ML17312B334

Forwards Request for Addl Info Re Plant Pilot Proposal for risk-informed Inservice Testing

ML17312B334
Person / Time
Site:	Palo Verde
Issue date:	03/21/1997
From:	Clifford J NRC (Affiliation Not Assigned)
To:	James M. Levine ARIZONA PUBLIC SERVICE CO. (FORMERLY ARIZONA NUCLEAR
References
TAC-M94139, TAC-M94140, TAC-M94141, NUDOCS 9703260237
Download: ML17312B334 (67)

Text

.~~~ ~~00~

1,~

~o C,

nO cA

<c ~

+~

~O

+**++

UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 2055&4001 March 21,,

1997 Mr. James M. Levine Executive Vice President, Nuclear Arizona Public Service Company Post Office Box 53999

Phoenix, Arizona 85072-3999

SUBJECT:

REQUEST FOR ADDITIONAL INFORMATION REGARDING RISK-INFORMED INSERVICE TESTING FOR PALO VERDE NUCLEAR GENERATING STATION (TAC NOS.

M94139, M94140, AND M94141)

References:

l.

Arizona Public Service letter 102-03554-WLS/AKK/GAM, from W. L. Stewart to the NRC, dated November 27, 1995 2.

Arizona Public Service letter 102-03573-WLS/SAB/GAM, from W. L. Stewart to the NRC, dated December 20, 1995 3.

NRC letter from Charles R.

Thomas to W. L. Stewart,

APS, dated March 15, 1996 4.

Arizona Public Service letter 102-03714-JAB/AKK/GAM, from Jack A. Bailey to the NRC, dated June 7,

1996 5.

Arizona Public Service letter 102-03752-WLS/AKK/GAM, from W. L. Stewart to the NRC, dated August 7, 1996 6.

Arizona Public Service letter 102-03763-AKK/GAM, from A. K.

Krainik to the NRC, dated August 23, 1996 s

Arizona Public Service (APS) submitted a request to the NRC (references 1 and

2) to utilize a risk-informed inservice testing (RI-IST) program to determine inservice test frequencies for certain valves that were categorized as low safety significant.

The request was part of a pilot plant effort with TU Electric.

'The NRC staff provided an initial request for additional information (RAI) to APS related to the proposed RI-IST program via reference 3.

The NRC staff met with APS at the Palo Verde Nuclear Generating Station (PVNGS) site on April'3 and 24, 1996, to discuss the RAI.

APS provided a

partial response to the NRC staff's initial RAI via reference 4.

In Reference 5,

APS committed to provide a revised schedule for fully responding to the staff's initial RAI to the NRC staff by September 15, 1996.

APS submitted som'e additional information to the NRC in support of their proposed RI-IST program via reference 6.

The NRC Staff used the information provided by both pilot plant licensees to help develop a dr aft RI-IST Regulatory Guide (DG-1062) and Standard Review Plan section (SRP Section 3.9.7).

Enclosed is an additional RAI aimed at determining the extent to which the RI-IST program proposed by APS is consistent with the guidance being considered by the staff in the draft RI-IST HIIIC~ CEIItI~CIA' 6"""970qeg0237 970Sas PDR ADQCK 05000528 P

PDR

Hr. James H. Levine March 21, 1997 Regulatory Guide and Standard Review Plan section (which will soon be made available for public comment).

Several of the questions in the RAI were based on information provided to the NRC by Oak Ridge National Laboratory (ORNL) in letter reports.

These letter reports (Attachments to the Enclosure) document ORNL's review of Nuclear Plant Reliability Data System (NPRDS) failure records associated with Palo Verde.

Because of ongoing work on the draft risk-informed Regulatory Guides and Standard Review Plan sections, the staff may need to ask the pilot plant licensees questions in addition to those contained in the attachment to this letter.

These additional questions may relate to the policy issues'discussed in the January 22, 1997, Staff Requirements Memorandum.

It is anticipated that the final RAI will be sent to the RI-IST pilot plant licensees shortly after the draft RI-IST regulatory guide (RG) and standard review plan (SRP) are sent out for public comment. 'hile'we regret that a comprehensive set of RAIs cannot be provided to the pilot plant licensees at this time, we are confident that significant progress will continue to be made towards implementing RI-IST programs at Palo Verde'.

Sincerely, Original Sigped By Charles'Thomas for James W. Clifford, Senior Project Manager Project Directorate IV-2 Division of Reactor Projects III/IV Office of Nuclear Reactor Regulation Docket Nos.

STN 50-528, STN 50-529 STN 50-530

Enclosure:

Request for Additional Information cc w/encl:

See next page DISTRIBUTION:

Docket File PUBLIC JRoe EAdensam WBateman JClifford EPeyton CThomas ACRS PDIV-2 Reading

OGC, 015B18

KPerkins, RIV/WCFO AKowell, RIV

DKirsch, RIV/WCFO RWessman DOCUMENT NAME:

PV94139.RAI OFC NAME DATE PDIV-2/PH CThomas:

3/

/97 PDIV-2/PH JCliffor DIV-2 LA EPeyt n

3al 97 3/2 /97 OFFICIAL RECORD COPY

'I C

~

C 1

Hr. James M. Levine March 2i, 1997 Regulatory Guide and Standard Review Plan section (which will soon be made available 'for public comment).

Several of the questions in the RAI were based on information provided to the NRC by Oak Ridge National Laboratory (ORNL) in letter reports.

These letter reports (Attachments to the Enclosure) document ORNL's review of Nuclear Plant Reliability Data System (NPRDS) failure records associated with Palo Verde.

Because of ongoing work on the draft risk-informed Regulatory Guides and Standard Review Plan sections, the staff may need to ask the pilot plant licensees questions in addition to those contained in the attachment to this letter.

These additional questions may relate to the policy issues discussed in the January 22, 1997, Staff Requirements Memorandum.

It is anticipated that the final RAI will be sent to the RI-IST pilot plant licensees shortly after the draft RI-IST regulatory guide (RG) and standard review plan (SRP) are sent out for public comment.

While we regret that a comprehensive set of RAIs cannot be provided to the pilot plant licensees at this time, we are confident that significant progress will continue to be made towards implementing RI-IST programs at Palo Verde.

Sincerely, Docket Nos.

STN 50-528, STN 50-529 and STN 50-530 J

e W. Cli ford, Senior Project Manager P oject Directorate IV-2 Division of Reactor Projects III/IV Office of Nuclear Reactor Regulation

Enclosure:

Request for Additional Information cc w/encl:

See next page

t3 I

C

Mr. James M. Levine March 2i, 1997 cc w/encl:

Mr. Steve Olea Arizona Corporation Commission 1200 W. Washington Street

Phoenix, Arizona 85007 Douglas Kent Porter Senior Counsel Southern California Edison 'Company Law Department, Generation Resources P.O.

Box 800

Rosemead, California 91770 Senior Resident Inspector USNRC P. 0.

Box 40 Buckeye, Arizona 85326 Regional Administrator, Region IV U. S. Nuclear Regulatory Commission Harris Tower

& Pavillion 611 Ryan Plaza Drive, Suite 400 Arlington, Texas 76011-8064

Chairman, Board of Supervisors ATTN:

Chairman 301 W. Jefferson, 10th Floor

Phoenix, Arizona 85003 Mr. Aubrey V. Godwin, Director Arizona Radiation Regulatory Agency 4814 South 40 Street

Phoenix, Arizona 85040 Ms. Angela K. Krainik, Manager Nuclear Licensing Arizona Public Service Company P.O.

Box 52034

Phoenix, Arizona 85072-2034 Mr. John C. Horne, Vice President Power Supply Palo Verde Services 2025 N. Third Street, Suite 220

Phoenix, Arizona 85004 Mr. Robert Burt Los Angeles Department of Water

& Power Southern California Public Power Authority 111 North Hope Street, Room 1255-B Los Angeles, California 90051 Mr. David Summers Public Service Company of New Mexico 414 Silver SW, ¹0604 Albuquerque, New Mexico 87102 Mr. Bob Bledsoe Southern California Edison Company 14300 Mesa 'Road, Drop D41-SONGS San Clemente, California 92672 Mr. Robert Henry Salt River Project 6504 East Thomas Road Scottsdale, Arizona 8525l Terry Bassham, Esq.

General Counsel El 'Paso Electric Company 123 W. Mills El Paso, Texas 79901

r I

"I e

,I

COMNENTS ND E UEST OR ADDITIONAL INFORNATION PALO VERDE PILOT PROPOSAL FOR RISK-INFORNED INSERVICE TESTING DOCKET NOS.

