ML092590242
| ML092590242 | |
| Person / Time | |
|---|---|
| Site: | Turkey Point  |
| Issue date: | 09/11/2009 |
| From: | Scholl P Sargent & Lundy |
| To: | Office of Nuclear Reactor Regulation |
| References | |
| 45343 | |
| Download: ML092590242 (5) | |
Text
09/111/2009 U.S. Nuclear Regulatory Commission Operations Center Event Report Pajee I General Information or Other (PAR)
Event#
45343 Rep Org: SARGENT & LUNDY LLC Notification Date / Time: 09/11/2009 16:00 (EDT)
Supplier: SARGENT & LUNDY LLC Event Date I Time: 09/11/2009 (CDT)
Last Modification: 09/11/2009 Region:
3 Docket #:
City: CHICAGO Agreement State:
Yes County:
License #:
State: IL NRC Notified by: PAULA SCHOLL Notifications:
DAVID PROULX R4DO HQ Ops Officer: PETE SNYDER ALAN BLAMEY R2DO Emergency Class: NON EMERGENCY OMID TABATABAI NRO 10 CFR Section:
PART 21 GROUP NRR 21.21 UNSPECIFIED PARAGRAPH PROBLEM WITH SEISMIC SOIL STRUCTURE ANALYSIS SOFTWARE "ACS SASSI (Sargent & Lundy Program No. 03.7.848-2.21) by Ghiocel Predictive Technologies Inc. (GPT) is a program which performs seismic soil-structure interaction (SSI) analysis. Sargent & Lundy (S&L) purchased licenses to use ACS SASSI NQA Version 2.2.1 from GPT in 2008 and verified and validated the software under S&L's QA program.
"S&L used ACS SASSI for seismic analysis activities for South Texas Project Units 3 & 4 (STP 3&4).
"On July 14, 2009, during SSI analysis of the Control Building performed to address the STP 3&4 Shear Wave Velocity Departure, irregularities were observed in the results obtained using ACS SASSI and the analysis was halted. The irregularities in turn led us to question the results for the Ultimate Heat Sink / Reactor Service Water Pump House (UHS/RSWPH) completed earlier.
"Our investigations, including a review of other uses of the software, indicate that numerical instability may occur with high numbers of soil layers, even when soil layer properties and number of soil layers (less than or equal to 100) are within the parameters stated in the User's Manual. Our investigations indicate that the results for the Control Building with 94 layers, the Reactor Building with 84 layers, and the UHS/RSWPH with 100 layers are not reliable. Some individual results are conservative, others are not.
"S&L recommended to the South Texas Project Nuclear Operating Company (STPNOC) that the work performed using ACS SASSI NQA Version 2.2.1 relied upon to support licensing positions and to support the design basis be reanalyzed using SASSI 2000, Version 3.0. STPNOC accepted this recommendation.
"S&L is no longer using ACS SASSI for production work. In addition, we have investigated whether SASSI 2000
09/11/2009 U.S. Nuclear Regulatory Commission Operations Center Event Report Pae 2 General Information or Other (PAR)
Event#
45343 has similar irregularities. We concluded that SASSI 2000 does not have this condition and is fundamentally sound."
Received at: 09/11/2009 14:19 O-1 1-09 01:23pm From-3122692453 T-954 P.01/03 F-001 Sa'te~f
,I L.andy L=
PAUIA t.. SCOHOL Senpr V"eg Prorident (312) 269-6645 pavwiQ i.$choiI@sargentiuudy corn September 11, 2009 10 CFR 21 Report ACS SASSI Computer Program Document Control Desk United Sates Regulatory Commission Washington, 0. C. 20555-0001
Dear Sir:
This report is being macle in accordance with 10 CFR 21.21 (d)(1),
ACS SASSI (Sargent & Luncly Program No. 03.7.845-2.21) 0y Ghiocel Predictive Technologies Inc. (GPT Inc-) is a program which performs seismic soil-structure interaction (551) analysis.
Sargent & Lundy (S&L) purchasecl licenses to use ACS SASSI NQA Version 2.2.1 from GPT Inc. in 2008 and verified and validated the software under S&L's QA program.
Sargent & Lundy used ACS SASSI for seismic analysis activities for South Texas Project Units 3&4 (STP 3&4).
