BECO-86-062, Responds to NRC Confirmatory Action Ltr 86-10 for Restart of Plant Re 860404,11-12 Events at Plant

Responds to NRC Confirmatory Action Ltr 86-10 for Restart of Plant Re 860404,11-12 Events at Plant

ML20236E687
Person / Time
Site:	Pilgrim
Issue date:	05/15/1986
From:	Harrington W BOSTON EDISON CO.
To:	Murley T NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I)
Shared Package
ML20236E691	List: BECO-86-062, Responds to NRC Confirmatory Action Ltr 86-10 for Restart of Plant Re 860404,11-12 Events at Plant ML20236E712 ML20236E725
References
BECO-86-062, BECO-86-62, CAL-293-86-10, NUDOCS 8710290315
Download: ML20236E687 (285)

Text

_ _ _ _ _ _ _ _ _

BOSTON EDISON COMPANY GOD BOYLBTON STREET BOSTON, MAsSACHUBrns 02199

(

WILLIAM D. HARRlNOTQ H esusen viss posesosat a u = ua May 15, 1986 BEco Ltr. #86-062 Dr. Thomas E. Murley Regional Administrator U.S. Nuclear Regulatory Commission 631 Park Avenue - Region 1 King of Prussia, PA 19406

}

License No. DPR-35 Docket No. 50-293 i

SUBJECT:

Response to NRC Confirmatory Action Letter 86-10, Regarding the Events Which Occurred on April 4, 11-12, 1986, at Pilgrim Nuclear Power Station

Dear Dr. Murley:

This letter provides the written report required by the subject Confirmatory Action Letter prior to restart of Pilgrim Nuclear Power Station.

The course of our investigation ',r.cluded the assignment of evaluation teams to each of the three potential problem areas:

(1) Intersystem leakage through the motor-operated injection valves (including the check valve) of the residual heat removal system (hereinafter referred to as the "RHR Team");

(2) The primary containment isolation which occurred during shutdown after the reactor mode switch was repositioned from the run mode to the start-up mode (hereinafter referred to as the " Mode Switch Team");

(3) The failure of the outboard main steam isolation valves to reopen after resetting the primary containment isolation signal (hereinafter referred to as the "HSIV Team").

Our initial investigative efforts were conducted in close conference with members of your staff, an NRR representative, and an NRC AIT (Augmented Investigation Team). At the end of the NRC Team's two week inspection, it was concluded that Boston Edison's three evaluation teams had been appropriately approaching their individual areas of concern and should continue their root cause analysis in what was observed to be a thorough and diligent manner.

Those team efforts continued and the results of their ef forts are included as Attachments 1, 2, and 3 to this letter.

After the reports had been reviewed for technical accuracy by the Engineering Department, they were reviewed by the Operations Review Committee (ORC) for 8710290315 0460515 / ) {

PDR ADOCK05000g3 QQ

Dr. Thomas E. Murley 1 Page 2 '

cause, corrective actions, and safety significance. Subsequently, a BECo j management review team was formed and reviewed each report to disposition the recommendations and their alternatives. Those decisions are documented in Attachments 4, 5, and 6 to this letter.

These reports include the elements required by the subject Confirmatory Action '

Letter which are: ,

- Summary of events  !

- Assessment of events being analyzed to previous events

- Root cause determination (including methodology)

- Corrective actions taken/ recommended

- Basis for restart (including criteria used and analysis associated with these criteria)

The following is a team-by-team summary of root cause determinations and corrective actions taken and planned.

RHR Team Due to repeated annunciation of RHR High Pressure alarms as observed while starting up the plant on 4/10/86, it was suspected that two primary containment isolation valves the RHR outboard and inboard injection valves, (M0-1001-28B and 298) were leaking. Although the appropriate actions were I taken by the control room staff, there was no method of accurately quantifying leakage through the system. Therefore, a shutdown was initiated.

A historical assessment of the system was then conducted and included a review of a recent valve lineup change, written operating records, Failure and Malfunction Reports, the RHR flow recorder records, and equipment i

qualification modifications work done on the M0-1001-288 valve of the subject system.

Root cause was determined through an investigative approach that involved two perspectives. First, a chronological approach was used to analyze cause by examining the factors that could have contributed to a change when the trend in pressurization began to deviate from normal. Second, an approach that I considered the variables unique to the 'B' RHR injection loop (versus the ' A' l RHR injection loop) was useJ to analyze the intersystem leakage. The root i cause was subsequently determined to be either.  ;

l (1) degradation of the M0-1001-288 valve (2) improper seating of the MO-1001-688 (a check valve) coupled with improper local venting, or (3) a combination of causes (1) and (2).

The recommendations made by the team, which will be implemented and completed prior to start-up, include:

Disassembling the MO-1001-288 valve in order to facilitate its full inspection and subsequent rebuild (if necessary);

l

- D'r. ; Thomas. E. Murley.

IPage.3-a

' Mitigate the ef fect of the expected. . nominal rate of leakage by allowing a controlled, acceptable amount of leakage to exit the:

system; Install additional pressure. gauges and:make available system

1. ' ' ' temperature monitoring devices _to enhance our system monitoring

. capabilities; Remove a pressure gauge installed on the RHR injection line to )

. allow for a vent path; ,

Develop a system venting program which wi11' include' adequate vent location for local highpoints; Revise the RHR procedure to allow for control of pressurization and to provide enhanced assessment'and corrective action

• capabilities for the control room staf f.

p '.The team's recommendations to be dispositioned after start-up include:

Reduce the fhequency'of testing on both the M0-1001-28A&B valves and the 1001-68A&B valves to reduce system problems which' result from overtesting. This will be done via Technical Specifications change requests, as deemed necessary; Perform tests on the 1001-68 valves for pressure drop capability at a once per. refuel outage frequency; Study the available options to replace or redesign the check valves to provide positive position indication; Trend the surveillance history of the 400# RHR injection valve interlocks for reliability.

