Text

Category 1 Root Cause Evaluation 95-12 for Undetected Loss of Unit 1 Reactor Coolant Sys Inventory While at Cold Shutdown, for Sept 1995
	ML18152A125
Person / Time
Site:	Surry
Issue date:	09/30/1995
From:	Beth Brown, Crist M, Modlin D; VIRGINIA POWER (VIRGINIA ELECTRIC & POWER CO.)
To:	;
References
	95-12, NUDOCS 9607240126
	Download: ML18152A125 (29)
	v • d • e

' .

SURRY POWER STATION CATEGORY 1 ROOT CAUSE EVALUATION 95-12 UNDETECTED LOSS OF UNIT 1 REACTOR COOLANT SYSTEM INVENTORY WHILE AT COLD SHUTDOWN September, 1995 DRAFT DRAFT DRAFT DRAFT DRAFT DRAFT Root Cause Team Brad Brown, Shift Supervisor, NAPS Mike Crist, Director Corporate Nuclear Safety Danny Modlin, Training, SPS Robert Scanlan, Shift Technical Advisor, SPS Francis Terminella, Corporate Nuclear Safety Prepared by: _!_/_

Robert J. Scanlan Date Reviewed by: _!_!_

Jerry A. McGinnis Date Approved by: _/_/_

John T. Swientoniewski Date Concurrence: _!_/_

Chairm~n, SNSOC Date

. -. -. 1i n ,._ _!_/_

- .9607240126 950930 .'

PDR ADOCK 05000280 I S PDR 1 (0

Management Review Date 2

EXECUTIVE

SUMMARY

RCE 95-12 On September 14, 1995, Unit 1 was at cold shutdown and drained to a Reactor Coolant System (RCS) standpipe level indication of 18 feet, which-corresponds to a level just below the reactor vessel flange. Maintenance personnel began de-tensioning the reactor vessel head at 0547 hours0.00633 days 0.152 hours 9.044312e-4 weeks 2.081335e-4 months and, by 0920 hours0.0106 days 0.256 hours 0.00152 weeks 3.5006e-4 months , de-tensioning had progressed sufficiently to allow communication between the RCS vessel and the containment atmosphere. RCS standpipe level indication rapidly dropped from 18 feet to 13.3 feet indicating a loss of reactor vessel inventory. Approximately 5000 gallons of borated water were added to the RCS to restore the RCS standpipe level indication to 18 feet. At 1100 hours0.0127 days 0.306 hours 0.00182 weeks 4.1855e-4 months , Operation's personnel identified that the reactor vessel head vent valves were tagged closed which had allowed a pressure differential to develop between the vessel head and the pressurizer.

A Root Cause Evaluation (RCE) team was formed to determine the root cause of the event and develop recommendations to prevent recurrence. The RCE team also reviewed plant conditions, evolutions, and operator actions which may have contributed to the loss of RCS inventory and the failure of the shift team to detect the loss of RCS inventory.

EVENT

SUMMARY

The reactor vessel head vent was isolated and tagged closed at approximately 1500 hours0.0174 days 0.417 hours 0.00248 weeks 5.7075e-4 months on September 13th to remove the reactor vessel head vent spoolpiece in support of cavity seal ring installation. The reactor vessel head vent was isolated which severed communication between the reactor vessel head and the pressurizer and rendered the RCS standpipe inoperable for accurate level indication. Prior to head vent isolation, reactor vessel level was being maintained just below the reactor vessel flange with a nitrogen pressure of 11 psig in the reactor vessel, the pressurizer and the pressurizer relief tank (PRT).

Consequently, isolation of the head vent also confined approximately 460 cubic feet of nitrogen gas at 11 psig in the reactor vessel head. The PRT was subsequently vented and RCS pressure decreased from 11 psig to atmospheric pressure. As the RCS depressurized, the nitrogen volume in the reactor vessel head expanded displacing RCS inventory. During this evolution, which was in progress during shift turnover, the operator controlling RCS inventory increased the letdown rate to maintain an indicated RCS standpipe level of 18 feet.

The increased letdown rate caused Volume Control Tank (VCT) level and pressure to increase. At approximately 1900 hours0.022 days 0.528 hours 0.00314 weeks 7.2295e-4 months , VCT level reached the automatic divert setpoint and the VCT began diverting to the Primary Drains Tank. The loss of RCS inventory resulting from the VCT divert continued for approximately 3 1 /2 hours until 2230 hours0.0258 days 0.619 hours 0.00369 weeks 8.48515e-4 months when the RCS was depressurized.

THE ROOT CAUSE OF THIS EVENT WAS:

3

TRAINING/QUALIFICATIONS - Operations personnel and Shift Technical Advisors (STAs)

did not consider RCS standpipe level to be inoperable with the head vent isolated. Personnel interviewed believed that the standpipe would continue to provide reliable indication as long as RCS level changes were not made.

4

e e CONTRIBUTING CAUSES TO THIS EVENT WERE:

WORK PRACTICES - Shift personnel did not display the appropriate questioning attitude with respect to indications that were available and, if properly assessed, would have detected the loss of RCS inventory or the status of the head vent. Shift personnel did not perform adequate self checks in reconciling inventory balances. Shift Supervisor and STA involvement in shift activities was inadequate to maintain a broad perspective of unit activities focusing on core safety, and to identify that unexpected conditions existed.

WRITTEN COMMUNICATIONS - The mass balance procedure was ineffective in that it did not provide for reconciliation of individual inventory changes. Additionally, it did not properly account for all sources of inventory changes for the existing unit conditions. Isolation of the reactor vessel head vent was not documented in the unit log or included on written turnover documents.

CONTRIBUTING FACTORS TO THIS EVENT WERE:

WORK PRACTICES - Status control of the reactor head vent was lost in .that neither the Reactor Operators nor the Unit SRO recorded the fact that the head vent was isolated in the unit log. Turnover was ineffective in that key members of the shift operating team were not made aware that the reactor vessel head vent was isolated.

SUPERVISORY METHODS - Command and control of shift activities by the Unit SRO and Shift Supervisor were inadequate to ensure that equipment status and plant conditions were known and understood by shift personnel and that plant conditions and the main control room environment were appropriate for shift turnover. The Unit SRO and Shift Supervisor did not integrate the STA into shift activities.

WRITTEN COMMUNICATIONS - There were no procedural controls to remove and return the head vent to service when the head vent spool piece was removed for cc1vity seal ring installation.

