ML18096A238

Auxiliary Feedwater System RISK-BASED Inspection Guide for the Salem Nuclear Power Plant

ML18096A238
Person / Time
Site:	Salem
Issue date:	08/31/1991
From:	GORE B F, PUGH R, VO T V Battelle Memorial Institute, PACIFIC NORTHWEST NATION
To:	Office of Nuclear Reactor Regulation
References
CON-FIN-L-1310 NUREG-CR-5761, PNL-7518, NUDOCS 9109050280
Download: ML18096A238 (36)

Text

.----* NUREG/CR-5761 PNL-7518 Auxiliary Feedwater System Risk-Based Inspection Guide for the Salem Nuclear Power Plant Prepared by R. Pugh, B. F. Gore, T. V. Vo Pacific Northwest Laboratory Operated by Battelle Memorial Institute Prepared for U.S. Nuclear Regulatory Commission

---

Docket #.5;;:...-zz;z_

Accession Cf/OC/

1

.. :. . . . of-Ltr-----"-)

I Reoulatmv Dooket File I I I I i i I I I I I I I -NOTICE-THE ATTACHED FILES ARE OFFICIAL RECORDS OF THE INFORMATION

& REPORTS MANAGEMENT BRANCH. THEY HAVE BEEN CHARGED TO YOU FOR A LIMITED TIME PERIOD AND MUST BE RETURNED TO THE RE* CORDS & ARCHIVES SERVICES TfON P1-22 WHITE FLINT. PLEASE.DO NOT SEND DOCUMENTS CHARGED OUT THROUGH THE MAIL. REMOVAL OF ANY PAGE(S) FROM DOCUMENT FOR REPRODUCTION MUST BE FERRED TO FILE PERSONNEL.

I ---1 -I ...... 9109050280 910831 PDR ADOCK 05000272 Q PDR \ ) -NOTICE-'* )' ,.,

I_ "i\. AVAILABILITY NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited In NRC publications wlll be avallable from one of the following sources: 1. The NRC Public Document Room, 2120 L Street, NW., Lower Level, Washington, DC 20555 2. The Superintendent of Documents, U.S. Government Printing Office, P.O. Box 37082, Washington, DC 20013-7082

3. The National Technical Information Service, Springfield, VA 22161 Although the listing that follows represents the majority of documents cited In NRC publications, It Is not Intended to be exhaustive.

Referenced documents available for Inspection and copying for a fee from the NRC Public Document Room Include NRC correspondence and Internal NRC memoranda; NRC bulletins, circulars, Information notices, Inspection and Investigation notices; licensee event reports; vendor reports and correspondence; sion papers; and applicant and licensee documents and correspondence.

The following documents In the NUREG series are avallable for purchase from the GPO Sales Program: formal NRC staff and contractor reports, NRC-sponsored conference proceedings, and NRC booklets and brochures.

Also available are regulatory guides, NRC regulations In the Code of Federal Regulations, and Nuclear Regulatory Commission Issuances.

Documents available from the National Technical Information Service Include NUREG-serles reports and technical reports .prepared by other Federal agencies and reports prepared by the Atomic Energy sion, forerunner agency to the Nuclear Regulatory Commission.

Documents available from public and special technical libraries include all open literature Items, such as books, journal articles, and transactions.

Federal Register notices, Federal and State legislation, and gressional reports can usually be obtained from these libraries.

Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference ceedings are available for purchase from the organization sponsoring the publication cited. Single copies of NRC draft reports are available free, to the extent of supply, upon written request to the Office of Administration, Distribution and Mall Services Section, U.S. Nuclear Regulatory Commission, Washington, DC 20555. Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, for use by the public. Codes and standards are usually copyrighted and may be purchased from the originating organization or, If they are American National Standards, from the American National Standards Institute, 1430 Broadway, New York, NY 10018. DISCLAIMER NOTICE This report was prepared as an account of work sponsored by an ageney of the United States Government.

Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or assumes any legal liability of responsibility for any third party's use, or the results of such use, of any information, apparatus, product or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights. f Auxiliary Feedwater System Risk-Based Inspection Guide for the Salem Nuclear Power Plant Manuscript Completed:

July 1991 Date Published:

August 1991 Prepared by R. Pugh, B. F. Gore, T. V. Vo Pacific Northwest Laboratory Richland, WA 99352 Prepared for Division of Radiation Protection and Emergency Preparedness Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, DC 20555 NRC FIN Ll310 NUREG/CR-5761 PNL-7518

SUMMARY

This document presents a compilation of auxiliary feedwater

{AFW) system failure information which has been screened for risk significance in terms of failure frequency and degradation of system performance.

It is a prioritized listing of failure events and their causes that are significant enough to warrant consideration in inspection planning at Salem Generating Station. This information is presented to provide inspectors increased resources for inspection planning at Salem Generating Station. The information presented is specific to Unit 2, however it may also be applied to Salem Unit 1, as the AFW systems are nearly identical.

The risk importance of various component failure modes was identified by analysis of the results of probabilisitic risk assessme.nts (PRAs) for many pressurized water reactors (PWRs). However, the component failure categories identified in PRAs are rather broad, because the failure data used in the PRAs is an aggregate of many individual failures having a variety of root causes. In order to help inspectors to focus on specific aspects of component operation, maintenance and design which might cause these failures, an extensive review of component failure information was performed to identify the rank and root causes of these component failures.

Both Salem and industry-wide failure information was analyzed.

Failure causes were sorted on the basis of frequency of occurrence and seriousness of consequence, and categorized as common cause failures, human errors, design problems, or component failures.

This information is presented in the body of this document.

Section 3.0 provides brief descriptions of these risk-important failure causes, and Section 5.0 presents more extensive discussions, with specific examples and references.

The entries in the two sections are cross-referenced.

An abbreviated system walkdown table is presented in Section 3.2 which includes only components identified as risk important.

This table lists the system lineup for normal, standby system operation.

This information permits an inspector to concentrate on components important to the prevention of core damage. However, it is important to note that inspections should not focus exclusively on these components.

Other components which perform essential functions, but which are not included because of high reliability or redundancy, must also be addressed to ensure that degradation does not increase their failure probabilities, and hence their risk importances.

iii -I l CONTENTS

SUMMARY

• * * * * * * * * * * * * * * * * * * * . * * * * . * . * * * * * * * * * * * * * * . * * * * * * * * * * * * * * * * * * * *

• i 11 1. 0 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2. 0 SALEM AFW SYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2.1 SYSTEM

DESCRIPTION

...........................................

2 2.2 SUCCESS CRITERION

............................................

4 2.3 SYSTEM DEPENDENCIES

..........................................

4 2. 4 OPERATIONAL CONSTRAINTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3.0 INSPECTION

GUIDANCE FOR THE SALEM AFW SYSTEM ......................

5 3.1 RISK IMPORTANT AFW COMPONENTS AND FAILURE MODES ..............

5 3.1.1 MULTIPLE PUMP FAILURES DUE TO COMMON CAUSE ............

5 3.1.2 TURBINE DRIVEN PUMP 23 FAILS TO START OR RUN ..........

6 3.1.3 MOTOR DRIVEN PUMP 21 OR 22 FAILS TO START OR RUN . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

• . . . . . . . . . . . . . . . . . .

7 3.1.4 PUMP 21, 22 OR 23 UNAVAILABLE DUE TO MAINTENANCE OR SURVEILLANCE . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3.1.5 AIR OPERATED LEVEL CONTROL VALVES FAIL CLOSED .........

7 3.1.6 MANUAL SUCTION OR DISCHARGE VALVES FAIL CLOSED ........ 7 3.1.7 LEAKAGE OF HOT FEEDWATER THROUGH CHECK VALVES .........

8 3.2 RISK IMPORTANT AFW SYSTEM WALKDOWN TABLE ............... ..... 9 4.0 GENERIC RISK INSIGHTS FROM PRAs ...................................

