ML20245B191

From kanterella
Jump to navigation Jump to search
Presents Composite Set of NRC & Industry Recommended Remedial Actions Based on 1985 AEOD Case Study Rept C503 Re DHR Problems for PWR Operation.Table 1 Presenting Chronology of 37 events,cost-benefit Analysis & Related Info Encl
ML20245B191
Person / Time
Site: Diablo Canyon Pacific Gas & Electric icon.png
Issue date: 05/18/1987
From: Jordan E
NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To: Beckjord E, Murley T
Office of Nuclear Reactor Regulation, NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
Shared Package
ML20195G243 List:
References
CON-IIT05-184-90, CON-IIT5-184-90, FOIA-87-714, TASK-099, TASK-99, TASK-A-45, TASK-OR NUDOCS 8705280091
Download: ML20245B191 (32)


Text

\\

f gg(

Y

UNITED STATES NudLEAR REGULATORY COMMIS$10N mamam.o.c.aosas NAY 181987 PEMORANDUM FOP: Thomas E. Murley, Director Office of Nuclear Reactor Regulation Eric S. Beckjord, Director Office of Nuclear Regulatory Research FROM:

Edward Jordan, Director Office for Analysis and Evaluation of Operational Data

SUBJECT:

LOSS OF DECAY HEAT REMOYAL FUNCTION AT PRESSURIZED WATER REACTORS WITH PARTIALLY DRAINED REACTOR COOLANT SYSTEMS

, Introduction On April 10,1987 with the reactor coolant system partially drained, the residual heat removal (RHR) pumps at Diablo Canyon 2 were tripped due to vortexing/ cavitation.

As a result, the plant lest its ability to remove decay heat for 85 minutes.

(RCS) heated up and bulk boiling was present in the RCS.During that 85 minu The loss of the decay heat removal (DHR) function which occurred at Diablo Canyon 2 is one of 37 such events that have been reported to have occurred at U.S PWRs over the last 10 Table 1 presents a chronology of these 37 events. These events have years.

the potential for leading to more serious events.

of NRC and industry actions in the area of DHR system losses. Table 2 presents a chrono This memorandum p(resents a composite set of NRC and industry recomended remedial actions ) which are based upon the 1985 AE00 case study report C503 dealing with decay heat removal problems for PWR operation, analysis of events subsequent to the case study, and related industry recommendations including INPO SOER 85-4, INPO SER 79-84, and'NSAC-52.

In its transmittal letter of the case study to the Director ONRR, AE00 reconnended that the recommendations contained within the report be considered in the resolution of Unresolved Safety Issue A-45.

In a response the Director, ONPR believed that the AE0D recommendations were not directly appitcable to the resolution of A-45, but instead planned to include them in the resolution of -

Generic Issue No. 99 "RCS/RHR Suction Line Interlocks."Thisissuewasspeci-fically concerned with loss of the RHR system during cold thutdown or refueling.

61-99 was subsequently modified to evaluate these issues.

Loss of D:iR during shutdown is clearly not a new issue. However, the continued occurrence of loss of DH? events, the apparent lack of effectiveness of licensee-corrective action in response to past NRC and industry actions, re-assessment of the estimated risk of such events, and the dependence of the risk estimates on human perfonnance, all indicate that prompt regulatory action is now needed to cinimize the loss of DHR during periods with a partially drained-primary ovatam and to help assure its rapid recovery should it be lost.

('

f 9'

m (Ot

\\

52fMW W th f/'

D

V.

~

Tho2as E. Murley '

Discussion U.S. PVR experience has shown that loss of DHR events have been occurring at a rate of approximately one every 3 to 4 reactor years, and in particular there have been 7 loss of DHR events in the last 2 years when the RCS was in a drained-down condition.

Human errors were the root causes of most of those events.

Plants may be subjected to relatively high risks when they undergn partially drained (mid-loop) operations.

It is standard procedure for PW to drain-down the RCS during shutdown to allow for steam generator maintenance, inspection, and tube plugging, and/or reactor coolant pump seal inspection / maintenance.

Factors which contribute to the acrident risk during such operations are:

1.

The containment is likely not to be isolat9d (the equipment hatch is often open).

2.

Plant design may dictate a ver drained-down operations (e.g.,y narrow band of allowable RCS levels during RCS levels was only a few inches - the constraints being the eleva the steam generator nozzle and the suction Lead required by the RHR pump topreventairbinding.)

3.

RCS level measurement during drained-down operations frequently depends upon jury-rigged equipment which is unanal may exceed the required control band (e.g.yzed and prone to errors which

, at Diablo Canyon the level measurement error was on the same order as the range of acceptable operation possibility 3 to 12 inches).

4.

Generally, procedures for operation during modes 3, 4, and 5 are of an ad hoc nature, scant or even nonexistent.

from a loss of DHR are not necessarily well thought out.Similarly, procedure operators may not be trained in recovery from a loss of DHR.In addition, During shut-down operations, operators may not be fully aware of what equipment is out of service vs. what alternative equipment is available for recovery from a loss of DHR.

Operators are not necessstily aware of time available for recovery from loss of DHR events.

thought that if the DHR function was lost the RCS heatup rate wou l'F/ minute.

However the RCS heat up rate was 2.7'f/sinute.

Therefore, the operators were not expecting bulk boiling to begin as soon as it did.

5.

Plants may not have adequate instrumentation available to determine RCS temperature in the reactor during a loss of DHR event.

For example, Diablo Canyon 2 had disconnected the core thermocouple prior tc the loss of DHR event in anticipation of head removal.

Analysis Center (NSAC) published a report on RHR exper (NSAC-52).

tions to industry to improve the situation.NSAC-52 provided data on loss of O Similarly, numerous industry reports (e.g., INPO SERs 17-86, 79-84, INPO SOER 85-4 mation on loss of DHR events, including recomm)endations for improv situation.

ficant industry-wide improvement in DHR loss experiences.Nonethel

...A Thomas E. Murley.

In July 1984 EPRI's Nuclear Safety Analysis Center published NSAC-84, a PRA which addressed operation at Zion 1 and 2 during shutdown.

That PRA stilized maintenance and operation records and control room logbook information to estimate equipment availabilities and recovery times.

To our knowledge, it was the first comprehensive PRA to address operations at U.S. PW during modes 4, 5, and 6.

That study shows that the likelihood of a core damage event in non power modes is comparable to that during power operat. ion.

NSAC-84 notes that,10 days after shutdown, if the p16 ', is in a drained-down (mid-loop) condition, fuel damage" can occur 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> after losing the DHR func-tion.

Assuming the same decay heat curve, we conclude that if a loss of DHR -

were to occur during drained-down operations at Zion 4 days'after shutdown, fuel damage could occur within about 80 minutes.

(NSAC-84 data indicates that for some maintenance outages, drain-down of the RCS to sid-loop operation was,

reached within 4 days from time of shutdown).

Recent experience at other U.S. PWRs has shown that there have been many loss of DHR events during drained-down conditions which were caused by level measure-ment errors.

