact Ratio The overall value-impact ratio of this alternative is from about 190 to 62 man-rem averted per million dollars. Because this alternative is not cost effective, does not reduce the probability or consequences of the dominant accident sequences, and does not provide defense-in-depth, this alternative is not recommended. See Section 4.1.7 for a discussion of improved ADS reliability combined with a backup water supply.

l t

II ibid, #7 12 ibid, #8 12 t

L_______.____

u 1

TABLE 1 CostBenefitsofAlternatives(1)-(v)

(Man-rem averted per million dollars )

Alternative (i) - do nothing 0 990 Alternative (ii) - Accelerate Rule Implementation (ARI)

Alternative (iii) - ARI & venting 4060 to 9630*

Alternative (iv) - ARI & ADS 72 to 290 *

=

Alternative (v) - ARI & containment 62 to 190

• sprays

• Ranges due to effects of TW frequency and two installation cost estimates.

l I

13 I 1

l l

I j

4.1.6 Alternative (vi)

While each of.the proposed improvements, individually, have some benefit in prevention or mitigation of one or more severe accident scenarios, taken together the improvements have greater benefits because of the effects of the interaction each enhancement has on the others. For example, providing' containment sprays will not affect the accident sequence or timing. Even providing an alternate low pressure reactor vessel injection capability will provide no benefit when the reactor is pressurized. Combining this enhancement with improved ADS operability provides greater assurance that the reactor can be depressurized and thus permit operation of the alternate low pressure reactor vessel injection system. If injection is not posr,1ble, using the ADS results in cooler corium when it does exit the vessel and thereby improves the effect of the containment sprays. Combining this enhancement with hard pipe venting of the wetwell provides assurance that the ADS valves will operate by reducing the back pressure on the valves which could, otherwise, prevent the.

valves from opening. Venting has been identified as the means to reduce the probability of a severe accident from the loss of long term decay heat removal.

This sequence, with venting only, has been dependent upon the operators realigning the suction of the RHR pumps to a source of water other than the suppression pool prior to venting containment. With the proposed combination of enhancements, the alternate water supply system provides an independent water system that would be available even if the RHR pumps are lost due to inadequate net positive suction head (NPSH).  ;

4.1.6.1 Value: Risk Reduction Estimates For station blackout the reduction in risk has been determined to be 33 man-rem /RY and the reduction in the core melt frequency is anticipated to be 6.3 x 10-6/RY.

While it is reasonable to consider that these proposed enhancements would have benefits for the TC sequence, no credit is taken for those benefits.

For example, if the TC is not a full power failure to scram (i.e. all control rods remain withdrawn at their full power position) but is instead 14 l_

4 a partial power failure to scram (such as at Browns Ferry), the diesel driven water pump with ADS and venting may be adequate to prevent degradation indefinitely.

For those plants which have not properly eliminated TW as the dominant severe accident scenario, the benefits associated with reducing the consequences from the TW and TB sequences have been determined to range from of 1,153 man-rem /RY to 144.9 man-rem /RY for risk reduction and from 1 x 10~4/RY to 1.6 x 10-5/RY for core melt frequency reduction. These reductions corre' pond to an initial TW frequency of 1 x 10~4/RY, and 1 x 10-5/RY, respectively, and a TB frequency of 6.6 x 10-6/RY. For a plant similar to Peach Bottom with a core melt frequency of 1.8 x 10-5/RY(which includes a TW f requency of 1 x 10-5/RY), the emergency procedures and operator training is expected to reduce the core melt frequency to approximately 3.0 x 10-6/RY for a net core melt frequency reduction of approximately 1.6 x 10-5/RY.

4.1.6.2 Impacts: Cost Estimates Installation of a hard pipe vent in a plant similar to Peach Bottom has been estimated to cost $690,000 13 A similar installation at Pilgrim has cost $2,909,000 I4 .

Installation of the backup water supply for containment sprays and low pressure injection into the reactor vessel at a 15 A

plant similar to Peach Bottom has been estimated to cost $810,000 .

16 similar installation at Pilgrim has cost $2,438,000 . Installation of the supplemental power supply and nitrogen gas supply at a plant similar 17 to Peach Bottom has been estimated to cost $500,000 . A similar 13 1 bid, #7 I4 ibid, #8 ibid, #7 16 ibid, #8, except that half of the costs are used because the spray nozzles j are not to be modified.

17 1 bid, #7 15

e 18 installation at Pilgrim has cost $1,993,000 . Together, the installation of these proposed modifications at a plant similar to Peach Bottom has been. estimated to cost $2,000,000 19 A similar insta11ation'at Pilgrim 20 The estimated. total cost for industry (for the 24 has cost $7,340,000 .

MarkIplants)toinstalltheproposedenhancementsrangesfrom$48 million 21 to $176 million 22 Actualtotalcostsmaybelesssincesom[

Mark I plants may already have some of the proposed features.

.The averted cost associated with prevention and mitigation of.an accident can be discussed as five separate costs: replacement power, cleanup, onsite health impacts,_offsite health impacts, and offsite property damage. To estimate the costs of averting plant damage.and cleanup. the reduction in accident frequency was multiplied by the discounted onsite 23 property costs. The following equations from NUREG/CR-3568 were used to make this calculation:

V = NdFU gp r

U=C/m[(e -rt(i))/r][1-e(t(f)-t(i))3(y,g-rm) 2 where:

V = value of avoided onsite property damage gp H = number of affected facilities = 24 dF

= reduction in accident frequency = 1.6 x 10-5/RY 181 bid, #8, except half of the nitrogen supply system cost is used and the additional AC and DC capacity cost is used instead of the third diesel cost.

19 1 bid, #7 20 ibid, #8, 16, and 18 L

i 'l i

ibid, #7 22 ibid, 18 23 NUREG-CR-3568, "A Handbook for Value-Impact Assessment", December 1983, pages l 3.29-3.31.

