05000219/LER-2006-002
| Docket Number06 20 2006 2006 - 002 - 01 09 22 2006 | |
| Event date: | 06-20-2006 |
|---|---|
| Report date: | 09-22-2006 |
| 2192006002R01 - NRC Website | |
Unit Condition Prior to Discovery of Event:
Oyster Creek Generating Station was in REFUEL mode when the MSIV surveillance test was performed on February 3, May 6, and May 9, 2006. The plant was in RUN mode between February 5 and May 6, 2006.
There were no structures, systems, or components out of service that contributed to this event. The actual discovery date of the failed surveillance tests was June 20, 2006.
Description of Event:
The surveillance data for MSIV NSO4A using the stopwatch surveillance timing method was 3.6 seconds in both 1F09 and 1F10. This stopwatch time is acceptable as compared to the minimum acceptance criteria of 3.5 seconds. The Operator performing the test step to document the PPC timed surveillance data in 1F09 incorrectly recorded the data as 13:28:19 to 13:28:22 on the test document. A review of the PPC data files shows that the times should have been 13:28:19 to 13:28:21. The result of the PPC time calculation was originally recorded as 3 seconds in the test procedure. The correct calculation based on actual PPC data should have indicated 2 seconds. The plant was started up from the 1F09 outage in February without identifying that MSIV NSO4A had failed the surveillance test and was inoperable.
The PPC full closure timing for the test performed during 1F1 0 in May 2006 resulted in a time of 2 seconds. At this time the valve was declared inoperable, the operator was adjusted, and the valve was retested to pass the surveillance test.
Analysis of Event:
The stroke time is measured two ways in the MSIV surveillance test. The first is by stopwatch from the time the control switch is positioned to CLOSE until the closed indication light comes on. The second method is by recording the PPC message time for the valve's open indication and the valve's closed indication, and then subtracting to get valve closure time. The PPC points come from the 5 percent closed limit switch and the 95 percent closed limit switch. The data taken and methodology used for obtaining the data during the test only provided 1-second time resolution.
The surveillance data for NSO4A using the stopwatch method was 3.6 seconds in both 1F09 and 1F10.
The Operator performing the test step to document the PPC surveillance data in February incorrectly recorded the data as 13:28:19 to 13:28:22 on the test document. A review of the PPC files shows that the times should have been 13:28:19 to 13:28:21. The result of the time calculation was originally recorded as 3 seconds in the test procedure. The correct calculation based on actual PPC data should have indicated 2 seconds.
The PPC surveillance data for the as-found test in 1F10 was listed as 17:31:32 to 17:31:34 and calculated to be 2 seconds. The PPC has 0.1-second time resolution data available if it is saved within 26 hours3.009259e-4 days <br />0.00722 hours <br />4.298942e-5 weeks <br />9.893e-6 months <br />.
Otherwise the permanent data storage only contains 1-second data. The PPC 0.1-second resolution data was saved for the 1F10 performance and it verified the time to be 2.7 seconds. No PPC 0.1-second data was saved for the 1F09 performance. Since the actual data recorded per the procedure for both methods were the same in 1F09 and 1F10 (3.6 seconds for switch-to-light and 2 seconds for light-to-light), it is assumed the actual PPC 0.1-second time is also essentially identical.
The Oyster Creek Technical Specification, Section 4.5.F.1, states "All automatic primary containment isolation valves shall be tested for automatic closure by an isolation signal during each REFUELING OUTAGE and the isolation time determined to be within its limit." Specifically, for the Main Steam Line isolation valves the isolation time has been specified as: a 3 seconds and s 10 seconds. Therefore, a full stroke time (to isolation) is the Technical Specification requirement.
The 2.7-second stroke data comes from the 5 percent closed (95% open) and 5 percent open (95% closed) limit switches and only indicates the time to travel 90 percent of full valve stroke. Since the 2.7 second stroke was measured across only 90 percent of the stroke and conservatively assuming a uniform rate of travel through the entire stroke, the following calculation can be performed to represent the time from start of movement until end of movement (i.e., full stroke): divide 2.7 seconds by 90 percent, which provides a full stroke time to close of 3.0 seconds.
2.7 seconds� =� 3.0 seconds 90% of full stroke� 100% of full stroke Performing a check of the "as-left" location of the open and closed indication switches from the work order performed in 1R20 (October 2004) obtains additional margin. This shows that the actual timed stroke represents an 84 percent stroke, which would provide additional margin when calculating the minimum full stroke times. Using the same calculation method described above, the following results can be obtained:
2.7 seconds� = 3.214 seconds 84% of Full Stroke� 100% of Full Stroke The 3-second stroke time Technical Specification limit is the minimum allowable time from the beginning of valve movement until the valve is fully closed. The minimum closure time is an assumption in the Reactor Pressure Vessel (RPV) overpressure protection safety analysis. The overpressure protection system is designed to accommodate the most severe pressurization transient, which involves complete, simultaneous closure of both steam lines. Evaluations have shown that the MSIV closure event is the most severe when the MSIV position indication Reactor Protection System (RPS) input is assumed to fail (MSIVF event). This event is analyzed each refueling cycle to ensure that the primary system design pressure is not exceeded. Three seconds is the MSIV stroke time typically used for this analysis.
