IR 05000249/1996008
| ML17187A489 | |
| Person / Time | |
|---|---|
| Site: | Dresden  |
| Issue date: | 06/14/1996 |
| From: | Axelson W, Hiland P NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | |
| Shared Package | |
| ML17187A488 | List: |
| References | |
| 50-249-96-08, 50-249-96-8, NUDOCS 9606250249 | |
| Download: ML17187A489 (24) | |
Text
U.S. NUCLEAR REGULATORY COMMISSION REG ION I I I Docket No:
License No:
~eport No:
Licensee:
Facility:
Location:
Dates:
Inspectors:
9606250249 960614 PDR ADOCK 05000249 G
PDR AUGMENTED INSPECTION TEAM 50-249 DPR-25 50-249/96008 Commonwealth Edison Company Dresden Nuclear Station Unit 3 Opus West III 1400 Opus Place - Suite 300 Downers Grove, IL 60515 May 16 through May 23, 1996 P~ Hiland, Team Leader; Chief, Reactor Projects Branch 1 J. Hopkins, Regional Project Engineer D. Butler, Senior Reactor Insp~ctor (Electrical)
P. Louden, Senior Radiation Specialist J. Guzman, Senior Reactor Inspector (Mechanical)
D. Roth, Resident Inspector, Dresden J. Stang, Senior Licensing Project Manager, NRR P. Hiland, Team Leader, Chief, Reactor Projects, Branch W. Axelson, Direc or, Division of Reac or Projects
SUMMARY On May 15, 1996, the failure of a Unit 3 feedwater regulating valve (FRV)
resulted in a complete loss of feedwater to the reacto The reactor automatically tripped and emergency systems, such as the high pressure coolant injection (HPCI) and Group 1 primary containment isolation, actuated as designe Following the initial transient, control room operators reset the Group 1 isolation signal in order to restore the main condenser as an alternate heat sin When the Group 1 isolation signal was reset, an inboard main steam isolation valve (MSIV) and a recirculation sample isolation valve (220-45), unexpectedly opened due to a failed relay in each of the valves'
control circuit The operators manually reclosed the valves and reverified the other Group 1 primary containment isolation system (PCIS) valves had remained close After the two Group 1 isolatiori valves unexpectedly opened, an Unusual Event was declared and the emergency plan was activated due to the potential degradation in the level of plant safet The plant was cooled down using the is6lation condenser and the shutdown cooling syste The Unusual Event was terminated after the plant was in Cold Shutdown and reactor coolant chemistry samples verified fuel integrit The AIT concluded that the control room operators performed the appropriate immediate actions, stabilized reactor water level and pressure, and placed the plant in ~ shutdown condition in accordance with plant procedure Observations of control room activities during the event noted that operators monitored control room indications, supervisors maintained proper command and control, and 3-way communications were use The licensee formed a "Task Force" to review the event, to identify the root causes of the equipment malfunctions (FRV and the Group 1 relays), and to develop immediate and long term corrective actions for the failure The Task Force initially appeared to be narrowly focused on repairing the failed equipment and correcting specific causes for the failure Additionally, some members of the Task Force were initially involved with both the repair efforts and the root cause investigatio However, licensee management recognized that a larger team with broad and distinct responsibilities was needed and subs~quently assigned additional resources to the Task Forc The failure mechanism for the loss of feedwater transient was determined to be a stem to disk separation of the in-service feedwater regulating valve (FRV 38) due to high-cycle, low-stress fatigue cracking due to flow induced vibratio The primary root cause was the flow induced vibrations which may not have been accounted for in the design of the ste The unexpected opening of the main steam isolation valve and the reactor recirculation sample isolation valve resulted from a failed relay in each of the valves' seal-in logic control circui The relays were mechanically bound due to an interference fit of the phenolic block at the armature support bracket pivot point The dimensional variations could have occurred during the manufacturing proces The licensee's engineering root cause determination teams were effective in investigating and identifying the technical problems related to the feedwater valve failure and the PCIS control relay failure The licensee provided sufficient engineering resources and developed a multi-discipline team consisting of site and corporate engineering and technical specialist Assembly and disassembly of the feedwater valve were well controlled and new information was methodically evaluated. Assumptions made by the feedwater valve failure investigative team were challenged and verified to be reasonabl In addition, a final independent review of the feedwater valve failure and the proposed corrective action was performed by two technical design engineering consultant Throughout the PCIS relay testing process, an approved test p~an was used and the instructions were written in_a logical manner that preserved the as-found condition of the relay The inspectors concluded that the failure mechanisms and the root causes were methodically determined and were technically founde Overall, the inspector's concluded the licensee's investigation was self critical and intensiv The root cause determinations were reasonable and technically soun The systematic approach and depth of the investigation assured appropriate corrective actions were develope However, the inspectors also concluded that the engineering modification backlog contributed to the delays in completing planned modifications to the 38 FR In addition, the iri~pectors concluded the station's maintenance backlog contributed to the delays in repairing the isolated 3A FRV and in the missed opportunities to identify a generic trend in the failed PCIS control relays in 1994 and 199 The licensee's decision io isolate the 3A.FRV two weeks prior to the even focused primarily on the existing steam leak and personnel protection and less on the relationship of the FRV to the reacto The licensee placed the plant in a configuration that required ECCS initiation following the failure of the 38 FR The inspectors concluded that the decision to operate on a single FRV for an extended period of time was non-conservative and resulted from unresolved material condition issue *
REPORT DETAILS Purpose of the Augmented Team Inspection Following initial review of the May 15, 1996, loss of feedwater transient to the Dresden Unit 3 (U-3) reactor, the NRC formed an Augmented Inspection Team (AIT) to examine the circumstances surrounding the even The AIT Charter included evaluations of plant equipment performance, operator response to the event, the effectiveness of the licensee's root cause investigation, and the effectiveness of the corrective actions. The AIT Charter is.included as Attachment 1 to this repor.0 Event Description [Charter Item No. l]
The Event Description and Sequence of Events were independently developed and validated by the inspectors using the following information:
Review and evaluation of control room instrument chart recorders; plant alarm and data print outs; and log books from the nuclear station operators (NSO), unit supervisors (US), and shift manage *
Interviews with personnel directly involved in the even *
Observations by onsite NRC inspectors at the time of event occurrenc *
Review of the licensee's investigation and time lin On May 15, 1996, at 10:43 a.m. (CDT), a valve failure (the plug separated from the stem) on the U-3 38 feedwater regulating valve (FRV).
resulted in a complete loss of feedwater to the reacto The subsequent automatic reactor trip and emergency core cooling systems (EC£S)
actuation occurred as designe (See Section 3.1 for the root cause o~
the FRV failure.)