16

U present value of onsite property damage 24 C = cleanup and repair costs = $1.0 billion t(f)= years remaining until end of plant life = 25 t(i)= years before reactor begins operation = 0 r = discount rate = 10%

m = period of time over which damage costs are paid out (recovery period in years) = 10 Using the'above values, the present value of avoided onsite property damage is estimated to be $2,23 million.

Replacement power costs can_be estimated several different ways.

25 has used a cost of $500,000 per day. NUREG/CR-4012 26 lists NUREG-1109 the replacement power costs for each nuclear power reactor by season.

Using this information for only Mark I reactors 1.veraged over the four years of projected data, escalated by 6% for 1987 dollars, and normalized for the numerically average size reactor (in megawatts electric), the replacement power cost is $335,000 per day. A draft BNL report dated August 16, 1988, related to a current RES re-evaluation of the cost benefit considerations in backfit analysis, has indicated that the replacement power cost used in regulatory analysis should be $400,000 per day. Using the $335,000 per day for the 24 Mark I plants is conservative and therefore is used here. This represents a replacement power cost of

$771 millian for 10 years. Thus, the present value of avoided onsite property damage and replacement power is estimated to be $3.95 million.

24 NUREG/CR-2723, " Estimates of the Financial Consequences of Nuclear Power Reactor Accident", September 1982, page 10.

25NUREG-1109, " Regulatory /Backfit Analysis for the Resolution of Unresolved Safety Issue A-44, Station Blackout", June 1988, page 23.

26NUREG/CR-4012, " Replacement Energy Costs for Nuclear Electricity-Generating Units in the United States: 1987-1991", Volume 2, January 1987, Table S.1, pages 2 - 5.

17 l

t _ - - - _ _ _ _ _ _ _ _ _ _

6 .

,, 4 The change in public health risk associated with the installation of the The proposed enhancements is expressed as total man-rem avoided exposure.

27 following equations-from NUREG/CR-3568 were used to make this calculation:

VPH = NT (Dp x R) where:

V = value of public health risk avoided for net-PH benefitmethod($)

N = number of affected reactors = 24 T = average remaining lifetime of affected facilities (years)=25 Dp = avoided public dose per reactor-year (man-rem /RY) = 144.9 and 1153 R = monetary equivalent of unit dose ($/ man-rem) = 1000 u

I Using the above values, the present value of avoided public health exposure is estimated to be $86.9 million (or $692 million using the greatest anticipated core melt frequency for the TW sequence).

4.1.6.3 Value-Impact Ratio The overall value-impact ratio, not including onsite accident avoidance costs, is about 1810 man-rems averted per million dollars for those plants with a core melt frequency of 1.6 x 10-5/RY and 3930 man-rems averted per million dollars for those plants with a core melt frequency of 1 x 10~4/RY. If the savings to industry from accident avoidance (cleanup and I

k 27NUREG/CR-3568, "A Handbook for Value-Impact Assessment", December 1983, pages "

3.11-3.12.

18 i l 1

O Table 2 COSTBENEFITS*forAlternative(vi)

(man-rem averted per million dollars)

No Cleanup, Repair, Repl Pwr Averted Onsite Cost ** Averted Onsite Costs (10% Discount)

TW = 1 x 10-5 Low Ind. Costs 1,810 1,970 High Ind. Cost 490 500 TW = 1x 10~4:

Low Ind. Costs 14,400 29,600 High Ind. Cost 3,930 4,570

• Cost Benefit = Averted Exposure (InstallationCost-AvertedOnsiteCosts)

• • Rate discounts are not applicable when averted costs are not included in the cost benefit ratio.

Conclusions:

1. Value/ Impact is not significantly affected by assumed value of real interest rate.

2. Value/ Impact is little affected (10%) by inclusion of averted onsite costs.

3. Value/ Impact is affected by a factor of about 4 depending on estimated industry installation costs, however the Value/ Impact results support implementation of enhancements.

l 19 l

)

q ..

repair of onsite damages and replacement power which has an estimated value of $3.95 million at a discount rate of 10%) were included, the overall value-impact. ratio would be about 1970 man-rem averted per million dollars.for the low core melt frequency plants. These values, which exceed $1000/ man-rem, indicate that the proposed enhancements are cost beneficial for all the BWR Mark I plants.

4.1.7 Sensitivity Studies Table 1 indicates that alternative (iii), which primarily consists of the hardened vent system, is a cost effective alternative. The benefit of this option is largely due to a reduction in.the TW core damage frequency. A range of'TW probabilities was used in the analysis which is believed to cover the expected range of core damage probabilities due to TW at Mark I plants.

Further, the cost-benefit ratios are high enough so that other plant-specific factors wnuld not be expected to change the conclusion that the hardened vent

' system -is . cost-ef fective.

Alternative (vi) consists of alternative (iii) plus the addition of improved ADS reliability and backup water supply to the containment sprays and for vessel injection. Table 2 indicates that alternative (vi) is cost-effective, although less cost-effective than alternative (iii). This raises a question about the cost effectiveness of the incremental benefit of the ADS and backup water beyond the benefit that would be obtained with the hardened vent system alone. The incremental benefit of the ADS and backup water supply is primarily due to prevention and/or mitigation of SB0 sequences. As discussed previously, the SB0 risk reduction is estimated to be 53 man-rem /RY associated with a reduction in core melt frequency of 6.3x10-6/RY for S80. This estimate is a

-single, plant-specific value determined from simplified containment event trees using assuinptions from the NUREG-1150 study of Peach Botte. Unlike the e analysis of TW sequences, no s' sensitivities to plant-specific factors or sensitivities due to uncertainties is included.

t i

20 L...

Three major sensitivities of risk due to SB0 were examined to further investigate the incremental benefit of ADS and backup water.

1. Plant specific probability of core damage due to SBO,

2. Site population, and

-l Assumptions used in the analysis. J 3.

Estimates of SB0 core damage frequency, assuming a 4 hour4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> coping period thet would result from the SB0 rule, were determined for all Mark I plants. These values ranged from 3.5x10-5/RY to 0.1x10-5/RY with an average value for Mark I plants of 1.0x10-5/RY. This compares to the 0.6x10-5/RY used in the previous analysis which is representative of Peach Bottom. Thus, using consideration of

.SB0 frequency alone, the risk due to SB0 could be higher by a factor of five at some Mark I plants and may be at least 50% higher on average than that determined using Peach Bottom assumptions.

