ML20035B892

Forwards Completed Analysis of Public Comments Re Implementation Guidance for Maint Rule for Use & Info

ML20035B892
Person / Time
Issue date:	03/18/1993
From:	Correia R Office of Nuclear Reactor Regulation
To:	Will Smith NUCLEAR ENERGY INSTITUTE (FORMERLY NUCLEAR MGMT &
References
NUDOCS 9304050254
Download: ML20035B892 (1)

Text

.

MAR 181993 Mr. Walter J. Smith Senior Project Manager Nuclear Management and Resource Council Suite 300 1776 Eye Street, NW Washington, D.C.

20006-2496

Dear Mr. Smith:

The NRC staff has completed the analysis of public comments regarding implementation guidance for the maintenance rule.

For your use and information, I have enclosed an advanced copy of the analysis package.

Please let me know if you have any questions regarding the enclosed information.

Sincerely, Oricinal signed byr Richard P. Correia NRC Coordinator for NRC/NUMARC Maintenance Interactions Performance and Quality Evaluation Branch Division of Reactor Inspection and Licensee Performance Office of Nuclear Reactor Regulation

Enclosure:

As stated DISTRIBUTION:

[ Central Files f p 2 L

PDR RPEB R.F.

310097

[ a j/n'/. X m

N 0Y IO REGO [ QRY ((gr g y j)

(,,./;

9304050254 930318 PDR ORG NRRB i

ef 4

o UNITED STATES i

E E

. NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20lili5 s.,*****/

MAR 161993 i

MEMORANDUM FOR:

John T. Larkins, Acting Executive Director j

Advisory Comittee On Reactor Safeguards FROM:

C. J. Heltemes, Jr., Deputy Director for Generic Issues and Rulemaking j

Office of Nuclear Regulatory Research

SUBJECT:

TRANSMITTAL OF THE NRC STAFF'S ANALYSIS OF PUBLIC COMMENTS REGARDING IMPLEMENTATION GUIDANCE FOR THE MAINTENANCE RULE,-10 CFR 50.65 j

The subject material is forwarded for your review. The analysis, enclosure 1, was reviewed by the NRC Staff's Steering Comittee on Regulatory Guidance for-l the Maintenance Rule. Also enclosed is a copy of the For-Comment Regulatory Guide / Draft Regulatory Analysis /Backfit Analysis that was offered for public coment on November 24, 1992 and copies of the 11 public coment replies l

received. The public coments from NUMARC included a copy of their latest l

proposed revisions to the industry guidelines (NUMARC 93-01).

i The NRC Staff made presentations to the ACRS on October 6 and 8, 1992 on the.

I proposed public coment package and the ACRS provided comments in a' letter to the EDO dated October 15, 1992. The NRC Staff considers that the issues raised by the ACRS have been addressed, and will be resolved as outlined in the analysis of public coments to the regulatory guidance.

As before, we are requesting ACRS review and comments on an expedited basis.

4 Our schedule is to resolve any outstanding issues and publish the final Regulatory Guide by June 30, 1993.

.j

'I g~

]

f C

. Helte Jr., Deputy Director Generic Issues and Rulemaking l

Office of of Nuclear Regulatory Research E

Enclosures:

-l 1.

NRC Staff's Analysis of Public Coments 2.

For-Coment Regulatory Guide / Regulatory Analysis /Backfit Analysis 1

3.-

Public Coments j

cc: See next sheet, j

i

Contact:

Owen 0. Rothberg, RE' Mail Stop NLS314, 49-23924 l

t f i E. b,e v m

-t

/-

w 2

1 cc: (w/ enclosure 1 only)

C. Ader NLS 324 F. Akstulewicz 11 F23 H.- Alderman (20 copies w/all enclosures)

P 315 R. Baer NLS 302 S. Bahadur NLS 129 P. Baranowsky MNBB 9112 E. Beckjord NLS 007 Re ')n II J. Blake s

N. Blumberg Region I F. Cherny

?

NLS 302 D. Cleary NLS 314 K. Connaughton 16 H3 J. Conran MNBB 3701 R. Correia 10 D18 J. Craig NLS 007 M. Dey NLS 169.

E. Doolittle 16 H3 i

T. Foley 10 A19 G. Grant 17 G21 D. Hickman MNBB 9112

- t F. Jablonski Region III C. Johnson NLN 316 i

T. King NLS 007 M. Lesar P 223 M. Malsch 15 B18

' I i

E. McKenna 16 GIS 4

W. Minners NLS 007 G. Mizuno 15 B18 T. Murley 12 G18 P. Narbut Region V 4

T. Novak MNBB 3701 P. O'Reilly MNBB 9112 i

J. Page NLS 302 i

C. Petrone 10 A19 i

B. Richter NLS 129 M. Ring Region III C. Rossi 9 A3

0. Rothberg NLS 302 W. Russell 12 G18 J. Scarborough 16 H3 A. Serkiz-NLS 314-l L. Shao NLS 007 J. Sniezek 17 G21 t

T. Stetka Region IV M. Taylor 17 G21 S. Treby 15 B18 D. Trimble 16 H3 i

A. Vietti-Cook 16 G15 J. Yora NLS 2178 G. Zech 10 A19 Public Document Room (w/all enclosures) l i

i f

t I

j i

I 1

ENCLOSURE 1 i

j i

l i

MARCH 15, 1993 ANALYSIS OF PUBLIC COMMENTS l

REGARDING REGULATORY GUIDANCE FOR THE MAINTENANCE RULE, i

i 10 CFR 50.65

=

.i jj f ;

w

SUMMARY

The NRC Staff solicited public comments (57 FR 55286) on a Draft Regulatory Guide, DG-1020, "Honitoring the Effectiveness of Maintenance at Nuclear Power Plants." By DG-1020 the NRC Staff proposes to endorse an industry guidance document, NUMARC 93-01.

Eleven responses to the request for public comments were received.

In the Federal Register Notice..the NRC Staff specifically requested comments on four queskions. These questions related to the relationship of the guidance tollicense renewal, clarity of the guidance with respect to the requirements of the rule, the use of the concept of " inherent reliability,"

and the use of probabilistic risk assessment methods in the guidance. Five of the licensees made specific responses to the four questions. The questions are quoted and t, hen the public comments are listed below each, followed by an analysis.

The next section contains other public comments and NRC Staff analysis and responses. The last section contains additional NRC Staff comments on several issues such as scope and diesel generator reliability.

A a.

b DISCUSSION In Draft Regulatory Guide DG-1020, "Honitoring the Effectiveness of Haintenance at Nuclear Power Plants," the NRC Staff proposes to endorse an industry guidance document, NUMARC 93-01, " Industry Guideline for Monitoring the Effectiveness Of Maintenance at Nuclear Power Plants," - July 1992.

Public-comments were solicited on the regulatory guidance during the public comment.

period that was noticed in the Federal Register on November 24, 1992 (57 FR 55286) and officially. closed on January 15, 1993.

The maintenance rule,10 CFR 50.65, requires that commercial nuclear power plant licens'ees monitor the effectiveness of maintenance activities for plant equipment w,ithin the scope of the rule, 550.65(b).- Systems, structures, and (or) components (SSCs) that are considered under 550.65(a)(1) are to be monitored against licensee-established goals to ensure that the SSCs perform their intended function. SSCs considered under 550.65(a)(2) need not be monitored againsilicensee-established goals, provided that the performance of the SSCs is controlled by effective preventive maintenance such that the SSC performs its intended function. The provisions of 550.65(a)(3) require that licensees periodically evaluate their goal setting and monitoring efforts and make adjustments where necessary.

Further, 550.65(a)(3) requires that licensees balance SSC maintenance against. availability, and assess the impact of taking equipment out of service for maintenance on plant safety.

Eleven public comments were received.

Those who commented are:

1.

State of Illinois, Department of Nuclear Safety (IDNS) 2.

Yankee Engineering Services, Yankee Atomic Electric Company (Yankee) 3.

Westinghouse Electric Corporation, Energy Systems 4.

Entergy Operations, Inc. (Licensee for Arkansas Nuclear One, Grand Gulf and Waterford 3) 5.

Nuclear Management and Resources Council (NUMARC) 6.

Northeast Nuclear Energy Company (NNECO) and Connecticut Yankee Atomic Power Company (CYAPCO) 7.

Tennessee Valley Authority (TVA) 8.

Centerior Energy / Toledo Edison (Operator of Davis-Besse) 9.

Arizona Public Service Company (Operator of Palo Verde Nuclear Generating Station (PVNGS)) (APSC) 10.

NRC, Office of Research, Civision of Engineering (RES/DE) 11.

Consolidated Edison Company of New York, Inc. (Operators of Indian Point Station) (CON ED)

RESPONSE AND ANALYSIS OF OUESTIONS PUBLISHED IN THE FEDERAL REGISTER Public comments were solicited in general, and on four particular questions regarding NRC's draft regulatory guide, DG-1020, " MONITORING THE EFFECTIVENESS OF MAINTENANCE AT NUCLEAR POWER PLANTS," November 1992. The four questions were derived from a number of issues that were identified by the ACRS and NRC Staff organizations and compiled in a memorandum from E. Beckjord to S. Treby and E. Jordan, " Resolution if Issues Identified by ACRS, CRGR, and OGC Regarding Drpft Regulatory Guidance To Implement the Maintenance Rule,10 CFR 50.65," dated November 10, 1992. The four questions are reproduced below in bold type. yPublic comments and the NRC Staff's analysis of each question follow.

1.

"TheliceNserenewalrule,10CFR54,containsrequirementsthatare related to the maintenance rule.

Is it possible to apply NUMARC 93-01 as written, or to modify the guidance, in order to not only satisfy the maintenance rule but also to address the requirements of the license renewal rule?"

1DNS Reply:

"The answer to the first question is best discussed in relation to the NRC's own comments on the effects of aging as stated in the 10 CFR Part 54 Statement of Consideration sections IV.e.(ii)(1), (2), and (3).

In consideration of these comments, an applicant for a license renewai may be able to use the structures, systems, and components (SSCs) as determined via the maintenance rule as a starting point for determining the SSCs that are important to license renewal. But such a determination alone should not be used to define the scope of SSCs for license renewal. The pitfall of relying solely on the NUHARC document as the determinant of SSCs important to license renewal is that the analysis under the NUMARC document may not account for failure mechanisms that only reveal themselves subsequent to the current 40-year license. As an example of this potential conflict with the license renewal rule, see section 8.2.1.5 in NUMARC 93-01 that states, in part:

'As indicated in the above paragraph (8.2.1.4), the determination of potential failures that could hypothetically occur but have not been previously experienced is not required.'"

Yankee Reply Yankee indicated that, in their opinion, no modifications are needed to the guidance for the maintenance rule in order to accommodate the license renewal rule.