Prior to the event, U-3 was operating at 82 percent of full power (677 megawatts electrical (MWe)) and had been on-line since November 11, 199 Normal. offsite power and all emergency equipment were operabl Routine surveillance testing on the.U-3 containment cooling service water (CCSW) system was in progress. The 3A FRV and the associated inlet isolation valve (M0-3206A) had been closed since April 27, 1996, due to a body-to-bonnet leak on the 3A FR The 38 FRV was controlling reactor water level in single element control at the normal operating level (about +30 inches (in.)).
After the 38 FRV failed, all feedwater flow to the reactor was blocked and water level rapidly dropped to the automatic low level scram setpoint (+17 in.). All control rods fully inserted and all other equipment and isolation signals operated as designe Reactor water level continued to decrease until the Low-low water level setpoint was reached (-51 in.). At this point, the high pressure coolant injection (HPCI) system started, main steam isolation valves (MSIV) closed (Group 1), and other safety equipment automatically functioned as designe *
Reactor water level continued to decrease until the HPCI and the control rod drive (CRD) systems began to restore vessel leve The lowest level reached was about -59 in., which was 84 in. above the top of active fue As water level increased, control room operators reduced HPCI flow to control vessel water level and reactor pressur The isolation condenser was manually placed in service to control reactor pressure and to commence a plant cooldow At 11:11 a.m., the Groups 1, 2, and 3 isolation signals were reset in order to restore the main condenser as an alternate heat sink (if needed) and the reactor water cleanup (RWCU) system for level contro When the signal was reset, the IA MSIV (inboard) and the 220-45 recirc~lation sample isolation valve, both Group 1 valves, unexpectedly*
reopene The cause for the valves to reopen was later determined to be failed relays in the valves' control circuit The operators manually reclosed the two valves and reverified the other Groups 1, 2, and 3 isolation valves were close (See Section 4.1 for additional information on the relay failures.)
After the two Group 1 isolation valves une~pectedly reopened, the licensee declared an Unusual Event (11:22 a.m.) and activated the emergency plan due to the potential degradation in the level of plan safet All notifications were performed as require The licensee continued the plant cooldown using the isolation condenser and the shutdown cooling (SOC) syste The Unusual Event was.terminated at 11:08 p.rn. after the plant was in Cold Shutdown and reactor coolant che~istry samples verified fuel integrit A detailed Sequence of Events is included as Attachment 2 to this report.. Failure of 3B Feedwater Regulating Valve [Charter Item No. 7 and No. 8] Root Cause Investigation The inspectors' assessment of the licensee's root cause investigation and Engineering Department's response was based on the following:
Observing portions of the 3B FRV valve and actuator disassembl *
Inspecting the failed valve internal *
Observing portions of the 3B FRV reassembl *
Observing the licensee's root cause team discussions and meeting *. Reviewing the material analysis failure report, the root cause investigation final report, the design documents, and applicable calculations evaluating stern fatigue for short term operatio Based on the laboratory material analysis of the stem's fracture surface and the as-found condition of the valve, the licensee concluded that the tailure mechanism was low-stress, high-cycle fatigue cracking that originated from bending load These loads initiated and propagated two cracks on the ste The inspectors reviewed the material analysis with regional specialist and concluded the analysis was technically soun *
The primary root cause that initiated the stem cracks was flow induced vibrations that were not necessarily accounted for in the design of the ste Although the valve plug was within vendor clearance tolerances, based on wear indications on the plug's o~ter circumference, the plug was moving within the c~ge in a translation and angular motio This flow induced vibration and the loads imposed on the stem may have been aggravated when the 3A FRV was taken out of service two weeks prior to the even With total system flow passing through the 3B FRV, the potential existed for excessive localized vibration of the valve plu Another possible contributor was a previously unnoticed eccentricity between the stem and plug of approximately 0.006 inch (the stem was not perfectly concentric with the plug). Analytically, this eccentric connection contributed little. to the bending moments; however, it contributed to the overall loads at the stem and plug interfac.2 Corrective Actions The licensee performed an analysis and determined that a like-for-like stem replacement was acceptable for the 6 months of operation remaining in the fuel cycle. A liquid penetrant test was performed on the replacement stem and no indications were identifie The licensee imposed an operational limit of 25 percent open on the 3B FRV to minimize valve stresses. Also, vibration monitoring of the valve was planned to identify the flow level at which the greatest frequency levels were reache Finally, the licensee planned to modify the 3B FR~
during the upcoming refueling outage in September 199 The analysis performed was a stress calculation to determine the induced stresses by the applied loads from the plug and guide to the valve ste Based on this stress level, the number of cycles needed to fatigue the stem material to failure was predicted from American Society o Mechanical Engineers (ASME) Stress versus Number of Cycles" graphs for the stem materia Based on the number of cycles to failure and the expected maximum operational time of 6 months, a maximum allowable frequency was calculated to analytically demonstrate that the Valve stem will not reach its fatigue life in 6 month Overall, the inspectors concluded that the licensee's short term corrective actions provided reasonable assurance that the 3B FRV woul.d operate as designed until the upcoming refueling outag.3 Engineering Response After the failure of the 38 FRV, the licensee quarantined the valve and formed a Task Force to review the event, to identify the root causes of the equipment malfunctions (such as the FRV and the Group 1 relays), and to develop imm~diate and long term corrective actions for the failure Senior licensee managers were assigned to provide dedicated oversight of the Task Forc The engineering root cause team was effective in investigating and identifying the technical problems related to the valve failur The licensee provided sufficient engineering resources and developed a multi-discipline team consisting of site and corporate engineering and technical specialists. Assembly and disassembly of the valve were well controlled and new information was methodically evaluate Assumptions
made by the investigative team were challenged and included a final independent review of the evaluation and of the corrective actions by two technical design engineering consultant The inspectors concluded that the failure mechanism and the root cause were methodically determined and were technically founde.4 Licensee's Decision to Operate with the 3A FRV Closed The inspector's assessment of the licensee's decision to operate U-3 with total system flow passing through the 38 FRV was based on the following reviews:
Licensee's engineering evaluation to close the 3A FR *
Updated Final Safety Analysis Report (UF_SAR).,..
