ML17333A014

Responds to IE Bulletin 79-07, Seismic Stress Analysis of Safety-Related Piping. Eleven balance-of-plant Piping Sys Analyzed Using Algebraic Summation & Found within Code Limits for Pipe Stres Allowables

ML17333A014
Person / Time
Site:	Columbia
Issue date:	08/08/1979
From:	Renberger D WASHINGTON PUBLIC POWER SUPPLY SYSTEM
To:	Engelken R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION V)
References
NUDOCS 7910030789
Download: ML17333A014 (5)

Text

p,~

Washington Public Power Supply System A JOINT OPERATING AGENCY P. O. nOX 9ee 3000 GEO, WASHINCZON WAY RICHLAND, WASHINCTON 99352 I NON+ (509'75'5000 Docket No. 50-397 August 8, 1979 c 'l('J" ?jgg<,

Nuclear Regulatory Commission Suite 202, Walnut Creek Plaza 1990 N .Cal.ifornia Boulevard

~

Walnut Creek, CA 94996 I7 Attention: Mr. R. H. Engelken, Director

Subject:

WPPSS NUCLEAR PROJECT NO. 2 IE BULLETIN NO. 79-07

Dear Mr. Engelken:

In response to the subject bulletin, WPPSS submits the following infor-mation for its project WNP-2..

It has been determined that, eleven WNP-2 balance-of-plant piping systems were analyzed utilizing algebraic summation but have subsequently been re-analyzed and have been found to be within code limits for pipe stress allowables and all supports and supplemental structures are within current rated load capacity.'See Attachment I)

With respect to GE, the NSSS supplier, review of the IE Bulletin resulted in the determination that none of the computer codes used f'r the seismic analysis of GE piping systems important to safety employed the techniques identified in Item 1 of subject bulletin. With regard to Item 3, GE will use SAP/PISYS computer programs for the seismic piping analysis. A description of these programs and the verification procedure is presented in Attachment II.

Should you have any questions regarding this response, please contact me.

Very truly yours,

~6~p, D. L. RENBERGER Assistant Director-Technology DLR:FJ:ct Attachment po/S

t-Attachment I (Bombers refer to IE Bulletin item number~.)

1 ) A version of the ADLPIPE Computer Program which includes an algebraic summation technique, has been used to do seismic analysis of safety-related piping.

The items listed under Item 1) in the subject bulletin do not describe explicitly the summation techniques used in ADLPIPE for determining final mathematical model responses.

This version of ADLPIPE Computer Program uses the intramodel algebraic summation for co-directiona1 components resultinq from multiple earthquake directional motions. The Square Root of the Sum of Squares summation is used to determine the combined response for all modes to obtain directional forces at a point in the pipe system. Closely spaced modes (within 10K) are absolutely sunned.

2) The ADLPipe Computer Program is a proprietary program owned and technically supported by:

Arthur D. Little, Inc.

Acorn Park Cambridge, HA 02140 A <<h The Computer. facilities, where Burns and Roe engineers utilize ADLPIPE Computer 'Program",'ave been, directed to forward to the U.S. HRC.,'1istings as called for in Item (2) of the bulletin."

e This direction specifically asks that this listing contain Old Versions of ADLPIPE from 1972 up to the present version.

3) For operational piping analyses, Burns and Roe has exclusively used Arthur D. Little piping analysis program, ADLPIPE. This computer program was the ~onl program in the public domain which considered the nuclear piping'requirements for the Code, ANSI-B31.7.

In the year 1972, an updated version enabled the user to produce a partial stress report required to meet the requirements of the BGPV AStlE Section III, Division 1 Code, as well as ANSI B31.1 Code for piping and components. It also contained, for the first time, a spectra modal response analysis for redundant pipe systems.

Burns and Roe also has an ongoing validation program. Typical spectra response and time history analyses from the ADLPIPE Program have .been compared with the results from AHSYS and STARDYHE Computer Programs Hinor differences, due to different numerical techniques used within the programs, have been identified but verification has been demonstrated by these comparisons.

In addition, the fol'Iowing comparisons have been made of the different summation options of ADLPIPE against other recognized industry methods and benchmarks.