S 50-528 STN 50-529 STN 50-530 The following are the supplemental questions and comments that have been developed by NRC staff reviewers who have been evaluating the proposed risk-informed inservice testing (RI-IST) program for Palo Verde.

These questions and comments are comprised of two parts:

(1) additional questions that remain from our review of the responses received for the first set of RAIs, and (2) relatively new issues that have been identified during the ongoing development of the NRC's guidance documents on risk-informed regulations.

We recognize that for some of these questions, the licensee's response to the first RAIs provided part of the answer that is being sought, and we encourage the licensee to refer to the previous RAI responses where appropriate.

What components are the major contributors to the change in core damage frequency (BCDF) associated with the proposed RI-IST program at Palo Verde?

What testing or other measures (including possibly taking credit for activities that would tend to reduce risk) can the licensee take to reduce the negative impact on ACDF?

Can the licensee make a more realistic estimate of the aggregate effect on CDF of the proposed RI-IST program (i.e.,

as opposed to a "conservative" estimate where component failure rate (A) is linearly extrapolated)?

This reassessment should include an identification of potential areas in which there was an overly conservative treatment in the quantification of both the baseline PRA risk levels and the change in risk associated with the proposed RI-IST program.

gualitative information should be provided for those areas that cannot be quantified.

2.

3.

Does the licensee's PRA assume that the current Code-required testing is 100 percent effective in assessing a component's operational readiness?

Does the current Code-required testing provide adequate information relative to the failure modes modeled in the licensee's PRA (e.g.,

failure of a valve to remain open for a 24-hour period)?

What consideration has been given to test effectiveness in establishing the proposed risk-informed IST program?

Are any PRA model or test strategy adjustments warranted?

On a component-specific

basis, the licensee should identify each instance where the proposed IST program change will affect the licensing basis of the plant (e.g.,

commitments made in response to NRC generic letters such as GL 89-10, THI action plan items, components relied on by the staff in concluding that the system and plant designs were acceptable).

These commitments, which may be incorporated into plant-procedures, may not be modeled in the licensee's PRA.

The licensee'hould identify the source and nature of the. commitment (or requirement),

and document the basis for the acceptability'f the proposed change.

.The licensee should consider the original acceptance

I

conditions, criteria, and limits as well as the risk significance of the component.

Consideration should also be given to diversity, redundancy, defense in-depth, and other aspects of the General. Design Criteria. If the licensing basis is not affected by the proposed IST program changes, the licensee should so indicate in its risk-informed IST Program description.

Provide any component-specific exemption requests, technical specification amendme'nt

requests, and relief requests necessary to implement the proposed RI-IST program.

Has the licensee submitted revised relief requests for high safety significant components (HSSCs) that were the subject of previously approved relief requests?

These relief requests should be reevaluated in light of the components risk significance.

Has the licensee submitted relief requests for high and low safety significant components not tested in accordance with the Code test method requirements or methods described in an NRC endorsed Code Case?

Has the licensee submitted relief requests for HSSCs that will not be tested in accordance with the Code test frequency requirements7 Please provide a detailed risk-informed IST Program implementation plan

.(i.e., for both HSSCs and LSSCs).

This implementation plan should contain details on how each component, or group of components, categorized as being LSSC, will have its test interval extended.

For

example, the staff needs to see a detailed description, or draft procedure, documenting how component test intervals will be extended in a step-wise manner (i.e., not just the "speed limit" test interval).

The implementation plan should describe how various component groupings were selected (e.g.,

using the guidance contained in NRC Generic Letter 89-04, Position 2 for check valves; Supplement 6 to NRC Generic Letter 89-10 and Section 3.5 of ASME Code Case OMN-I for motor-operated valves).

The implementation plan should document how the licensee proposes to use past performance, service condition, etc., in establishing the test strategy for specific components (see question 10 below). If the licensee wants to take credit for other operations and maintenance activities to justify less frequent inservice testing, then the details of these other activities and how they relate to the IST strategy needs to be described explicitly.

Provide a detailed description of how the licensee's integrated decision making process addressed each of the following issues:

shutdown and low power modes of operation seismic risk fires flooding other external events It is not sufficient to say that a "deliberative [undefinedj" process was used to account for these PRA scope issues.

If a well-defined process was not used, offer suggestions on how each issue might be addressed to produce well-defined, systematic, and scrutable results.

Please identify (if any) human actions that were used to compensate for a basic event probability increasing as a result of a test interval extension, specify the human failure probability used and describe how the licensee will ensure performance at this functional level.

How specifically will each of the following factors be considered by the licensee's integrated decision making process to establish an appropriate test strategy (i.e., test frequency and test method) for components:

past performance history, service condition,

design, and safety significance?

Either describe in detail the process that was used by APS to factor these variables into the test strategy determination or propose a

process.

The staff recognizes that, to some extent, these factors are embedded in the models and data supporting the licensee's PRA.

However, the staff expects licensee's to augment its PRA with a component-specific evaluation of performance, conditions, and design to arrive at an appropriate test strategy (including test interval).

A November ll, 1996 (Attachment 2), letter report from A. B. Poole (ORNL) to J.

E. Jackson (NRC) indicates that ORNL did a brief review of the available NPRDS failure records and performance data for approximately 228 "low risk significant" check valves (i.e., data from 1986 to 1995) at Palo Verde.

Of the 106 NPRDS failure records on check valves at Palo Verde during this period, 55 were associated with check valves categorized as LSSC.

Seventeen percent of the LSSC check valves listed in the licensee's RI-IST program submittal have experienced repeated failures.

Some valves had as many as seven repeat failures.

(all three units considered).

At least 16 of the check valves had failed or degraded internals caused at least in part by some age-related failure mechanism such as "wear " or "cyclic fatigue."

How were these types 'of failure causes considered when evaluating whether, and the extent to which, the testing interval could be extended't least 75 percent of the check valves categorized as LSSC come from either AFW, diesel starting air, containment isolation, CCW, main steam, or RHR systems, "which in previous ORNL studies have been shown to have some of the highest relative failure rates by system for significant failures (in terms of component degradation)."

The number of repeat failures and the type of failures listed in NPRDS seems to indicate that age-related failure mechanisms are present in the CVCS, diesel starting air, main steam, and RHR systems.

Unmitigated component aging can significantly increase component unavailability and the risk of undetected failure due to decreased testing.

Unavailabilities of all check valves in applications susceptible to aging should be simultaneously increased by the appropriate factor to cover the

simultaneous effects of aging.

This should be completed to show that the impact on risk remains low even for unmitigated aging.

10.

A November 6, 1996 (Attachment 1), letter report from A. B. Poole (ORNL) to J.

E. Jackson (NRC) indicates that ORNL reviewed NPRDS failure records for low risk significant motor-operated valves (MOVs) (i.e.,

data from 1987 to 1995) at Palo Verde.

This letter report states that "ample evidence exists to question the technical validity of extending the inspection interval for the requested valves" particularly those in certain systems (e.g., auxiliary feedwater and safety injection systems).

While a failure rate for MOVs derived from NPRDS data since 1987 may be overly conservative (i.e.,

because it does not adequately reflect improvements to MOVs made as a result of GL 89-10), it may also be non-conservative (i.e.,

because MOV testing has not typically evaluated MOV performance under dynamic conditions).

Describe the basis for the selection of failure rates used in licensee's PRA.

How were these failure rates adjusted based on plant-specific experience and operating environment?

11.

The licensee should describe in detail its performance monitoring plan and explain how sufficient data will be developed to facilitate PRA and risk-informed IST Program updates.

Will there be sufficient monitoring of both HSSC and LSSC to support the periodic updates?

As noted in RAI tl, have the components that contribute most to risk increase been identified and a monitoring program specifically planned that could be used to modify assumed failure rate data that is currently either under or overly conservative?

Does the proposed performance monitoring process ensure:

~

enough tests are included, over gradually extending time periods, to provide meaningful data;

~

incipient degradation is likely to be detected and corrective action taken; and

~

appropriate parameters, as required by the ASME Code or ASME Code

case, are trended as necessary to provide validation of the PRA?

Does the proposed performance monitoring process ensure that degradation is not significant for components that are placed on an extended test

interval, and that failure rate assumptions for these components are not compromised by test data?

12.

Does the licensee's corrective action program:

a.