Westinchouse Reply:

i Westinghouse stated that the NUKARC guidance could be referenced in the guidelines for renewing plant licenses.

_4_

[nteray Reply:

"It is not possible to apply NUMARC 93-01 as written to address the requirements of the license renewal rule. However we do feel it is important to coordinate the two efforts. We do not believe the NUMARC 93-01 guidance should be modified because of the considerations listed i

below. We strongly recommend that 10CFR54 be modified to facilitate a coordinated. maintenance rule and license renewal effort. The license renewal rule scope should only address those aspects of plant' aging that are not addressed by offective maintenance programs. Modification of 93-01 should not be made given these following considerations:

Adding license renewal implementation requirements would require a gewrite to maintenance rule guideline (93-01). This would delay

-publication of a finished product to the nuclear industry for each tplant's rule implementation.

Putting.the maintenance rule and licensing renewal into the same guideling would make monitoring and goal setting prescriptive as 10CFh54 is currently written. To meet the requirements of 10CFR54, the performance monitoring would be at the component level. The establishment of the NUMARC guidance and the spirit of 10CFR50.65 is that monitoring could be performed on the plant, system, or train level. Component level performance monitoring -

would require significantly more manpower and resources to l

implement.

Some plants are not presently (or in the future) pursuing license renewal. Adding license renewal to the maintenance rule guidance would be burdensome and unnecessary for these plants."

t NUMARC Repiv:

HUMARC stated that it would be premature to modify NUMARC 93-01 to satisfy the license renewal rule requirements at this time because the license renewal rule implementation guidance has not been developed and the maintenance rule will have to be implemented long before most utilities had decided to seek license renewal.

NRC Staff's Response:

As noted above, the consensus of the public comments was that no modifications should be made to the proposed maintenance rule guidelines in order to better satisfy the requirements of the license renewal rule.

s The NRC continues to evaluate several alternatives regarding this issue.

~

However, the NRC Staff does not plan to modify the guidance for the maintenance rule to address specific provisions of the license renewal rule at this time.

i

)

~

2.

"IN 10CFR50.65(a)(1), the maintenance rule calls for monitoring performance or conditions of structures, systems, and components against licensee - established goals, and 10CFR50.65(a)(2) states that such monitoring is not required if the performance or condition of structures, systems, and components is effectively controlled through the performance of appropriate preventive maintenance such that the structure, system, or components remains capable of performing its intended function. The guidance being provided emphasizes the establishment of performance criteria to demonstrate that structures, systems, and components are effectively controlled through preventive maintenance.

Is the guidance sufficiently clear that an affirmative demonstration is necessary that the established performance criteria have been met if a structure, system, or component is to be considered to be Icontrolled under 10CFR50.65(a)(2)? If not, how could the clarity

+

and consistency of the guidance be improved?"

IDNS Reply:

IDNS indicates,that the industry guidelines are sufficiently clear that an affirmalive demonstration of acceptable performance is necessary to initially place SSCs under (a)(2).

Yankee Reolv:

Yankee indicated that current guidance is clear that an affirmative demonstration is necessary to assure that the established performance criteria have been met if an SSC is to be considered to be controlled under (a)(2). However, they stated that it should be clarified that existing surveillance and testing to ensure that present plant level performance goals are satisfactery and that additional system or train level monitoring is not required.

i Westinahouse Reply:

Westinghouse indicated that, with the proposed changes to the NUMARC l

guidance document (see NUMARC's public comments), the guidance is clear l

that an affirmative demonstration of satisfactory performance is necessary if a SSC is to be placed in the (a)(2) category.

The suggested word change is considered unnecessary because the next i

sentence in the NUMARC guidance clearly indicates that performance criteria must be satisfied.

Enteroy Reply:

"We feel that the guidance, as it is written, is clear. Our

^

interpretation is that an affirmative demonstration is necessary, and is clearly identified in the maintenance rule guideline. The guideline-supports the language used in 10CFR50.65."

NUMARC Reolv:

NUMARC stated that their guidance document is sufficiently clear to.

demonstrate that criteria must met to affirmatively demonstrate the control of SSCs under (a)(2) of the rule.

j NRC Staff's Response:

}

w

.~

The above comments indicate that the proposed guidance is sufficiently --

- - = -

clear. The NRC Staff notes that the industry' guidance was clarified on this. issue and now. states that an affirmative demonstration of 1

acceptable performance is necessary to place or retain a particular SSC l

in (a)(2).

SSCs.are not considered under:a separate' program under q

(a)(1), rather, for SSCs under (a)(1), goals are established and SSCs.

monitored against those goals in addition to the normal maintenance that is required under-(a)(2).

l 3.

"Both the statement of considerations for the maintenance rule (56 FR i

31308, July 10,1991) and NUMARC 93-01 refer to the concept of inherent i

reliabflity.

Is this concept sufficiently. clear, given the examples and discussion to describe the concept in NUMARC 93-01, or are there i

improyements that would help to better define this concept?"

r IDNS Reolv:

"Further clarification of this concept. appears to be warranted. The problem.is'the ' subjective nature of the term 'high reliability.'

Perhaps a better definition would be to say that an SSC is.of high-i reliability if it meets the licensee-established performance criteria with no preventive maintenance. Otherwise high reliability is too subjective; eg., is it 80%, 90%, 99%?"

y_ankee Reply:

Yankee stated that the concept of inherent reliability is sufficiently clear but the _ terminology could be enhanced to better indicate the relationship.

L Westinahouse Repiv:

Westinghouse indicated that the concept of inherent reliability is

.l sufficiently clear.

Enterav Reply:

"The concept of inherently reliable was understood by the V&V group and'-

interpretation was similar. The NUMARC guidance document defines inherently reliable as having high reliability without having preventive mai,ntenance.

j Examples of inherently reliable would be:

Buildings Cable = Trays Raceways'

'i Cable / Wiring -

j Piping"

+

NUMARC Reply:

j NUMARC stated that the concept of inherent reliability is sufficiently clear in their document; the verification and validation program resultsD did not indicate that further clarification is needed.

i

'I i.

l l

a--.

NRC Staff's Resoonse:

Public comments on this issue were mixed in that most commentors believed that the concept of inherent reliability was sufficiently clear. Yet, one commentor believed that further clarification was needed.

The concept of inherently reliable components was introduced in the Statement of Considerations that was published with the maintenance rule (57 FR 31308, July 10, 1991). The statement was:

"The purpose of paragraph (a)(2) of the rule is to provide an alternate approach (a preventive maintenance program) for those SSCs where it is not necessary to establish the monitoring regime required by (a)(1).

For example, this provision might also be used where an SSC, without preventive maintdnance, has inherently high reliability and availability (e.g.,

electrical cabling) or where the preventive maintenance necessary to achieve high reliability does not itself contribute significantly to unavailability (e.g. moisture drainage from an air system accumulator)."

The concepl of'an inherertly reliable SSC is described in Section 9.3.3 of NUMARC 93-01 as "one that, without preventive maintenance, has high reliability.

Inherently reliable components might include, but are not limited to, the following: electrical cables, wiring, certain manual valves and piping, pipe supports, and the reactor vessel."

The definition itself may be sufficiently clear that the document need not be revised. However, the NRC staff believes there are very few SSCs within the scope of the rule that will perform their intended functions without preventive maintenance. The definitions of maintenance and l

preventive maintenance are offered in Appendix B of NUMARC 93-01 as follows:

" maintenance: The aggregate of those functions required to preserve or

(

restore safety, reliability, and availability of plant structures, systems, and components. Maintenance includes not only activities traditionally associated with identifying and correcting actual or potential degraded conditions, i.e., repair, surveillance, diagnostic examinations, and preventive measures; but extends to all supporting functions for the conduct of these activities.

(Source:

Federal Register /Vol. 53, No. 56/ Wednesday, March 23,1988/ Rules and Regulations /page 9340.)

maintenance, preventive: Predictive, periodic, and planned maintenance actions taken prior to SSC failure to maintain the SSC within design operating conditions by controlling degradation or failure."

It is clear from these definitions that very few SSCs can continue to perform their intended functions without preventive maintenance.

For example, the reactor vessel requires rigorous inspection and surveillance (preventive maintenance) in order to ensure acceptable performance, Performance includes availability, reliability or condition, as appropriate, in accordance with footnote 10 of NUMARC 93-.

l

01. Without such main'tenance, the performance of the reactor vessel cannot be determined to be acceptable. The reactor _ vessel, if considered under either (a)(1) or (a)(2), would be monitored against the standard of meeting the established inspection and surveillance criteria -

that demonstrate acceptable condition.

(Performance criteria based on j

the appearance of visible cracks, or rupture. failures,~ or plant performance would not be useful.)

If inspection or. surveillance indicated that a trend in performance may reasonably be leading to unacceptable degradation, a_ licensee would be expected to set goals (e.g., temperature history, crack growth, etc.) and mv.,; tor the performance of the reactor vessel against those goals until the performance trends proved acceptable. Of course, _if the goals are not met, corrective action must be taken.

4.

"NUM ltc 93-01 outlines methods based on probabilistic risk assessment.

(PRA) to determine risk significance of structures, systems, and components _.

Is this guidance clear and does it satisfactorily address low frequency, high - consequence contributors (e.g., inter-system loss-of-coolant accidents and boiling water reactor anticipated transients without scram events), or are there improvements that would add to the clarity and completeness of this guidance?"

IDNS Reply:

"It would appear the Sections 9.3.1.1 and 9.3.1.2 of NUMARC 93-01 entitled, Criteria Determination Method I, and Criteria Determination Method 2, respectively, adequately encompass the questioned events."

l litokee Repiv:

?

Yankee indicated that the guidance should clarify that PRA results-.

should only by used as an input to an expert panel and should not be-relied upon solely to determine risk significance.

Westinahouse Reply:

Westinghouse stated that the NUMARC guidance document, as revised (see i

NUMARC public comment) adequately addresses the risk significance 1

issues.

Enteroy Reply:

"The guidance is sufficiently clear. The criteria prescribed in the NUMARC guideline yields specific results for each plant. However, using 1

the same criteria, the results may differ from plant to plant based on the varying risk significance of systems at different plants.. NUMARC.

1 93-01 should be revised to place less emphasis on the direct use of the '

PRA and more emphasis on expert opinion as a result of the V&V.

Plants i

should have the flexibility to utilize and rely on the, existing plant experience gained through operation of the plant and completion of the.

V&V-effort in making final determinations of a system's risk i

significance. While insights-for the plant's PRA are' valuable and j

should be factored into the final decision, the plant experience and knowledge of plant operations can not and should.not be ignored.

For l

l

-g-ij a

't'

~

f example, a list of systems (which could be-derived from PRA) should be used as input to an Expert panel. Decisions from this Expert panel could be the final filter for determining systems that are risk significant."

