ML20235T085

Comments on Draft NUREG/CR-4696 Re Containment Venting Analysis for Plant.Conclusion Re Zero Probability of Successfully Venting Following Station Blackout Totally Misleading

ML20235T085
Person / Time
Site:	Peach Bottom
Issue date:	07/18/1986
From:	Macdonald P IDAHO NATIONAL ENGINEERING & ENVIRONMENTAL LABORATORY
To:	Joel Jenkins NRC
Shared Package
ML20235T023	List: ML20235T021 ML20235T068 ML20235T085 ML20235T115 ML20235T139
References
CON-FIN-A-6842, FRN-52FR7950, RTR-NUREG-1150, RTR-NUREG-CR-4696 52FR7950-00027, 52FR7950-27, NUDOCS 8710090256
Download: ML20235T085 (11)

Text

.

H /NEA, %mr idaho National Engineering Laboratory --

hc July 18, 1986 f',8 5..,., ,

N , ,,,

Dr. J. P. Jenkins U. 5. Nuclear Regulatory Commission 5650 Nicholson Lane Rockville, MD 20852 TRANSMITTAL OF DRAFT NUREG/CR-4696, CONTAINMENT VENTING ANALYSIS FOR THE i

3 PEACH _ BOTTOM NUCLEAR POWER PLANT, JULY 1986, FIN A6842 - MacD-85-86

Dear Dr. Jenkins:

The enclosed report describes the results of research and analyses performed to determine the effectiveness of containment venting at Peach Bottom Unit 2 for specific accident sequences which are postulated threats to containment integrity. This transmittal completes Node 85-11 of the Accident Management Analysis milestone chart.

Since this is a Draft NUREG, it has not received patent clearance. _The report Diould not_be ouoted. disseminated. or referenced. We suggest that your review comments and those of othfar reviewers be available to us by August 26th. Please call Doyle Batt (FTS 583-9836) is you have any q u e s tiTDlEh----

Very truly yours,

~P. E Mubd gy.gg/- /

P. E. MacDonald, Manager Risk Analysis & Equipment Qualification JC:eb

Enclosure:

As Stated cc: N. Bonicelli, DOE-ID G. R. Burdick, NRC/DRA0 (10) /

A. R. Diederich, PECo R. O. DiSalvo, BCL F.Eltawila,NRC/ SORI (j5 F. E. Haskin, SNL (2) V S. A. Hodge, ORNL -

M. T. Leonard, BCL P. E. Litteneker, DOE-ID F. L. Sims, DOE-ID J. O. Zane, EG&G Idaho, w/o Enc.

9 EGrGsa. . p.o. sox 1s2s ideho ratis, to e3415 8710090256 070925 PDR NUREG 1150 C PDR

t y

COMMENTS ON DRAFT NUREG/CR-4696 DATED JULY 1986 1.- In a ntnt>er of locations this report states that the PBAPS Venting procedures do not serve their purpose in the avoldance or mitigation of severe accidents. Such inferences are inappropriate and do not. reflect an understanding of the basis for these procedures or an appropriate perspective from past PRA evaluations of BWR's. The BWROG Emergency Procedure Guidelines (EPG's) were developed in response to NUREG-0737. They represent a structured approach to the use of existing plant capabilities in response to out of nonel plant parameters. All EPG guidance is based on response to symptcms. One of the actions directed to manage increasing containment pressure is venting.

Previous BWR PRA's (WASH-1400, BFNP IREP, etc.) have Indicated that the risk dominant accident sequences were translent initiated events accompanied by a loss of decay heat' rerroval .

These sequences were found to be risk significant because It was assuned that containment fallure due to overpressure would lead directly to core melt. This asstmpt!on was made because procedures did not exist for the control of high containment pressures, definitive Information was not,available regarding containment pressure capability, and no analyses had been performed to define plant response to a containment failure event. , , .

This simplifying asstmption and the insight which it Impiles was actually in large part the basis for another major NRC research effort, that associated with USI A-45 Decay Heat Rerroval . IDCOR evaluations have indicated that consideration of existing containment venting procedures substantially alters the risk sign 1ficance of such sequences. This major insight is recognized at one location in the report but.this perspective is totally absent from the suTmary and conclusions. .

