ML20246Q175
| ML20246Q175 | |
| Person / Time | |
|---|---|
| Site: | McGuire, Mcguire  |
| Issue date: | 05/01/1989 |
| From: | Tucker H DUKE POWER CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| TASK-2.D.1, TASK-TM TAC-44593, TAC-54601, NUDOCS 8905220402 | |
| Download: ML20246Q175 (5) | |
Text
.. _ _ _ _
i DUKE PowEn Gomwwy r
P.O. DOX 33189 CIIAHLOTTE, N,0. 28242 IIALD. TUCKER Tet.rrsimrz
, vecs enewspeart (7o4) 373-4531 wuos.m.m ruoovsmon May 1. 1959 U. S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, D.C.
20555
Subject:
McGuire Nuclear Station Docket Nos. 50-369 and 50-370 NUREG-0737 Item II.D.1, Performance Testing of Relief and Safety Valves NRC Review / Technical Evaluation Report - Response (TACS 44593 and 54601)
Gentlemen:
Mr. D. B. Matthew's (NRC/0NRR) February 2, 1989 letter transmitted a Technical Evaluation Report (TER) reflecting the present status of the NRC's review of the McGuire Nuclear Station response to NUREG-0737 Item II.D.1, " Performance Testing of Relief and Safety Valves." My letter of March 13, 1989 submitted a response-I to the deficiencies noted in the TER (i.e. with respect to Item 8 and part of Ttem 7 of NUREG-0737 Item II.D.1), indicating that additional information would be submitted by May 1 (for Item 8a) and December 1, 1989 (for Items 7 and 8b).
Accordingly, please find attached additional information regarding Item Ba.
Note that during the process of preparing this additional information, Duke's contracted vendor (Impe11 Corporation, Inc.) identified an error in the SUUGGER computer program used for the analysis. Duke was verbally notified of the possibility of this error on April 20, 1989, and a written confirmation was provided on April 26, 1989. Thus, the Duke structural calculation results previously reported are incorrect. Duke has initiated a Problem Investigation Report to formally address McGuire support / relief valve piping system operability, although our initial assessment is that the system is operable in its current configuration. The NRC Resident Inspector will be advised of the situatioa and informed of the results. Further details on this matter are contained in the attachment.
As a result of the above, the verification information regarding the updated SUUGGER program will be revised and provided by June 16, 1989. The Items 7 and 8b additional information will continue to be supplied by December 1, 1989.
]
3 4
p gp2ggg g;gggp o
e
Document Control D:sk
.L a-Page Two May 1, 1989 Should there be any questions concerning this response or if further information is desired, contact Bruce Nardoci at (704) 373-7432.
Very truly yours,
/
d6/
43.<__
Hal B. Tucker PBN166/lcs Attachment cc: W/ Attachment Mr. S. D. Ebneter Regional Administrator U. S. Nuclear Regulatory Commission Region II 101 Marietta Street, NW, Suite 2900 Atlanta, Georgia 30323 Mr. D. S. Hood, Project Manager Division of Licensing Office of Nuclear Reactor Regulation U. S. Nuclear Regulatory Commission Washington, D.C.
20555 Mr. P. K. VanDoorn NRC Resident Inspector McGuire Nuclear Station l
i Duke Power Company McGuire Nuclear Station Response to NRC Concerning Thermal Hydraulic j
Questions for NUREG-0737
)
As noted in our letter dated March 13, 1989, Duke Power l
contracted with its vendor, Impell Corporation, Inc.
I (formerly EDS Nuclear, Inc.) to review the analysis of record for a response to NRC's question item 8a raised in David B. Matthews' letter to H.B. Tucker of 2/2/89.
Impell is the vendor Duke selected to perform the original analysis and subsequent reviews.
Item 8a of the Technical Evaluation Report attached to NRC's letter referenced above, deals with verification of Impell's computer code EDSFLOW which was used in Duke's thermal hydraulic analysis.
Impell was requested to provide additional information on the computer code to Duke to assist in-developing a response to NRC by May 1, 1989.
The following information is offered to describe what has I
been learned from that review.
EDSFLOW is a modified version of the Idaho National Engineering Laboratory RELAP4/ MOD 5 computer code.
For applications in the Support / Relief Valve (S/RV) discharge transients, the code is equivalent to RELAP4/ MOD 5 Version 4.601.