14 4.1 RISK IMPORTANT ACCIDENT SEQUENCES INVOLVING AFW SYSTEM FAILURE ...............................................

14 4.2 RISK IMPORTANT COMPONENT FAILURE MODES ........*..............

15 v

• I I 1 CONTENTS (Continued)

5.0 FAILURE

MODES DETERMINED FROM OPERATING EXPERIENCE

................

16 5.1 SALEM EXPERIENCE

.............................................

16 5.1.1 AFW PUMP CONTROL LOGIC, INSTRUMENTATION AND ELECTRICAL FAILURES ...................................

16 5.1.2 AFW PUMP FAILURES .....................................

16 5.1.3 FAILURE OF AFW PUMP DISCHARGE LEVEL CONTROL VALVES ................................................

16 5.1.4 OTHER AFW VALVE FAILURES ..............................

16 5 .1 . 5 HUMAN ERRORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17 5.2 INDUSTRY WIDE EXPERIENCE

.....................................

17 5.2.1 COMMON CAUSE FAILURES .................................

17 5. 2. 2 HUMAN ERRORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20 5.2.3 DESIGN/ENGINEERING PROBLEMS AND ERRORS ................

20 5.2.4 COMPONENT FAILURES ....................................

22 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

25 vi .

---

1.0 INTRODUCTION

This document is the fifth of a series providing plant-specific inspection guidance for auxiliary feedwater (AFW) systems at pressurized water reactors (PWRs). This guidance is based on information from probabilistic risk assessments (PRAs) for similar PWRs, industry-wide operating experience with AFW systems, plant-specific AFW system descriptions, and plant-specific operating experience.

It is not a detailed inspection plan, but rather a compilation of AFW system failure information which has been screened for risk significance in terms of failure frequency and degradation of system performance.

The result is a risk-prioritized listing of failure events and their causes that are significant enough to warrant consideration in inspection planning at Salem Nuclear Generating Station. Information presented is specific to Salem Unit 2, however it may also be applied to Salem Unit 1, as the AFW systems are nearly identical.

This inspection guidance is presented in Section 3.0, following a description of the Salem Unit 2 AFW system in Section 2.0. Section 3.0 identifies the risk important system components by Salem identification number, followed by brief descriptions of each of the various failure causes of that component.

These include specific human errors, design deficiencies, and hardware failures.

The discussions also identify where common cause failures have affected multiple, redundant components.

These brief discussions identify specific aspects of system or component design, operation, maintenance, or testing for inspection by observation, records review, training observation, procedures review, or by observation of the implementation of procedures.

An AFW system table identifying risk important components and_ their lineup for normal, standby system operation is also provided.

The remainder of the document describes and discusses the information used in compiling this inspection guidance.

Section 4.0 describes the risk importance information which has been derived from PRAs and its sources. As review of that section will show, the failure categories identified in PRAs are rather broad (e.g., pump fails to start or run, valve fails closed). Section 5.0 addresses the specific failure causes which have been combined under these categories.

AFW system operating history was studied to identify the various specific failures which have been aggregated into the PRA failure mode categories.

Section 5.1 presents a summary of Salem failure information, and Section 5.2 presents a review of industry-wide failure information.

The industry-wide information was compiled from a variety of NRC sources, including AEOD analyses and reports, information notices, inspection and enforcement bulletins, and generic letters, and from a variety of INPO reports as well. Some Licensee Event Reports and NPRDS event descriptions were also reviewed individually.

Finally, information was included from reports of NRC-sponsored studies of the effects of plant aging, which include quantitative analyses of reported AFW system failures.

This industry-wide information was then combined with the plant-specific failure information to identify the various root causes of the PRA failure categories, which are identified in Section 3.0. 1

---

---

2.0 SALEM UNIT 2 AFW SYSTEM This section presents an overview description of the Salem AFW system, including a simplified schematic system diagram. In addition, the system success criterion, system dependencies, and administrative operational constraints are also presented.

2.1. System

Description The AFW system provides feedwater to the steam generators (SG) to allow secondary-side heat removal from the primary system when main feedwater is unavailable.

The system is capable of functioning for extended periods, which allows time to restore main feedwater flow or to proceed with an orderly cooldown of the plant to where the residual heat removal (RHR) system can remove decay heat. A simplified schematic diagram of the AFW system is shown in Figure 2.1. The system is designed to start up and establish flow after receipt of an automatic start signal. All pumps start on receipt of a steam generator low level signal. (The motor driven pumps start on low level in one SG, whereas, two low level signals are required for a turbine driven pump start.) The motor driven (MD) pumps start on a trip of both MFW pumps or a safety injection signal and the single turbine driven (TD) pump starts on undervoltage on either of the 4-KV buses, loss of 125VDC, or loss of control air sources. The normal AFW pump suction is from the auxiliary feedwater storage tank (AFWST). Each pump from a common header through a locked-open isolation valve and a check valve. Power, control, and instrumentation associated with each motor-driven pump are independent from one another. Steam for the turbine driven pump is supplied by steam generators 21 and 23 from a point upstream of the main steam isolation valves, through valve 2MS132. Each AFW pump is equipped with a continuous recirculation flow system, which prevents pump deadheading.

Each auxiliary feedwater pump discharge is provided with a check valve. Each motor-driven pump normally supplies feedwater to only two steam generators, but the headers may be cross-connected.

The turbine-driven pump normally supplies all four steam generators.

Manual isolation valves are located downstream of the pump discharge check valve in the lines to each of the steam generators, each followed by an air operated level control valve, a check valve, another manual isolation valve, and a final stop-check valve. The AFWST is the normal source of water for the AFW System and is required to store sufficient demineralized water to maintain the reactor coolant system (RCS) at hot standby conditions for 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> with steam discharge to atmosphere.

Alternate sources of water for the AFW system include the supply for the alternate unit, the Demineralized Water Storage Tanks, the Fire Water Storage Tanks, and the Service Water System. 2 w Fig. Z.1 SALEM AUXILIARY FEEDWATEH SIMPLIFIED SYSTEM DIAGRAM r.w. AtfJ mE l'ROT. SER\'1CE WATER AUX ruo ...L STOOAOE -,-TAii< r -23AF52 23 22"F-"8 zz

• • _2-tiAf923 22"F923 2mr1t . . .

L 2.2 Success Criterion System success requires the operation of at least one pump supplying rated flow to at least two of the four steam generators.

2.3 System

Dependencies The AFW system depends on AC power for motor-driven pumps and level control valves, DC power for control power to pumps and valves, and an automatic actuation signal. Instrument Air is required for the pneumatic flow control valves. In addition, the turbine-driven pump also requires steam availability.

2.4 Operational

Constraints When the reactor plant is in Mode 1, 2, or 3 the Salem Technical Specifications require that all three AFW pumps and associated flow paths are operable with each motor-driven pump powered from a different vital bus. If one AFW pump becomes inoperable, it must be restored to operable status 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> or the plant must shut down to hot standby within the next six hours. If two AFW pumps are inoperable, the plant must be shut down to hot standby within six hours. With three AFW pumps inoperable, corrective action to restore at least one pump to operable status must be initiated immediately.

The Salem Technical Specifications require that 200,000 gallons of water are stored in the AFWST, with back up supplies maintained in the demineralized water storage tank (200,000 gallons), and in the fire protection/domestic water storage tank. 4

3.0 INSPECTION

GUIDANCE FOR THE SALEM AFW SYSTEM In this section the risk important components of the Salem AFW system are identified, and the important modes by which they are likely to fail are briefly described.

These failure modes include specific human errors, design problems, and types of hardware failures which have been observed to occur for these types of components, both at Salem and at PWRs throughout the nuclear industry.

The discussions also identify where common cause failures have affected multiple, redundant components.

