ML20149L921
| ML20149L921 | |
| Person / Time | |
|---|---|
| Site: | Maine Yankee |
| Issue date: | 09/16/1996 |
| From: | Wolf L MARYLAND, UNIV. OF, COLLEGE PARK, MD |
| To: | Lauben G NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML20148C622 | List: |
| References | |
| NUDOCS 9612040259 | |
| Download: ML20149L921 (10) | |
Text
,
.(... .... . .. ... v..-:-: .-:.
- . . . . . - . - ...- hn e-e Dr. Lothar T. Wolf I University of Maryland Depanment of Materials and Nuclear Engineering Bldg. 090, Room 2303 College Park, MD 20 742-2115 I
I 4
Mr. G. Norman Lauben
- U S Nuclear Regulatory Commission l
! _ OfTice of Nuclear Regulatory Research l RES
, Washington, D.C. 20555 - 0001 l
- i September 16,1996
Subject:
Copy of summary letter report on
, " Peer Review of Maine Yankee Independent Safety Assessment Team Findings"
}
} RE: LMrrCO Purchase Order Number: C 96 - 176102 1
] Appendix - A, Statement of Work l 4
Dear Mr. Lauben:
Attached, please find a copy of my summary letter repon on the evaluation of ISAT-findings for your information and considerations as deliverable for the purchase order i number cited above. , ;
As member of the peer review panel,1 like to thank you and your colleagues for the efforts and patience to provide us with the requested informations and answer all our questions.
Sincerely yours,
/ Enclosure L c%3 Y f
[_+ofmA ,
_ 3: .; .n-
. :: _ >.. a. -: ;- - ::# :: : > :J NGE 01 i !
1
-e
SUMMARY
LETTER REPORT l
l PEER REVIEW OF MATNE YANKEE INDEPENDENT SAFETY ASSESSMENT TEAM FINDINGS i
l l
l I
t l
l l
l L WOLF Rochille, MD s September 16,1996 l
l l
The findings and conclusions provided in the following express the personal opinion of the author based upon the written technical informations provided and the two peer
- ~ review panel meetings with the ISAT, l
I j
I i i i
I
- .g..'..
- -
. . ::_- .. n. :-: ..:. .
. .__' -~ __ . :E -;-
['
i FOREWORD This summary letter report solely focusses on the key findings on major techn cal and organisational findings by the peer review panel.
Therefore, the conclusions, recommendations and suggestions address key issises in .
the context of the Maine Yankee ISAT findings from the perspective of the peer review panel and highlights some of them which deem of generic nature.
j Technical details of computer code aspects, modeling, nodalization and validation have been discussed at length during the two meetings with the ISAT and advice prov;ded accordingly. They are beyond the scope of this summary letter repon 1
l i
l i
l l
l l I
t t 1 l
l l
.s - . -- . .. ... ..-:-: . =
. ~~_ r *,x. c__
3 1 BACKGROUND i LMITCO has been chanered by U.S NRC with the task of assembling a small, )
independent peer review team for evaluating the findings of the Independent Sefety )
Assessment Team (ISAT), which conducted an assessment of the Maine Yankee l Nuclear Power Plant. ISAT reviewed, assessed and audited the methodology. l I
computer codes, their validation, application and results 1 l
Members of the peer review included: l l
Dr. H. Sullivan, LANL, Los Alamos Dr. M. Thurgood, Richland Dr L. Wolf, UMCP, College Park Dr. N Zuber, ACRS observer The peer review team met twice with the members of the 1 SAT, who were'responsibi:
for the assessment of the analytical support provided by Yankee Atomic Electric Company (YAEC) for the Maine Yankee Nuclear Power Company in the area of non.
LOCA safety analyses as per Chapter 14 in the FSAR and specifically the main stean-line break and control rod assembiy drop. The two meetings took place August 7 through 8 and September 12 through 13 at NRC's headquarter in Rockville, wt ere ISAT briefed the peer review team about its findings of field trips to YAEC. Prior to these meetings and during these meetings an enormous amount of material was provided to the peer review team members. Additional informations were provided b e ISAT during the meetings upon the request of the peer review team members.
Based upon the technical repons, written answers to ISAT questions, copies of YAEC presentation slides, publications, interview sheets, computer code summaries, briefings by and discussions with ISAT, the following observations, conclusiors and recommendations emerge.