2. It is stated in the report that IDCOR studles "... were not sufficiently det. plied to reliably conclude that existing E0P's, vent path systems, and appropriate operator actions could be effective in. lessening an overpressure threat to containment i integrity." This is certainly arguable. IDCOR Tasks 21.1 and 23.1 demonstrate conclusively that the consideration of venting substantially alters the classical perceptions of BWR risk. This observation is actually supported by the facts presented in the INEL report.

l l

l 3 Venting is a last resort accident mitigation measure which is l Included in the Peach Bottom procedures to provide the operating staff the ability to preserve containment functionability (drywell intact and a scrubbed pathway for the release of any radionuclides). Venting as implemented at Peach Bottom has a

i

' ~

)

nurber of beneficial features which are nenifested in different ways depending on the sequence type. For example:

o Loss of contalnnent heat renoval sequences (which were identified in WASH-1400 as one of the highest risk contributors at Peach Bottom) can be successfully mitigated with containment venting. This action provides a redundant, diverse controlled nethod of containment heat renoval that has no impact on the successful operation of equipment required for safe shutdown.

/

o ATWS related events that could cause the containment to be breached due to overpressure can also be mitigated using the

. Peach Botton procedures since the contalnnent functionability as a fission product scrubber is preserved and the release of radionuclides is directed through the suppression' pool. I J

o Other accident sequences which can lead to core nelt prior to containment challenge leave substantial nergin for recovery of containment heat renoval capability before the containment is challenged. For these sequences there is substantial time to recover power to the containment heat renoval systen (l .c. - RHR or venting) before containment breach is assuTed to occur. Recovery of power to allow operation of RHR or venting within a typically available tine frane of'24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> has a conditional failure probability on the order of 0.01. Therefore, c*ontalnnent heat renoval should not be a Ilmiting paraneter, even for station blackout events.

4. The report concludes that "... the effectiveness of venting as a neans to reduce risk has not been demonstrated" while the actual INEL evaluations show that it does effectively reduce risk while not entirely eliminating it. Venting has not been touted by the industry as a panacea for all BWR risk issues. The intent of this procedure is not to demonstrate mitigation capability for all possible accident scenarlos. It is to allow the operating staff additional flexibility to cope with sequences which ney arise and can benefit from the existing hardware and plant capability. Sequences such at TW, TQUX, TWUV, LOCA, LOOP, and certain ATWS cases all can benefit from the venting capability.

Since these sequences can represer,t a substantial fraction of the risk as identified in published industry sponsored BWR PRAs, these procedures represent a real positive increase in safety.

5. The report states that "The primary impediment to success is the low likelihood that the operator can successfully implement l

venting as currently specified." This is an inappropriately

definitive statenent regarding a relative issue. The Ilkelihoods l of venting success detennined by INEL are quite substantial for all but two specific evaluations. For these two evaluations snell changes in sequence definition or assumptions would dramatically inprove the assessed likellbood of success.

I w - - _ _ . _ _ _ _ _ - _ _ _ _ _ _ . _ _ _ _ _ _

_3.

6. The report concl udes that ". . . If venting is part of the strategy for reducing the risk associated with BWR Mark I reactors, means nust be identlfled to increase the probability of success for venting, either through hardware or procedural changes." This i sweeping conclusion cannot possibly be reached based on this l limited scope study. It can only be reached after a definitive study of all sources of BWR risk and subsequent to a detennination that this risk is unacceptably high. [

The risk at Peach Bottom is believed to be low enough. Therefore, no significant hardware changes have yet been considered appropriate to increase the effectiveness of these procedures. It is acknaaledged that venting capability could be somewhat increased if additional plant modifications were nede.

7. The evaluation of risk at a nuclear plant requires the integrated I assessnent of all the types of accidents which could result in core danege or neiting of the fuel. The accident mitigation neasures and operator actions for the different types of accidents can be substantially different, depending on Itans such as the folloalng:

- Support system failures (e.g. power, service water)

- Operator Indication Procedures

- Training

- Technical support center assistance A proper evaluation of the success and benefits associated with r specific operator actions, such as venting, requires such an i Integrated examination of the plant and its response. Due to the limited scope of the INEL review, this type of an evaluation has not been perforned. Examination of the body of the report Indicates that risk was not included as one of the basic questions to be addressed in the INEL analyses.

8. The INEL report identifies a nunber of uncertainties which can impact the results of the detenntnistic calculations perforned:

o The ability of the vent valves to be open, c Other containment failure mechanisms.

The power level at TAF.  ! l I

o Alternative severe accident strategies.

The ability to control water level at TAF.

Containment failure pressure.

l l

.- t, - -

~~

Given these uncertainties, it appears that the NUREG conclusion "that easy procedural <and hardware fixes must be required" as stated in the abstract cannot be supported. Rather, it appears that the NUREG has confinned the positive benefit that is achieved under a nud>er of scenarios, but also has identified areas of uncertainty which nay limit the noxintm benefits that can be achieved regardless of how the venting procedure and hardware are improved.

