ML23159A110

PR-050 - 61FR05318 - Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment

ML23159A110
Person / Time
Issue date:	02/12/1996
From:	Hoyle J NRC/SECY
To:
References
PR-050, 61FR05318
Download: ML23159A110 (1)

Text

ADAMS Template: SECY-067 DOCUMENT DATE: 02/12/1996 TITLE: PR-050 - 61FR05318 - REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT CASE

REFERENCE:

PR-050 61FR05318 KEYWORD: RULEMAKING COMMENTS Document Sensitivity: Non-sensitive - SUNSI Review Complete

STATUS OF RULEMAKING PROPOSED RULE: PR-050 OPEN ITEM (Y/N} N RULE NAME: REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT PROPOSED RULE FED REG CITE: 61FR05318 PROPOSED RULE PUBLICATION DATE: 02/12/96 NUMBER OF COMMENTS: 31 ORIGINAL DATE FOR COMMENTS: 06/11/96 EXTENSION DATE: I I FINAL RULE FED. REG. CITE: FINAL RULE PUBLICATION DATE: I I NOTES ON: REQUIRES OPERATORS OF NUCLEAR POWER PLANTS TO PROVIDE NRC WITH REL STATUS IABILITY & AVAILABILITY INFO. FOR SELECTED RISK-SIGNIFICANT EQUIP.

OF RULE: PR WITHDRAWN, SEE 63FR66497, 12/2/98.

HISTORY OF THE RULE PART AFFECTED: PR-050 RULE TITLE: REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT PROPOSED RULE PROPOSED RULE DATE PROPOSED RULE SECY PAPER: 95-215 SRM DATE: 12/19/95 SIGNED BY SECRETARY: 02/02/96 FINAL RULE FINAL RULE DATE FINAL RULE SECY PAPER: SRM DATE: I I SIGNED BY SECRETARY: I I STAFF CONTACTS ON THE RULE CONTACT!: DENNIS ALLISON, AEOD MAIL STOP: T4-A9 PHONE: 415-6835 CONTACT2: MAIL STOP: PHONE:

DOCKET NO. PR-050 (61FR05318)

In the Matter of REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT DATE DATE OF TITLE OR DOCKETED DOCUMENT DESCRIPTION OF DOCUMENT 02/02/96 FEDERAL REGISTER NOTICE - PROPOSED RULE 03/02/96 COMMENT OF MARVIN I. LEWIS ( 1) 03/19/96 03/12/96 COMMENT OF NUCLEAR ENERGY INSTITUTE (WILLIAM H. RASIN) ( 2) 04/01/96 03/27/96 COMMENT OF CLEAN WATER FUND OF NORTH CAROLINA (CARL RUPERT) ( 3) 04/09/96 03/14/96 COMMENT OF DUQUESNE LIGHT CO (SUSHIL C. JAIN) ( 4) 06/06/96 05/31/96 COMMENT OF JAMES A. PERRY ( 5) 06/10/96 06/06/96 COMMENT OF STEVEN E. FARKAS ( 6)

.6/11/96 06/06/96 COMMENT OF FLORIDA POWER & LIGHT COMPANY (W.H. BOHLKE) ( 7) 06/11/96 06/08/96 COMMENT OF DR. ZINOVY V. REYTBLATT ( 8) 06/11/96 06/10/96 COMMENT OF OHIO UTILITY RADIOLOGICAL SAFETY BOARD (STEVEN T. NOURSE) ( 9) 06/11/96 06/11/96 COMMENT OF NORTHERN STATES POWER CO (JOSEPH ALAN GONYEAU) ( 10) 06/11/96 06/11/96 COMMENT OF NUCLEAR ENERGY INSTITUTE (THOMAS E. TIPTON) ( 11) 06/12/96 06/11/96 COMMENT OF DETROIT EDISON (LYNNE GOODMAN) ( 12) 06/12/96 06/11/96 COMMENT OF NUCLEAR UTILITY BACKFITTING & REFORM GRP (DANIEL F. STENGER) ( 13) 06/12/96 06/11/96 COMMENT OF YANKEE ATOMIC ELECTRIC CO (JANE M. GRANT) ( 14) 06/13/96 06/11/96 COMMENT OF OMAHA PUBLIC POWER DISTRICT (T.L. PATTERSON) ( 15)

DOCKET NO. PR-050 (61FR05318)

DATE DATE OF TITLE OR DOCKETED DOCUMENT DESCRIPTION OF DOCUMENT 06/13/96 06/10/96 COMMENT OF DEVONRUE (STEPHEN MALONEY) ( 16) 06/13/96 06/09/96 COMMENT OF BOB CHRISTIE ( 17) 06/14/96 06/11/96 COMMENT OF TENNESSEE VALLEY AUTHORITY (RAUL R. BARON) ( 18) 06/14/96 06/11/96 COMMENT OF PECO ENERGY CO (G.A. HUNGER, JR.) ( 19) 06/17 /96 06/10/96 LETTER FROM SUSHIL C. JAIN, DUQUESNE LIGHT CO, SUPPLEMENTING 3/14/96 LETTER (COMMENT NO. 4) 06/17 /96 06/10/96 COMMENT OF DUKE POWER CO (M.S. TUCKMAN) ( 20) 06/17 /96 06/11/96 COMMENT OF ROCHESTER GAS AND ELECTRIC CORP (ROBERT C. MECREDY) ( 21) 06/18/96 06/11/96 COMMENT OF TU ELECTRIC (D.R. WOODLAN) ( 22) 06/21/96 06/11/96 COMMENT OF SOUTH CAROLINA ELECTRIC &GAS CO (GARY J. TAYLOR) ( 23) 06/24/96 06/17/96 COMMENT OF ARIZONA PUBLIC SERVICE CO (WILLIAM L. STEWART) ( 24) 06/24/96 06/20/96 COMMENT OF VIRGINIA POWER (M. L. BOWLING) ( 25)

.7/03/96 06/28/96 COMMENT OF BALTIMORE GAS AND ELECTRIC CO (CHARLES H. CRUSE) ( 26) 07/03/96 07/03/96 LETTER FROM JANE GRANT, YAEC, PROVIDING COMMENTS ON DRAFT REGULATORY GUIDE DG-1046 AND SUPPLEMENT-ING YAEC'S 6/11/96 LETTER (COMMENT NO. 14) 07/09/96 07/03/96 COMMENT OF GEORGIA POWER COMPANY (J. T. BECKHAM, JR.) ( 27) 07 /11/96 07/03/96 COMMENT OF ILLINOIS POWER COMPANY (MICHAEL W. LYON) ( 28) 07/12/96 07/10/96 COMMENT OF FLORIDA POWER CORPORATON (B. GUTHERMAN) ( 29) 07/12/96 07/05/96 COMMENT OF CAROLINA POWER & LIGHT COMPANY (T. D. WALT) ( 30) 07/16/96 07/03/96 COMMENT OF SOUTHERN NUCLEAR OPERATING COMPANY (DAVE MOREY) ( 31) 11/27 /98 11/24/98 FEDERAL REGISTER NOTICE - PROPOSED RULE:

WITHDRAWAL

DOCKET NO. PR-050 (61FR05318)

DATE DATE OF TITLE OR DOCKETED DOCUMENT DESCRIPTION OF DOCUMENT 01/11/99 12/10/98 FEDERAL REGISTER NOTICE - PROPOSED RULE:

WITHDRAWAL; CORRECTION

' DOCKET NUMBER PROPOSED RULE PR 5O

( <,I Ff< 5E! s'} 7590-01-P DOCKETE D US NRC "99 JAN 11 P3 :as Nuclear Regulatory Commission OFhr_,

10 CFR Part 50 RLL. ~=

ADJUJ RIN 3150-AF33 Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment; Correction AGENCY: Nuclear Regulatory Commission.

ACTION: Notice of proposed rulemaking: Withdrawal; Correction.

SUMMARY

This document corrects a notice appearing in the Federal Register on December 2, 1998 (63 FR 66497), that withdraws a notice of proposed rulemaking that requested public comments on proposed amendments to its regulations that would have required licensees for commercial nuclear power reactors to report to the NRC, plant-specific summary reliability and availability data for certain risk-significant systems and equipment. This action is necessary to correct an erroneous telephone number.

FOR FURTHER INFORMATION CONTACT: David L. Meyer, Chief, Rules and Directives Branch, Division of Administrative Services, Office of Administration, telephone (301) 415-7162.

/µ/;. m 1a./16/1&'

J' t,~FRb'10:/.b

SUPPLEMENTARY INFORMATION:

On page 66498, in the first column, in the third line from the top, the telephone number, "(202) 512-2249" is corrected to read "(202) 634-3273."

Dated at Rockville, Maryland, this /Q~ day of December 1998.

For the Nuclear Regulatory Commission.

David L. Meyer, Chief Rules and Directives Branch Division of Administrative Services Office of Administration

DOC KETED L[?5f0-{)1-P]

NUCLEAR REGULATORY COMMISSION

• 9s NOV 27 P2 :40 10 CFR Part 50 RIN 3150-AF33 Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment AGENCY: Nuclear Regulatory Commission. So

(/p/FR53/8 ACTION: Proposed rule: Withdrawal.

SUMMARY

The Nuclear Regulatory Commission (NRC) is withdrawing a notice of proposed rulemaking that solicited comments on proposed amendments to its regulations that would have required licensees for commercial nuclear power reactors to report to the NRC, plant-specific summary reliability and availability data for certain risk-significant systems and equipment. The proposed rule would have also required licensees to maintain onsite, and to make available for NRC inspection, records and documentation that provide the basis for the

• summary data reported to the NRC. The systems and equipment for which data would be provided are a subset of the systems and equipment within the scope of the NRC's maintenance rule. The Commission has decided to accept industry's proposed alternative to the rule to voluntarily provide reliability and availability information for risk-significant systems and equipment and, therefore, withdraws this rulemaking.

ADDRESSES: The Commission paper, the staff requirement memoranda (SRM), and associated documents are available for public inspection, and copying for a fee, at the NRC Public Document Room located at 2120 L Street NW . (Lower Level), Washington, DC 20012-7082, telephone: (202) 512-2249.

pp-/;. 9Yl I :;i,/{),/(jg o.f lt,3Ff<fpi,'-/91

FOR FURTHER INFORMATION CONTACT: Dennis Allison, Office for Analysis and Evaluation of Operational Data, U.S. Nuclear Regulatory Commission, Washington, DC 20555-0001, telephone (301) 415-6835, e-mail dpa@nrc.gov.

