- - - - - - - - - -

December 29, 2005

Motor-operated valve starter failures at Cooper

On September 23, 2005, at 2:48 a.m., Cooper operators manually scrammed the reactor in response to lowering main condenser vacuum. The loss of vacuum was caused by the failure of a drain line passing through the main condenser which allowed an unacceptable amount of air in-leakage. In order to repair the leak, the licensee determined that the reactor would have to be placed in cold shutdown using the residual heat removal (RHR) system.

At approximately 11:49 p.m., operators attempted to close the suppression pool suction valve for RHR Pump C (RHR-MOV-MO13C) in order to align Pump C for shutdown cooling. Operators placed the control switch for the valve in the closed position but the position indication lights in the control room did not indicate any valve movement. An auxiliary operator was dispatched to the reactor building and reported an acrid odor in the vicinity of the motor control center (MCC) containing the motor starter for RHR-MOV- MO13C. At approximately the same time, the control room lost position indication for the valve due to a blown fuse in the control power supply for the motor starter.

RHR-MOV-MO13C is a containment isolation valve which is normally open in Modes 1, 2, and 3 in order to satisfy the low pressure coolant injection function of RHR by aligning the Pump C suction to the suppression pool. When RHR Pump C is aligned for shutdown cooling, this valve is required to be closed in order prevent a drain path from the reactor coolant system to the suppression pool. Based on the failure of RHR-MOV-MO13C to operate on demand, operators declared RHR Loop A inoperable, as well as the associated containment penetration, and appropriately implemented the compensatory actions required by Tech Specs.

At approximately 2:40 a.m. on September 24, operators attempted to open the shutdown cooling suction valve for RHR Pump B (RHR-MOV-MO15B) in order to align Pump B for shutdown cooling. Operators placed the control switch in the open position but observed no indication of valve movement. The switch was taken back to the close position. A short time later, operators attempted to re-open the valve and were successful, but the decision was made not to use Pump B for shutdown cooling due to the previous failure. At approximately 6:53 a.m., operators were successful in establishing shutdown cooling using RHR Pump D.

Eight similar valve failures had occurred at Cooper over a 6-year period; however, corrective actions for those failures did not prevent the two similar failures in September 2005. The inspector identified this as a violation of 10 CFR Part 50, Appendix B, Criterion XVI, regarding inadequate corrective actions for motor-operated valve failures. NRC concluded that the finding was more than minor since it affected the cornerstone attributes of availability and reliability of mitigating equipment as well as the operational capabilities of primary containment. The safety significance was assessed using Phase 2 of NRC's Significance Determination Process. NRC concluded that the performance deficiency did not increase the initiating event frequencies or degrade the mitigating functions described in the Phase 2 analysis. Therefore, the finding was determined to be of very low safety significance. NRC noted crosscutting aspects associated with problem identification and resolution based on the fact that it was within the plant's capability to have determined and corrected the valve failure mechanism two months prior to the failures in September 2005, yet they failed to do so. The finding was classified as a noncited violation.

During troubleshooting on RHR-MOV-MO13C and RHR-MOV-MO15B, the plant determined that the mechanical interlock between the contactors in each of the starter assemblies was deformed. This deformation caused the interlock to bind such that when a valve operation was demanded from the control room, the associated contactor would energize but would not have sufficient force to move the interlock. In the case of RHR-MOV-MO13C, this binding caused the close contactor to remain energized in an attempt to close the valve until it overheated, causing the control power fuse to open. In the case of RHR-MOV-MO15B, the successful operation of the valve on the second attempt indicated that the failure mechanism was intermittent in nature. Plant staff also determined the most likely cause for the deformation was a manufacturing issue with the starter assemblies. The interlock is mounted to the backplate of the starter assembly with an ÒLÓ bracket which is held in place by a single screw. The mounting holes for the interlocks in these two starter assemblies were off-center; therefore, external force had to be applied to the edges of the interlocks in order to mount them using these holes. Over time, the external force caused stress relaxation in the nylon plastic case of the interlock, which resulted in misalignment of the internal components and caused the interlock to bind.