9. Published BWR PRAs have identified that there nuy be a spectrun of potential contributors to core nelt or containment challenge that can arise for a wide variety of reasons. In addition, sufficient analysis has been done to indicate that these sequences are highly uncertain and, therefore, they ney be more or less important on an absolute scale and relative to each other.

This uncertainty is related to assumptions, training, unique maintenance occurrences, weather, etc. This uncertainty means that we cannot dismiss portions of the spectrun from consideration, particularly when trying to assess the benefits and conpeting risks associated with a change or a plant feature, (i.e. - we cannot be too narrow in our focus).

The use of two (2) ATWS sequences and a unique definition of station blackout can only lead to the conclusion that, for these uniquely defined sequences, the venting capability nay have a sna11 positive benefit. By concentrating on these sequences, the analysis has a predestined result. By ignoring the more likely sequences as identified in published PRAs and their reviews, it is clear that the report suffers fran a scope which is too narrow to provide adequate insights into the efficacy and benefits of venting.

The doninant accident sequences are not known with sufficient accuracy te limit the scope of analysis to three sequences, disregard < % " seau 7:es, and then come to conclusions which purport to 6escribe the risk associated with Peach Botton. Even the three sequences considered are so narrowly defined that sna11 variations (i .e. - lower initial power, sone rods inserting, sone other partial successes) could have significant positive benefits which are not acknowledged anywhere in the report. By bcIng very prescriptive in the accident scenario definition, the NUREG has apparently overlooked other probabilistic variations of the sequences which can be effectively mitigated through venting.

This narroa focus limits the ability for this evaluation to be used to nake sweeping conclusions and generalizations.

10. INEL reconnends that additional work needs to be done to address

.. . equipnent needed to restore the plant to a safe, stable state." Nothing has been identified which would indicate that a vented configuration is not a safe, stable state.

11. INEL reconnends that additional study is needed to determine "how to reisolate a containment. once venting has occurred, and how to proceed if reisolation is not possible." As stated above, there

r,

.is not known reason that a vented configuration should not be considered a safe-stable state. Additionally, merely allowing the valves to close (deenergized) will cause reisolat. ton.

12. INEL reccnmends additional study to address "the relationship of venting to emergency management, including guidance to the surrounding populace if reclosure is not possible." There is no reason to believe that existing envargency manage: Tent plans, including alert and evacuation plans, are not suitable for this or any other situation which involves the release of radioact;vity from the plant.

13. A substantial variation in the deterministic evaluation performed of station blackout results if the AC power recovery is included in the assessment. Specifically, there is an extended period of time available for the recovery of AC power between the time of core melt and time of containment' overpressure failure. This time frame (between where the NUREG 1150 calculation leaves off and the point at which containment is asstmed to fall) affords an extended time frame for recovery of AC power and restoration of venting if required (i.e. - If not successfully manually initiated). The conditional probability of restoring AC power would be expected to be quite high, particularly at Peach Bottom where hydroelectric units with "blackstart" capability exist nearby.

The conditional probabilltles without consideration of the hydroelectric capability are approximately:

2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> .25 success 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> .5 to .75 success 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> .98 success it is clear from these estimates (based on restoration of offsite or onsite AC power) that there could be a substantial venting success probability depending on the precise sequence of events.

Therefore, the conclusion regarding a zero probability of successful venting following a station blackout is totally mlsicading. It may be true that for the sequence identified in the NUREG that the likelihood of successful implementation of the asstmed vent procedure would be low, however it is also true that the procedure is different and the likelihood of restoration of AC power during the windcw of time before containment failure would allow successful venting to be carried out.

14. Much is nude of the fact that the ductwork downstream of the 18" vent valves is likely to fall vhen these vent paths are utilized. Information provided by PECo relative to their experience in using this ductwork for depressurization from ILRT pressures has apparently been ignored. Even so , ductwork failure was recognized in the IDCOR evaluations and, in fact, resulted in significant fission product disposition as the gases ficwed

~ " - - - - - - - - _ - - , _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _

through the reactor building. It is stated that closure of the fire damper in the ductwork between the 18" vent valves and SGTS would "inproperly terminate" venting. Damper closure could only l be another factor which would cause releases to pass through the reactor building before-release.