SUPPLEMENTARY INFORMATION:

On February 12, 1996 (61 FR 5318), the NRG published in the Federal Register proposed amendments to 10 CFR Part 50 that would have required operating reactor licensees to report reliability and availability information for certain risk-significant systems and equipment. The reporting requirements would have applied to the event-mitigating systems and equipment that have or could have a significant effect on risk in terms of avoiding core damage accidents or preserving containment integrity. The data that would have been reported would have included: the number of demands and the number of failures to start associated with those demands, along with additional descriptive information; the number of hours of operation following each successful start including whether or not the run was terminated by equipment failure, along with additional descriptive information; the number of hours equipment is unavailable, along with additional descriptive information; for each period equipment is unavailable due to component failure, descriptive information on that failure; and the number of hours when two or more trains from the same or different systems were concurrently unavailable, along with additional descriptive information.

The public comment period closed on June 11, 1996. The NRG received 31 comment letters. One comment letter supported the rule, stating that the public and industry could expect significant benefits. Most of the remaining comments opposed the rule, stating that the 2

proposed reporting requirements costs were underestimated, benefits were overestimated, the rule would be overly burdensome, the rule would be premature, and that the rule is not justified.

The Commission SAM dated June 28, 1995, issued in response to SECY-95-129, and the SAM on SECY-95-215 dated October 24, 1995, directed the NRC staff to continue to work with industry on voluntary submittal of reliability data under a program that will meet the needs of all parties. On October 1, 1996, the Institute of Nuclear Power Operations (INPO) provided the NRC with a sample of data available from its Safety System Performance Indicator (SSPI) system, as part of a voluntary nuclear industry data sharing initiative. A revised Memorandum of Agreement (MOA) between INPO and the NRC was signed on December 24, 1996, providing NRC with access to SSPI data. In addition, on March 21, 1997, the Nuclear Energy Institute (NEI) provided the NRC with a description of a new INPO data collection system, Equipment Performance and Information Exchange (EPIX). Based upon a review of data available in SSPI and EPIX, as well as the information available from Licensee Event Reports and Monthly Operating Reports, the Commission has determined that under the voluntary approach, the NRC can estimate risk parameters and construct a reliability database that reflects the parameters needed for effective use in risk-informed applications. Thus, the intended benefits of the proposed rule would be realized and the main advantages of the voluntary approach (i.e., the lower cost, schedule, and industry support) outweigh any disadvantages. The NRC will continue to work with industry representatives to improve the 3

content of the voluntary data. Because of industry's voluntary alternative approach to the rule, the Commission is withdrawing this proposed rulemaking.

Dated at Rockville, Maryland, this ~ day of November, 1998.

For the Nuclear Regulatory Commission.

ary of the Commission.

4

i Southern Nuclear Operating Company Post Office Box 1295 Birmingham, Alabama 35201 Telephone (205) 868-5131 DOCK ETE D DOCKETED Dave Morey US ,'R

'96 JUL 16 PlZ :1 3 US RC Southern Nu~~a~pwgr~~g f ompany

.\

Vice President Farl ey Project the southern electnc system Docket Nos. 50-348 50-364 DOCKET NLMBER PB 5 0 PROPOSED RULE -

U. S. Nuclear Regulatory Commission (tol FR 53\ <o)

ATTN: Document Control Desk Washington, DC 20555 @

Comments on the Proposed Rule "Reporting Reliability and Availability Information for Risk-significant Systems and Equipment" (Federal Register Vol. 61, No. 29, dated February 12, 1996)

Ladies and Gentlemen:

Southern Nuclear Operating Company (SNC) has reviewed the proposed rule "Reporting Reliability and Availability Information for Risk-significant Systems and Equipment,"

published in the Federal Register on February 12, 1996. SNC is in agreement with the NEI comments which are to be provided to the NRC.

Should you have any questions, please advise.

Respectfully submitted,

/: Ll ) }l,ht'.,;("'

V  !/

Dave Morey DNM/AAF

"'JUL 1 9 1996

-'\cknowledged by cara ........,uauu,tt .....,t*..-v..

.J.S. NUCLEAR REGULATORY COMMISSIO~

DOCKETING & SERVICE SECTION OFFICE Of THE SECRET ARY OF THE COMMISSION Ooalnlnt Sta!ilb Polbnlfk Date 1fa /j l.o Coples Received_-'-/_'_'_ _ _ __

.\dd'I Coples Reproduced --¥J____

5f)l0lm OiSibJIIOn RIDSI PI>B , 1/Ju>iaa

U. S. Nuclear Regulatory Commission Page Two cc: Southern Nuclear Operating Company R. D. Hill, FNP Plant Manager U.S. Nuclear Regulatory Commission. Washington. DC B. L. Siegel, NRR Senior Project Manager U. S. Nuclear Regulatory Commission. Region II S. D. Ebneter, Region II Administrator T. M. Ross, FNP Senior Resident Inspector

DOCKETED CP&L US RC Carolina Power & Light Company *96 JUL 12 P3 :13 PO Box 1551 411 Fayetteville Street Mall Raleigh NC 27 602 OFFIC E OF SECRETAR Y DOCK ETI NG & SER' I C E

) t, .,,,..,

July 5, 1996 Serial um * : PE&RAS96-042 DOCKET Nu.tBER PR r .,...,

Document and Service Branch PROPOSED RULE--=-~~ L/- -

U. S. Nuclear Regulatory Commission ( 6 / F~ 5 3li-)

Washington, DC 20555-0001

Subject:

Comments on Proposed Rule 10 CFR 50.76, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment" and Draft Regulatory Guide DG-1046, "Guidelines for Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment in Nuclear Power Plants" Carolina Power & Light Company (CP&L) is pleased to provide the Nuclear Regulatory Commission (NRC) with the attached comments regarding the subject rule and its associated draft regulatory guide. In summary, CP&L believes that the objectives of the proposed rule can be achieved within the framework of existing voluntary data collection efforts.

If you have any questions, please contact me at (919) 546-6901.

Sincerely, T. D. Walt Manager, Performance Evaluation and Regulatory Affairs TDW/mlm Attachments c: Mr. Thomas E. Tipton, Nuclear Energy Institute

'JUI. 1l 1995 Acknowteaged by cara _..,,.,.,,tlttttMll/1441 ~ -

?/.$. NUCLEJ,Ji rfr ,.;U111TORY COMMISSIO~

OOCKEl 1NG & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSIOO Doa.ment Statistics

'Postmark Date 7 /{-I?£ Copies Received _ __,/"--- - - -

.\dd'I Copies Reproduced-~- ----,,,...- -

Speclal Dislributiofl Yl-:J; 05;, fJI}~

A L) l S O V'v

Attachment 1 Carolina Power & Light Company Comments on Proposed Rule 10 CPR 50.76, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment" (as published in 61 FR 5318-2325, February 12, 1996)

Page 5320, Move to Risk-Based Regulation In general, this section argues that the proposed rule would enhance and improve the Nuclear Regulatory Commission's (NRC's) "oversight capability with respect to public health and safety

- i.e., the ability to maintain or enhance safety by identifying and reviewing indications of increased risk and, if appropriate, taking generic or plant-specific action." Carolina Power &

Light Company (CP&L) agrees that the use of such data can indeed enhance the oversight of licensee activities. We disagree, however, that reporting this data routinely will accomplish this end in the most mutually beneficial way, both for the NRC and for the industry.

The argument is put forth that, "In order to move towards risk-based regulation and the increased use of PRA information, the NRC needs scrutable, plant-specific and generic reliability and availability information." CP &L's general response to this argument is that the industry voluntarily collects and reports data of this nature. With implementation of the Maintenance Rule, additional data is now being collected and trended. CP&L supports the NRC's desired goal but questions whether the process implemented by the proposed rule is the most efficient manner in which to achieve this goal. Further reporting of data duplicates efforts already in place in the industry that, properly utilized or with minor changes, would easily and more readily achieve the NRC's goals.

- Page 5321, Accident Sequence Precursor (ASP) and Event Analysis This section states that, "Plant-specific data are needed to better understand an event and calculate the associated conditional core damage probability." CP&L believes that a more efficient means of accomplishing this goal would be to address such impacts via existing reporting requirements, i.e., LER's.

Page 5321, Aging This section states in the second column:

It is anticipated that a significant number of additional requests will be received that rely upon risk-based arguments. These changes could adversely affect the level of safety achieved by plants if the risk evaluations are flawed or the changes improperly executed or the changes involve synergistic effects that are not covered by the risk models or captured by historical data.

, Proposed Rule Comments, Page 2 of 3 CP&L does not believe this to be a compelling reason for additional reporting of data. If the NRC wishes to validate or verify the inputs to such requests, requests for the supporting data should be specific to the request.

Page 5321, Inservice Testing This section states, "For example, less frequent valve testing might lead to an increase in the demand failure rate because the valve actuating mechanism tends to bind or freeze after extended periods of idleness." While this may be the case for this specific example, in general, one must acknowledge that surveillance testing is generally deleterious to plant equipment reliability and has been accepted for lack of better means to establish equipment operability. Again, the data requested by the NRC should be specific to the request being made.

Pages 5321-5322, NRC Maintenance Rule This section states:

The reliability and availability information that would be required by the proposed reporting rule would improve the NRC's oversight of licensee's implementation of the maintenance rule. It would also enhance licensee's capabilities to implement the evaluation and goal-setting activities required by the maintenance rule by providing licensees with access to current industry-wide reliability and availability information for some of the systems and equipment within the scope of the maintenance rule.

CP&L believes that the current trending conducted in compliance with the Maintenance Rule is sufficient to allow the NRC to oversee licensee implementation. We further believe that current voluntary data collection and reporting is adequate to support implementation and goal-setting activities within the framework of the Maintenance Rule.

It is further argued that "improving NRC's oversight of the maintenance rule could strengthen one of the bases for the scope of the license renewal rule." While it is clear that an effective maintenance program will enhance the prospects of successful license renewal activities, and an effective Maintenance Rule program can strengthen such a maintenance program, it does not necessarily follow that improved NRC capability to oversee Maintenance Rule compliance will strengthen license renewal prospects.