A review of a similar issue at Quad Cities (documented within NER QC-06-0646) indicates that MSIV stroke time can change (slows when its cold) by as much as 0.3 seconds between testing while the valve is cold vs. testing while the valve is hot. Cold, for purposes of this testing has been interpreted as less then 115°F, for instance during final testing of the MSIV toward the end of a refueling outage, when the valve has cooled. In the case of this specific valve, NSO4A, it is an outboard valve located in the Trunnion Room. MSIV testing during 1F09 was performed at the front end of the Outage. The room temperature was confirmed to be greater than 126°F on 1/29/06 via the CORA database (Operator Round Data). By the end of the outage, (02/05/06), while the temperature had dropped to 105°F, by the day of the test, 02/03/06, the room and the valve could still be considered "essentially hot", that is at or about 115°F. With this consideration, no additional temperature correction is needed. The temperature at which the valve can be considered "essentially hot" was established by testing and documented within a technical evaluation performed for the Quad Cities NER (NER QC-06-046, EC# 0036062). Another 0.1 seconds of speed increase has been attributed to steam assist when closing. The tested valve did not have steam to assist it in closing, so this increase will need to be considered. Conservatively assuming the valve cooled somewhat and subtracting another 0.1 seconds from the valve stroke time, the margin provided by the actual position of the limit switches, 0.214 seconds, there is still sufficient margin to ensure operability.
This provides sufficient margin for the valve stroke time to be considered successful in meeting its Technical Specification requirement.
In the event that the valve temperatures were as high as projected (since there was no temperature element installed on the actuator), the 3.0-second acceptance criteria may not have been met. In that case, a review was performed of an OC Cycle 19 Technical Evaluation, "Evaluation of OC Mid-Cycle As Found MSIV Stroke Times" relative to data obtained during the 1F09 Outage. The Cycle 19 evaluation assumed an actual average MSIV closure time (full open to full closed by stopwatch of the fastest MSIV in each steam line assuming a 1.3 second delay between control switch actuation and actual start of valve closure) of 2.6 seconds, which is 0.4 seconds faster than the 3.0 second assumption in the safety analysis. A similar review of the 1F09 data yields an actual average MSIV closure time of 2.85 seconds, which is 0.15 seconds faster than the analysis basis of 3.0 seconds.
The Cycle 19 evaluation estimated that the 0.4 seconds reduction in the MSIV closure time relative to the analysis basis would result in only a minor increase in peak reactor vessel pressure ( within the available margin of 55 psig for the Cycle 19 Over-pressurization Analysis. For Cycle 20 the Over-pressurization Analysis peak pressure is 1335 psig, which yields a 40 psig margin to the vessel pressure limit (1375 psig). The minor increase in peak reactor vessel pressure expected for a 0.15 second reduction in MSIV closure time is well within the available margin for Cycle 20.
The Cycle 19 evaluation concluded that OC operation with the reduced MSIV closure time was acceptable. The 1F09 MSIV closure time is bounded by the Cycle 19 data, and sufficient margins are maintained relative to the reactor vessel pressure limit. Therefore, it can be concluded that OC has operated within the established safety analysis basis throughout operating Cycle 20.
Cause of Event:
The cause of Oyster Creek Generating Station starting up from an outage and running with an MSIV inoperable is a transcription error during the performance of the PPC timing section of the surveillance test in February 2006.
Corrective Actions:
The full closure MSIV surveillance timing procedure is being revised to require the operator to print out a PPC alarm summary to determine the closing time from the PPC. This revision will remove the possibility of future transcription errors during the performance of this surveillance.
The surveillance procedure is also being revised to provide guidance to keep the full closure time of the MSIVs toward the middle of the 3 to 10 second technical specification full closure time range. This revision will prevent future surveillance test failures as a result of the valves being maintained near the top or bottom band of the acceptance criteria.
Additional Information:
A. Failed Components:
Main steam isolation valve NSO4A (plant component ID V-1-9) was determined to be inoperable per surveillance test acceptance criteria.
B. Previous Similar Events:
Oyster Creek Nuclear Generating Station has previously failed MSIV full closure surveillance tests. Of these previous failures all were identified in a timely manner and work performed to restore the respective valve's operability. No previous events at Oyster Creek have involved unknowingly starting the plant up with an MSIV inoperable.
C. Identification of Components Referred to in this Licensee Event Report:
Components� IEEE 805 System ID�IEEE 803A Function Main Steam Isolation Valve (MSIV)� EIIS: SB� EIIC: RV Plant Process Computer (PPC)� EIIS: ID� EIIC: CPU Reactor Protection System (RPS)� EIIS: JC� EIIC: XC-RCT