Many of these events lasted more than 80 minutes. There have also been many similar shorter duration events which resulted in the initiation of bulk boiling (see Table 3).

Review of plant operations during modes 4, 5, and 6 have shown that the key to prevention, nitigation and recovery from loss of the DHR function depend strongly cn operators and their ability to dependency upon human performance,perfors certain tasks. Because of the strong and the large error bands inherent in quanti-fying human reliability, the results of risk assessments for operations (estimated to be in the range of 2 to 5 x 10.s/RY) in modes 4, 5, and 6 are subject to large uncertainties.

This is noted in both C503 and NSAC-84.

While there may have been over a hundred loss of DHR function events that have been successfully mitigated in the past 10 years at U.S. PWRs, the potential for a sicious event is apparent particularly during drained-down conditions.

The freqmency of such events continues to be several per year even after exten-sive NRC and industry communications; the estimated probability is in the range of 10 5 core damage /RY and there is no assurance' that containment would be available; and often the operator, being the key element in loss of DHR func-tion events, is not provided with adequate information (instrumentation), or well thought out procedures, and training.

The cost-benefit analysis for the implementation of remedial actions shows that improvements can be made at modest cost and that the cost / benefit ratio justifies action (Enclosure 2).

The total cost range from $13 million to a savings of $321 million.

The benefits from averted doses range from 59,000 person rem to 177,000 person-rem.

"NSAC-84 assumes that fuel damage occurs when the RCS boils off to the cid plane of the core.

  • a 4

4 Conclusion l

1 Adequate justification exists for an appropriate generic communication requiring proept corrective action to minimize the loss of RHR during periods when the RCS is partially drained.

We trust that the composite list of recommended remedial cetion and the cost-benefit analysis will assist you in preparation of the l

g2neric communication.

{

4 AE00 is ready to assist your offices in the preparation and implementation of the generic communication.

Sidd f.v 3 is L 01x' des Edward Jordan, Director Office for Analysis and Evaluation of Operational Data

Enclosure:

As stated i

Distribution:

VStello

.N(Taylor i

JHSniezek F14fraglia j

RStarostecki Dross TSpeis natussell JNGrars JGKeppler RDMartin i

JBMartin EJordan CHeltemes i

TNovak t

YBonaroya i

JResenthal PLas FtDrnstein AEOD R/F ROAB R/F 1

DCS 2SEE PREVIOUS CONCURRENCES ROAB:05P: AE00 ROAB:DSP:AE00 *ROA8:DSP: AE00 *D:DSP: AE00 *DD: AE00

  • D:AE00 FtDrnstein:md Plan JRosenthal TNovak CHeltenes EJordan S/14/87 5/14/87 5/14/87 5/14/87 5/14/87 5/15/87

-J

Table 1 Chronology of 37 loss of DHR Events Attributed to inadequate RCS Level Docket Plant Duration Heatup l

344 Trojan 5/21/77 55 min.

Unknown 3/25/78 10 min.

Unknown

~

3/25/78 10 min.

Unknown 4/17/78 Unknown Unknown 334 Beaver Valley 1 9/4/78 60 min.

145 - 175'F l

366 Millstone 2 3/4/79 Unknown 150 - 208'F 272 Salem 1 6/30/79 34 min.

Unknown 334 Beaver Valley 1 1/17/80 Unknown Unknown 4/8/80 35 min.

0 4/11/80 70 min.

101 - 108'F 3/5/81 54 min.

102 - 168'F 344 Trojan 6/26/81 75 min.

140 - 150'F 369 McGuire 1 3/2/82.

50 min.

105 - 130'F 339 North Anna 2 5/20/82 8 min.

Unknown 5/20/82 26 sin.

Unknown 5/20/82 60 min.

Unknown 7/30/82 46 min.

Unknown 338 North Anna 1 10/19/82 36 min.

Unknown 10/20/82 33 min.

Unknown 369 McGuire 1 4/5/83 Unknown Unknown 339 North Anna 2 5/3/83 Unknown Unknown 280 Surry 1 5/17/83 Unknown Unknown 328 Sequoyah 2 8/6/83 77 min.

103 - 195'F 370 McGuire 2 12/31/83 43 min.

Unknown 1/9/84 62 min.

Unknown 344 Trojan 5/4/84 40 min.

105 - 201'F 316 DC Cook 2 5/21/84 25 min.

Unknown 368 ANO-2 8/29/84 35 min.

140 - 205'F 295 Zion 1 9/14/84 45 sin.

110 - 147'F 339 North Anna 2 10/16/84 120 min.

Unknown 413 Catawba 1 4/22/85 81 min.

140 - 175'F 327 Sequoyah 1 10/9/85 43 min.

<1'F 296 Zion 2 12/14/85 75 min.

  • 15' 361 San Onofre 2 3/26/86 49 sin.

114 - 210'F 382 Waterford 3 7/14/86 221 min.

138 - 175'F 327 Sequoyah 1 1/28/87 90 min.

95 - 115'F 323 Diablo Canyon 2 4/10/87 85 min.

100 - 220'F

t.

Table 2 Chronology of NRC and Industry Actions A - Chronology of NRC Actions USI A-45 (circa - 1980) originally focused on all phases of shutdown for PWRs and BWRs redirected in 1986, no longer concerned with modes 4. 5, and 6.

IEB 80-12/IE-IN 80-20 requested licensees to review Davis-Besse 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> loss of DHR (4/19/80), and to analyze their own plant's procedures, focusing on redundancy, administrative controls, and technical specifications.

Generic Letter 6/11/80 - Requested licensees to review St. Lucie's upper.

head voidireg event, amend technical specifications regarding DHR capability.

IE IN 81-09 discussed Beaver Yalley's loss of RHR (drain-down - Tygon).

NUREG/CR 4005 (Parameter, Inc., 6/85) closeout of IE Sulletin 80 Stated-that the issue of DHR operability was closed out at 75% of affected facili-ties (did not address operation during drained-down conditions, Tygon etc.).

AE00 Case Study C503 (12/85) - Addressed loss of DHR, included 32 events during drained-down conditions (1976-1984).

Indicated that the situation is not improving. Five major recommendations were made, including:

reliable level measurement, operator aids, improved procedures for DHR operations, improved procedures / training for-recovery fros loss of DHR events, improved technical specifications.

In response to CFC3, NRR noted it would include the recommendations of CSLs in GI-99 (interlocks).

To resolve this issue, Brookhaven National Lab is -

to extrapolate the Zion DHR PRA (NSAC-84) to other PWRs and assess the effect of implementing C503's recommendations.

A prelimin' / report is I

due in June, 1987.

Preliminary results indicate that cort neit frequency due to shutdown may be as high as 5.4 x 10.s/Ry (which is three times higher l

than NSAC-84's result). Brookhaven's preliminary results indicate that' implementing C503's recommendations say reduce the core seit frequency to about half that value.

IE IN 86-10112/86 " Loss of DHR due to loss of Fluid Levels in RCS" discussed events at SONGS 2 (3/86), Zion 2 (12/85), Sequoyah I (10/85),

and Catawba 1 (4/85).