2 TECHNICAL OBSERVATIONS AND SUGGESTIONS a) The reactor physics codes and code suites CASMO, SIMULATE - 3 and STAR receive superior licensee attention, effective control and state - c f- the
- an validation. The comparisons between measured and predicted power distributions for several MY fuel cycles demonstrate excellent agreement and a state . of- the - an knowledge base in this area, b) The fuel performence code FROSSTEY, after previous review by BPN'-staff and resultant modifications receives superior licensee attention, efTective control and has been verified by comparisons with a large PWR fuel data base c) The application of the core / bundle / subchannel code COBRA Ill C in combination with the unique YAEC - 1 DNBR correlation has been previously reviewed by BPNL and receives superior licensee attention, effective control and has been consistently applied together with the statistical combination of uncertainties.
. . -~ - - . .. - -. . - . - - - - . ------- -
ci .i . tii c: : Oc. +;.-iai casi ::: :: ._:
. : :: P.4E OE 2
3 4 l
Although the MY cycle 15 core consists of a combination of fliel from sciveral .
vendors, the implemented fuel procurement procedure enforces conformance !
l -with geometrical details and fabrication tolerances such. that DNBR as Ggure l of merit remains uneffected. J i
( l d) The simplified system code GEMINI - 11 for the analysis of pressure and
{
temperature changes for symmetrical system behavior has received acceptabh i licensee attention. Benchmarking and validation against test data is con tidere.1 j incomplete and minimal. Benchmarks against other codes are limited, some l datmg back 20 years. Additional activities habe been launched upon the ISAT-request.
l e) The CHIC-KIN code, originally developed by W, has been expanded by YAEC and is applied for CEA Ejection and Loss of RCS Flow events. It received good licensee attention and necessary control. Benchmarking to the .
SPERT- test data was performed by W. YAEC benchmarking against other !
l codes, such as HERMITE and STAR, covers comparisons with 3 - D space -
l time kineties codes primarily for CEA Ejection events. No comparisons have j- been performed with experimental loss of flow data.
i l
f) The RETRAN 02 MOD 2 code is a complex system analysis code whic's has l been developed by EPRI for the utility industry. The code is applied worldwide, supported by the RETRAN User Group and results have been widely published in the open literature and international RETRAN-Conferences. YAEC was founding member and since 1981 applies RE' IRAN L for the analysis of the Main Steam Line Break event for licensing purposes.
YAEC has obtained NRC-approval for analysing all non-LOCA SRP chapter i 15 transients with REMN for both PWR and BWR conditions.
The code has received insufficient licensee attention as well as unsatisfactory control in some areas, because of too much reliance on " black box" code -
application and benchmarking / verification efforts by others. A clear understanding of potential safety implications of some thermohydraulic phenomena in important components has not been fully detaonstrated a.4 i incorrect models were applied in the first place, omissions, code errors :md questionable computational results remained undetected. The combinatmn of i the above suggests YAEC is not cognizant of model and result inadequacies of i this RETRAN version. No-in-house RETRAN benchmarkint; against well-
! known separate effects and integral test data existed.
l
e . / . + -i - 'L
- ce u ;.- m -e m :::: . :: . _ ;c ; :_: run .:c
, 3 .
Upon ISAT request and persistance. YAEC launched a substantial effort in ,
parametric sensitivity studies and presented comparisons ofits modeling approach with the 100% MSLB-experimem BT 12 in the LOBI-MOD 2 test facility. For the time being these additional justifications indicate overall conservatively low primary system temperature for reactivity computation and the potential for retum to power. Further validation deems warranted by analyzing the complete LOBI-MOD 2 facility for BT 12 as well as BT C 1 (10% MSLB)
In addition, pressurizer data (MIT, Netherlands, Shippingport) should b: fully evaluated and RETRAN-analyses performed accordingly.
The issue of single core versus split core modeling primarily depends upon thi:
availability of representative mixing data in cold leg, downcomer, lower and upper plena and in the core. Proprietary vendor data for mixing may not be always representative because of scaling issues, thermal boundary cond:tions, j core and plena representations. In order to keep MSLB analysis with sy:: tem l cods scrutinable and traceable, it is suggested to treat mixing as parameter i foi ; . 'mding. Experimental plant data for asymmetric cooldown were measured in Occonee and more detailed data were obtained in various facilities, including the HDR-vessel, examining PTS.
Refined lumped-parameter approach nodalization of the RPV upper hesd region may potentially lead to too high computed recirculation flows in this region which are unrealistic and afTect important primary system quantities. f t. !
is therefore suggested to keep nodalization simple but adequate and
, scrutinable.
j When cold leg temparature is used for reactivity computation, the effect and degree of mixing with the HPI and boron concentration should be accot nted for.