These brief discussions identify specific aspects of system or component design, operation, maintenance, or testing for observation, records review, training observation, procedures review or by observation of the implementation of procedures.

Table 3.1 is an abbreviated AFW system walkdown table which identifies risk important components.

This table lists the system lineup for normal, standby system operation.

Inspection of the components identified addresses essentially all of the risk associated with AFW system operation.

3.1 Risk Important AFW Components and Failure Modes Common cause failures of multiple pumps are the most risk-important failure modes of AFW system components.

These are 'roll owed in importance by single pump failures, -level control valve failures, and individual check valve backleakage failures.

The following sections address each of these failure modes, in decreasing order of importance.

They present the important causes of these component failure modes which have been distilled from historical records. Each item is keyed to discussions in Section 5.2 which present additional information on historical events. 3.1.1 Multiple Pump Failures due to Common Cause The following listing summarizes the most. important multiple-pump failure modes identified in Section 5.2.1, Common Cause Failures, and each item is keyed to entries in that section .. Incorrect operator intervention into automatic system functioning, including improper manual starting and securing of pumps, has caused failure of all pumps, including overspeed trip on startup, and inability to restart prematurely secured pumps. CCI.

• Valve mispositioning has caused failure of all pumps .. Pump suction, steam supply, and instrument isolation valves have been involved.

CC2. Steam binding has caused failure of multiple pumps. This resulted from leakage of hot feedwater past check valves into a common discharge header, with several valves involved including a motor-operated discharge valve. (See item 7 below.) CCIO. Multiple-pump steam binding has also resulted from improper valve lineups, and from running a pump deadheaded.

CC3. 5 Pump control circuit deficiencies or design modification errors have caused failures of multiple pumps to auto start, spurious pump trips during operation, and failures to restart after pump shutdown.

CC4. Incorrect setpoints and control circuit calibrations have also prevented proper operation of multiple pumps. CC5. Loss of a vital power bus has failed both the turbine-driven and one motor-driven pump due to loss of control power to steam admission valves or to turbine controls, and to motor controls powered from the same bus. CC6. Simultaneous startup of multiple pumps has caused oscillations of pump suction pressure causing multiple-pump trips on low suction pressure, despite the existence of adequate static net positive suction head (NPSH). CC7. Design reviews have identified inadequately sized suction piping which could have yielded insufficient NPSH to support operation of more than one pump. cca. 3.1.2 Turbine Driven Pump 23 Fails to Start or Run Improperly adjusted and inadequately maintained turbine governors have caused pump failures, both at Salem and elsewhere.

HE2. Problems include worn or loosened nuts, set screws, linkages or cable connections, oil leaks and/or contamination, and electrical failures of resistors, transistors, diodes and circuit cards, and erroneous grounds and connections.

CFS. Terry turbines with Woodward Model EG governors have been found to overspeed trip if full steam flow is allowed on startup. Sensitivity can be reduced if a startup steam bypass valve is sequenced to open first. DEi. Condensate slugs in steam lines have caused turbine overspeed trip on startup. Tests repeated right after such a trip may fail to indicate the problem due to warming and clearing of the steam lines. Surveillance should exercise all steam supply connections.

DE2. Trip and throttle valve (2MS52) problems which have failed the turbine driven pump include physically bumping it, failure to reset it following testing, and failures to verify control room indication of reset. HE2. Whether either the overspeed trip or TTV trip can be reset without resetting the other, indication in the control room of TTV position, and unambiguous local indication of an overspeed trip affect the likelihood of these errors. DE3. Turbines with Woodward Model PG-PL governors have tripped on overspeed when restarted shortly after shutdown, unless an operator has locally exercised the speed setting knob to drain oil from the governor speed setting cylinder (per procedure).

Automatic oil dump valves are now available through Terry. DE4. .6

3.1.3 Motor

Driven Pump 21 or 22 Fails to Start or Run Control circuits used for automatic and manual pump starting are an important cause of motor driven pump failures, as are circuit breaker failures.

CF7. Similar failures have occurred at Salem. Mispositioning of handswitches and procedural deficiencies have prevented automatic pump start. HE3. Low lubrication oil pressure resulting from heatup due to previous operation has prevented pump restart due to failure to satisfy the protective interlock.

DES. 3.1.4 Pump 21. 22 or 23 Unavailable Due to Maintenance or Surveillance Both scheduled and unscheduled maintenance remove pumps from operability.

Surveillance requires operation with an altered line-up, although a pump train may not be declared inoperable during testing. Prompt scheduling and performance of maintenance and surveillance minimize this unavailability.

3.1.5 Air Operated Level Control Valves Fail Closed TD Pump Trains : 21AFII. 22AFII, 23AFII. 24AFII MD Pump Trains : 21AF21, 22AF21. 23AF21, 24AF21 These normally-open air operated valves (AOVs) control flow to the steam generators.

They fail open on loss of Instrument Air. Control circuit problems have been a primary cause of failures, both at Salem and elsewhere.

CF9. Valve failures have resulted from blown fuses, failure of control components (such as current/pneumatic converters), broken or dirty contracts, misaligned or broken limit switches, control power loss, and calibration problems.

Degraded operation has also resulted from improper air pressure due to air regulator failure or leaking air lines. CFII. Out-of-adjustment electrical flow controllers have caused improper valve operation, affecting multiple trains of AFW. CC12. Leakage of hot feedwater through check valves has caused thermal binding of flow control MOVs. AOVs may be similarly susceptible.

CF2. Multiple flow control valves have been plugged by clams when suction switched automatically to an alternate, untreated source. CC9. 3.1.6 Manual Suction or Discharge Valves Fail Closed 7 TD Pump 23: MD Pump 21: MD Pump 22: Valves 23AF3. 22AF10, 22AF86, 24AF10, 24AF86, 21AF10, 21AF86, 23AF10, 23AF86 Valves 21AF3, 24AF20, 24AF22, 23AF20, 23AF22 Valves 22AF3, 21AF20, 21AF22, 22AF20, 22AF22 These manual valves are normally locked open. For each train, closure of the first valve listed would block suction fromt he AFWST. Closure of the following valves block AFW discharge to the S/G indicated by the first two digits of the valve number, i.e. valve 24AF10 would block flow to #24 S/G. Valve mispositioning has resulted in failures of multiple trains of AFW. CC2. It has also been the dominant cause of problems identified during. operational readiness inspections.

HEl. Events have occurred most often during maintenance, calibration, or system modifications.

Important causes of mispositioning include: Failure to provide complete, clear, and specific procedures for tasks and system restoration Failure to promptly revise and validate procedures, training, and diagrams following system modifications Failure to complete all steps in a procedure Failure to adequately review uncompleted procedural steps after task completion Failure to verify support functions after restoration Failure to adhere scrupulously to administrative procedures regarding tagging, control and tracking of valve operations Failure to log the manipulation of sealed valves Failure to follow good practices of written task assignment and feedback of task completion information Failure to provide easily read system drawings, legible valve labels corresponding to drawings and procedures, and labeled indications of local valve position 3.1.7 Leakage of Hot Feedwater through Check Valves: Between Pump 23 and MFW: Valves 21AF921, 22AF921. 23AF921, 24AF921 Between Pump 21 and MFW: Valves 23AF920. 24AF920 Between Pump 22 and MFW: Valves 21AF920, 22AF920 At Pump Discharges:

Valves 21AF8, 22AF8, 23AF8 Leakage of hot feedwater through check valves in series has caused steam binding of multiple pumps. Leakage through a closed level control valve in series with check valves has also occurred, as would be required for leakage to reach the motor driven pumps 21 and 22. CClO. Slow leakage past the final check valve of a series may not force upstream check valves closed, allowing leakage past each of them in turn. Piping orientation and valve design are important factors in achieving true series protection.

CFl. 8 3.2 Risk Important AFW System Walkdown Table Table 3.1 presents an AFW system walkdown table for Salem Unit 2 including only components identified as risk important.