NSAC-52.

Referenced AE00 Case Study C503, IE IN 81-D9, AEOD is presently contacting a foreign country for information on improved level measurement equipment.

IRS report #659 (8/86)' indicates that a foreign country is testing improved level gauges based on "different physical principles."

n i

4 Table 2 (Contined)

B - Chronology of Industry Actions NSAC-52 " Residual Heat Removal Experience and Safety Analysis, Pressurized Vater Reactors," January 1,1983.

NSAC-52 reported on 96 loss of RHR events that occurred at US PWRs from 1977-1981.

It concluded that procedures are the key to RHR system performance.

The report provided many suggestions for improv-ing RHR operations.

The suggestions addressed procedures and administrative controis relating to: maintenance and evolution planning; monitoring of reactor vessel level during partially drained operations; control over plant status, maintenance decisions, and outage coordination.

In addition NSAC-52 suggested improvements in human engineering and herdware, including: control room indica-tion; audible alarms for low RHR flow; redundant independent RCS level indica-ting systems; improved instrumentation; and improved data collection for shutdown operations.

INPO SER 79-84 " Loss of Shutdown Cooling Due To Inaccurate Level Indication" -

November 1984.

The SER discussed numerous events in which the DHR function was lost due to inaccurate RCS level indication and air-binding of the RHR pumps.

The SER noted the need for accurate RCS level indication and discussed methods I

for improving RCS level control.

The SER provided comments on the problems I

associated with using tygon tubing.

It also discussed air entrainment and vortexing, and it noted that methods for recovery from loss of DHR cooling should be included in operator training and procedures.

NSAC-84 " Zion Nuclear Plant Residual Heat Removal PRA," July 1985. The report presented a PRA for Zion during modes 4, 5, and 6.

It indicated that there were large uncertainties in the estimates of risk for shutdown operations.

It concluded that modes 4, 5, 6 may present significant risk reistive to operating modes 1, 2, and 3.

Core melt frequency for shutdown operations was estimated at 1.8 x 10 5/Ry.

IMPO SOER 85-4 " Loss of Degradation of Residual Heat Removal Capability in PWRs,"

August,1985.

The 50ER noted that probabilistic risk studies had identified loss of RHR as a significant contributor to the potential for core damage.

Other areas addressed in the 50ER w,ere automatic suction valve closures and loss of RHR pumps. The report stated that analyses had shown that under adverse conditions with a partially drained reactor it is possible to uncover the core within 15 to 30 minutes after loss of DHR due to boiling off the RCS.

The 50ER noted that controlling RCS level in the " required narrow range is a difficult evolution."

It referr6d to INP0 SERs 60-83, and 79-84 which point out the need for reliable RCS level information.

The 50ER stated that the use of cortain procedures, operational controls, training and hardware could have prevented many of the referenced loss of RHR events.

Specific recommendations addressed training, operating procedures and emergency procedures relating to drained-down operations.

l

)

INPO SER 17-86 " Loss of Shutdown Cooling Flow," May 1986. The SER discussed errors inherent in the tygon tube manometer system that was used for RCS level 1

s t

Table 2 (Continued) 8 - Chronology of Industry Action measurement:

the routing of the tygon tubing, and'the lack of operator awareness potential for vortexing.

The SER also presented potential corrective actions.

1 I

l

's a

Table 3

.Recent loss of DHR Events Which Occurred Durino Drained-down Operations Attributed to Inadequate RCS tevel Measurement Plant Date Duration Boiloff initiated Vaterford 3 7/14/86 221 minutes Yes North Anna 2 10/16/84 120 minutes No Sequoyah 1 1/28/87 90 minutes No Diablo Canyon 2 4/10/87 85 minutes Yes Catawba 1 4/22/85 81 minutes No San Onofre 2 3/26/86 49 minutes Yes I

ANO-2 8/29/84 35 minutes Yes i

-l i

l 1

J i

l i

i I

l l

l l

l

ENCLOSURE 1-Recommended Remedial Action for Reducing Risk from DHR Operations

~

(Based Upon NRC and Industry Sources)

(1)

Licensees should maintain containment integrity to the maximum extent practicable during periods of highest DHR risk (i.e., early stages of shutdown and drain-down operations).

It is recognized that the containment _ equipment hatch sust be open to allow cajor inspections or repairs during maintenance and refueling outages. Never-the-less licensees should take actions to minimize the risk to the public by:

delaying the time of opening the equipment hatch following shutdown, and improv-ing the procedures and training to minimize the time required to re-establish containment integrity during a loss of DHR event. ' For example, task analyses to integrate equipment hatch opening with the maintenance and refueling operations should be performed. Measures to permit reclosing of the equipment hatch during outages should be' developed based on the task analysis.

It should be recognized that operability of the containment purge valves.is relied upon during shutdown operations. We also note that during an accident inoperable containment purge valves could compromise containment integrity.

Therefore the task analyses should address the containment purge valves-and any other valve whose operation is needed to re-establish containment integrity during periods of highest DHR risk.

This item reflects the staff risk analyses based on NSAC-84 and BNL's on going work in support of GI-99.

The risk analyses contained in Enclosure 2 focused staff attention on the importance and benefit of containment integrity during shutdown operations.

(2)

Licensees should improve planning, coordination, procedures, and personnel training during shutdown to ensure the availability of hiR.

NRC C503, INPO SOER 85-4, NSAC-52, INPO SER 79-84 all recognized the importance of this issue and contained recommendations, suggestions and observations to this effect.

Wa believe that significant improvements in DHR system availability and reliability can be achieved by focusing on human factors aspects of plant shut-down.

Emphasis should be placed on detailed planning of test, surveillance and maintenance activities, and the equipment or system interactions which have frequently caused loss-of-DHR events.

In addition, plant practices regarding the procedures and training of personnel for performance of normal (non emergency) operations during shutdown should be evaluated.

For example: all operations and maintenance staff (licensed and non-licensed) should receive training to assure that they become sensitized to the risks associated with plant shutdown. Emphasis should be placed upon understanding the risks and high vulnerability associated with times of.high decay heat rate, drain and fill operations, disabling redundant safety equipment, etc.

ENCLOSURE 1

.g.

(3)

Licensees should have a reliable method of measuring and monitoring reactor vessel level during shutdown modes of operation and corresponding technical specification requirements for operability.

NRC C503, NSAC-52, INPO SER 79-84, INPO SOER 85-4 and INPO SER 17-86 all addressed the importance of reliable level instrumentation.

Common industry practice using unanalyzed makeshift devices such as tygon tube sight gages to monitor RCS level during plant shutdown should be modified or discontinued.

Reliable, redundant level indication should be required during modes 4, 5, and 6 to ensure availability of trending data, and to warn operators in advance of unacceptably low RCS level.

In addition, plant procedures should b2 modified to assure that the frequency of RCS level monitoring is commensurate 1

uith plant status (e.g., as noted in section 4.1 of C503, one plant could have j

conitored vessel level as infrequently as once every 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />, whereas fuel uncovery could occur only a few hours after a loss of DHR).

