ML20055B054
| ML20055B054 | |
| Person / Time | |
|---|---|
| Site: | Yankee Rowe |
| Issue date: | 07/13/1982 |
| From: | Denise Edwards YANKEE ATOMIC ELECTRIC CO. |
| To: | Mattson R Office of Nuclear Reactor Regulation |
| References | |
| FYC-82-12, NUDOCS 8207200301 | |
| Download: ML20055B054 (2) | |
Text
1 4
Tecphons (617) 872-8100 TWX 7103807619 YANKEE ATOMIC ELECTRIC COMPANY 2.C.2.1
~*
j[
FYC 82-12 M*L r a 1671 Worcester Road, Framingham, Massachusetts 01701
,YAN KEE,
~~
July 13, 1982 Dr. Roger J. Mattson, Director Division of Systems Integration United States Nuclear Regulatory Commission Washington, D. C.
20555
Reference:
YAEC Submittal Letter to USAEC Directorate of Licensing, License DPR-3, Docket 50-29
Subject:
ATWS Model Verification Using LOFT Test L9-4
Dear Dr. Mattson:
Yankee Atomic Electric Company appreciates the offer in your letter of May 10, in which you advised that a forthcoming LOFT Test (L9-4) "...will be an anticipated transient without scram test intended to qualify models used to develop emergency procedure guidelines for ATWS events." We regret that we were unable to provide a more prompt response to your request. After careful consideration of the merits of the proposed post-test verification program, we must decifne to participate. We discuss our decision in the following paragraphs.
Existing Analytical Capability--Impending, But Uncertain, ATWS Requirements We developed ATWS computer models in 1974, in accordance with WASil-1270 guidelines. They are described in the Reference, which also submitted results of an ATWS evaluation for our PWR plant in Fowe, Massachusetts.
Since that i
time, Yankee Atomic has continued to pursue st' ate-of-the-art capabilities, I
including RETRAN and RELAP models, for analyzing nuclear reactor transients and accidents. This capability is being continually exercised to demonstrate l
compliance with existing regulations for our power plant in Rowe, and for l
other plants in the Northeast, including Vermont Yankee, Maine Yankee, and Seabrook.
We are mindful that the final ATVS rule may contain qualification l
requirements for licensees' ATWS evaluation models.
In view of the uncertain l
regulatory requirements for ATWS evaluations, however, we believe that to commit additional resources to revising and verifying our currently-licensed computer rodels would be premature and imprudent.
()cy[
Our ATWS Position, Restated In our comments on the three proposed ATWE rules, we endorsed the Utility Group's proposal as the superior alternative.
The Utility Group consists of twenty-two companies who, unlike the NSSS vendors or NRC Staff, actually own and operate nuclear power plants, with B&W, 9207200301 820713 DR ADOCK 05000029 p
1 Dr. Roger J. Mattson July 13, 1982 Page 2 CE, CE, and Westinghouse reactors.
In sum, the group's commentary on ATWS rulemaking emphasized the opinion:
(1) that ATWS risks are extremely low; (2) that Industry's existing analytical capabilities are adequate for predicting ATWS consequences; and, finally, (3) that no favorable cost-benefit relationship exists to support new code development for plant-specific analysis of all conceivable ATWS conditions.
Questionable Value of "Elind" Post-Test Analysis We agree the LOFT facility provides a credible source for data for computer code verification.
However, we do not believe the process of code verification should necessarily involve " blind" predictions.
The adequacy of computer code modeling can be equally demonstrated, by comparisons to data independent of the timing of a specific test or experiment.
Knowledge of LOFT's transient behavic', during the test, does not limit the analyst's responsibility to justify specific modeling approaches.
Post-test knowledge of actual transient behavior, however, can significantly reduce the resources required to perform data comparisons with computer code simulations. This is particularly true where the timing and characteristics of control systems, valve action, and operator action are crucial in affecting the overall system response.
In these cases, " blind" test predictions become more of a hit-or-miss proposition, with regard to how the control systems and valves will function during the test, rather than verifying the various thermal-hydraulic models employed in the prediction.
Poor comparisons of " blind" test predictions to LOFT operational transients in the past have not resulted from the inability of system codes to predict the transient thermal-hydraulics, but have resulted from the inability to know in advance the timing and characteristics of systems that influenced the transient betavlor. This fact has been proven in post-test comparisons, when the actual behavior of key systems was factored into the analysis.
Concluding Remarks In summary, the thrust of our decision not to voluntee'r for a " blind" ATWS Model Verification Program is that it would not be the best possible use of our resources at present. This is especially the case when the uncertainty of impending rulemaking, at best, makes the future requirements for ATWS computer models questionable.
Very truly yours, YANKEE ATOMIC ELECTRIC COMPANY D. W. Edwards, Directt r Operational Projects and Licensing cc:
W. F. Conway P. A. Bergeron R. E. Helfrich P. C. Higgins (AJI)
T. J. Tipton (AIF) l