05000260/LER-2009-004
| Browns Ferry Nuclear Plant | |
| Event date: | 06-11-2009 |
|---|---|
| Report date: | 02-12-2010 |
| Reporting criterion: | 10 CFR 50.73(a)(2)(i)(A), Completion of TS Shutdown 10 CFR 50.73(a)(2)(iv)(A), System Actuation |
| 2602009004R01 - NRC Website | |
I. PLANT CONDITION(S)
Prior to the event, Browns Ferry Nuclear Plant (BFN) Units 1 and 3 were operating in Mode 1 at 100 percent thermal power (approximately 3458 megawatts thermal). BFN Units 1 and 3 were unaffected by the event. BFN Unit 2 was in Mode 1 at approximately twelve percent and in power ascension following a refueling outage.
II. DESCRIPTION OF EVENT
A.� Event:
During reactor startup from the BFN Unit 2 Spring refueling outage, a failure of a Main Steam (MS) Line B Safety Relief Valve (SRV) [SB] to fully close was revealed. Steam leakage through this SRV stopped when reactor pressure decreased to approximately 850 psig. The Tennessee Valley Authority (TVA) initially thought the steam leakage was due to pilot valve leakage because of observed discharge tailpipe indications and past experiences with pilot leakage. However, following destructive testing, it was determined to be steam leaking by the main valve body. As a result of this steam leakage, two MS SRV tailpipe vacuum breakers, 2.5 inch and 10 inch, were cycling. This SRV failure and vacuum breaker cycling allowed steam- to enter the drywell instead of going to the torus.
experienced an increase in drywell leakage during reactor startup. The four-hour unidentified leakage from 0800 to 1200 hours0.0139 days <br />0.333 hours <br />0.00198 weeks <br />4.566e-4 months <br /> CDT on June 10, 2009, was 0 gallons per minute (GPM), the four-hour unidentified leakage from 0800 to 1200 hours0.0139 days <br />0.333 hours <br />0.00198 weeks <br />4.566e-4 months <br /> CDT on June 11, 2009, increased to 3.88 GPM. This increase in Reactor Coolant System (RCS) operational leakage exceeded the Technical Specifications (TS) 3.4.4, RCS Operational Leakage, limit of a 2 GPM increase in unidentified leakage within the previous 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period.
Therefore, the TS Limiting Condition for Operation (LCO) 3.4.4 was not met and at 1555 hours0.018 days <br />0.432 hours <br />0.00257 weeks <br />5.916775e-4 months <br /> CDT on June 11, 2009, Unit 2 Operations personnel initiated a manual reactor scram to comply with TS 3.4.4 LCO Condition C, to be in Mode 3 in 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> and to be in Mode 4 within 36 hours4.166667e-4 days <br />0.01 hours <br />5.952381e-5 weeks <br />1.3698e-5 months <br />.
During the reactor shutdown, all automatic functions resulting from the manual scram occurred as expected. All control rods [AA] inserted. No primary containment isolation system (PCIS) [JE] isolations were received.
Subsequently, at 1609 hours0.0186 days <br />0.447 hours <br />0.00266 weeks <br />6.122245e-4 months <br /> CDT on June 11, 2009, a full reactor scram occurred due to Intermediate Range Monitor 'C' spiking high concurrent with the inability to reset Reactor Protection System [JC] (RPS) 'B' scram channel.
Following verification that the 2-A0I-100-1, Reactor Scram, actions were completed, the reactor mode switch was placed in shutdown.
WA is submitting this report in accordance with 10 CFR 50.73(a)(2)(i)(A), as the completion of any nuclear plant shutdown required by Technical Specifications, and in accordance with 10 CFR 50.73(a)(2)(iv)(A), as any event or condition that resulted in manual or automatic actuation of the RPS including: reactor scram or reactor trip.
B. Inonerable Structures. Components. or Systems that Contributed to the Event:
None.
C. Dates and Approximate Times of Maior Occurrences:
June 11, 2009, at 1555 hours0.018 days <br />0.432 hours <br />0.00257 weeks <br />5.916775e-4 months <br /> CDT� Unit 2 reactor manually scrammed.
June 11, 2009, at 1609 hours0.0186 days <br />0.447 hours <br />0.00266 weeks <br />6.122245e-4 months <br /> CDT� Unit 2 full reactor scram occurred due to Intermediate Range Monitor 'C' spiking concurrent with inability to reset RPS 'B' scram channel.
June 11, 2009, at 1724 hours0.02 days <br />0.479 hours <br />0.00285 weeks <br />6.55982e-4 months <br /> CDT� Operations made an Emergency Notification System report in accordance with 10 CFR 50.72(b)(2)(i)(B).
D. Other Systems or Secondary Functions Affected
None.