As a sinimum, each plant's safety review committee should review the instrumentation and procedures used for RCS level measurement during modes 4, 5, and 6 to ensure that a high level of reliability is achieved.

(4)

Licensees should perfore a task analysis of DHR operation.

NSAC-52 recognized the need for improvements in human engineering.

Perfomance of a task analysis per se is a specific AE00 recommendation.

We recognize that all DHR losses cannot be totally eliminated by good planning good procedures, well-trained personnel, etc.

We believe that if all licensees, l

would perform human factors anal normal and abnormal conditions) yses of their plant's DHR operations, (including and modify their plant practices and man / machine A codel to use for such human factors analyses is one us Reference 1 requires licensees to perfore specific task analyses, and to inte-dures for transients and accidents occurring during power ope should be required to perform sfallar reviews for shutdown operations, with Licensees emphasis on detection and mitigation of loss-of-DHR events.

available for recovery from postulated loss of-OHR events a as Figure 4 of C502, parametric curves of uncovery time Examples of such transients are:

primary system filled at maximus DHR system temperature primary system drained to minimum level and open to the atmosphere RCS at refueling temperature, etc.

Information on time margins available would assist operators in recognizing the potential seriousness of the event, and assist them in choosing appropriate methods for restoring the DHR function.

.S. Nuclear Regulatory Commission, " Clarification of TMI A ti U

Requirements, II.F.2 Instrumentation for Detection of Inadequate Core c on Plan Ceoling, (NUREG 0737), November 1980.

4

=-

3 1

(5)

Autoclosure interlocks should minimize loss-of-DHR events.

i NSAC-52 and NRC C503 both address interlocks.

1 1n order to prevent inadvertent DHR suction / isolation valve closures (during DHR system operation) it_is recommended that' NRR consider either requiring the removal of the autoclosure interlocks to the DHR suction / isolation valves, cr requiring removal of power to the DHR suction / isolation valves when valve motion.is not required. Prior to implementing this recommendation, it is necessary to ensure that there is adequate relief capacity to prevent _o'er-pressurization of the DHR system.

(6)

Plant technicai specific'ations should be modified to ensure that 'the DHR ~

system is available during mode 4 and the early stages of mode 5.

i While INPO SER 17-86 acknowledged shortcomings in plant technical specifications, modification of the technical specifications was recommended in AE00 C503.

Even-though NRR's generic letter of 1980 on DHR addressed DHR"systes redundancy, plant technical specifications do not require DHR redundancy' throughout periods when it is most.needed (mode 4 and the early stages of mode 5). Since test, mainte-nance, and other shutdown activities can be initiated during these periods it is apparent that as a result, a DHR loss could occur at a time when the risk is highest.

We recommend that NRR address the DHR system operating requirements and that plant technical specifications be modified to:

Ensure all plants have proper shutdown mode definitions (as discussed in sections 4.3 and 5.3 of C503); and Ensure that both trains of the DHR system are operable during periods of

)

high decay heat load, i.e. ; sode 4 and the early stages; of mode '5.

(The 1980 generic letter permits one train to be inoperable during-this time.)

Since the loss-of-DHR experience has not greatly improved following the issuance of NSAC-52 and NRR's generic letter, we believe that technical specification modifications are necessary to ensure adequata redundancy.

(7)

Licensees should analyze the hydraulics associated with drained-down operations.

Lovel measurement errors observed at Diablo Canyon 2 (April 1987), preliminary information from the AIT assigned to Diablo Canyon INP0 SER 79-84, and SER 17-86 which reported on probless resulting from gas, entrainment in tygon 4

measurement equipeent etc, all indicate that this issue should be addressed.

Large errors in RCS level measurements have been observed during drained-down cperations because of air or gas entrainment which resulted from draining or vanting operations, RNR pump vortexing, etc.

At many plants the elevations of the ' steam generator nozzles, pressurizer surge Ifne, reactor hot legs, and reactor coolant pump discharge are such that the there is little margin for measurement error prior to gas entrainment/vortexing.

The Diablo Canyon i

i

4 licensee ran tests which indicated gas entrainment caused erratic level

)

measurements.

We recommend that licensees perform a detailed hydraulic analyses of their plants' drain-down configuration to assure that the RCS leve1 measuring equipment remains accurate, and operators are aware of the range allowable RCS levels which will assure reliable operation of the RHR pumps.

j j

i i

i I

i l

l i

~

i

l ENCLOSURE 2 Cost-Benefit Analysis for Proposed NRC Generic Communication Loss of Decay Heat Removal Function in PWRs i

I.

Introduction This analysis provides an estimate of the costs and benefits associated with implementing plant and procedural modifications intended to reduce the likelihood I

of loss of the DHR function in modes 4, 5, and 6 at U.S. NRs.

The analysis was parformed based on the NRC's value impact methodology and it employed data which was extrapolated from the most comprehensive probabilistic risk assessment presently available for pressurized water reactors during shutdown (NSAC-84 July i

1985).

NSAC-84 presented the results of work that was performed by Pickard, Lowe and Garrick to quantify core melt frequency for the Zion nuclear plants i

l during modes 4, 5, and 6.

It reviewed operating experience at Zion 1 and 2 during shutdown.

It utilized detailed plant and maintenance logbook records to estimate availability and performance of systems and subsystems during modes 4, l

5, and 6.

Preliminary results from an NRC contractor working on this issue (Brookhaven National Laboratory), and AE00's review of recent operating experience indicate that the core melt frequencies appearing in NSAC-84 may be overly optimistic cnd the value of DHR system improvements recommended by AE00 may be signifi-cantly greater than the values listed in this cost-benefit analysis.

II.

Analysis Benefit - averted dose:

Based upon NSAC-84:

Core melt frequency due to operations during shutdown:

1.8 x 10 5/RY Installing a " perfect alarm system" to guarantee the operators are aware of loss of cooling would halve the core damage frequency to.9 x 10 5 The benefit of such a system is quantified as follows:

The equipment hatch is assumed open 1/2 of the time while the plant is shut down. The release is either a category 2 or 3 release.

l 4.8 x IOS person rem / accident 'L avg. = 5.1 x 105

,7 i

5.4 x 10 person res/ accident J Averted Dose = (.9 x 10.s) x (.5) x 5.1 x IOS = 23 person-rem RY

ENCLOSURE 2 4

Alternatively, per Generic Issue 99's prioritization in NUREG-0933, the core melt from a loss of RHR system would proceed as-T MLU of the Oconee RSSMAP analysis.

The releases would be as follows:

Category Probability O_ose (person-res) 3

.5 5.4 x 108 5

.0073 1.0 x 108.