5

TABLE OF CONTENTS EXECUTIVE

SUMMARY

I. DETAILED DESCRIPTION OF THE EVENT A. INITIAL OUTAGE ACTIVITIES

8. PRE~EVENT ACTIVITIES

C. EVENT D. INITIAL CORRECTIVE ACTIONS

11. EVENT ANALYSIS AND CAUSES A. PROBLEM STATEMENT

8. INVESTIGATIVE TECHNIQUES USED C. RESULTS OF THE CATEGORY 1 ROOT CAUSE ANALYSIS

1. Training/Qualifications D. CONTRIBUTING CAUSES

1. Work Practices

2. Written Communications E. CONTRIBUTING FACTORS

1. Work Practices

2. Supervisory Methods

3. Written Communications 111.

SUMMARY

OF ROOT CAUSE AND CONTRIBUTING CAUSES AND FACTORS IV. RECOMMENDATIONS V. ENHANCEMENTS VI. PERIPHERAL ISSUES ATTACHMENTS 1 . Event & Causal Factors Chart

2. Sequence of Events 6

3. Control Room Logs

4. Station Deviations 7

I. DETAILED DESCRIPTION OF THE EVENT A. INITIAL OUTAGE ACTIVITIES

1. Summary of Unit Maneuvers Prior to the Event At 2205 hours0.0255 days 0.613 hours 0.00365 weeks 8.390025e-4 months on September 7, 1995, a Unit 1 rampdown was initiated from 73% reactor power and 565 MWe for a scheduled refueling outage. Unit 1 was taken off line at 0241 hours0.00279 days 0.0669 hours 3.984788e-4 weeks 9.17005e-5 months on September 8th, and the reactor was manually tripped at 0320 hours0.0037 days 0.0889 hours 5.291005e-4 weeks 1.2176e-4 months . Hot Shutdown outage activities were completed and a cooldown was commenced at 1600 hours0.0185 days 0.444 hours 0.00265 weeks 6.088e-4 months . On September 9th, at 1230 hours0.0142 days 0.342 hours 0.00203 weeks 4.68015e-4 months , the Residual Heat Removal (RHR) system was placed into service and at 1315 hours0.0152 days 0.365 hours 0.00217 weeks 5.003575e-4 months the Reactor Coolant System (RCS) was filled solid.

Cold Shutdown was achieved at 1455 hours0.0168 days 0.404 hours 0.00241 weeks 5.536275e-4 months .

A through wall leak was identified in the II A II RHR pump casing at 2230 hours0.0258 days 0.619 hours 0.00369 weeks 8.48515e-4 months on September 9th. The "A" RHR Pump was declared inoperable and remained inoperable until code relief was granted by the NRC at approximately 1700 hours0.0197 days 0.472 hours 0.00281 weeks 6.4685e-4 months on September 12th.

At 1937 hours0.0224 days 0.538 hours 0.0032 weeks 7.370285e-4 months on September 12th, with two operable RHR pumps, the remaining operating Reactor Coolant Pump, 1-RC-P-1 C, was secured and RCS depressurization was initiated. Once the RCS was depressurized, the pressurizer Power Operated Relief Valves (PORVs) were opened to allow communication between the pressurizer and the Primary Relief Tank (PRT).

Nitrogen gas was added to the PRT in preparation for RCS draindown.

Draindown of the RCS to 5% pressurizer level was initiated at 2335 hours0.027 days 0.649 hours 0.00386 weeks 8.884675e-4 months .

The reactor vessel head vent was placed in service at 0325 hours0.00376 days 0.0903 hours 5.373677e-4 weeks 1.236625e-4 months on September 13th, and the RCS draindown was secured approximately one hour later at 0416 hours0.00481 days 0.116 hours 6.878307e-4 weeks 1.58288e-4 months .

2. Schedule Status The original outage schedule had called for RCS loop isolation to be completed by 0300 hours0.00347 days 0.0833 hours 4.960317e-4 weeks 1.1415e-4 months on September 13th. Thus as Dayshift assumed the watch, the station was approximately 12 hours1.388889e-4 days 0.00333 hours 1.984127e-5 weeks 4.566e-6 months behind the original Unit 1 outage schedule.

3. Unit Conditions Prior to SI Accumulator Check Valve Testing On the morning of September 13th, the following conditions existed at Operations Shift Turnover:

RCS level was being maintained at 8 % cold calibrated pressurizer level with the reactor vessel head vent in service.

8

e RCS pressure was being maintained at a positive pressure by PRT pressure indication via the pressurizer and PRT.

Charging was secured, the unit was on a VCT float, with RCS inventory being controlled both by makeup via the blender and RWST gravity feed and by letdown.

8. PRE-EVENT ACTIVITIES

1. RCS Level Indication Discrepancies The normal cold calibrated pressurizer level indicator, 1-RC-Ll-462, was inoperable prior to the beginning of this refueling outage. As a result, 1-RC-Ll-461 was cold calibrated September 8, 1995 with Unit 1 at Hot Shutdown.

It is the normal practice to perform the cold calibration activity at Cold Shutdown and to fill and vent the level transmitter and reference leg following calibration. In this instance, RCS pressure precluded filling the reference leg via the normal method. The transmitter, 1-RC-LT-*461, was returned to service without being filled and vented.

Early on Dayshift, September 13th, preparations were being made to conduct accumulator check valve test, 1-0PT-Sl-006, "SI Accumulator Discharge Check Valves Full Open Test." The only operable pressurizer level indicator, 1-RC-Ll-461, would not indicate below 8%. This was discovered while attempting to lower pressurizer level to 5% in accordance with step 6.1.4 of 1-0PT-Sl-006. With 1-RC-LT-1461 indicating 8%, the Shift Supervisor noted that RVLIS had decreased to approximately 95% from 100%. RCS standpipe level remained off-scale high. Draining of the pressurizer was stopped and the situation was evaluated. The Shift Supervisor observed the RCS wide range pressure, as indicated on the P-250 computer, was reading -8 psig. He concluded that the RCS was under a vacuum and directed that the PRT pressure be increased to 11 psig to ensure that the RCS was not under a vacuum. The operating shift then maintained this pressure throughout this evolution and the subsequent RCS draindown to 18 feet. Following this event, the Operator at the Controls (OATC) indicated that he was not aware that RVLIS had come on scale with 1-RC-LT-461 indicating 8%, nor was the STA.