NUMARC Reolv:

As a result of the'V&V program, the risk significance guidance has been l

extensively revised to address low frequency high consequence o

contributors.

j NRC Staff's Response:.

Some commentors believed that revisions in NUMARC 93-01 were warranted j

concerning the concept and guidance on risk significance.

In this.

regard [ the industry guidelines were revised by NUMARC to clarify that Indivjdual Plant Examination (IPE) or PRA results serve as one source,

.among several, to determine risk significance.

Further, the text of l

NUMARC 93-01 regarding risk significance was extensively re-written 1

based on NUKARC's Verification and Validation (V&V) Program results.

The revise'd text calls for licensees to establish risk' significance by.

several parallel paths if PRA methods are used..The procedure consists i

of establishing lists of risk-significant systems by-three different 1

PRA-based methods and then having an expert panel. review the results and

.i arrive at a final, comprehensive list.

In this way, any of the j

shortcomings of any individual method of determining risk-significance-l will be mitigated. The method has been tested at several plants that participated in NUMARC's V&V effort, with acceptable results.

j I

i i

r i

i i

i l

1

'l I

t h j i

f f

4 4

OPAER PUBLIC COMMENTS (IN ORDER DOCKETED 1 AND NRC STAFF ANALYSIS / COMMENTS fpITORIAL NOTE:

The public comment (summarized or paraphrased in some cases) is shown in bold type and is followed by the NRC Staff's response.

I.

State of Illinois, Department of Nuclear Safety (IDNS) 1.

I...(T)he (maintenance) rule is not detailed enough to ensure yconsistent performance by all nuclear power plant owners.

Historically, there have been widespread levels of performance in the nuclear industry, including several plants that were shut down because of poor performance. We do not believe a generic mainl:enan'ce rule, with little detail on iconitoring, preventive maintenance, performance or acceptance criteria, will ensure optimum maintenance practices nationwide, or provide the NRC with a clear technical basis to regulate maintenance in the nuclear power plants."

The maintenance rule is results-oriented and is intended to be non-prescriptive. The Commission considered prescriptive regulation regarding maintenance as one alternative and chose to adopt the rule as it appears now, as explained in the statement of considerations that was published with the final maintenance rule (56 FR 31306, July 10, 1991). The Commission specifically designed the maintenance rule to provide maximum flexibility to licensees to implement its provisions utilizing their existing programs to the extent possible, in the interest of efficiency.

The regulatory guidance is intended to reflect the provisions and intent of the maintenance rule.

l The key to the provisions of the maintenance rule is performance of the equipment under consideration. The effectiveness of maintenance is to be evaluated in comparison the how well structures, systems and components (SSCs) perfcrm their intended functions. Thus, detailed or prescriptive guidance about how maintenance is to be conducted by licensees is, by intention, not provided by the maintenance rule. The existing provisions of regulations and guidance already provided by the NRC remain in place to provide specific instructions to licensees, where appropriate.

l 2.

"It appears to IDNS that either the rule or regulations associated with the rule need to be thorough and descriptive so that licensees know what is expected of them and the rule can be 1,

m readily enforced. Otherwise, maintenance performance will i

continue to vary widely throughout the industry.t IDNS then listed a number of specific items that they thought the rule t

should contain.

IDNS indicated that the rule is too general and i

that it lacks specific criteria. Their view is that guidance on i

specific maintenance practice is needed to provide for acceptable maintenance.

The Commission did not agree with this view and passed the i

performance-based, results-oriented maintenance rule in its present form. The Commission recognized that maintenance methods i

and practices vary between licensees. They also recognized that maintenance at nuclear plants is generally acceptable. The lommission seeks to support and maintain this situation, and do so

.f n an efficient manner. Maintenance practices are expected to i

vary between plants, as they do now, but licensees must l

demonstrate satisfactory performance, condition and availability of SSCs, which is the objective of the maintenance rule.

Licehsees will have to comply with the rule and the performance of SSCs will be assessed by licensees and the NRC accordingly. The regulatory guidance to be issued is expected to aid licensees and the NRC in achieving a common understanding regarding acceptable -

i methods for implementing the maintenance rule.

3.

IDNS made several suggestions and observations concerning the use of PRA and IPE analyses. These were as follows: All licensees should be required to use the IPE/PRA methodology in determining risk significance of SSCs. As NUMARC 93-01 is currently written, licensees can implement any standard they feel is appropriate and the mandatory use of IPE results would lead to more uniform and i

efficient use of resources.

IDNS made several suggestions-concerning the structure of IPEs. Finally, IDNS believes that NRC l

should review each licensees assumptions concerning determination -

of performance.

The NRC encourages the use of PRA methods to determine risk significance, however, such use is~not mandatory, as noted in the Statement of Considerations that was published with the maintenance rule on July 10, 1991 (56 FR 31308). The NUMARC guidance that the NRC staff proposes to endorse follows that-same philosophy. The NRC staff's observation of the V&V program indicates that most licensees plan to use the results of their l

IPEs to establish risk significance and establish goals or i

performance criteria. However, the NRC does not require use of PRA/IPE results. The NRC specifically does not impose standard methods of practice under the maintenance rule. These decisions

_j are left to licensees. The NRC will evaluate the success of each licensee's efforts to comply with the rule based on the i

performance, condition and availability of SSCs, as appropriate.

j 4.

IDNS questions the statement in the NUMARC guidance document that i j

" Documentation developed for implementation of the guideline is not subject to the utility quality assurance program unless the documentation used has been previously defined a within the scope of the quality assurance program." Will the NRC endorse NUMARC 93-01 but not audit the documentation resulting from the implementation of NUMARC 93-01, in accordance with 10 CFR 50, Appendix B.? To do so seems inconsistent with current practice and intent of the regulation.

Certain non-safety related SSCs covered by the maintenance rule, as defined in 10 CFR 50.65(b)(2)(i), (ii), and (iii), are not covered by Appendix B of 10 CFR 50. The NRC does not intend to extend the scope of Appendix B or the licensee's quality assurance

! program.

Further, the NRC does not want licensees creating

.. unnecessary paperwork and documentation to implement the

. maintenance rule.

It will be necessary for licensees to document certain.information, in order to demonstrate to themselves, as well as the NRC, that SSCs are performing their intended functions. This aspect is addressed in the proposed maintenance guidance.

II.

Yankee Atomic Engineering Services - Yankee Rowe and Connecticut Yankee 1.

These comments supplement the NUMARC comments, which they endorse.

See NRC Staff responses to the NUMARC comments, below.

2.

Guideline does not clearly limit goal setting and performance monitoring at the plant or system, or in some few cases, the train level.

The guideline intentionally does not specify or limit a licensee's choices regarding goal setting. The rule allows each licensee to decide the appropriate level of goal setting for their individual plant. The objective is to allow licensees flexibility to implement the performance-oriented maintenance rule.

Each licensee's implementation of the maintenance rule will be evaluated based on equipment performance, condition, and availability.

l 3.

The NUMARC guideline lacks definition of repetitive failure, endorses NUMARC proposed definition.

See the NRC Staff's comments on NUMARCs Comment 9 below.

4.

Guideline should clarify the limitations of the use of " industry experience" as a direct comparison for performance measure purposes. -

The provisions concerning the use of industry-experience were revised by NUMARC in the guidance document, as indicated in their comments. Refer to the NRC Staff's analysis of NUMARC's comment 10.

5.

NRC inspection and enforcement guidance should recognize that differences in plant design, system boundaries, etc., will result in differences in each licensee's implementation of the rule.

The NRC staff recognizes that this situation will be the case and this will be reflected in the inspection guidance when it is developed.

$he definition of the scope which includes all SSCs in the 6.

.. emergency operating procedures (EOPs) is too broad.

It should be

. limited to "SSCs identified as the principal means to control key plant safety parameters or mitigate the effects of core damage or radi,oacti,ve release."

The concept adopted is to include those SSCs that are of significant value to completing an E0P.

Limiting SSCs to the

" principal means" of completing an E0P is be read in the context of the examples that follow the discussion.

The following comment is, to some extent, editorial. On page 8, lines 7-11, NUMARC 93-01, the sentence reads; "For a nonsafety-related SSC to be considered important, it must add significant value to the mitigation function of an E0P by providing the total (or a significant fraction of the) functional ability required to mitigate core damage or radioactive release (e.g., required quantity of water per minute to fulfill the safety function)."

The NRC Staff's concern is that licensees will focus on the word

" total" and ignore or discount the parenthetical phrase that follows.

It is suggested that the sentence be revised to make it clear that a "significant fraction" need not be at or near the total functional ability.

7.

Yankee does not agree that new types of monitoring need to be established to satisfy performance monitoring requirements for J

SSCs under (a)(2).

The NRC Staff agrees that most performance monitoring requirements will be satisfied by existing technical specification and ISI/IST surveillance and testing. However, some new monitoring may be needed where the licensees determines that existing monitoring is inadequate to ensure the performance requirements for SSCs under the maintenance rule. The NUMARC guidance outlines this approach i

to SSC monitoring under (a)(2).

8.

Yankee does not agree that all the monitoring characteristics (system-specific criteria, surveillance results, or plant level,

1

+

i

~

performance) for SSCs ur. der (a)(2) of the rule need to be assessed annually because (a)(3) of the rule refers to evaluation of performance and condition monitoring activities and associated goals. Since the term goals only applies to SSCs under-(a)(1),

the evaluation described in (a)(3) only applies.to-(a)(1), not (a)(2).

Section (a)(3) of 10 CFR 50.65 specifically requires that

"(p)erformance and condition monitoring activities and associated goals and Dreventive maintenance activities (emphasis added) shall i

be evaluated...". Further, it is clear that the intent of (a)(3) is for licensees to evaluate their entire scope of activities prformedunder(a)(1)and(a)(2).

III. Westinghouse Electric Corporation, Energy Systems j

1.

Westinghouse made 17 connents and suggestions about the NUMARC guidance ' document to improve clarity.

l These comments and suggestions will be considered in the NRC Staff's review of NUMARC 93-01. The comments appear to be editorial except:

(1)

Guidance to licensees to consider performance data for two refueling cycles or 36 months for the licensees initial evaluation of performance. Westinghouse observed that some licensees may not have this data for some SSCs. The NRC i

Staff's view is that the existing guidance is reasonable.

If there are SSCs included under the scope of the rule for which data has not been collected previously, the licensee should commence data collection.

l (2)

A change in the text of NUMARC 93-01 (Section 9.3.2).that indicates that "All risk significant SSCs determined to have acceptable performance are placed in (a)(2) and monitored under plant level performance criteria..." The NRC Staff.

believes that such an approach would result in practical

^l difficulties in determining acceptable performance; see the NRC Staff's response to NUMARC Comment 22 below.

l l

r IV.