Page 5322, Licensee Implementation In the second column on this page it states that, " ... utilities do not use industry operating experience in a systematic and consistent way for goal setting purposes under the maintenance rule." There are a number of ways in which utilities do not effectively use operating experience data and a number of reasons why not. It does not follow, however that if the NRC has large volumes of additional operating experience data, that utilities will then become more effective in its use.

, Proposed Rule Comments, Page 3 of 3 In fact, it would be more efficient and germane to the overall goal of public health and safety if the NRC would share with the industry the bases for these conclusions in a manner that facilitates more effective use of operating experience data. One of the reasons data may not be effectively used is the labor required to assimilate, analyze and apply it. Adding another layer of collection and reporting requirements will only aggravate this situation. Typically, even for a few key components, initiating events and test and maintenance unavailabilities, PSA models are relatively insensitive to failure data. Requiring extensive data collection on multiple components has a very low benefit to cost ratio.

It is further stated that, "Prudent on-line maintenance decisions depend on a full appreciation of the risk-significance taking equipment out of service (individually or collectively) and use of plant-specific and generic reliability and availability data would play a significant role in improving such decision making." Again, it does not follow that reporting this data to the NRC will enhance this decision making ipso facto because it is available. As with the availability of additional operating experience data, it is only useful if the utility considers and uses it.

Page 5323, Description of Proposed Rule It is stated that, "The hours of operation following successful starts are needed to estimate the probability the equipment will function for a specified period of time." If in the case of emergency diesel generators, for example, reliability data from monthly surveillance tests that last one to two hours could be used to extrapolate the probability of successful functioning over a seven day period, then there would be no valid reason to perform 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> refueling interval runs.

Attachment 2 Carolina Power & Light Company Comments on Draft Regulatory Guide DG-1046, "Guidelines for Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment in Nuclear Power Plants" Section 1.2, Other Reportable Systems, page 11 The second bullet states that, "Systems and equipment that are risk-significant only because of operator error for actuation, operation, or termination of a risk-significant safety function need not be considered reportable ... However, this exclusion would not apply to operator or other personnel errors that could unintentionally make equipment inoperable (unavailable) with respect to its risk-significant safety function." While a desire to attain some understanding of actual system and equipment unavailability may be the issue of concern here, reading the literal sense of the paragraph, its interpretation remains unclear. The issue of potential misalignment of motor operated valves debated in relation to Generic Letter 89-10, "Safety-Related Motor-Operated Valve Testing And Surveillance" and its supplements comes to mind. The NRC staff tended not to grant such leeway with regard to operator error in reviewing licensee compliance with this generic letter. Thus, the consistent application of this paragraph is of concern to CP&L.

Section 1.3, Boundaries of Systems, Trains, and Equipment Groups, page 12 This section suggests that licensees should submit diagrams for all reportable systems, trains, and equipment groups that indicate the principal active and passive components. Diagrams of this type are already submitted as part of the licensee's Final Safety Analysis Report. With sufficiently detailed logs, the diagrams previously submitted should be sufficient. In fact, Appendix E of DG-1046 refers the reader to these diagrams as published in the NRC's Plant Information Book and available on the World Wide Web.

Section 5, Reportable Failures, page 16 This section defines reportable failures as " ... reportable when a reportable system, train, or equipment group fails to perform its risk-significant safety function in response to a reportable demand," or " ... when discovered by means other than a reportable demand (e.g., it is found to be unable to perform its risk-significant safety function because of actual or incipient failure by inspection or other nonreportable demands." Failures such as those described in this section are already captured as part of compliance with the Maintenance Rule, and corrective action is required by the Rule and achieved by compliance with it. The impact of such failures on reliability and availability is already trended to achieve this compliance, and the industry voluntarily collects these data for its own use.

i

• Attachment 2, Draft Regulatory Guide Comments, Page 2 of 2 Section 10, Onsite Data Storage, page 23 This section states that, "Alternatively, licensees may wish to state their methods or references for linking each reportable data element (demand, failure, or unavailable period) to existing plant records." It further stipulates that, "These records should cover a period of at least the 5 most recent calendar years."

Does the Regulatory Guide intend that licensees reconstruct this data from records dating back over the five years preceding the effective date of the Rule? If so, such a requirement would be exceedingly cumbersome and costly. Furthermore, allowing licensees to tie summary reports to existing plant reporting merely confirms the industry's assertion that the data the NRC wishes to accumulate is already available to the NRC via current voluntary collection efforts. Minimal benefit will be derived by duplication of this reporting.

, DOCKETED DOCKEi NlMBER us~rn PROPOSED RULE PR 5 !J (ti PY2..G3LJj

'96 JUL 12 A9 :17 Florida Power CORPORATION Cryllal River Unit 3 June 10, 1996 NL96-0079 U. S. Nuclear Regulatory Commission Attn: Docketing and Services Branch Washington, D. C. 20555-0001

Subject:

Comments on Proposed Rulemaking for 10 CFR 50.76, "Reporting Reliability and Availability Information for Risk-significant Systems and Equipment"

Reference:

Federal Register, 61FR5318, Monday, February 12, 1996

Dear Sir:

This letter submits Florida Power Corporation's (FPC) comments regarding the proposed rule for 10 CFR 50.76, "Reporting Reliability and Availability Information for Risk-significant Systems and Equipment".

This proposed rule would require holders of operating licenses for nuclear power reactors to report reliability and ava ilability data for certain risk-significant systems and equipment. The proposed reporting requirements would apply to the event-mitigating systems and equipment that hav~ .or could have a significant effect on risk in terms of avoiding core damage accidents or preserving containment integrity.

If this proposed rule is promulgated, changes to procedures and training will be required. Additionally, plant modifications and organizational changes may also be necessary. Therefore, the position of the NRC as stated in the*proposed rule t hat the requirements of the Backfit Rule, 10 CFR 50 .1 09 , are not applicable is inv alid . A backfit analysis is required as a prerequisite to the NRC implementing these new requirements. Promulgation of this rule appears premature and would result in unnecessary utilization of utility resources.

ti~~

, I

~lldbycard ...

-- I

• ~~~-

CRYSTAL RIVER ENERGY COMPLEX: 15760 W Power Line St

• Crystal River, Florida 34428-6706 * (352)795-6486 A Florida Progress Company

1J.S. NUr-i l:H" AEGU! ..l.T,Jt:W COMMISSIOt,,

00(,. .~i & Sr.H' 1 *:E SECTION O: * ,; OF THE: SECRETARY c:* THE COMMISSION C*c- -* "'Tlent Statistics Jlostmark Da~ ~ "'{;.1'1-rei>--" ~ -,G--

Coples Rci/4!;1,ec! '"'-"--i L I

.\dd'I Copies Reprocuced _ .___ _ _.....

Special Distri!xJtionU/,)~ P{JfL./ All 1.S,0 1L,

U. S. Nuclear Regulatory Commission NL96-0079 Page 2 Specifically, Crystal River Unit 3 (CR-3) opposes this proposed rulemaking for the following reasons:

• Redundant Information. The purpose of the proposed rule is to require licensees to report reliability and availability information for risk-significant systems including:

1. the number of demands, the number of failures with such demands, and the dates of any such failures.

2. the number of hours of operation following each successful start.

3. the number of hours equipment is unavailable.

4. a failure record for each period equipment is unavailable due to component failure(s).

5. the number of hours when two or more trains from the same or different systems were concurrently unavailable.

The CR-3 risk-significant systems encompassed by the proposed rule include Emergency Feedwater, Emergency Power, Reactor Protection, High Pressure Injection, and Low Pressure Injection. However, the majority of required information for these systems is already being compiled by other programs.

The numbers of estimated demands, failures per demand, and hours of operation are submitted quarterly to the Institute of Nuclear Power Operations (INPO) for the Safety System Performance Indicator (SSPI) program. Except for Reactor Protection, the number of unavailable hours for the required risk-significant systems is compiled by the SSPI program and the Maintenance Rule. Diesel demand information is captured in a procedure as part of the Station Blackout Coping Assessment. The Nuclear Plant Reliability Data System (NPRDS), that includes the required risk-significant systems, provides more than adequate component failure data. Therefore, the only information requirement that is not already available is the number of hours two trains are concurrently unavailable.

Since the majority of this required information is immediately available through INPO, creation of a new rule for this purpose appears superfluous. The apparent solution is to evaluate the worth of missing data items, and include them in the INPO program.

• Compilation Burden. The proposed rule requires compilation of the number of actual demands while the SSPI program requires the number of estimated demands. This is a major implementation difference.

Currently, approximately two weeks per quarter are required to compile the SSPI program information. The Shift Supervisor Logs and Replacement Emergency Dose Assessment System (REDAS) are reviewed to determine the numbers of demands and hours of operation. Although a thorough method of compiling information, this tedious process is subject to errors. Therefore, if an accidental omission of a system demand will result in regulatory enforcement actions, an improved methodology for demand counting will be required. Several possible methodologies include collection of the information via procedures,

U. S. Nuclear Regulatory Commission NL96-0079 Page 3 performance of an independent review, or implementation of demand counters.

Since makeup and emergency feedwater pump starts are not documented as well as diesel starts, implementing a procedure to log demands may not provide the solution in determining the actual number of demands. Performance of an independent review may result in more accurate information, but is not consistent with past efforts to improve business processes. Demand counters, unless they fail, will produce an accurate number of demands. However, they will be subject to initial installation and continual maintenance costs.

Also, no methodology is currently available at CR-3 to determine the number of hours when two or more trains are concurrently unavailable. This information is desired to monitor licensees' risk management during on-line maintenance activities. A methodology could be implemented to compile this information from the equipment clearance log. However, caution must be exercised when using the clearance log since the data is not verified and may not always be accurate.

Again, if regulatory enforcement actions for incorrect information result from the promulgation of this proposed rule, it would become necessary to implement changes to the plant which certainly meets the definition of backfit under 10 CFR 50.109.

• Information Purpose. Apparently, there is a misconception that utilities desire to perform comparisons of reliability and availability information against other utilities. Even pl ants designed by the same Nuclear Steam Supply System (NSSS} vendor are not identical in construction or operation. However, similar comparisons have been performed by INPO and the NSSS vendors to identify outliers. These comparisons used SSPI program information due to the level of detail required.

One primary reason for the promulgation of this rule is the proposed risk based rulemaking. However, it would appear premature to require the industry to report this particular information for risk based rulemaking since the specifics have not been identified and the data requirements are not known.

