ML20149L924

From kanterella
Jump to navigation Jump to search
Expresses Appreciation to NRC & Inel Consultants Who Spent 2 Days Answering Questions,Sharing Info Obtained During Reviews,Insights & Work Done in Developing Conclusions. Submits Comments on Areas of Improvement for Yaec
ML20149L924
Person / Time
Site: Maine Yankee
Issue date: 09/19/1996
From: Sullivan H
LOS ALAMOS NATIONAL LABORATORY
To: Lauben G
NRC
Shared Package
ML20148C622 List:
References
NUDOCS 9612040280
Download: ML20149L924 (3)
v • d • e


Text

,.  !

Ed SE: 9-18-86 : 1:28PM ;- N-6 LANL-  ;# 2/ 4 ,

. +

Los Alamos N ATioN AL L ABOR AtoRY l

T 2 :' a Annonemerst Dudenors

. TSA-11, C:- W f and Hazard Analysts, MS K557 1 1.cs Alamos.New Meidco 87546 (505)8574E31, FAX (505)667 52n1  ;

_ - _ _ _ _ - _ l l September'19,1996 )

I l

i l j

Dear Mr. Lauben.  :

t l L l l Independent Review of Yankee Atomic Electric Company Non-Loss-of-Coolant 1

Accident Safety Analysis i

i I would like to start by thanking the Nuclear Regulatory Commission (NRC) and Idaho National i

! Engineering Laboratory (INEL) consultants who were with us for 2 days answering, questions; j i sharing information obtained during their mviews, their insights, and work they had donc in j l ~ developing their concInsions: pmviding suggestions; and guiding us through a large amount of I

!- . documentation and technical issues. I was appreciative of their candor, technical competence, I l and ability to identify major issues, which guided me to a good understanding of the overall Issues. .

l

! - It was difficult to keep from focusing solely on the interesting technical issues that the staff i presented and, instead, to keep focused on determining root causes and how processes can be

, improved so that the Yankcc Atomic Electric Company (YAEC) reviews would not have beca i necessary. The staff did find some issues that need to be addressed in YAEC and at the NRC, j and my comments on these are given below. I have divided my comments into thme areas: areas of improvement for YAEC, areas of impruvement for NRC, and conclusions.

l i

! AREAS OF IMPROVEMENT FOR YAEC i . Several improvements were identified during ISAT reviews of non-loss-of-coolant accident

! (non-LOCA) transients. Validating codes using experimental data to ensure that models and i nad*Hntions can adequately represent the nuclear plant and ensuring that restrictions are properly 4 implemented in code application are the major areas ofimprovement.

Formal validation of codes related to areas of application The lack of a formai validation effort appears to be a probicm in the YAEC analyticel effort for non-LOCA transients. I believe that it is necessary to validate code calculations by comparing them with experimental separate-effects and integral-system experiments before using a thermal-hydraulic code and associated nodalization in modeling reactor system components. A great deal of care is required in selecting the number of nodes for a reactor system, and the nodalization should be vahdated using experimental data that provide valuable mformation related to the analytical methodx. By performing these validations, the personnel obtain a better undentanding of code models and the nodalization requirements of applying the code to a plant tmnsient.

-Understanding of code models and use in applications ,

h appears that YAEC c ional knowledge of computer pode analytical models in order to co nm a rma h

? ydraulic codes. Details of the models in the

( p ogordD N .

iEy B'i: 9-16-36 : 1:2 BPM . .\-6 LtNL-  :# 3/ 4 2-r code have caused erros and inconsistencies in the pressurizer and upper head calculations. It is suggested that YAEC obtain additional expertise to suppoit them in this area.

NRC restrictions implemented in a formal process A number of restrictions have been placed on the YAEC computer codes used in the licensing of Maine Yankee. There appears to be no fonnal process for responding to these limitations or ensurmg that restrictions are meluded pmperly m their calculations. In the steady-state fuel i perfonnance code FROSSTEY, the restrictmn on burnup was 60,000 mwd /MTU and is I controlled through a technical specification. In the thermal-hydraulic code area YAEC said that only steam was removed form the secondary side of the steam generators, whereas, in actuality, small amounts of liquid are released. YAEC needs to have a formal process to resolve how restrictions are enforced.