The second site-specific factor considered is the effect of population on the risk. The Peach Bottom site is typical of the most populated site. Less populated sites may result is risk reductions of about a factor of five due to population consideration alone. The combined effects of population and SB0 frequency would. tend to be off setting for Peach Bottom, but depending on the plant-specific values, could result in risk due to SB0 up to five times higher than the Peach Bottom value for some Mark I plants with both a high population and a high probability of SB0 sequences.

The third sensitivity involves those due to assumptions used in the analysis.

The major sensitivity in the SB0 simplified containment event trees examined was the effect of assumptions about liner melt. The event trees used a conditional probability of liner melt of 0.5 given a station blackout core damage event. The issue of liner melt was examined by experts during the NUREG-1150 program and no consensis was reached - hence the value of 0.5 which l' is an average of the differing opinions of experts. If one makes a bounding I: assumption that liner melt would be a certainty (probability = 1.0) given a core ,

21 j

melt with no water in the containment to cool the core debris, as would be the case during a SB0. Under this assumption, the risk due to SB0 would be increased from the base value of 33 man-rem /RY to 56 man-rem /RY at a core damage frequency of 0.6x10-5 for SB0 and 93 man-rem /RY at a core damage frequency of 1.0x10-5 ,

SB0 risks could be as high as 165-280 man-rem /RY for plants with a SB0 frequency of 3.5x10-5/RY depending an assumptions about liner melt.

The regulatory analysis did not previously examine the benefit of the proposed improvements due to prevention or mitigation of ATWS sequences. This was due to the a very low core damage frequency assumed for ATWS based on the NUREG-1150 Peach Bottom study. For Mark I plants with a higher ATWS core damage frequency, the backup water supply would provide mitigative benefits.

Given an ATWS, the containment is expected to fail and core melt would result from failure of the pumps. Assuming a failed containment, a pool of water overlying the core debris woulti help to scrub fission products and thus mitigate the consequences of the event. The following discussion provides an estimate of the risk reduction resulting from mitigation of ATWS events by having a backup water supply to the drywell. i The man-rem consequences of an SST1 release occurring at each plant site has been evaluated in NUREG/CR-2723. For the BWR Mark I plants, the man-rem consequences, averaged over all sites, was 2.6x10*7 man-rem, given an SST1 release. Based upon the radionuclides release calculations reported for a BWR Mark I (NUREG/CR-4624), Vol 1.) an ATWS event is expected to result in release fractions roughly equal to one-third of an SST1. As noted in NUREG/CR-2723, Table 10, the person-rem from such a release would be about 50% that of an SST1 release, further, the person-rem given in NUREG/CR-2723 is for an infinite radius, whereas the value-impact requires that this he estimated over 50 miles.

As noted in NUREG-1109, the total person-rem exposure within a 50 mile radius is approximately one-fourth the person-rem exposure for an infinite radius.

Based on the above, it can be calculated that the consequences of an unmitigated ATWS event for an average BWR Mark I over a 50 mile radius is estimated to be 46 l 2.6x10+7 man-rem x 0.5x0.25 = 3.25x10 man-rem.

22 l

+

If the probability of an ATWS event, after the ATWS rule, is assumed to be 2x10-5 per reactor-year, the risk prior to any improvements, is 3.25x10+6 man-rem x 2x10-5/RY=65 man-rem /RY.

Given an ATWS event, the containment is expected to fail and a core-melt will result from failure of the pumps. Assuming liner failure, but with the presence of an overlying pool of water, fission products will be scrubbed by the overlying pool with an overall DF=3. Therefore, the consequences of an ATWS with the backup water supply present is 23 man-rem /RY, and the rlsk reduction benefit of water, as a mitigative effect for ATWS, is 65 - 23 = 42 man-rem /RY.

The cost-benefit of the incremental addition of ADS improvements and a backup water supply a:; a function of assumptions concerning SB0 sensitivities and ATWS mitigation is shown in Table 3. This table indicates that the incremental additions of ADS improvements and a backup supply of water are cost effective based on prevention of SB0 if the Mark I average SB0 frequency is used and the low cost estimate is assumed. For Mark I plants with high SB0 frequencies, the improvements are cost effective even with the high cost estimate. The additional benefits of possible ATWS mitigation further support the cost-effectiveness of the ADS and backup water supply.

4.1.8 Summary Alternative (vi) is recommended because this alternative is cost-effective and provides a substantial reduction in risk. While the improvements form a package in the sense that the benefit is maximized when the improvements are combined, the individual improvements contained in alternative (vi) have also been shown to be generally cost-effective on an incremental basis. Containment venting reduces core damage frequency due to TW sequences while the combination of venting, ADS and backup water reduces core damage frequency due to SB0 sequences.

, Venting would also mitigate the consequences of core damage events by protecting the containment from failure due to long-term overpressure. The backup water supply also provides mitigation by preventing, or at least delaying the contain-ment shell from melting, by attack of the core debris, or failing that, by providing scrubbing of fission products.

23

;;9 Table 3 Cost-Benefit For Incremental Addition of-ADS and Backup Water Supply (Man-remavertedpermilliondollars):

MAN REM . LOW INST. HIGH INST.

. AVERTED- COST COST' ASSUMPTION BASE CASE

• 33 630 190 AVERAGE SB0 PROB.** 55 1,050 310 93 1,780 530-

! AVERAGE'S30 PROB.-

HIGH LINER MELT. PROB.***-

135 2,580 760 AVERAGE SBO. PROB. .

HIGH LINER MELT PROB.

ATWS MITIGATION ****

~HIGH'SB0 PROB.***** 165 3,150 930 HIGH'SB0 PROB. 280 5,344- -1,580

.HIGH LINER MELT PROB.

HIGH SB0 PROB. 322 6,145 1,816 HIGH LINER MELT PROB.