On September 25, 2005, Cooper modified the mechanical interlocks inside the motor starters for two safety-related motor-operated valves without following the requirements in their modification procedure. As a result, the required torque values specified in the seismic qualification report for this equipment were not used during the modification. NRC concluded that the finding was more than minor since configuration control and the maintenance of the plantÕs design basis is a basic principle of safe plant operation and, if left uncorrected, could become a more significant safety concern. However, the finding was determined to be of very low safety significance since it only involved a design or qualification deficiency that did not result in the loss of a safety function. NRC noted that the finding also had cross-cutting aspects associated with human performance based on the fact that appropriate administrative barriers were in place to ensure that the modifications were performed in accordance with procedures. The finding was classified as a noncited violation.

For more detail, see Cooper special inspection rept 50-298-2005-14

April 22, 2005

* Arkansas - valve maintenance failure risked overfeed; a unique motor on Limitorque feedwater block valve didn't get manufacturer-recommended preventive maintenance despite procedural mention which had been appropriately applied by workers to less unique designs

December 12, 2003

* McGuire - PORV block valve testing insights and requested frequency change

December 11, 2003

* Peach Bottom-2 - HPCI system Torus check valve found to allow reverse flow

November 6, 2003

Valve failure due to obstruction by loose fasteners for cartridge mounting plate

On February 22, 2003, at Pilgrim, the breaker for residual heat removal system valve MO-1001-29A tripped open on thermal overload when the valve was taken to the full open position for plant operation in shutdown cooling. Immediate actions were taken to investigate the cause of the failure, repair the motor operator and to restore the RHR loop to an operable status. Plant staff determined that the valve failed because four fasteners in the cartridge mounting plate backed out, which impacted the proper operation of the valve limit switches.

The plant performed a "weak link analysis" to determine the impact on valve MO-1011-29A, and to evaluate the event for reportability. Plant staff also reviewed the industry operating experience for the failure and took action to determine the extent of the condition in other safety related motor operated valves at Pilgrim. A list of potentially susceptible valves was developed and the valves were prioritized for inspection based on risk significance. Pilgrim inspected the valves during plant operations and during the refueling outage, and expanded the scope of the inspections as necessary depending on the inspection results. Although other cartridge plate and torque switch fasteners were found slightly loose (for example, by a 1/4 turn), none were significant enough to impact proper valve operation. The plant reviewed the industry standard maintenance practices for motor operated valves and enhanced the Pilgrim procedures to check the fasteners during preventive maintenance checks and to apply loctite on the fasteners to ensure they remain secure.

[Source: W. Raymond (Senior Resident Inspector, NRC) et al., Inspection at Pilgrim Nuclear Power Station conducted June 29, 2003 - September 27, 2003, Inspection Report 05000293/2003007, November 6, 2003]

August 29, 2003

Undersized air-operated valve actuators has been a common problem in nuclear plants, sez Davis-Besse

Here's part of the section titled "Apparent Cause of Occurrence" from the Licensee Event Report (LER 50-346-2003-001-01) describing the finding that eight valves a Davis-Besse were not capable of performing their intended safety functions for all required conditions:

"Lessons learned from the nuclear power industry's motor-operated and air-operated valve programs indicate that AOV performance can be enhanced by improvements in valve and actuator sizing, setting, testing, and maintenance. It was found that during the original procurement cycle, many AOV actuators were undersized. This was a result of vendors being provided with inaccurate system conditions in combination with less than conservative sizing methodology used at the time, and a lack of formal calculations supporting the design basis and appropriate settings for AOV actuators. There was also the practice of sizing AOV actuators with minimum built-in margin. Similar analytical deficiencies resulted in the design of the air accumulators, used to provide a source of motive power in the event of a loss of non-safety related instrument air, not being sufficient to ensure the valves would perform their intended safety function under all design conditions. This apparent cause applies to valve CC1495."

Other causes were identified for the other seven valves. The full LER is available as a pdf.

August 18, 2003

MSIV LLRT approach - potential generic issue at BWRs

NRC inspectors questioned Columbia Generating Station's practice of using instrument air to close main steam isolation valves (MSIVs) before local leak rate testing (LLRT). The instrument air system provides more pressure than the safety-related air accumulators that serve as design basis for MSIV operation, so the seal tightness conditions being tested weren't the same conditions desired to be tested. Calls to five other BWRs revealed that none of them actually tested the design basis conditions. Columbia's MSIVs did pass proper test when performed. For more info, see Columbia - MSIV LLRT had never been done right, and error may be pervasive at BWRs.

* Columbia - MSIV closure tests since 1989 used inadequate GE SIL instead of required ASME method [note: this is different than the LLRT story described above]