7

.5 2.3 x 10:

Averted Dose =.9 x 10 5 x.5 x 5.4 x 105 = 24.3 person res RY PVR population (present plus future plants) y 55 reactors 1785 RY B&W 10 reactors 298 RY CE 15 reactors 485 RY 80 reactors 2568 RY Total Averted Dose = 23 person rem x 2568 RY = 59,000 person-rem RY averted Cost:

NRC Labor:

from NUREG-0933 Resolution of Generic Issue-99 (Interlocks)

Foryonly: 8 man-wks = $38,000 Assume CE & B&W require similar efforts 2 x $38,000 = $76,000 Total cost for interlocks

= $114,000 Assume a similar effort is needed for level measurement _$214,000 but that issue is more complex, and plant specific inspections will be Each plant will need to be inspected, procedures reviewed, etc.

neces sary.

Assume 300 hrs / plant x 80 plants x $50/hr = $1.2 M Total NRC labor cost = $1.4 M

CNCLOSURE 2

(

Cost:

Industry labor and hardware:

from NUREG-0933 Resolution of Generic Issue-99 (Interlocks)

NUREG 0933 estimated resolution of interlocks At W plants the cost would be $47,200/ plant (including hardware, licensing, reviev, technical specifications, etc).

Assume this cost would exist at all PWRs 80 plants x 47,200/ plant = $3.8 M Assume other hardware would also be used perfect alars," level instrumentation, improved planning, procedures etc. - ' assume these items cost 2 x as much as the interlocks (add $7.6 M)

Total industry cost and hardware labor = $11.4M Benefit:

Onsite property damage cost avoidance

- $2 x 10'/ core-melt x.9 x 10 5 core melt RY x 2,568 RY

= - $46 M ; however the present worth assuming 15 yrs avg and 5% discount rate is

- $23 M Benefit:

avoidance of non core-melt loss of DHR eventsCost reductio Shorten outages due to better planning estimate 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> /RY Avoidance of non core-melt loss of DHR events - frequency of non core-melt losses of DHR is one every 4 RY - assume such losses cause on average a 4-hour delay (extension of outage for a more severe event which includes investigation of the causes of inoperability and the effort required to assure that adequate corrective action is taken) -

the net delay = 1/4 x 4 = 1 hr/RY.

I 4 hrs x - $500,000 replacement Total

=

F x256B RY power cost i

per 24-hour day

= -$213 M present worth (15 yr. avg & 5% discount rate)

= _$107 M i

Onsite dose and Onsite dose avoidance are neglected.

I i

.ENCLOSilRE 2 b

Uncertainties 4

1 BNL has reviewed NSAC-84 and has added one or more accident scenarios.and has reexamined the models used for NSAC 84.

BNL has found that the core melt frequency presented in NSAC.84 is low by a factor a 3.

If BNL is correct then the benefit from averted dose should be 3 times'that listed-in this analysis.

69 )erson res 177,000 person-rem total'

.tY l

2 Time available for successful operator actions.to recover from loss of DHR.

HSAC-84 data indicates drain-down during maintenance outages were completed in 4 days or less from time of rod insertion.

The decay heat after 4 days is such that the drained-down system could.heatup and boiloff to the. fuel mid plane (criteria used for core damage in 4

NSAC-84) in under 80 minutes!

i However, the loss of cooling event trees assume operator recovery in 1-8 hours with mean error rates of lx10 6 to 2x10.a.

These rates appear to be overly. optimistic for actions which allow as little as 80 minutes for recovery from a high stress situation especially if the operators have no procedures, no training and inadequate information regarding the status'of. equipment availability...

severe loss of DHR events during drained-down operation w more than 80 minutes and there have been many shorter duration events which resulted in the initiation of boiloff.

For. example:

Plant Date Duration

  • Waterford 3 7/14/86 221 min North' Anna 2 10/16/84 120 min' Sequoyah 1 1/28/87 90 min
  • Diablo Canyon 2 4/10/87 85'ain

. Catawba 1 4/22/85 81 min i

  • San Onofre 2 3/26/86 49 min
  • ANO-2 8/29/84 35 min 4

3 NSAC-84 assumes that operator. recovery taproves with shift change, i.e., if there is a shift change, discovery / recovery from the casualty is assured.

DHR loss event experienceThis assumption does not agree with recent DHR event which was exacer; bated by the shift change.e.g., on 3/26/86

  • Denotes initiation of boiloff.

l

ENCLOSURE 2 III. Summary Cost Benefit NRC labor:

$1.4 M Industry labor + equipment:

$11.4 M Sum:

$12.8 M Offsite doses:

59,000*

person rem averted Propert

' Damage:y

- $23 M (could be as high as (could be as high

- $69 M) as 177,000 person-rem averted)

Replacement

- $107 M l

Cost:

(could be as high as

- $321 M)

I Total Cost Total Benefit

- $321 to $13 million

$59 - 177 million l

l l

w__-___-_-_-________.

Table 2 Chronology of NRC and Industry Actions A - Chronology of NRC Actions USI A-45 (circa - 1980) originally focuied on all phases of shutdown for FWRs and BWRs redirected in 1986, no longer concerned with modes 4, 5, and 6.

IEB 80-12/IE IN 80-20 requested licensens to review Davis-Besse 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> loss uf DHR (4/19/80), and to analyze t'seir own plant's procedures, focusing on redundancy, administrative controls, and technical specifications.

Generic Letter 6/11/80 - Requested licensees to review St. Lucie's upper head voiding event, amend technical specifications regarding DHR capability.

IE IN 81-09 discussed Beaver Valley's loss of RHR (drain-down - Tygon).

NUREG/CR 4005 (Parameter, Inc., 6/85) closeout of IE Bulletin 80 Stated that the issue of DHR operability was closed out at 75% of affected facili-ties (did not address operation during drained-down conditions, Tygon etc.).

AE00 Case Study C503 (12/85) - Addressed loss of DHR, included 32 events during drained-down conditions (1976-1984).

Indicated that the situation is not improving.

Five major recommendations were made, including:

reliable level measurement, operator aids, improved procedures for DHR 1

operations, improved procedures / training for recovery from loss of DHR events, improved technical specifications.

In response to C503, NRR noted it would include the recommendations of C503 in GI-99 (interlocks).

To resolve this issue, Brookhaven National Lab is to extrapolate the Zion DHR PRA (NSAC-84) to other PWRs and assess the ef fect of implementing C503's recommendations.

A preliminary report is due in June, 1987.

Preliminary results indicate that core melt frequency due to shutdown may be as high as 5.4 x 10 5/Ry (which is three times higher than NSAC-B4's result).

Brookhaven's preliminary results indicate that implementing C503's recommendations may reduce the core melt frequency to about half that value.

IE IN 86-101 12/86 " Loss of DHR due to Loss of Fluid Levels in RCS" discussed events at SONGS 2 (3/86), Zion 2 (12/85), Sequoyah 1 (10/85),

and Catawba 1 (4/85).

Referenced AE00 Case Study C503, IE IN 81-09, NSAC-52.

AEOD is presently contacting a foreign country for information on improved level measurement equipment.

IRS report #659 (8/86) indicates that a foreign country is testing improved level gauges based on "different physical principles."

I m____.-___

m-c I

j Table 2 (Contined)

B - Chronology of Industry Actions NSAC-52 " Residual Heat Removal Experience and Safety Analysis, Pressurized-Water Reactors," January 1, 1983.