ATWS-MITIGATION HIGH SB0 PROB. 64 1,221 361 3

HIGH LINER MELT PROB.

.ATW'HITIGATION

. LOW POPULATION L

24

}:

[.~

c NOTES FOR TABLE 3

• Based on SB0 frequency of 6x10-6/RY and a conditional liner melt probability of 0.5 given a core melt.

• • Based on a SB0 frequency of 1x10-5/RY which is an average for Mark I plants.

• • • Conditional liner melt probability of 1.0 given a core melt.

• • • • Assumes ATWS mitigation of 42 man-rem per RY.

• • • • • Based on a SB0 frequency of 3.5x10-5/RY.

l

.1 1

25 l

I 1

i

4.2 Impacts on Other Requirements There are six programs related to severe accidents. These programs are:

Individual Plant Examinations (IPE), Conteir. ment Performance Improvements (the topic of this regulatory analysis), Improved Plant Operations, Severe Accident Research Program, External Events, and Accident Management. Each of the five programs related to Containment Performance Improvements (CPI) will be discussed briefly below 28 ,

4.2.1 Individual Plant Examinati.ons The IPE involves the formulation of an integrated and systematic approach to an examination of each nuclear power plant now in operation or under construction for possible significant plant-specific risk contributors that might be missed without a systematic search. The examination will pay specific attention to containment performance in striking a balance between accident prevention and consequence mitigation. It is anticipated that the IPE program may take from three to five years until the last plant has performed the IPE and incorporated the appropriate plant modifications. Since the staff has already identified cost-effective improvements that are generic, there is no need to wait for the IPE to be completed. The modifications related to the CPI program are expected to be installed in approximately 30 months.

4.2.2 ImprovedPlantOperations(IPO)

The IPO includes consideration of the continued improvements in the Systematic Assessment of Licensee Performance (SALP) program; regular reviews by senior NRC staff managers to identify and evaluate those plants that may not be meeting NRC and industry standards of operating performance; diagnostic team inspections; improved plant Technical Specifications; improved operating procedures; expansion of the Emergency i

28For additional information, refer to SECY-88-147, " Integration Plan for Closure of Severe Accident Issues", dated May 25, 1988.

26

)

OperatingProcedures(EOPs)toincludeguidanceonsevereaccident management strategies; industry's programs to reduce transient and other challenges to engineered safety feature systems; feedback from the IPE program of experience and improvements in operational areas, such as maintenance'and training; and continued research to evaluate the sensi-tivity of risk to human errors, the contribution of management to the-level of human errors, and the effectiveness of operational. reliability methods to help identify potential problems early and prevent their-occurrence. The IPO is related to the CPI program's recommendations since we recommend improved procedures and operator training.

4.2.3 Severe Accident Research Program (SARP)

The SARP was begun after the TMI-2 accident in March 1979 to provide the Commission and the NRC staff with the technical data and analytical methodology needed to address severe accident issues. This program has provided input to the NUREG-1150 program and to the CPI program.

Additional research is required to evaluate the need for and feasibility

of core debris controls. Research is also needed to confirm and quantify the benefits of having water in the containment to either scrub fission products or to prevent or delay shell melt by core debris.

4.2.4 External Events The Commission's Severe Accident Policy Statement does not differentiate between events initiated within the plant and externally initiated events.

Typically, external events have not been incorporated in the staff PRAs.

Procedures for external events examinations are under development and the evaluation of external events will proceed separately. The CPI program  ;

only addresses internally initiated events and it is not anticipated that future consideration of external events will adversely affect the j recommendations of the CPI program. 1 4.2.5 Accident Management l 27 l

~'

A, , , ..,-,s -m, .G- - - - -~~

l

-4

, j 1

l

" The accident management program is concerned with addressing certain preparatory and recovery measures that can be taken by the plant operating - ]

and technical staff that could prevent or significantly mitigate the l consequences of a severe accident. This includes measures taken by the plantstaffto1)preventcoredamage,2) terminate'theprogressofcore damage if it begins and retain the core within the reactor vessel, 3) failing that, maintain containment integrity'as long as possible, and finally 4). minimize the consequences of offsite releases. The CPI program recommended plant enhancements would provide the accident management program with additional capabilities to achieve their goals by providing improved hardware with which to deal with severe accidents.

4.3 Constraints The backfit rule (10 CFR 50.109) as published by the Commission on September-

'20,'1985 sets forth restrictions on imposing new requirements on currently licensed nuclear power plants and specifies standard procedures that must be applied to backfitting decisions. The backfit rule states:

"The Commission shall require a systematic and documented analysis pursuant to paragraph (c) of this section for backfits which it seeks to impose....(10 CFR 50.109(a)(2)).

"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 protectipn of the public health and safety or the common defense and security to be derived from the backfit and that the direct and indirect costs of implementation for that facility are justified in view of this increased protection. (10 CFR 50.109(a)(3))".

In order to reach this determination, 10 CFR 50.109(c) sets forth nine specific h

factors which are to be considered in the analysis for the backfits it seeks to

[

l impose. These nine factors are among those discussed in the main body of this report. Appendix A provides a discussion summarizing each of these factors.

The Commission also states in the backfit rule that "any other information relevant and material to the proposed backfit" will be considered.

28

g-_ - _ _ _ _- ___ -.

l

.; j This report provides' additional relevant information concerning the proposed

, containment performance enhancements. This analysis supports a determination i L

that a substantial increase in-the protection of the public health and safety will be derived from backfitting the containment performance enhancements, and that the backfit is justified in view of the direct and indirect costs of i

implementing the enhancements. It is also noted that the Commission directed the NRC staff to provide all potential improvement and related recommendations to the Commission for their consideration regardless of the results of the backfit analysis.

No other constraints have been identified that affect this program.

l 5.0 DECISIDH RATIONALE

.The evaluation of the CPI program included deterministic and probabilistic analyses. Calculations to estimate the core damage frequency and the consequences of TB and TW sequences were performed based on using simplified containment event trees and the information available from the NUREG-1150 program. These estimates were used to give insights, along with engineering judgement, to develop the recommendations to improve containment performance.