The lineup indicated is for normal power operation.

This information allows inspectors to concentrate their efforts on components important to prevention of core damage. However, it is essential to note that inspections should not focus exclusively on these components.

Other components which perform essential functions, but which are absent from this table because of high reliability or redundancy, must also be addressed to ensure that their risk importances are not increased.

Examples include the (open) steam lead isolation valves upstream of 2MS132, an adequate water level in the AFWST, and the (closed) valves cross connecting the discharges of the two motor-driven AFW pumps. 9 L

1. 21 22 2AF92 21AF3 21AF10 21AF11 21AF20 21AF21 21AF22 21AF23 Table 3.1. Risk Important AFW System Walkdown Table Salem Unit 2 Component Name Electrical Motor-Driven Pump Motor-Driven Pump Valve AF B/U FROM SW AF PMP 21 SUCT VALVE AF PMP 23 DISCH TO SG 21 SG 21 AF LVL CONT VALVE SG 21 AF CNTL ISOL VALVE SG 21 AF LVL CONT VALVE SG 21 AF CNTL VALVE ISOL SG 21 AF INLET STOP CHECK 10 Required Position Racked In/ Closed Racked In/ Closed LOCKED CLOSED LOCKED OPEN LOCKED OPEN OPEN HANDWHEEL LOCKED IN NEUTRAL AIR TO OPERATOR LOCKED OPEN CLOSED HANDWHEEL LOCKED IN NEUTRAL AIR TO OPERATOR LOCKED OPEN LOCKED OPEN Actual Position t*

Component

1. 21AF40 21AF52 21AF86 21AF923 22AF3 22AF10 22AF11 22AF20 22AF21 22AF22 22AF23 22AF40 Table 3.1. Risk Important AFW.System Walkdown (Continued)

Component Name AF PMP 21 RECIRC VALVE AF PMP 21 B/U SUCT VALVE AF PMP 23 DISCH VALVE TO SG 21 AFP 21 & 22 DISCH X-CONN AF PMP 22 SUCT VALVE AF PMP 23 DISCH TO SG 22 SG 22 AF LVL CONT VALVE SG 22 AF CONT ISOL VALVE SG 22 AF LVL CONT VALVE SG 22 AF CONT VALVE ISOL VALVE SG 22 AF INLET STOP CHECK VALVE AF PMP 22 RECIRC VALVE 11 Required Position OPEN AIR TO OPERATOR TAGGED CLOSED LOCKED OPEN CLOSED LOCKED OPEN LOCKED OPEN OPEN HANDWHEEL LOCKED IN NEUTRAL AIR TO OPERATOR LOCKED OPEN CLOSED HANDWHEEL LOCKED IN NEUTRAL AIR TO OPERATOR LOCKED OPEN LOCKED OPEN OPEN AIR TO OPERATOR Actual Position *.* '**;' **!.J. ,' *:. ., .

Table 3.1. Risk Important AFW System Walkdown Table (Continued)

Required Actual Component

1. Component Name Position Position 22AF52 AF PMP 22 B/U SUCT VALVE TAGGED CLOSED 22AF86 AF PMP 23 DISCH VALVE TO SG 22 LOCKED OPEN 22AF923 AFP 21 & 22 DISCH X-CONN CLOSED 23AF3 AF PMP 23 SUCT VALVE LOCKED OPEN 23AF10 AF PMP 23 DISCH TO SG 23 LOCKED OPEN 23AF11 SG 23 AF LVL CONT VALVE OPEN HANDWHEEL LOCKED IN NEUTRAL AIR TO OPERATOR 23AF20 SG 23 AF CONT ISOL VALVE LOCKED OPEN 23AF21 SG 23 AF LVL CONT VALVE CLOSED HANDWHEEL LOCKED IN NEUTRAL AIR TO OPERATOR 23AF22 SG 23 AF CONT VALVE ISOL VALVE LOCKED OPEN 23AFZ3 SG 23 AF INLEl STOP CHECK VALVE LOCKED OPEN 23AF52 AF PUMP 23 B/U SUCT VALVE TAGGED CLOSED 23AF86 AF PMP 23 DISCH VALVE TO SG 23 .LOCKED OPEN 24AF10 AF PMP 23 DISCH TO SG 24 LOCKED OPEN ;,;< ' . ' 12 Table 3.1. Risk Important AFW System Walkdown Table (Continued)

Required Actual Component

1. Component Name Position Position 24AF11 SG 24 AF LVL CONT VALVE OPEN HANDWHEEL LOCKED IN NEUTRAL AIR TO OPERATOR 24AF20 SG 24 AF CONT ISOL VALVE LOCKED OPEN 24AF21 SG 24 AF LVL CONT VALVE CLOSED HANDWHEEL LOCKED IN NEUTRAL AIR TO OPERATOR 24AF22 SG 24 AF CONT VALVE ISOL LOCKED OPEN 24AF23 SG 24 AF INLET STOP CHECK VALVE LOCKED OPEN 24AF86 AF PMP 23 DISCH VALVE TO SG 24 LOCKED OPEN TURBINE VALVES 21MS45 STEAM TO 23 AFW TURBINE LOCKED OPEN 23MS45 STEAM TO 23 AFW TURBINE LOCKED OPEN 2MSl32 TURBINE STARTUP VALVE CLOSED AIR TO OPERATOR -2MS52 TURBINE TRIP VALVE . OPEN/RESET 2MS53 TURBINE GOVERNOR VAlVE OPERABLE 2AF42 23 AFW PUMP MINIMUM FLOW VALVE LOCKED OPEN 2AFI03 23 AFW PUMP MINIMUM FLOW VALVE LOCKED OPEN . ' 13

.4.0 GENERIC RISK INSIGHTS FROM PRAs PRAs for 13 PWRs were analyzed to identify risk-important accident sequences involving loss of AFW, and to ident.ify and .risk-prioritize the component failure modes involved.

The results of thii analysis are in this section. They are consistent with resu.lts reported by INEL and BNL (Gregg et al 1988, and Travis et al, 1988}. 4.1 Risk Important Accident Sequences Involving AFW System Failure Loss of Power System A loss of offsite power is followed by failure of AFW. Due to lack of actuating power, the PORVs cannot be opened, preventing adequate feed-and-bleed cooling, and resulting in core damage. A station blackout fails all AC power except Vital AC from DC inverters, and all decay heat removal systems except the driven AFW pump. AFW subsequently fails due to battery depletion or hardware failures, resulting in core damage. A DC bus fails, causing a trip and failure of the power conversion system. One AFW motor-driven pump is failed by the bus loss, and the turbine-driven pump fails due to loss of turbine or valve control power. AFW is subsequently lost completely due to other failures.

Feed-and-bleed cooling fails because PORV control is lost, resulting in core damage. Transient-Caused Reactor or Turbine Trip A transient-caused trip is followed by a loss of PCS and AFW. Feed-and-bleed cooling fails either due to failure of the operator to initiate it, or due to hardware failures, resulting in core damage. Loss of Main Feedwater A feedwater line break drains the common water source for MFW and AFW. The operators fail to provide feedwater from other sources, . and fail to initiate feed-and-bleed cooling, resulting in core damage. A loss of main feedwater trips the plant, and AFW fails due to operator error and hardware

• The operators fail to initiate feed-and-bleed cooling, resulting in core damage. ,*. *steam Generator Tube Rupture A SGTR is followed by failure of AFW. Coolant is lost from the primary until the RWST is depleted .. HPI fails since recirculation*

cannot be established from the empty sump, and core damage results. 14 . I 4.2 Risk Important Component Failure Modes The generic component failure modes i dent i fi ed from PRA ana 1 yses as important to AFW system failure are listed below in decreasing order of risk importance.