Supposedly, reporting this information would enhance the capability of utilities to implement the evaluation and goal setting activities and imp rove the NRC' s oversight of the Maintenance Rule.

Additionally, this improved oversight could assist with the license renewal rule. However, the Maintenance Rule requires utilization of operating experience in goal setting not industry availability or reliability information, which, as pointed out above, would be of little value. Moreover, if this type of information is desired, it is available through the SSPI program.

Also, the majority of the information requirements for License Renewal have been included in studies of passive long lived components, or in age and time related degradation. The information

U. S. Nuclear Regulatory Commission NL96-0079 Page 4 required in the proposed rule applies to active components.

Utilities involved in the license renewal process have already taken credit for active components being covered by the Maintenance Rule or the SSPI program. Nothing in the proposed rule specifically addresses license renewal issues, or adds anything to the Maintenance Rule or the license renewal process that does not already exist.

In conclusion, the SSPI program information can be made available for immediate NRC use. Additionally, NPRDS and the SSPI program can be modified if additional information is required. Therefore, promulgation of this proposed rule would appear premature and unnecessary.

e:;~-__,

B. G~x:; ;~n Manager Nuclear Licensing BG/JLO xc: Records Management

Illinois Power Company DOC KE TED Clinton Power Station P.O. Box 678 US J C Clinton, IL 61727 Tel 217 935-8881 ILLIN915 U-602607 PGWER '96 JUL 11 P4 :24 Ll0-96(07-03)-LP lA.120 OFFI CE OF SE CRE T, ~y DOCK ETI NG &; ': --11* J E BRA~CH u Y 3 , 1996 DOCKET NlNBER 1111 Docket No. 50-461

~noPOSED RULE rn 5 0 (6/f/2- f;31~)

Chief, Rules Review and Directives Branch Division of Freedom oflnformation and Publication Services U.S. Nuclear Regulatory Commission Washington, D .C. 20555-0001

Subject:

Illinois Power's (IP's) Comments on the Draft Regulatory Guide DG-1046, Guidelines for Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment in Nuclear Power Plants"

Dear Sir:

Draft Regulatory Guide DG-1046 was issued in April 1996 in its draft form to encourage public involvement in the early stages of this regulatory development. Illinois Power appreciates the NRC's encouragement of public involvement in the rulemaking process and values the opportunity to comment on this draft regulatory guide. Listed below are IP's comments regarding Draft Regulatory Guide DG-1046, "Guidelines for Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment in Nuclear Power Plants."

IP has determined that implementation of the proposed rule would be manpower intensive. With the manhours expended implementing the Maintenance Rule to date, the reliability and availability of risk-significant systems will be achieved. In addition, the majority of this information will be maintained on site, per Maintenance Rule requirements, and will be available for NRC's inspection at any time.

IP questions whether the amount of review time the NRC will charge back to the utilities reviewing the information from over 100 plants will increase safety commensurate with the expense. The Maintenance Rule, which goes into effect July 10, 1996, will already ensure the reliability of safety significant systems. IP believes that the Backfit Rule as defined in 10 CFR 50. 109 should have been applied, vice a regulatory analysis, as this will require a modification of procedures in addition to those changed due to the Maintenance Rule.

JUL 1 6 1996..

~wiec:lgeci by cardN ....,,.,, .... , , .... ....... 14;

-8;..~ NU*~,_.- . '..*,. : .. L,;TORY COMMISSIOt-v DOCk : . *;*,:i & .:ERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Dorunent Statistics

~ stma k Date 7 l5/ ? £ .

Cop;es Received t e....op y 1-Yl;"t -t-. PJ./3

.\dd't Copies Reproduced t:, r-i;p ~ t"o (lfl 08 Special Distribution ,3

__ztt__P1; << .roavA JJ tJ{J.,,

IP has currently identified twenty-four "Risk Significant Systems" in preparation for the Maintenance Rule implementation. Since the NRC has identified seven "Risk Significant" safety functions, it would appear that this is a disagreement or contradiction between the Maintenance Rule and this proposed Reliability Rule. In addition, it would appear as if the unavailability definition contained within the Reliability Rule is more prescriptive than that defined in NUMARC 93-01, Revision 2, "Industry Guideline for Monitoring the Effectiveness of Maintenance at Nuclear Power Plants." This could be subject to misinterpretation and could result in notices of violation that are not equally applied in all cases.

In conclusion, IP believes that this is unnecessary rulemaking. It does not add significant safety benefit for the resources expended. The subject systems are well

• controlled by Technical Specification out-of-service times .

Sincerely yours, 1* /_J ,,~/;: /}

~Y:il/L-Michaet W. Lyon Director-Licensing JSP/krk cc:

Georgia Power Company 40 Inverness Center Parkway Post Office Box 1295 Birmingham, Alabama 35201 Telephone 205 877-7279 Georgia Power

,\

J. T. Beckham, Jr.

Vice President - Nuclear *96 JUL -9 A11 :52 the southem electric system Hatch Project Docket Nos. 50-321 50-424 HL-5197 50-366 50-425 LCV-0834 DOCKET NUMBER PR ()

PR POSED RULE.:.:;......

~

U. S. Nuclear Regulatory Commission 6 I F Y2. 5 J 1 fr)

ATTN: Document Control Desk Washington, D. C. 20555 Comments on the Proposed Rule "Reporting Reliability and Availability Information for Risk-significant Systems and Equipment" (Federal Register Vol. 61, No. 29. dated February 12, 1996)

Gentlemen:

Georgia Power Company (GPC) has reviewed the proposed rule "Reporting Reliability and Availability Information for Risk-significant Systems and Equipment," published in the Federal Register on February 12, 1996. GPC is in agreement with the NEI comments which are to be provided to the NRC.

Should you have any questions, please advise.

Sincerely, AAF/af i -. Beckham, Jr.

JUL f 6' 1996__

~wfeoged by cera.. ,,"' "'* ...

1. . CLE.1..;-; R.::, -~*; ~.;GflY C0~.JtMISStON OO~r\Ern 1G ; ,. }-NICE SECTION o;:n:E CJF FIE ~ECRETARY 0t Tf-JE COMMISSION Document Statistics Postmark Date 7 h /16 CcresR~C#JI\Jed _ __,/_ _ _ __

.\dd'I Cooies Fle~roducad 3 SpeciJI Dlstnt--Jtion Yl.1:. ~ Y.J.-!5--:...,-:,,.--:;;;--

<:zaos a 7' OJ I 1 -~O >\-

Georgia Power <<\

U.S. Nuclear Regulatory Commission Page Two July 3, 1996 cc: Georgia Power Company C. K. McCoy, Vice President Nuclear - Vogtle Project J.B. Beasley, Nuclear Plant General Manager - Vogtle H. L. Sumner, Jr., Nuclear Plant General Manager - Hatch U.S. Nuclear Regulatory Commission, Washington, DC K. N. Jabbour, Licensing Project Manager - Hatch L. L. Wheeler, Licensing Project Manager, Vogtle U. S. Nuclear Regulatory Commission. Region II S. D. Ebneter, Regional Administrator B. L. Holbrook, Senior Resident Inspector - Hatch C. R. Ogle, Senior Resident Inspector - Vogtle HL-5197 LCV-0834

Telephone (508) 779-6711 YANKEE ATOMIC ELECTRIC COMPANY TWX 710-380-7619 OOCt{ETEO USN C

~~-~ 580 Main Street, Bolton, Massachusetts 01740-1398

\..;(ANKE!,/

---

* July 3, 1996

'96 JUL -s A11 :54 FYC 96-005 Pft U.S. Nuclear Regulatory Commission Washington, D.C. 20555-0001 Attention: Chief, Rules Review and Directories Branch DOCKET NU-'BER PROPOSED RULE l6' Ffl r 3 t ';-:)

5' O su.,~ p I -e.,,~-r ta Fl= I'-/

Division of Freedom oflnformation and Publication Services

Subject:

Comments on U.S. NRC Draft Regulatory Guide DG-1046, "Guidelines For Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment" (Task DG-1046, dated April 26, 1996)

References:

1) USNRC Proposed Rule: Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment (61 FR 5318, dated February 12, 1996)

2) Letter, Yankee to USNRC, FYC 96-002, dated June 11, 1996

Dear Sir or Madam:

Yankee Atomic Electric Company (Yankee) appreciates the opportunity to provide comments in response to the subject Draft Regulatory Guide DG-1046. Yankee is the owner of the Yankee Nuclear Power Station in Rowe, Massachusetts and provides engineering and licensing services to the nuclear power plants in the New England Region. Yankee is also submitting these comments on behalf of Northeast Utilities.

This Draft Regulatory Guide is associated with the NRC's Proposed Rule published in the Federal Register, dated February 12, 1996 [Reference (l)] that, if promulgated, will require licensees of commercial m.1c!ear power reactors report plant-Spi;;cific summaries of reliability and availability data for selected risk-significant systems and equipment to NRC.

By letter dated June 1 f, 1996 [Reference (2)], Yankee submitted a letter in ~esponse to the comment opportunity associated with the USNRC' s Proposed Rule on this subject and stated our intent to submit comments on the Draft Regulatory Guide. Accordingly, we are providing, as an attachment to this letter, our specific comments on the Draft Regulatory Guide DG-1046, "Guidelines for Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment in Nuclear Power Plants.,,

In our comment letter on the Proposed Rule, we stated our support for moving forward with enhancing nuclear safety and reliability through risk-based and performance-based regulation.

klvlowftagldbycard JUL 1 6. 199~

NUCU:AR 9[8ULATORY COMMISSl01'1 DOCrC.f i:~*:3 &. SERVICE SECTION OFFICE OF THE SECRETARY Of THE COMMISSION Document Statistics p~c.L Postmark Date "'111 l "l 6 71 3 /'!-£ Copies Received_ __ _ _ _ __

.\dd'I Copies Reproduced ,,. . 3

:..----...-~ -

S~~ I Distribution A/ Jz G¢n, POf?.,

_ rtIOS ' '

However, we also stated that the associated data needs should be developed and fulfilled as part of the NRC's PRA Pilot Implementation Plan and other risk and performance-based regulatory initiatives. Accordingly, we opposed promulgation of the Proposed Rule on the basis of several reasons, including that: I) the requirements were never intended as necessary by NRC for implementation and oversight of current regulatory requirements; 2) no bacldit analysis was conducted; 3) the proposed rule is premature as there are no current regulatory requirements for licensees to conduct PRA's; and 4) the proposed requirements impose an excessive and unreasonable burden on licensees.