A review of the available BWR Mark I PRAs provided only limited information.

However, the highest core damage frequency ideatified was for a plant which had only been reviewed as part of the A-45 study for the TW accident sequence which was 3x10~4 per reactor-year and the IREP. Millstone Unit 1 PRA yielded a core melt frequency of 3 x 10~4/ reactor-year. The lowest core damage frequency identified from all dominant accident sequences was identified to be 2.3x10-6 per reactor year with successful venting. For those plants where TW is the i dominant contributor to the plant core melt frequency, a range of core melt frequencies from 1x10~4 to 1x10-5 per reactor-year was used in the risk analysis. An assumed core damage frequency, excluding TW and with compliance with the ATWS and SB0 rules, was taken to be 8.2x10-6 per reactor-year. Sensi-tivity studies were also performed to investigate the impact of the proposed f- Implemen-improvements for Mark I plants with higher SB0 and ATWS frequencies.

tation of the proposed recommendations will result in TW being a minsr contributor 29

and SB0 being a sniall contributor to the total core damage frequency.

5.1 Commission's Safety Goal On August 4,.1986, the Commission published in the Federal Register a policy statement on " Safety Goals for the Operations of Nuclear Power Plants" (51 FR 28044). This policy statement focuses on the risks to the public from nuclear power plant operation and establishes goals that broadly define an acceptable level of radiological risk. The discussion below addresses the CPI program recommendations in light of these goals.

The two quali_tative safety goals are:

(1) Individual member of the public should be provided a level of protection from the consequences of nuclear power plant operation such that individuals bear no significant additional risk to life and health.

(2) Societal risks in life and health from nuclear power plant operation should be comparable to or less than the risks of generating electri-city by viable competing technologies a1d should not be a significant addition to other societal risk.

The following quantitative objectives are used in determining achievement of the above safety goals:

(1) The risk to an average individual in the vicinity of a nuclear power plant of prompt fatalities that might result from reactor accidents should not exceed one-tenth of one percent (0.1%) of the sum of prompt fatality risks resulting from other accidents to which members of the U.S. population are generally exposed.

E (2) The risk to the population in the area near a nuclear power plant of l

cancer fatalities that might result from nuclear power plant I

30 1

l 1

I - _ - _ _ . _-_-_-__ _ -__ -

, -- _ _ _ - _- .- . _ - . = = _ - . _ _ _ .

)

J operationshouldnotexceedone-tenthofonepercent(0.1%)ofthe sum of cancer fatality risks resulting from all other causes.

Results of analyses published in draft HUREG-1150 for the BWR Mark I (Peach Bottom Atomic Power Station, Unit 2) indicated that the Mark I plant meets the risk criteria for prompt fatalities and latent cancer fatalities stated above, even considering the large uncertainties involved. Implementation of the CPI recommendations will result in the total core damage frequency being reduced by about a factor of five to ten by reducing the two dominant sequence frequencies to below the estimated core melt frequency for the Mark I plant in NUREG-1150.

The Comniission also stated the following regulatory objective relating to the frequency of core melt accidents at nuclear power plants.

Severe core damage accidents can lead to more serious accidents with the potential for life-threatening offsite releases of radiation, for evacuation of members of the public, and for contamination of public property. Apart from their health and safety consequences, such accidents can erode public confidence in the safety of nuclear power and can lead to further instability and unpredictability for the industry. In order to avoid these adverse consequences, the Commission intends to continue to pursue a regulatory program that has as its objective providing reasonable assurance, giving appropriate consideration to the uncertainties involved, that a severe core damage accident will not occur at a U.S. nuclear power plant.

With the implementation of the CPI reconaendations, it is expected that the total core melt f requency can be reduced by a factor of between 1.6x10-5 to 9.7x10-5 per ceactor-year. Therefore, implementing the recommendations for CPI significantly reduces the likelihood that a severe core melt accident will occur at a U.S. BWR with a Mark I containment.

Additional rationale for implementing the CPI recommendations over other I alternatives is discussed as part of the value-impact analysis (Section 4.1).

l 31

{

w - _ _ _ _ - _ - - _ _ _

This action represents the staff's position' based on a comprehensive analysis of~the containment performance improvement issues.

6.0 IMPLEMENTATION-6.1 Schedule for Implementation Within 60 days after issuance of the final rule, ifcensees will submit to the NRC a schedule for implementing any necessary equipment and procedural modifications to meet the performance goals and to provide adequate defense-in-depth. All plant modifications are to be installed, procedures revised, and operators trained not later than 30 months from the issuance of the final rule.

Other schedules were considered; however, the staff believes the proposed implementation of the CPI recommendations can be performed with minimum interfacing with containment and engineered safety feature systems and thus with the plant online and therefore is achievable without unnecessary financial burden on licensees for plant shutdown. The schedule allows reasonable time for the-implementation of necessary hardware items to achieve a reduction in risk of severe accidents. Shorter or le:s flexible schedules would be unnecessarily burdensome.

6.2 Relationship to Other Existing or Proposed Requirements Several NRC programs are related to the CPI program; these are discussed in Section 4.2. These programs are compatible with the recommendations of the CPI program.

The electrical power requirements of the proposed improvements have been coordinated with the electrical power reliability requirements of the Station BlackoutRule(10CFR50.63).

~

s 32

~ '

7.0 REFERENCES

' -- , NUREG-1150, " Reactor Risk Reference Document". February 1987.

-- , NUREG-1109, " Regulatory /Backfit Analysis for the Resolution of Unresolved Safety Issue A-44,' Station Blackout", June 1988.

-- , NUREG-1032, " Evaluation of Station Blackout Accidents at Nuclear Power Plants, Technical Findings Related to Unresolved Safety Issue A-44", June 1988.

-- , NUREG/CR-4012, " Replacement Energy Costs for Nuclear Electricity-Generating Units in the United States: 1987-1991", January 1987.

-- , NUREG/CR-3568, "A Handbook for Value-Impact Assessment", December 1983.