Comply with 10 CFR 50, Appendix B, Criterion XVI, Corrective Action?

0

~

1 l

b.

Evaluate IST components that fail to meet the test acceptance criteria as well as IST components that are otherwise determined to be in a nonconforming condition.

c.

For each component failure:

(i) comply with 10 CFR 50, Appendix B, Criterion XVI, Corrective Action d.

(ii) determine the impact of the failure or nonconforming condition on system/train operability since the previous

test, (iii) determine and correct the root cause of the failure or nonconforming condition (e.g.,

improve testing practices, repair or replace the component),

(iv) assess the applicability of the failure or nonconforming condition to other components in the IST program (including any test sample expansion that may be required for grouped components such as relief valves),

(v) correct other susceptible similar IST components as necessary, (vi) assess the validity of the PRA failure rate 'and unavailability assumptions in light of the failure(s),

and (vii) consider the effectiveness of the component's test strategy in detecting the failure or nonconformirig condition.

Adjust the test frequency and/or methods, as appropriate, where the component (or group of components) experiences repeated failures or nonconforming conditions.

Provide the licensee's PRA group with the corrective action evaluations so that any necessary model changes and re-grouping are done as might be appropriate.

Is any credit taken for the corrective action program in the PRAT If not, do you think that it is feasible and justified to do'sot 13.

Are there any RI-IST program changes that the licensee proposes to make without prior NRC approval other than changes explicitly described by the licensee in RI-IST program submittals and approved by the staff (e.g.,

component categorization/re-categorization in accordance with an NRC approved methodology, gradual extension of a component's test interval in a step-wise fashion as approved by the staff in its safety evaluation)?

Ooes the licensee have an adequate process or procedures in place to ensure that RI-IST program changes of the following two types get reviewed and approved by the NRC prior to implementation:

~

]

J'

(1) test method changes that involve deviation from the NRC-endorsed Code requirements.

(2) changes to the risk-informed IST program that involve process changes (e.g.,

changes to the plant probabilistic model assumptions, changes to the grouping criteria or figures of merit used to group components, changes in the Acceptance Guidelines used by the licensee's integrated decision-making process

[e.g.,

expert panel]).

14.

Does the licensee's RI-IST program test components in the HSSC category that are not in the licensee's current IST program commensurate with their safety significance?

These components should be tested in accordance with the ASNE Code where practical, including compliance with all administrative requirements.

Where ASNE Section XI or 0&N testing is not practical, has the licensee proposed alternative test methods to ensure operational readiness and to detect component degradation (i.e.,

degradation associated with failure modes identified as being important in the licensee's PRA)?

15.

Are IST components in the RI-IST program (with the exception of check valves) exercised or operated at least once every refueling cycle?

Are components in the following categories exercised more frequently than once per operating cycle, if practical:

(a) components with high risk significanqe, (b) components in adverse or harsh environmental conditions, or (c) components with any abnormal characteristics (operational,

design, or maintenance conditions)?

16.

How does the licensee plan to address, or deal with, the synergistic effects of implementing its risk-informed IST program and other risk-informed initiatives?

How does the licensee plan to maintain the level of comm'itment of plant resources (e.g.,

gA or maintenance) that was assumed in justifying extended IST intervals?

17.

Does the licensee have procedures for conducting the periodic risk-informed IST program review to ensure that it

~

prompts the licensee to conduct overall program assessments periodically (i.e., at least once every two refueling outages) to

'reflect changes in plant configuration, component performance, test results, industry experience, and to reevaluate the effectiveness of the IST program,

jl II ll~

'li I

prompts the licensee to compare actual component performance to predicted levels to determine if component performance and conditions are acceptable (i.e.,

as compared to predicted levels).

If performance or conditions are not acceptable then the cause(s) should be determined and corrective action implemented, prompts the licensee to review and revise as necessary the assumptions, reliability data, and failure rates used to group components to determine if component groupings have changed, and prompts the licensee to reevaluate equipment performance (based on both plant-specific and generic information) and.test effectiveness to determine if the inservice test program should be adjusted (Plant-specific data should be incorporated into the generic data using appropriate updating techniques)7 Does the licensee have procedures to ensure that the results of its corrective action program for IST program components get fed back into its periodic IST program reassessment?

Does the licensee have procedures in place to identify the need for more emergent RI-IST program updates (e.g., following a major plant modification, or significant equipment performance problem).

18.

To avoid being overly prescriptive in its guidance, yet still ensure that certain topics having major.safety importance for all risk-informed programs are addressed in licensee's propoSals, the staff has identified a set of five key safety principles in the draft risk-informed guidance documents.

It is currently intended that the 5 key principles given

~

below must be explicitly addressed in all licensee applications for risk-informed programs.

The regulatory guides that are under development are to provide an example of acceptable means for satisfying these key principles.

Because that guidance has not been finalized, it would be useful to have the pilot plant licensees describe how their proposed RI-IST program satisfies each of the following key safety principles:

(a)

The proposed change meets the current regulations.

[This principle applies unless the proposed change is explicitly related to a requested exemption or rule change.]

(b)

The defense in depth philosophy is maintained.

(c)

Sufficient safety margins are maintained.

(d)

Proposed increases in risk, and their cumulative effect, are small and do not cause the NRC Safety Goals to be exceeded.

(e)

Performance-based implementation and monitoring strategies are proposed that address uncertainties in analysis models and data and provide for timely feedback and corrective action.

1 'I l

II) f, 1

In addressing these principles, the licensee should describe how:

All safety impacts of the proposed changes were evaluated on a

component-specific basis as well as in an integrated manner as part of an'verall risk management approach in which the licensee uses risk analysis to improve operational and engineering decisions broadly and not just to eliminate requirements that the licensee sees as undesirable.

The approach used to identify changes in requirements should be used to identify areas where requirements should be increased as well as where they could be reduced.

~

The acceptability of proposed changes should be evaluated by the licenseein an integrated fashion that ensures that all principles are met."

Core damage frequency (CDF) and large early release frequency (LERF) can be used as suitable metrics for making risk-informed regulatory decisions.

Increases in estimated CDF and LERF resulting from proposed CLB changes will be limited to small increments.

The scope and quality of the engineering analyses (including traditional and probabilistic analyses) conducted to justify the proposed CLB change should be appropriate for the nature and scope of the changes proposed and should be based on the as-built and as-operated and maintained plant.

~

Appropriate consideration of uncertainty is given in analyses and interpretation of findings.

~

The plant-specific PRA supporting decisions has been subjected to quality controls such as an independent peer review.

Data, methods, and assessment criteria used to support the proposed IST program changes (e.g.,

those used by the licensee's expert panel) must be available for public review.

19.

Please summarize any reviews (e.g.,

peer review, industry-wide

- comparison, etc.) that was performed on the PRA used to support the licensee's proposed RI-IST program.

Attachments:

l.

ORNL Letter Report dtd. Il/6/96 2.

ORNL Letter Report dtd. ll/ll/96

'ne important element of integrated decision making can be the use of an "expert panel."

Such a panel is not a necessary component of risk-informed decision making; but when it is used, the key principles and associated decision criteria still apply and must be shown to have been met or to be irrelevant to the issue at hand.

OAK RIDGE NATIONALLABORATORY

~

MANAGEDBY LOCKHEED MARTINENERGY RESEARCH CORPORAT)ON FOR THE U.S. DEPARTMENTOF ENERGY POST OFFlCE BOX2009 OAKROG5, TN 3783HIM Attachment 1

PHONE; (423) 5744734 FAX: (423) 5754493 INTERNET: AOPomLgoy November 6, 1996 Jerry E. Jackson U. S. Nuclear Regulatory Commission MS Tlo E10 Washington, D.C. 20555

Dear Mr. Jackson:

As was discussed in the October 8, 1996 meeting with NRC personnel, Palo Verde motor-operated valves (MOVs) listed in Appendix D to 13-NS-CO5, Rev. 0 were reviewed relative to failures. The time period covered was &om 1987 through 1995.

NPRDS was searched for failures ofthe above components at Palo Verde. Atotal of28 MOVs from the Appendix D population were identified with failures during the time period. This represents about 4% ofthe total MOVpopulation at all ofthe Palo Verde units and about 17% ofall the MOV failures.

These'components and the tabulation of all failures by year are shown in the attached summary report. Ifpacking leaks and external leakage are excluded the results are provided in Table 1. Ifthe 21 valves listed in Table 1 are evaluated for mean time between failures (MTBF),

then this value is determined to be 3.81 years.