Overconservatism in the thermohydraulic quantities during overcooling event :
may lead to safety concems regarding piping, nozzle and vessel integrity.
e) The CONTEMPT / LT - 028 containment code with input from R 5 M3 YA 91 (YAEC version of RELAP 5 / MOD 3) for LOCA mass and energy releases it YAEC's code of choice for the containment spray pump NPSH analysis For this purpose CONTEMPT was expanded with heat exchanger network capability (see publication 1996 National Heat Transfer Conference, Houston TX). An elaborate iterative cycle between R5M3YA92 and CONTEMPT was used to converge on the containment pressure as input for the NPSH-evaluation. The RELAP-model for MY is extremely sophisticated while the MY-containment was only presented by one single node, lumping everything into one point. In addition, the T-flash option was exercised . A parame ric study, encompassing 15 different cases for break location, size, type and various discharge coefficients and one case for 2440 MWth, was perfonned and the NPSH margin values documented.
4
~,
I d . i . :- H 7:n :c.-ic.-i --G b :::: :: . _ : w: .: _:
. :ME 05 i Y i
These large effons were done to take full credit for the efTect of contair-ment overpressure as MY is considered to be a pre-Reg. Guide I plant.' ,
However, the approach, especially for modeling the containment, must be questioned, because of the tight correlation between computed sump temperature and equivalent NPSH (1 F is equivalent I foot NPSH).This level ;
of accuracy is unachievable by the iterative procedure choseri an'd containment ;
model developed for MY.
The application of a single-node, lumped-parameter approach for the containment including spray and sump, together with "T-Flash" or "P-Flash" ,
option generates final computational results which are not scrutinable. !
Complete mixing with non-condensables is already achieved when using a !
single-node approach The iterative solution for the containment pressure between two, so much '
differing models deems to be inappropriate, may have convergence difT:cultie e and seems to be an " overkill" in view of the overly simplistic containment model approach,' added conservatisms and uncenainties involved. It has been reponed at the 1996 National Heat Transfer Conference, Houston, TX., that two iterations between primary systems and containment codes sutTice for cases where more efTons and focus went into the cantainment modelling with several nodes to cope with the key phenomena. It is therefore suggested to redirect efforts and focus towards the containment modelling as sump, liquid level, sump temperatures and containment atmospheric conditions are the figures of merit for the available NPSH. Whatever containment modelling approach is applied, it is recommended to validate it against sump data obtained from containment experiments in the HDR and the Battelle Model Containment.
The prediction of NPSH should also realistically account for the post-L(X'A-conditions in the sump, e.g. rather than taking credit for dearcated water, i debries panicles from failed insulation, paint and coating particles and t.ludge should be accounted for.
In summary of the technical observation:ron YAEC's analytical capabilities, support and applications it is worthwile mentioning that in areas which are driven by economics YAEC's performance is very good to excellent, whereas system safe analyses are seemingly only performed on a satisfactory bases.
- 3. OBSERVATIONS REGARDING YAEC's PROCEDURES AND SUGGES'110NS Upon the arnval oflSAT, the situation at YAEC with respect to in-house procedures can be summarized as follows:
't. Generic RETRAN conditions werejudged unimportant on the sole basis of engineering judgement; no sensitivity analyses were performed, no documentation existed which quantify thesejudgements. YAEC should formally be requested to establisch a procedure for tracking Generic SER items and restrictions which is scrutinable.
l
- 1 . i . 3 H r: 5: .:.;-ic. :-i : m :::.: ._:c ; ~: t.: P 3E 05 I i
5 l I
l 2 No in house procedure existed to review new code versions and to track _
l disposition of errors and assess potential discrepancies upon licensing analysis by comparing with the previous version
- 3. Lack of a cycle-specific data reference document as a unique and encompassing collection of all input data. (YAEC is developing this capability now).
- 4. Inability to reconstmet decks for older code versions; their model and solutions procedures because of poor decumentation two decades ago, although YAEC relics heavily on the benchmarks performed then in view of a foreseeabic generation change over, it is suggested that YAEC updates the old documentation and Edds all moditications and benchmarkings to it. This is best done in the overall context of a Documentation Management System.