NSAC-52 reported on 96 loss of P.HR' events i

that occurred at US PWRs from 1977-1981..It concluded that procedures are the key to RHR system performance.

The report provided many suggestions for improv-i ing RHR operations.

The suggestions addressed procedures and administrative controls relating to:

maintenance and evolution planning; monitoring of reactor vessel level during partially drained operations; control over plant status, t

maintenance decisions, and outage coordination.

In addition'NSAC-52 suggested improvements in human engineering and hardware, including:

control room indica-tion; audible alarms for low RHR flow; redundant independent RCS level indica-ting systems; improved instrumentation; and improved data collection for shutdown operations.

I INPO SER 79-84 " Loss of Shutdown Cooling Due To Inaccurate Level Indication" -

November 1984.

The SER discussed numerous events in which the DHR function was

)

lost due to inaccurate RCS level indication and air-binding of the RHR pumps.

The SER noted the need for accurate RCS level indication and discussed methods for improving RCS level control.

The SER provided comments on the problems associated with using tygon tubing.

It also discussed air entrainment and 1

vortexing, and it noted that methods for recovery from loss of DHR cooling should be included in operator training and procedures.

I NSAC-84 " Zion Nuclear Plant Residual Heat Removal PRA," July 1985.

The report presented a PRA for Zion during modes 4, 5, and 6.

It indicated that there were large uncertainties in the estimates of risk for. shutdown operations.

It concluded that modes 4, 5, 6 may present significant risk relative to operating n' odes 1, 2, and 3.

Core melt frequency for shutdown operations was estimated at 1.8 x 10 5/Ry.

INP0 50ER 85-4 " Loss of Degradation of Residual Heat Removal Capability in PWRs,"

August, 1985.

The 50ER noted that probabilistic risk studies had identified loss of RHR as a significant contributor to the potential for core damage.

Other areas addressed in the 50ER were automatic suction valve closures and loss of RHR pumps.

The report stated that analyses had shown that under adverse conditions with a partially drained reactor it is possible to uncover the core within 15 to 30 minutes after loss of DHR due to boiling off the RCS.

The 50ER noted that controlling RCS level in the " required narrow range is a difficult evolution."

It referred to INPO SERs 60-83, and 79-84 which point out the need for reliable RCS level information.

The SOER stated that the use of certain procedures, operational controls, training and hardware could have prevented many of the referenced loss of RHR events.

Specific recommendations addressed training, cperating procedures and emergency procedures relating to drained-down operations.

INP0 SER 17-86 " Loss of Shutdown Cooling Flow," May 1986.

The SER discussed errors inherent in the tygon tube manometer system that was used for RCS level

.. q Table 2 (Continued)

B - Chronology of Industry Action measurement:

gas bubbles in the l tubing, lack of procedural ' controls regarding the routing of the tygon tubing, and the lack.of operator awareness of the.

potential for vortexing.

The SER also presented potential corrective actions.

1 i

l l

l 4

l i

1

.. +

Table 3 Recent loss of DHR Events Which Occurred During Drained-down Operations' i

Attributed to Inadequate RCS Level Measurement Plant Date Duration-Boiloff initiated Waterford 3 7/14/86

.221 minutes Yes North Anna 2 10/16/84 120 minutes No i

Sequoyah 1 1/28/87.

90 minutes-No i

Diablo Canyon 2 4/10/87 85 minutes Yes Catawba 1 4/22/85 81 minutesx No 4

San Onofre 2 3/26/86 49 minutes-Yes

-ANO-2 8/29/84 35 minutes Yes.

i

-1 l

1

.l i

1 f

v ENCLOSURE 1 Recommended Remedial Action for Reducing Risk from DHR Operations (Based Upon NRC and Industry Sources)

(1)

Licensees should maintain containment integrity to the maximum extent practicable during periods of highest DHR risk (i.e., early stages of shutdown and drain-down operations).

1 It is recognized that the containment equipment hatch must be open to allow major inspections or repairs during maintenance and refueling outages.

Never-the-less licensees should take actions to minimize the risk to the public by:

delaying the time of opening the equipment hatch following shutdown, and improv-ing the procedures and training to minimize the time required to re-establish containment integrity during a loss of DHR event.

For example, task analyses to integrate equipment hatch opening with the maintenance and refueling operations should be performed.

Measures to permit reclosing of the equipment hatch during outages should be developed based on the task analysis.

4 It should be recognized that operability of the containment purge valves is relied upon during shutdown operations.

We also note that during an accident inoperable containment purge valves could compromise containment integrity.

Therefore the task analyses should address the containment purge valves and any other valve whose operation is needed to re-establish containment integrity during periods of highest DHR risk.

This item reflects the staff risk analyses based on NSAC-84 and BNL's on going work in support of GI-99.

The risk analyses contained in Enclosure 2 focused staff attention on the importance and benefit of containment integrity during l

shutdown operations.

(2)

Licensees should improve planning, coordination, procedures, and personnel training during shutdown to ensure the availability of DHR.

HRC C503, INPO SOER 85-4, NSAC-52, INPO SER 79-84 all recognized the importance of this issue and contained recommendations, suggestions and observations to this effect.

We believe that significant improvements in DHR system availability and reliability can be achieved by focusing on human factors aspects of plant shut-down.

Emphasis should be placed on detailed planning of test, surveillance and maintenance activities, and the equipment or system interactions which have frequently caused loss of-DHR events.

In addition, plant practices regarding the procedures and training of personnel for performance of normal (non emergency) operations during shutdown should be evaluated.

For example:

all operations and maintenance staff (licensed and non-licensed) should receive training to assure that they become sensitized to the risks associated with plant shutdown.

Emphasis should be l

placed upon understanding the risks and high vulnerability associated with times of high decay heat rate, drain and fill operations, disabling redundant l

safety equipment, etc.

m ENCLOSURE 1 e,

4 (3)

Licensees should have a reliable method of measuring and monitoring reactor vessel level during shutdown modes of operation and corresponding technical specification requirements for operability.

NRC C503, NSAC-52, INPO SER 79-84, INP0 50ER 85-4 and INP0 SER 17-86 all addressed the importance of reliable level instrumentation.

Common industry practice using unanalyzed makeshift devices such as tygon tube sight gages to monitor RCS level during plant shutdown should be modified or discontinued.

Reliable, redundant level indication should be required during modes 4, 5, and 6 to ensure availability of trending data, and to warn operators in advance of unacceptably low RCS level.

In addition, plant procedures should i

be modified to assure that the frequency of RCS level monitoring is commensurate with plant status (e.g., as noted in section 4.1 of C503, one plant could have monitored vessel level as infrequently as once every 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />, whereas fuel uncovery could occur only a few hours af ter a loss of DHR).

As a minimum, each plant's safety review committee should review the instrumentation and procedures i

used for RCS level measurement during modes 4, 5, and 6 to ensure that a high level of reliability is achieved.

(4)

_ Licensees should perform a task analysis of DHR operation.

t NSAC-52 recognized the need for improvements in human engineering.

