Headquarters Daily report AUGUST 02, 1995 *************************************************************************** REPORT NEGATIVE NO INPUT ATTACHED INPUT RECEIVED RECEIVED HEADQUARTERS X REGION I X REGION II X REGION III X REGION IV X *************************************************************************** PRIORITY ATTENTION REQUIRED MORNING REPORT - REGION I AUGUST 2, 1995 Licensee/Facility: Notification: Gpu Nuclear Corp. MR Number: 1-95-0098 Oyster Creek 1 Date: 08/02/95 Forked River,New Jersey SRI PC Dockets: 50-219 BWR/GE-2 Subject: POTENTIAL MANUFACTURING DEFECT OF GENERAL ELECTRIC CR120AD DC RELAYS Reportable Event Number: N/A Discussion: During Oyster Creek's 1994 Fall refueling outage, 19 safety related relays were replaced as part of the Relay Maintenance Program. The replacement relays, General Electric CR120AD 120 Vdc relays, are upgraded versions of the existing GE CR120A obsolete relays. An alternate replacement engineering evaluation was performed. One relay failed immediately after it was installed and energized. Another relay failed within several days of replacement. Followup investigation of the failed relays revealed a short circuit in the dc coils; with no evidence of excessive heatup or overcurrent damage. One was a normally energized relay, and the other was a normally de-energized relay. A third failure (normally de-energized) occurred about three months later during performance of a surveillance test that exercised the relay. GPUN sent two of the failed relays to an offsite laboratory for failure analysis. The laboratory determined that the principal problem in each failed relay was a manufacturing defect. Specifically, the start-end of the coil was off the insulating lead pad and was in contact with the outer layer of the coil winding (near the finish-end of the coil). The affected length of coil wire and insulation was placed in a high electrical stress condition that led to dielectric breakdown at the failure site, thus effectively short circuiting the coil. The resulting high current through the short circuit caused the coil wire to fuse and produce an energetic arc at the failure site. The laboratory concluded that it may be possible to identify failure-prone relays prior to installation by repetitive bench testing, since it appeared certain that a relay containing the defect would fail. GPUN purchased the GE CR120AD relays as safety-related components from a vendor, Farwell & Hendricks, Inc. (F&H). However, F&H originally purchased the relays as commercial-grade components, and then commercially dedicated them. F&H developed a Dedication Plan for the relays, which was approved by a GPUN representative. F&H performed relay inspection and functional testing of all of the relays provided to GPUN (23 in total). There were no failures. GPUN did not perform functional testing upon receipt of the relays, however, their Control Relay Replacement, Adjustment and Test procedure performed functional relay testing as a part of the relay installation and turnover process. Currently, the remaining GE CR120AD relays from the original purchase order are located in the warehouse and are on "QA-hold" to prevent installation. There was a subsequent purchase order of 18 relays, and they are also in the warehouse. However, there is no QA-hold on those relays as GPUN has not confirmed a problem with the second "batch" of relays. Engineering is developing an enhanced pre-service testing program for those relays in the event they are installed in the plant. GPUN anticipates future relay replacements in the next refueling outage, and is pursuing actions to resolve this issue. GPUN notified the nuclear industry of their experiences with these relays via Nuclear Network. They are also reviewing this issue for 10CFR21 applicability and sent the laboratory failure analysis report to F&H for further investigation and corrective action. Regional Action: Routine resident inspector followup. Contact: John Rogge (610)337-5146 PRIORITY ATTENTION REQUIRED MORNING REPORT - REGION IV AUGUST 2, 1995 Licensee/Facility: Notification: Houston Lighting & Power Co. MR Number: 4-95-0094 South Texas 2 Date: 08/02/95 Wadsworth,Texas SRI Dockets: 50-499 PWR/W-4-LP Subject: INADVERTENT START OF STANDBY DIESEL GENERATOR Reportable Event Number: 29147 Discussion: On August 1, 1995, at 3:05 p.m. (CDT), the Unit 2, Train A, Standby Diesel Generator 21 inadvertently started from a standby condition. The Train A engineered safety features bus, Switchgear E2A, continued to be powered from the offsite grid. Indications locally and in the control room verified that the start was in the normal mode and had not been initiated by the engineered safety features actuation circuitry. No work was taking place in areas or on systems that would have potentially caused the start. The standby diesel generators at South Texas Project have a long history of inadvertent starts caused by noise in the dc system actuating the nonclass, fiber optics, test start circuitry. The NRC has previously reviewed this problem and determined that, while these inadvertent, normal mode starts are not desirable for the long-term maintenance of the standby diesel generators, they do not impact the safety function of the machines. After an initial evaluation of the engine, operators attempted to shut down the diesel by placing it in the cooldown sequence. However, following the cooldown run, the engine inadvertently restarted and returned to the 600 rpm, normal running speed. Since the plant had previously taken multiple Train B components out of service for planned maintenance, the licensee decided not to shut down the Train A diesel by placing the engine hand switch in "pull-to-lock," thus avoiding a multitrain outage. A previous engineering analysis had indicated that long-term unloaded runs of the diesel were undesirable. Therefore, operators loaded the generator on the grid to approximately 60 percent power. Presently, the diesel generator remains running and loaded on the grid. Licensee engineers have stated that a complete analysis has been performed to indicate that the diesel generators remain operable while tied to the grid. There are no concerns for grid stability at this time. The licensee does not intend to shut down the Train A diesel until all Train B components have been returned to service to avoid a multitrain outage. Since the time of the event, all Train B components have been returned to service with the exception of the component cooling water pump, which is out of service for replacement of the rotating element. This work is expected to be completed by 4 p.m. August 2, 1995. Once the Train A diesel is shut down, the licensee will commence troubleshooting to determine the cause of the inadvertent start signal. Regional Action: The Resident Inspectors are following the licensee's actions. Contact: L. A. Yandell (817)860-8182