1. Turbine-Driven Pump Failure to Start or Run. 2. Motor-Driven Pump Failure to Start or Run. 3. TOP or MOP Unavailable due to Test or Maintenance.

4. AFW System Valve Failures steam admission valves trip and throttle valve flow control valves pump discharge valves pump suction valves valves in testing or maintenance.

5. Supply/Suction Sources AFW storage tank stop valve hot well inventory suction valves. In addition to individual hardware, circuit, or instrument failures, each of these failure modes may result from common causes and human errors. Common cause failures of AFW pumps are particularly risk important.

Valve failures are somewhat less important due to the multiplicity of steam generators and connection paths. Human errors of greatest risk importance involve: failures to initiate or control system operation when required; failure to restore proper system lineup after maintenance or testing; and failure to switch to alternate sources when required.

15

5.0 FAILURE

MODES DETERMINED FROM OPERATING EXPERIENCE This section describes the primary root causes of component failures of the AFW system, as determined from a review of operating histories at Salem and at other PWRs throughout the nuclear industry.

Section 5.1 describes experience at Salem. Section 5.2 summarizes information compiled from a variety of NRC sources, including AEOD analyses and reports, information notices, inspection and enforcement bulletins, and generic letters, and from a variety of INPO reports as well. Some Licensee Event Reports (LERs) and NPRDS event descriptions were also reviewed individually.

Finally, information was included from reports of NRG-sponsored studies of the effects of plant aging, which include quantitative analyses of AFW system failure reports. This information was used to identify the various root causes expected for the broad PRA-based failure categories identified in Section 4.0, resulting in the inspection guidelines presented in Section 3.0. 5.1 Salem Nuclear Generating Station Experience The AFW system at Salem Nuclear Generating Station has experienced failures of the AFW pumps, turbine trip and throttle valve, turbine start and stop valve, pump discharge level control valves to steam generators, and pump suction and discharge valves. Failure modes include electrical, instrumentation, and hardware failures, and human errors. 5.1.1 AFW Pump Control Logic, Instrumentation and Electrical Failures There have been more than 20 failures of the AFW pumps to start and/or reach desired speed and discharge pressure at Salem since 1984. These events resulted from failures of steam admission valve control circuits, turbine governor control circuits, pressure transmitters, breaker failures, limit switch failures, and air regulator malfunctions.

The failure causes are mechanical

wear, or inadequate test and surveillance procedures.

Problems with the turbine governor control have resulted in failure of the turbine-driven pump to achieve rated speed or discharge pressure.

5.1.2 AFW Pump Failures Both motor-driven AFW pumps have experienced bearing failure and/or pump seizure since 1984. 5.1.3 Failure of AFW Pump Discharge Level Control Valves There have been more than 15 level control valve failures since 1984 at Salem. Failures have resulted from control circuit failures caused by air regulator failure, I/P transducer failure, control line air leaks, and valve and valve operator mechanical failures.

5.1.4 Other

AFW Valve Failures There have been more than 20 failures of AFW valves other than the level control valves since 1984. Valves stems have separated from valve operators, 16 l ' I valves and valve stems have bound, and valve operator bolts and nuts have vibrated lose. All these failure mechanisms have resulted in either total or partial loss of AFW valve operability.

5.1.5 Human

Errors Improper or inadequate calibration of AFW component has resulted in inaccurate flow, pressure, and speed indications.

Inadequate procedural guidance has resulted in both valve damage and a reactor trip. In one case improper use of procedures resulted in damage to flow transmitter bellows. Misunderstanding of system interfaces has resulted in exceeding Technical Specification limits. 5.2 Industry Wide Experience Human errors, design/engineering problems and errors, and component failures are the primary root causes of AFW System failures identified in a review of industry wide system operating history. Common cause failures, , which disable more than one train of this operationally redundant system, are highly risk significant, and can result from all of these causes. This section identifies important common cause failure modes, and then provides a broader discussion of the single failure effects of human errors, design/engineering problems and errors, and component failures.

Paragraphs presenting details of these failure modes are coded (e.g., CCI) and referenced by inspection items in Section 3. 5.2.1 Common Cause Failures The dominant cause of AFW system multiple-train failures has been human error. Design/engineering errors and component failures have been less frequent, but nevertheless significant, causes of multiple train failures.

CCI. Human error in the form of incorrect operator intervention into automatic AFW system functioning during transients resulted in the temporary loss of all safety-grade AFW pumps during events at Davis Besse (NUREG-1154, 1985) and Trojan (AEOD/T416, 1983). In the Davis Besse event, improper manual initiation of the steam and feedwater rupture control system (SFRCS) led to overspeed tripping of both turbine-driven AFW pumps, probably due to the introduction of condensate into the AFW turbines from the long, unheated steam supply lines. (The system had never been tested with the abnormal, connected steam supply lineup which resulted.)

In the Trojan event the operator incorrectly stopped both AFW pumps due to misinterpretation of MFW pump speed indication.

The diesel driven pump would not restart due to a protective feature requiring complete shutdown, and the turbine-driven pump tripped on overspeed, requiring local reset of the trip and throttle valve. In cases where manual intervention is required during the early stages of a transient, training should emphasize that actions should be performed methodically and deliberately to guard against such errors. CC2. Valve mispositioning has accounted for a significant fraction of the human errors failing multiple trains of AFW. This includes closure of 17 normally open suction valves or steam supply valves, and of isolation valves to sensors having control functions.

Incorrect handswitch positioning and inadequate temporary wiring changes have also prevented automatic starts of multiple pumps. Factors identified in studies of mispositioning errors include failure to add newly installed valves to valve checklists, weak administrative control of tagging, restoration, independent verification, and locked valve logging, and inadequate adherence to procedures.

Illegible or confusing local valve labeling, and insufficient training in the determination of valve position may cause or mask mispositioning, and surveillance which does not exercise complete system functioning may not reveal mispositionings.

CC3. At AN0-2, both AFW pumps lost suction due to steam binding when they were lined up to both the CST and the hot startup/blowdown demineralizer effluent (AEOD/C404, 1984). At Zion-I steam created by running the driven pump deadheaded for one minute caused trip of a motor-driven pump sharing the same inlet header, as well as damage to the turbine-driven pump (Region 3 Morning Report, 1/17/90).

Both events were caused by procedural inadequacies.

CC4. Design/engineering errors have accounted for a smaller, but significant fraction of common cause failures.

Problems with control circuit design modifications at Farley defeated AFW pump auto-start on loss of main feedwater.

At Zion-2, restart of both motor driven pumps was blocked by circuit failure to deenergize when the pumps had been tripped with an automatic start signal present (IN 82-01, 1982). In addition, AFW control circuit design reviews at Salem and Indian Point have identified designs where failures of a single component could have failed all or multiple pumps (IN 87-34, 1987). CCS. Incorrect setpoints and control circuit settings resulting from analysis errors and failures to update procedures have also prevented pump start and caused pumps to trip spuriously.

Errors of this type may remain undetected despite surveillance testing, unless surveillance tests model all types of system initiation and operating conditions.

A greater fraction of instrumentation and control circuit problems has been identified during actual system operation (as opposed to testing) than for other types of failures.

CC6. On two occasions at a foreign plant, failure of a balance-of-plant inverter caused failure of two AFW pumps. In addition to loss of the motor driven pump whose auxiliary start relay was powered by the invertor, the turbine driven pump tripped on overspeed because the governor valve opened, allowing full steam flow to the turbine. This illustrates the importance of assessing the effects of failures of balance of plant equipment which supports the operation of critical components.