_
___......
Plant operating procedure *
Feedwater system modification *
Feedwater system lesson plan *
Control room operation log *
Interviews with operations personne The body-to-bonnet steam leak on the 3A FRV was initially identified in September 199 Repair parts were ordered, but were not available, when the plant was shut down twice in October 199 During November 1995, the plant implemented a temporary modification to inject sealant to control the leakag This sealant controlled the leakage but was not fully successful. Since November 1995, six sealant injections were made in an attempt to further control the leakag Further, during planned load reductions in March 1996, the licensee again decided to continue sealant injections instead of repairing the leak, even though the materials for the repair were on sit On April 27, 1996, the 3A FRV was closed and removed from service due to increased body-to-bonnet leakag The plant was licensed to operate in this configuration in accordance with the UFSAR arid plant operating procedure Both the UFSAR and the procedures identified that feedwater level control with only the 38 FRV may produce flow oscillations due to irregular feedwater piping configuration Section 10.4.7.2.1 of the UFSAR stated that 100 percent rated flow through a single FRV was not recommended since flow induced vibrations increased to unacceptable levels above 700 MW This statement was considered during the licensee's decision to operate with the 3A FRV close The licensee confirmed that the 38 FRV was passing less than 100 percent rated flow with power initially at 712 MW Additionally, engineering monitored and evaluated the vibrations after the feedwater system was operated in this configuratio The inspectors agreed with the licensee's conclusion that the decision to isolate the 3A FRV focused primarily on the leak and personnel protection and not enough on the relationship of the FRV to the reacto The licensee placed the plant in a position that required ECCS initiation following a failure of the 38 FR The inspectors concluded that the decision was non-conservative and was influenced by the unresolved material condition issues and the maintenance backlog on the feedwater syste Further, the inspectors concluded that the decision to isolate 3A FRV did not fully consider the potential impact on risk as
described in the licensee's individual plant evaluatio Specifically, the feedwater system was a "key system" in minimizing core damage frequency, and operating with a single FRV increased the probability of system failur.5 Historical Review of the FRVs and the Feedwater System The inspectors review of the history of the FRVs and the feedwater system was based on the following:
Previous modifications.
Valve manufacturer technical information.
Valve and actuator corrective and preventive maintenance work reques ;_.,,.,Component failure history information and corrective actio *
Industry failure informatio *
Valve design documents and applicable calculation. Copes Vulcan FRV The original design of the U-2 and U-3 FRVs was a Copes Vulcan series D-100 valve with Copes Vulcan internals (stem, plug, cage, and guides).
Valve stem related problems were noted with Copes Vulcan valves throughout the nuclear industry especially during the 1980' The failure mode for most reported failures was related to retaining roll-pin cracks or loo~ening of a holding nut that allowed the plug to work fre Various modifications recommended by Copes Vulcan appeared to have been effective since very few failures of these types were reported since the early 1980' In 1987, a Dresden Unit 2 FRV plug separated from the stem after being in service for approximately 30 month (The valve stem separated at approximately the same location as the 38 FRV in 1996.)
The 1987 failure mechanism was attributed to fatigue, but no further review was performed to identify the root caus [Note that an AIT was formed to review the 1987 event (Inspection Report 50-237/249-87029).]
The stem and plug of the 2A FRV were replaced and welded together. Other corrective action included replacement of the all of the original Copes Vulcan FRV's internals with a modified trim (cage and plug assembly) to minimize vibration and fatigu The Dresden Unit 3 38 FRV internals*
were modified in 198 In late 1991, based on the failure of similar type of Copes Vulcan valve at a different utility, the licensee inspected the FRVs and found cracking on the 38 FRV plug adjacent to the retaining pi The licensee replaced the stem and modified the stem-to-plug connection (completed in March 1992).