15 The particular event sequences considered in the report are characterized both as " generic" (p. 8) and " unique" to PBAPS (p. 6). It should be noted that other NRC sponsored studies investigating containment response issues have selected the TQUX (loss of high pressure nakeup and failure to.depressurize) sequence as a surrogate for all challenges to BWR containments.

16. -The contalnnent failure pressure assumed for this study is based on an outdated and overly simplistic evaluatlon of the Browns Ferry containment by Anes Laboratory. The shortcomings of this study are recognized in Section 4.2.5 but the significance of this deficiency is overlooked when insights and conclusions are discussed. The effect is further conpounded by placing an uncertainty on the failure pressure, giving credence to values lower than what is, in effect, a lower bound failure pressure.

17. For the station blackout evaluation it is assuned that DC power is exhausted in 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> causing a loss of high pressure injection at that tinc. This assumption, which we have advised does not reflect a realistic assessnent of PBAPS battery capabl.11ty, conpletely dominates the course of this sequence and the evaluation based on it. Evaluations by IDCOR and others have demonstrated that only slightly longer high pressure injection

• will cause the sequence to change such that the containment will reach the venting inttration pressure before core nelt occurs.

For this minor change of assumption, the results of INEL's evaluation will be dramatically different.

18. The operator task descriptions in the INEL report indicate that a l decision to vent is node after reading the venting procedure.

This is not correct; the venting procedure does not call for such a Judgement to be nede.

19. . The report states that the PBAPS venting procedure ". .. specifies the use of the 18" torus vent paths .. ." for the blackout sequence. This is an error; the 6" vent path is specified preferentially to the 18" path for all but ATWS sequences.

Correction of this error would substantially alter INEL's evaluation.

20. The report says that a " procedural analysis revealed neny items which require correction before the Peach Bottom venting procedure could be effectively used ...". The specific convents  ;

provided by INEL on procedure T-200 are generally insignificant or natters of style. Even if uncorrected, they would have little bearing on the effectiveness of the procedure. Hoaever, the procedure has been revised to incorporate the nore substantial of -

the connents provided. It is not stated if these Itens are

assmed to be corrected for purposes of INEL's evaluation. For some unstated reason the report advises that "... an HRA performed on the final venting procedure would result in different HEP's."

21. Footnote a. on page 26 seemingly raises doubt as to the effectiveness of venting in breaking the asstmed '11nk between containment overpressure failure and core melt. The statements made, however, ignore the existence of sources of makeup which do not rely on the torus as a source of water and which are not located in the Reactor Building (1.e. - CRD punps, condensate ptnps, HPSW ptmps, etc.).

22. This report states that, under TCl sequence conditions, "It is

!c_ difficult, lf not impossible, for. the control rocm operators to l; control reactor power level". There is no basis for such a I conclusion within the scope of this study. This complex topic has been studied by others in grebt depth and different conclusions have been reached.

23. INEL asstmes that HPCI and RCIC fall at a suppression pool temperature of 200F despite Information provided by PECo which j indicated operation would be expected up to at least 260F. From

> this technical information, and changes in Peach Bottom procedures

,I to allow HPCI suction to be swltched to the CST, it is clear that

! adequate water supply to the reactor can be available for extended periods of time. The impilcation of this is to reduce the likelihood of the TC2 ATWS sequence as defined in the NUREG analysis and make reaching venting initiation pressure achievable prior to core melt. This is a crucial issuTption in the analysis not even discussed in the uncertainty section. -

24. Treatment of CRD ptmp operation appears to be incorrect for the blackout sequence. It is indicated that "CRDHS is asstmed not to continue operating after battery exhaustion." This is not battery pcwered equipment - AC power is required for CRD ptmp operation.

4 25. It is stated that reactor building equipment is quallfled for conditions of "no htmidity and temperatures less than 200F".

This is incorrect. Much of the reactor building equipment is qualified to withstand harsh environments due to factors such as

high energy line breaks.

26. Concern is expressed regarding the possibility that the Yarway

! reactor level instruments could beccme unreliable during depressurization due to reference leg flashing. These Yarway Instruments are currently being replaced. Additionally, EPG and

, TRIP guidance currently exists to handle such a situation for all events (including ATWS). This topic was not discussed with the NRC contractors since they didn't raise the subject and since it was believed to be outside of their defined scope of review.

(

27. Section 4.2.4 raises a concern regarding torus-drywell vacutm breaker operability. The substantial design and qualification

, information which exists regarding these components was not s ., d

discussed with:INEL since they didn't raise the subject and since -

it was believed to be outside of their defined scope of review.