Initial comparisons of EDSFLOW (RELAP4/ MODS) against the EPRI test results showed that EDSFLOW underpredicted transient fluid forces immediately downstream of the safety and relief valves.
A comprehensive effort was undertaken to conservatively supplement the EDSFLOW generated force-time history in those piping regions which showed a non-conservative benchmark against the EPRI tests.
The 3
following describes the methodology employed to ensure that the force-time history used in the analysis bounds the expected plant transients.
The release of data from the EPRI/C-E tests in 1982 initiated a review of the EDSFLOW (i.e. - RELAP4/ MOD 5) model used for the safety and relief valve piping analysis.
concern was centered around the cold loop seal tests of the Crosby safety valves.
These tests measured very high peak j
~
loads in the early pait of the transient persumably caused by the discharge loop seal water in the discharge piping.
The basis of concern was the application of EDSFLOW's j
homogeneous equilibrium model to represent discharge of the cold water loop seal.
An effort was undertaken to benchmark EDSFLOW against the EPRI cold water loop seal tests.
]
J l
i The EPRI/C-E tests clearly showed the highest peak loads for the case of steam discharge through the Crosby 6M6 safety valves when a subcooled water loop seal was used.
The tests pertaining to this situation were C-E tests 908 and 1406.
The test results stated that these two tests were conducted under identical conditions except that the 908 was performed while an orifice plate was installed at the end of the test rig piping.
The test report recommends use of test 908 results rather than test 1406 because they contained more data.
Also, the resultant forces from test 908 are slightly i
higher than for test 1406.
Otherwise, the results of the two tests were very similar.
An EDSFLOW model of the EPRI/C-E test rig was constructed for benchmarking against the test results.
The EDSFLOW model for the test rig was constructed using the same assumptions that were used in the original analysis performed for the McGuire S/RV piping system.
An effort was made to match key parameters such as node sizing and time steps to the McGuire plant model as well.
The results of EPRI/C-E test analysis showed that EDSFLOW produced a poor benchmark for forces near the safety valve outlet where the loop seal water was presumed to be discharged as a solid " slug" of water.
However, EDSFLOW did produce conservative results downstream where the water density has been decreased through mixing with steam and air.
This mixing in the vnstream port.'an of pipe is more accurately modeled by EDSFLOW's homogeneous model.
In response to the EDSFLOW benchmarking results, an ans. lysis technique was developed specifically to predict the peck loads caused by loop seal water discharge as measured in the EPRI/C-E tests.
This technique was implemented in a computer program called SLUGGER.
The SLUGGER program was benchmarked against the results of C-E tests 908 and 1406.
A successful benchmark was achieved thereby validating I
assumptions made about the dominant physical phenomena that i
caused the peak loads measured in the test.
The same I
technique was then applied to the safety valve discharge l
analysis for McGuire.
During the process of preparing this response to your questions, Impell's engineers identified an error in the SLUGGER run of record.
The error was detected duri_ng a I
review of program logic when it was discovered that the i
program did not properly calculate steam mass downstream of expanders (pipe area changes).
This has the effect of l
reducing the driving steam pressure for the loop seal water l
slug, therefore, reducing slug velocities and momentum l
forces.
1 w _ __-
For piping configurations in which the expander is located well away from the safety valve, the error has little effect, since the slug reaches high velocities prior to reaching the expander.
This is the case for the C-E test configuration.
However, for the McGuire configuration, the expanders are located close to the safety valves.
Preliminary results studies using a corrected version of the
" SLUGGER" computer program have yielded higher than previously analyzed force-time history profiles for several locations in the McGuire configuration.
Thus the Duke structural calculation results previously reported are in error.
Impell-verbally notified Duke of the possibility of this circumstance on 4/20/89 and a written confirmation was provided on 4/26/89.
Duke has looked at the preliminary load changes and our initial assessment is that the system is operable in it current configuration.
Duke Power has initiated a Problem Investigation Report which is formally addressing operability and has also
-requested Impell to provide the revised force-time histories based on their review of this situation.
As part of the PIR process, Duke is reviewing the impact of these revised force-time histories on the structural analysis.
The NRC Resident Inspector will be notified of the events described above and informed of the results.
As a result of the above, the verification information which correlates the CE/EPRI test data against the updated SLUGGER program will be revised and provided by June 16, 1989.
1 1
1
_ _ _ _ _ - _ _ - _ _ _ _