- 5. Deficiencies in benchmarking and validation of computer codes and models against mtegral ar.d separate afTect tests. Comparisons performed thus far l
separated the component from the system behavior by supplying the appropriate l boundary conditions it is suggested that YAFC shows a more proactive attitude '
towards validation efforts with data from scaled experimental facilities and :)lant data when available
- 6. Insights and learning efrects from reactor safety experiments, some of whict even co-sponsored by YAEC, were not used and implemented into current code version ;
- 7. Lack of understanding and documentation of thermalhydraulic key phenomena, parameters and their sensitivities
, 8. Weaknesses in interpreting important code results in the context of underlying model assumptions
- 9. Inconsistent conformance with QA-procedures ,
10.YAEC failed to compley with GL-83-11
- 11. Numerous added conservatisms seem to indicate that the best-estimate valu:s of important quantities are unknown and that the major content of physical and engineering processes gets lost j i
- 12. Existing documentation did not reflect current status, degree of margin 13 Lack of self-assessment in certain, mostly T/H related areas During the assessment period and as result ofISATs detailed questions, YAEC staff and management changed attitude and launched the following activities:
)
l c; .i . m .t ac; +;. -ni-c; H ::.:r . :: . ._0 = "es mc 1 :,
i 6
I a) Initiation of the writeup of a " Methods Overview Manual"(MOM) which lists all -
important key inputs, quality control and safety analysis methods in a startured.
l consistant manner. It also addresses the respective code capability and applicability SER restrictions and conformance are listed.
i b) Additional sensitivity studies and comparisons with data, other codes and other code versions c) Detailed answers to wntten questions. by ISAT which provided additional insights and informations not otherwise documented d) Proactive responses, initiatives and issue resolutions (SG tube repture, LOCA PLHGRs) for which YAEC can be commended
- 4. OBSERVATIONS REGARDING NRC's PROCEDURES AND SUGGESTIQHS The Undings of the ISAT and the review by the peer review team can be summanzed as follows:
- 1. A formal procedure, QA procedures and guidance should be established for the disposition of restrictions / conditions which are imposed on codes, methods, assumptions in SERs tranmitted to licensees.
2 The current lack of follow- up on requirements and requests necessitates th.: )
implementation of a coherent internal procedure j
- 3. Presently, striking inconsistencies exist between plant specific and generic SERs.
~
In the future, NRC should ensure consistancy among SERs for different plants l using the same computer code (and version) for the same transient. A procedure should be established which guarantees that plant-specific SERs always contain the total set of conditions as specified in the generic SERs.
- 4. Specifically for MY, the ISAT findings must be followed up, addressed and efTectively controlled when YAEC uses RETRAN 02 MOD 5 in the future.
- 5. As follow-up on GL-83.11, NRC must make shure that industry carries the full responsibility of proof, quality and adequate resources to assure safety. An enforcement and control mechanism must be in place.
- 6. For transients and non-LOCA events, no rule and no basis for the control of validation exist. However, these must be correctly analyzed in the context o-physical and engeneering processes. The performance of a peer review process is recommended.
l
@ ; 5. ; S H ;": h 001-ici- ni-0:M :::F.:7..:s ; ,ctF
. P.:6E ..
4 F
7 Currently, no structure exists at NRC which would allow to collect all informations about a specific code such, that a unique set of concems could be established. It is recommended to improve intra-agency communication for his purpose and to establish such a document data bank.
8 NRC should focus attention to successful plant cooldown and demonstration of long-term coolability following all kinds of events and accidents.
- 9. At the occasion of MY ISA, it deems appropriate to review the following items-e tracking of restrictions a validation ofcodes for transients E code relicensing issue W NRC review process a SRP update E . quality of reviews (LOCA vs. transients) e quality of communications with industry a completeness of documentation
- 10. After closure of MY issues, the ISAT should document strengths and weaknesses of the exercised assessment process for further and future optmization and application.
- 5. OBSERVATIONS ON ISAT EFFORT AND ACCOMPLISHMENTS Never before has NRC undertaken such an efTort covering such broad spectrum of issues during a brief period of tirne.
ISAT should be commanded for its efforts, because of: .
a) excellent combination ofdifferent skills, long time experience and detailed code and code application knowhow b) focussed approach on CEA drop and MSLB covered the whole spectrum of codes c) Exhaustive list of RFI, questions, requests for sensitivity studies parameter change .
d) uncovering new safety-related open issues (NPSH) e) demonstrating deficencies in YAEC understanding, code application, result interpretation, code benchmarking and validation f) efforts for MY ISA can serve future activitics needed at different sites g) collection ofinformations on current status may be the basis for necessary improvements both in industry and agency.