Maintain consistency in code applications l

YAEC's technical capabilities are not consistent in the technical arcas required for non-LOCA l transients. The quality of their thermal-hydraulic code applications and validations needs to be improved. YAEC's documentation did not support the conclusions that were developed, and the conclusions werejustified only after IS AT requested sensitivity studies. Often subdc changes in

options in anal cal models and/or changes in n<xlalization affect the results. Application I pctsonnel n to be aware of these subtle differences when modeling complex systems.

AREAS OF IMPROVEMENT FOR NRC l I have noted six areas where NRC can improve their systems that may help prevent this situation l fnnn happening in the future. These six areas were discussed in the meetmg. Some are l relatively easy to implement in an action plan; others will require changes to documentation such as the Standard Review Plan (SRP) and may take longer.

l Code restrictions required by SERs not tracked It appears fmm the discussions in the meetings that several SER code restrictions were identified in NRC reviews. These apparently were not closed out with YAEC or tracked within the NRC systems related to code reviews. Because the restrictions were not tracked, they were not closed i out in later reviews.

[

Code validation not required for applications of laterest ,

NRC should consider requiring the vendors and utilities to validate their ccxles using cxperimental data (mm separate-effects and integral systems experiments. This was suggested to the code users after the Three-Mile island accident in a memo from NRC. It is suggested that NRC consider developing requirements and ensuring that a validation process is used.

NRC revtew process not complete and documentation needs to be upgraded Fmm the limited documentation given for the RETRAN SERs, it is difficult to determine the extent of the revicw that was wrormed and how the 39 restrictions were developed in September 1984. It is also difficult (o determine why no restrictions were included in the SER in l October 198.5. Including specific documentation requirements in a m(xlified SRP would limit j this issue.

NRC has a review process that develops a SER containing important restrictions These restrictions are developed and documented. In the RETRAN review process of September 1984, 39 restrictions were issued; in the October 1985 RETRAN review, no restrictions applied.

l

iSENT BY: 9-16-96 : 1:26B1 : N-6 LeBL-  :- 4/ 4 4

e A number of crrors and inconsistencies were identified in the ISAT review of RETRAN. Based

, on the discussion, it appears that the mot causes are not covered in enough detail in the SER and the reviews are not fully documented. It is suggested that the review quality needs to be improved

by improving the SRP and confonning to the updated SRP. It also appears that management reviews did not find the problems and independent reviews had similar results. The updated SRP review needs to include details on what phenomena are imponant, the depth of reviews needed, j

and what guidance is needed. An EPRI review plan was availabic, and a quick review of their l process indicated it was more rigorous because EPRIincludes discussion of topics such as l assumptions made, modeling phenomena, safety systems to be considered, use of analyscs, and useful suggestions. Tids document could be the starting point for a rewrite of the SRP.

NRC documentation did not support the conclusions Before the IS AT review, the documentation at both the vendors and within NRC did not support the modeling assessments or the degree to which the models predicted the phenomena associated with the transients. The ISAT staff requested several sensitivity studies before the conservatism could be determined. As noted earlier, several enors and inconsistencies were identified.

Computer code licensing is a disincentive for both vendors and NRC New computer codes are becoming available from a number of sources. These codes have improved abilities to predict important phenomena in non-LOCA transients. Because of the extensive reviews for the curn:nt code licensing process, new and improved codes tend not to be considered when an analysis tool has been approved, which can be a disadvantage for the vendor, the utilitics, and the NRC. This may become a more important issue as conservatism is reduced. NRC should mexamine its procedures and requirements for code licensing, try to establish a more efficient process, and try to mduce appmval time. Vendors and utilities need to have an incentive to improve license codes that could lead to improved safety.

CONCLUSIONS YAEC and NRC have an o additional confidencemthat, pportunity the future, non-LOCA to improve transients methods are fully justified and and review pmccsses to provide representative. I believe, as do most of the independent reviewers and NRC staff, that the YAEC power plant is safe to continue runting at the cunent power limits. The problems identified arc (1) fully demonstrating the degree of conservatism, (2) the lack of code modeling validation to demonstrate that nodali7.ation is adequate for the models, and (3) improving consistency in modeling related to non-LOCA transients. Both YAEC and NRC need to consider how to improve quality in both applications and reviews and how to develop new ways to reduce time and funding for getting codes approved or new analytical methods into place. -

If you have any questions about my comments, please feel free to contact me at (505) 667-6231 (phone), (505) 667-5531 (fax), or hsullivan@lant. gov (e-mail).

Sincerely, Harold Sullivan LHS/mct

._

Retrieved from "https://kanterella.com/w/index.php?title=ML20149L924&oldid=2973608"