E. Method of Discovery
The annunciator for Drywell Floor Drain Sump Pump Excessive Operation was received in the Main Control Room.
F. Operator Actions
Operations personnel completed the shutdown as required by Technical Specifications 3.4.4 and entered 2-A0I-100-1, Reactor Scram.
G. Safety System Responses
The RPS logic responded to the manual reactor scram. All control rods inserted. No PCIS isolations were received.
RPS 'B' scram channel did not reset after the manual scram.
III. CAUSE OF THE EVENT
A. Immediate Cause
The immediate cause of the excessive RCS operational leakage was the failure of a MS Line B SRV to fully close. Also, two MS SRV tailpipe vacuum breakers, 2.5 inch and 10 inch, were cycling. This cycling allowed steam to enter the drywell instead of going to the torus.
B. Root Cause
There are two root causes for the excessive RCS operational leakage.
The first root cause of this event was identified as an inadequate original manufacturer-threaded main joint design that develops a fretting condition after years of service of a main body valve on a Target Rock Two or Three Stage SRV. Destructive examination at Wyle Laboratories identified that mating threads on the main valve piston-to-main valve stem were damaged to the point that the shaft appeared to be cocked approximately 1/4 inch, which would prevent the main body from cycling correctly. General Electric (GE) Service Information Letter (SIL) 646 documents this same condition, which previously occurred at Plant Hatch. GE SIL 646 is applicable to Target Rock Two or Three Stage SRVs, and BFN has Target Rock Two Stage SRVs. The design detail dimensioning tolerance for the parts, when originally manufactured and assembled, permitted the lead thread of the piston to prematurely contact the under-cut area between the load bearing shoulder and the final thread on the stem of the main valve disc. In this condition, the torque (or applied preload) between the jam nut and the piston was lost during certification testing of the main valve body. Subsequent vibration of the loosened piston during normal plant operations allows the piston to fret the threads of the stem which can result in mechanical binding of the SRV.
The SRV steam leakage into the drywell occurred because the associated 2.5 inch vacuum breaker was found stuck open and the 10 inch vacuum breaker was found open with the spring mechanism found to be weak. These conditions were determined to be caused by excessive cycling due to the leaking SRV. Steam leakage was flowing down the tailpipe of the MS Line B SRV, through the open vacuum breakers, and into the drywell.
The second root cause of this event deals primarily with the failure to fully implement GE SIL 646 at BFN. Organizational to Organizational Interface Deficiencies were identified as the underlying root cause for failure to fully implement GE SIL 646. The GE SIL 646 recommended actions consisted of inspections and, as needed, interim modifications, which Target Rock would provide, over the next three refueling outages with follow-up inspections in the succeeding 6 to 10 years of service. For installed SRVs and spares, the population of valves to be inspected was recommended to be approximately one-third of the SRVs each outage. The inspection of other SRVs not installed or spares was also prescribed.
Per the cause analysis, previous implementation of GE SIL 646 at BFN consisted of two main parts: generation of work orders for valves outside of the SIL requirement and preventative maintenance (PM) work orders generated on the remaining valves. Review of both parts of the implementation revealed that a breakdown occurred during initial development of the refueling outage scope. The MS System Engineer generated the appropriate documentation to have GE SIL 646 implemented; however, the Valve Engineer identified only one main body to be replaced during refuel outages to coincide with past practices of changing only one main body in any given outage. Therefore, refuel outage work scope development and control were deficient.
The root cause for the failure of the RPS 'B' scram channel to not reset was found to be a loose scram relay/contactor terminal connection. On June 11th at 1602 CDT, the RPS 'B' scram channel did not reset as expected. Investigation found that the 5A-K14H scram relay/contactor was not re-energized due to a loose connection block. The connection was tightened and the relay/contactor operated correctly. This relay had been recently operated in support of the BFN Unit 2 refuel low power startup. Failure analysis has identified that the friction connection (pressure pad) of the power feed connection block was loose with the likely cause of less than adequate tightening during coil replacement two weeks prior to the event.
C. Contributing Factors
A significant component of the unidentified RCS operational leakage was a packing leak on the Reactor Vessel Drain Valve. The packing leak was caused by ineffective maintenance.
IV. ANALYSIS OF THE EVENT
During reactor startup from the BFN Unit 2 Spring refueling outage, a failure of a MS Line B SRV to fully close was revealed. Steam leakage through this SRV stopped when reactor pressure decreased to approximately 850 psig. TVA initially thought the steam leakage was due to pilot valve leakage because of observed discharge tailpipe indications and past experiences with pilot leakage. However, following destructive testing, it was determined to be steam leaking by the main valve body.
The GE SIL 646 failure mode is described as loss of torque at the main valve piston-to-main valve stem threaded joint due to deformation of the leading edge of the piston threads. The failure mode occurs when the leading thread edge of the piston prematurely contacts the under- cut area between the load bearing shoulder and final thread on the stem of the main valve disc.