This value ofMTBF is considerably smaller than the 6 year test interval being requested.

The distribution ofleakage related and degraded normal operation failures as identified in NPRDS was 40 due to leakage and 51 due to degraded operation.

Table 2 identifies the distribution of failures between the actuator, electrical supply, and valve. This shows that one failure was due to the electrical supply and that the other failures were distributed with 45 in the actuator and 9 in the valve.

Table 3 provides a listing ofthe NPRDS symptom offailure for each year studied. Table 4 provides a listing ofthe NPRDS cause offailure for each year studied. Table 5 provides a listing ofsymptom offailure relative to actual cause offailure. The NPRDS evaluation shows that 51% ofthe failures were detected during some form oftesting.

Wringing c5cience lo knife

Jerry E. Jackson November 6, 1996 Page 2 Ifthe 28 valves are reviewed for system ofservice and type offailure the following evaluation is found:

AFW System Degraded Operation Leakage 18 2

Charging System Containment Purge Essential Cooling Water Reactor Vent Bc Drains 10 1

0 5

Rad. Waste Drain Steam Generator Bypass Safety Injection Total:

26 55 24 36 This evaluation shows that AFW accounts for 22% ofthe MOV failures in this population and the Safety Injection System accounts for 55% ofthe MOV failures.

Although this examination was rather cursory in nature, ample evidence exists to question the technical validityofextending the inspection interval for the requested valves. These failure rates would suggest an over all total MOVfailure rate at Palo Verde ofapproximately 1

>< 10'ailures per hour.

Jerry E. Jackson November 6, 1996 Page 3 We hope that this information willbe useful to you and should you need additional information we would be glad to provide further assistance.

More failure data information on MOVs will be provided later by special Letter Report.

Sincerely, g,B,P A. B. Poole ABP:jkc Attachments cdenc:

P. L. Campbell, NRC J. Colaccino, NRC D. F. Cox D. C. Fischer, NRC W. C. Gleaves, NRC T. G. Scarbrough, NRC W. E. Vesely, SAIC J. P. Vora, NRC G. H. Weidenhamer, NRC R, H. Wessman, NRC

I

An Examination ofMotor-Operated Valve Failures With Application to Increasing the Surveillance Testing Period at Arizona Public Service Company Palo Verde, Units 1, 2, and 3

'repared by D. F. Cow Oak Ridge National Laboratory Prepared for U.S. Nuclear Regulatory Commission

~

~

In a response to a request by Arizona Public Service, Palo Verde Nuclear Generating Stations 1, 2 and 3 to extend the inspection interval for low safety-significant motor-operated valves, data were collected from the Institute ofNuclear Power Operations (INPO) Nuclear Plant Reliability Data System (NPRDS).

The data were collected using the following search methodology and search parameters.

~

Selected Components are VALVEand VALVOP

~

Selected VALVETypes are Butterfly, Gate, and Globe

~

Selected VALVEOperator is Electric Motor/Servo

~

Selected VALVOPTypes are Electric Motor-AC and Electric Motor-DC Selected VALVOP Subtype is Geared

~

Selected Unit IDs are PALO VERDE 1, PALO VERDE 2, and PALO VERDE 3 This search yielded a total of 523 failures noted as motor-operated valve failures. It is important to note here that the term failure refers to component degradation that affects valve or actuator function. In common terms, ifthe motor-operated valve does not function as designed then that degradation is considered a failure.

The 523 failures are distributed among 159 MOVs and 11 years, from 1986 to 1996. The distribution of failures by vear ofoccurrence is shown in Figure I. This list ofvalves was compared to the list ofvalves for which an extension of the inspection period is requested.

This resulted in a total of 91 failures among 28 valves. The failure narratives for these failures were analyzed and data grouped by the following calegones.

Component The area offailure was limited to actuator, valve, or electrical. Actuator failures involve failures ofthe MOV that include the housing, motor, switches, etc. Basically this includes anything between the valve and actuator mounting flange to the, conduit containing power and control cables. Valve failures involve the valve body, bonnet, stem, and trim.

Electrical failures include componeilts in the motor control center, including breakers and thermal overloads Problem Three categories were used here to further segregate failures. The event was either leakage related, a problem that did not cause a loss ofoperability/functionality, or a problem that did cause a loss ofoperability/functionality.

Symptom This category lists the unusual circumstance that alerted utilitypersonnel that a degraded condition existed.

Cause What was the actual cause ofthe observed symptom.

Detection What activity led to the discovery ofthe failure. The method ofdetection was listed as either a failure on demand, discovered during maintenance, observation (usually limited to leaks), testing, or walkdowns.

'I After categorizing, packing leaks and external leakage records were removed from further analysis since they were not considered critical to MOVoperation. This resulted in a total of 55 failures among 21 valves from 1987 to 1995, with no failures noted for 1994. For the readers convenience these failures are listed below by component identifier and year offailure in Table 1.

r

Table 1 - Tabulation of failures by component identifier Total by Component ID 1987 1988 1989; 1990 1991 1992 1993 1995 comP>>ent Id AFAHV0032 AFBHV0030 AFBHV0031 AFBUV0035 AFCHV0033 AFCUV0036 CHAHV0531 CPAUV0002A EWAUV0065 EWAUV0145 GRAUV0001 RDAUV0023 SGEUV0169 SIAHV0657 SIAHV0698 SI AU V0634 SIAUV0644 S IBHV0658 S IBHV0699 SIBUV0665 S IBUV0667 Total by year 0

0 1

0 0

0 4

0 0

0 1

0 0

0 0

0 1

0 1

0 1

0 1

0 0

2 0

0 0

0 0

0 1

0 0

0 1

1 0

1 0

0 0

0 0

0 0

0 0

0 1

0 1

0 0

5 0

0 0

0 0

0 0

1 0

0 0

0 0

0 1

0 0

0 1

0 0

1 5

0 0

0 3

0 5

4 19 7

1 1

0 0

0 0

1 0

1 1

0 0

1 0

0 0

0 1

0 0

0 1

0 0

0 0

0 0

0 0

0 1

0 0

0 1

0 0

0 0

0 1

0 0

0 0

0 1

0 0

0 0

0 1

0 0

1 0

0 0

0 0

1 0

1 0

0 0

0 0

1 0

0 0

1 0

0 0

0 1

0 1

0 0

5 7

4 4

3 5

3 1

3 3

21' 3

1 1

1 7

1 2

1 2

2 7

4 55 The reader may wish to observe that 14 ofthe 21 components identified have multiple failures, and only two ofthese have a time period ofsix years between failures. The distribution ofthese failures is shown in Figure 1.

20 18 16 14 12 10 8

1987 1988 1989 1990

1991, 1992 1993 1995 Year of Failure Figure 1-Distribution of Failures by Year

l

vr r.

~

Examination ofthese failures identifies the actuator as the component area with the largest percentage of failures (82%), This is shown below in Table 2'. It should be noted that units 1 and 2 at Palo Verde started commercial operation in 1986, and unit 3 started commercial operation in 1988. Aging ofvalve components in contact with the fluid may not see aging related failures without additional service wear.

Table 2 - Failure tabulation by major component and year of failure Major component Total by 1987 1988 1989 1990 1991 1992 1993 1995 area Actuator

.'alve, Electrical Total by year 5

3 15 6

5 4

3 4

45 0

1 4

1 0

2 1

0 9

0 0

0.