The instrument air system is another example of such a system. CC7. Multiple AFW pump trips have occurred at Millstone-3, Cook-I, Trojan and Zion-2 (IN 87-53, 1987) caused by brief, low pressure oscillations of suction pressure during pump startup . These oscillations occurred despite the availability of adequate static NPSH. Corrective actions taken include: 18 extending the time delay associated with the low pressure trip, removing the trip, and replacing the trip with an alarm and operator action. CC8. Design errors discovered during AFW system reanalysis at the Robinson plant (IN 89-30, 1989} and at Millstone-I resulted in the supply header from the CST being too small to provide adequate NPSH to the pumps if more than one of the three pumps were operating at rated flow conditions.

This could lead to multiple pump failure due to cavitation.

Subsequent reviews at Robinson identified a loss of feedwater transient in which inadequate NPSH and flows less than design values had occurred, but which were not recognized at the time. Event analysis and equipment trending, as well as surveillance testing which duplicates service conditions as much as is practical, can help identify such design errors. CC9. Asiatic clams caused failure of two AFW flow control valves at Catawba-2 when low suction pressure caused by starting of a motor-driven pump caused suction source realignment to the Nuclear Service Water system. Pipes had not been routinely treated to inhibit clam growth, nor regularly monitored to detect their presence, and no strainers were installed.

The need for surveillance which exercises alternative system operational modes, as well as complete system functioning, is emphasized by this event. Spurious suction switchover has also occurred at Callaway and at McGuire, although no failures resulted.

CCIO. Common cause failures have also been caused by component failures (AEOD/C404, 1984}. At Surry-2, both the turbine driven pump and one motor driven pump were declared inoperable due to steam binding caused by backleakage of hot water through multiple check valves. At Robinson-2 both motor driven pumps were found to be hot, and both motor and steam driven pumps were found to be inoperable at different times. Backleakage at Robinson-2 passed through closed motor-operated isolation valves in addition to multiple check valves. At Farley, both motor and turbine driven pump casings were found hot, although the pumps were not declared inoperable.

In addition to multi-train failures, numerous incidents of single train failures have occurred, resulting in the designation of "Steam Binding of Auxiliary Feedwater Pumps" as Generic Issue 93. This generic issue was resolved by Generic Letter 88-03 (Miraglia, 1988}, which required licensees to monitor AFW piping temperatures--each shift, and to maintain procedures for recognizing_

steam binding and for restoring system operability.

CCII. Common cause failures have also failed motor operated During the, total loss of feedwater event at Davis Besse, the normally-open AFW isolation valves failed to open they were closed. The failure was due to improper setting_ of the torque switch-bypass switch, whic_h prevents motor trip on the high torqua required to unseat a closed valve.

• Previous problems with these valves had been addressed by increasing the torque switch trip setpoint -a fix which failed during the event due. to 'the higher torque required due to high differential pressure across the valve. Similar common mode failures MOVs have also occurred in other systems, resulting in issuance of Generit Letter 89-10, "Safety Related Motor-Operated Valve and Surveillance

{Partlow, 1989}." This generic letter requires licensees to develop and implement a program to provide for the testing, 19 inspection and maintenance of all safety-related MOVs to provide assurance that they will function when subjected to design basis conditions.

CC12. Other component failures have also resulted in AFW multi-train failures.

These include out-of-adjustment electrical flow controllers resulting in improper discharge valve operation, and a failure of oil cooler cooling water supply valves to open due to silt accumulation.

5.2.2 Human

Errors HEI. The overwhelmingly dominant cause of problems identified during a series of operational readiness evaluations of AFW systems was human performance.

The majority of these human performance problems resulted from incomplete and incorrect procedures, particularly with respect to valve lineup information.

A study of valve mispositioning events involving human error identified failures in administrative control of tagging and logging, procedural compliance and completion of steps, verification of support systems, and inadequate procedures as important.

Another study found that valve mispositioning events occurred most often during maintenance, calibration, or modification activities.

Insufficient training in determining valve position, and in administrative requirements for controlling valve positioning were important causes, as was oral task assignment without task completion

• feedback.

HE2. Turbine driven pump failures have been caused by human errors in calibrating or adjusting governor speed control, poor governor maintenance, incorrect adjustment of governor valve and overspeed trip linkages, and errors associated with the trip and throttle valve. TTV-associated errors include physically bumping it, failure to restore it to the correct position after testing, and failures to verify control room indication of TTV position following actuation.

HE3. Motor driven pumps have been failed by human errors in mispositioning handswitches, and by procedure deficiencies.

5.2.3 Design/Engineering Problems and Errors DEI. As noted above, the majority of AFW subsystem failures, and the greatest relative system degradation, has been found to result from turbine-driven pump failures.

Overspeed trips of Terry turbines controlled by Woodward governors have been a significant source of these failures (AEOD/C602, 1986). In many cases these overspeed trips have been caused by slow response of a Woodward Model EG governor on startup, at plants where full steam flow is allowed immediately.

This oversensitivity has been removed by installing a startup steam bypass valve which opens first, allowing a controlled turbine acceleration and buildup of oil pressure to control the governor valve when full steam flow is admitted.

DE2. Overspeed trips of Terry turbines have been caused by condensate.

in the steam supply lines. Condensate slows down the turbine, causing the governor valve to open farther, and overspeed results before the governor valve can respond, after the water slug This was to be the of 20 the loss-of-all-AFW event at Davis Besse (AEOD/602, 1986), with condensation enhanced due to the long length of the cross-connected steam lines. Repeated tests following a cold-start trip may be successful due to system heat up. DE3. Turbine trip and throttle valve (TTV) problems are a significant cause of turbine driven pump failures (IN 84-66). In some cases lack of TTV position indication in the control room prevented recognition of a tripped TTV. In other cases it was possible to reset either the overspeed trip or the TTV without reseting the other. This problem is compounded by the fact that the position of the overspeed trip linkage can be misleading, and the mechanism may lack labels indicating when it is in the tripped position (AEOD/C602, 1986). DE4. Startup of turbines with Woodward Model PG-PL governors within 30 minutes of shutdown has resulted in overspeed trips when the speed setting knob was not exercised locally to drain oil from the speed setting cylinder.

Speed control is based on startup with an empty cylinder.

Problems have involved turbine rotation due to both procedure violations and leaking steam. Terry has marketed two types of dump valves for automatically draining the oil after shutdown (AEOD/C602, 1986). At Calvert Cliffs, a 1987 loss-of-offsite-power event required a quick, cold startup that resulted in turbine trip due to PG-PL governor stability problems.

The short-term corrective action was installation of stiffer buffer springs (IN 88-09, 1988). Surveillance had always been preceded by turbine warmup, which illustrates the importance of testing which duplicates service conditions as much as is practical.

DES. Reduced viscosity of gear box oil heated by prior operation caused failure of a motor driven pump to start due to insufficient lube oil pressure.

Lowering the pressure switch setpoint solved the problem, which had not been detected during testing. DEG. Waterhammer at Palisades resulted in AFW line and hanger damage at both steam generators.

The AFW spargers are located at the normal steam generator level, and are frequently covered and uncovered during level fluctuations.

Waterhammers in top-feed-ring steam generators resulted in main feedline rupture at Maine Yankee and feedwater pipe cracking at Indian Point-2 (IN 84-32, 1984). DE7. Manually reversing the direction of motion of an operating valve has resulted in MOV failures where such loading was not considered in the design (AEOD/C603, 1986). Control circuit design may prevent this, requiring stroke completion before reversal.

DES. At each of the units of the South Texas Project, space heaters provided by the vendor for use in preinstallation storage of MOVs were found to be wired in parallel to the Class IE 125 V DC motors for several AFW valves (IR 50-489/89-11; 50-499/89-11, 1989). The valves had been environmentally qualified, but not with the non-safety-related heaters energized.

21

5.2.4 Component

Failures Generic Issue 11.E.6.l, "In Situ Testing Of Valves" was divided into four sub-issues (Beckjord, 1989), three of which relate directly to prevention of AFW system component failure. At the request of the NRC, in-situ testing of check valves was addressed by the nuclear industry, resulting in the EPRI report, "Application Guidelines for Check Valves in Nuclear Power Plants (Brooks, 1988)." This extensive report provides information on check valve applications, limitations, and inspection techniques.

