05000247/LER-2005-002
| Docket Number | |
| Event date: | 2-18-2005 |
|---|---|
| Report date: | 4-19-2005 |
| Reporting criterion: | 10 CFR 50.73(a)(2)(i)(B), Prohibited by Technical Specifications |
| 2472005002R00 - NRC Website | |
Note: The Energy Industry Identification System Codes are identified within brackets {
DESCRIPTION OF EVENT
On February 18, 2005, while in mode 1 at 100' power, engineering analysis determined that the 23 Safety Injection {BO Pump {P} had been inoperable for an indeterminate period between the successful pump surveillance on 12/24/04 and 01/27/05 when gas was vented from the 23 SI pump casing due to gas build-up within the casing. The event was recorded in the IPEC corrective action program (CAP) as CR-IP2-2005-00398.
An evaluation of the cause looked at the sequence of events that occurred following refuel Outage 16 (2R16) which ended November 22, 2004.
- Out-leakage fiom the 24 SI Accumulator began following startup from 2R16 at an approximate value of 0.14 gpm.
- November 21, 2004, CR-IP2-2004-06364 was issued to document that the 24 SI Accumulator was being pressurized every shift and "topped off" every two days. Initially it was thought the leakage was through the packing leak-off of MOV-894D (Accumulator discharge valve) to the Pressurizer Relief Tank (PRT). While troubleshooting, a local drain valve was opened and water was present which supported the initial belief. The packing of 894D was tightened but the Accumulator continued to leak water at the same rate.
- December 1, 2004, CR-IP2-2004-06531 was issued which noted that 24 SI Accumulator continued to leak and Operations was burdened with "topping off" the tank on a daily basis.
- December 10, 2004, CR-IP2-2004-6679 was issued and identified that troubleshooting confirmed that AOV-839H, the 24 Accumulator test valve, was leaking.
- On December 24, 2004 the quarterly surveillance test for the 23 SI pump was satisfactorily run.
- December 30, 2004 it was determined following a detailed troubleshooting plan that intrusive repairs would be needed to terminate the leakage through 839H.
- On January 12, 2005 additional troubleshooting to identify the leak path through 839H to the RWST was initiated to identify the specific check valve providing the leak path back to the RWST. The repair of 839H was rescheduled to the next refuel outage. Recognizing that long-term Accumulator leakage could lead to gas intrusion in the SI System the SI system engineer subsequently, on 1/18/05, had two work orders initiated to perform UT inspections at specified locations in the SI System on a priority basis.
- On or about January 22, 2005 the continuing troubleshooting identified leakage through check valves 858A and B (Safety Injection test line checks) and 849B (23 SI Pump discharge check valve)as the leakage path from the 24 accumulator to the RWST.
- January 26, 2005 UT was performed and discovered gas pockets in the discharge headers of the system. CR-IP2-2005-00370 was issued.
- January 27, 2005 venting of the SI pump suction header and the SI pump casings was conducted. 23 SI pump casing contained gas; 21 and 22 SI pump casings did not contain gas; gas was also vented from the common suction header(there is an ongoing evaluation as to the past operability of the 21 and 22 SI pumps based on the gas vented from the common suction header).
- January 28, 2005 gas was vented from those portions of the SI discharge piping system within the VC which had been determined by UT to contain gas (an evaluation concluded that the gas vented from the SI discharge headers did not pose an operability concern for the discharge piping/supports).
The accumulator inventory is maintained at approximately 650psig through the use of a nitrogen overpressure blanket. This results in nitrogen gas entrainment within the process fluid. As this fluid left the accumulator and transmigrated to the piping system in the area of the safety injection pumps the nitrogen came out of solution. This is due to the fact that this piping system is normally maintained only under static head pressure. An evaluation of the past operability of the SI pumps and system was initiated because the effect of the gas was not readily determinable. This included calculations to estimate the , approximate amount of gas that was vented from the discharge headers, common suction of the SI pumps as well as the 23 SI pump casing. An ongoing evaluation has concluded that the 23 SI pump was inoperable for some period of time due to gas build-up within its' casing following the December 24, 2004 successful surveillance test. Evaluations of the gas vented from the SI discharge headers have concluded that the condition did not pose a past operability concern for the discharge piping/supports. Evaluation of the gas vented from the common SI suction header in terms of its affect on the past operability of the 21/22 SI pumps is ongoing.
CAUSE OF EVENT
The apparent cause was the organization was too narrowly focused and overconfident. The focus of all involved station personnel was to identify and repair a leak from 24 SI Accumulator. There was a latent organizational weakness in that gas intrusion from the SI accumulator to the SI pump suction was not considered a credible event.