2. Performance of 1-0PT-Sl-006 Due to the behavior of 1-RC-Ll-461, a procedure change (PAR) was processed to start the Safety Injection (SI) accumulator discharge check valve testing at 10% pressurizer level verses 5%. This PAR was approved as an intent change by the Supervisor Shift Operations (SSO). The initial pressurizer level recorded for the first check valve test was 8%. The test procedure was 9

e completed satisfactorily and during its performance, Operations personnel and the STA monitored pressurizer level, observing~it to trend as expected;-' An increase in pressurizer level was observed during the testing of each accumulator check valve and a decrease in pressurizer level was observed during the subsequent draindown. However, the lowest indicated pressurizer level observed was 8%.

3. Unit Conditions Prior to the Event Operation's management discussed the performance of 1-RC-Ll-461 and concluded that it would be acceptable to begin the RCS draindown to 18 feet RCS standpipe level from 11 % indicated pressurizer level. This decision was based upon the fact that both RVLIS and RCS standpipe level instrumentation were available to support draindown of the RCS. This decision was not supported by the entire shift crew. Just prior to initiating the draindown to 18 feet, the RCS was being maintained at 11 psig via PRT nitrogen gas overpressure, pressurizer level was indicating 11 % and the reactor vessel head vent and RCS standpipe were in service with their valves locked open as recommended by RCE 91-06, "Surry Unit 2 Head Vent."

C. EVENT

1. RCS Draindown to 18 Feet A Pre-Job Brief was held at 1100 hours0.0127 days 0.306 hours 0.00182 weeks 4.1855e-4 months for the draindown from 11 %

Pressurizer Level to 18 feet on the RCS standpipe. The procedure for the draindown, 1-0P-RC-004, "Draining the RCS to Reactor Flange Level,"

provided 30 minutes as the expected time from 5% pressurizer level to when the RCS standpipe level should,. be on scale at a drain rate of 40 gpm.

Draindown commenced at 65 gpm and the standpipe came on scale in approximately 12 minutes. The shift team did not estimate when RCS standpipe level should have come on scale given the higher draindown rate (24 minutes), however, they were surprised that the standpipe came on scale so quickly. Station Deviation Report, S-95-2103, was submitted by the STA to document the decreased accuracy at the lower end of the span of 1-RC-Ll-461.

The draindown was secured at 18 feet with PRT pressure being maintained at 11 psig. At this point, an 11 psig nitrogen gas bubble existed in the reactor vessel from 18 feet (just below flange level) to the top of the reactor vessel head. The actual time 18 feet was reached was not logged, however, the STA turned over at 1305 hours0.0151 days 0.363 hours 0.00216 weeks 4.965525e-4 months and observed RCS standpipe level indicating 18 feet.

The draindown procedure, 1-0P-RC-004, step 5. 7. 19, requires that the PRT 10

be vented to the Process Vent system following the draindown to 18 feet.

Completion of this step prior to entry into, 1-0P-RC-007, "Isolation and Drain of All Reactor Coolant Loops with All Loop Stop Valves Closed and RHR in Service," would have established a slight vacuum in all portions of the RCS.

Step 5. 7.19 was initiated although not completed. The step did not contain an operator sign-off line. The following step, 5. 7.20, instructed that all conditions and required sign-offs be satisfied prior to isolating one or more RCS loops. As a result of not completing step 5. 7.19, the RCS remained at 11 psig.

During the Unit 1 outage in December 1994, it was desired to maintain nitrogen gas overpressure on the PRT. A One-Time-Only PAR was processed to change 1-0P-RC-004 to provide this option. Station Procedures personnel stated that the intent of step _5. 7.19 is to depressurize the PRT prior to exiting 1-0P-RC-004.

2. Key Outage Activities Prior to RCS Depressurization At 1313 hours0.0152 days 0.365 hours 0.00217 weeks 4.995965e-4 months the shift commenced isolating RCS loops in accordance with 1-0P-RC-007 beginning with the "A" loop. The pre-job brief for this evolution was inadequate and not controlled by the unit supervisor. At approximately 1330 hours0.0154 days 0.369 hours 0.0022 weeks 5.06065e-4 months , work orders (WOs) to retract the incore flux thimbles and remove the pressurizer code safeties were authorized by the Desk SRO.

The STA was in the Main Control Room (MCR) for the RCS loop isolation activities. Shortly after arriving in the MCR, he reviewed 1-0P-RC-10.2, "Reactor Coolant System Cold Shutdown Inventory Balance," and calculated that the change in Boron Recovery Tank level attributed to the RCS draindown to 18 feet was 4%. This represented an increase of 5000 gal in BRT inventory. Reconstructed estimates of the actual change in RCS inventory during the draindown, after factoring in the inaccuracy of 1-RC-Ll-461, was 4206 gal.

In preparation for setting of the cavity seal ring, a tagout was prepared to isolate the head vent to allow removal of the head vent spoolpiece for setting the cavity seal ring. The tagout for the head vent spoolpiece referenced three WOs. Two of the WOs involved the removal and reinstallation of the head vent spool piece. The third was for the master WO for reactor disassembly and reassembly. The latter WO was added by the Desk SRO to support removal of the RVLIS bracket. The maintenance foreman responsible for the first two WOs was not aware the third WO had been added to the tagout.

A pre-job brief for head vent isolation associated with cavity seal work was conducted by the Desk SRO with the ROs at the board. The Unit 1 SRO was not present at the beginning of the brief and was briefed separately by the Desk SRO. The STA was not included in the pre-job brief and was unaware 11

it was planned to isolate the head vent. The pre-job brief emphasized to the maintenance foreman the need to expeditiously complete reinstallation of the head vent spoolpiece following lowering of the cavity seal ring. The Shift Supervisor (SS) did not maintain cognizance of the progress of this evolution once the reactor ,vessel head vent was isolated.

At 1500 hours0.0174 days 0.417 hours 0.00248 weeks 5.7075e-4 months , the reactor head vent was isolated and tagged to support removal of the spoolpiece for setting the cavity seal ring. Isolation of the head vent was well controlled and communicated to the control room via radio. This evolution was not documented in the Unit 1 log nor was the Unit 1 SRO or the STA notified when the head vent isolation valves were closed.

