ML20080B279

Confirms 830714 Telcon Responding to NRC Questions Re Ghosh Computer Code.Discrepancies in Ghosh Program & Labeling of Rocking Acceleration Data Unrelated Issues

ML20080B279
Person / Time
Site:	Shoreham File:Long Island Lighting Company icon.png
Issue date:	08/03/1983
From:	James Smith LONG ISLAND LIGHTING CO.
To:	Harold Denton Office of Nuclear Reactor Regulation
References
SNRC-944, NUDOCS 8308050395
Download: ML20080B279 (5)

Text

.

1 J LONG ISLAND LIGHTING COMPANY w SHOREHAM NUCLEAR POWER STATION P.O. BOX 618, NORTH COUNTRY ROAD e WADING RIVER, N.Y.11792 Direct Dial Number August 3, 1983 SNRC-944 Mr. Harold R. Denton, Director Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Response to Additional NRC Questions on the GHOSH Computer Code Discrepancies Shoreham Nuclear Power Station - Unit 1 Docket No. 50-322

References:

1. Telphone conversation between E. Weinkam, et al (NRC) and R. W. Grunseich, et al (LILCO), dated July 14, 1983

2. Notes of telephone conversation between R. Caruso and R. W. Grunseich dated June 28, 1983

3. LILCO letter, SNRC-875 (. M . H. Milligan) to the NRC (J. M. Allan) , dated April 20, 1983.

Dear Mr. Denton:

As requested, the following documents our verbal responses given

to the NRC on July 14, 1983 (Ref. 1) to questions concerning the GHOSH Computer Code discussed in References 1 and 2.

Question 1:

"On page 2 of LILCO's letter (SNRC-875), the units of rocking acceleration (g/ft and rod /sec) have been mentioned. Explain the difference between these units in actual computation and provide conversion for one unit to the other."

Response

On page 2 of Reference 3, the units of rocking acceleration are

stated in g/ft and radians /sec2 , where g represents the gravita-tional acceleration and equals 32 2 ft/sec2 To convert the value in g/ft unit to radians /sec I unit, one has to multiply the

\

eaceosoaes esoeoa 79

^PDR ADOCK 05000322 PDR I i

1 FC 8935.1 l

e ,

August 3,'1983 SNRC-944 Page 2 r value in g/ft" unit by 32.2. i Question 2:

i ."In view of observation that the 2 errors that occurred are an

' indication to some degree of a breakdown in'the QA program in

, design,' describe action taken after the discovery of the 2 errors as far as overall QA in design is concerned."

Response

As stated in the report transmitted.via Reference 3, the dis-crepancy in the GHOSH program and the discrepancy in the labeling of rocking acceleration data are unrel.ated issues.

The GHOSH program is a commercially available finite element .

program. Although the program has been modified and improved by our architect engineer (SWEC), the discrepancy was within an original, unmodified subroutine that generated the triangular element stiffness matrix which was unique to this program.

The other discrepancy was due to an isolated error in the labeling Jof the units of rocking accleration data transmitted to General Electric. The report addressed both issues together upon request by_the NRC only because they were both concerned with the: development of the Mark II Confirmatory Loads.

The occurrence of these two events, since they were related.

~

both'in nature and cause, cannot be reasonably considered as indicators of any programmatic weakness in our QA program for design. As stated in reference 3 the following actions have been initiated:

The discrepancy in the GHOSH program has been rectified and

.an assessment of the effects has been completed.

Both SWEC and GE have completed evaluations of the revised i

Mark II confirmatory loads.

GE has also considered the rocking acceleration data mis-labeling-in their evaluation.

Additionally, SWEC will perform an independent review of their

~QA. Category I load transmittals which are generated from the

, Shoreham' building dynamic analyses, to prevent future trans-mittal discrepancies.

4

--,-,+e e . .- ,wm.e.,,,- ,, -..,-.-...-,.i,, - .y .n,.%,,. - - -y,.. , . , - - , , , ...-,%. e--.<w_.,-y- ,, - .,.,m, ,,< _ce,-ww,.w, .

o ,,

August 3, 1983 SNRC-944 Page 3 Question 3:

"In light of deficiency found in GHOSH program, provide a description of steps taken to assure the validity of this program as a whole. An individual check of the analysis through the use of another validated computer program is required."