-- ,NUREG/CR-4651, " Evaluation of Severe Accident Risks and the Potential for Risk Reduction: Peach Bottom, Unit 2", Volume 3 Draft, May 1987.

-- , NUREG/CR-2723, " Estimates of the Financia: Consequences of Nuclear Power Reactor Accidents", September 1982.

-- , NUREG/CR-4624, " Radionuclides Release Calculations for Selected Severe Accident Scenarios", July 1986.

-- , NUREG/CR-5225, "An Overview of Boiling Water Reactor Mark 1 Containment Venting Risk Implications", October 1988.

U.S. Atomic Energy Commission, WASH-1400, " Reactor Safety Study", October 1975 (also.re-issuedasNUREG-75/014)

Victor Stello Jr, SECY-88-147, " Integration Plan for Closure of Severe Accident Issues", May 25, 1988 Idaho National Engineering Laboratory, " Sensitivity Results for Mark I Containment Improvements Program - RDJ-53-88", latter report to NRC, November 18, 1988.

Science and Engineering Associates, Inc. Report 87-253-07-A:1," Cost Ana?fsis for Potential BWR Mark I Containment Improvements", November 1988.

Letter from Boston Edison Company, DPU 88-28, Request No. AG 13-6.

' i 33

~ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _

>~.

APPENDIX A BACKFIT ANALYSIS

/

Analysis and Determination That the Recommended Safety Enhancements for Containment Performance Improvements Complies With the Backfit Rule 10 CFR 50.109 The Commission's existing regulations establish requirements for the design and testing of containment and containment cooling systems (10 CFR 50, Appendix A, General Design Criteria 50, 52, 53, 54, 55, 56, and 57) with respect to design basis accident conditions. As evidenced by the accident at TMI Unit 2, accidents could progress beyond design basis considerations and result in a r severe accident. Such an accident could pose a challenge to the integrity of containment. Existing regulations do not require explicitly that nuclear power

' plant containments be designed to withstand the severe accident conditions.

This issue has been studied by the staff and our consultants as-part of the severe accident program for the General Electric Company Boiling Water Reactors (BWRs) with Mark I containments. The BWRs with Mark I. containments have been reviewed first because the perceived susceptibility of the Mark I containments have been reviewed first because other perceived susceptibility of the Mark I containment to failure based, in part, on the small containment volume of the Mark I containment design._ Both deterministic and probabilistic analysis were performed to determine the dominant challenges to containment integrity and potential. failure modes affecting the likelihood of core melt, reactor vessel failure, containment failure, and risk to the public health and safety.

Although the risk analysis shows that there is no endue risk to the overull plant risk from station blackout (TB) and loss of long-term decay heat removal capability (TW) can be significantly reduced.

The estimated benefit from implementing the proposed plant performance improvements is a reduction in the frequency of core melt due to TB and TW events and the associated risk of offsite radioactive releases. The risk reduction for the 24 operating BWR reactors with Mark I containments is estimated to between 86,900 and 692,000 man-rem and supports the Commission conclusion implementation of the proposed safety enhancements for Mark I plants provides a substantial improvement in the level of protection of the public health and safety.

The cost to licensees to implement the proposed safety enhancements would vary depending on the existing capabilities of each plant. The costs would be primarily for licensees (1) to assess the plants capabilities, (2) modify L existingequipmenttoprovideadditionaloperationalflexibility,(3)to L retrofit plants with additional components or systems, as necessary, to meet ll ro j.

the and p(5) to posed providecapabilities, opertor training (4)related to reviseto mitigating the emergency operating procedure severe accidents.

t- \

I- l

The estimated total cost for the 24 BWRs with Mark I containments to provide the proposed safety enhancements is between $48 and ,$176 million. The cost per reactor would be between $1.6 and $7.3 million.

The overall value-impact ratio, not including accident avoidaxe costs, is estimated to between 1,810 and 3,930 man-rem averted per million dollars based on a range of core melt frequencies for TW and a range of installation costs.

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 estimated overall value-impact ratio would improve slightly to between 1,970 and 4,570 man-rems averted per million dollars.

These values support proceeding with the proposed Mark I containment performance improvements.

The preceding quantitative value-impact analysis was one of the factors considered in evaluating the proposed improvements, but other factors also played a part in the decision-making process. Probabilistic risk assessment (PRA) studies performed for this issue have shown that station blackout (TB) and loss of long term decay heat removal (TW) events can be significant contributors to core melt frequency, and, with consideration of the conditional containment failure probability, TB and TW events can represent an important contribution to reactor risk. In general, active systems required for reactor and containment heat removal are unavailable during the postulated severe accidents. Therefore, the offsite risk is higher from a severe accident than it is from many other accident scenarios.

Although there are licensing requirements and guidance directed at providing a containment and support systems intended to contain any release of material from the reactor vessel, containment integrity may be significantly challenged under severe accident conditions. The challenge to containment integrity is primarily by over-pressure and over-temperature. Failure of the containment can initiate core degradation (as in the TW sequence) or can be the result of core degradation (a'; in the TB sequence).

The estimated frequency of core melt from TB and TW events are directly proportional to the frequency of the initiating events. Estimates of TB frequency was based on the information provided in draf t HUREG-1150, " Reactor Risk Reference Document", for Peach Bottom Atomic Power Station, Unit 2. This is assumed to be a realistic estimate of the core melt frequency for Mark I plants after the plant is in compliante with 10 CFR 50.63, the station blackout rule. For the TW sequence, a range of frequencies was used based on the PRAs available for the Mark I plants. This range is assumed to be typical of those plants where TW is the dominant contributor to core melt, as suggested in WASH-1400.

The factors discussed above support the determination that the additional defense in-depth provided by the ability of a Mark I plant to cope with a severe accident would provide a substantial increase in the overall protection of the public heslth and safety, and the direct and indirect costs of implementation are justified in view of this increased protection. The staff has considered how this bbci. it should be prioritized and scheduled in light of other related regulatory activities. The proposed rule to implement these improvements would require notification of plans and schedules within 60 days 2

of the final rule and implementation of the improvement within 30' months of~the final rule.