The subject Draft Regulatory Guide is intended to provide guidance to licensees on methods that would be acceptable to the NRC staff for implementing the NRC Proposed Rule. The Draft Regulatory Guide identifies the summary data to be reported to the NRC and the basic data to be maintained onsite and available for NRC inspection. The Guide states the intent to provide " ... a flexible framework to obtain consistent data yet allow licensees to use existing data collection programs to the extent they are applicable." As a general comment we do not support the issuance of the Draft Regulatory Guide for the reasons we oppose promulgation of the Proposed Rule. Our specific comments focusing on the monitoring and reportability burdens imposed by the Regulatory Guide are provided in the attachment to this letter.

Yankee appreciates the opportunity to provide comments on this Draft Regulatory Guide.

Sincerely, YANKEE ATOMIC ELECTRIC COMPANY M. Grant ger, Regulatory and Industry Affairs Attachment

.c: USNitC Docume,1t Control Desk M. Fairtile, NRC, NRR J. White, NRC, Region I

- - _ _ _J Attachment Letter, YAEC to U.S. Nuclear Regulatory Commission July 3, 1996 Pagel ATTACHMENT COMMENTS ON DRAFf REGULATORY GUIDE DG-1046 "GUIDELINES FOR REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT IN NUCLEAR POWER PLANTS" Section 1.2 Other Reportable Systems The process described in this Section is identical to that used by New England nuclear units for determining risk significance under the requirements of 10 CFR 50.65 and their programs implementing the Maintenance Rule. The need to perform an essentially duplicative process is unnecessary. The Regulatory Guide should reference the reportability requirements associated with the Maintenance Rule and should not require the reporting of any risk-significant systems beyond those associated with Maintenance Rule compliance.

Section 1.3 Boundaries of Systems, Trains, and Ewiipment Groups This Section requests licensees to " ... submit diagrams for all reportable systems, trains, and equipment groups that indicate principal active and passive components." Preparation and submittal of such diagrams are an additional and unnecessary burdens on licensees and are not related to a clear regulatory application. The Regulatory Guide should be limited to existing documentation that is already available for NRC inspection and use. For example, principal active and passive components and equipment groups can be determined by NRC from existing plant surveillance procedure information.

Further, it is not clear what level of"safety function" is relevant for NRC's equipment grouping in Section 1.3 with respect to grouping by testing. The language could be interpreted as relating to the performance of a single portion of a surveillance or the entire surveillance. For example, the Engineered Safeguards Actuation Systems provide a number of functions which include actuation of various plant equipment depending on plant conditions. The equipment groupings could be comprised of the specific equipment that provides each singular function or the group could consists of all the actuated equipment. The Regulatory Guide should provide specific unambiguous guidance on what level of safety function is relevant for equipment groupings.

Attachment Letter, YAEC to U.S. Nuclear Regulatory Commission July 3, 1996 Page2 Section 2 Reportable Plant Operational States Commercial nuclear power plants are operated in modes as specified in Technical Specifications. This Section applies the concept of"plant operational states." The establishment of new and different reportable plant operational states that are not consistent with existing and established plant mode definitions introduces additional operational and reporting burdens on operations personnel. These new "plant operational states" present an additional set of regulatory enforcement conditions on operations personnel that may adversely impact plant safety. The Regulatory Guide provides no analysis justifying the basis or benefit for imposing such an additional operational burden, nor does it clearly state the required degree of accuracy needed for the operational state information. The Guide should be limited to existing and established Technical Specification definitions of operational modes.

Table 3 Plant Operational States for PWRs This table requires reliability data be documented for Reactor Protection Systems in shutdown modes. Technical Specifications provide the plant conditions under which Reactor Protection Systems and their channels must be functional in shutdown modes.

These Technical Specification requirements, for PWRs, apply typically whenever the reactor trip breakers are closed. They are not associated with " ...RHR initiations." This requirement introduces additional burden and presents confusion. The Regulatory Guide should be restricted to referencing currently approved Technical Specification definitions of operational modes.

Table 3 also requires unavailability information be tracked in all four plant operational states (POS) for emergency ac power system. In general, only one of two emergency ac power system trains is required in POS other than Pl. It is not clear from Section 1.3 that unavailability should be tracked even when the train or equipment is not required. The NRC workshop of June 4, 1996 indicated that unavailability should be tracked whether the equipment is required or not. The Regulatory Guide should provide additional guidance on this and address the additional operational documentation burden.

Section 3 Reportable Demands This section defines the conditions under which demands are reportable. These conditions differ from those required by INPO. Such differences create additional operational burden and data reporting confusion. The Regulatory Guide should reference the current information reported for the Emergency Electric Power System in accordance with NUMARC 87-00, Appendix 0.

Attachment Letter, YAEC to U.S. Nuclear Regulatory Commission July 3, 1996 Page3 Section 5 &portable Failures This section requires the reporting of failures for "reportable systems. 11 Certain nuclear units in New England have implemented the Maintenance Rule, 10 CFR 50.65, using the count of functional failures as a measure of the effectiveness of maintenance. The Regulatory Guide should provide a linkage to 10 CFR 50.65 and the reportability of functional failure requirements under that process. For example, the required reporting of failures on demand and on failure to run are incorporated as functional failures under the Maintenance Rule. The Regulatory Guide (Sections 5.1 and 5.2) should accept documentation of such Maintenance Rule functional failures .

Section 5.3 Recoverable Failures This section provides details on "recoverable failures." The section addresses operator action to nullify a failure (render it not reportable) if the equipment is started "... within a few minutes. 11 What constitutes a "few minutes" is subjective and, not counting those failures which are recoverable, is non-conservative. For nuclear units in New England, implementation of the Maintenance Rule counts all such failures as functional failures. To impose an additional reporting category will lead to operator and data reportability confusion. The Regulatory Guide instead should relate to required documentation of functional failures required under Maintenance Rule implementation.

Section 6 Unavailable Hours This Section requires that unavailability be reported as either planned or unplanned.

Consistent with the Maintenance Rule, certain nuclear units in New England, report all unavailability as one category. The Regulatory Guide should reference unavailability data requirements consistent with those required in meeting the Maintenance Rule.

This section also relies on the time in 11 *** plant operational states." See our comment regarding this subject contained in Section 2 above.

Section 7 Hours in Plant Operational States This section requires licensees to "... report the number of hours during each quarter that the plant was in each of the plant operational states ... " See our comment regarding this subject contained in Section 2 above.

CHARLES H. CRUSE DOCKE1EO Baltimore Gas and Electric Company Vice President !J, 'Rf'

'\,_ \ J Calvert Cliffs Nuclear Power Plant Nuclear Energy 1650 Calvert Cliffs Parkway Lusby, Maryland 20657 410 495-4455

'96 JUL - 3 A1 :13 DOCKET N BER PR PROPOSED RULE 50 BGE (bn=e.55\~J June 28, 1996 U.S. Nuclear Regulatory Commission Washington, DC 20555 ATTENTION: Document Control Desk

SUBJECT:

Calvert Cliffs Nuclear Power Plant Unit Nos. 1 & 2; Docket Nos. 50-317 & 50-318 Comments on Proposed Rule: Reporting Reliability and Availability Information for Risk-significant Systems and Equipment (61 FR 5318)

The Baltimore Gas and Electric Company is pleased to provide comments on the subject proposed rule.

We have reviewed the comments submitted by the Nuclear Energy Institute and the Nuclear Utility Backfitting and Reform Group. We endorse the comments submitted by both groups.

- We support Nuclear Energy lnstitute's commitment to work with the NRC to ensure that data needs are fulfilled in an effective and efficient manner. We also support greater sharing of the reliability and availability data that are already collected in the industry. Therefore, we believe that the proposed rule is premature and should not be promulgated.

Industry's burden regarding record keeping and reporting of data is already large. Any proposed rulemaking that will require submitting additional data should include a review of existing data requirements. The review should justify why each existing data requirement should not be eliminated if new data requirements are promulgated.

Finally, we are troubled by the NRC's determination that the proposed rule does not require a backfitting analysis. It seems intuitive that any rule that will require significant modification to a licensee's information collection and reporting process is a backfit. At a minimum, we believe that a backfitting analysis is required for proposed rules that will establish additional data collection and reporting criteria.

JUL 1 6 199t A\cknowlectged by card ..................... .... _

J.S. NUCLEAR ::...~~1l..r\TCflY COMMISSIOI\

DOCKET!N*"' & SERVICE SECTK)N OFFICE OF THE SECRETARY OF THE COMMISSION Document Statistics Postmark Date V\~...______

Coples Rece,ved_li _ ,__ _ _ __

,\dd't Copies Reorod* *ed -*- - - - -

Speclal Dislritx.Jtio ,,...l~---=,-,,..:l.:~ ::-:.;-

Document Control Desk June 28, 1996 Page2 Should you have questions regarding this matter, we will be pleased to discuss them with you.

CHC/JMO/dlm cc: D. A. Brune, Esquire Resident Inspector, NRC J.E. Silberg, Esquire R. I. McLean, DNR Director, Project Directorate 1-1, NRC J. H. Walter, PSC A. W. Dromerick, NRC A. R. Pietrangelo, NEI T. T. Martin, NRC

1,u_ .>1<,uJ-.' ,i 1../11 ,h*o/ c_.en l<'r

",ii/HI /)um/11/,,11 8u1 i/cu1 nl

<,'1<'11 .11l<'II I 1ni,i11 /11 -3,i!H,(I OOC ~{ ETE D USNRC

'96 JUN 24 P2 :58 June 20, 1996 U. S. Nuclear Regulatory Commission OF FICE OF St_:R~_f"A RY OOCKETIt~ G & :: f i<\' IC F BHANCH

• VIRGINIA POWER Serial No. GL 96-019A ATTN.: Docketing and Service Branch NL&OS/EJL Washington, DC 20555-0001 DOCKET NlMBER pR(l()SED RULE PR so

Dear Sir:

(~\FR.._53\~J @

10 CFR PART 50.76- PROPOSED RULE REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT FEDERAL REGISTER/ VOL. 61, NO. 29 / FEBRUARY 12, 1996 / P. 5318 This letter provides Virginia Power's comments regarding the subject proposed rule.