In-situ testing of MOVs was addressed by Generic Letter 89-10, "Safety Related Motor-Operated Valve Testing and Surveillance" (Partlow, 1989) which requires licensees to develop and implement a program for testing, inspection and maintenance of all related MOVs. "Thermal Overload Protection for Electric Motors on Related Motor-Operated Valves -Generic Issue 11.E.6.l (Rothberg, 1988)" concludes that valve motors should be thermally protected, yet in a way which emphasizes system function over protection of the operator.

CFl. The common-cause steam binding effects of check valve leakage were identified in Section 5.2.1, entry CClO. Numerous single-train events provide additional insights into this problem. In some cases leakage of hot MFW past multiple check valves in series has occurred because adequate valve-seating pressure was limited to the valves closest to the steam generators (AEOD/C404, 1984). At Robinson, the pump shutdown procedure was changed to delay closing the MOVs until after the check valves were seated. At Farley, check valves were changed from swing type to lift type. Check valve rework has been done at a number of plants. Different valve designs and manufacturers are involved in this problem, and recurring leakage has been experienced, even after repair and replacement.

• CF2. At Robinson, heating of motor operated valves by check valve leakage has caused thermal binding and failure of AFW discharge valves to open on demand. At Davis Besse, high differential pressure across AFW injection valves resulting from check valve leakage has prevented MOV operation (AEOD/C603, 1986). CF3. Gross check valve leakage at McGuire and Robinson caused overpressurization of the AFW suction piping. At a foreign PWR it resulted in a severe waterhammer event. At Palo Verde-2 the MFW suction piping was overpressurized by check valve leakage from the AFW system (AEOD/C404, 1984). Gross check valve leakage through idle pumps represents a potential diversion of AFW pump flow. CF4. Roughly one third of AFW system failures have been due to valve operator with about equal failures for MOVs and AOVs. Almost half the MOV failures were due to motor or switch failures (Casada, 1989). An extensive study of MOV events (AEOD/C603, 1986) indicates continuing inoperability problems caused by: torque switch/limit switch settings, adjustments, or failuies; motor burnout;-*improper sizing or use of thermal overload devices; premature degradation related to inadequate use of protective devices; damage due to misuse (valve throttling, valve operator hammering);

mechanical

• problems (loosened parts, improper assembly);

or the torque switch bypass* circuit improperly installed or adjusted.

The study concluded that current 22

• l' methods and procedures at many plants are not adequate to assure that MOVs will operate when needed under credible accident conditions.

Specifically, a surveillance test which the valve passed might result in undetected valve inoperability due to component failure (motor burnout, operator parts failure, stem disc separation) or improper positioning of protective devices (thermal overload, torque switch, limit switch). Generic Letter 89-10 (Partlbw, 1989) has subsequently required licensees to implement a program ensuring that MOV switch settings are maintained so that the valves will operate under design basis conditions for the life of the plant. CFS. Component

• problems have caused a significant number of turbine driven pump trips (AEOD/C602, 1986). One group of events involved worn tappet nut faces, loose cable connections, loosened set screws, improperly latched TTVs, and improper assembly.

Another involved oil leaks due to component or seal failures, and oil contamination due to poor maintenance activities.

Governor oil may not be shared with turbine lubrication oil, resulting in the need for separate oil changes. Electrical component failures included transistor or resistor failures due to moisture intrusion, erroneous grounds and connections, diode failures, and a faulty circuit card. CF6. Electrohydraulic-operated discharge valves have performed very poorly, and three of the five units using them have removed them due to recurrent failures.

Failures included oil leaks, contaminated oil, and hydraulic pump failures.

CF7. Control circuit failures were the dominant source of motor driven AFW pump failures (Casada, 1989). This includes the controls used for automatic and manual starting of the as opposed to the instrumentation inputs. Most of the remaining problems were due to circuit breaker failures.

CFS. "Hydraulic lockup" of Limitorque SMB spring packs has prevented proper spring compression to actuate the MOV torque switch, due to grease trapped in the spring pack. During a surveillance at Trojan, failure of the torque switch to trip the TTV motor resulted in tripping of the thermal overload device, leaving the turbine driven pump inoperable for 40 days untilthe next surveillance (AEOD/E702, 1987). Problems result from grease changes to EXXON NEBULA EP-0 grease, one of only two greases considered environmentally qualified by Limitorque.

Due to lower viscosity, it slowly migrates from the gear case into the spring pack. Grease changeover at Vermont Yankee affected 40 of the older MOVs of which 32 were safety related. Grease relief kits are needed for MOV operators manufactured before 1975. At Limerick, additional grease relief was required for MOVs manufactured since 1975. MOV refurbishment programs may yield other changeovers to EP-0 grease. CF9. For AFW systems using air operated valves, almost half of the system degradation has resulted from failures of the valve controller circuit and its instrument inputs (Casada, 1989). Failures occurred predominantly at a few units using automatic electronic controllers for the flow control valves, with the majority of failures due to electrical hardware.

At Turkey Point-3, controller malfunction resulted from water in the Instrument Air system due to maintenance inoperability of the air dryers. 23 CFlO. For systems using diesel driven pumps, most of the failures were due to start control and governor speed control circuitry.

Half of these occurred on demand, as opposed to during testing (Casada, 1989). CFll. For systems using AOVs, operability requires the availability of Instrument Air, backup air, or backup nitrogen.

However, NRC Maintenance Team Inspections have identified inadequate testing of check valves isolating the safety-related portion of the IA system at several utilities (Letter, Roe to Richardson).

Generic Letter 88-14 (Miraglia, 1988), requires licensees to verify by test that air-operated safety-related components will perform as expected in accordance with all design-basis events, including a loss of normal IA. 24

6.0 REFERENCES

Beckjord, E. S. June 30, 1989. Closeout of Generic Issue II.E.6.1, "In Situ Testing of Valves". Letter to V. Stello, Jr., U.S. Nuclear Regulatory Commission, Washington, DC. Brooks, B. P. 1988. Application Guidelines for Check Valves in Nuclear Power Plants. NP-5479, Electric Power Research Institute, Palo Alto, CA. Casada, D. A. 1989. Auxiliary Feedwater System Aging Study. Volume 1. Operating Experience and Current Monitoring Practices.

NUREG/CR-5404.

U.S. Nuclear Regulatory Commission, Washington, DC. Gregg, R. E. and R. E. Wright. 1988. Appendix Review for Dominant Generic Contributors.

BLB-31-88.

Idaho National Engineering Laboratory, Idaho Falls, Idaho. Miraglia, F. J. February 17, 1988. Resolution of Generic Safety Issue 93, "Steam Binding of Auxiliary Feedwater Pumps" (Generic Letter 88-03). U.S. Nuclear Regulatory Commission, Washington, DC. Miraglia, F. J. August 8, 1988. Instrument Air Supply System Problems Affecting Safety-Related Equipment (Generic Letter 88-14). U.S. Nuclear Regulatory Commission, Washington, DC .. Partlow, J. G. June 28, 1989. Safety-Related Motor-Operated Valve Testing and Surveillance (Generic Letter 89-10). U.S. Nuclear Regulatory Commission, Washington, DC. Rothberg, 0. June 1988. Thermal Overload Protection for Electric Motors on Safety-Related Motor-Operated Valves -Generic Issue II.E.6.1.

NUREG-1296.

U.S. Nuclear Regulatory Commission, Washington, DC. Travis, R. and J. Taylor. 1989. Development of Guidance for Generic, Functionally Oriented PRA-Based Team Inspections for BWR Plants-Identification of Risk-Important Systems, Components and Human Actions. TLR-A-3874-T6A Brookhaven National Laboratory, Upton, New York. AEOD Reports AEOD/C404.