Isolation of the reactor vessel head vent confined approximately 460 cubic feet of nitrogen gas at 11 psig in the reactor vessel head.

At approximately 1515 hours0.0175 days 0.421 hours 0.0025 weeks 5.764575e-4 months , Westinghouse notified the Desk SRO that they were getting water out of a high pressure seal on an incore flux thimble. The Desk SRO called the MCR to verify plant status and discovered that, although RCS Standpipe level was 18 feet, the RCS was still pressurized to 11 psig via the PRT. This pressure was sufficient to force water out of the incore flux thimble seal. The Desk SRO directed Westinghouse to retighten the HP seal and, recalling that he had authorized removal of the pressurizer code safeties, recalled the associated WO prior to it being released to the craft for work.

Preparations were initiated to vent the PRT in accordance with 1-0P-RC-011, "Pressurizer Relief Tank Operations." This event was not documented in the unit logs, no station deviation was submitted, and the STA was not aware that it had occurred.

Maintenance personnel also completed transmitter venting and filling the reference leg for pressurizer level transmitter 1:-RC-LT-461 at approximately 1515 hours0.0175 days 0.421 hours 0.0025 weeks 5.764575e-4 months .

By 1629 hours0.0189 days 0.453 hours 0.00269 weeks 6.198345e-4 months , all three RCS loops has been isolated and the shift commenced draining "A" RCS loop to the Primary Drains Transfer Tank (POTT). Operations personnel experienced problems with the pumping rate of the POTT throughout this evolution, therefore, draining of "A" RCS loop was intermittent. Resolution of the POTT pumping rate was the main focus of the SS, who was actively performing flow calculations in an attempt to troubleshoot the problem. The SS did not request STA .assistance in resolving this problem and did not inform him that a problem existed. Difficulty in pumping the POTT was the major topic of the subsequent turnover to night shift.

3. RCS Depressurization Venting the PRT to the Process Vent system in accordance with Section 5.6 of 1-0P-RC-011 commenced at 1710 hours0.0198 days 0.475 hours 0.00283 weeks 6.50655e-4 months . The Gaseous Mixed Mode 12

e - - - - - - - - - - - - - - - - - - - - - - - - -- -

Release Permit required by step 5.6.4 was obtained. The PRT release rate to the Process Vent system was slow and controlled to prevent alarming the Process Vent radiation monitor and resulted in a slow RCS depressurization.

With the head vent isolated, the volume of nitrogen gas trapped in the reactor vessel head began expanding as the RCS depresurized. As the gas in the reactor vessel head expanded, RCS standpipe indicated level increased as water was pushed out of the reactor vessel. In order to maintain standpipe stable, the RO increased his letdown flow rate to the VCT.

The Unit 1 SRO directed the PRT be drained to the POTT to provide an additional vent path to gaseous waste and potentially speed up the PRT venting process. This was performed without a procedure and was not identified to the STA so that the volume of water could be factored into shift inventory balance calculations.

By 1800 hours0.0208 days 0.5 hours 0.00298 weeks 6.849e-4 months the reactor vessel cavity seal ring had been set and reinstallation of the head vent spoolpiece completed. Mechanical maintenance personnel signed off the two WOs

for spoolpiece removal/installation, and the signing of the associated tagout release form was turned over to the night shift. The operating shift was not informed by maintenance personnel that the spoolpiece had been reinstalled.

With the PRT venting rate stabilized, the operating shift estimated that the PRT would be fully vented at approximately 2200 hours0.0255 days 0.611 hours 0.00364 weeks 8.371e-4 months . This information was communicated to the Shift Outage Coordinator and was inconsistent with a previous estimate of completion by 1900 hours0.022 days 0.528 hours 0.00314 weeks 7.2295e-4 months .

At approximately 1815 hours0.021 days 0.504 hours 0.003 weeks 6.906075e-4 months , the Superintendent of Planning called the Superintendent of Operations and the Supervisor Shift Operations and requested that conflicting estimates of PRT venting completion be resolved.

The Supervisor Shift Operations proceeded to the control room, and after evaluating the existing PRT venting rate, told the Shift Supervisor that he desired to vent the PRT via a parallel path to the containment. This proposal was accepted by the Shift Supervisor. The Unit 1 SRO objected to venting the PRT to containment and voiced his concern to the Shift Supervisor and the Supervisor Shift Operations. The decision was made to vent to containment. The Supervisor Shift Operations contacted Radiation Protection (RP) during the RP shift turnover to inform them that Operations was planning to vent the PRT to containment. The RP Shift Supervisor stated that he would "check on it" and subsequently turned over to night shift to investigate.

At approximately 1830 hours0.0212 days 0.508 hours 0.00303 weeks 6.96315e-4 months , the Supervisor Shift Operations contacted an Operator assigned to the Type "C" Testing Group who was in containment and directed him to open 1-RC-ICV-5052, the instrument vent valve associated with PRT pressure transmitter 1-RC-PT-147 2. The operator located 1-RC-ICV-5052, informed the Supervisor Shift Operations that it was 13

not labeled and asked whether it was still desired that the valve be operated.

A prior senior station management memo had provided direction that valves not be operated unless properly labeled. This direction was provided as part a

of corrective action response to misposition events. The operator said he was confident he had properly identified the vent valve based upon other properly labeled valves and components. The Supervisor Shift Operations directed the operator to open the valve. The PRT vent was opened directly to the containment atmosphere, left unattended, and the operator returned to his Type "C" testing duties.

A pre-job brief was not held prior to venting the PRT to containment and the STA was not informed that this evolution was to be performed. The evolution was performed without using a procedure although the required procedure, 1-0P-RC-011, was open and in use in the control room for venting to the PRT to the Process Vent system. Both Reactor Operators on the control board provided the Supervisor Shift Operations with the procedure for venting the PRT to the Vent Vent system. The procedure was subsequently handed back to them and was not used during venting of the PRT to containment. As a result, the Gaseous Ground Level Release Permit required by step 5.5.4 was not obtained, a poly hose specified in step 5.5.5 was not used to direct the discharge from 1-RC-ICV-5052 to the nearest Containment Purge Exhaust, and the PRT vent path to the Process vent system was not secured as required by step 5.5.6.a. Station Deviation Report S-95-2122 was later submitted by RP personnel after identifying that a Gaseous Ground Level Release Permit was not issued prior to the release.

