ML20214W581
| ML20214W581 | |
| Person / Time | |
|---|---|
| Site: | Seabrook  |
| Issue date: | 09/25/1986 |
| From: | Speis T Office of Nuclear Reactor Regulation |
| To: | Novak T Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML20214K191 | List: |
| References | |
| FOIA-86-678 NUDOCS 8610030138 | |
| Download: ML20214W581 (3) | |
Text
-____ ____ __ __________________________ _ _______ _ ___
DISTRIBUTION:
CENTRAL FILES e
, SEP 2 51986 RRAB RDG FIbE TPSpeis BW5 heron LSoffer MEMORANDUM FOR: Thomas Me N6vak, Acting Director FJCongel Division of PWR Licensing-A RJBarrett Office of Nuclear Reactor Regulation FROM: Themis P. Spets, Director Division of Safety Review and Oversight Office of Nuclear Reactor Regulation
SUBJECT:
ADDITIONAL COMMENTS ON THE BNL REVIEW 0F THE SEABROOK EMERGENCY PLANNING STUDY REF.ERENCE: T. SPEIS TO T. NOVAK, "BNL REVIEW 0F SEABROOK EMERGENCY PLANNING STUDY," SEPTEMBER 15, 1986.
In response to your request of August 12, 1986, we reviewed the proposed tasks for the BNL evaluation of the Seabrook Emergency Planning Study. Our comments were forwarded in Reference 1.
\
Since that time we have had an opportunity to examine the Seabrook study results and to attend an August 27, 1986 briefing by the applicant and their PRA consultants. We now have further suggestions on areas which should receive emphasis in the BNL review.
A key element of the Seabrook study is the finding that early containment failure with large-scale releases of fission products would occur in only 0.1% of all core melt sequences. This estimate is based primarily on analyses which show low probability of a failure to isolate containment in the event of a core melt
, ( failure), an exceptionally robust containment design and an unusually low probability of an interfacing systems LOCA (V sequence). These are all important results which should be carefully checked and verified by the NRC contractor. It is particularly important to verify the low V sequence probability, since this result appears to be a consequence of improved analysis techniques rather than er.hanced design or operation of the plant.
' Seismically initiated sequences leading to containment failure or bypass should also be carefully reviewed.
The estimated frequency of an early containment failure with large-scale release of fission products (3E-7 per reactor year) is very low compared with staff assessments of similar plants particularly when external events are included and phenomenological uncertainties are accounted for. This raises the question whether small effects, which are ordinarily neglected in PRAs, might become significant contributors to risk for Seabrook. Assuming that the NRC contractor review upholds the bases for the low estimates of early release probability, NRC should also obtain assurance that heretofore neglected issues do not become dominant contributors to risk. The issues to be examined should include, but not necessarily be limited to, the following two areas:
- ~
y _ _
, OFC - -
T "*
NAME : : : .
_____:-____.......:...._______:___________:___________. ____--_-_g________.-__:-__---___-
DATE : : : ,: : emmum 5 -
0FFICIAL RECORD COPY h i oo3o138 Xh Sep -
8
/
T. Novak e
- 1. Off. Normal Operation In general, PRAs concentrate on accidents that would occur during power operation, because decay heat is much lower during non-power operation.
- However, from the viewpoint of early releases, there are some significant potential contributors from operation in modes 4, 5 and 6. Typically, technical specifications do not address the status of containment isolation in mode 5, and require isolation in mode 6 only during periods of fuel handling. Consequently, it is possible to have a core melt accident with the containment wide open.
Even with the containment isolated, the potential exists for an interfacing systems LOCA during RHR operation with a water-solid PCS.
Inadvertent PCS pressurization due to charging pump operation or heatup by the reactor coolant pumps, combined with inadequate response of the LTOP features, could lead to a pipe rupture or pump seal LOCA outside of containment. The probability of such a core melt sequence is low, s because additional failures of isolation valves and injection systems m0st be postulated. However, it remains to be shown that the frequency is less than 3E-7 par reactor year.
- 2. Exceptional Containment Failure _ Modes There are several potential modes of containment failura which are generally not examined in PRA analyses because they are Judged to be less 4
likely or less credible than the conventional failure modes. However, given the very low base case estimate of containment failure .
. probability, the NRC should consider whether such phenomena represent a containment failure probability of 0.1% or greater. Two examples of such phenomena are:
A. Local Hydrogen Detonations - although the Seabrook containment is sized and designed to withstand global hydrogen burns of any credible magnitude, tN potential always exists for hydrogen detonations in small comparfu;.ts which would produce shock wave pressures capable of causing a local containment. breach. The potential for pockets of high hydrogen concentrations in areas such as the pressurizer relief tank enclosure should be addressed.