The Nuclear Regulatory Commission (NRG) has proposed to amend its regulations to require that licensees for commercial power nuclear reactors report plant-specific summary reliability and availability data for risk-significant systems and equipment to the NRG. The NRG has stated that the systems and equipment for which data would be provided are a subset of the systems and equipment within the scope of the Maintenance Rule. The NRG has determined that reporting of reliability and availability information is necessary to substantially improve the NRC's ability to make risk-effective regulatory decisions consistent with the Commission's policy statement on the use of probabilistic risk assessments (PRAs). The NRG claims that this information would assist the NRG in improving its oversight capabilities with respect to public health and safety and becoming more efficient by focusing its regulatory program on those issues of greatest risk significance and reducing unnecessary regulatory burdens on licensees.

The proposed reporting requirements would apply to the event-mitigating systems and equipment which have or could have a significant effect on risk in terms of avoiding core damage accidents or preserving containment integrity.

Our comments follow.

We support the establishment of industry data for PRA usage in order to encourage constructive PRA applications for enhancing safety while reducing licensee burden. We agree that the proposed required information would likely provide additional risk insights, and we fully support the NRC's initiatives to move towards risk-based regulation and the increased use of PRA information. We also appreciate the NRG staff's need for scrutable, plant-specific reliability and availability data compiled for risk-significant systems and equipment. However, we believe that the proposed requirements are unnecessarily burdensome, and that a more modest voluntary program could better accomplish the objectives of the proposed rule.

'JUN. 2 8" 199?-

Aclmowfedged by card ........................ w""....,."

IJ.S. NUCLEAR REGULATOOY COMMISSIOf\

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Documeni Statistics Poamark Date ~ -~ ~'-""'--- -

Coples Received _ __,___ __ _ __

,\def! Copies Reproduced -=-- - - --

Speolal DisbibutionCQQ, J>~ J 7>):;.@>

--R,::;ps I

Comments on Proposed Rule 10 CFR Part 50.76 Serial No. GL 96-019A page 2 of 3 The proposed rule would require collecting, recording, consolidating, and reporting a substantial amount of detailed information. This will require substantial resources . The proposed required information is typically not logged automatically by computer, and a considerable amount is not currently recorded. Plant staffing would have to be augmented or diverted from current tasks in order to perform the tasks required by the proposed rule. Alternatively, the design and installation of automated data collecting and logging systems would require a significant capital investment without a commensurate safety benefit. See Attachment 1 for additional comments on the burdens associated with the proposed rule.

The management and staff of the North Anna and Surry Power Stations have identified the risk-significant plant systems and equipment, are sensitive to the importance of the reliability and availability of these systems and equipment, and ensure that the Technical Specifications operability requirements for these systems are met. We believe that in general the reliability and availability performance of the risk-significant systems and equipment at our plants is excellent, and that those systems and equipment already receive a significant level of regulatory attention and scrutiny from the NRC's resident inspectors. Therefore, additional regulation is not needed in this area. Further, we believe that the benefits associated with the proposed rule would be incremental and not in proportion to the substantial effort needed to meet the requirements of the proposed rule.

We are also concerned that the proposed rule may have a deficiency for dual unit plants such as our North Anna and Surry Power Stations. The proposed rule does not appear to address dual unit sites with shared support systems and special backup features for safety systems (the recovery of safety functions using the opposite unit's systems).

Following the requirements of the proposed rule would potentially result in improperly collecting data to model these shared support systems and special backup features.

Some examples of the shared support systems are the Surry emergency switchgear room cooling system and the North Anna service water system. Some examples of the special backup features for safety systems are the Surry auxiliary feedwater cross-connect between units 1 and 2 and a similar cross-connect for high pressure injection at both North Anna and Surry. (Examples of the specialized data required for these Virginia Power dual unit PRA models are provided in our responses to Generic Letter 88-20 for the North Anna and Surry Power Stations.)

Finally, the rule requirements seem to be too broad in certain areas. For example, paragraphs 50.76(b)(iii) and (v) request unavailability information. However, the rule does not address whether or not this information is required for all time periods or only for the time periods when the equipment is required to be operable by the plant's Technical Specifications. As an example, it is not clear why the availability of the reactor protection system (RPS) or the auxiliary feedwater system (AFWS) would be of interest when the plant is shutdown and the RPS and AFWS are not required to be operable. In addition, our automated Technical Specification Action Statement tagout logs could not be fully utilized since the majority of the risk-significant systems and

Comments on Proposed Rule 10 CFR Part 50.76 Serial No. GL 96-019A page 3 of 3 equipment are not required to be operable during shutdown, and their availability is not tracked when they are not required to be operable. We recommend that this issue be clarified.

We fully endorse and support the comments sent separately to the NRC by the Nuclear Energy Institute. We appreciate the opportunity to provide comments on this proposed rule.

Very truly yours, M. L. Bowling, Manager Nuclear Licensing & Programs Comments To Office Of Management And Budget, Proposed Rule - 10 CFR Part 50.76, Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment, Federal Register/ Vol. 61, No. 26 / February 7 1996 IP. 1686 cc: Mr. A. R. Pietrangelo Project Manager Licensing & Performance-Based Regulation Nuclear Energy Institute 1776 I Street NW Suite 400 Washington, DC 20006-3708

Attachment 1 Comments To Office Of Management And Budget

• Proposed Rule -10 CFR Part 50.76 Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment Federal Register/ Vol. 61, No. 26 / February 7 1996 / P. 1686

',(){)() /)(Jll/111/CJI/ l/,111i, *1,11 *,,

G/('11 Alll'11. \°i1~1111u .!,rH ,,,

March 15. 1996 Mr. Troy Hiller Office of Information and Regulatory Affairs VIRGINIA POWER Serial No. GL 96-019 NL&OS/EJUDJB (3150-0011)

NEOB-10202 Office of Management and Budget Washington, DC 20503 e

Dear Mr. Hiller:

DOCUMENTS CONTAINING REPORTING OR RECORDKEEPING REQUIREMENTS:

OFFICE OF MANAGEMENT AND BUDGET (0MB) REVIEW PROPOSED RULE - 10 CFR PART 50,76 REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SVSTEMS AND EQUIPMENT FEDERAL REGISTER / VOL, 61, NO, 26 LFEBRUARY 7 1996 / P. 1686 This letter provides Virginia Power's comments regarding the subject notice of the 0MB review of information collection. The Nuclear Regulatory Commission (NRC) has recently submitted to 0MB for review a proposal for the collection of information under the provisions of the paperwork Reduction Act of 1995 (44 U.S.C. Chapter 35). The title of the information collection is 10 CFR Part 50.76, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment."

Our comments follow.

1. The estimates of the Industry and Licensee cost and resource burdens to implement programs to gather, analyze, file and report*the requested data may be substantially underestimated in the NRC's Supporting Statement to 0MB. The NRC has stated that they expect licensees to monitor the performance of an average of 11 risk-significant systems. The NRC went on to describe monitoring and reporting requirements for "ensembles of equipment" and failure records at the component level. The NRC did not define "ensemble of equipment", nor did they provide an estimate for the number of ensembles or the number of components that would fall within the scope of the proposed rule. We are concerned that this scope could tum out to be substantial, and result in a burden that far exceeds the NRC's estimate.

The NRC should be asked to provide additional information regarding "ensembles of equipment" and the number of components that are risk significant in order to allow NRC licensees and 0MB to better evaluate the burden associated with this proposed rule. The burden estimates should not be considered accurate or complete until this additional level of detail has been provided by the NRC. For reference, we estimate that we are currently expending approximately 300 hours0.00347 days <br />0.0833 hours <br />4.960317e-4 weeks <br />1.1415e-4 months <br /> per year per unit on monitoring the performance (reliability and availability) of 17 risk significant systems at the train level.

0MB Review of 10 CFR Part 50. 76 Serial No. GL 96-019 page 2 of 2

2. The NRC states that there is no similar information available to the NRC. We believe that similar information is available. Reliability and availability data (down to the train level) for risk-significant systems is being collected by licensees in accordance with the provisions of 10 CFR 50.65, the Maintenance Rule. This data is available to the NRC as part of their inspection activities.

Finally, we fully endorse the comments sent separately to the 0MB by the Nuclear Energy Institute. We appreciate the opportunity to provide comments on this proposal tor the coiiection of information.

Very truly yours, JN_/!,~

M. L. Bowling, Manager Nuclear Licensing & Programs cc: Mr. A. A. Pietrangelo Project Manager Licensing & Performance-Based Regulation Nuclear Energy Institute 1 ns I Street NW Suite 400 Washington, DC 20006-3708

DOCKETED USNRC Arizona Public Service PALO VERDE NUCLEAR GENERATING STATION '96 JU 24 A9:23 P.O. BOX 52034 PHOENIX, ARIZONA 85072-2034 102-03718-WLSIRA8cJFFI'.> OF sr:::RET.A.RY WILLIAM L. STEWART EXECUTIVE VICE PRESIDENT June 17, 1996 OO CK [Tlt 1C 8. SERVICE NUCLEAR 5RP,~ :H oocm tUIBEA111t -~

U. S. Nuclear Regulatory Commission PROPOSED RULE rn 5 0

• Washington, DC 20555-0001 (~\f"R_.5~\"e,j

• ATTN.: Docketing and Service Branch

Subject:

Palo Verde Nuclear Generating Station (PVNGS)

Units 1, 2, and 3 Comments on Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment Proposed Rule These comments are being submitted by Arizona Public Service Company (APS), on behalf of itself and participants in the Palo Verde Nuclear Generating Station (PVNGS), in response to the request for comments on the proposed rule, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment," noticed in the Federal Register dated February 12, 1996 (Vol. 61, No. 29 Federal Register 5318).

- In the Regulatory Analysis, the NRC states that the backfit rule does not apply to information collection and reporting requirements. The backfit rule, 10CFR50.109, states in part: "Backfitting is defined as the modification of . . . the procedures or organization required to . . . operate a facility ... which may result from a new or amended provision in the Commission rules ... " The Staff estimates that the data collection burden will be 1375 person-hours per year per reactor.

This represents more than one half of an employee's time for one year, which is an impact on a licensee's organization required to operate the facility, without a demonstrated "substantial increase in the overall protection of the public health and safety." Contrary to the Staff's position, APS believes that the backfit rule applies to this proposed rule, does not meet requirements, and should not be promulgated.

APS believes that the proposed rule does not contain a demonstrated improvement in plant safety.