28. Section 4.2.6 raises a question regarding steam condensation effectiveness in suppression pools. This topic has been extensively studled but was not discussed with INEL since they didn't raise the subject and since it was believed to be outside of thelr identified scope of. review. ,

29. .The possibility of containment failure by corlun nelt-through or overtenperature failure is raised in Section 4.2.5. Ava11able evaluations of these topics were not reviewed with INEL since they didn't raise the issues and they weren't believed to be within their defined scope of review.

30. SLIM analysis as used in the INEL human factors ' analysis has been considered in detall for other appilcations and rejected by both PECo and IDCOR. There 1s no reason to believe that the " experts" employed were provided with appropriate infonnation and perspective, or were qualified to render expert opinion based on relevant experience and background. Other published and more widely accepted methods of analysis should be anployed.

The calculated value for failure to vent during the postulated TC2 sequence is too high based upon the tine available, the training, and the written procedures. Hunan response model199 provides substantial reduction in procedural response error rates as a function of time. For the TC2 sequence, the dominant contributor appears to be the SLIM methodology itself. A high stress response curve such as used in the Shoreham, Limerick, and Big Rock Point PRA's would give conditional failure probabilities on the order of .01 to .001 for the available time.

This is compared with the greater than .1 used as the "best estimate" in the INEL report. The methodology employed on these PRA's has previously been reviewed and accepted by the NRC and its Contractors. ,

i

31. It is assumed that no preparatory work would be done in anticipation of venting (i.e. - first consideration of venting is assumed at 60 i psig). This noy be appropriate for the fast-moving ATWS sequences, but It is overly conservative for other sequences.

32. A note on p. A-41 Indicates that 4-18" vent paths will be insufficient to stabilize containment pressure for sequence TC1 12% power. This is inconsistent with previous PECo, IDCOR, i and NRC analyses. No explanation of the source of this Information is provided.

33. Appendix A concludes that "such things as thorough familiarity ,

with the procedures and overlearning through simulator training I in the events and procedures would greatly assist." INEL did not pursue this subject with us. The described approach to training h employed for all PECo emergency procedures.

4

34. With regard to the timing of core melt onset and containment pressure reaching venting initiation pressure, the ATWS sequences considered by INEL do not agree with other rrodels.

35. The vent area determined by INEL to be required to successfully stabilize containment pressure under TC1 ATWS conditions is much larger than shown in previous analyses. No plausible explanation is offered. it appears that overly conservative assumptions were made. Failure pressures in excess of the lower bound 117 psig used in the analysis should be considered.

36. A statement is made that operator action time estimates were obtained fran PECo. This is only partially true since conments provided by PECo on the time estimates used by INEL were not resolved. The total sequence timing differences are substantial for both the ATWS and blackout scenarios. These differences do not pertain to anticipatory preparation for venting.

37. The report states that "... Instructions for local operation of vent valves apply only to the 18" vent valves." This is an error. Instructions for manual operation of all vent valves are included.

38. The report is poorly organized and written. The various sections are inconsistent and in a ntmber of instances contradictory.

Materials are repeated in a ntmber of locations and the swmary and conclusions do not reflect the work actually performed.

39. Confidence limits on success probabilities appear to be derived frau using the high and low probabilities for operator actions in the fault trees, although there is no description of how the confidence limits were obtained in the report. This is not mathematically correct and the values presented could only be viewed as upper and lower bounds.

40. A ntmber of the operater action probabilities shown in the Appendix A logic trees do not seem to be reasonable. For exanple, " fails to sultup in anit-Cs" should have no effect on the success of venting, and the likelihood of opening the second valve in a vent path should be higher than for opening the first.

41. The values employed by INEL for suppression pool scrubbing effectiveness are unreasonably low. For reasons that cannot be discerned from the report, the SPARC code as onployed for this study has provided suppression pool DF's which are substantially lower than experimental evidence would indicate. DF's applied for "invessel" releases are low by at least a factor of 4, while DF's for "exvessel" releases are also low by a substantial margin.

Additionally, accurate DF's cannot be determined without performing species-specific evaluations (i.e. - heavy species such as Ba 1 and Sr will have higher DF's).

1

~

l L

_ 10, _

~ -

f te 2. The tredels for rnelt progression / reactor.-vessel failure and i

! hydrogen generation used in this analysis appear to be arrong the 1 rrost adverse. These topics are currently under discussion as l- part of the IDCOR-NRC issue resolution process and will therefore not be expanded upon here.

I DRH/pdO8258602 1

1 1

l i

I

_ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ . _ . _ _ _ _ _ _ . _ _ _ - - - -- --