A schedule to inspect the valve stem every second refueling outage was develope The 38 FRV was scheduled to be inspected in the Fall 199 The licensee could not determine the basis for choosing a two outage inspection interva Since the repair, the 38 FRV had been in service approximately 32 month The inspectors concluded that the licensee had not utilized available information to fully evaluate the proper inspection interval for the modified 38 FRV ste The history of Copes Vulcan FRVs at Dresden was
well documented, including in the 1987 AIT report, and the 3B FRV vibration problems were clearly described in the station's FSA However, no detailed engineering analysis was performed to support the inspection interval determinatio.5.2 Feedwater System The review of the feedwater system identified concerns with flow induced vibrations believed to be caused by irregular piping configurations throughout the syste Interactions between the feedwater control system and the FRVs, and the system's hyd~aulic instabilities have historically been a concer A part of the licensee's 19ng term corrective actions for the 1987 failure of the 2A FRV was a comprehensive review of the feedwater syste The review identified the need for system-wide, modification The modifications were done slowly until 199 However, since 1994, the modifications were more comprehensive and included additional valve internal replacements, valve operator replacements, and control and process system improvement At the time of the M~y 15 event occurrence, the licensee was in the final stages of these modification As noted earlier, the valve internals for the 3B FRV were scheduled to be replaced during the Fall 1996 outag While significant progress was made in improving the feedwater system since 1987, the modification to the 3B FRV was still outstanding 9 years late The 3B FRV stem had not been inspected for cracks since 1992 due to extended outage The inspectors concluded that the licensee's modification backlog contributed to the event initiatio.0 Failure of the Group 1 Primary Containment Isolation System (PCIS) Valve Relays [Charter Item No. 5 and No. 9] The inspector's assessment of the licensee's root cause investigation, corrective actions, and the Engineering Department's response was based on the following:
Observing removal of two failed HGA relay *
Observing the licensee's root cause testing performed by relay specialist *
Reviewing the electrical schematics for all U-2 and U-3 Group 1 valve *
Reviewing the U-2 relay operability evaluatio *
Reviewing the HGA relay dedication packag *
Reviewing the post maintenance verification test for the reinstalled relay *
Reviewing the HGA relay failure history.
Design The purpose of the PCIS was to provide automatic isolation of certain systems which penetrate the primary containment whenever setpoints were exceede The Group 1 isolation consisted of eight MSIVs, two main
steam line drains, two isolation condenser steam line vents, and two recirculation loop sample valves. Whenever a setpoint was reached, the Group 1 isolation valves automatically close The circuit design used
a control relay to provide a seal-in feature to prevent a valve from reopening following reset of the Group l isolation signa When resetting the Group l isolation signal, these control relays should remain de-energized with the armature contacts "dropped-out" until the valves were manually reopene The licensee determined that the seal-in relays for the two valves that unexpectedly reopened were de-energized but the armature contact had not "dropped out." With the armature contacts still closed, the two valves automatically reopened when the Group I isolation signal was rese.2 Root Cause Investigation When the Group I isolation signal was reset, the IA MSIV and thi 220-45 recirculation sample isolation valve* unexpectedly reopened due to failed relays in these two.valves' control circuit The operators manually reclosed the two valves and reverified the other Groups l isolation valves were close The licensee quarantined the.two control relay *
cabinets to preserve the as-found condition of the relay The safety consequences of the valves reopening were mi~imal because redundant valves remained close The relays used to perform the seal-in function for the Group l alternating current (AC) operated solenoid valves were General Electric (GE) model number 12HGA17S6 The HGA relays were "commercial grade"-
prqducts that had been dedicated for safety relat~d application At Dresden, model 12HGA17S63 relays were only used in the PCIS application, Twelve relays*of this type were used in each Uni Both of the failed relays had been in service since 199 One failed relay from the IA MSIV and one relay from a valve circuit that had operated properly were removed and taken to Commonwealth Edison's Central Receiving, Inspection and Testing Facility (CRIT) for
- root cause analysi The CRIT team identified the failed relay had mechanical binding of the moveable phenolic contact finger carrier block at the armature support bracket pivot point The balanced armature was hinged at the support bracket pivot points and was supposed to freely move at this point to insure good contact wip The good relay's armature freely pivoted when mechanically manipulate Disassembly of the failed relay required a screwdriver to pry the phenolic block from the armature plate and armature support bracke The hinged armature plate freely pivoted around the pivot points once the phenolic block had been remove Critical relay part dimensions were measured by the CRIT tea Inside dimensional measurements of the failed relay phenolic block were found to be smaller than the good relay. In addition, the armature support bracket pivot point dimensions of the failed relay were larger than the good rela The CRIT team concluded that manufacturing tolerances may have caused an interference fit at the pivot point The CRIT team reassembled the failed relay and was able to recreate the mechanical bindin The licensee concluded that the two Group 1 relays were mechanically bound due to an interference fit of the phenolic block at the armature support bracket pivot point The dimensional variations could have
occurred during manufacturing, such as the mold used to manufacture the phenolic block and/or stamping of the armature support bracke The inspectors concluded that the licensee's cause determination was reasonabl.3 Corrective Actions.5 The licensee initiated prompt corrective actions to prevent unexpected valve repositioning upon resetting a Group 1 isolation signal. Caution tags were placed on each unit's Group 1 Main Steam Isolation Reset switches with instructions to manually close all Group 1 isolation valves prior to resetting a Group 1 signa The inspectors reviewed the operability evaluation for the Unit 2 HGA relays and the electrical schematics for all Unit 2 and Unit 3 Group 1 valves and concluded that these actions would prevent spurious opening of the valves if other Group 1 seal-in relays were mechanically boun The CRIT team "dedicated 12 HGA relays to be reinstalled in U-Thi included seven new relays and five relays originally installed since 199 The inspectors reviewed the relay "dedication" package and post maintenance verification tests and concluded that the relays and the contacts operated satisfactorily. Other actions included preparation of a preliminary industry notification letter and review of the potential generic binding problems by the Station's 10 CFR Part 21 committe Engineering Response Overall, the inspectors concluded that the licensee's root cause determination team was effective in investigating and identifying the technical problems with the relays. There was a good mix of station and corporate engineering and relay specialists to perform the root cause analysi The analysis was conducted in.a controlled, careful manne Throughout the testing process, an approved test plan was used and the instructions were written in a logical manner that preserved the as-found condition of the relay Historical Review of the HGA Relays The model 12HGA17S63 relays were installed in 1991 by Modification Package Ml2-3-88-06 Similar relays were installed in U-2 around the same time fram The relays were dedicated for safety related application The inspectors reviewed the 1991 U-3 post modification test and concluded that the relays had been satisfactorily teste At Dresden, the 12HGA17S63 model relay was only used in the PCIS applicatio Twelve relays of this type were used in each Uni The relays were fully tested during each refueling outage during PCIS logic functional testing. Unit 3 Group 1 HGA relays were last tested on July 25, 1994, and U-2 relays were last tested on March 18, 199 All relays operated as designed during the test The inspectors reviewed industry notifications pertaining to the HGA relay No problems associated with armature binding at the pivot points were identifie In addition, a search of industry failure records identified 1 out of 43 entries that involved a binding HGA rela The binding relay was not a model 12HGA17S6 No binding
problems were identified during a review of the failure history of the other model HGA relays used in safety.related applications at Dresde The inspectors identified a concern with several missed opportunities to identify the relay binding proble During 1994 and 1995, three corrective work requests (WRs) were written to repair sticking Group 1 HGA relays on U-A total of six relays were repaired or replaced as neede These failures were not captured by the licensee's trending or generic corrective action program The two 1994 work requests had a job step for the system engineer or work analyst to determine if a generic problem existe In both WRs, this job step was marked "NR" (not required).