One U.S. facility, Pilgrim, has installed a hardened vent system, but it.is not yet operational.

Analysis of 10 CFR 50.109(c) Factors (1) Statement of the specific objectives that the backfit is designed to achieve Tne objective of the proposed Mark I containment performance improvements is to reduce the risk of severe accidents by reducing the likelihood of core melt and by improving the ability to mitigate the consequences in the -

event.of a severe accident. Specifically, the proposed improvements is for all BWRs with Mark I containments to enhance t'ie. reliability of the automatic depressurization system (ADS), to provide an alternate supply of water for injection into the reactor vessel and for operation of the containment sprays, and to provide a hardened vent capability from the containment wetwell to plant stack and to implement improved procedures and training. These modifications are to be operable during a station blackout by means of an additional power supply.

(2) General description of the activity required by the licensee or applicant in order to complete the backfit In order to comply with the proposed containment improvements, licensees will be required to

• Evaluate the plant's actual capabilities relative to the proposed  !

safety enhancements. This evaluation will include

- verifying the ability of the ADS cables inside containment to remain operable in the anticipated environment during a severe accident.

- verifying the avad lability of a diesel powered water pump and the capacity of the pump to p avide low pressure cooling to the core with a depressurized reactor and to the drywell sprays during a severe accident

- verifying the capability of the containment vent system to withstand the anticipated containment pressures without failing any portion of the vent path to the stack

- Verifying the capability of the containment isolation valves to be ,

opened and reclosed under all antic',,ated containment pressures and vent flow rates during severe accidents .

l

)

• Determine the 'necessary plant modifications to comply with the proposed containment improvements, develop a schedule for plant f

modification, and submit the schedule to the NRC within 60 days from I issuance of the final rule.

3

• Complete necessary plant modifications within 30 months from issuance of the final rule.

Depending on the plant's existing capability to cope with severe accidents, licensees may or may not need to backfit hardware modifications, (Seeitem8 of this analysis for additional discussion.) Licensees will be required to have procedures and training to cope with and recover from severe accidents.

These procedures should conform to Revision 4 of the BWROG Emergency Procedure Guidelines. ,

(3) Potential change in the risk to the public from the accidental offsite release of radioactive material Implementation of the proposed BWR Mark I containment improvements is expected to result in an estimated total risk reduction to the public rar4ing from 86,900 to 629,000 man-rems over an assumed 20 years of remaining life for the 24 Mark I plants, based on the particular plant's TW frequency.

f (4) Potential impact on radiological exposure of facility employees For the 24 operating BWRs with Mark I containments, the estimated total reduction in occupational exposure resulting from reduced core damage frequencies and associated post accident cleanup and repair activities is 1,500 man-rem. The estimated total occupational exposure for installation r of the proposed improvements is assured to be negligible. No increase in occupational exposure is expected from operation and maintenance activities associated with the proposed improvements.

(5) Installation and continuing costs associated with the backfit, including the cost of facility downtime or the cost of construction delay No Mark I containments are under construction, thus there are no costs associated with construction delays. All plant modifications are expected to be capable of being made either with the plant operating or during normal plant outages, thus there are no costs associated with additional plant downtime.

The cost of the proposed improvements has been estimated to range between

$1.6 and $3.2 million per plant with a best estimate of $2.0 million. For the 24 Mark I plants, this represents an industry cost range of $38.4 to

$76.8 million with a best estimate of $48 million. Pilgrim has instituted a Safety Enhancement Program which incorporated some of the recommended containment improvements. The Pilgrim costs for those elements identified as part of the containment performance program is $7.3 million. This would translate to an industry cost to $176 million, (6) The potential safety impact of changes in plant or operational complexity, including the relationship to proposed and existing regulatory requirements The proposed containment improvements to be able to cope with severe accidents should not add to plant or operational complexity. The ,

l 4

l l

l 1

)

containment performance improvement (CPI) program is related to several f HRC programs as the following discussion indicates.

• Station Blackout Rule (10 CFR 50.63) the electrical power {

requirements of the proposed improvements have been coordinated with l the requirements of 10 CFR 50.63 for electrical power reliability. J

• IndividualPlantExamination(IPE) 1 The IPE involves the formulation of an integrated and systematic approach to an examination of each nuclear power plant now in operation or under construction for possible significant plant-specific risk contributors that might be missed without a systematic search. The examination will pay specific attention to containment performance in striking a balance between accident prevention and consequence mitigation. It is anticipated that the IPE program may take from three to five years until the last plant has performed the IPE and incorporated the appropriate plant ,

modifications. Since the staff has already identified cost-effective improvements that are generic to BWRs with Mark I containments, there is no need to wait for the IPE to be completed. The modifications related to the CPI program are expected to be installed approximately 30 months.

• ImprovedPlantOperations(IOP)

The IOP includes consideration of the continued improvements in the Systematic Assessment of Licensee Performance (SALP) program; regular reviews by senior NRC staff managers to identify and evaluate those plants that may not be meeting NRC industry standards of operating performance; diagnostic team inspections; improved plant Technical Specifications; improved operatin Procedures (E0Ps)g procedures; expansion of the Emergency Operating strategics; industry's programs to reduce transient and other challenges I to engineered safety feature systems; feedback from the IPE program of experience and improvements in operational areas, such as maintenance and training; and continued research to evaluate the sensitivity of risk to human errors, the contribution of management to the level of human errors, and the effectiveness of operational prevent their occurrence. The IPO is related to the CPI program's recommendations since we recommend improved procedures and operating training.

• Severe Accident Research Program (SARP)

The SARP was begun after the THI-2 accident in March 1979 to provide the Commission and the NRC staff with the technical data and analytical methodology needed to address severe accident issues. This program has provided input to the NUREG-1150 program and to the CPI program.

Additional research is needed to confirm and quantify the benefits of having water in the containment to either scrub fission products or to

! prevent or delay shall melt by core debris.

!