0 0

1 0

0 1

5 4

19 7

5 7

4 4

55 As can be seen in Figure 2, ifdata for 1989 is not averaged then actuator failures are relatively constant in number.

16 14

~~~~

Actuator Failures Average Actuator Failures 10 8

LL 6

1987 Excluding 1989 1988 1989 1990 1991 Year of Failure 1992 1993 1995 Figure 2 - Distribution of actuator failures by year In examining the symptom that led to discovery ofthe failure it can be seen that failure to move to either the fully open or closed position or change positions with adequate thrust margin accounts for approximately 51% ofthe failures. These categories have been highlighted in Table 3 below. The term "Failure to Close" means that the actuator would not move in the closed direction. "Failure to Close Completely'eans that actuator moved the valve in the desired direction but was incapable of accomplishing a full stroke. The same, applies to failures in the open direction. Ifthe failure narrative did not specify the desired direction oftravel the term "Failure to Operate" was used.

A

1

~

Table 3 - Tabulation of failures by symptom and year of failure Symptom of failure 1987 1988 1989 1990 1991 1992 1993 1995 Total by symptom Breaker Trip Broken Ls Rotor Cycling Declutch Nol Disengaging Degraded Stroke Time Failure To Declutch Failure To Close Failure To Close Completely Failure To Open Failure To Open Completely Failure To Operate High Run Current Improper Ls Setup Inadequate Voltage Internal Leakage Locked Rotor None Over Thrusting Partial Rotor Rotation Power Imbalance Smoke Under Thrusting Total By Year Of Failure 3

1 0

1 0

1 0

4 0

1 2

0 0

1 0

0 0

1 1

1 1

19.,

0 0

0 0

1 0

0 2

1 0

0 0

0 0

2 0

~

0 1

0 0

0 0

7 0

0 0

0 1

1 0

0

'0 0

0 0

0 1

0 1

0 0

0 1

0 1

0 0

0

~

0 1

0 0

0 0

0 1

0 1

1 0

0 0

0 0

0 1

1 5

7 0

0 1

0 0

0 0

0 4

0 1

0 0

0 0

0 3

1 2

1 3

1 2

10 8

2 3

2 1

1 3

1 2

3 1

1 1

3 55 tfne ciaminc the cause of the failure that generated the symptoms noted above we see that setpoint shift accounted for 29'r> of the failures. Thc reader inay note that the limit switches for this particular ty~ ofactuator are adjusted n'ith gears. not a slid:ng slop Tlicrcforc a change in liiniiswitch setting requires,n'ear of the limitsivitch adjusting gears, which are located in a sealed gear boy packed niiligrcasc. Thc reason fnr this aniount of near is not explained in the failure narratives.

Table 4-Tabulation of failures by cause and year of failure Cause 1987 1988 1989 1990 1991 1992 1993 1995 Total by Cause Bent Stem Blown Fuse Degraded Packing Design Error Grease Migration Improper Assembly/Operation Loose Valve to Actuator Mount LS Setpoint Shift Motor Short Normal Wear/Aging Travel Stop Drift TS Roll Pin Shear TS Setpoinl Shift Under Thrusting Unknown Total by Year 0

0 0

0 2

0 0

3 0

0 0

0 0

0 0

5 2

0 2

0 1

1 0

1 2

2 0

0 3

0 5

19 0

0 0

0 0

0 1

4 0

1 0

1 0

0 0

7 0

0 0

2 0

0 0

0 0

0 0

0 1

0 2

5 0

1 0

1 0

1 0

0 0

0 1

0 1

1 1

7 0

0 0

0" 0

0 0

0 1

0 0

1 0

1 1

4 0

0 0

1 0

0 0

1 1

0 1

0 0

0 0

4 2

1 2

5 3

3 1

10 4

3 2

2 6

2 9

55

I e

t

~

~

The information in Table 5 allows the reader to examine the coded symptom offailure relative to the coded cause offailure.

Table 5 - Tabulation of Failures bySymptom and Cause of Failure CO Symptom of Failure ID Vl u

I I I C

o.

g cn P

th 0

O CO II C9 0

E e

E 0

mO Co C

K M

to Cb 2M O

g z z CD V)

IO CO O.

OK CO I

Co Io I-I- I-C C

Total by Symptom Breaker Trip Broken LS Rotor Flow Oscillations Declutch Not Disengaging Degraded Stroke Time Failure to Declutch Failure to Close Failure to Close Completely Failure to Open Failure to Open Completely Failure to Operate High Run Current Improper LS Setup Inadequate Voltage Internal Leakage Locked Rotor None Over Thrusting Partial Rotor Rotation Power Imbalance Smoke Under Thrusting Total b Cause of Failure 1

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 1

0 0

0 0

1 0

0

,0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

2 1

0 0

0 0'

0 0

0 0

1 0

0 0

0 0

0 0

1 0

0 0

0 2

0 0

0 0

1 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

2 2

5 0

0 0

0 0

0 0

0 0

0 1

0 0

1 0

0 0

1 0

1 1

0 0

0 0

0 0

0 0

0 1

0 0

0 0

0 0

0 0

0 0

0 0

0 3

3 0

0 0

0 0

0 0

0 1

0 0

0 0

0 0

0 0

0 0

0 0

0 1

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

1 0

0 1

0 0

0 0

0 0

0 0

0 0

0

,7 0

0 0

0 0

2 0

0 0

0 0

0 1

0 0

0 0

0 0

0 0

0 0

0 0

0 0'0 0

0 0

0 0

0 2

0 1

0 0

0 0

0 0

0 0

0 0

0 1

0 0

0 0

1 0

0 1

0 0

0 1

0 0

0 0

1 0

0 0

0 0

1 0

0 10 4

3 2

2 2

0 0

0 0

0 0

0 0

0 0

0 0

1 2

1 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 2

0 0

0 0

0 0

0 0

0 6

2 0

3 1

1 2

2 1

1 0

3 0

1 0

2 0

10 2

8 0

2 1

3 0

2 1

1 0

1 0

3 0

1 1

2 0

3 0

1 0

1 0

1 0

3 9

55 Although the above examination was rather cursory in nature, ample evidence exists to question the technical validity ofextending the inspection interval for the requested valves. The degree ofwear displayed at these units does not support extending the inspection interval without further analysis ofthe failures, their causes, and actions implemented to prevent recurrence.

~

I

Attachment 2

OAKRIDGE NATIONALLABORATORY, MANAGEDBY LOCKHEED MARTINENERGY RESEARCH CORPORATION FOR THE L.S. DEPARTMENTOF ENERGY POST OFFCE BOX2009 OAKRIDGE. TH 37831463S November 11, 1996 PHONE: (423) 5744734 FAX: (423) 5764493 INTERNET:AOPeml.gov Jerry E. Jackson U. S. Nuclear Regulatory Commission MS T10 E10 Washington, D.C. 20555

Dear Mr. Jackson:

As we discussed in the October 8, 1996 meeting withNRC personnel, selected Palo Verde check valves listed in Appendix D to 13-NS-CO5, Rev. 0, have been reviewed relative to failure history.

The time period covered was &om 1986 through 1995.

Palo Verde has requested IST extension &om their current Code requirements (usually quarterly) to an interval of 6 years on approximately 228 "low risk significant" check valves (76 valve applications were listed in the submittal; it is assumed that all three units affected). In an effort to provide information needed to evaluate potential candidate risk based inservice test (RBIST) check valves at Palo Verde for extended IST intervals, Oak Ridge National Laboratory (ORNL) has done a brief review of the available NPRDS failure records and performance data for the valves in question. The results ofthis review are provided in the attached summary report.

The most significant findings of this study resulted &om a brief review of the 106 raw NPRDS failure records listed for check valves at Palo Verde Units 1, 2, 3 during the time period 1986-1995.

Ofthe 106 raw NPRDS failure records, 55 were associated with the candidate valves.

Seventeen percent of the valve applications listed in the relief request submittal have experienced repeat failures.

Some valves had as many as seven repeat failures (all three units considered).

It is important to note that at least 16 ofthe failed valves listed in Table 10 ofthe attachment had failed or degraded internals caused at least in part by some age-related failure mechanism such as "wear" or "cyclic fatigue." These types offailure causes need to be considered when evaluating whether to extend inservice testing intervals. Ofthe 55 failure records associated withthe deferral candidate valves, 11 involved external leakage, while a characterization ofthe remainder according to extent ofdegradation resulted in 29 (66%) moderate and 15 (34%) significant failures. These results are comparable to those found industry-wide during previous ORNL studies for check valve failures occurring during 1991 and 1992.

The system ofservice for candidate IST deferral should also be considered. At least 75 percent of the candidate "lowsafety significance" valves for deferral come &om either AFW, Diesel Starting Air,Containment Isolation, CCW, Main Steam, or RHR systems, which in previous ORNL studies have been shown to have some ofthe highest relative failure rates by system for significant failures (in terms ofcomponent degradation),

