ML20066E733

Interim Deficiency Rept Re Discrepancy in Ghosh Computer Program Used to Develop Mark II Response Spectra.Initially Reported on 820927.Problem Subroutine Corrected to Include Stiffness Matrices of Triangular Subsections

ML20066E733
Person / Time
Site:	Shoreham File:Long Island Lighting Company icon.png
Issue date:	10/29/1982
From:	Milligan M LONG ISLAND LIGHTING CO.
To:	Haynes R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I)
References
10CFR-050.55E, 10CFR-50.55E, SNRC-785, NUDOCS 8211150418
Download: ML20066E733 (2)

Text

-

o a e u :v me ns ,* t- -a r 29 .,J

, faaramww g SHOREH AM NUCLEAR POWER STATION P.O. DOX 618 NOR1:8 """jNTu f ROAD e WADING RIVER, N.Y.11792

_ m. _

October 29, 1982 SNRC-785 Mr. Ronald C. Haynes

• Office of Inspection & Enforcement Region I U.S. Nuclear Regulatory Commission 631 Park Avenue King of Prussia, PA 19406 LONG ISLAND LIGHTING COMPANY Shoreham Nuclear Power Station - Unit 1 Docket No. 50-322

Dear Mr. Haynes:

On September 27, 1982 we notified Region I verbally of an apparent discrepancy in the GHOSH computer program used to develop Shoreham Mark II response spectra which could be potentially reportable in accordance with 10CFR50.55 (e) . We were informed of this situation by Stone & Webster Engineering Corporation on September 24, 1982.

This letter serves as our 30 day written report on this potentially reportable condition.

Description of the Potential Deficiency The GHOSH program is a commercially available finite element program used at Shoreham in the development of the response spectra for Mark II loadings. It should be noted that the GHOSH methodology was not used in developing seismic response spectra for our facility. The problem was first reported to the NRC by Gulf States Utilities who have used GHOSH in calculating Mark II and seismic response spectra.

The discrepancy occurs in an internal subroutine which calculates stiffness matrices for triangular finite elements. The program internally breaks each triangular element into three triangular subsections in order to determine the centroid of the element and thuc its stiffness matrix. In performing this function, it has incorrectly ignored the stiffness of two subecctions, assigning 8211150418 821029 PDR ADOCK 05000322 s PDR FC-8 935

W ,

,4

• SNRC-785 October 29, 1982 Page Two the stiffness matrix of one subsection to the entire triangular element. This tends to present a lower relative stiffness than actually exists. These triangular elements were used in combin-ation with rectangular elements in modeling of the soil beneath the Reactor Building. No triangular elements were used in the superstructure, thereby reducing the effects of the error.

Corrective Action The problem subroutine has been corrected to include the stiffness matrices of all triangular subsections. We are presently perform-ing study runs to determine the extent of original inaccuracies.

Due to the limited use of triangular elements in our model, we do not anticipate a significant impact of this error. Our investi-gation is scheduled to be completed by December 30, 1982 at which time we will inform your office of its outcome.

If you have any questions concerning this matter, please contact us.

Very truly yours,

) --

N .%. pn M. H. Milliga Project Engineer Shoreham Nuclear Power Station DJH/cmw cc: Mr. Richard DeYoutg, Director NRC Office of Inspection & Enforcement Division of Reactor Operators Inspection Washington, DC 20555 Mr. J. Higgins, Site NRC All Parties A