Entergy Operations, Inc. (Licensee for Arkansas Nuclear One, Grand Gulf I

and Waterford 3) 1.

"Entergy Operations does not endorse changing the maintenance rule to incorporate requirements of the license renewal rule. However, we strongly believe the license renewal rule should be reconsidered in light of.the inherent benafits derived by maintenance rule programs. We believe the set of components

'l monitored for age related degradation can be significantly reduced

) 1 i

by plant level and system level performance monitoring via the maintenance rule. The scope of the license renewal rule should only address those aspects of plant aging that are not addressed by effective maintenance programs."

The NRC Staff agrees that the maintenance rule should not be changed to incorporate requirements of the license renewal rule.

The NRC is currently evaluating methods to coordinate the implementation of both rules.

2.

"The NUMARC 93-01 guidance places too much emphasis on specifically identifying the scope of SSCs subject to the requirements of the rule. This emphasis could result in Enwarranted and unproductive regulatory effort in scrutinizing

.fwhich SSCs are in the scope of the rule. The primary objective should be to identify the important performance criteria and then to match equipment to that criteria."

The'NRC Staff believes that the guidance provided is appropriate.

The scope of SSCs under the maintenance rule is clearly established and the NUMARC guidance provides an acceptable method for licensees to identify which plant-specific SSCs are covered by the rule. The focus of NRC inspection activities will be related to ascertaining reliable performance of important functions that

~

relate to the overall plant, system, and train safety, and the overall effectiveness of maintenance activities in maintaining those functions.

Should a licensee fail to include SSCs whose function is important to maintaining a safety function, and as a result are not adequately maintained as evidenced by unacceptable performance, then the scope of the licensee's program may receive specific NRC review.

3.

"The guidance places too much emphasis on using the PRA. The rule can be adequately addressed with no reliance on the PRA. More direction should be provided on alternate means of satisfying the requirements."

The emphasis placed on PRA in the guidance is provided to encourage uniform risk-based methodologies.

The process for determining which SSCs within the scope of the maintenance rule _

are to have specific or plant-level performance criteria applied is based on their risk significance. Since all licensees prepared a PPA or IPE it would appear reasonable that it should be used in implementation of the maintenance rule.

4.

"The guidance states that risk significant and standby systems will have specific performance criteria and implies that this will be at the system level.

It may well be that the most appropriate performance criteria for some of these systems would be at the plant level.".

~

In the industry guidance, NUMARC specifies that risk significant and standby systems are to have specific performance criteria, and permits setting the criteria at the system level.

Performance criteria is also to be set at the component or train level when system performance is not acceptable. However, a plant level performance criteria would be inappropriate for a standby system because its performance would not be reflected in plant performance.

r Hasking of true performance because performance criteria are set at too high a level could apply to any level of performance criteria. Performance criteria must be set at the appropriate Jevel for the SSC and its related conditions. NUMARC 93-01 l

provides guidance to licensees regarding the appropriate level for yperformance criteria; however licensees must decide, based on the risk significance, the operating mode, and the actual or expected performance, what level of performance criteria is appropriate.

x 5.

"Section 8.2.1.6 addresses SSCs that are not within the scope of the rule. This section is irrelevant and should be deleted."

Section 8.2.1.6 of the NUHARC guidance document addresses those SSCs outside the scope of the rule and provides examples of those i

SSCs. While this section may not be critical to implementing the intent of the maintenance rule, it is appropriate in an industry guideline document to clarify by example what SSCs are or are not within scope when such decisions may be uncertain, and to emphasize that continued attention to maintenance on those SSCs outside the scope of the rule is desirable.

6.

"As mentioned above, the method for determining risk significance using the PRA should be modified to place more emphasis on expert opinion."

~

The methods outlined in the industry guidance for determining risk significance were substantially revised by NUMARC as a result of i

the NUMARC V&V effort (see NUKARC's comments). The proposed method uses the expert opinion of an on-site review group to screen the results of three PRA methods for identifying risk-significant SSCs.

7.

"Section 11 addresses the removal of equipment from service for maintenance. This section appears overly prescriptive. Section 11.2 provides guidance for the development of an approach.

Section 11.2 should be revised to make it clear that this is not the only method that licensees may use.

It should state that licensees may use other approaches, provided they satisfy the intent of 10CFR50.65(a)(3)."

In section 2, Purpose and Scope, of the industry guidance it is b.

i i

clearly stated that licensees "may elect other suitable methods or-approaches for implementation." With respect to the comment that Section 11 is prescriptive, it is true that licensees are specifically required by the maintenance rule to assess the effects of equipment being out of service for maintenance or monitoring on performance of required safety functions.

V.

NUHARC NUHARC responded to the request for public comments and also enclosed a copy of their guidance document that they have revised, based on the result.eoftheirV&Veffort. The revised NUHARC guidance document is enclosed.

't NUHARC proposed 39 changes to their industry guidance document, NUHARC 93-01, based on their V&V program as well as other industry reviews.

These are. analyzed below and numbered according to the NUHARC representalion. The NUHARC comment is quoted (bold) and the NRC Staff's response is stated below each comment.

4 1.

" Comment 1 (Section: Executive Summary, Page: iii, Line: 22) For clarification add after the first sentence: To be placed in (a)(2), the SSC will have been determined to have acceptable l

performance."

The NRC Staff agrees. See Question 2 and the responses from the Federal Register Hotice (FRN questions) above.

2.

" Comment 2 (Section: Executive Summary, Page: 111 Line: 27) The V&V program results indicate that a better definition and examples of standby systems and trains would improve the understanding and intent of the guidelines.

Indicate a footnote number at line 27 after the words '... standby model' and describe the footnote as follows: Refer to the Appendix B definition and examples of standby systems and trains."

The NRC Staff agrees with the definition of standby system or train.

3.

" Comment 3 (Section: Executive Summary, Page: iii, Line: 28)

~

Phrase beginning 'For example.. ' is not a complete sentence and should be clarified as follows:

'The high pressure coolant injection system is an example of a system that is in a standby mode during normal plant operations and is expected to perform its safety function on demand.'"

i The NRC Staff agrees.

j 4.

" Comment 4 (Section: Executive Summary, Page: iv, Line: 2) To make the executive summary consistent with footnote 7 on page I change the word after '... trains,' from the word 'and' to the word 'or'."

l.

The NRC Staff agrees.

5.

" Comment 5 (Section: Executive Summary, Page: iv, Line: 6) To clarify the scope of IPE add a footnote number after the word

... Examination' and describe the footnote as follows:

'As used in this document the scope of IPE includes both internal and external events.'"

The NRC Staff agrees.

6.

" Comment 6 (Section: Figure 1, Page: vi, Line: Not applicable) The logic diagram as presented in Figure 1 is intended as an overview

,g}ogic flow and as such does not describe the full text of the

. uideline. However, industry and NRC questions indicate that some

-concepts would be more clear if added to the logic diagram.

Two such concepts are:

Establishing specific performance criteria for non risk signjficantstandbysystems;and More efearly demonstrating risk significant and non risk significant logic flow paths.

The logic diagram has been modified as indicated by a cloud around the changed areas."

The NRC Staff agrees.

7.

" Comment 7 (Section: 4.0, Page: 3, Line 6) The inclusion and treatment of changes to the plant configuration would be more explicit if addressed in the applicability section of the guideline. After the last sentence in the existing paragraph add a new sentence as follows:

Periodically, as a result of design changes, modifications to the plant occur that may affect the maintenance program. These changes should be reviewed to assure i

the maintenance program is appropriately adjusted."

The NRC Staff agrees.

8.

" Comment 8 (Section: 8.2.1.4, Page: 9, Line: 12) The meaning of this sentence can be improved by adding the word ' subsequent' before the word ' failure'".

The NRC Staff agrees.

9.

" Comment 9 (Section: 8.2.1.4-Page: 9, Line: 14) V&V participants recommended that the meaning of the term ' repetitive failure' should be defined or made more explicit in its use. At line 14 delete the words '... repetitive failures.' and insert the words

...the loss of a safety function due to a subsequent MPFF for the same maintenance related cause'".

The NRC Staff agrees that a definition of " repetitive failure" is needed, however, the restriction in the NUMARC definition to a.

j j

l loss of " safety function" is too narrow. Failures ~that result in j

a loss of function that is within the scope of the maintenance rule should be evaluated.

The NUMARC recommended definition of repetitive failure is "...

I the loss of safety function due to a repetition of a failure for j

the same maintenance related cause."

The definition focuses on i

the failure of the " safety function" for each case described.

l This could be interpreted to mean that only a failure of a safety i

related SSC would be a maintenance preventable functional failure (MPFF) and that failure of any SSC that serves no direct " safety-function" would not be an MPFF. For example, repetitive failure i

pf a feedwater pump would not be a MPFF. Failures of equipment j

-used in E0Ps that contribute (substantially) to another SSC's 3

t safety function, or provide an alternate means of providing a safety function would not be a MPFF.

ThesMaintenance Rule focuses on maintaining function commensurate l

with' safety of all SSCs within the scope of the rule. The rule does not address just those SSCs that are safety related.

Both (a)(1) and (a)(2) of the rule emphasize the intent of maintaining i

the functional capability of all SSCs within the scope of the rul e.

1 Repetitive MPFFs (on a single SSC) within a short period of time are seldom random. After corrective maintenance is performed on a SSC and before that SSC is returned to service, it should be tested and placed in service totally functional.

Post-maintenance j

testing should be conducted in a thorough and comprehensive manner such that any other outstanding deficiencies would be identified, 1

and corrective actions should be taken prior to returning the SSC to operation. Otherwise the overall maintenance was not offective.

The main focus of the maintenance rule is the effectiveness of i

maintenance. Several failures in a given period of time on a single component could indicate ineffective maintenance even if each separate failure is for a different reason. ~ Although a specific piece-part may have failed which caused inoperability, maintenance should not be performed solely to correct that piece-.

i part failure, but to ensure that the total component /SSC is functional.

Nuclear plant maintenance should focus on the total SSC oerformance (availability, reliability and condition) not piece part replacement.

10.

" Comment 10 (Section: 8.2.1.4, Page: 9, Line: 16) The maintenance-i rule is a performance based rule and the term industrywide experience is not defined either in the :J1e or the existing i

guideline. To minimize the potential ~ for diverse interpretation after issuance of the guideline many corr.nentors requested clarification. On line 16 after the words 'industrywide j b

4

~

experience (') add a footnote number and describe the footnote as follows:

'Industrywide experience is appropriately considered in utility specific programs and procedures when the experience is determined applicable to the utility.

It is. appropriate to use this information to the extent practical to preclude unacceptable performance experienced in the industry from being repeated.