Performance of a task analysis per se is a specific AE00 recommendation.

t We recognize that all DHR losses cannot be totally eliminated by good planning, good procedures, well-trained personnel, etc.

We believe that if all licensees would perform human factors analyses of their plant's DHR operations, (including normal and abnormal conditions) and modify their plant practices and man / machine interfaces accordingly, the risks from DHR losses would be significantly reduced.

A model to use for such human factors analyses is one used by NRR (Ref.1).

Reference 1 requires licensees to perform specific task analyses, and to inte-grate instrumentation, alarms and annunciators into normal and emergency proce-dures for transients and accidents occurring during power operation.

Licensees should be required to perform similar reviews for shutdown operations, with emphasis on detection and mitigation of loss-of-DHR events.

The operators should be provided with information outlining the time margins available for recovery from postulated loss of-DHR events as a function of time from reactor trip for a representative set of DHR loss transients (such as Figure 4 of C502, parametric curves of uncovery time vs. shutdown time).

Examples of such transients are:

primary system filled at maximum DHR system RCS at refueling temperature, etc. temperature primary system drained to m Information on time margins available would assist operators in recognizing the potential seriousness of the event, and assist them in choosing appropriate methods for restoring the DHR function.

1 U.S. Nuclear Regulatory Commission, " Clarification of TMI Action Plan Requirements," II.F.2 Instrumentation for Detection of Inadequate Core Cooling, (NUREG 0737), November 1980.

c'.

ENCLOSURE 1 (5)

Autoclosure interlocks should minimize loss-of-DHR events.

NSAC-52 and NRC C503 both address interlocks.

In order to prevent inadvertent DHR suction / isolation valve closures (during DHR system operation) it is recommended that NRR consider either requiring the l

removal of the autoclosure interlocks.to the DHR suction / isolation valves, or requiring removal of power to the DHR suction / isolation.alves when valve motion is not required.

Prior' to implementing this recommendation, it is necessary to ensure that there is adequate relief capacity to prevent over-pressurization of the DHR system.

{

i (6)

Plant technical specifications should be modified to ensure that the DHR

- i system is available during mode 4 and the early stages of mode 5.

i While INP0 SER 17-86 acknowledged shortcomings in plant technical specifications, modification of the technical specifications was recommended in AE00 C503.

i Even though NRR's generic letter of 1980 on DHR addressed DHR system redundancy, plant

)

technical specifications do not require DHR redundancy throughout periods when it is most needed (mode 4 and the early stages of mode 5).

Since test, mainte-nance, and other shutdown activities can be initiated during these periods it is apparent that as a result, a DHR loss could occur at a time when the risk is highest.

We recommend that NRR address the DHR system operating requirements and that plant technical specifications be modified to:

Ensure all plants have proper shutdown mode definitions (as discussed in sections 4.3 and 5.3 of C503); and Ensure that both trains of the DHR system are operable during periods of high decay heat load, i.e. ; mode 4 and the early stages of mode 5.

(The 1980 generic letter permits one train to be inoperable during this time.)

i Since the loss-of-DHR experience has not greatly improved following the issuance of NSAC-52 and NRR's generic letter, we believe that technical l

specification modifications are necessary to ensure adequate redundancy.

(7)

Licensees should analyze the hydraulics associated with drained-down operations.

l Level measurement errors observed at Diablo Canyon 2 (April 1987), preliminary information from the AIT assigned to Diablo Canyon, INPO SER 79-84, and SER 17-86 which reported on problems resulting from gas entrainment in tygon measurement equipment etc, all indicate that this issue should be addressed.

Large errors in RCS level measurements have been observed during drained-down operations because of air or gas entrainment which resulted from draining or venting operations, RHR pump vortexing, etc.

At many plants the elevations of the steam generator nozzles, pressurizer surge line, reactor hot legs, and reactor coolant pump discharge are such that the there is little margin for measurement error prior to gas entrainment/vortexing.

The Diablo Canyon

ENCLOSURE l'

, e licensee ran tests which indicated gas entrainment caused erratic level We recommend that licensees perform a detailed hydraulic' analyses measurements.

of their plants' drain-down configuration to' assure that the RCS level measuring equipment' reinains accurate, and operators are' aware of the range ' allowable RCS levels which will assure reliable operation.of the RHR pumps.

1 i

1 l

I l

l

i e,

a*.

j

,t i

ENCLOSURE 2 Cost-Benefit Analysis for Proposed NRC Generic Communication Loss of Decay Heat Removal Function in PWRs i

j I.

Introduction This analysis provides an estimate of the costs and benefits associated with implementing plant and procedural modifications intended to reduce the likelihood I

of loss of the DHR function in modes 4, 5, and 6 at U.S. PWRs.

The analysis was performed based on the NRC's value impact methodology and it employed data which was extrapolated from the most-comprehensive probabilistic risk assessment presently available for pressurized water reactors during shutdown (NSAC-84 July 1985).

NSAC-84 presented the results of work that was performed by Pickard, Lowe and Garrick to quantify core melt frequency for the Zion nuclear plants during modes 4, 5, and 6.

It reviewed operating experience at Zion 1 and 2 1

during shutdown.

It utilized detailed plant and maintenance logbook records.to estimate availability and performance of systems and subsystems during modes 4, 5, and 6.

Preliminary results from an NRC contractor working on this issue (Brookhaven National Laboratory), and AE00's review of recent operating experience indicate that the core melt frequencies appearing in NSAC-84 may be overly optimistic and the value of DHR system improvements recommended by AEOD may be signifi-cantly greater than the values listed in this cost-benefit analysis.

II. Analysis 1

Benefit averted dose:

Based upon NSAC-84:

Core melt frequency due to operations during shutdown:

l 1.8 x 10 5/RY Installing a " perfect alarm system" to guarantee the operators are aware l

of loss of cooling would halve the core damage frequency to.9 x 10 5 1

The benefit of such a system is quantified as follows:

)

i The equipment hatch is assumed open 1/2 of the time while the plant is

'i shut down.

The release is either a-category 2 or 3 release.

4.8 x 108 person rem / accident S r

5.4 x IOS person rem / accident [ avg. = 5.1 x IOS Averted Dose = (.9 x 10 5) x (.5) x 5.1 x 108 = 23 person-rem RY l

l l

s I

m ENCLOSURE 2 4

8

\\,

Alternatively, per Generic Issue 99's prioritization in NUREG-0933, the core melt from a loss of RHR system would proceed as T MLU of the Oconee RSSMAP 1

analysis.

The releases would be as follows:

Category Probability Dose (person-rem) 3

.5 5.4 x 108 5

.0073 1.0 x 108 7

.5 2.3 x'108 Averted Dose =.9 x 10 5 x.5 x 5.4 x 108 = 24.3 person rem RY k

PWR population (present plus future plants)

{

i W

55 reactors 1785 RY 8&W 10 reactors 298 RY

\\

CE 15 reactors 485 RY l

i 80 reactors 2568 RY Total Averted Dose = 23 person rem x 2568 RY = 59,000 person-rem RY averted 3

4 Cost:

NRC Labor:

from NUREG-0933 Resolution of Generic Issue-99 (Interlocks)

For W only:

8 man-wks = $38,000 Assume CE & B&W require similar efforts 2 x $38,000 =

$76,000 Total cost for interlocks

= $114,000 Assume a similar effort is needed for level measurement $114.000 i

but that issue is more complex, and plant specific inspections will be Each plant will need to be inspected, procedures reviewed,.etc.

j necessary.

j Assume 300 hrs / plant x 80 plants x $50/hr = $1.2 M.