The work packages were reviewed by _the Quality Control org.anization; however, the reviewer had not identified a concern that
- the generic problem review step was not complete In addition, the WRs lacked sufficient detail to determine what corrective work was performed and that all contacts of the repaired or replaced relays had been thoroughly teste The work packages focused on the MSIV white.
indicating lights not going out rather than the safety related drop-out of the seal-in contact following a Group 1 isolation signa The inspectors concluded that the three U-2 corrective WRs were an i.ndicat~r of potential binding problems that.subsequently effected U-3 Group 1 HGA relay.0 Isolation Condenser Performance and Radiological Impact of Event
[Charter Item No. 10]
The inspector's assessment of the isolation condenser's (lsoCondenser)
performance and the ~ubsequent radiological impact were based on the fol.lowing reviews:
Radiological survey *
Offsite dose calculation *
Chemi~try sample result *
Interviews with Radiation Protection (RP) and Chemistry personne *
Direct field observation Due to historical uses of contaminated condensate water to provide makeup water to the shell side of the lsoCondenser, residual low levels of radioactivity resided within the condense When the lsoCondenser was placed in operation, low levels of radioactivity were exhausted into the environment through the IsoCondenser steam vent The RP and Chemistry departments dispatched technicians to isolate the area where the IsoCondenser steam was being release This action mitigated any unnecessary exposures (even though small) to plant personne Chemistry technicians collected water samples from various locations near the steam release poin Analytical results indicated that 10 of the samples contained detectable amounts of cobalt-6 The highest sample concentration reported was 1.4 E-06 µcuries/cm 3 *
This concentration was used to quantify the total activity and resultant dose associated with the releas Calculations were performed to determine a conservative value of the volume of water exhausted from the lsoCondenser as stea This volume was multiplied by the highest sample concentration and resulted in a
total activity release of 447 µcuries of cobalt-6 This total was used to quantify an offsite release throtigh airborne and liquid pathway The calculated offsite dose equivalent for the airborne pathway to an adult was 7.12 E-03 mrem and for the liquid pathway for a child was 4.65 E-07 mre These offsite doses were below the limits specified in 10 CFR Part 50, Appendix I (3 mrem/year whole body).
The inspectors independently verified the calculations and methodology for determining the offsite dose The NRC PCDOSE program was used during the verification proces The NRC calculations indicated that the licensee's methods were reasonable for the quantity of radioactivity considered release.0 Performance of Plant Equipment [Charter Item No. 4]
In general, plant equipment performed as designed during the event and through the point of achieving Cold Shutdow Based on interviews. with the operators, there were no work-arounds, degraded equipment, out-of-service equipment, or malfunctioning equipment other than the PCIS relays, whith interfered with the operators' ability to respond to the even There were some unexpected alarms which the licensee determined were indicative of normal wear or were caused by the changing state (Start/Stop) of equipmen The following equipment anomalies were
'identified:
Two Group 1 PCIS valves unexpectedly reopened when the Group 1 isolation signal was rese The cause was a stuck relay in each of the valves' control circuits. There were no consequences because redundant Group 1 valves remained close These failed relays are discussed in Section *
The 38 FRV stem failure resulted in complete loss of feedwater to the reacto The cause was fatigue crackin This issue is discussed in Section *
One control rod display lost indication during the a~tomatic reactor trip. There were no consequences because alternate indication was-available on the Rod Worth Minimize The display returned to normal and functioned correctly' during subsequent troubleshootin.0 Operators' Responses and Procedure Adequacy [Charter Item No. 2]
The inspectors' assessment of the operators' responses to the event and of the procedural adequacy was based on a review of the following:
Plant procedures (normal, abnormal, annunciator response, emergency).
- Control room logs - nuclear station operators (NSO), unit supervisors (US), and shift manage *
Control room instrument chart recorder *
Plant alarm and data print out *
Interviews with control room personne *
Licensed operator training lesson plan *
Simulator scenario *
Field observations during the even.1 Control Room Operators' Responses to the Event The inspectors reviewed the information above to determine if the operators' initial responses and subsequent actions to control room indications were timely and in accordance with plant procedure *
The NSOs' (licensed reactor operators) and US [unit supervisor (licensed senior reactor operator)] initial responses to the control room alarms were to evaluate all of the other indications and alarms on the main control boar Reactor water level dropped to the automatic reactor trip setpoint 7 seconds after the first alarm *
After the reactor trip, the NSOs and US performed the immediate actions for a reactor trip, continued to monitor reactor water level, assessed the condition of the feedwater system, and entered multiple abnormal and annunciator response procedure The NSOs were reviewing the HPCI initiation procedure when HPCI automatically started on Low-Low water leve *
After water level was restored to the normal band and HPCI injection had been reduced to minimum, HPCI was used to control reactor pressur Level continued to increase until the HPCI turbine automatically tripped on high water leve The operators were hesitant to trip the HPCI turbine before the automatic setpoint because HPCI was a functioning source for pressure control and high pressure water injectio *
The Groups l, 2, and 3 isolation signals were reset in order to restore the main condenser as an alternate heat sink (if needed)
and the RWCU system for level contro When the signal was reset, two Group 1 valves unexpectedly reopened due to failed relays in the valves' control circuit (See Section 7.2 for information on procedural inadequacy.)