The leakage and associated effects of leakage, such as overheating of the piping, will have been adequately addressed through corrective actions taken

- or planned prior to start-up and described in this Attachment 4.

MSIV Team On April 4 and April 12, 1986, while reactor shutdowns were in progress, the reactor. operators tried to open the outboard MSIV's numerous times without success following a PCIS Group I isolation. The root cause analysis for the April 4 event concluded that the inability of the outboard MSIV's to open was due-to' low air pressure which resulted in partial opening. The symptoms exhibited during the April.12 event were identical to those experienced on

April 4. However, on April 12, additional data on the MSIV performance was obtained due to our heightened awareness of the problem and indicated that the outboard MSIV stems moved upward only approximately 1/2 inch and then j stopped. After reviewing previous scram reports and the refueling outage #6 i records, it became apparent the problem had occurred on April 4 and April 12, 1986 but apparently at no other time in PNPS operating history.

i (L..., I

, 1 o

1 Dr. Thomas'E.-Murley lPage 4' .

The team then utilized the-Kepner-Tregoe (K-T) change analysis technique which j Lis a1 systematic approach to problem solving. The system involves comparing a.

/ problem-freeisituai, ion with a problem situation in order to: isolate causes and effects of change.. Through further analysis, the team determined that the- ,

-pilot poppet disassociation from their stems was the most probable mode of

. failure. ' A test was subsequently conducted to test this theory and indeed, it became apparent.that the stem and pilot assemblies were no longer

attached as. designed.

-Subsequently, all eight MSIV's were disassembled and examined. .A combination-of visual examinations, dimensional checks, and further analysis indicated that the. rotational and vibrational, forces present, are. apparently strong l i

enough to cause the pilot and lock nut' assembly to disassociate. 1 1

The three contributing factors to the disassociation were-therefore determined I

'to be in: d j

(1) the design, .I l

(2) the assembly process. and

]

1 (3) the~gove ning pre iure The recommendations made by the team, which will. be completed prior to j start-up,'are:

A new design PDC 86-28 which will provide the redundancy of two set screws locking the pilot poppet via counterbored machined land

~

areas'on the pilot poppet nut'. { ;

I The assembly process now includes specific, sequenced steps with )

measurements-and torques specified, resulting in a governing 1 procedure'with discrete steps, and quantified values being 1 verified.

Mode Switch Team On. April 12, 1986, during an orderly shutdown, the turbine generator was tripped of f line and the power descent continued. Before the APRM downscale j alarms and main steam line low pressure alarms were reached, the Reactor Mode '1 Switch was moved from "Run" to the."Startup" position. As the orderly ~i shutdown continued, the main steam line low pressure (less than 880 psi in

~

{

Run) alarm Channel "B" was received and a short time later the main steam line

. low pressure (less than 880 psi in Run) alarm channel "A" was received.  ;

. Approximately 30 to 40 seconds af ter the Channel "A" was received, the Main l Steam-Isolation Valves went closed causing a Reactor ' Scram. .This event is j similar to an event on April 4,1986, as discussed in the team report. l L  !

L An evaluation of:the similarity of the. subject events to previous events was H conducted by the Mode Switch evaluation team. The evaluation showed that (4)

L MSIV closures have occurred previous to the April 4 event with the plant in other than the "run" node. The dates of those events were 10/3/72, 7/27/75, 4/9/78 und 3/22/86, 1

l I i

r  ?

Dr. Thomas E.' Murley Page 5 Using the Kepner Tregoe (K-T) change analysis process and other investigative techniques, exhaustive testing and evaluation of the potential root causes was performed by the team. All tests resulted in the instrumentation and associated circuitry being shown to operate properly. Physical inspection and l

walkdowns of the relays, contacts and wiring identified loose grounding connection on the PCIS relays and instrument sensitivity to vibration could have contributed to the subject event. Other items which could not be eliminated as potential root causes are the reactor water level instrumentation and the Reactor Mode Switch. To address the mode switch concern, Boston Edison will replace the unit with an improved, G.E. l recommended model prior to start-up. This action is also in response to the l G.E. SIL issued on this sebject.

Per the Mode Switch team's recommendation, the plant will conduct a limited power ascent and descent through the problem area. This approach was agreed to by INPO representatives who concurred that this method of testing would be the only appropriate method to determine root cause. In order to perform this testing, we would need to secure your approval for re-start. Under our current shutdown condition, it is impossible to duplicate the conditions that would lead to possible excess vibration, or water level transients, since they can only be duplicated during plant operation. Performance of such a test should indicate whether water level perturbations or instrument rack sensitivity to vibration were contributors to the spurious isolation signals.

The test coupled with the extensive monitoring of PCIS logic circuits should identify the source of the PCIS initiation if it has not already been corrected.

General Electric has been contracted to supply the Transient Analysis Recording System (GETARS) for this startup test. This monitoring equipment will be lef t in place and utilized to supply performance monitoring evaluation data until the EPIC computer becomes operational.

During this test, twenty-eight individual digital signals will be monitored in panels C915 and C917. These signals will be one of each of the following from the four PCIS sub-channels:

Reactor Low Low Water Level Steam Tunnel High Temperature Steam Line High Flow Main Steam Line Low Pressure j PCIS Group I Isolation Relay l Reactor High Water Level 1 Main Steam Line High Radiation i These signals will be extracted f rom unused " dry contacts" on the associated relays, wired out to terminal blocks and monitored by the GETARS.

1 Additionally, eight resistive voltage divider circuits (two per PCIS l sub-channel) will monitor mode switch bypass circuitry.

i