Initiation of a parallel PRT vent path increased the rate of RCS depressurization which, in turn, increased the rate of reduction in vessel inventory. The unit 1 reactor operator continued to adjust letdown to maintain standpipe level. VCT level began rising more rapidly and continued to rise until the VCT began diverting to the PDT due to high level at approximately 1900 hours0.022 days 0.528 hours 0.00314 weeks 7.2295e-4 months .

During turnover in the Main Control Room, the Unit 1 Dayshift ROs and SRO felt frustration in turning over to Nightshift with a member of Operations management directing activities. However, none of them voiced this frustration. The Dayshift SS and the Desk SRO were aware of the venting of the PRT and did not see a problem with performing the evolution. Operation's Nightshift RO and SRO assumed their watches at 1904 hours0.022 days 0.529 hours 0.00315 weeks 7.24472e-4 months and 1907 hours0.0221 days 0.53 hours 0.00315 weeks 7.256135e-4 months , respectively. Difficulty in pumping the POTT was the major topic of the turnover to night shift. The Dayshift RO of record recalls verbally turning over to the Nightshift RO of record that the reactor vessel head vent was isolated, however, the Nightshift RO of record does not recall hearing this information. No other control room personnel discussed the status of the Unit 1 head vent during turnover. Written communications tools such as the Unit 1 Log and turnover sheets did not contain any information concerning the 14

status of the reactor vessel head vent for use during turnover. Shortly after assuming the watch, the RO assigned responsibility for RCS inventory control initiated actions to stabilize VCT level and pressure. VCT level was subsequently stabilized about the VCT high level divert setpoint of 60% and the VCT continued to divert for 3 1 /2 hours until 2230 hours0.0258 days 0.619 hours 0.00369 weeks 8.48515e-4 months . The RO incorrectly attributed the increase in letdown rate to steam generator tube "burping."

During turnover in the main control room annex, the Dayshift Desk SRO verbally turned over twice to the Nightshift Desk SRO that the Unit 1 reactor vessel head vent was tagged out and it was one of eight priority tagouts discussed. The Dayshift SS did not verbally communicate any information on the status of the Unit 1 reactor vessel head vent to the Nightshift SS and written turnover documentation did not contain any information concerning the Unit 1 reactor vessel head vent.

The STA turnover did not include any information on the status of the reactor vessel head vent because the Dayshift STA was not aware that the head vent had been tagged out.

Consequently, the only individual on Nightshift cognizant that the reactor vessel head vent was tagged out was the Desk SRO. He did not communicate this information to the shift team during the Shift Brief.

Clearing of the reactor vessel head vent spoolpiece WOs was accomplished following maintenance turnover by obtaining the tagout in the main control room annex and signing the tag release form indicating the two WOs had been completed. The third WO was not signed off and would not normally be until the reactor was reassembled or Operations requested a release.

Because the Desk SRO was not aware that maintenance had signed the tag release form, Operations did not request a release for the third WO and the tagout remained in the stack of open tagouts.

The loss of RCS inventory continued until the PRT vent was complete at 2230 hours0.0258 days 0.619 hours 0.00369 weeks 8.48515e-4 months . After that there may have been some reduction in inventory from off gassing and the draining of A loop with leakage through the loop stop valves, but the VCT divert had stopped.

4. Outage Activities Prior to Restoration of Standpipe Level Indication At 2316 hours0.0268 days 0.643 hours 0.00383 weeks 8.81238e-4 months , the Nightshift STA reported to the Unit 1 SRO that the RCS Leakrate was a 4. 6 gpm (leakage out of the RCS). This was the initial performance of 1-0P-RC-10.2 for this outage. The calculation spanned a period of time where multiple RCS inventory changes were known to have been made and satisfied the 1 0 gpm acceptance criteria contained in the procedure. Consequently, neither the STA nor the Unit 1 SRO pursued any 15

further investigation of the source of the leakage.

A review of the completed calculation after the event identified that, because the procedure did not properly account for draining an isolated loop, the Nightshift STA had attempted to compensate for this condition. As a result, the inventory increase in the BRT due to the VCT divert (the VCT diverted to the PDT which was subsequently pumped to the BRT) was attributed to the II II II draining of the A RCS Loop. Since the actual volume drained from the A" RCS Loop was unknown and small due to problems with transferring water from the POTT to the PDT, a procedurally acceptable leak rate was obtained.

Recalculating the initial leak rate following the event and after understanding the sequence of events associated with the RCS inventory loss, yielded a RCS leak rate of 12 gpm (leakage into the RCS).

Nightshift increased RCS inventory from 18 feet to 20 feet twice to leak check the Seal Table for leakage. During both evolutions approximately 600 gallons of water was added and subsequently drained as compared to what should have been 1 600 gallons. The Unit 1 SRO recalls that he was surprised at how fast they could reduce standpipe level from 20 feet to 18 but did not investigate or communicate this to the STA. The STA was not present during the floodup to 20 feet and draindown to 18 feet and only computed the change in inventory to provide RCS inventory balance data for the RCS leakrate calculation due later in the shift. The STA was not informed of the second floodup to 20 feet and draindown to 18 feet and, consequently, did not include this information in his next RCS leakrate calculation. The second RCS leakrate calculation performed determined the RCS leakrate to be 3.25 gpm (leakage out of the RCS) at 0516 hours0.00597 days 0.143 hours 8.531746e-4 weeks 1.96338e-4 months on September 14th. Again, this RCS leakrate satisfied the 10 gpm acceptance criteria contained in the procedure, represented a downward trend in RCS leakrate and received no further attention from the STA or the Unit 1 SRO.

A review of the second completed calculation after the event identified an error in recording BRT level. Recalculating the second leak rate following the event using the correct BRT level yielded a RCS leak rate of 7. 2 gpm (leakage into the RCS).

The pressurizer safeties were also removed on night shift.

The turnover at 0700 hours0.0081 days 0.194 hours 0.00116 weeks 2.6635e-4 months on September 14th did not include any communications concerning the reactor head vent. The cavity seal ring was set, all three RCS loops were isolated, the p"roblems with pumping the POTT 11 were resolved, and "A RCS loop was still being drained.

5. RCS Level Standpipe Indication Restoration Detensioning of the head commenced at 0547 hours0.00633 days 0.152 hours 9.044312e-4 weeks 2.081335e-4 months and at 0920 hours0.0106 days 0.256 hours 0.00152 weeks 3.5006e-4 months 16

e standpipe level dropped from 18 feet to 13.3 feet. Calculations show this volume to be approximately 4233 gallons.

