Part 21 Report - 1997-620

ACCESSION #: 9706190261 GPU Nuclear, Inc. G P U One Upper Pond Road NUCLEAR Parsippany, NJ 07054-1095 Tel 201-316-7000 June 12, 1997 6730-97-2176 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, DC 20555 Gentlemen: Subject 10CFR Part 21 30-Day Report Oyster Creek Nuclear Generating Station Docket No. 50-219 Potential Failure of General Electric Type CR120AD Relays Oyster Creek has experienced failures of the General Electric Type CR120AD relays. The failure mechanism was identified as a manufacturing defect; specifically, part of the start-end pigtail of the relay coils was off the insulating lead pad and, therefore, lying directly on the outer layer of the winding. This configuration placed insulation on the affected length of magnet wire in a high-electrical-stress condition that led to dielectric breakdown at the failure site, thus effectively short circuiting the coil. Although our evaluation of all ten potentially defective relays installed at Oyster Creek concluded that their fail-safe mode does not impact plant safety, this potential defect is being conservatively reported in accordance with 10 CFR Part 21.21, (d) (3) (ii). Should you have questions regarding this matter, please contact Yosh Nagai of our staff at (201) 316-7974. Sincerely, T. G. Broughton President YN/plp c: Administrator, Region I NRC Resident Inspector Oyster Creek NRC Project Manager 10 CFR 21 - 30 DAY WRITTEN REPORT CONCERNING POTENTIAL FAILURE OF GENERAL ELECTRIC TYPE CR120AD RELAYS 1. Name and Address Of The Individual Informing The Commission: Name: T. G. Broughton President Address: GPU Nuclear, Inc. One Upper Pond Road Parsippany, NJ 07054-1149 II. Basic Component Affected: General Electric Type CR120AD Relays III. Firm Supplying Components:, General Electric Nuclear Energy 175 Curtner Avenue San Jose, California 95125 IV. Nature of the Potential Defect: On March 26, 1996 Oyster Creek Generating Station issued a deviation report which identified an open coil in three CR120AD relays dedicated for safety-related use: one relay failed immediately after the installation when energized, another relay failed within two days during startup test and a third failed some months later during a surveillance test. Following these incidents a coil from one of the failed relays was sent to General Electric (GE) for analysis and the two remaining relays were sent intact to an independent research laboratory (Southwest Research Institute) for failure analysis. After the analysis, Southwest Research Institute concluded that the principal problem observed in both relay coils was a manufacturing defect: specifically, part of the start-end pigtail of the coil was off the insulating lead pad and, therefore, lying directly on the outer layer of the winding, which was near the finish end of the coil. This configuration placed insulation on the affected length of magnet wire in a high-electrical-stress condition that led to dielectric breakdown at the failure site, thus effectively short circuiting the coil. Resulting high current through the short circuit caused the magnet wire to fuse and produced an energetic arc at the failure site. There are ten (10) potentially defective relays installed at Oyster Creek. GPU Nuclear has concluded that their fail-safe mode does not Impact plant safety. However, there may be other nuclear facilities with these relays installed or kept in their warehouse without knowing of or about this potential failure. V. Date on which Potential Defect Was Identified. We determined on May 9, 1997 that this potential failure of the GE relay CR120AD could have a generic implication and it was reported to NRC Operations Center, by phone, on May 9, 1997. VI. Number and Location of Components. There are ten (10) potentially defective relays installed at Oyster Creek. VII. Corrective Actions Taken If a premature failure of any of these relays occurs, with an "open" coil as the failure mechanism, we may consider replacing the remaining nine relays. However, replacement of the ten relays is not scheduled at this time for the following reasons: i) The previous failures of the CR120AD relays at Oyster Creek occurred either immediately upon energization or within a few months of operation. Also, none of the installed relays have failed since their installation approximately seven months ago. ii) The probability of these relays having a potential manufacturing defect is extremely low. iii) Our evaluation concluded that there is no impact to plant safety due to their fail-safe mode feature. GPU Nuclear, Inc. G