On July 14, 2009, during SSI analysis of the Control Building performed to address the STP 3&4
$near Wave Velocity Departure, irregularities were observed in the results obtained using ACS SASSI and the analysis was halted. The irregularities in turn led us to qLestion the results for the Ultimate H-eat Sink/Reactor Service Water Pump House (UHSIRSW Pump House) completed earlier Our investigations, including a review of oter uses of the software, indicate that numerical instalility may occur with nigh numbers of soil layers, even when soil layer properties and number of soil layers ($100) are within the parameters statecl in the User's Manual. Our investigations indicate that the results for the Control Building with 94 layers, the Reactor Building with 84 layers and the UHS/RSW Pump House with 100 layers are not reliable. Some individual results are conservative, others are not.
55 East Monroe Street - Cficago, IL 60603-5780 UA - 312-269-2000
09-11-09 01:23pm From-3122692453 T-954 P.02/03 F-001 September 11, 2009 Page 2 of 3 United Sates Regulatory Commission ACS SASSI Computer Program Sargent & Lundy has used ACS SASSI only for STP 3&4. ACS SASSI was used in the applications shown in the following Table:
SSI Analysis Program Purpose Status 1, Preliminary seismic analysis ACS Study This work was clone as a of the Reactor Building.
SASSI study. It is not relied upon for any downstream analysis or design.
- 2. BenchmarKing analyses in ACS Licensing Reanalyze with SASSI 2000 support-of seismic models SASSI Support (due September 30. 2009)
- 3. SS analysis of the Control SASSI Licensing Complete - ACS SASSI was Building to address the STP 2000 Support not used after suspected 3&4 Shear Wave Velocity error - results were based Departure on SASSI 2000
- 4. $M1 analysis of the Reactor ACS Licensing Reanalyze with SASSI 2000 5uiijirng to acsaress me STP
$ASSI Support (due September 30, 2009) 3&4 Shear Wave Velocity Departure
- 5. 5SI analyses of the ACS Design Reanalyze with SASSI 2000 UHS/RSW Pump House SASSI Basis &
(due November 24, 2009)
Licensing Support Sargent & Lundy recommended to the South Texas Project Nuclear Operating Company (STPNOC) that worK performed using ACS SASSI NQA Version 2.2.1 relied upon to support licensing positions and to support the design basis be reanalyzed using SASSI 2000, Version 3.0. STPNOC accepted this recommendation. The schedule for completion of the reanalysis is noted in the Table above.
We believe that this non-compliance is reportable under the provisions of 1 OCFR21. On March 16, 2009, S&L issued safety-related Calculation No. U7-UHS-C-CALC-6001 titled "Soil-Structure Interaction Analysis of the UHS-Pump House Buildings" for STP 34. This calculation utilized the ACS SASSI NQA Version 2.2.1. It provides the design basis seismic parameters used for overall stability evaluations, confirmation of member sizing, design of concrete reinforcement and generation of response spectra for the UHSIRSW Pump House. The S51 reanalysis (Item 5 above) will be used to revise calculations used for design of the UHS/RSW Pump House. These calculations are expected to be complete by December 31, 2009.
Further, comparison of results using ACS SASSI and comparable software, i.e., SASSI 2000, indicate that the ACS SASSI results are not conservative in all cases. Therefore, the safety hazard that could be created using Calculation No. U7-UHS-C-CALC-6001 or ACS SASSI is the design of insufficient reinforcement.
,Sargent & Lundy is no longer using ACS SASSI for procluction worK. In addition, we have investigated whether SASSI 2000 has similar irregularities. We concluded that SASSI 2000 does not have this condition and is fundamentally sound.
For your information, we have discussed this issue directly with Dr. Dan Ghiocel, President &
Chief of Engineering, GPT Inc. They issued a software error notice letter on this issue dated Septsmoer 10, 2009.
09-11-09 01:23pm From-3122692453 T-954 P.03/03 F-001 United Sates Regulatory Commission September 11, 2009 ACS SASS1 Computer Program Page 3 of 3 If you have any questions or require further clarification on this issue, please feel free to contact me at the above listed number or our QA Director, Randall Kurtz, at (312) 269-6562.
Yours very truly, Paula L. Scholl Director of Engineering