B. Direct Core Debris Attack On Containment Penetrations - A large fraction of postulated core melt accidents will lead to vessel meltthrough with the primary system at high pressure. This will probably result in a forcible ejection of core debris into regions of the containment beyond the reactor cavity. The presence of hot core debris (molten or solid) in contact with metallic or elastomer penetrations holds the potential for structurally weakening components which are relied upon in the analysis of containment failure pressure.
~C : : : : : : :
ME : : : : : : :
TE : : : : : : :
0FFICIAL RECORD COPY
T. Novak A particular potential problem is the containment sump, because of its low elevation in the containment building and the presence of large diameter recirculation lines which penetrate containment.
A final point relates to the role of uncertainties in probabilistic risk assessment. The methods of probabilistic risk assessment have improved greatly in the past decade, but substantial uncertainty remains in the estimation of core melt frequency, containment performance and fission product behavior.
The BNL review should include an attempt to identify the major areas of uncertainty in risk and to evaluate potential improvements to the plant design and. operation which could significantly reduce those uncertainties.
"RTCIRAL srcuED BY Themis P. Speis, Director Division of Safety Review and Oversight Office of Nuclear Reactor Regulation cc: H. Denton E. Rossi V. Noonan E. Doolittle V. Nerses j
J 0FC :DSRO:RP SR0: RIB : #RO:D /: :
_ DSRO:RRAB)
_ _ _ _ _ : ._ ___:__________p____________:-_..60SRj
__ _G_ _ _[/ : ,
'______.____.f____:..__________:_
NAME :RJBarrett:sj:FJCongpl :LSoffer :BWSt n :TPS is : : :
JATE :09//(/86 :09/Jh/86 :09/ /86 :09/IY/86 :09 /86 : :
OFFICIAL RECORD COPY
,* "*feq'o UNITED STATES y
- f. If/C 8
g o
. ,i NUCLEAR REGULATORY COMMISSION WASHINGTON. D. C. 20555 4 :c/
AUG 121986 MEMORANDUM FOR: Harold R. Denton, Director Office of Nuclear Reactor Regulation FROM: Thomas M. Novak, Acting Director Division of PWR Licensing-A
SUBJECT:
REVIEW OF EPZ SENSITIVITY STUDY FOR SEABROOK On July 21, 1986, Public Service of New Hampshire (PSNH) submitted a sensitivity study on the emergency planning zone (EPZ). The study provides a comparison of dose versus distance curves for the Seabrook plant and site with similar generic curves from NUREG-0396 which were used in developing the EPZ regulation in 10 CFR 50.47. The study concludes that a 1 mile evacuation radius at Seabrook provides for a similar or greater degree of public protection than was shown
~ by NUREG-0396 for a 10-mile evacuation radius around the plants considered by WASH-1400.
The study is largely based on the Seabrook Probabilistic Safety Assessment that PSNH submitted about 3 years ago. The source terms used in the Emergency Planning Sensitivity study were drawn from the source terms used in the WASH-1400 calculations, with some modifications under specific scenarios.
Also, some of the probabilistic models have been changed from the Safety Assessment. Thus, the report is intended to examine differences made by the Seabrook design and site, plus the improvements in accident sequence modeling capabilities, without credit for source term reductions that may result from recent studies. The EPZ study attributes reductions in the offsite dose pre-dictions to the higher strength of the Seabrook containment, a more refined failure modes analysis for the containment, and a more realistic treatment of the initiation and progression of interfacing systems LOCA sequences. Along with the Emergency Planning Sens.civity Study, PSNH has also submitted a report titled "Seabrook Station Risk Management and Emergency Planning Study,"
which provides results of Seabrook specific calculations with new source terms based upon the recent IDCOR work.
The conclusions of the EPZ Sensitivity Study are based upon comparison of the results of the study to three acceptance criteria that were drawn from NRC documents. One of the criteria is a comparison of the individual risk of early fatality in the population within 1-mile of the plant, assuming no l
-y Pa** N eA 6/,3 a
Harold R. Denton AUG 181985 immediate protective action, to the NRC proposed safety goal. A second criterion is the comparison of early fatalities at the Seabrook site, assuming a 1-mile evacuation, to the early fatalities results of WASH-1400, which assumed a 25 mile evacuation. The third criterion is the comparison of the risks of exposure to 1, 5, 50, and 200 rem whole body doses at various distances from the Seabrook site to the corresponding NUREG-0396 results at 10 miles, assuming no immediate protective actions. It should be noted that a presentation on this general subject was made to NRC a few months ago by the AIF Subcommittee on Emergency Planning on behalf of the nuclear industry. The AIF proposal is currently under review in DSRO and IE; furthermore, it has been combined with the NRC initiated changes in
- EPZ related rules and regulatory practice.