- The plant was *taken to Cold Shutdown using the IsoCondenser and the shutdown cooling (SOC) syste The cooldown was performed in a controlled fashion and no technical specification (TS) or administrative limits were exceede *
After the two Group 1 isolation valves unexpectedly reopened, the licensee declared an Unusual Event and activated the emergency plan due to the potential degradation in the level of plant safety. All notifications were performed as require The inspectors concluded that the operators performed the appropriate immediate actions, stabilized reactor water level and pressure, and placed the plant in a shutdown condition in accordance with plant procedure The inspectors observed that NSOs monitored control room indications, supervisors maintained proper command and control, and 3-way communications were used throughout the even *
- Procedure Adequacy The inspector's identified minor discrepancies in four procedures that were used during the even *
Dresden Annunciator Procedure (DAN) 903-5 D-4 "GROUP 1 ISOLATION INITIATED," Revision Procedure DAN 903-5 D-4 was unclear regarding the actions necessary to reset a Group 1 isolation signa Section A,
"Automatic Actions," of the DAN listed the Group 1 valves which were expected to close on an isolation signal ~nd had an asterisk (*) beside 12 of the 14 valves (e.g., AO 2(3)-203-lA*).
There was a "Note" in the section B, "Operator Act} ons, ~*. wh-i-Gh stated,
"Control switches for asterisked (*) G~o~p 1 valves must be in CLOSE before the i so 1 at ion s i gna 1 can be reset."
The intent of the "Note," was to direct the operator to place the individual valve control switches to "close" prior to resetting the isolation signa There were no interlocks, devices, or administrative aids to prevent resetting the isolation signal. Additionally, "Notes" in procedures were for information only and were not used to direct operator actio There was no specific step to reset the isolation-signa The licensee revised the DAN on May 22, 1996, to correct the deficienc *
Dresden General Procedure (DGP) 02-03,
"R~actor Scram P~ocedure,"
Revision 2 After the scram, two control rods did not indicate "00" on the full core displa The NSOs used the Rod Worth Minimizer (RWM) to verify that one control rod was fully inserted after the scra Step D.2 of the scram procedure stated to check that all rods were inserted using 00-7 or the full core di~play. The RWM was not mentioned in the procedur However, use of-the RWM w~ consistent with the guidance in Lesson Plan (LP) 295L-Sl, "DEOP 100 Reactor Control," Revision 1, which stated to look at the full core display, RWM, and/or run an OD-7 to obtain the control rod positio The licensee revised the procedure to incorporate the use of the RWM for verification of rod position The inspectors concluded that the use of the RWM was acceptabl The second control rod was verified fully inserted using an approved operator ai As discussed above, two Group 1 valves reopened because relays in the control circuits faile Based on additional reviews, the inspectors concluded that no specific guidance was provided on how to reset the Group 1 isolation *signal or if any prerequisites were neede The status of the relays could have been verified by checking the MSIV Pilot Solenoid Lights in the auxiliary electric equipment roo The licensee revised DGP 02-03 to place Group 1 valve switches to close and to check all MSIV Pilot Solenoid Lights prior to resetting the Group 1 isolation signa *
Dresden Annunciator Procedure (DAN) 902(3)-3 A-10, "HPCI THRUST BRG WEAR ACTIVE FACE," Revision During the HPCI start, the HPCI Thrust Bearing Wear Alarm momentary actuate The DAN stated that the probable causes for this alarm were thrust bearing failure, abnormal shaft movement, or switch failur Additionally this alarm indicated abnormal turbine rotor movement and may indicate impending turbine failur The licensee determined that this was an expected alarm for sudden HPCI star The licensee planned to revise the DA *
Dresden Operating Procedure (QOP) 2300-04, "HPCI System Shutdown,"
Revision _ *. *-'_;r!
The steps in DOP 2300-04 were inadequate to correctly realign the HPCI auxiliary cooling water s~bsyste By performing the steps in DOP 2300-04, the auxiliary cooling pump's flow path was isolate The licensee revised the procedure to include the correct valve line u The inspectors concluded that the.procedural deficiencies had not prevented the operators from placing the plant in a safe conditio.0 Evaluation of the Licensee's Response to the Event [Charter Item No. 3]
Prior to the NRC's decision to conduct an AIT, the li~ensee formed a IO-person Task Force to review the event, to identify the root causes of the equipment failures, and to develop immediate and long term corrective actions for the failures. -Senior licensee managers were assigned to provide dedicate*d oversight of the Task Forc The licensee added personnel to the Team and formed separate groups to investigate the failures of the FRV, HGA relays, and other equipment; to assess the control room operators response to the event; and to evaluate the
- plant's overall response to the even The inspectors' evaluations of the licensee's specific investigations were discussed earlier the repor As part of the licensee's evaluation of the plant's overall response to the event, the condensate -and feedwater systems were inspected and the licensee determined that no external damage existe However, the licensee identified that the condensate booster pumps were operating warmer than usua Subsequent borescope inspection of the inboard and outboard wear rings and pump internals found no problem Desi~n Engineering personnel evaluated the feedwater and condensate piping and various components and confirmed that the event's pressure transient was within the design basis of the syste The inspectors independently walked down portions of the condensate and feedwater system's piping, supports, and equipmen No visible external damage was identifie In addition to the investigations described above, the licensee performed fuel integrity assessments to determine what effect the pressure and temperature transients had on an existing leaking fuel pin (identified in April 1995).
