05000341/LER-2001-003
| Fermi 2 | |
| Event date: | |
|---|---|
| Report date: | |
| Reporting criterion: | 10 CFR 50.73(a)(2)(i)(B), Prohibited by Technical Specifications |
| 3412001003R00 - NRC Website | |
Initial Plant Conditions:
Mode Reactor Power Reactor Pressure Reactor Temperature 5 (Refueling) 0 Percent 0 psig 90 Degrees Fahrenheit
Description of the Event
On November 9, 2001, leak rate testing of Pressure Isolation Valve (PIV)[ISV] El 1 00F050A, Division 1 Residual Heat Removal (RHR)/Low Pressure Coolant Injection (LPCI)[BO] System injection line inboard isolation check valve, was performed in accordance with Technical Specification (TS) Surveillance Requirement SR 3.4.5.1. The resultant leak rate was determined to be in excess of the specified leakage criteria of 10 gpm. While attempting to pressurize the inboard side of the valve, the test pressure could not be achieved and, therefore, leakage past the PIV was categorized as through seat leakage. During the testing, Division 1 systems were out of service for maintenance activities on that Division; therefore, there was no impact on the plant in the refueling configuration. The upstream RHR/LPCI System injection line outboard isolation motor-operated valve, El 1 50F015A, was successfully leak tested and met its specified leakage criteria.
This event is being reported in accordance with 10 CFR 50.73(a)(2)(i)(B) as a condition prohibited by TS Limiting Condition for Operation (LCO) 3.4.5, RCS Pressure Isolation Valve Leakage. LCO 3.4.5 requires PTV leakage to be within the specified limits during plant operation in Modes 1 and 2.
Cause of the Event
Valve E1100F050A and its air-operated test actuator were disassembled to determine the cause of the excessive leakage. Examination revealed that the valve disk was being prevented from fully closing by the test actuator. Previous assembly of the actuator had resulted in the actuator spur gear and the actuator gear rack being misaligned by one tooth such that the actuator shaft could not complete its full rotation.
This prevented the valve disk from fully closing under no/low flow and low differential pressure conditions. The leak testing method used on November 9, 2001 involved very low flow conditions.
Valve E1100F050A had passed its leak test coming out of the previous refueling outage by employing another test method consisting of pressurizing the line beyond the check valve to the next isolation valve, El 1 50F015A, and then opening El 150F015A, simulating a break downstream. Valve El 1 00F050A fully seated during this previous test because of the resulting differential pressure exerted on the disk.
The cause of the failure of the disk to fully close was determined to be improper reassembly of the actuator during the previous refueling outage because of inadequate craft skills and insufficient craft supervision during this period. The procedure used to reassemble the actuator following disassembly provided detailed instructions concerning the importance of proper fit, and required the match marks made during the removal of the actuator to be properly aligned upon its restoration. This was not properly performed during the previous refueling outage, resulting in the actuator spur gear and the actuator gear rack being misaligned by one tooth such that the actuator shaft could not complete its full rotation.
This valve had failed its leak rate test during the previous refueling outage because of degradation of the elastomer soft seat, which was believed to have been caused by hot water leakage through the soft seat.
The apparent high temperature degradation of the soft seat observed during the previous refueling outage was not observed this cycle. Thermocouples were installed during the previous refueling outage and the valve was monitored during this cycle for high temperature conditions. No high temperature conditions were observed, and the soft seat was found to be in good condition.
The opposite division counterpart, E1100F050B, successfully passed its as-found leak rate test this outage; however, it was also disassembled, inspected, and its soft seat replaced. With both actuators properly assembled, and with new soft seats installed, historical performance indicates that both valves should pass their as-found leak rate tests at the end of the next operating cycle.
Analysis of the Event
The purpose of the PIVs is to provide isolation at the interfaces between high pressure and low pressure systems. The affected check valve is one of two valves in series in the LPCI System injection line. The valve provides for isolation of the high pressure Reactor Coolant System from the low pressure LPCI System. The other valve in the injection line, a motor-operated valve, E1150F015A, was successfully tested and met its leakage criteria. Therefore, isolation of the high pressure to low pressure interface was maintained during power operation.