In this condition, the torque (or applied preload) between the jam nut and the piston was lost during certification testing of the main valve body at a limited steam supply test facility. The loss of torque condition is undetectable without disassembly of the certified main valve body. When the main valve body is installed on the steam line header, the steam flow-induced vibration allows the piston to fret the threads of the stem. If the main valve body is subjected to this degradation process long enough, the entire threaded joint is compromised. The alignment between the piston and cylinder cannot be maintained when the SRV is required to open which can result in the mechanical binding of the SRV. The mechanism by which the main valve body opens is identical for both the mechanical and electrical opening mode. Therefore, the condition that resulted in the failure of 2-PCV-1-23 is applicable to both the mechanical and electrical operating modes of the SRV. The valve was installed at the 2-PCV-1-23 position since April 1999 and had not been inspected per the GE SIL 646 recommendations.
With respect to the vacuum breaker cycling, typically, these vacuum breakers do not cycle under normal plant conditions. The vacuum breakers were cycling due to the leaking 2-PCV-1-23.
Normally, during a transient situation, the vacuum breakers could potentially cycle once and each time the SRV opened and closed. Since this was a unique event, the vacuum breakers cycled continually during the time of the leaking SRV or approximately 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br />.
V. ASSESSMENT OF SAFETY CONSEQUENCES
The safety consequences of this event were not significant. The manual scram was not complicated. The operational impact was manageable during these valve failures. Operations reset the reactor scram at 1602 hours0.0185 days <br />0.445 hours <br />0.00265 weeks <br />6.09561e-4 months <br /> CDT.
Two cases identified in the industry where the SRV failure mechanism has occurred resulted in valves that either did not open or only partially opened. BFN Updated Final Safety Analysis Report (UFSAR) Section 14.5, Analysis of Abnormal Operational Transients — Uprated, includes various analyses, which result in a scram of the reactor from low power operation. Included are analyses for transients and accidents with only 12 of the 13 SRVs available for opening and for the inadvertent opening of a MSRV at limiting conditions, beginning of core life at rated core flow conditions. These analyses bound this actual condition or event. Further, based on the infrequent occurrence of this type of valve failure and the lack of a history of the GE SIL 646 failure mechanism from past inspections of BFN MS SRVs, there is a high confidence that the installed SRVs will perform their safety function.
With the exception of the RPS failure to reset, all safety systems operated as required during the manual scram. As expected, there were no PCIS Group 2, 3, 6, or 8 isolations. Although the Emergency Core Cooling Systems were available, none were required. No MS SRVs [SB] actuated. The turbine bypass valves [JI] maintained reactor pressure. The main condenser remained available for heat rejection.
Reactor water level was recovered and maintained by the reactor feed water [SJ] and condensate [SG] systems. Therefore, TVA concludes that there was no significant reduction in the protection of the public by this event.
VI. CORRECTIVE ACTIONS
A. immediate Corrective Actions Operations performed the immediate actions of operating procedure "Relief Valve Stuck Open." The immediate actions were to identify the stuck open SRV by observing Safety Relief Valve Tailpipe Flow or Main Steam Relief Valve Discharge Tailpipe Temperature. Operations attempted to close the MS SRV, but it still indicated partially open and a work order was initiated.
B. Corrective Actions to Prevent Recurrence - The corrective actions to prevent recurrence are being managed by BFN's corrective action program.
The corrective actions to prevent recurrence are to complete an inspection and refurbishment of all affected main body valves installed on Units 2 and 3 at BFN in accordance with the GE SIL 646 recommended action. TVA will fully implement these recommendations.
The second root cause corrective actions are to revise procedures to ensure appropriate PM work orders are scheduled and to require additional rigor and documentation to initial outage scoping. A training needs analysis will be performed to determine training needs with regards to engineering responsibility for outage scope.
For the extent of condition evaluation, corrective actions also include performance of a review of previous, associated corrective action documents to determine if applicable GE SILs and GE Technical Information Letters (TILs) were appropriately implemented at BFN.
To address the relay/contactor loose connection, periodic verification of coil power termination tightness was added for each relay/contactor being inspected in the procedures.
VII. ADDITIONAL INFORMATION
A. railed Components Failed components are a MS Line 'B' SRV, associated vacuum breakers, and Reactor Vessel Drain Valve stem packing.
B. PREVIOUS LERS ON SIMILAR EVENTS
None.
C. Additional Information
Corrective action documents for this report are Problem Evaluation Reports '173480, 174037, and 174044.
D. Safety System Functional Failure Consideration; This event is a not a safety system functional failure in accordance with NEI 99-02.
E. Scram With Complications Consideration; This event was not a complicated scram according to NEI 99-02.
VIII. COMMITMENTS
None.