In order to review the EPZ Sensitivity Study, it will be necessary to identify the baseline against which comparisons are made, to identify the appropriate criteria for making the comparisons, and to review the basic assumptions and the more significant aspects of the probabilistic calculations. We have met with representatives of IE, and they have agreed to provide guidance on the baselines and comparison criteria. They will be responsible for determining whether the study accurately portrays the principal conclusions of and technical material contained in NUREG-0396.
When the Seabrook Station Probabilistic Safety Assessment was submitted three years ago, the staff engaged in a review that was discontinued in January 1985.
This occurred due to funding restrictions on the part of the utility. The current review is intended to focus on those aspects of the PRA that contribute most to the differences in the results for public risks and doses. In this regard, it is noted that the core melt frequency of the updated Seabrook study is somewhat higher than the frequency estimated by WASH-1400 because of a more complete assessment of dependent events and component failure rates by the Seabrook study. However, the percentage of core melt scenarios of principal concern to emergency planners (i.e., early gross containment failure and con-tainment bypass scenarios) is more than 300 times less at Seabrook, primarily due to credit granted based on the strength of the containment. Therefore, our review should carefully evaluate the assumptions and analyses regarding the bebsvior of the containment and the probability of the containment bypass sequences. ;
Several areas that have already been identified for review are:
early containment failure frequencies and the sensitivity to assumptions of loading (e.g. , hydrogen detonation) and containment behavior (e.g. ,
local versus global response), ,
1 I treatment of source term for bursting type containment failure, ;
i l -
severe accident sequences involving containment bypass due to human l factors and hardware problems (for example, malfunction of air operated valves due to high ambient pressure inside containment),
l l t
1 1
I l
. , , _ . . . - - - . . ~ , _ - . . - .- . - _ _ . - - - - - - - , - . . - , . - - . . - - . . _ - _ . . . - - . _ . - _ - - - . - . . . . . - _ . . - . - -
. l Harold R. Denton MJG 121986 interfacing LOCA sequences which result in core melt and simultaneous 1
breach of containment with consideration of procedures used for successful isolation of containment, depth of treatment of severe accident sequences resulting from '
external events (for example, earthquakes),
consideration of conditional probabilities.
We expect to expand and refine the above list early in the technical review process.
The Division of PWR Licensing-A will coordinate the review. DSRO has essen-tial expertise in the appropriate issues and techniques, and it has familiarity with the Seabrook Probabilistic Safety Assessment and the previous review by the Lawrence Livermore Laboratory. We would expect that the recently formed Oversight Committee on Source Tern Related Technology would be working to define its role in this effort.
A technical assistance contract with Brookhaven National Laboratory will be used to support the staff effort. We have met with personnel from Brookhaven
~
National Laboratory, and they have proposed a three-month effort to review the Seabrook submittal. BNL has identified six tasks necessary to assist NRC in evaluating the technical validity of the applicant's conclusions regarding the Emergency Planning Sensitivity Study for Seabrook. Enclosure 1 contains a description of those tasks proposed by BNL. By copy of this memo we are requesting DSRO review of the proposed BNL tasks.
l A meeting was held on Wednesday, August 6, with PSNH personnel to brief the NRC staff on the basic content and conclusions of their EPZ study. BNL personnel were present for this presentation.
We have informed the ACRS of the PSNH EPZ study and our review of it. By a letter to the Commission dated April 19, 1983, the ACRS requested that they be kept informed of the staff's review of the Probabilistic Safety Assessment, and this has been done to date. At this time, the ACRS has provided a letter approving only 5% power operation for Seabrook.
Preliminary discussions have been held with ACRS staff for the purpose of scheduling a subcommittee meeting on the EPZ study some time in September.
The applicant has requested that the technical merits of the EPZ study be reviewed with respect to its adequacy to support a change to the emergency .
response process. The exact nature of the change has not yet been specified.
PSNH has further requested that the review be completed on an expedited basis.