Reactor coolant samples collected following the reactor trip indicated that dose equivalent iodine levels remained consistent with pre-trip levels and were below TS limit However, a
spike indicating increased concentrations of neptunium-239 (Np-239) was note Analysis of prior reactor coolant samples collected during power reductions confirmed that similar spikes had occurred in the pas Evaluation of the spike observed during this shutdown determined that the Np-239 increase was proportional to the magnitude of the power reductio Based on reviews and analyses performed by a qualified nuclear engineer and by corporate fuel integrity engineers, the licensee determined that the reactor trip had no detrimental effect on the leaking fuel pi Region based specialists reviewed the licensee's assessment and concluded the licensee's determination was reasonabl The inspector's concluded that overall, the licensee's investigation was self critical and intensiv The root cause determinations were reasonable and technically sound. 'The systematic approach and depth of the investigation as~ured appropriate corrective ~ctions were develope The inspectors concluded that the existing engineering modification backlog contributed to the delays in completing the modifications on the 38 FR In addition, the station's maintenance backlog contributed to the delays in repairing the 3A FRV and in the missed opportunities to identify a possible trend in the failed HGA relays in 1994 and 199.0 Potential Generic Implications [Charter Item No. 6]
One of the line items of the AIT Charter was to determine if there were any generic issues identified as a result of this even The inspectors identified three issues with potential generic implication.1 GE Model 12HGA17S63 Relays The HGA relays were "commercial grade" products that had been
"dedicated" for safety related application At Dresden, this model was only used in the PCIS applicatio The licensee prepared a preliminary industry notification letter and the mechanical binding problem was being reviewed by the Station's 10 CFR Part 21 committe.2 Procedure to Reset Group 1 Isolation Signal The procedures to reset a Group 1 Isolation signal from two other ComEd boiling water reactors (BWRs) and from one non-ComEd utflity were reviewe One of the ComEd BWR procedures specifically directed taking the individual valves' control switches to close BEFORE resetting th Group 1 signa The procedures from the.other two sites directed that the individual valve control switches be closed AFTER resetting the Group 1 signa.3 Copes Vulcan FRV Failure The licensee reviewed equipment history to determine if other Copes Vulcan valves were used at Dresden which incorporated a similar design to the 38 FR No additional valves were identified; however, Copes Vulcan indicated to the licensee that four other nuclear power plants had similar valves and those plants were informed of the event through the "Nuclear Network."
10.0 Assessment of Licensee's Management Oversight of Followup to Event and Reaction to Issues and Problem [Charter Item No. 11]
1 As stated above in Section 8.0, the licensee immediately formed a 10-person Task Force to review the event, to identify the root causes of the equipment failures, and to develop immediate and long term corrective actions for the failure Senior licensee managers were assigned to provide dedicated oversight of the Task Forc The Task Force initially appeared to be narrowly focused on identifying and correcting the causes of the FRV and Group l isolation relay failure Additionally, some members of the Task Force were involved with both the repair efforts and the root cause investigat~on. Licensee management recognized that a larger team with broad and distinct responsibilities was needed and subsequently assigned additional resources to the Task Forc During the Task Force review of the transient, 37 "issues" (such as equipment failures, unexpected alarms, or other abnormal indications)
were identified that required a more thorough investigatio The licensee identified 29 of the issues and the inspectors identified The inspectors concluded that the licensee had adequately addressed each of the issues both in scope and in dept The inspector's also concluded that management's oversight of the Task Force resulted in a thorough investigation with a broad r~view that looked for generic implication Exit Interview The team met with licensee representatives (identified below) during a public meeting on May 23, 1996, and summarized the purpose of the AIT, AIT charter items, and inspection finding The team discussed the likely informational content of the inspection report with regard to documents or processes revi~wed by the team during the inspectio PERSONNEL CONTACTED Commonwealth Edison Company (ComEd}
T. Maiman, Senior Vice President-Nuclear Division S. Perry, Dresden Site Vice President M. Heffley, Dresden Station Manager T. O'Conner, Operations Manager R. Freeman, Plant Engineering Superintendent P. Swafford, Maintenance Superintendent P. Planing, Shift Operations supervisor R. Fisher, Work Control Superintendent R. Whalen, Staff Assistant-Station Manager C. Howland, Radiation Protection Manager Other members of ComEd Corporate and Dresden Site personne *
U. S. Nuclear Regulatory Commission H. Miller, Regional Administrator, RIII G. Grant, Director, Division of Reactor Safety (ORS}, RIII C. Pederson, Director, Division of Nuclear Materials Safety, RIII P. Hiland, Chief, Division of Reactor Projects (ORP}, Branch l, RIII J. Jacobson, Chief, Engineering Specialists Branch, ORS, RIII J. Hopkins, Branch I Project Engineer, DRP, RIII 0. Butler, Reactor Inspector (Electrical), ORS, Riii P. Louden, Senior Radiation Specialist, ORS, Riii J. Guzman, Reactor Inspector (Mechanical}, ORS, Riii 0. Roth, Resident Inspector, Dresden, ORP, Riii J. Stang, Licensing Project Manager - Dresden, Office of Nuclear Reactor Regulation (NRR}
J. Strasma, Public Affairs Officer, Riii Attachments: Augmented Inspection Team Charter Sequence of Events
Attachment 1 Augmented Inspection Team Charter - Dresden Unit 3 5/16/96 Examine the circumstances surrounding the May 15, 1996, Dresden Unit 3 reactor trip event including but not limited to the following: Develop and validate the sequence of events and activities occurring just before and after the even.
Interview plant personnel and evaluate operator's response to the event and their ability to stabilize the plant and place it in a shutdown conditio Determine if personnel actions and procedural guidance were adequat ~
Evaluate the licensee's actions during and following the event; include initial indicators, their response to the initial indicators, system operations that may have contributed to the event, management's response and the root cause determinatio.