W. D. Lanning. July 1984. Steam Binding of Auxiliary Feedwater Pumps. U.S. Nuclear Regulatory Commission, Washington, DC. AEOD/C602.

C. Hsu. August 1986. Operational Experience Involving Turbine Overspeed Trips. U.S. Nuclear Regulatory Commission, Washington, DC. AEOD/C603.

E. J. Brown. December 1986. A Review of Motor-Operated Valve Performance.

U.S. Nuclear Regulatory Commission, Washington, DC. 25 AEOD/E702.

E. J. Brown. March 19, 1987. MOV Failure Due to Hydraulic Lockup From Excessive Grease in Spring Pack. U.S. Nuclear Regulatory Commission, Washington, DC. AEOD/T416.

January 22, 1983. at Trojan on January 22. 1983. Washington, DC. Information Notices Loss of ESF Auxiliary Feedwater Pump Capability U.S. Nuclear Regulatory Commission, IN 82-01. January 22, 1982. Auxiliary Feedwater Pump Lockout Resulting from Westinghouse W-2 Switch Circuit Modification.

U.S. Nuclear Regulatory Commission, Washington, DC. IN 84-32. E. L. Jordan. April 18, 1984. Auxiliary Feedwater Sparger and Pipe Hangar Damage. U.S. Nuclear Regulatory Commission, Washington, DC. IN 84-66. August 17, 1984. Auxiliary Feedwater Train. DC. Undetected Unavailability of the Turbine-Driven U.S. Nuclear Regulatory Commission, Washington, IN 87-34. C. E. Rossi. July 24, 1987. Single Failures in Auxiliary Feedwater Systems. U.S. Nuclear Regulatory Commission, Washington, DC. IN 87-53. C. E. Rossi. October 20, 1987. Auxiliary Feedwater Pump Trips Resulting from Low Suction Pressure.

U.S. Nuclear Regulatory Commission, Washington, DC. IN 88-09. C. E. Rossi. March 18, 1988. Reduced Reliability of Steam-Driven Auxiliary Feedwater Pumps Caused by Instability of Woodward PG-PL Type Governors.

U.S. Nuclear Regulatory Commission, Washington, DC. IN 89-30. R. A. Azua. August 16, 1989. Robinson Unit 2 Inadequate NPSH of Auxiliary Feedwater Pumps. Also, Event Notification 16375, August 22, 1989. U.S. Nuclear Regulatory Commission, Washington; DC. Inspection Report IR 50-489/89-11; 50-499/89-11.

May 26, 1989. South Texas Project Inspection Report. U.S. Nuclear Regulatory Commission, Washington, DC. NUREG Report NUREG-1154.

1985. Loss of Main and Auxiliary Feedwater Event at the Davis Bes.se Plant on June 9, 1985. U.S. Nuclear Regulatory Commission, Washington, DC. 26 11*

No. of Copies OFFSITE U.S. Nuclear Regulatory Commission B. K. Grimes OWFN 9 A2 F. Congel OWFN 10 E A. El Bass on i OWFN 10 A2 W. R. Butler OWFN 14 El9 10 S. M. Long OWFN 10 A2 K. Campe OWFN 1 A2 J. Chung OWFN 10 A2 2 K. S. West OWFN 12 HS U. S. Nuclear Regulatory Commission

-Region 1 M. W. Hodges C. W. Hehl P. Swetland DI STR IBUTI ON NUREG/CR-5761 PNL-7518 No. of Copies 4 Salem Resident Inspector Office OFFSITE J. H. Taylor Brookhaven National Laboratory Bldg. 130 Upton, NY 11973 R. Travis Brookhaven National Laboratory Bldg. 130 Upton, NY 11973 R. Gregg EG&G Idaho, Inc. P. 0. Box 1625 Idaho Falls, ID 83415 Dr. D. R. Edwards Professor of Nuclear Engineering University of Missouri -Rolla Rolla, MO 65401 ONSITE 27 Pacific Northwest Laboratory Distr.l S. R. Doctor L. R. Dodd B. F. Gore (IO) R. Pugh (5) N. E. Moffitt B. D. Shipp F. A. Simonen T. V. Vo Publishing Coordination Technical Report File (5)

NRC FORM J!l5 12*8'31 ' , NRCM 1102, 3201. 3202 2. TITLE AND SUBTITLE U.S. NUCLEAR REGULATORY COMMISSION BIBLIOGRAPHIC DATA SHEET (See insrructions on the rt1verse)

Auxiliary Feedwater System Risk-Based Inspection Guide for the Salem Nuclear Power Plant 5. AUTHOR (S) R. Pugh, B. F. Gore, T. V. Vo 1. REPORT NUMBER 3. (Aaigned b\I NRC. Add Vol., Supp ** R8'1., .,d AdOondum Num-1, ii ony.) NUREG/CR-5761 PNL-7518 DATE REPORT PUBLISHED MONTH I YEAR August 1991 4. FIN OR GRANT*NUMBER Ll310 6. TYPE OF REPORT Technical

7. PERIOD COVERED flnclusive Dare*/ 8/90 to 7/91 8. PERFORM I NG 0 RGAN IZA T ION -NAME AND ADDRESS llf NRC, provide Division, Offico or Rt19ion, IJ.S. Nucleor R"!lularory Comrniuion, and mailing llddrPss:

if conrri!JCror, proviae

• nd mailini Mltirea.J Pacific Northwest Laboratory Richland, WA 99352 9. SPONSOR ING ORGANIZATION

-NAME AND AODR ESS l/f NRC, rype Samtt .,.,,,,.., .. :if conrncror, provide NRC DwWon, Office or Rt19ion, IJ.S.

Rt19ularory Commi.,.ion, *nd mailing *ddreu.) Division of Radiation Protection and Emergency Preparedness Office of Nuclear Reactor Regulation

u. s. Nuclear Regulatory Commission Washington, DC 20555 10. SUPPLEMENTARY NOTES 11. ABSTRACT 12ooworo.or1ess1 In a study sponsored by the U.S. Nuclear Regulatory Commission (NRC), Pacific Northwest Laboratory has developed and applied a methodology for deriving plant-specific risk-based inspection guidance for the auxiliary feedwater (AFW) system at pressurized water reactors that have not undergone probabilistic risk assessment (PRA). This methodology uses existing PRA results and plant operating experience information.

Existing PRA-based inspection guidance information recently developed for the NRC for various plants was used to identify generic component failure modes. This information was then combined with plant-specific and industry-wide component information and failure data to identify failure modes and failure*mechanisms for the AFW system at the selected plants. Salem was selected as the fifth plant for study. The product of this effort is a prioritized listing of AFW failures which have*occurred at the plant and at other PWRs. This listing is intended for use by NRC inspectors in the preparation of inspection plans addressing AFW risk-important components at the Salem plant. 12. KEY WORDS/DESCR!PTORS ILi.r worrls orphrwt!S rhar will ...Ur,....,.,.,h,,,.

in Joe.ring rhe report.I Inspection, Risk, PRA, Salem, Auxiliary Feedwater (AFW) NRC FORM 335 (2-891 13. AVAILABILITY STATEMENT Unlimited

14. SECURITY CLASSIFICATION (This Pagel Unclassified Unclassified

15. NUMBER OF PAGES 16. PRICE THIS DOCUMENT WAS PRINTED USING RECYCLED PAPER

. 1 UNITED STATES REGULATORY COMMISSION WASHINGTON, D.C. 20555 OFFICIAL BUSINESS PENALTY FOR PRIVATE USE, $300 120555139531 1 .lANlRG US NRC-OADM . . c DIV FO!A &

TPS-PDR-NUREG WASHINGTON DC .__ --------SVCS 20555 SPECIAL FOURTH-CLASS RATE POSTAGE & FEES PAID USN RC PERMIT No. G-67