However, an event that has occurred at a similarly configured plant need not be considered as the basis for requiring SSCs involved in the event to be considered within the scope of the maintenance rule by other utilities that have not experienced the event.'"

ficensees should not disregard industry-wide experience that findicates that certain SSCs should be within the scope of the maintenance rule.

650.65(a)(3) states that licensees are to consider, where practical, industry-wide operating experience in evalyating their maintenance program effectiveness, 11.

" Comment 11 (Section: 8.2.1.5, Page: 10, Line: 9) Although section 8.2.1.4 is referenced in Section 8.2.1.5, it would be clearer to state the expectation directly.

In this way, more consistent implementation would result. Delete the second sentence of the second paragraph beginning on line 9 that begins

'As indicated in the above paragraph...' Replace the deleted sentence with text copied from 8.2.1.4 beginning on line 16 and ending on line 19 with the words

'...is not required.'"

The NRC Staff agrees. The last sentence might be clearer if constructed as follows:

" Licensees need not postulate hypothetical failure scenarios resulting from system interdependencies."

12.

" Comment 12 (Section: 8.2.1.6, Page: 11, Line: 9) Add example as follows:

' Examples of categories of equipment that are outside the scope of the Maintenance Rule Fire Protection SSCs Fire protection SSCs that are used under applicable CFR Part 50, Appendix R requirements are nonsafety-related and, therefore, are not included within the scope of the Maintenance Rule unless they meet the guidance of paragraphs 8.2.1.2, 8.2.1.3, 8.2.1.4, or 8.2.1.5.

Seismic class II SSCs installed in proximity with seismic class I SSCs Seismic class II SSCs are not included within the scope of the Maintenance Rule unless they meet the guidance of paragraphs 8.2.1.2, 8.2.1.3, 8.2.1.4, or 8.2.1.5.

Security SSCs.

The SSCs used for the security of nuclear power plants are nonsafety and their maintenance provisions are addressed separately under the requirements of 10 CFR Part 73. Security SSCs are not included within the scope of the Maintenance Rule unless they meet the guidance of paragraphs 8.2.1.2, 8.2.1.3, 8.2.1.4, or 8.2.1.5.

Emergency facilities described in the emergency plan Examples include the technical support center (TSC), operations support center (OSC), and other emergency operating facilities (EOFs).'"

Jhe examples may be misleading. All of the SSCs cited in the examples could fall under 550.65(b)(2)(i), (ii), or (iii).

7 Although the text following the examples states this, the inference is that their specific functional category as either fire protection SSCs, seismic class II SSCs, security SSCs, or emersency facilities inherently excludes them from the scope of the rule. This is not the case. SSCs should be directly compared to the scope of the rule to determine if they are within the scope of the maintenance rule. For example, fire protection systems failures have resulted in scrams or trips, and fire protection inoperability can very likely cause a loss of safety SSC function.

Therefore, fire protection systems would, in general, be within the scope of the maintenance rule and are not specifically excluded as the example implies.

13.

" Comment 13 (Section: 9.2, Page: 12, Line: 22) For clarity add the word '... initially' between the words 'to' and ' determine...'"

The NRC Staff agrees.

14.

" Comment 14 (Section: 92, Page: 12, Line: 30) To better describe the scope of (a)(2) the following sentence should be added:

'SSCs that are within the scope of paragraph (a)(2) of the rule could be included'in the formal PM program, be inherently reliable (monitored by walkdowns, etc.), or be allowed to run to failure (provide little or no contribution to system safety function).'"

The NRC Staff agrees that licensees should have the option to run equipment to failure, provided that regulations or licensee commitments are not violated, and SSCs under the maintenance rule can perform the functions as described in the scope of the rule.

However, licensees are expected to estimate or consider, in advance, the consequences of running equipment to failure.

15.

" Comment 15 (Section: 9.3.1, Page: 13, Line: 36) An example of risk significant and non risk significant function for the same SSC should be added to enhance understanding. Add the following after the first sentence:

'An example of an SSC that is risk significant for one failure mode and non-risk significant for

, l 4

f i

another is as' follows: Blowdown valves on steam generators perform a safety function to close on isolation. However, the open position function is to maintain water chemistry which is a nonsafety function.'"

r Once a SSC is determined to be risk significant for any reason, specific performance criteria or goals are expected to be established for that function under the industry guidelines.

16.

"Coment 16 (Section: 9.3.1, Page: 14, Line: 8) For clarity, the word 'but' should be replaced with the word 'and'."

The NRC Staff agrees.

17. ;"Coment 17 (Section: 9.3.1, Page: 14, Line: 22)The meaning of the sentence beginning 'For risk significant SSCs...' can be made more precise. There is an implication that risk significance dete,rmination per se is a requirement of GL 88-20, which is not the case. Modtfy the sentence to read as follows:

' Risk significant SSCs may be determined in accordance with a PRA similar to that used for response to GL 88-20, Individual Plant Examination for Severe Accident Vulnerabilities. The assumptions developed for GL 88-20 could also De used in the calculation of the total contribution to core damage frequency (CDF) and 10 CFR Part 100 type releases as a basis for establishing plant-specific risk significant criteria.'"

The NRC Staff agrees.

18.

"Coment 18 (Sections: 9.3.1, 9.3.1.1, 9.3.1.2, Page: 14, Line:

28) As a result of V&V program activities for determining risk significant SSCs it was determined that each of the methods

- described in revision 2A of the guideline were not alternative methods because the methods resulted in different SSC selection.

Additionally, PRA methodologies have various limitations that would be easily recognized by an expert panel.

To implement the results of the V&V program the following changes to the guideline are recomended.

Delete all text from page 14 line 28 to page 15 line 15 and replace it with the text that follows." Thereafter, HUMARC provided 2 alternative pages of text. The text is not reproduced here, however, a copy of the NUMARC guidance is enclosed."

The NRC Staff agrees.

19.

"Coment 19 (Section: 9.3.2, Page: 15, Line: 19-32) Additional clarity can be achieved by making the first paragraph three paragraphs and adding the following text that is underlined.

' Performance criteria for evaluating SSCs are necessary to identify the standard against which performance is to be measured.

Criteria are established to provide a basis for determining satisfactory performance and the need for goal setting. The actual performance criteria used should be availability, reliability or condition.'

'The performance criteria could be quantified to a single value or range of values. For example, if a utility wanted to maintain an availability of 95 percent for a particular system because that was the assumption used in the PRA then the 95 percent value would be the performance criteria. If the performance criteria are not met, then a goal could be sat at a value equal to or greater than the 95 percent. Additionally. an example of condition as a Serformance criteria would be a case in which a utility wanted to haintain the wall thickness of a pipino system to comply with the

ASME code reouirements. The utility would establish some acceptable value for wall thickness and monitor of ultrasonic testino or other means.'

'If)erfo'rmancecriteriaarenotmet,thebasisforthecriteria should be reviewed to determine if goal setting is required and the appropriate goal value established.

It should be recognized that while performance criteria and goals may have the same value and units, goals are only established under (a)(1) where performance criteria are not being met and are meant to provide reasonable assurance that the SSCs are proceeding to acceptable performance.'"

The NRC 5taff agrees.

20.

" Comment 20 (Section: 9.3.2, Page: 15, Line: 35) The reason for non-risk significant SSCs in a standby mode needing to have specific performance criteria is not stated. For additional clarity add the following text after the sentence that ends on line 9.

' Standby systems (either risk significant or non risk significant and safety related or non safety related) may only affect a plant level criteria if they fail to perform in response to an actual demand signal. This means that a standby system could be failed but its inability to perform the function for which it was designed is not known until it is required to perform in response to a demand signal or during test (e.g., a surveillance test to determine operability). The mode in which most standby system failures are observed is during testing. Because plant transients occur less frequently, failure on demand provides minimal information. For this reason, a plant level criteria may not be a good indicator or measurement of performance.'

'The acceptance criteria for a standby system can be qualitatively stated as " initiates upon demand and performs its design function for its required mission time". The reliability of a standby system to satisfy both criteria can be quantitatively established as calculated in PRA methodology.'"

The NRC Staff agrees.

21-

" Comment 21 (Section: 9.3.2, Page: 15, Line: 37) This paragraph states that ' Plant level performance criteria are established for all remaining non-risk significant normally operating SSCs.' This is not necessarily true as there may be other non-risk significant 1

SSCs whose performance can not practically be monitored by plant level criteria. To more properly bound SSC that would not necessarily affect a plant level criteria insert the following

... or other words after the acronym '...SSCs...' on line 38:

performance criteria are established if appropriate (e.g.,

repetitions of safety function failures attributable to the same maintenance related cause).'"

tThe NRC Staff agrees.

22.

Comment.22 (Section: 9.3.2, Page: 15, Line: 40) To improve the clarjty qf the text replace lines 40 through 42 ending with the word... criteria.' with the following text:

'All risk significant SSCs determined to have acceptable performance are placed in (a)(2) and monitored under plant level performance criteria (e.g., trips or unplanned safety system actuations or specific criteria as described in the example below (reference Section 12.2.2).

An example of the process is as follows:

o SSC is determined to be in scope of the Maintenance Rule; o

SSC is determined to be risk significant by expert panel; and o

SSC performance is determined to be acceptable to specific criteria.

An example of the criteria could be an acceptable level of availability or reliability relative to core damage 1

frequency contribution.

Therefore, the SSC may be addressed in (a)(2) under plant level l

criteria, plant specific criteria, e.g., availability, reliability, or failure rate as determined appropriate by the utility.'"

The statement is contradictory to the guidance provided for establishing performance criteria. Risk significant SSCs are to be identificd for the specific purpose of setting up system or train level performance criteria and monitoring the SSCs against those criteria under (a)(2). Only non-risk significant, active SSCs are to be monitored against plant level goals.

It is not consistent identify risk significant SSCs and establish performance criteria and then discontinue monitoring SSCs against that performance criteria immediately after acceptable performance is established.

Specific performance criteria are to be established and SSCs monitored against performance criteria in order to ensure the continuina effectiveness of maintenance for those SSCs.

23.

"Coment 23 (Section: 9.3.2, Page: 16, Line: 2) The use of the expression unacceptable performance should be amplified by adding after the word ' performance' the words 'as defined in Section 9.3.4'".

The NRC Staff agrees.

24.

"Coment 24 (Section: 9.3.1, Page: 16, Line: 37) Performance over time since an original PRA was performed can affect the original PRA assumptions. Therefore after the word ' maintained' on line 37 add the words '...or adjusted when determined necessary by the utility.'"

khe phrase "when determined necessary by the utility" does not fit ythe thought of the rest of the sentence. The need to adjust reliability and availability assumptions used in PRAs, etc. is driven by the need.to ensure that these assumptions are correct.

At b,est,,the phrase is redundant and should be deleted.

25.