The PIV leak test performed this outage consisted of pressurizing the test volume between the check valve and a downstream closed manual isolation valve, E1100F060A/B, with three positive displacement pumps of approximately 10 gpm capacity each. Leakage by the disk exceeded the test pumps capacity and prevented the test volume from being pressurized sufficiently to seat the disk, thereby preventing demonstration that the valve could meet the leak rate acceptance criteria. Valve El 100F050A passed its leak test at the end of the previous refueling outage by employing a different test method, consisting of pressurizing the line beyond the check valve to the next isolation valve, El 150F015A, and then opening EI 150F015A, simulating a line break downstream. The disk on valve E1100F050A fully seated during this previous test because of the resulting differential pressure exerted on the disk using this method. Thus, we believe that E1100F050A would have fully closed if a demand resulting from an actual intersystem loss of coolant accident (LOCA) would have occurred during the cycle. Additionally, it is believed that E1100F050A would have passed its leak rate test this outage had this more realistic leak test methodology been used. The soft seat which had been a problem during previous tests does not appear to have been a problem this cycle.
Regardless of the condition of E 1 1 00F050A, this event had no adverse impact on the health and safety of the public because the other PIV in the injection line, E 1 150F015A remained closed during Cycle 8 plant operation, and successfully passed its leak rate test this outage.
Corrective Actions
Following the seventh refueling outage (RFO-7), Detroit Edison recognized that the skill of the craft in the specialty work activity of valve repair was limited. Therefore, preparations for the most recent, eighth refueling outage (RFO-8) included a detailed screening process for valve workers consisting of pre-hire written and practical tests. A 100% fidelity mockup of the E1100F050 valve was purchased from the vendor, Anchor Darling and was used for pre-outage training. Valve workers practiced on the mockup for 3 weeks prior to the outage under the scrutiny of Detroit Edison and contractor supervision, as well as the vendor representative, to validate the procedure and to hone their skills. Additionally, Detroit Edison and contractor supervision was increased this outage to ensure that valve repairs would be correctly performed. As a result of this pre-outage training and increased supervision, the problem associated with the previous reassembly of E 1 1 00F050A was recognized and corrected well within the window scheduled for this work.
The air-operated test actuator on valve E1100F050A was properly reassembled in the presence of the vendor, and a new soft seat was installed. The removed soft seat appeared to be in good condition. The post maintenance leak rate was within Technical Specification acceptance criteria.
Check valve E1100F050B, the Division 2 counterpart to E1100F050A, successfully passed its as-found leak rate test. However, due to past concerns, the soft seat on E1100F050B was also replaced following the as-found testing. The removed soft seat appeared to be in good condition. Fit-up gap measurements between the disk and the valve body were taken to determine whether the valve actuator and hinge mechanism were properly aligned. These measurements showed that the fit-up was good, and no further work was performed on the actuator.
With both valve actuators properly assembled, and with new soft seats installed, historical performance indicates that both valves should pass their as-found leak rate tests at the end of the next operating cycle.
Temperature monitoring instrumentation installed on check valves E1100F050A and E1100F050B to monitor operating conditions for these valves during the previous cycle has been removed. Temperatures recorded during the previous cycle remained below 300 degrees Fahrenheit, well within the soft seat specifications.
Further corrective actions relating to this event to obviate the need for periodic soft seat replacement are being considered, and will be developed and implemented commensurate with established priorities and processes of the Fermi 2 corrective action program. This event is documented in the Fermi 2 corrective action program (CARD 01-20111).
Additional Information
A. � Failed Components Component: Division 1 Residual Heat Removal (RHR)/Low Pressure Coolant Injection (LPCI) System Inboard Isolation Testable Check Valve (E1100F050A) Description: 24 inch-Full Exercisable Air Operated Swing Check Valve Manufacturer: Anchor Darling Type: � Model 2229-3 B. � Previous LERs On Similar Problems This LER documents the previous failure of the as-found leak rate test for E1100F050A which occurred during the seventh refueling outage. This failure was caused by the degradation of the soft seat.
This LER and its supplement document previous failures of the as-found leak rate tests for E1100F050A/B which occurred during the fifth and sixth refueling outages. In 1998, E1100F050B failed its as-found LLRT in the sixth refueling outage due to degradation of the soft seat and was reported in LER 98-008. During the investigation for this LER, it was discovered that E1100F050A had failed its as-found leak rate test during the fifth refueling outage in 1996, but had not been reported. LER 98-008-01 was submitted to document this failure. This failure was caused by the degradation of the soft seat which was exacerbated by a minor misalignment of the valve disc.