P U One Upper Pond Road NUCLEAR Parsippany, NJ 07054-1095 Tel 201-316-7000 June 12, 1997 6730-97-2176 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, DC 20555 Gentlemen: Subject: 10 CFR Part 21 30-Day Report Oyster Creek Nuclear Generating Station Docket No. 50-219 Potential Failure of General Electric Type CR120AD Relays Oyster Creek has experienced failures of the General Electric Type CR120AD relays. The failure mechanism was identified as a manufacturing defect; specifically, part of the start-end pigtail of the relay coils was off the insulating lead pad and, therefore, lying directly on the outer layer of the winding. This configuration placed insulation on the affected length of magnet wire in a high-electrical-stress condition that led to dielectric breakdown at the failure site, thus effectively short circuiting the coil. Although our evaluation of all ten potentially defective relays installed at Oyster Creek concluded that their fail-safe mode does not impact plant safety, this potential defect is being conservatively reported in accordance with 10 CFR Part 21.21, (d) (3) (ii). Should you have questions regarding this matter, please contact Yosh Nagai of our staff at (201) 316-7974. Sincerely, T. G. Broughton President YN/plp c: Administrator, Region 1 NRC Resident Inspector Oyster Creek NRC Project Manager 10 CFR 21 - 30 DAY WRITTEN REPORT CONCERNING POTENTIAL FAILURE OF GENERAL ELECTRIC TYPE CR120AD RELAYS 1. Name and Address Of The Individual Informing The Commission: Name: T. G. Broughton President Address: GPU Nuclear, Inc. One Upper Pond Road Parsippany, NJ 07054-1149 II. Basic Component Affected: General Electric Type CR120AD Relays III. Firm Supplying Components: General Electric Nuclear Energy 175 Curtner Avenue San Jose, California 95125 IV. Nature of the Potential Defect: On March 26, 1996 Oyster Creek Generating Station issued a deviation report which identified an open coil in three CR120AD relays dedicated for safety-related use: one relay failed immediately after the installation when energized, another relay failed within two days during startup test and a third failed some months later during a surveillance test. Following these incidents a coil from one of the failed relays was sent to General Electric (GE) for analysis and the two remaining relays were sent intact to an independent research laboratory (Southwest Research Institute) for failure analysis. After the analysis, Southwest Research Institute concluded that the principal problem observed in both relay coils was a manufacturing defect: specifically, part of the start-end pigtail of the coil was off the insulating lead pad and, therefore, lying directly on the outer layer of the winding, which was near the finish end of the coil. This configuration placed insulation on the affected length of magnet wire in a high-electrical-stress condition that led to dielectric breakdown at the failure site, thus effectively short circuiting the coil. Resulting high current through the short circuit caused the magnet wire to fuse and produced an energetic arc at the failure site. There are ten (10) potentially defective relays installed at Oyster Creek. GPU Nuclear has concluded that their fail-safe mode does not impact plant safety. However, there may be other nuclear facilities with these relays installed or kept in their warehouse without knowing of or about this potential failure. V. Date on which Potential Defect Was Identified. We determined on May 8, 1997 that this potential failure of the GE relay CR120AD could have a generic implication and it was reported to NRC Operations Center, by phone, on May 9, 1997. VI. Number and Location of Components. There are ten (10) potentially defective relays installed at Oyster Creek. VII. Corrective Actions Taken If a premature failure of any of these relays occurs, with an "open" coil as the failure mechanism, we may consider replacing the remaining nine relays. However, replacement of the ten relays is not scheduled at this time for the following reasons: i) The previous failures of the CR120AD relays at Oyster Creek occurred either immediately upon energization or within a few months of operation. Also, none of the installed relays have failed since their installation approximately seven months ago. ii) The probability of these relays having a potential manufacturing defect is extremely low. iii) Our evaluation concluded that there is no impact to plant safety due to their fail-safe mode feature. *** END OF DOCUMENT ***

Page Last Reviewed/Updated Wednesday, March 24, 2021