Total NRC labor cost = $1.4 M i

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

ENCL'05URE'2 l

. c.

- 3.-

i i

Cost:

Industry labor and hardware:

from NUREG-0933 Resolution of Generic Issue-99 (Interlocks)

~NUREG 0933 estimated resolution of interlocks At W plants the cost would be $47,200/ plant (including hardware, licensing, review, technical specifications, etc).

Assume this cost would exist at all PWRs 80 plants x 47,200/ plant = $3.8 M Assume other hardware wou'ld also be used'

" perfect alarm," level instrumentation, improved planning, procedures etc. -

assume these items cost 2 x as much as the interlocks (add $7.6 M)

Total industry cost and hardware labor = $11.4M Benefit:

Onsite property dan; age cost avoidance

- $2 x 109/ core-melt x.9 x 10 5 core melt RY x 2,568 RY

= - $46 M ; however the present worth assuming 15 yrs avg and 5% discount rate is

- $23 M-Benefit:

Cost reduction from having shorter outages due to better planning and avoidance of non core melt loss of DHR events i

I Shorten outages due to better planning estimate 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> /RY l'

Avoidance of non core-melt loss of DHR events - frequency of non core-melt losses of DHR is one every 4 RY - assume such losses cause on average a 4-hour delay (extension of outage for a more severe event 3

I which includes investigation of the causes of inoperability and the

)

effort required to assure that adequate corrective action is taken) -

the net delay = 1/4 x 4 = 1 br/RY Total =

4 hrs x - $500,000 replacement x2568 RY E

power cost per 24-hour day

= -$213 M present worth (15 yr. avg & 5% discount rate)

= -$107 M Onsite dose and Onsite dose avoidance are neglected.

I

4

,. 5 Uncertainties 1

BNL has reviewed NSAC-84 and has added one or more accident scenarios and has reexamined the models used for NSAC 84.

BNL has found that the core melt frequency presented in NSAC 84 is low by a frtor a 3.

If BNL is correct then the benefit from averted l

dose should be 3 times that listed in this analysis.

i 69 3erson-rem 177,000 person-rem total-RY 2

Time available for successful operator actions to recover from' loss j

of DHR.

~

NSAC-84 data indicates drain-down during maintenance outages were completed in 4 days or less from time of rod insertion.

The decay heat af ter 4 days is such that the drained-down system could heatup and boiloff to the fuel mid plane (criteria used for core damage in NSAC-84) in under 80 minutes!

However, the loss of cooling event trees assume operator recovery in 1-8 hours with mean error rates of 1x10 s to 2x10 3 These rates appear to be overly optimistic for actions which allow as little as 80 minutes for recovery from a high

)

stress situation especially if the operators have no procedures, no training and inadequate information regarding the status of equipment availability...

Recent experience has shown that there have been many severe loss of DHR events during drained-down operation which lasted q

more than 80 minutes and there have been many shorter duration events i

which resulted in the initiation of boiloff.

For example:

Plant Date Duration

  • Waterford 3 7/14/86 221 min North Anna 2 10/16/84 120 min Sequoyah 1 1/28/87 90 min
  • Diablo Canyon 2 4/10/87 85 min Catawba 1 4/22/85 81 min
  • San Onofre 2 3/26/86 49 min
  • ANO-2 8/29/84 35 min 3

NSAC-84 assumes that operator recovery improves with shift change, i.e., if there is a shift change, discovery / recovery from the casualty is assured.

This assum DHR loss event experience; e.g.,ption does not agree with recent on 3/26/86 SONGS 2 had a loss of DHR event which was exacerbated by the shift change.

  • Denotes initiation of boiloff.

v-1

~ ENCLOSURE 2-

,,.3,

III. Summary

'(

Cost Benefit NRC labor:

$1.4 M I

Industry labor + equipment:

$11.4 M.

Offsite doses:

59,000*

Sum:

$12.8 M person rem averted 1

Property

- $23 M-(could be as high Damage:

(could be as high as as 177,000 person-

$69 M) rem averted)

Replacement

- $107 M Cost:

(could be as high as

- $321 M) l Total Cost Total Benefit

$321 to $13 million

$59 177 million j

I I

I i

)

a

m a+o.

-Y i

  • l t

s r, e

Table 1 Chronology of 37 loss of DHR Events Attributed to Inadequate RCS Level Docket Plant Date Duration Heatup i

344 Trojan 5/21/77 55 min.

Unknown 3/25/78 10 min.

Unknown 3/25/78 10 min.

Unknown 4/17/78 Unknown Unknown 334 Beaver Valley 1 9/4/78 60 min.

145 - 175'F 366 Millstone 2 3/4/79 Unknown 150 - 208'F 272 Salem 1 6/30/79 34' min.

Unknown i

334 Beaver Valley 1 1/17/80 Unknown Unknown 4/8/80 35 min.

O 4/11/80 70 min.

101 - 108*F 3/5/81 54 min.

102 - 168'F-344 Trojan 6/26/81 75 min.

140 - 150'F 369 McGuire 1 3/2/82 50 min.

105 - 130*F I.

339 North Anna 2 5/20/82 8 min.

Unknown l

5/20/82 26 min.

Unknown i

5/20/82 60 min.

Unknown 7/30/82 46 min.

Unknown I

i 338 North Anna 1 10/19/82 36 min.

Unknown 10/20/82 33 min.

Unknown 369 McGuire 1 4/5/83 Unknown Unknown 339 North Anna 2 5/3/83 Unknown Unknown 280 Surry 1 5/17/83 Unknown Unknown 328 Sequoyah 2 8/6/83 77 min.

103 - 195'F-i 370 McGuire 2 12/31/83 43 min.

Unknown i

i 1/9/84 62 min.

Unknown 344 Trojan 5/4/84 40 min.

105 - 201'F 316 DC Cook 2 5/21/84 25 min.

Unknown 368 ANO-2 8/29/84 35 min.

140 - 205'F 295 Zion 1 9/14/84 45 min.

110 - 147'F i

339 North Anna 2 10/16/84 120 min.

Unknown 413 Catawba 1 4/22/85 81 min.

140 - 175'F 327 Sequoyah 1 10/9/85 43 min.

<1'F 296 Zion 2 12/14/85 75 min, s15' 361 San Onofre 2 3/26/86 49 min.

114 - 210 F 382 Waterford 3 7/14/86 221 min.

138 - 175'F 327 Sequoyah 1 1/28/87 90 min.

95 - 115*F 323 Diablo Canyon 2 4/10/87 85 min.

100 - 220*F l

l

__m________-_____-ua---

-