6. Recovery From Loss of RCS Inventory The shift team responded to the loss in level by restoring inventory both via blending and gravity feed from the RWST. Entry into 1-AP-27, "Loss of Decay Heat Removal Capability" was initiated following a prompt from the Assistant Station Manager Operations and Maintenance. Indicated RCS standpipe level was restored to 18 feet at 1036 hours0.012 days 0.288 hours 0.00171 weeks 3.94198e-4 months . At 1100 hours0.0127 days 0.306 hours 0.00182 weeks 4.1855e-4 months Operations personnel identified that the reactor vessel head vent isolation valves were tagged closed. The tags were cleared, the head vent valved in and the head vent and standpipe verified to be properly aligned by an SRO at 1214 hours0.0141 days 0.337 hours 0.00201 weeks 4.61927e-4 months . No change in indicated RCS standpipe level was observed after the head vent was unisolated.

17

D. INITIAL CORRECTIVE ACTIONS

1. The reactor vessel head vent valves were tagged open and will remain tagged open during all periods when the RCS standpipe is needed for RCS level indication.

2. The configuration of important safety systems was confirmed by the review of status control documents and, for low head safety injection, a walkdown of the system.

3. The station managers held face to face meetings with on-shift SROs and STAs to ensure that performance expectations and operation of the RCS standpipe are clearly understood.

Management expectations of Operations leadership in the area of command and control and monitoring of critical evolutions were clarified.

Management expectations for STA involvement in shift activities were clarified. Critical monitoring of plant parameters and the independent evaluation of the impact of changes in equipment condition and schedule changes on plant safety and plant activities were emphasized.

4. The backshift managment coverage schedule was revised such that only SRO qualified individuals who have been coached on the above expectations will perform this function.

5. The core onload procedure was revised to require Station Manager approval prior to core on-load. The pre-core on-load assessment will emphasize equipment status, status control and actions taken in response to this event.

6. The outage schedule was revised to reduce multiple RCS inventory activities.

18

11. EVENT ANALYSIS AND CAUSES A. PROBLEM STATEMENT What was the cause of an undetected loss of Surry Unit 1 RCS inventory which was identified at 0920 hours0.0106 days 0.256 hours 0.00152 weeks 3.5006e-4 months on September 14th by a decrease in RCS standpipe level from approximately 18 feet to 13.3 feet.

B. INVESTIGATIVE TECHNIQUES USED

1. Interviewing

2. Event and Causal Factor Charting

3. Barrier Analysis C. RESULTS OF THE CATEGORY 1 ROOT CAUSE ANALYSIS The reactor vessel head vent was isolated which severed communication between the reactor vessel head and the pressurizer and rendered the RCS standpipe inoperable. Isolating the reactor vessel head vent also confined a pressurized volume of nitrogen in the reactor vessel head. When the RCS was subsequently depressurized, the nitrogen in the reactor vessel head expanded displacing RCS inventory.

1. Training/Qualifications - Operations personnel and Shift Technical Advisors (ST As) did not consider RCS standpipe level to be inoperable with the head vent isolated. Personnel interviewed believed that the standpipe would continue to provide reliable indication provided RCS parameters were not changed.

a. Information Notice 94-36, "Undetected Accumulation of Gas in Reactor Coolant System," covers the natural off-gas phenomena which will cause a loss of RCS inventory with the head vent closed.

Recommendations:

1) Conduct remedial training for Operations personnel and STAs on the interrelationship of the standpipe and head vent.

2) Revise continuing training lesson plans for Operations personnel and STAs to enhance discussion of the interrelationship between the standpipe and head vent, including simulator lesson plans.

3) Modify Initial RO/SRO license class integrated plant operations simulator training to include actual performance of draindown training.

4) Evaluate the use of the see-through reactor for drain down training to 19

e include demonstration of industry and company events related to loss of inventory/standpipe and head vent problems.

D. CONTRIBUTING CAUSES

1. Work Practices - Shift team work practices did not meet expectations in that:

a. Shift personnel did not display the appropriate questioning attitude with respect to indications that were available and, if properly assessed, would have detected the loss of inventory or the status of the head vent.

b. Shift personnel failed to be constantly on the alert for the unexpected.

c. Shift personnel did not perform adequate reconciliation of changes in RCS inventory.

d. Shift Supervisor and STA involvement in shift activities was inadequate to maintain a broad perspective of unit activities focusing on core safety, and to identify that unexpected conditions existed.

Recommendations:

1) Enhance pre-outage training to include industry and company events which will help focus Operations personnel and STAs to be alert to plant conditions in the event that something unexpected is occurring or has occurred which could affect core safety.

2) Coach all licensed operators and STAs on the need to maintain a questioning attitude.

3) Coach all licensed operators and STAs in the need to self check when making inventory changes. This is to include defining what to expect prior to the change and then comparing the amount added or removed against the change in level.

4) Coach all Shift Supervisors and STAs on the expectation that they ensure they are sufficiently involved in shift activities to maintain a broad perspective of unit activities focusing on core safety, and the identification of unexpected plant conditions.

2. Written Communications - Shift team work practices did not meet expectations in that:

a. The mass balance procedure was ineffective in that it did not provide for reconciliation of individual inventory changes. Additionally, it did not properly account for all sources of inventory changes for the existing unit 20

conditions.

b. Isolation of the reactor vessel head vent was not documented on written turnover documents.

Recommendations:

1) Revise the mass balance procedure to:

a) Provide an attachment for the ROs to record all inventory changes, b) Provide for reconciliation of individual inventory changes, c) Account for any input from the opposite unit, draining of isolated loops, etc, d) Delete the current 10 gpm acceptance criteria. Each leak rate taken should receive a review by the Unit SRO and the Shift Supervisor regardless of the leak rate value.

2) Develop an Operations Standard which clearly defines management expectations for the conduct of shift turnovers including required written communications necessary to ensure quality shift turnovers.

E. CONTRIBUTING FACTORS

1. Work Practices - Loss of status of the reactor vessel head vent.

a. Failure of the Shift Supervisor and Unit SRO to properly control the activities of the shift. The tagout of the head vent was directed by the Desk SRO.