Since one cannot predict in advance with any confidence the scope, level of detail, or precision of data required to support any future risk-based or performance-based regulatory application, we believe that the proposed rule will not satisfy future needs. Data required to support submittals relying on risk insights will be application specific, and will require submittal of specific data rather than using the data proposed to be provided under the proposed rule. APS

illmowledged by card ...JJL~..a.*.~"~

l

I J.S. NUCLEAH ~ < * . * ..,.; , COMMISSIO~

OOCKETING 6 !:il::RVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Doa.ment Statis1ics Pottmart Dale --t'-4-<i->->-----

Copea Received ___________

Men r

U.S. Nuclear Regulatory Commission ATTN.: Docketing and Service Branch Comments on Reporting Reliability and Availability Information Proposed Rule Page2 also believes the proposed rule will not reduce the licensee burden for information submitted to support any licensing action, because all information to support such an action would be provided in a stand-alone document.

Another area of concern with the proposed rule is that the data required are duplicative of other data sources available to the NRC to support its activities. The WANO Safety System Performance Indicator (SSPI) program includes much of the data requested under this rule.

Licensees already bear a significant burden with regard to record keeping and data reporting. The proposed rule does not eliminate any of the existing data requirements. Therefore, we believe that the additional burden imposed by the proposed rule is unnecessary and not justified.

In summary, APS believes that the proposed rule is unnecessary and should not be promulgated.

Additionally, APS supports the Nuclear Energy Institute's comments on behalf of the industry.

We appreciate the opportunity to comment on the proposed rule and the Staff's consideration of our comments.

WLS/RAB/dpr cc: L. J. Callan K. E. Perkins J. W. Clifford K. E. Johnston NEI

6 South Carolina Electric & Gas Company Gary J. Taylor PO. Box 88 Vice President Jenkinsville, SC 29065 (803) 345-4344 DOCK E~u~,r Operations SCE&G June 11, 199GI St-JRC ASCANIICompany RC-96-0155

'96 JUN 21 P3 :35 Document Control Desk U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 ooc,ETtUIERtlft"'---

Attn: Docketing and Service Branch PROPOSED RULE rn 50 -

Dear Sir:

(lo\FR.So\~) @

Subject:

VIRGIL C. SUMMER NUCLEAR STATION DOCKET NO. 50/395 OPERATING LICENSE NO. NPF-12 COMMENTS ON PROPOSED RULE, "REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT" This letter provides South Carolina Electric & Gas' (SCE&G) comments on the Nuclear Regulatory Commission's proposed rulemaking on reporting reliability and availability information for risk-significant systems and equipment. SCE&G fully endorses the Nuclear Energy Institute comments regarding this proposed rule.

Overall, the proposed rule is totally unnecessary and represents an additional regulatory burden which has not been cost justified or demonstrated to significantly reduce the risk to the health or safety of the general public.

SCE&G appreciates the opportunity to review and provide comments on this proposed rule.

Very truly yours,

do.;;~ rv/L C-,

5i-S;e",~,,E,,/ A- 6)'1i?1VE Gary J. Taylor MJZ/GJT /nkk Attachment c

J. L. Skolds R.R. Mahan R. J. White A. R. Johnson NSRC RTS (PR 960008)

DMS (RC-96-0155)

File (811.02, 50.083) ,_ .. .,.,_ .

-JI![ 2I' 19S:

A&;knowtedged by card ....:....................;..,..,,**

NUCLEAR EXCELLENCE -A SUMMER TRADITION!

tJ.S NUCLEAR REGULATORY COM ISSt('~

DOCKtTl!'lG & SERVICE SECTION OFFICE Of THE SECRET ARY OF THE COMMISS Docunent Statisb Postmark Date ---.µS!!~~ - - --

Co~s Received. _ _.__ _ _ _ __

,\dd'1 Coples ReprodOOld ~ ~ - --

Speclal Dts1Jtoo1lon ~ I rwos:-,, ,~ ;,

Log # TXX-96385 File # 10185 Ref.# 61FR5318 DOCKETED us~mc

• 96 JUN 18 A9 :11 1UELECTRIC OFFICE [ii~ SEC RF TJ\ Y June 11, 1996 DOC KE T!~~~ ~< :.:::n ICF 1

C. Lance Terry BRAHLH Group Vice President U. S. Nuclear Regulatory Commission Attn: Docketing and Service Branch Washington, DC 20555-0001

SUBJECT:

SUBMITTAL OF COMMENTS ON PROPOSED RULE REGARDING REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK SIGNIFICANT SYSTEMS AND EQUIPMENT REF: 1) Reporting Reliability and Availability Information for Risk Significant Systems and Equipment, Proposed Rule, Federal Register, Volume 61, No. 29, February 12, 1996, Pages 5318 through 5326.

Gentlemen:

By Federal Register notice dated February 12, 1996 (Reference 1), the NRC solicited comments for amending its regulations to require that licensees for commercial nuclear power reactors report plant-specific summary reliability and availability data for risk-significant systems and equipment. These comments are submitted by TU Electric in response to the NRC notice of February 12, 1996 (Reference 1).

TU Electric commends the Commission for its commitment toward the use of Probabilistic Risk Assessment (PRA) methodology in nuclear regulatory activities. TU Electric supports the use of PRA methodology in the nuclear industry and is actively involved as part of a pilot plant project on Risk-Based Inservice Testing. TU Electric recognizes the need for accurate and consistent quantitative data to support the development of risk-based programs. However, TU Electric believes this information can be attained through efforts that do not require rulemaking. TU Electric also believes that it would be useful to suspend the rulemaking while the NRC evaluates easily obtainable data from a sample of utilities (possibly through pilot plants) and examines that data to determine how plants are currently maintaining reliability data.

The rulemaking process should be suspended to adequately address the following issues:

• the benefit to the NRC for having the information versus the cost and current capabilities to supply the information to the NRC (details encountered during implementation may result in more costs and implementation time than estimated) and Energy Plaza 1601 Bryan Street Dallas, Texas 75201 -3411

I J.S. NUCLEAR tit.uvlA rORY COMMISS~

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Document Statistics PostmmDate ~ ( < ' ! ' b ~ ~

Co~s Received,---'1_ _ _ _ _ __

.\dd1 Copies Reproduced ..:::

~,:___ _ __

Special DlstrlbutiooCW* *= ~

TXX-96385 Page 2 of 3

• data required under the proposed rule may duplicate other data sources and add to the reporting burden of licensees.

TU Electric also questions some of the general assumptions in the regulatory analysis such as the 1% availability increase per plant and the estimate that 80% of all plants are already collecting the required data.

The data requested by the proposed rule exists in separate documents such as work orders and manual logs. It would require a significant effort to manually gather and evaluate the data. Even if part of this process were automated, it would require a significant amount of time to assure that events were correctly categorized.

TU Electric opposes the proposed rule in its present format. However. if the NRC pursues methods of obtaining data to support PRA methodology the following objectives should be paramount.

• Greater use of the PRA applications in the nuclear regulations.

Examples are graded QA, risk-based IST/ISI, and risk-based Technical Specifications.

• Eliminate or reduce duplicate reporting in 10CFR50.73, "Licensee Event Report System," and 10CFR50.72, "Immediate Notification Requirements for Operating Nuclear Power Reactors."

• Eliminate existing regulatory requirements that have marginal importance to safety based on the PRA insights and plant-specific reliability/availability data. This is consistent with the NRCs Annual Planning and Program Guidance (PPG) document dated 1984.

• Use of PRA models. supported by the plant-specific reliability/

availability data. to facilitate performance-based regulation and to develop new plant performance indicators to replace the existing ones.

In addition. the NRC should minimize the documentation and reporting requirements of the data collection so that the utilities can use their existing tools and programs to the maximum extent possible in order to minimize additional costs.

Comments on the draft Regulatory Guide DG-1046, "Guidelines for reporting Reliability and Availability Information for Risk-Significant Systems and Equipment in Nuclear Power Plants," (Federal Register, Volume 61, No. 86.

May 2. 1996. Pages 19645 through 19646) are due July 5. 1996.

TU Electric is continuing the review of the proposed rule and the draft regulatory analysis and will provide any additional comments at that time.

TXX-96385 Page 3 of 3 TU Electric endorses the responses submitted by the Nuclear Energy Institute on behalf of the nuclear energy industry (letters dated March 12, 1996, and June 11, 1996). TU Electric appreciates this opportunity to comment on the proposed rule and requests that the Commission or NRC staff contact Carl Corbin at (214) 812-8859 if there are any questions.

Sincerely, C. L. Terry By,wrwJ~

D. R. Woodlan Docket Licensing Manager CBC/cbc c- Mr. L. J . Callan, Region IV Ms. L. J. Smith, Region IV Mr. P. M. Ray, NRR Resident Inspectors. CPSES

OCKETE.0 U.. 'RC

~

ROCHESTER GAS AND ELECTRIC CORPORATION

• 89 EAST AVENUE, ROCHESTER, N. Y. 14649-0001 AREA CODE 716 546-2700

.96 JUN 17 p 2 :22 ROBERT C. MECREDY Vice President Nucl ear Operations June 11, 1996 DOCKET NlNBER pn

)

_U.S . Nuclear Regulatory Commission pROf>()SEO RULE R SQ -

Attn: Docketing and Service Branch Washington, D .C. 20555 lto\FR}5S\~J

Subject:

Comments on Proposed Rule for Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment

Reference:

(a) Federal Register, Volume 61, Number 29, 5318-5325 Gentlemen, By Reference (a), the NRC published a proposed rule (10CFR50.76) concerning a potential new requirement for holders of operating licenses for nuclear power reactors to report reliability and availability data of risk-significant systems and equipment. The purpose of this letter is to provide comments on this proposed rule from the standpoint of a licensee of a nuclear power reactor.

The proposed rule as written would essentially require the following :

1. Compilation on a quarterly basis of equipment reliability data for an expected average of 11 "event mitigating" systems identified as risk-significant by a licensee. The reliability data would include the following (specific requirements are identified in parentheses):

a. Equipment demands (number);

b. Equipment failure information (date, train affected, type of demand, plant mode, and plant failure record which identifies these items including duration of unavailability and cause and effect);

c. Equipment operating time ( date, train affected, statement of whether the run was successful);

d. Equipment unavailability (train affected, plant mode, type of unavailability, and identification of support system if this was the cause of the unavailability); and

e. Number of hours when two or more trains from same or difference systems were concurrently unavailable characterized by trains that were unavailable.