2fringiny cocci ence fo knife

/

t c'

~ ~

Jeny E. Jackson November 11, 1996 Page 2 The number ofrepeat failures and type offailures listed in NPRDS (see page 7, 8, and 9 ofthe attachment) certainly seems to indicate that age-related failure mechanisms are present in the followingsystems:

~

CVCS

~

Diesel Starting Air

~

Main Steam

~

RHR The recently provided Draft NUREG/CR-6508 "Component Unavailability Versus Inservice Test

'ST)

Interval: Evaluations ofComponent Aging Effects With Applications to Check Valves," has shown that unmitigated component aging can significantly increase the unavailability and risk due

.to decreased testing. The Palo Verde submittal has not addressed aging-related effects on the risk analysis completed.

Although this examination was rather cursory in nature, ample evidence exists to question the technical validity of extending 'the inspection interval for the requested check valves.

Unavailabilities ofall check valves in applications susceptible to aging should be simultaneously increased by the appropriate factor to cover the simultaneous effects of aging.

This should be completed to show that the impact on risk remains low even for unmitigated aging.

We hope that this information willbe useful to you. Should you need additional information we would be glad to provide further assistance.

More failure data information on check valves willbe provided later by special Letter Report.

Sincerely, ggP A. B. Poole ABP:jkc Attachments cc;enc:

= P. L. Campbell, NRC J. Colaccino, NRC K. L. McElhaney D. C. Fischer, NRC W. C. Gleaves, NRC F. Grubelich, NRC W. E. Vesely, SAIC J. P. Vora, NRC R. H. Wessman, NRC

I I

1']

Evaluation of Candidate LSSC Check Va1ves for Risk Based IST Extension at Palo Verde Units IPP K.L. McElhaney Oak Ridge National Laboratory Oak Ridge, Tennessee November 12, 1996 NRC Job Code W6324

1 h,

I I,

t I

if i

Evaluation of idate LSSC Check Valves for Risk Ba Extension at palo Verde Units 1g,3

Background

Palo Verde Nuclear Generating Station has recently submitted to the NRC a request for relief&om the Insemce Test.(IST) intervals etrrently required by the ASME Code for certain check valves based on a probabilistic analysis ofthe valves'mportance to safety. In theory, this analysis methodology results in the ranhng ofcomponents into two basic categories, those ofhigh safety signi6cance and those with low safety significance. The goal is to ensure that the components more important to plant safety are to be tested in a manner that provides a high level ofassurance oftheir operability. Another goal ofthe tusk Based approach (RBIST) is to show that IST intervals may be extended beyond the current requirements without resulting in signi6cantly increased safety risks. One consideration in this type ofanalysis is supposed to be component performance history, both from the speci6c plant as well as &oman industry persI)ective.

Palo Verde has requested IST extension from their current Code requirements (usually quarterly) to an interval of6 years on approximately 228 (assunung 76 valve application groupslunit x 3 units) "lowrisk signi6cant" check valves. In an effort to provide information needed to evaluate potential candidate check valves at Palo Verde for extended 1ST intervals, Oak Ridge National Laboratory (ORNL) has done a brief review ofthe available performance data for the valves in question.

The followingis a sutnmary report on that analysis.

Analysis Results An analysis was done on available component performance data using both the characterized data &om the ORNL check valve performance database and raw NPRDS data. The ORNL database is composed of over 2000 check valve failure records derived &om NPRDS from 1984-1992, and manually reviewed, 6ltered to remove non-failures, non+heck valves, and external leakage type failures, and characterized according to consistent criteria for a number ofparameters, such as failure mode, failure area, failure cause, specific valve type (where possible; e.g., swing check, liftcheck), etc. Raw (uncharacterized)

NPRDS data is not generally preferred for analysis puqeses due to the lack ofsome data and inconsistency in data input practices between plants, but for some portions ofthis analysis, raw Mure data for all Palo Verde check valve failures recorded in NPRDS &em 1986-1995 was also used.

CEOG Generic Valve Grou Where possible, it is particularly bene6cial to compare check valve performance based on specific application. Unfortunately, speci6c valve application information is rarely available, due to difFerences in plant designs and terminology and a hck ofinformation available &om NPRDS. When this type of comparison is desired, it is generally necessary to reviewplant~c FSARs and attempt to develop some type ofgeneric valve application groups. This task is usually both time~nsuming and &ustrating, since comIerisons can usually only be made among plants with the same NSSS and very similar system con6gurations.

A recent report issued by the Combustion Engineering Owners Group (CEOG), CE NPSD-1048, "Deinonstration Project to ApplyRisk Based hservice Testing (IST) to ECCS Check Valves," 'ttempted to develop certain generic application categories forECCS check valves in a number ofCE plants. Six utilitieswith a total often plants participated in the CEOG study. In order to gather data on check valves withinthe scope ofthe study, CEOG sought to Mlitate crossglant contparisons by developing a set of generic check valve con6guration diagrams with a corresponding set ofgeneric check valve groups based on location and function. Since Palo Verde was one ofthe ten plants participating in the CEOG study, ORNL was able to cross-reference the valves that appeared in both the CEOG report and the IST relief request in order to review both plant~c and industry valve performance based on the generic

S

Evaluation of ldate LSSC Check Valves for Risk Based Extension at Palo Verde Units 1,2,3 groupings.

Table 1 lists the CEOG generic group descriptions for those groups of "lowrisk significant" Palo Verde valves which also included candidate reliefrequest valves.

Table 1

CEOG Report CE NPSD-1048 Generic Valve Group Descriptions CEOG

~Rc rt Grou Group B Group D Group E Group H Group J Group K Group M Group 0 Group P Group R Group S Group T Description SIT Outlet Chock Valves LPSI Pump Discharge Check Valves LPSI Pump Suction Check Valves LPSI Pump MiniflowCheck Valves Hot Leg Injection Line Check Valves Hot Leg Injection Line to RCS Loop Check Valves HPSI Pump Discharge Check Valves HPSI Pump MiniflowCheck Valves Containment Spray Header Check Valves Containmcnt Spray Pump Discharge Check Valves Containment Spray Pump Suction Check Valves Containment Spray Pump MiniflowCheck Valves Palo Verde Valve Application Groups SIEV215, SIEV225, SIEV235, SIEV245 SIAV434, SIBV446 SIAV201, SlBV200 SIAV451, SIBV448 SIAV523, SIBV533 SIAV522, SIBV532 SIAV404, SIBV405 SIAV424, SIBV426 SIAV164, SIBV165 SIAV485, SIBV484 SIAV157, SIBV158 SIAV486, SIBV487 The 1984-1992 ORNLcharactcrized failure and 1991 population databases were used to review the performance history ofthe 13 groups ofvalves listed fiom the ten plants included in the CEOG report. A sutnmary ofthe initial findings is as follows:

Industry Failures Based on CEOG Report Generic Application Groups:

No Failures:

Groups E, H,O, P, R, S, T.

Group B:Ifailure; St. Lucie 2. Borg-Warner 12" DWG 73060 check valve. Valve was stuck open.

Significant.

Group D:Ifailure; San Onofre 2. Anchor/Darling 10" DWG 3454-3 check valve. Broken tack welds and binding between disc shrt and valve stem. Significant.

Group J:

)failure;PaloVctdc2. Borg-Warncr3" DWG77700checkvalve.

LLRThilure. Moderate.

Group K:

9failures-St. Lucio 2. AII3" Westinghouse Model 03000CS8800007 swing check valves.

Exccssivc seat leakage due to stcam erosion ofthe discs. Moderate.

I I

t I

I

Evaluation of idate LSSC Check Valves for Risk Ba Extension at Palo Verde Units 1,2,3 Group M:

6failuresg1) Palo Verde 2, Borg-Warner 4" DWG 79120-1 check valve. Scat leakage.

Moderate.