Even though the pre-job brief for the tagout of the head vent was conducted by the Desk SRO without the Shift Supervisor and Unit SRO present, both were cognizant of the evolution and did not exhibit sufficient command and control to ensure status of the head vent was known to shift personnel and the evolution was properly controlled.

Removal of significant components from service should only occur under the direction of the Unit SRO with concurrence of the Shift Supervisor.

b. The Reactor Operator failed to log that the reactor head vent valves were closed and did not inform the Unit SRO.

c. The Nightshift Desk SRO was informed, during turnover, the. status of the reactor vessel head vent and failed to: communicate the status to other shift team members at shift brief, pursue restoration of the head vent to service, and failed to turnover to his relief that the head vent had not been restored.

d. The Unit SRO and Shift Supervisor did not integrate the STA into shift activities.

21

e Recommendations:

1) Develop an Operations Standard which clearly defines management expectations for use of "Super Crews" which ensures that the roles of the Unit SRO, Desk SRO, Shift Supervisor and other SROs assigned to the shift team are well defined.

2) Develop an Operations Standard which clearly defines management expectations for equipment status control including: log keeping, acceptable methods for maintaining status of systems and components, and the communications required to ensure changes in equipment status are properly conveyed to shift personnel.

3) Include in LORP training, including outage readiness training, focus on STA responsibilities and expectations for the use of the STA in normal operating and outage situations.

2. Supervisory Methods - Command and control of shift activities by the Unit SRO and Shift Supervisor was inadequate to:

a. Ensure equipment status and plant conditions were known and understood by shift personnel

b. Ensure the level of activity in the control room was manageable for the experience and expertise of the shift team

c. Ensure plant conditions and main control room environment were appropriate for shift turnover.

Recommendations:

1 ) Develop an Operations Standard which clearly defines management expectations the for use of "Super Crews" and which ensures activities are performed under the direction and cognizance of both the Unit SRO and the Shift Supervisor.

2) Communicate these standards for command and control to the SROs and reinforce these expectations in simulator training.

3) Ensure these expectations are being met both on shift and in simulator training through Operations department line management observations.

Develop corrective actions plans to correct deficiencies observed.

4) Develop an Operations Standard which clearly defines management expectations for the conduct of shift turnovers including: the expectations 22

e for control room activity level and demeanor during turnover and the SRO's responsibility to enforce these expectations.

3. Written Communications - Lack of procedural control of the reactor vessel head vent.

a. The reactor vessel head vent was routinely isolated for cavity seal ring set in past outages without incident thereby creating a false sense of security.

The RCS was typically depressurized to containment atmosphere when this evolution was performed and it was the normal practice to use operations standby versus a tagout to control the isolation and subsequent reopening of the head vent piping to the containment atmosphere. Setting the cavity seal ring was considered a maintenance evolution and did not receive enough focus by shift personnel to ensure that: the time the head vent was out of service was monitored, maintenance personnel received. specific instructions on contacting the control room when the spool piece was reinstalled, and that the entire control room team was aware of the evolution.

b. RCE-91-06, "Surry Unit 2 Head Vent," Recommendation #4 was not fully implemented in that it recommended that a specific procedure be developed for removal and reinstallation of the reactor head vent spoolpiece to include specific Shift Supervisor sign-offs for maintaining a head vent path. Had this procedure been developed it would have required Shift Supervisor involvement in the event and may have increased the awareness of the head vent status.

c. Procedural controls for removal and reinstallation of the head vent spool piece did not include specific steps for the manipulation of the head vent isolation valves. These valves were controlled by either Operations standby or a tagout. Lack of procedural guidance contributed to this event in that the entire evolution was not procedurally guided to ensure completion of one task lead directly to the next task. Therefore, the only barrier to ensure the head vent was placed back in service was verbal communications.

d. Refueling activities are not procedurally controlled to ensure activities that are required to support one another are properly sequenced, or that precautions and limitations are completed or met prior to proceeding to the next activity. This lack of control does not ensure schedule changes are adequately controlled to ensure all prerequisite activities or conditions are met prior to performing the activity. Also, activities that involve multiple disciplines are not consistently conducted to ensure that the activities are performed in a timely manner and are controlled.

Recommendations:

23

1) Develop a specific Operations procedure for controlling manipulation of the reactor vessel head vent valves, that is directly linked to any other procedure which controls an evolution which requires the head vent valves be manipulated (such as spoolpiece removal). This procedure shall:

a) Standardize the method of control for the head vent isolation valves to ensure consistent performance of this evolution, b) Require a pre-job brief with all disciplines involved prior to taking the

head vent out of service, including the STA, c) Maintain RCS vessel vent path anytime RCS level is out of the pressurizer. If head vent is required to be removed from service then:

i. Establish methodology for monitoring RCS inventory during any period the head vent is out of service, ii. Declare the RCS standpipe inoperable with the head vent out of service, and iii. Require placement of out of service tag on the control room RCS standpipe indication, d) Require any change in head vent status be logged in the unit log.

2) Revise all maintenance procedures which control evolutions that require the head vent valves be manipulated to include a pre-job brief conducted in the annex. The procedures should be linked to the Operation's department procedure for head vent operations.

3) Develop a procedure to be used by Operations to include all the refueling activities to ensure Operations department maintains control of these activities, multi-discipline activities are properly orchestrated, and pre-requisite activities and plant conditions are met prior to authorizing activities.

24

Ill.

e *

SUMMARY

OF ROOT CAUSES, CONTRIBUTING CAUSES AND CONTRIBUTING FACTORS A. ROOT CAUSE:

TRAINING/QUALIFICATIONS - Operations personnel and Shift Technical Advisors (ST As) did not consider RCS standpipe level to be inoperable with the head vent isolated. Personnel interviewed believed that the standpipe would continue to provide reliable indication as long as RCS level changes were not made.

B. CONTRIBUTING CAUSES:

WORK PRACTICES - Shift personnel did not display the appropriate questioning attitude with respect to indications that were available and, if properly assessed, would have detected the loss of RCS inventory or the status of the head vent.

Shift personnel did not perform adequate self checks in reconciling inventory balances. Shift Supervisor and STA involvement in shift activities was inadequate to maintain a broad perspective of unit activities focusing on core safety, and to identify that unexpected conditions existed.

WRITTEN COMMUNICATIONS - The mass balance procedure was ineffective in that it did not provide for reconciliation of individual inventory changes.