Coment 25 (Section: 9.3.2, Page: 16, Line: 41) A guideline document should not use the word 'shall' unless it is specified in the associated regulation. Change the word 'shall' to the word

'should' and to clearly state the type of failure, add 'MPFF' after the word ' previous'.

The NRC Staff agrees.

26.

"Coment 26 (Section: 9.3.4, Page: 18, Line: 22 and 30) To improve the consistency of the text, add to line 22 between the words

'significant* and 'SSC' the words 'or non risk significant.'

In addition, on line 30 insert the acronym 'MPFF' after the word

' repetitive'."

The NRC Staff agrees, however, the " bullet" would be clearer and more concise if it read; "A MPFF caused a SSC performance criteria not to be met".

27.

"Coment 27 (Section: 9.4.4, Page: 23, Line: 8) Insert a new paragraph after the bullet that concludes on line 7 to read as follows:

'During initial implementation of the maintenance rule, repetitive failures that have occurred in the previous two operating and refueling cycles should be considered. After the initisl rule implementation utilities should establish an appropriate review cycle for repetitive MPFFs (e.g., during the periodic review, during the next maintenance or test of the same function, or in accordance with Section 9.4.3).'"

The NRC Staff agrees, provided the words "at least" are added before the word "two" in the new paragraph.

Licensees should not be arbitrarily' advised to limit their historical investigaticns if they consider it necessary to look further back.

6 i

28.

" Comment 28 (Section: 9.4.4,- Page:- 23,1 Line: '23-25)L'The.

l determination to allow failure...' at line 24: implies that u

failure may be accepted even with'no corrective action to reduce

]

the failure rate. This would not likely be the decision for a risk significant.SSC. To clarify the intent change at line'25 the

-l sentence beginning with the words 'Forl example'... ' to'. read:

~j

'For example, a decision to replace ~ a failed. component 1h31 i

orovides little or no contribution to safety function rather. ti.an:

performance of a preventive maintenance activity may reduce -

j exposure, contamination, and cost without impacting safety

.(Section 10.2).'"

1 t

Jhe NRC Staff agrees that it is acceptable for a licensee to make 4 conscious and realistic (engineered) decision.to let a component.

vfail. However, licensees should make these decisions before-the-fact, based on realistic estimates of the expected results and consequences of failure.

l 29.

"CoNnent'29(Section: 9.4.4, Page: 23, Line: 29) To make line'29 consistent with line 22 of the same paragraph after the words

... evaluated until the' add the words ' performance criteria or...'"

j The NRC Staff notes that according to the industry guidelines, 1

SSCs will continue to be monitored whether or not performance '

criteria or a goal is met.

30.

" Comment 30 (Section: 9.4.4, Page: 23, Line: 38) To clarify-the.

text on lines 38 and 39, delete:. 'The cause determination'should be documented for failures of risk significant and non risk' 1

significant SSCs.' Add: 'The cause determination results 'should be.

j documented for failures of SSCs under the scope of the Maintenance-j Ruie.'"

1 The NRC Staff agrees.

It-is. suggested that. reference be.made to' l

Section 13, DOCUMENTATION, of NUMARC 93-01.

j 31.

" Comment 31 (Section: 10.2, Page: ~26, Line: 21) Add the following _

sentence: SSCs that are within the scope of' paragraph (a)(2) of the rule could be included in the formal PM program, be inherently..

reliable (monitored by walkdowns, etc.), or be allowed to run to '

a failure (provide little or no' contribution to system safety

~

function)."

Refer to comment 3 for a discussion of inherent reliability and.

comment 28 for additional remarks on running equipment to' failure.

The~ decision to allow a system or component (the. acronym, "SSC" is inappropriate here) to "run to failure" should be rationally evaluated before the failure occurs and should be based _ on realistic estimates of the expected results of the decision.

i Unanticipated results of either letting' a system or component operat( in a deteriorated condition as a consequence of the.

Y

.g

A

~

decision of allowing it to run to failure should be investigated and disposed in accordance with the guidelines.

32.

"Coment 32 (Section: 10.2.2, Page: 28, Line: 6) To improve the accuracy of the text insert the word ' operating' between 'the' and

'SSCs'."

The NRC Staff agrees.

33.

"Coment 33 (Section: 11.2, Page: 29, Liae: 27) For clarity add the word '...pl ant... ' after the word '...during... '"

peNRCStaffagrees.

34. y"Coment 34 (Section: 12.2.3, Page: 33, Line: 10) To improve the accuracy of the text change the word ' evaluated' to the word

' performed.'"

The hRC S'taff agrees with the change made in the body of the NUMARC guidance document (the word " evaluated" was substituted for the word " performed"). The description of the change is correct in the text but is incorrectly described in comment 34.

35.

"Coment 35 (Section: 12.2.4, Page: 33, Line: 27) To improve the accuracy of the text insert the word ' operating' after the word

'For....' and insert after the word '... criteria

• insert the words

...and for standby systems to measure performance against specific criteria.'"

The NRC Staff agrees.

36.

"Coment 36 (Section: 13.2.1, Page: 35 Line: 29) Delete the text on lines 29 and 30 and replace the text with the following sentences:

' Changes to plant configuration because of plant modifications should be reviewed to determine SSCs that have been added to or deleted from the scope of the maintenance rule. Plant modifications could also change risk significance.'"

The NRC Staff agrees.

37.

"Coment 37 (Section: Appendix B, Page: B-1, Line: 41) A definition is needed for the acronym MPFF and for the term

' repetitive' as it relates to an MPFF. Add a new definition as follows:

' Maintenance Preventable Functional Failure (MPFF) -

initial and repetitive:

An MPFF is the failure of an SSC that causes the loss of any safety function where the cause of the failure of the SSC is attributable to s maintenance related activity. The maintenance related activity is intended in the broad sense of maintenance as defined above.

The loss of the safety function can be either direct, i.e., the SSC that performs a design safety function fails to perform the safety function or indirect, i.e., the SSC fails to perform its intended safety function as a result of the failure of another SSC (either safety related or nonsafety related).

An initial MPFF is the first loss of a safety function that is attributable to a maintenanca related cause. An initial MPFF is a failure that would have been avoided by maintenance activity that has not been otherwise evaluated as an acceptable result (e.g.,

" allowed to run to failure due to an acceptable risk or determined to be inherently reliable).

A ' repetitive' MPFF is the second or subsequent loss of a safety function (as defined above) that is attributable to the same paintenance related cause that has previously occurred (e.g. an AiOV fails to close because a spring pack was installed backward -

ethe next time this MOV fails to close because the spring pack is installed backward: the MPFF is repetitive and the previous corrective action did not preclude recurrence). A second or subsequent loss of a safety function that results from a different mainl'enance related cause is not considered a repetitive MPFF (e.g., an MOV initially fails to close because a spring pack was installed backward -- the next time it fails to close, its failure to close is because a set screw was improperly installed: the MPFF is not repetitive.'"

The definition of MPFF is too narrow as written because it is confined to loss of safety function. The maintenance rule covers both safety-related and non-safety-related SSCs and the failures of these SSCs to perform their intended functions must be directly addressed by the definition.

The definition of MPFF should include unacceptable degradation to cover passive SSCs.

(See the response to NUMARC's Comment 9 above.) Section 9.3.4 of NUMARC 93-01, "retermining Whether an SSC Level Goal is Required," does not cover either unacceptable

• e.

degradation of passive SSCs or unacceptable availability. The distinctions in the definition should provide licensees with guidance on how to best investigate repetitive failures rather than instructions about when such investigations can be avoided.

38.

"Coment 38 (Section: Appendix B, Page: B-2, Line: 28) Provide a

~

definition and examples of standby SSCs as follows:

'A standby system or train is one that is not operating and only performs its intended design function when initiated by either an automatic or manual demand signal. The following cases illustrate standby systems and trains:*

(Three examples were provided. These are not repeated here. The reader is referred to the copy of NUMARC 93-01 that is provided as an appendix to this analysis.)

The NRC Staff agrees.

39.

" Comment 39 (Section: Appendix B, Page: B-2, Line: 31) The.

sentence as stated is unclear. Revise the sentence to read 'An SSC that is required to be available for automatic operation must be available and respond without human action.'"

The NRC Staff agrees.

VI.

Northeast Nuclear Energy Company (NNECO) and Connecticut Yankee Atomic Power Company (CYAPCO)

NNEC0/CYAPC0 provided the following essay coment on the regulatory guidance:

"In 9.eneral, NNECO and CYAPCO support those views provided by NUMARC, and those of Yankee Atomic Electric Company (YAEC).

In addition, NNECO and CYAPC0 have provided coments on the following points whi,ch we considered most significant."

"NNECO and rYAPCO's single largest area of concern regarding the Maintenance Rule / Guidelines involves the definition and use of the terms " Maintenance Preventable Functional Failure (MPFF)" and

" Repetitive." Special requirements related to goal setting and cause determinations are specified by the guidelines on the basis of these terms. NNECO and CYAPCO consider the current terminology subject to a wide range of interpretation. Viewed in the extreme, this could result in overly programatic treatment of structures, systems, and components (SSCs) independent of their safety signi ficance. "

"NNECO and CYAPC0 consider a primary objective of the Maintenance Rule to be the focus of effective maintenance efforts upon SSCs comensurate with their safety significance. Consistent.with this interpretation, NNECO and CYAPCO advocate definition of an MPFF as a failure, at a specific functional threshold, based upon the consequences of this failure. This functional threshold should be determined by the individual utility based on the SSC's relative safety significance.

It is expected that the appropriate functional threshold would be most often at the train level or higher."

"In short, NNECO and CYAPCO consider prevention of repetitive SSC failure a necessary element of a maintenance rule program.

However, the manner and degree to which related efforts are applied, regarding failure analysis and cause determination, should concentrate on preservation of safety function at the train level."

"As stated, Question 2, posed by the NRC, i.e., "is the guidance sufficiently clear that an affirmative demonstration is necessary...," has generated some concern. The question could be -

construed to mean that specific performance criteria are necessary as a justifying basis for any and all SSCs to be placed in category 1(2). So interpreted, this requirement could become counterproductive, in that it would force essentially equivalent treatment of all SSCs independent of safety significance. NNECO and CYAPC0 envision an approach wherein the majority of non-risk significant, non-standby SSCs are assessed, via plant level criteria, as part of the periodic assessment process.

SSCs would be transferred to category a(1) and/or more effective criteria developed in response to emerging negative performance trends."

"The Draft Regulatory Guide DG-1020 is currently scheduled for final approval in July 1993. At that time, DG-1020 should endorse a newIrevision to NUMARC 93-01 (i.e., Rev. 3) which integrates Rev. JA and all applicable modifications.

Endorsement of Rev. 2A as written, in combination with separate modifyino documents, should be avoided."