(5) St. Lucie 2, Anchor/Darting 4" DWG 3527-3 check valve:- (1) stuck openmuse unknownwgnificant; (I)damaged internal parts resulted in restricted motion due to wear-signi6cant; (1) stuck open due to packing bindingwgnificant; (I)internal damage (galling) caused restricted motion - combination design problem and operating errormgnificant; (1) internals galling resulted in restricted motion - material incompatibility and excess tightening of thc valve to the closed position - signi6cant.

==

Conclusion:==

St. Lucie has had operational problems with this valve due to a combination of causal factors. Failures were related to galling and binding ofinternal parts.

Ind Parameters Reviewed Usin the 1984-1992 ORNL Database Two additional industry-wiCk performance parameters were also investigated using the ORNL database.

This review focused on the specific candidate valves by manufacturer and design.

I Industry Failures ByManufacturer/Model Number:

~

AllPalo Verde valves reviewed were manufactured by Borg-Warner (now BW/IP).

~

AllBorg-Warner check valves in the "lowsafety significant" groups identified at Palo Verde are swing check valves, empt those valves in CEOG Group 0 (SIAV424, SIBV426), which are Borg-Warncr littcheck valves.

~

No other plants have Borg-Warner valves with modeVdrawing numbers corresponding to those at Palo Verde, since Borg-Warner apparently uses unique drawing numbers instead ofmodel numbers for each plant, so results ofa failure history search by model number were inconclusive. Additional design information is necessary to evaluate hilures ofspecifi Borg-Warner valves.

Industry Failures-Borg-Yarner Valve Failures at AIIPlants:

Borg-Warner (including Borg-Warner Corp., Byron-Jackson Pumps DivJBorg-Warner, Nuclear Valve Division/Borg-Warner, and Weston Hydraulics DivlBorg-Warner) totals 749 valves installed as recorded in the 1991 NPRDS database.

This makes Borg-Warner 13th ofover 150 valve manufacturers in terms of actual number ofvalves installed. (Note: some model numbers listed in the database forBorg-Warner may actually be Kerotest valves, which may differin design &om the other Borg-Warner valves.) Tables 24 show the Mure distributions ofall Borg-Warner check valves in all plants by various parameters,

&om 1984-1992. It should be noted, however, that in order to establish any relative failure rates, the population distribution based on each parameter must also be determined.

Any conclusions drawn without considering population effects would almost certainly be misleading.

Evaluation of idate LSSC Check Valves for Risk Ba Extension at Palo Verde Units 1,2,3 Table 2 Industry Borg-Warner Failure Distribution by Component hge Group Component Age Group at time of failure No. Failures Percent of Total

~Bar -Warner Failures

<<5

>>=5 and <<10

>>"-10 and <<15 19 10 1

63 33 3

Table 3 Industry Borg-Warner Failure Distribution by Plant hge Group Plant Age Group (at time of failure/

<<5

>>=10 and <<15

>>=5 and <<10 No. Failures 15 4

11 Percent of Total Borg-Warner Failures 50 13'7 Table 4 Industry Borg-Warner Failure Distribution By Extent of Degradation Extent of Degradation Moderate Si nificant No. Failures 20 10 Percent of Total Bor -Warner Failures 67 33 Table 5 Industry Borg-Warner Failure Distribution by Valve She Group Component Size Group

<<=2

>>2 and <<=4

>>10 No. Failures 15 13 2

Percent of Total Bor -Warner Failures 50 43 7

r r

t t

Evaluation of Idate LSSC Check Valves for Risk Ba~T Extension at Palo-Verde Units 1,2,3 Table 6 Iadustry Borg-Warner Failure Distribution by System ORNL Standard S stem Name AFW Containment Isolation Control Rod Drive CVCS Feedwater HPSI Reactor Recirculation RHR Standb L ukl Control No. Failures 5

~

10 3

2 3

3 1

2 Table 7 1ndustry Borg-Warner Failure Distribution by Manufacturer Model/Drawing Number and Extent of Degradation Manufacturer Model Drawin Number None listed 116CCB1-004 118FCB1-005 3-75500 316DCBL-005 485QBB1-002 74730 74750 75560 76790-1 77680-1 77700 79120-1 80200 82530 C F00-1206H-203 CN-1500-1009J-255 CN-1500-1206J-230 DWG 73060 MC4900-1206H-101 No. Moderate Failures No. Significant Failures 1

0 0

0 1

1 1

1 1"0 1

0 0

1 0

1 0

0 0

l I

Evaluation of idate LSSC Check Valves for Risk Ba Extension at Palo Verde Units 1,2,3 Table 8 Industry Borg-Warner Failure Distribution by Unit Unit Name No. failures Ext. of De radation CATAWBA2 BYRON 2 PAI.O VERDE 2 PERRY 1 COMANCHE PEAK 1 MCGUIRE 1 MCGUIRE 2 ARKANSASNUCLEARONE 2 BRAIDWOOD2 BYRON 1 ST. LUCIE 2 SUSQUEHANNA1 SUSQUEHANNA2 WNP-2 2 WOLF CREEK 1 7

3 3

3 2

2 2

1 1

1 1

1 1

Moderate B; Significant 1 Moderate 2; Significant 1 Moderate 2; Significant 1 Moderate 3; Significant 0 Moderate 1; Significant 1 Moderate 1; Significant 1 Moderate 2; SignHicant 0 Moderate 0; Significant 1 Moderate 1; Significant 0 Moderate 1; Significant 0 Moderate 0; Significant 1 Moderate 0; Significant 1 Moderate 0; Significant 1 Moderate 1; Significant 0 Moderate 0; Si nificant 1 Palo Verde Check Valve Failure Histo Usin Raw NPRDS Data 1986-1995 Allfailures ofPalo Verde check valves occiirring during the time Game 1986-1995 (inclusive) were downloaded Gom NPRDS. (This was done for completeness, since the current ORNL check valve performance database contains failures only through 1992, and many ofthe Palo Verde failures were assumed to have occurred after 1992. Palo Verde Units 1 and 2 began commerrial service in 1986, while unit 3 began commercial service in 1988.)

Failures were manually reviewed (and characterized only according to extent ofdegradation to the component), and external leakage type failures were included for most ofthe followinganalysis.

The followingare the results ofa cursory evaluation ofthe 106 NPRDS check valve failures fmm Palo Verde for this time period:

~

Fifty-fiveofthe 106 NPRDS failure records involved the deferral candidate check valves. Ofthe 76 valve application groups represented, 13 groups experienced repeat Mures. There were nine individual valves that experienced repeat failures and three valves that had repeat significant failures. Table 9 shows the number offailures by valve application group, unit, and extent of degradation.

~

Ofthe 55 failure records associated with the deferral candidate valves, 11 involved external leakage.

The remaining 44 failures were reviewed and characterized in terms ofextent ofdegradation in accordance with the criteria used in previous ORNL analyses.~

There were 29 (66%) failures deemed moderate in nature and 15 (34%) termed significant. This ratio is very close to that exhbited irMhstry-widefor 1991 and 1992.~

~

The set of228 deferral candidate valves (76 valve application gmups/urut x 3 units) accumuhted approximately 2100 valve-years ofservice during the period 1986-1995. Ifthe munber ofsignificant Mures only is considered, this represents a hilure rate ofappmximately 7x10 /yr. forthe seL "Significant" in terms ofthe degradation to the valve's abilityto function. These failures include those withbmken and/or detached intermris, restricted motion, stuck open, and stuck closed cases.

Evaluation of klate LSSC Check Valves for RLsk Based Extension at Palo Verde Units 1g,3 Table 9 Dlstnbution of Palo Verde Check Valve Failures (1986-1995) by Unit and Eztent of Degradation Valve Application System Grou Unit )

Failures Unit 2 Failures Unit 3 Total by Failures Group AFAV137 hFBV138 CHAV177 CHAV190 CHBV331 CHEV334 CHNV154 CHNV494 DGAV066 DGAV067 DGAV397 DGBV068 DGBV069 GAEV015 HPAV002 NCEV118 NPBV004 SGAV043 SGAV044 SGEV005 SGEV642 SGEV693 SIAV404 SIAV434 SIAV485 SIAV523 SIBV405 SIBV446 SIBV484 SIBV533 AFW AFW CVCS CVCS CVCS CVCS CVCS CVCS Diesel Starting Air Diesel Starting Air Diesel Starting Air ~

Diesel Starting Air Diesel Starting Air Containment Isolation Combustible Gas Control CCW Combustible Gas Control Main Steam Main Steam Main Steam Main Steam Main Steam RHR RHR RHR RHR RHR RHR RHR RHR 1S 1E 1E 3M 1M IM 1M IE IM IS 2E IE 1M 1M IE IM 1S 2M 1S 2S IM 3S 1S 1S 1M IM 2S IE IM IM 1E 2M 1M 2M 1M 1M IM IM 1M 1S 1S, IE 1M IM I

3 1

I 2

I 3

I 3

2 I

1

.4 M-Moderate failure S4ignificant failure F Ezternal leakage (no internab degradation)

~