• Accident Management The accident management program is concerned with addressing certain preparatory and recovery measures that can be taken by the plant operating u_ _ --

L f

and technical staff that could prevent or significantly mitigate the consequences of a severe accident. This~ includes measures taken by the plant staff to 1) prevent r'~ e damage 2) terminate tide progress of core damage if it begins' and re' ' the core within the reactor vessel, 3) lfailing that, maintain cont oment integrity as long as possible, and finally 4) minimize the consequences of offsite releases. The CPI program recommended plant enhancements would provide the accident management-program with additional capabilities to achieve their goals by providing improved hardware and incorporating improved training and procedures with m which to deal with severe accidents.

.(7)' The estimated burden on the NRC associated with the backfit and the

~ '

availability, of such resources b The estimated total cost fce '$C review of industry submittals is $0.4 million based on submittals for 24 reactors and an estimated average of 200 man-hours. per reactor.

(8) The potential impact of differences in facility type, design, or age on the relevancy and practicality of the backfit The proposed improvements are only for BWRs with Mark I containments. All l

reactors and containments in the category are essentially similar and thus the recommendations are applicable to all 24 Mark I plants. However, a survey of Mark I plants has identified that, to varying degrees, each facility may already have some of the equipment which may be used to satisfy the requirements of the CPI program. For example, a dedicated safe shutdown facility may provide the needed power and alternate water supply. Some plants may have a diesel driven water pump for the fire protection system or a diesel driven service water pump. A hardened vent pipe may already exist from wetwell to outside of the reactor building, leaving only a small section of the vent path requiring replacement. l (9) Whether the backfit is interim or final and,_if interim, the justification f or imposing the backfit on an interim basis The CPI recommendations for Mark I plants is the final resolution for containment performance except as related to the liner meltthrough issue.

Confirmatory research on this issue will continue, and is expected to show >

that use of water in the drywell will have a beneficial effect of preventing or at least delaying liner meltthrough, in addition to mitigating any consequences. The proposed improvements are not an interim measure.

i s

)

6 L

N --_ _ - _ - - _ - _ - _ - _ _ -

ce Q4 l

Enclosure 7 DRAFT PROPOSED RULE SECTIO! '0.XX - SEVERE ACCIDENT REQUIREMENTS FOR BOILING-WATER REACTORS HAVING MARK 1 CONTAINMENTS a) Applicability The requirements of this section apply to all boiling water reactors (BWR) having Mark I' containments.

b) Requirements

1. Backup Water Suppy for Drywell Spray / Core Injection All BWRs with a Mark I containment shall provide at least one backup water supply system for the containment drywell spray which shall be functional during an extended station blackout.1 Water to the spray system from this backup supply shall be available by remote manual operation or. by simple procedures for connection and startup which can be implemented during a severe accident scenario.

The backup water supply system shall also be capable of being diverted to the reactor vessel to provide an alternate source of water to cool the core once the vessel has been depre.surized. The flow rate shall be at least equivalent to provide decay heat removal of 1% of full power and all required valve realignments shall be functional during an extended station blackout.

1 An extended station blackout is defined as loss of all normal and emergency AC power and loss of DC power due to depletion of station batteries. Operability of controls and valves during such an event may require an independent source of power such as a dedicated battery set or a means to recharge the station batteries.

l m______.____-__________ _ - _ . _ _ _ _ _ _ . _ - . _ _ _ _

(

Instrumentation needed to accomplish both above functions shall be operable in the expected accident conditions and should, as a minimum, include [to redetermined].

l

2. Containment Venting l

k BWR Mark I containments shall be provided with an exhaust line from the wetwell vapor space to a suitable release point (e.g., plart stock) which is capable of retaining expected vent pressure. This "hard vent" system shall meet the following criteria:

a) The basic design objective shall be to provide sufficient venting capacity to prevent long-term overpressure failure of containment (sizing of vent shall accommodate no less than 1% of full power).

b) The venting setpoint shall be set at or above containment design pressure. Capability of ADS valve s, torus vent valves, or other equipment needed to cope with an accident should not limit vent operability to less than containment design tr":ssure, c) The venting capability shall be available during station blackout extending for a period up to 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> beyond onset of blackout.

d) The hardened vent path should provide a mecns to prevent premature or inadvertent actuation.

e) The vent path up to and including the second containment isolation barrier should be designated safety Class 2.

f) The hard vent path shall accommodate effects of potential combustion phenomena and remain operable.

2

_ ______a

O g) The hardened vent path shall have a radiation monitor, clarmed in control room and operable during extended station blackout.

3. ADS Enhancements l

All licensees having BWR Mark I containments shall examine the automatic depressurization system (ADS) and make modifications to ensure its operability during severe accidents, including performance during extended station blackout. The ability of the ADS cables and components to survive the environment in the containment during a severe accident aad the adequacy of the power supply as discussed in Section 4.C.2, as a minimum, shall be reviewed.

4. Procedures and Training All BWRs with Mark I containments shall implement procedures for use of the backup water supply, the vent system, and depressurization equipment to prevent or citigate severe accidents.

c) Implementation

1) Schedule All licensees to whom this section applies shall submit their plans and anticipated schedule within 60 days after a final rule is issued which identifies any actions taken and those needed to be taken to comply with the requirements of this section. The requirements of this section shall be fully implemented within 30 months after a final rule is issued.

2) Co-ordination with requirements of the Station Blackout Rule (10 CFR 50.63) .

l t

Any sources of electrical power required to assure the operability of the backup water supply, containment venting system, and ADS during on extended station blackout, as required in part (b) above, should be coordinated with the requirements of 10 CFR 50.63, as follows:

l b_____________________ l

l o

f t

h a) Those licensees who choose to implement the requirements of 10 CFR f 50.63 by the use of an alternate AC (AAC) source need not provide any additional power supplies to comply with the provisions of this section, provided that the capacity, capability, and duration of the L AAC can be shown to meet the requirements of both 10 CFR 50.63 and this section.

b) Those licensees who choose to impiement the requirements of 10 CFP, 50.63 solely by means of a coping analysis, mr.at provide additional power supplies of sufficient capacity and reliability to assure the operability of the backup water supply, containment venting system and ADS during an extended station blackout.

)

_-