Contributing to this concern was ineffective use of operating experience in the operability evaluation of the 24 SI accumulator inventory loss.
Operating experience relating to backleakage was not assimilated and acted upon in a timely fashion.
CORRECTIVE ACTIONS
The following corrective actions have been or will be performed under the CAP to address the causes of this event and prevent recurrence.
- The SI piping system and pump casing were vented and continue to be vented on a periodic basis.
- A program for periodic monitoring using ultrasonic testing for gas pockets was initiated.
- A design change was initiated to add a new isolation valve to the test line to terminate the accumulator backleakage.
- A Station Clock Reset was initiated by the General Manager Operations.
- All Shift Managers were briefed on the inadequacy of the operability evaluation associated with CR-IP2-2004-06531.
- A briefing of all engineers regarding the cause of the event as evaluated under CR-IP2-2005-00370.
- Surveillance and operating procedures have been or will be updated to formalize ongoing inspections for possible gas intrusion in both Units SI systems as well as to notify the responsible engineer when it is required to top off accumulators.
EVENT ANALYSIS
The event is reportable under 10CFR50.73(a)(2)(i)(B). The licensee shall report any operation or condition which was prohibited by the plant TS. This event meets the reporting criteria because one or more trains of ECCS were presumed inoperable for greater than the TS allowed completion time of 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> for TS LCO 3.5.2. The inoperable 23 SI pump was due to test line valve leakage resulting in gas build-up rendering one or more trains of ECCS inoperable. The 23 SI pump was successfully tested on 24 December 2004; the 23 SI pump casing and suction piping was vented on 27 January 2005. This period bounds the inoperability window. The Safety Injection Pumps are Pacific (Flowserve) Type JTCH 10 stage diffuser type pumps with opposed impellers and axially split case. Gas bindng of centrifugal pumps is a condition where the pump casing is full of gases or vapors to the point where the impeller is no longer able to contact fluid to function correctly. The impeller spins in the gas bubble but is unable to force liquid through the pump. This can lead to cooling problems for the pump's seals and bearings. Centrifugal pumps are designed so that their pump casings are completely filled with liquid during pump operation. Most centrifugal pumps can still operate when small amounts of gas accumulate in the pump casing, but pumps in systems containing dissolved gases that are not designed to be self-venting should be periodically vented manually to ensure that gases do not build up in the pump casing if a gas intrusion mechanism is present. The pump manufacturer (Flowserve) calculated the internal free volume of the SI pumps to be approximately 1.8 cubic feet (each). Furthermore, the pump manufacturer states that a 20% void fraction inside the pump could cause loss of prime. Therefore, a conservative value for the maximum acceptable gas volume inside the pump would be 0.2 x 1.8 = 0.36 cubic feet per vendor information.
Per calculation IP-CALC-05-00059 Rev. 1, the volume of gas (at system pressure of 23.8 psig) vented from the 23 SI pump casing was calculated to be 2.87 cubic feet. Since the 23 SI pump was operated satisfactorily on 12/24/04 during its quarterly surveillance test without any degraded head or adverse vibration it is reasonable to state that the pump was operable from 11/21/04 (start date of system leakage past 839H) thru the 12/24/04 surveillance test. Subsequent to 12/24/04 gas built up in the pump casing and at some point reached an amount that challenged pump operability. Since the period of time is not known with certainty the time for pump inoperability will be taken as T/2 where T is 34 days(time-frame between 12/24/04 and 01/27/05). Therefore the pump was inoperable for 17 days.
PAST SIMILAR EVENTS
A review of the past two years of Licensee Event Reports (LERs) for events that involved inoperable components that exceeded the TS allowed completion time identified two issues when AOTs were exceeded. LER 2003-002 described an event where a boric acid transfer pump flow path could not be verified due to an improperly installed diaphragm valve. LER 2005-001 described an event where an auxiliary component cooling water pump was determined to be inoperable due to an improperly serviced discharge check valve. However, there are distinctly different causes for these events. The corrective actions associated with these events would not have prevented this particular event from occurring.
SAFETY SIGNIFICANCE
This event had no effect on the health and safety of the public.
There were no actual safety consequences for the event because there were no accidents or transients requiring the SI or ECCS.
There were no significant potential safety consequences of this event under reasonable and credible alternative conditions. A risk evaluation determined that the incremental Core Damage Probability is 7.92E-9 and the incremental Large Early Release Probability is 1.95E-9 for the 17 day inoperability period indicating there was no significant risk to public health and safety. This evaluation is based on only 23 SI pump being adversely affected by the gas intrusion into the SI system. It is noted that past operability evaluations for 21/22 SI pumps are ongoing and that the risk evaluation for this LER assumes availability of these two pumps.