Table 10 shows a list ofall candidate Palo Verde "lowsafety signi6cant" check valves application groups for IST interval deferral. Italso lists the number offailures recorded in NPRDS &om 1986-1995, and the number ofrepeat failures. Where applicable, the corresponding CEOG report generic valve application group is also listed.

Table i0 Verde RBIST Relief Request Check V alo e

T Deferral Candidate Cbeck Valve Application Grou AFAV007 AFAV015 AFAV137 AFBV022 AFBV024 AFBV138 CHAV177 CHAV190 CHAV328 CHBV331 CHEV334 CHEV433 CHEV440 CHNV118 CHNV154 CHNV155 CHNV494 CHNV835 CTAV016 CTAY037 CTBV020 CVCS CVCS CVCS CVCS CVCS CVCS CVCS CVCS CVCS CVCS CVCS CVCS S

em Failures in NPRDS 1986-1995) 1 nit 1 1 (Unit 1 1

nit 1 1

nit 2 2

nit 2, 1

nit 3 1

nit2,2(Unit3 2 (Unit 2 1

nit 2) peat lraiiures per Application Grou Notes 4,9 Valve Listed in CEOG rt2 CiSV038 DGAY066 DGAV067 DGBV068 DGBV069 GAEVOl1 GAEV015 HPAV002 HPBV004 NCEV118 SGAV043 SGAV044 SGEV003 SGEV005 SGEV006 SGEV007 SGEV642 SGEV652 SGEV653 SGEV693 SGEV887 Diesel Startin Air Diesel Startin Air Diesel Startin Air Diesel Startin Air Containment Isolation Containment Isohtion Combustible Gas Control Combustible Gas Control CCW Main Steam Main Steam Main Steam Main Steam Main Steam Main Steam Main Stcam Main Stcam Main Steam 3 (Unit 1), 3 (Unit 2),

1(Unit 3 1(Unit 1, 1(Unit 2 1(Unit 1, 1 (Unit 3 1 (Unit 1) 1 nit2), 1 nit 3 1 (Unit 3 1

nit 3) 1 nit 2 1

nit 3 1 (Unit I, 2 (Unit 2) 1 (Unit 3) 1 (Unit 1 2

nl't 3 2,6 10

t Table 10 Palo Verde RBIST Relief Request Check Valves 0

e e

T Deferral Candidate Check Valve Application Grou SGEV888 SIAV157 SIAV164 SIAV201 SIAV404 SIAY424 SIAV434 SIAV451 SIAV485 SIAV486 SIAV522 SIAV523 SIBV158 SIBV165 SIBV200 SIBV405 SIBV426 SIBV446 SIBV448 SIBV484 SIBV487 SIBV532 SIBV533 SIEV215 SIEV225 SIEV235 SIEV245 SPAV041

SPBV012, WCEV039 DGAV396 DGBV496 DGBV497 DGAV397 Main Steam Failures ln NPRDS (1986-199 1

nit 1 2

nit 1, 1

nit2 1

nit 1) 1 (Unit 1), 1 (Unit 2),

1 (Unit 3 1 (Unit 1, 1 nit 2 1

nit 2 1

nit 1) 1 (Unit 1), 2 (Unit 2),

1(Unit 3 ESW ESW Diesel Startin Air Diesel Startin Air Diesel Startin Air Diesel Startin Air 1

nit 2 peat failures per Application Grou

'?

Yes Valve Lbted in CEOG Notes Re rtf Grou S

Grou P

Grou E G

M Grou 0 Gro D

Gtou H Grou T Grou K 4

Grou J

Grou S

Grou E

4 Grou M Gro 0

Grou D

Grou H

Grou R

Grou T Grou K 2

Grou J

Grou B

Grou B

Grou B

Grou B

~

e 'jf"

Evaluation of idate LSSC Check Valves for Risk Ba Extension at Palo Verde Units 1,2,3 Table 10 Notes:

1 Stuck open failure due to corrosion and normal wear.

2 Repeat Mures duc to pitting and corrosion caused by debris and moisture buildup in system.

Rcpcat "Med to scat" failures.

3 Repeat leakage and binding failures attributed to corrosion buildup in system and normal operational and environmental wear/aging.

4 Internal leakage due to normal wear or aging/cyclic fatigue.

5 2/17/95: Leakage past seat attributed to wear. Disc stud broken due to cyclic fatigue.

3/27/95: Pieces ofvalve internals found to be missing, including a 2-inch length ofthe disc stud with the welded nut, stud sleeve, and washer.

Valve would not have functioned properly. Failure attributed to inadequate design and cyclic fatigue.

6'epeat stuck closed failures.

7 Broken hinge arm and loose internal parts attributed to possible cyclic fatigue.

8 Repeat hinge pin failure duc to wear.

9 Internal leakage caused by abnormal wear/cyclic fatigue.

10 Stuck open condition due to cyclic conditions.

11 Valve binding (restricted motion) due to inadequate assembly.

10

t t'

II I

Evaluation of idate LSSC Check Valves for Risk Ba Extension at Palo

'erde Units 1,2,3 Conclusions Thirteen oftbe generic valve application groups listed in the CEOG RBIST report contained valves for which Palo Verde has requested IST relief. Ofthese, seven groups had no recorded Mures for any ofthe ten plants in tbe CEOG study, while six groups did have failures recorded in tbc ORNL'failure database during the time period 1984-1992. Thc most Mures occiured in generic Group M, HPSI pump discharge check valves.

Failure histories based on valve manufacturer (Borg-Warner) and manuiacturer/model number were also reviewed with generally inconclusive results.

Since this particular manufacturer uses unique dmwing numbers rather than model numbers for its valves, making direct comparison based on design is dd5cult without additional information. Some other industry analyses based on various parameters related to Borg-Warner check valves were also presented.

Potentially the most significant findings resulted &om a briefreview ofthe 106 raw NPRDS failure records for Palo Verde Units 1,2,3 during the time period 1986-1995. Fifty-fiveofthe 106 NPRDS hilures involved defenal candidate valves. Thirteen valve application groups experienced repeat failures across all three units. Nine individual valves experienced repeat failures, and three valves had repeat significant Mures. It is important to note that at last 35 ofthe 44 failures involving internals degradation were attributed by Palo Verde (in the NPRDS nanatives) at least in part to some age-related Mure mechanism such as "wear," "cyclic fatigue," or "debris buildup." These types offailure causes must be considered when evaluating whether to extend inscrvice testing intervals.

Thc system ofservice for candidate IST deferral should also bc considered.

At least 75 percent ofthe deferral candidate "lowsafety significance" valves are located in either AFW, Diesel Starting Air, Containment Isolation, CCW, Main Steam, or RHR systems, which have been shown to have some ofthe highest relative failure rates by system for significant failures (in terms ofcomponent degradation).~

It is not clear &om tbe performance data reviewed so fiirthat IST interval extension isjustified for all the components listed in the Palo Verde reliefrequest.

Although both the Palo Verde reliefrequest itselfand the CEOG report identify specific component performance as a critical consideration in the determination ofboth level ofsafety significance and length ofinterval extension, how this criteria was applied is not straightforward. Neither document cites either plant~c or industry data as their source for check valve Mure rates used as input for the probabilistic analyses.

Instead, it appears that "generic" data was used as input for all the probabilistic analyses, which would fail to take into account any ofthe perfoimance history parameters reviewed herein.

ln order to fullyjustifyIST interval extension for any oftbe components listed as candidates for deferral in the Palo Verde reliefrequest submittal, a further review ofboth operational performance data and other plant practices should be undertaken.

For example, itmight be prudent to ask, " What measures have been taken to ensure that tbe Diesel Starting Airsystem is fiec ofcorrosion and debris caused by moisture inside the system?" This is esiiecially important, since fiem previous industry-wide studies it has been shown that Diesel Starting Aircheck valves have been especiaHy prone to failure (signljicant failure: both stuck open and stuck closed) &om this problem. The current ORNL review has also shown that several of tbc candidate Palo Verde valves in the Diesel Starting Airsystem have Med repeatedly forthe same, reason.

Other supporting programs such as plant maintenance and preventive maintenance should be reviewed also when considering IST deferral, since component performance and longevity are highly dependent upon these practices.

11

~

i

Evaluation of C ldate LSSC Check Valves for Risk Ba Extension at Palo Verde Units 1,2,3 References 1

Combustion Engino:ring Owners Group (CEOG), CE NPSD-1048, "Demonstration Project to Apply Risk Based Inservice Testing (IST) to ECCS Check Valves," June 1996.

2 Oak Ridge National Laboratory, NUREG/CR-5944, Vol. 2, "ACharacterization ofCheck Valve Degradation and Failure Experience in the Nuclear Power Industry - 1991 Failures," July 1995.

3 Oak Ridge National Laboratory, ORNVNRGLTR-96/11, "ACharacterization ofCheck Valve Degradation and Failure Experience in the Nuclear Power Industry - 1992 Failures," June 1996.

12

0

') rs lgI

~ '