ML20052H954
| ML20052H954 | |
| Person / Time | |
|---|---|
| Site: | Fermi  |
| Issue date: | 05/15/1982 |
| From: | Wells D DETROIT EDISON CO. |
| To: | James Keppler NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| References | |
| 10CFR-050.55E, 10CFR-50.55E, EF2-57467, NUDOCS 8205240248 | |
| Download: ML20052H954 (2) | |
Text
Donald A. Wells vc,
, rx r, A.. or.
Detroit
~~
Ecison !!EE=~
May 15, 1982 EF2-57467 Mr. James G. Keppler, Regional Administrator Frgion III U.S. Nuclear Regulatory Ccumission 799 Roosevelt Road Glen Ellyn, Illinois 60137
Subject:
Final Report of 10CFR50.55(e) Item on Deficiency in Ccnputer Program (#49)
Dear Mr. Keppler:
This is Detroit Edison's final report on the deficiency in a ccuputer program utilized by Detroit Edison Engineering for the Penni 2 Project.
This problem was originally reported to Mr. J.E. Konklin of HEC Region III by Detroit Edison's Mr. H.A. Walker, Supervisor-Construction Quality Assurance, on December 7, 1981.
On September 2,1981, Control Data Corporation notified Detroit Edison that errors had been detected in the ccuputer program Baseplate, version 1.0.
On Novenber 19, 1981, Control Data Corporation notified Detroit Edison that an additional error had been detected in the ccuputer program.
During tk period of June 29, 1981 through November 3,1981, the program was available as a standard product frcm Control Data Corporation and was used for pipe support analysis on Enrico Fermi Unit 2.
The subject errors were corrected in Baseplate, version 2, released on November 4,1981.
Control Data Corporation's explanation of the three problems discovered in Baseplate, version 1.0 is as follows:
1.
Negative angles are converted internally to positive angles. This results in incorrect nodal locations for attachments, and incorrect load transfers.
2.
Multiple load cases within a run with the S directive active is a potential problem because the postprocessor may overwrite itself.
This is caused by scue arrays exceeding their dimensions because subscripts are not re-zeroed for each load case. If the nurrber of load cases and the number of elenents are large enough, array dimen-sions will be exceeded and potential erroneous nurrbers may be outputted.
p*gy 17198k 820S240248 820S15 PDR ADOCK 05000341 9
PDR ZEM s, _o
Mr. James G. Keppler, Regional Administrator May 15, 1982 Page % o EF2-57467 3.
The pipe attachment locating dimensions define the pipe centerline coordinates on the surface of the baseplate regardless of whether or not the pipe is perpendicular to the plane of the baseplate. Version 1.0 of the program inadvertently included an offset in the locating dimensions when the pipe attachment was not perpendicular to the plane of the baseplate. This resulted in pipe attachment coordinates being incorrectly located.
It was concluded that the potential for excessive loads on pipe support baseplates existed, and could cmprcmise the integrity of OA Invel I pipe supports. Therefore, in accordance with 10CFR50.55(e), a reportable design deficiency existed.
All OA Level I previously designed pipe supports were reviewed to determine where the subject program had been used. In addition, the investigation included a review and cmputer re-analysis, if necessary, to determine if a nodification was required for each affected support. The results of this-investigation revealed that none of the OA IcVel I pipe supports required modification.
An Engineering assessment of the potential effects of the errors in,the program, Baseplate, version 1.0, concluded that the potential for support system structural failures and attendant overstress of the OA Invel I piping could exist.
We have evaluated the original design of the supports that could have been affected by this design problem. None of these supports require modification, and therefore the integrity of OA IcVel I pipe supports has been maintained.
~
If you have questions concerning this matter, please contact Mr. H.A. Walker, Supervisor-Construction Quality Assurance.
Very truly yours,
]
b a
DAW / HAW /cp i
cc: Mr. Richard DeYoung, Director Office of Inspection and Enforcement i
Division of Reactor Inspection Prograrrs
[
U.S. Nuclear Regulatory Ccmnission i
Washington, D.C. 20555 Mr. Bruce Little, Resident Inspector U.S. Nuclear Regulatory Ccmnission 6450 North Dixie Highway Newport, Michigan 48166 l
.