A number of internal staff meetings were held within DPL-A with members of DSRO and I&E to discuss a plan for review of the Seabrook submittal. A draft of this memo was provided to DSRO and I&E for comment. We have accepted the ce pents provided by I&E, With regard to DSRO comments (Enclosure 2), we
Marold R. Denton -4 AUG 121986 believe they are directed to legal and policy considerations. (In earlier discussions, OELD did indicate that the Commission regulations would permit tkut staff to consider the merits of an exemption to the Seabrook EPZ.) We believe the decision chart suggested by DSRO has been essentially satisfied and a technical review can start. We have identified the essential technical issues which would be addressed as part of the BNL effort and the staff would be prepared to provide its evaluation by the end of October. We are proceeding with this approach.
A list of pertinent submittals on this subject is included in Enclosure 3.
Original signed by:
ThomasM. Kovsk Thomas M. Novak, Acting Director Division of PWR Licensing-A
Enclosures:
- 1. Proposed Tasks
- 2. Note fm Speis dtd 8/11/86
- 3. List of References cc. R. Vollmer F. Miraglia R. Bernero T. Speis W. Russell E. Jordan DISTRIBUTION Docket File (
PDf5 R/F g,
PDf5 Seabrook File TNovak VNoonan ERossi VBenaroya GBagchi 0Doolittle SLong
- See previous concurrences PWR-A:PD#5 PWR-A:PD#5 PWR-A:PD#5 PWR-A:F0B PWR-A:EB PWR-A:AD
- Stong:aj *EDoolittle *VNoonan *VBenaroya *GBagchi *CERossi 8/5/86 8/5/86 8/5/86 8/5/86 8/5/86 8/5/86 g6 k
8/ 86
ENCLOSURE 1 TASKS PROPOSED FOR REVIEW OF ENERGENCY PLANNING SENSITIVITY STUDY FOR SEA 8 ROOK Task 1: System Evaluation BNL will review those portions of the Seabrook Emergency Planning Sensitivity Study related to system failure to determine the appropriateness of the cal-caJ1ated accident sequence probabilities. In particular, the probability for interfacing system LOCA will be carefully assessed to' determine the potential for containment bypass. BNL will also review the probability of equipment an1 functions, personnel errors or design errors resulting in containment ttypass at the time of a severe accident.
Task 2: Containment Event Tree Review BedL will review the conditional probabilities of early containment failure l given in the Seabrook submittals. In particular, the vulnerability of the Seabrook containment to uncertainties in containment loads will be carefully assessed. This task will be highly coupled to Task 3, which will assess the
, puerformance of the Seabrook containment under severe accident conditions.
- Task 3
- Evaluation of Containment Behavior Tim purpose of this task is to evaluate the technical validity of the applicant's conclusions regarding the behavior of the Seabrook containment under severe accident conditions. BNL will review and evaluate the relevant containment structural analyses performed by the applicant and its consultants. In addi-tion, a plant site tour and engineering audit at the applicant's (or consultants')
office will be conducted to better understand the containment analyses and ensign, and to identify any unique design features and/or analytical assumptions tkhet merit further investigation.
i Based on the above review, BNL will develop an axisymmetric finite element model armd perform analyses utilizing BNL's NFAP computer code to confirm the applicant's prediction of the overall capacity of the containment. Special attention will
- he given to the post-cracking behavior of the concrete which controls the shear failure mode of the containment. To expedite the performance of this task, BNL will utilize, to the maximum extent practical, the input parameters obtained
'from the applicant's analytical models. In addition, simplified hand calcula-tions will be performed to assess the applicant's conclusions regarding the tuehavior of selected containment penetration assemblies. Finally, BNL will
, puerform a qualitative assessment of the appitcant's seismic fragility analysis of the containment structures and components.
BNL will also support meetings with NRC management and the ACRS to describe
- the interim status of this review, as well as the final results.
[
i i
1
.g.
Task 4: Review of Source Terms 1
The appropriateness of the new source terms based on RSS methodology used !
in the Seabrook submittal will be reviewed. '
Task 5: Site Consequence Modelina Trut site consequence modeling will be reviewed to determine the appropriateness of the consequence calculations presented in the Seabrook submittal. In addi-tion, any consequence calculations found necessary as a result of the work to be performed under Tasks 1-4 will be performed.
l Task 6: Final Report A final report due by October 31, 1986 will be prepared based on the results of Tasks 1-5. The final report will address BNL's recommendations on procedures, testing or design modifications to reduce the probability of containment bypass in conjunction with a severe accident.
Task 7: Follow-on Effort Follow-on effort in terms of resolution of issues will be provided under this task.
- . - , - - , - - , , . - - - -