Additionally, it did not properly account for all sources of inventory changes for the existing unit conditions. Isolation of the reactor vessel head vent was not documented in the unit log or included on written turnover documents.

C. CONTRIBUTING FACTORS:

WORK PRACTICES - Status control of the reactor head vent was lost in that neither the Reactor Operators nor the Unit SRO recorded the fact that the head vent was isolated in the unit log. Turnover was ineffective in that key members of the shift operating team were not made aware that the reactor vessel head vent was isolated.

SUPERVISORY METHODS - Command and control of shift activities by the Unit SRO and Shift Supervisor were inadequate to ensure that equipment status and plant conditions were known and understood by shift personnel and that plant conditions and the main control room environment were appropriate for shift turnover. The Unit SRO and Shift Supervisor did not integrate the STA into shift activities.

WRITTEN COMMUNICATIONS - There were no procedural controls to remove and return the head vent to service when the head vent spool piece was removed for cavity seal ring installation.

25

e IV. RECOMMENDATIONS A. The Training Department shall:

1. Conduct remedial training for Operations personnel and STAs on the interrelationship of the standpipe and head vent.

2. Revise continuing training lesson plans for Operations personnel and STAs to enhance discussion of the interrelationship between the standpipe and head vent, including simulator lesson plans.

3. Enhance pre-outage training to include industry and company events which will help focus Operations personnel and STAs to be alert to plant conditions in the event that something unexpected is occurring or has occurred which could affect core safety.

4. Include in LORP training, including outage readiness training, focus on STA responsibilities and expectations for the use of the STA in normal operating and outage situations.

5. Modify Initial RO/SRO license class integrated plant operations simulator training to include actual performance of draindown training.

6. Evaluate the use of the see-through reactor for drain down training to include demonstration of industry and company events related to loss of inventory/standpipe and head vent problems.

B. The Operations Department shall:

1. Develop Operations Standards which clearly define management expectations for:

Briefs - Pre-job, Tagout, Turnover, and Shift Super Crews - SRO & RO roles and responsibilities Equipment Status Control

2. Communicate these standards to Operations Personnel.

3. Monitor the performance of Operations personnel to ensure the standards are being met.

4. Coach all licensed operators and STAs on the need to maintain a questioning attitude.

5. Coach all licensed operators and STAs in the need to self check when making inventory changes. This is to include defining what to expect prior to the 26

change and then comparing the amount added or removed against the change in level.

6. Coach all Shift Supervisors and STAs on the expectation that they ensure they are sufficiently involved in shift activities to maintain a broad perspective of unit activities focusing on core safety, and the identification of unexpected plant conditions.

C. The Procedures Group shall:

1. Revise the mass balance procedure to:

a. Provide an attachment for the ROs to record all inventory changes,

b. Provide for reconciliation of individual inventory changes,

c. Account for any input from the opposite unit, draining of isolated loops, etc,

d. Delete the current 10 gpm acceptance criteria. Each leak rate taken should receive a review by the Unit SRO and the Shift Supervisor regardless of the leak rate value.

2. Develop a specific Operations procedure for controlling manipulation of the reactor vessel head vent valves, that is directly linked to any other procedure which controls an evolution which requires the head vent valves be manipulated (such as spoolpiece removal). This procedure shall:

a) Standardize the method of control for the head vent isolation valves to ensure consistent performance of this evolution, b) Require a pre-job brief with all disciplines involved prior to taking the head vent out of service, including the STA, c) Maintain RCS vessel vent path anytime RCS level is out of the pressurizer. If head vent is required to be removed from service then:

i. Establish methodology for monitoring RCS inventory during any period the head vent is out of service, ii. Declare the RCS standpipe inoperable with the head vent out of service, and iii. Require placement of out of service tag on the control room RCS standpipe indication, d) Require any change in head vent status be logged in the unit log.

3. Revise all maintenance procedures which control evolutions that require the head vent valves be manipulated to include a pre-job brief conducted in the annex. The procedures should be linked to the Operation's department procedure for head vent operations.

4. Develop a procedure to be used by Operations to include all the refueling activities to ensure Operations department maintains control of these 27

activities. multi-discipline activities are properly orchestrated. and pre-requisite activities and plant conditions are met prior to authorizing activities.

V. ENHANCEMENTS 1 . Evaluate installing a vacuum breaker on the pressurizer to be used during draindown to eliminate the operator's concern of drawing a vacuum in the RCS.

2. Evaluate augmentation of the STA resources on shift during critical segments of an outage to enhance STA focus on core safety.

3. Evaluate other methods of staffing "super crews" to ensure control room team, unit SRO, OATC, and BOP operator are optimum for working together. Also ensure that the desk SRO activities have checks and balances and tagging is controlled by unit SRO.

4. Evaluate adequacy of instrumentation necessary to ensure shift capability to perform accurate mass balance.

5. Evaluate adequacy of Operations Department standards on management expectations for communications, command and control, use of "super crews,"

and configuration management to ensure standards are clear and expectations are communicated to shift personnel.

6. Evaluate accountability for management meetings decisions & communications to ensure items are communicated from second line and above to first line supervision.

7. Evaluate maintenance procedure methodology to determine if including specific procedure steps for valve manipulations or other operations department activities (i.e. performance of MOPs, venting of the RCS to atmosphere, etc.), with appropriate operations signoff, would improve control of activities.

28

VI. PERIPHERAL ISSUES:

Peripheral issues identified during the information gathering phase of the root cause evaluation are identified below. It is the expectation of the RCE team that management review these items and utilize the corrective action process to disposition. The RCE team has obtained from each of them the information needed that relates to the causal factors for this root cause.

¥ Decision to draindown into the "dead zone" with an erratic/inoperable pressurizer level indication, previous unexplained phenomena with RVLIS and an unproven standpipe.

¥ Resolution of 1-RC-Ll-461 indication problems.

¥ Supervisor Shift Operations involvement in directing activities of the shift.

¥ Venting of the PRT to containment without a procedure or release permit.

¥ Release of the work orders for removal of pressurizer safeties and retraction of the flux thimbles in a plant configuration that would not support the maintenance.

¥ Draining of the PRT to the POTT to provide an additional vent path for the RCS without a procedure.

¥ Operations Department management expectations for procedure usage.

29

ML18152A125

Contents

Text

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

Navigation menu

ML18152A125

Text

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

Navigation menu

Search