I. Comparison of ADLPIPE vs. HUPIPE Computer Program - This study included eight (0) typical pipe systems havinq various diameter pipe sizes.

ADLPIPE Algebraic Summation vs. NUPIPE ADLPIPE Option 'l92 vs. HUPIPE ADLPIPE Option 160 vs. NUPIPE

II. Second comparison of ADLPIPE Program was to selected problems in report Bl<L HUREG-21241-R2 issued by Department of >>uclear Energy Brookhaven tlational Laboratory.

ADLPIPE Algebraic Summation vs. BHL Hovgaard Problem ADLPIPE Option 192 vs. BHL Hovgaard Problem ADLPIPE Option 160 vs. BHL Hovgaard Problem ADLPIPE Algebraic Summation vs. Bl'IL Coffee Table Problem ADLPIPE Option 192 vs. BNL Coffee Table Problem ADLPIPE Option 160 vs. BHL Coffee Table Problem

4) The following list identifies those piping systems on Washington Public Power Supply System (WPPSS) Nuclear Project No. 2 which have been run on ADLPIPE where Algebraic Summation Intramode has been used.

~Sstsm Standby Liquid Control (SLC)

Condensate (COND)

Reactor Mater Cleanup (RWCU)

'II/2 Code/Class 831.1 III/2 Heating Steam (HS) ~

831.1 Heating Steam Condensate (HCO) 831.1 Diesel Exhaust (DE) 831.1 Diesel Exhaust (DE)

Diesel Exhaust (DE)-

Standby Service Water (SW) .

Standby Service Water (SW)

Spray Pond Ring Header (SW)

~

831. 1 831.1 III/2 III/2 III/2 the above piping systems have been rerun using ADLPIPE Option 160

'll'f and have been found to be within Code limits for pipe stress allowables and all supports and supplemental structures are within current rated Load Capacity.

The plant is fully capable of withstanding the seismic criteria as set forth in the Washington Public Power Supply System, nuclear Project t<o. 2, FSAR..

Attachment II SAP46 Verification Pro ram Descri tion SAP4G, a version of SAP, was originally developed for General Electric by F.A. Peterson and t;:J. Bathe of the Engineering Analysis Corporation at Berkely. The SAP program is a general purpose structure progr am used to perform static and dynamic analysis of mechanical and piping components by the finite element method.

Ve ri fi cation All GE p oduciion vers i ons of SAP are verified using a special benchmark problem that exercises all the important features of the program. The benchmark problem has been analyzed for the effects of constraint of free end, distributed forces, and is dynamically analyzed to determine mode shapes and natural frequencies using Swanson System's ANSYS program. ANSYS was also used to predict dynamic response spectra,and time histroy integration methods. The predicted frequencies, mode shapes, and loads were compared to the corresponding SAP ANSYS before SAP was qual!defied for production use.

In order to test unique, features of SAP that cannot be compared to the results of another program, a special problem is devised which has an equivalent computer or manually calculated solution. Before any new version of SAP is

, verified, for production application, the benchmark-problem is re-analyzed to verify that the program'changes have not changed predictions or reduced their accuracy.

PISYS Veri fication Pro ram Description PISYS is a computer program specializing to analyze pi ping systems. The PISYS program provides a highly flexible user oriented input format for pi ping system modeling. The analysis modules of PISYS are taken directly

.from the SAP4G program.'erification Since PISYS analysis modules are identical to SAP46, a SAP analysis of a typical BHR steam piping system is used as a benchmark problem for PISYS verification. The steam line is analyzed for thermal expansion, dead weight,

, and a variety of dynamic loads in order to exercise all the features of PISYS. PISYS was not verified as a production program unti 1 the predictions of SAP and PISYS were shown to be identical for practical purposes.

Before any new version of PISYS is verified for production application, the benchmark problem is re-analyzed to verify that the program changes have not changed the predictions or reduced their accuracy.

Five NRC benchmark problems will also be analyzed as a further verification of the PISYS code. This analysis is expected to be completed and submitted by GE to the Conmission for review.