Evaluate the performance of plant ~quipment during the event and following recovery from initial transient up to the point of achieving COLD SHUTDOW.
Determine the cause for the Main Steam Isolation valve and the Reactor Recirc~lation Sample valve reopening following reset of the Group 1 isolation logi.
Determine generic implications of the event, if an.
Assess engineering organization effectiveness to investigate and identify the event's technical problems and their control of the investigation and root cause determinatio.
Determine if appropriate attention had been given to the Feedwater *
Regulating valves (FRVs) including corrective and preventive maintenance, prior to the even Determine the problems associated with the FRVs and how long they have exi5te.
Determine if appropriate attention had been given to the Main Steam and Reactor Recirculation valve relays and other isolation logic relays including corrective and preventive maintenanc.
Evaluate operation of the Isolation Condenser and assess significance of any radioactive releas.
Assess licensee management oversight of the followup to the event and their reaction to the issues and problems identifie *
Attachment 2 Sequence of Events NOTE:
The time is listed using a 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> cloc The units are hours:minutes:seconds, unless otherwise note Information. in It a 7 i cs i ndi cat es an unexpected respons Plant conditions on May 15, 1996, prior to the even Unit 3:
About 82 percent power (677 MWe) in coastdow Unit 3 had been on-line since November 15, 1995:
Normal offsite power and all ECCS available (for both units).
Routine quarterly surveillance testing of the containment cooling service water (CCSW) in progres The 3A FRV and its associated inlet isolation valve, M0-3206A, had been closed since April 27, 1996, due to a body-to-bonnet leak on the 3A FR The 38 FRV was controlling reactor water level in automatic at the normal operating level (about +30 in.).
[Note that the top of active fuel (TAF) is -143 in.]
Unit 2:
About 55 percent power (457 MWe).
Unit.2 had been on-line since April 20,.199 Routine surveillance testing was in progres :43:24 T= 00:00 10:43:28 T= 00:04 10:43:31 T= 00:07 10:43:33 T= 00:09 10:44:43 T= 01:19 10:44:53 First Indications of a Problem were Alarms on the U-3 Main Control Boar *
Feedwater Regulating Station Vibration Hig *
38 Feedwater Regulating Valve Actuator Low Oil Leve Reactor Vessel Water Low Level Alarm (+20 in.)
Automatic reactor trip on low r~actor level (+17 in.). All rods fully in.serte Group 2 Isolations:
drywell, reactor building, and turbine building ventilation systems isolated; standby gas treatment system automatically starte Group 3 Isolations:
RWCU and SOC system All other equipment operated as designe Operators performed the immediate actio~s of reactor scram procedur Two control rods had not indicated position "00."
The two rods were verified fully inserted using alternate indication Reactor water Low-Low level trip signal (-51 in.).
Automatic initiation of HPCI and auto-start of the U-2 and U-2/3 EDG The EDGs were not required to supply power to the emergency busse Group 1 Isolations: 14 valves closed -
8 MSIVs, 2 Isolation Condenser vent valves, 2 reactor recirculation sample valves, and 2 main steam line drain valve The A and B reactor recirculation pumps trippe All other equipment operated as designe Lowest indicated vessel water level:
-59 i *
T= 01:29 10:45:02 T= 01:38 10:45:05 T= 01:42 10:45:08 10:47:00 10:47:35 10:47:50 T = 04 :*36 10:51
.10: 56 10:57:18 T= 13:55 11:01:05 11:01:37
. T= 18:13 11 :02 11: 11 :44 T= 28:20 11:15:12 11: 16 11: 22 T= 39:00
The HPCI system was injecting at full flow (about 6000 gpm) about 19 seconds after the start signal which was within UFSAR assumptions (UFSAR T~ble 6.3-7).
Reactor vessel water level begins to increase due to HPCI and CRD systems injectio Reactor vessel water level Low-Low trip signal rese As level increased, control room operators reduced HPCI injection flow to control vessel leve Main turbine generator tripped as designe Reactor vessel water level was at +7 i Reactor Vessel Water Low Level Alarm cleare The HPCI flow reduced to minimu The HPCI system remained in service for reactor pressure contro Level was +28 i Total volume of HPCI Injection was about 14,800 gallon Reactor pressure was at the lowest point prior to initiating a controlled cooldown:
about 845 psi Torus cooling was placed in service. Torus temperature rise during entire event was about 4°F (71 to 75 °F).
The HPCI turbine tripped on high vessel water level (+48 in).
The CRD pump continued to inject water into the reacto Condensate and Feedwater pumps secured because sump pumps for the 3A and 38 turbine building equipment drains sumps and 38 turbine building floor d~ain sump were continuously runnin (The 3C3 feedwater heater relief valve and a condensate/condensat~ booster pump suction relief valve had lifted.)
The 3C3 feed~ater relief valve had not reseated quickl Isolation Co~denser manually placed in service for reactor pressure control and cooldown.
Reactor pressure was at the highest level during event:
about 1000 psi Groups 1, 2, and 3 Isolation signals were reset...
The IA NSIV and recirculation sample valve 220-45 unexpectedly reopened when the Group 1 signal was rese The U-2 US identified that the IA NSIV was ope The positions of the remaining Groups 1, 2, and 3 valves were re-verified close The IA MSIV was closed by a NS Recirculation sample valve 220-45 was found open during the re-verification and closed by a NS An Unusual Event was declared due to the potential degradation in the level of plant safet All notifications were performed in a timely manne..
'
17:43 17:48 18:55 19:30 19:52 20:03 23:08 e
Secured Isolation Condense (T = 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />.)
Started 3B SOC Loop to continue cooldow Torus Cooling secure Reactor at Cold Shutdow (T = 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> and 43 minutes.)
Started 3C SOC Loop for mixin Maintaining reactor coolant temperature between 160 - 170 Reset reactor trip in accordance with procedur Terminated Unusual Even (T = 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> and 25 minutes.)
24