"Overall,'hNECd and CYAPC0 view the V&V process as successful in terms of yielding an effective guideline document. However, as total industry experience increases during detailed implementation efforts, many additional issues requiring disposition and/or clarification are expected to arise. To address these issues, NNECO and CYAPC0 stress that:

1.

NUMARC and NRC involvement, similar to the V&V, should continue, and that at least one change to the guideline (i.e., Rev. 4) should be scheduled for not later than July 1994 and, 2.

Maintenance rule workshops should be conducted for the appropriate industry and NRC personnel to ensure consistent interpretation of acceptable implementation approaches."

i,

4 s

It would appear that the majority, if not all of the comments applicable to the industry guidelines'have been addressed in the revised NUMARC guidance document (see NUMARC's response to public comments above). A detailed definition of HPFF was provided and commented on by the NRC Staff.

Goals are to be set commensurate with the SSC's importance to safety.

NNEC0/CYAPCO's interpretation appears consistent with the rule. The functional threshold is permitted to be determined by the individual utility, as long as it is reasonable, based on the SSC's relative safety significance.

See the response to NUHARCs Comment 37, above for a discussion of the NRC Sfaff's view with regard to the definition of " repetitive failure" (MPFF) offered by NUKARC.

The NRC Staff.has no plans to change the existing arrangements regarding the conduct of activities between the NRC and the industry representatives.

It is expected that NUMARC will make changes to their guidelines as a result of the planned workshops, and from lessons learned during the implementation process. These changes are under the control of NUMARC. The NRC Staff will evaluate the changes that NUHARC makes and then evaluate the need to change our regulatory guidance concerning endorsement of the NUHARC guidance accordingly.

The Maintenance Rule focuses on maintaining function commensurate with safety of all SSCs within the scope of the rule. The rule does not address just those SSCs that are safety related or possess a direct safety function.

Both (a)(1) and (a)(2) of the rule emphasize the intent of maintaining the functional capability of all SSCs within the scope of the rule.

VII. Tennessee Valley Authority (TVA)

TVA provided the following comment on the regulatory guidance:

"TVA supports the comments on this draft Regulatory Guide made by the Nuclear Management and Resources Council (NUMARC) regarding implementation of the Maintenance Rule. TVA further supports the NUKARC comments resulting from the Verification and Validation (V&V) program and review by the NUMARC Maintenance Working Group regarding the need for clarification of the Industry Guideline, NUMARC 93-01. However, TVA has reviewed Comment 26 to the guideline relative to repetitive failures that have occurred in the previous two operating and shutdown modes'to determine potential impact on TVA.

It is TVA's position that further clarification should be provided to limit review of repetitive failures to Maintenance Preventable Functional Failures (MPFF) that affect safety functions. The three-year retrospective period implied by the comment will have a significant resource impact on industry tending and analysis programs with little value added.

It is also TVA's position that the

~

review period specified in NUMARC 93-01 (one refueling cycle not to exceed 24 months) is a more realistic period."

The NRC views repetitive failures as a potential indication of-ineffective corrective actions, and thus, within the broad definition of maintenance, ineffective maintenance.

(See the NRC Staff's comments on the definition of " repetitive failure" offered by NUMARC in their public comment 9.) Repetitive HPFFs (on a single SSC) within a short period of time are seldom random. Post-maintenance testing should be conducted such that deficiencies would be identified, and corrective actions taken prior to returning equipment to operational status.

(Also, refer to the NRC Staff's comments on the definition of HPFF, Comment 37, in the respony to NUMARC's public comments, above.)

Thedinfocusofthemaintenanceruleistheeffectivenessof maintenance, as measured by performance (which includes condition and availability,.as well as reliability).

Several failures in a given period of stime on a single component could be a sign of ineffective maintenance eve'n if each failure is for a different reason. Although a specific piece-part may have failed which caused inoperability, maintenance should not be performed to correct only that piece-part failure, but to ens,ure that the whole component is functional.

The NRC is currently developing a rule change to the required evaluation frequency to endorse a review period of each refueling cycle and not to exceed 24 months.

VIII. Centerior Energy / Toledo Edison (Operator of Davis-Besse)

Centerior restated their position that the maintenance rule,10 CFR 50.65, and the backfit analysis performed in support of the meintenance rule, are flawed because the maintenance rule will place an unnecessary financial burden on operating nuclear plants without resulting in a substantive improvement in safety.

If the rule is not rescinded, Centerior supports the comments submitted by NUMARC.

The NRC Staff intends to implement the maintenance rule as approved by the Commission. Our comments on the industry guidelines, HUMARC 93-01 are noted above.

IX.

Arizona Public Service Company (Operator of Palo Verde Nuclear Generating Station (PVNGS))

PVGNS endorsed Baltimore Gas and Electric coments, however Baltimore Gas and Electric did not submit public comments.

PVGNS endorsed NUMARC comments with the following exceptions:

I l

a.

Use of cumulative risk reduction does not provide any additional insights.

b.

There is no need to impose specific performance criteria on Non-Risk Significant SSCs.

With respect to comment a, the results of NUMARC's V&V effort indicate that a single perspective on risk significance is not particularly useful. The guidance provided by NUMARC is to approach risk significance from several PRA perspectives, if PRA methods are used, and then screen the results using a panel of knowledgeable personnel. There appearoto be devotees of any number of variations of PRA methods for deterdining risk significance. NUMARC consulted several experts in the indusfry and after several trials, made revisions submitted with the public comments. The approach is acceptable to the NRC Staff.

In general individual performance criteria are not to be applied to non-risk si,gnificant SSCs, except for standby systems, under the industry guidelines.

Standby systems are discussed in the NUMARC public comments above.

l X.

NRC, Office of Research, Division of Engineering 1

1.

"The proposed draft Regulatory Guide (DG-1020) and the accompanying Industry Guideline (NUMARC 93-01, Revision 2A, July 9, 1992) for Monitoring the Effectiveness of Maintenance at Nuclear Power Plants do not explicitly address time dependent age-related degradation in SSC that are within the overall scope of the maintenance rule. These documents do not provide guidelines for evaluating the effectiveness of maintenance programs to include passive SSC such reactor coolant pressure boundary components and other 'long lived' components (that may degrade with time and compromise safety and may not experience failures) such as ' pipes' and ' cables'. We believe that these deficiencies can be addressed through the following considerations:

o Risk significant and important (Footnote in memo:

determined deterministically and based upon design and operating experience; expert's opinion; existing codes, standards and regulatory instruments.) SSC generally include 'long lived' passive structures and components of interest) should be treated under paragraph (a)(1) of the maintenance rule with established goals to manage age-related degradation.

o

'Short lived' active components identified for preventive maintenance and condition monitoring program and treated under paragraph (a)(2) of the maintenance rule should be evaluated for maintenance effectiveness to manage age-i

s

~

related degradation.

The inclusion of the aforementioned considerations in the NRC/ Industry Guidelines will help achieve greater harmonization and the integration of the maintenance rule and the license renewal rule."

Neither the maintenance rule nor its implementing guidance exolicitly address age-related degradation or, for that matter, any other degradation mechanism. The rule is performance based and results oriented; it does not specify or describe specific maintenance practices to mitigate specific degradation or failure mechanisms. This means that the standard by which the 7effectivenessofmaintenanceistobemeasuredistheparformance yof SSCs within the scope of the rule. Performance, as noted in the guidance, includes availability, reliability, and condition as appropriate. The ability to " explicitly address" age-t elated degradation is not required by the maintenance rule and, ther'efore, is not reflected in its implementing guidance.

Guidance is provided under the maintenance rule, although not directly, to evaluate the effectiveness of maintenance for all SSCs within the scope of the rule including 'long lived' passive structures and components as well as 'short lived' active components.

Licensees are to establish performance criteria for SSCs under the scope of the rule. These performance criteria are established and modified by the licensees, although such criteria must allow licensees to judge if performance of SSCs under the rule is acceptable.

Under the NUMARC guidelines performance criteria are set at the plant, system, or train level. This is done to avoid having to manage a very large number of performance criteria so that attention can be focused on significant issues. The performance criteria are based on the maintenance requirements and commitments, including inspection and t'esting that are currently in place.

SSCs within the scope of the rule are monitored against the performance criteria.

If the performance criteria are not met, goals are set and the SSCs monitored against the goals until the condition that caused the performance criteria to not be satisfied is corrected.

If the goals are not met, corrective action must be taken to remedy the situation. The performance criteria (and goals) are related to determining that maintenance is effective, and are not specifically focused on identifying ace-related degradation mechanisms for structures or components. Licensees will, no doubt, identify degradation and failure mechanisms as part of their cause analyses for failures or in their periodic evaluations.

However, they will not have to establish specific performance criteria or goals to identify age-related degradation t.

i f

.l S

l mechanisms, as opposed to other degradation or failure mechanisms, i;

unless failures or unacceptable degradations prompt them to do so.

j Obviously, for components such as the reactor vessel or primary

~

system piping, performance criteria based on failures such as j

rupture would not be useful or acceptable. Performance criteria i

for vital components such as the reactor vessel need.to be based

+

on satisfaction of current inspections, tests, and surveillance.

i i

Trends of. unsatisfactory results should be noted and goals should be established if necessary. - The goals could be similar.to the-performance criteria. Under the industry guidelines, performance-would be evaluated similarly under either (a)(1) or (a)(2). - The-i pajor difference would be that under (a)(1), utility management:

4ttention is to be focused on results of monitoring the

performance of the system or component until the situation that.

i caused concern is resolved. Thus, placing and keeping the reactor i

vessel and primary piping under (a)(1) does not fit the approach propgsed,in the guidance, nor-is such required by the maintenance l

rule.

+

2.

In the draft regulatory guide, DG-1020, page 1, line 7,

"(d)efinition of ' effective maintenance' should be included.

j Effective maintenance is defined by the requirements of 550.65(a)(1) and (a)(2).

In each of these paragraphs the standard for success of licensees in meeting the requirement is that SSCs t

I within the scope of the rule remain capable of performing their.

intended function. Maintenance is effective if this criterion.is met.

3.

The draft regulatory guide, DG-1020, page 1, line 10, "(m)entions

' safety equipment.' However, the overall scope of SSC goes beyond 4

' safety equipment.' Recommend consistency."

I The sentence referred to is paraphrased from the regulatory analysis that is referenced by footnote in the same sentence. The-remainder of the paragraph goes on to explain the additional implications of good maintenance, just as the comment recommends.

i In the' context of the paragraph and the complete sentence, the phrase is consistent and accurate.

t 4.

In the draft regulatory guide, DG-1020, page 2, line 2, "(w) hat is meant by ' unacceptably degraded?' Recommend definition or provide expl anation."

Again, unacceptable degradation is to be measured against the standard for the SSCs to perform-their intended function, as is explained in the maintenance rule.

5.