Response

Upon discovery of the GHOSH program error,the GHOSH program was corrected to eliminate the discrepancy. The revised code was then used on several typical problems including those that used only triangular elements. The results obtained were compared to solutions obtained using computer program ANSYS (a fully verified program) and other theoretical solutions. All comparisons were favorable.

Some of the results are presented in the forthcoming change to the FSAR.

Question 4:

"In Section 6.2.1 on page 9 of your April 1983 report you stated that in the assessment, the load data used are the representative sets of the revised confirmatory loads and in Figures 1 to 12 of the report you indicated selected time-histories were used. Since the responses of structural systems are affected by the inter-relationship between the stiffness of the supporting soil and the frequency content of the dynamic input, provide your assurance together with its basis that the responses obtained by you are the enveloping or the governing responses."

Response

In the reanalysis of the confirmatory Mark II loads, all input time-history traces were considered for the reanalysis.

The traces selected for further review were based on the significance of each input trace as it affects the responses of the structure. Figures 1 through 12 in the report (Ref.3) show a comparison of two curves. Both curves are developed using the old GHOSH program. The solid curves (labeled )

are ARS developed using the complete sets of the Mark II loads.

The dotted curves (labeled ---) are ARS developed using only the selected traccs used in the reanalysis. The differences between the two curves are the contribution of traces not used in the reanalysis. Although the ARS in Figures 1 through

- - - .~. . . . . - . - - - - -

August 3, 1983 SNRC-944 Page 4 12 are developed using the old GHOSH program, they also represent the results of a study of the frequency content of the input time history traces, since the only variable is the number of traces used. Figures 1 through 12 shows that the comparison of the two curves is favorable. Good comparison of the curves also implies adequacy, in terms of frequency content. Thus the results, as shown in Figures 1 through 12 in the report (Ref. 3), indicated the traces selected are adequate.

Changes in structural responses due to the discrepancy in the triangular element can be seen by comparing the original ARS in Figures 1 through 12 (labeled --) with the revised ARS in Figures 17 through 28 (labeled -- --). These ARS figures (1 through 12 and 17 through 28) were developed using the same input time-history traces. As can be seen from the figures, the over-all shape of the ARS remains the same with only minor shifts in ARS peaks. This indicates that the effects of the discrepancy in the triangular soil elements is minor. Thus, based on the comparison as shown in Figures 1 through 12, the contribution from the time-history traces not used in the reanalysis of the revised ARS will not be significant. The input time-history traces that were selected are adequate for the assessment.

LILCO believes the information stated herein should enable the NRC to verify the adequacy of the corrective actions and the actions taken to prevent recurrence that were initiated as a result of the discovery of these discrepancies.

If any additional information is required please contact this office.

Very truly yours,

. L. Smith Manager, Special Projects Shoreham Nuclear Power Station bc cc: J. Higgins All Parties Listed in Attachment 1

ATTACHMENT 1

~

t

\

Lawrence Brenner, Esq. Herbert H. Brown, Esq.

Administrative Judge Lawrence Coe Lanpher, Esq.

Atomic Safety and Licensing Karla J. Letsche, Esq.

Board Panel Kirkpatrick, Lockhart, Hill U.S. Nuclear Regulatory Commission Christoper & Phillips Washington, D.C. 20555 -8th Floor 1900 M Street, N.W.

Washington, D.C. 20036 4

Dr. Peter A. Morris Administrative Judge Atomic Safety and Licensing Mr. Marc W. Goldsmith Board Panel Energy Research Group U.S. Nuclear Regulatory Commission 4001 Totten Pond Road Washington, D.C. 20555 Waltham, Massachusetts 02154 Dr. George A. Fergus6n MHB Technical Associates School of Engineering 1723 Hamilton Avenue Howard University Suite K i 2300 Fifth Street San Jose, California 95125 Washington, D. C. 20059 Stephen B. Latham, Esq.

Twomey, Latham & Shea Daniel F. Brown, Esq. 33 West Second Street Attorney P.O. Box 398 Atomic Safety and Licensing Riverhead, New York 11901 Board Panel U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Ralph Shapiro, Esq.

Cammer and Shapiro, P.C.

9 East 40th Street Bernard M. Bordenick, Esq. New York, New York 10016 David A. Repka, Esq.

U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Matthew J. Kelly, Esq.

State of New York Department of Public Service James Dougherty Three Empire State Plaza 3045 Porter Street Albany, New York 12223 Washington, D.C. 20008

. _ . , _. _ _ _ _ _ . . _ _ _ _ . . - _ . . . - . _ _ _ _ . . _