ML20008F104
| ML20008F104 | |
| Person / Time | |
|---|---|
| Site: | Davis Besse, Oconee, Arkansas Nuclear, Crystal River, Rancho Seco, Crane  |
| Issue date: | 03/02/1981 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20008F101 | List: |
| References | |
| NUDOCS 8103120270 | |
| Download: ML20008F104 (2) | |
Text
.
^
fitCy d
I%.
UNITED STATES
(('
.( ',
NUCLEAR REGULATORY COMMISSION W ASHINGTON,0. C. 20555 j
g
/
i
?
%, *.. ".. p SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION CONCERNING ITEM II.K.2.19 " BENCHMARK ANALYSIS OF SEQUENTIAL AUXILIARY FEEDWATER FLV."
FOR BABC0CK & WILC0X REACTOR PLANTS DOCKETS NOS. 50-269, 50-270, 50-287, 50-289, 50-302, 50-312, 50-313 AND 50-T,46 Introduction 26, 1979, with the owners of Babcock and At a aceting) in Bethesda, Aprilreactor plants, we requested a benchnark analysis of seque Wilcox (B&W tial auxiliary feedwater flow to the stean generators following a loss of main feedwater.
This analysis was provided in a letter from J. Taylor (B&W) to R. flattson (NRC) dated June 15, 1979.
However, in this analysis the TRAP-2 Code with 6 node steam generator model was utilized. All small break analysis presented to the NRC have been performed using the CRAFT-2 Code with a 3 node steam generator model. We require a benchmark analysis for sequential auxiliary feedwater flow also be performed using CRAFT-2 with a 3 mode steam generator representation. By letter dated August 21, 1979 we requested such analysis.
Each licensee of D&W reactor plants responded with a report which presented analysis of sequential auxiliary I
feedwater flow to the steam generators for a loss of main feedwater trans-ients using the CRAFT-2 Code.
This issue was later identified as Item II.K.2.19 of the TMI Action Plan req ui rements.
Discussion & Conclusions B&M utilizes the CRAFT-2 computer progran in performing loss of coolant accident (LOCA) licensing evaluations for their nuclear stean supply systems (NSSS).
Subsequent to the T!11-2 accident, this computer program was used to confirm energency operator guidelines for all power plants with NSSSs designed by B&W. Our review of these confirmatory analyses have led to questions re-garding the ability of the CRAFT-2 program to adequately predict steam generator performance and its influence on the primary system thermal-hydraulic behavior.
In particular, we noted that the CPAFT-2 steam generator nodel did not contain the same degree of detail as the model used j
with_the TRAP-2 Code.
TRAP-2 is a computer code primarily used for non-l LOCA transients by D&W.
In order to validate the TRAP-2 transient code with actual plant data, an asymmetric cr sidown test was incorporated into the Crystal River Unit 3 power ascensicc. program.
Because of the simplified steam generator model in the CRAFT-2 Code, we also requested that the CRAFT-2 Code be assessed against the Crystal River Unit 3 asymmetric cooldown data.
l 8103120 M b-7 y-
-.- r y
--.p
-.--..v.
The comparative analyses of the startup test demonstrated that the simplified steam generator model used in the licensing code (CRAFT-2) predicted thennal-hydraulic behavior similar to the more detailed steam generator model utilized in the TRAP-2 Code.
However, comparisons with data for both codes were poor.
Further examination of the Crystal River Unit 3 asyrnetric startup test has indicated the test to be inappropriate for assessing com-puter codes.
This is attributed to inadequate instrunentation whereby key data required for code assessment were not obtained.
Reviews conducted by our B&O Task Force, follaving the TMI-2 accident, have concluded that further assessnent of the CRAFT-2 Code would be required.
The majority of the concerns identified are documented in NUREG-0565.
In particular, the neglect of a nechanistic, regine-dependent heat transfer model and the use of a corstant, stean generator heat transfer coefficient throughout the transient have been identified as requiring either revision or further justification.
This requirement for further justification and/or revision of the small break ECCS models is being performed under TMI Action Plan Iten II.K.3.30.
We believe that satisfactory resolution of code nodeling concerns as part of the Action Iten II.K.3.30 will :esolve the modeling concerns of II.K.2.19.
i The conclusions of our review of Action Item II.K.2.19 are as follows:
(a)
The intent of Item II.K.2.19 was accomplished, (b)
Results provided by CRAFT-2 were similar to those provided by the more detailed ' RAP-2 program.
However, both codes showed poor agreement when compared with the test data, (c)
The poor agreenent of the code prediction with test data has been attributed to the fact that the Crystal River ascension test data was not adequate for assessing thernal-hydraulic codes, and (d) A more rigorous assessnent of the B&W small break LOCA nodel is being performed under TMI Action item II.K.3.30.
Further code assessnent under TMI Action Item II.K.2.19 is therefore unnecessary.
Based on the above conclusions, we consider Item II.K.2.19 conpleted by all licensees with B&W NSSSs by issuance of this Safety Evaluation Report.
Moreover, we do not believe it necessary for Item II.K.2.19 to be addressed any further.
Dated: March 2,1981 7
-