In the draft regulatory guide, DG-1020, page 4, paragraph 1,

"(t)his paragraph should be expanded to include the treatment of l l i

passive structures and components associated with the reactor coolant pressure boundary, parameter trending, and maintenance effectiveness to manage age-related degradation effects."

The comment that the guidance should " include the treatment of passive structures," etc. appears to suggest that specific methods of establishing performance criteria or goals be specified. This is not in accordance with the rule or the NRC Staff's specific intention to avoid prescriptive instructions to licensees concerning how to conduct maintenance. The paragraph in question is considered to be an accurate description of the extent of monitoring that is required by the rule and reflected in the draft gulatory guidance.

6.

rIn the draft regulatory guide, DG-1020, page 5, lines 2 & 3, the

"(d)efinition of ' safety significance' should be included."

The gente,nce will be modified to substitute the term " risk i

significance" which is adequately described in the industry guidelines referenced by the regulatory guide.

7.

In the draft regulatory guide, DG-1020, page 5, last line,

"(r)evise the last line to read as follows:

' submittal for construction permits, operating licenses, and license renewal (as appropriate)'."

The sentence will be modified, however, the suggested word addition is unnecessary.

Licensees who adopt the industry guidelines will have to implement the guidance in accordance with the effective date of the maintenance rule (July 10,1996).

Licensees who apply for license renewal and use the industry guidelines will have to continue to comply with them during the license renewal period.

8.

In the draft regulatory analysis to DG-1020, page A-1, line 16, a

"(d)efinition of phrase ' safety significant' should be provided.

The phrase is intended to cover that equipment described in the scope of the maintenance rule, 550.65(b). The sentence will be modified to refer to equipment "within the scope of the maintenance rule."

9.

In the draft regulatory analysi:: to DG-1020, page A-1, line 17,

"(w) hat is meant by ' minimizing the likelihood of failure?' An explanation or definition would be of interest.

The phrase, as used in the context of the sentence, refers to the fact that failures can never be totally eliminated, and the objective is to minimize their likelihood.

10.

In the draft regulatory analysis to DG-1020, page A-1,-line 33

"(i)t is suggested that the performance criteria are to be established in rare cases at component levels.

In real world, for maintenance effectiveness (,) performance criteria are generally established at component level, especially for fluid-mechanical systems and for long-lived passive structures and components.

Recommend deleting the words 'in rare cases.'"

Performance criteria may be established at the plant, system, train, or component level. The foundation of the guidance is that acceptable performance at the component level will be reflected in successful performance at the plant, system, or train level, as appropriate. The reason to avoid setting performance criteria (or 7 goals) at the component level, to the extent possible, is to avoid t aving so many performance criteria or goals that the licensee h

cannot focus attention on the important ones.

11.

In the draft regulatory analysis to DG-1020, page A-5, lines 20 &

21, *(s)ubstitute the word 'should' for 'could' and delete the words 'with some exceptions'."

The comment was offered without an explanation of the concern.

The sentence is not intended as an instruction to licensees, but rather, a description of how the maintenance rule might be of benefit to licensees in areas outside the maintenance rule.

12.

In the draft regulatory analysis to DG-1020, page A-6, lines 6, 7,

& 8 "(r)ecommend deleting the statements (1) 'The full....from the rule.' and (2) 'The staff worked closely with NUMARC....would be t

addressed.'"

The comment was offered without an explanation of the concern.

13.

In the draft regulatory analysis to DG-1020, page A-7, Paragraph 6.2, " Relation to Other Existing or Proposed Requirements," RES/DE proposes to add text similar to that suggested in their comment I above.

For the reasons stated in the above response to RES/DE comment 1, the maintenance rule does not require that licensees place SSCs in,

the (a)(1) category or demand that licensees evaluate age-related degradation mechanisms, absent an identification of unacceptable SSC performance related to such mechanisms.

14.

In the backfit analysis to DG-1020, page B-6, line 16, " add 'and mitigation of age related degradation.'"

The suggested addition refers to a description of the guidance in the backfit analysis. The regulatory guide and industry guidelines do not call for such practices and the maintenance rule does not specifically require them. The description of periodic

s l

'~ '

performance assessment is sufficiently. descriptive as written in the current backfit analysis.

15.

In NUMARC 93-01, page IV, line 34, "after 12.0) add 'and mitie9 tion of age-related degradation.'"

Again, it is not a requirement of the rule to specifically perform i

performance assessments to address age-related degradation.

If-such practices are needed for license renewal, guidance should be provided that is related to license renewal.

1 16.

1[n NUMARC 93-01, page VI, Figure 1, diamond block related to fs3.3, " revise it to read 'is condition or performance acceptable?'"

r The definition of " performance" in NUMARC 93-01 includes

" availability, reliability, or condition as appropriate."

17.

In NDMARC' 93-01, page VI, Figure 1, block related to' 9.3.1,

[

" Supplemental consideration....and regulatory.nstruments."

Refer to the response to comments I and 13 above..

18.

In NUMARC 93-01, page VI, Figure 1, block related to 9.4.1, " Risk i

i significant.... age-related degradation."'

Refer to the response to comments 1 and 13 above.

e XI.

Consolidated Edison Company of New York, Inc. (Operators of Indian Point Station) t I

Consolidated Edison has participated in the review of the industry.

guidelines and accepted the NUMARC guidance document.

The NRC Staff noted the endorsement.

D

+

f F

T ADDITIONAL NRC STAFF COMMENTS i

1.

Geographic / Administrative Scope of the Maintenance Rule:

The following statement (in quotes)is not,.as of now, in the industry guidance, NUMARC 93-01, or the NRC Staff's implementing regulatory guide but should be included in one or the other. Originally the ACRS suggested that switchyards adjacent to the site, but not on it, might include SSCs that are within the scope of the maintenance rule. This led to the NRC Staff's conclusion that the physical extent of each The fol.pe's consideration of the maintenance rule needs to be defined.

licens lowing is offered in an attempt to bound the responsibility of licensees to define those SSCs that are to be considered when determining those SSCs that are within the scope of the maintenance rul e.

"The scope'of t'he maintenance rule, as defined in 10 CFR 50.65(b), is limited to SSCs that are within the cognizance of the licensee and that directly affect plant operations, regardless of what organization actually performs the maintenance activities.

For example, electrical distribution equipment out to the first inter-tie with the off-site distribution system should be considered for possible. inclusion in the scope of the maintenance rule. This would generally include equipment in the switchyard, regardless of its geographical location."

2.

On-site Review committee for the License Renewal Rule, SS4.33(c):

The following paragraphs will be added to the Discussion section of the NRC Staff's regulatory guide to implement the maintenance rule.

"Certain requirements for an a renewed license under 10 CFR 54, may be satisfied by taking credit for activities required by the maintenance rul e.

However, the renewal rule requires (10 CFR 54.21 (a)(6)(iii)),

among other provisions, that an effective program must be implemented by the facility operating procedures and reviewed by the on-site review committee. The maintenance rule does not have these requirements."

3.

Use Of The Maintenance Rule In License Renewal Activities:

The extent to which the maintenance rule can provide an effective tool toward meeting the requirements of the license renewal rule will be addressed in the implementing guidance for the license renewal rule.

4.

Diesel Generator Reliability:

The following paragraphs will be added to the Discussion section in the NRC Staff's regulatory guide for the maintenance rule.

i k

" Industry and NRC-sponsored PRAs have shown the risk significance of emergency AC power sources. The Station Blackout (SBO) rule (10 CFR 50.63) required the completion of plant specific analyses to determine appropriate actions to mitigate the effects of a total loss of AC power.

In the course of the SB0 reviews, most licensees either; (1) made a commitment to implement an EDG reliability program in accordance with NRC regulatory guidance but reserved the option to later adopt the outcome of Generic Issue B-56 resolution, or (2) stated that they had or will implement an equivalent program, or (3) endorsed the program embodied in NUHARC Initiative SA and NUKARC 87-00, Revision 1 (i.e.,

maintain the Emergency Diesel Generator (EDG) target reliability of 0.95 or 0.975)."

"The 6DG unavailability due to maintenance assumed in developing the SB0 rule yas 0.007; however, utilities were allowed to use plant specific EDG unavailability data in their plant specific analyses (Regulatory Guide 1.155).. Alternatively, EDG unavailability due to maintenance, as assumed in the plant specific IPE analysis, could be used as an EDG target unavailability."

"The maintenance rule (10 CFR 50.65(a)(3)) defines an objective of i

ensuring that preventing failures of structures, systems and components through maintenance is appropriately balanced against the unavailability nf SSCs due to maintenance activities. Consistent with licensee ~

commitments, EDG unavailability and reliability should be monitored against goals under 10 CFR 50.65(a)(1) or established as performance criteria in the plant's preventive maintenance program under 10 CFR 50.65(a)(2), taking into account the objective noted in 10 CFR 50.65(a)(3)."

"Under 550.65(a)(2) the utility would establish performance criteria for balancing EDG reliability and availability.

EDG performance criteria for reliability would be met by the absence of a maintenance preventable failure or the occurrence of a single maintenar.ce preventable failure followed by appropriate root cause determination and corrective action.

Performance criteria for unavailability would'be met by having fewer hours unavailable, on a rolling-one-year basis than that required by the established performance criteria."

"If any performance criterion is not met, or a second EDG maintenance preventable failure occurs, or if availability is unacceptable, the licensee should establish goals and monitor subsequent EDG reliability or availability under 10 CFR 50.65(a)(1). The utility would determine the appropriate balance between EDG reliability and unavailability and establish goals for each. The NRC would find it acceptable if the EDG reliability goals selected for compliance with 10 CFR 50.63 are monitored through the use of the triggers described in NUMARC's Initiative SA and the monitoring methods described in Appendix D of NUMARC-87-00, Revision 1, ' Guidelines and Technical Bases for NUHARC Initiatives Addressing Station Blackout at LWRs,' August 1991. An acceptable unavailability goal would be to have fewer hours unavailable j.

t

9

~

(on a rolling-one-year basis) than that established as acceptable by the licensee. Corrective action must be taken by the licensee if any goal is not met."

The following text will be added to the Regulatory Position section in the NRC Staff's regulatory guide for the maintenance rule.

"For purposes of monitoring EDG performance against EDG target reliability levels selected for compliance with 10 CFR 50.63, the NRC finds acceptable the use of the triggers described in NUMARC Initiative SA, and the guidelines outlined in Appendix D of NUMARC 87-00, Revision 1, August 1991, ' Guidelines and Technical Bases for NUMARC Initiatives Addressing Station Blackout at Light Water Reactors'."

I s

e l

e 4

file - coment.027 March 15, 1993

0. Rothberg, RES/EIB, 49-23924, NLS302 a

T.

G a

\\

o i

t i

a P

I 4

i l i

k,