U.S. NtJr~L _ _v._, . i 0HY COMMISSIO/1.

DO ,KETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Doalnent Statisb Postmark Date . . .<;~=>\->.>W..-

o - --

Coples Recelved_.e._ l _ _ _ _ __

,\dtfl Coples Reproduced -.::....- - - -

(!,-1.a

~

Olsibmon C)co,ei:r:n ~~,-

---RI Q>S

2. Submittal of an annual summary report of the quarterly compiled information.

3. Retention onsite for at least five years those records and documents related to each occurrence of a demand, failure, or unavailability that provided the basis for the data included in the annual report.

The NRC estimates that initial implementation of the proposed rule would require an average of 423 hours0.0049 days <br />0.118 hours <br />6.994048e-4 weeks <br />1.609515e-4 months <br /> per licensee with a recurring annual burden of 1375 hours0.0159 days <br />0.382 hours <br />0.00227 weeks <br />5.231875e-4 months <br /> for the required annual reports. The Federal Register notice states that this proposed rule is required to essentially support the following programs and issues:

a. Risk-based regulation

b. Resolution of generic issues C. Indicate if individual plants have plant performance concerns

d. Accident Sequence Precursor and Event Analysis program

e. Risk-based inspections

f. Equipment aging

g. Risk-based technical specifications

h. Inservice Testing I. Maintenance rule (I0CFRS0.65)

J. NRC monitoring in general Following a review of this proposed "data reporting" rule, NRC justification, and the burden estimate, RG&E would like to offer the following:

1. While RG&E supports the NRC's efforts with respect to factoring more risk-based information into areas such as regulations, inspections, testing, and technical specifications, the data reporting rule as proposed is overly burdensome. Specific examples of this are described below. However, RG&E would strongly encourage continued dialogue with the Nuclear Energy Institute (NEI) and Institute for Nuclear Power Operations (INPO) on this subject since it is easily conceivable that with the planned changes to the Nuclear Plant Reliability System (NPRDS) as maintained by INPO, the NRC could obtain most, if not all, of their requested information. Specifically, changes would allow the NRC to obtain reliability data electronically for all nuclear plants using a database with consistent trending, reporting formats, and standards. In addition, data needs may change in the future which if written into a formal rule as proposed, would require rulemaking to change. The use of available industry programs would therefore, result in the lowest burden to both licenses and the NRC.

2. Determination of specific component demands can be a very cost-intensive effort. For example, requiring plant personnel to document every component demand during a planned test could conceivably be accomplished through hand-written records or even computer tracking points. However, equipment challenges due to component failures or plant trips can be very difficult to obtain due to other more pressing priorities (e.g., stabilizing the plant).

In addition, the use of plant personnel for recording equipment demands is subject to human errors in counting. This raises the question that if an operator incorrectly misses a component demand, is a licensee subject to a violation? It is RG&E's opinion that educated estimates of component demands can and should be used in many instances without significantly impacting the final result. For example, assume that there are 3 failures of a component over 400 known demands. Suppose that it was estimated that there were only 380 demands of this same component. The failure rates between the known and estimated case differ by only 5% which is basis for including uncertainty in risk analyses. However, the cost between developing an educated estimate and actual demand value can be significant.

3. Being required to identify and report if equipment unavailability is due to support system inoperability to "properly capture concurrent outages and eliminate double counting" goes far beyond the maintenance rule requirements and the Improved Technical Specifications (ITS) recently implemented at Ginna Station. The maintenance rule effectively requires tracking of equipment unavailability which may be tracked either on the basis of direct equipment inoperability or equipment inoperability plus support system inoperability. This requirement would force licensees to specifically identify and track support system inoperability for a given component. Also, the ITS specifically do not require "cascading" equipment inoperabilities except when two or more sypport systems are inoperable (see LCO 3.0.6). Since the majority of licensees have committed to convert to the ITS, this reporting requirement is unnecessary and overburdensome.

4. Being required to report the number of hours in which two or more trains are inoperable for the purpose of identifying "subtle" risk concerns could generate an incomplete picture of risk. For example, Ginna Station has two auxiliary feedwater systems comprised of 5 pump trains due to high-energy line break and fire concerns. Therefore, a high unavailability of one AFW pump train at Ginna Station would be of much less significance than at a plant with a typical 3 pump train system design (note - this issue also impacts what is "risk significant"). In addition, dependencies between systems supported by the inoperable trains may not be readily apparent (e.g., DC power) . These dependencies must be specifically addressed in the ITS (see LCO 3.0.6) such that NRC concerns over improper plant configurations is now specifically addressed within technical specifications. The maintenance rule also requires evaluation of on-line maintenance activities. If the NRC is still concerned over licensee's inappropriate on-line maintenance activities, efforts to enforce the existing rules should be made versus creating a new rule.

5. Requiring the retention of plant records for up to 5 years in the proposed rule is redundant to existing requirements within the quality assurance (QA) program of most, if not all, licensees. Currently, similar record retention requirements are specified by Regulatory Guide 1.88 as referenced within the Ginna Station QA Program. The QA program as required by IOCFRS0.54 specifically requires NRC review of any changes which would reduce the effectiveness of the program. Therefore, it is unnecessary to specify this requirement within the proposed rule.

6. The NRC proposes that a 10CFRS0.109 Backfitting Analysis does not need to be performed.

Because of the relatively high manpower estimates calculated in the NRC's Regulatory Analysis for implementation of this proposed rule (7 /1 O's of a person per year), and the fact that new procedures and a new program would have to be created by each licensee, we believe that the proposed rule should not be exempted from IOCFRS0.109. Furthermore, if the NRC manpower estimates were updated following industry input based on past data collecting activities, the proposed rule would not meet the criteria of 10CFRS0.109.

We hope that the NRC finds these comments to be useful. RG&E has collected, and continues to collect, reliability data in several (but not all) of the areas which the proposed rule would impose requirements in. These efforts have gained valuable insights into the problems which exist with collecting reliability data which RG&E would be willing to share further with the NRC if so desired.

Nonetheless, RG&E does not believe that Ginna Station could implement the proposed rule in the burden estimates proposed the NRC.

Very truly yours,

'-t;;L_

Robert C. Mecredy MDF\860 xc: Mr. Tony Pietrangelo Nuclear Energy Institute 1776 I Street NW Suite 400 Washington, DC 20006-3708

II I Duke Power Company M. S TUCKMA1\'

P.O. Box 1006 Charlotte, NC 28201-1006 DOCKETED Senior Vice President Nuclear Generation US 'RC (704)382-2200 Office (704)3824360 Fax

'96 JUN 17 A9 :23 DUKEPOWER OFFIC E OF SECR_E_!l~ RY DOC KETli*F1 *, ~I:. RvICF sn {\i~cr1 June 10, 1996

-~ -...

U.S. Nuclear Regulatory Commission DOCKET NlMBERPR . . .____

Washington, D. C. 20555 PROPOSED RULE 59 *

(_<o\rR.5~\~J ~

Attention: Docketing and Service Branch

Subject:

Proposed Rule on Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment; Duke Power Company Comments In response to the subject Proposed Rule, which was published in the February 12, 1996 Federal Register, Duke Power Company offers these comments on the need for the rule and its potential impact.

The cost, both to the NRC and to licensees, has been seriously underestimated.

The NRC's own estimate of the data collection and reporting burden for this rule is 1375 hours0.0159 days <br />0.382 hours <br />0.00227 weeks <br />5.231875e-4 months <br /> per reactor per year.

(By comparison, data collection for the maintenance rule is approximately 330 hrs/yr/unit.) In reality, however, data collection will be much more onerous. It is estimated that a typical plant will have 20 systems that would be considered risk-significant, with each system typically having 100 components. Assuming that a highly automated method is available to record actual demands and operating hours, the time required to collect, record, and report data is estimated to be 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />s/year per component; about 4000 hours0.0463 days <br />1.111 hours <br />0.00661 weeks <br />0.00152 months <br /> for the 2000 components identified above. If manual data collection and recording are necessary, the time estimate becomes 15 hrs/yr per component; about 30,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />. The actual total time will likely be somewhere in between. But even at the low end of the estimate, the burden is almost triple the estimate given in the NOPR.

Acknowfed ed b .., JUn- 1-g* 199(:__

g vca, u ,..!_"...- ..;;...,;.."""'

U.S. t'UCU:Ar H:-: J L.r.'i ,>W CO , ISSIOt-..

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Documen1 Statisb Postmant Date _ 1..,;

=µJ-~ " -- - - - -

Copes Received,- - - - - -

.\dd'I Co?es Reprod~ed ~ - - - -

Special Oislriootion ~~;;i,QS:~..J.:::~ -

U. S. Nuclear Regulatory Commission June 10, 1996 Page 2 For the NRC, the cost associated with compiling, evaluating, and making meaningful use of the data received from about 100 plants would be similarly more burdensome than is anticipated at present.

The proposed rule is not needed to facilitate any current NRC regulatory function.

The Federal Register notice of proposed rulemaking (NOPR) offers several justifications for the proposed rule, mostly centered around improving the NRC's oversight capability in various areas, and the future development of risk-based regulation. No deficiencies in the NRC's current capabilities are identified which the proposed rule would address to enable the NRC to fulfill its statutory mission to protect the health and safety of the public. Therefore, the vastly increased resources needed to comply with the recordkeeping and reporting requirements of the rule are not justified by any need to assure the "adequate protection" standard of 10 CFR 50.109 (Backfitting). The NOPR stated that "a backfit analysis is not required because these amendments do not involve any provisions which would impose backfits as defined by 10 CFR 50.109(a) (1)." This is simply not true. Paragraph (a) (1) defines a backfit as, among other things, the additions to procedures required to operate a facility, which may arise from a new or amended provision in Commission rules. The recordkeeping and reporting requirements imposed by the rule would indeed necessitate significant procedural and programmatic additions.

It is not clear how the NRC will use the data.

The NOPR states that "plant-specific, train-level reliability and unavailability data would be used to improve the plant-specific ASP [accident sequence precursor] models which the NRC uses ... " It is unlikely that the NRC will create and maintain true plant-specific models for each plant in the country. Using plant-specific data in a generic model is of minimal value, and does not justify the effort required to collect it.

U. ~- Nuclear Regulatory Commission June 10, 1996 Page 3 Other alternatives are available to obtain reliability and availability information.

Licensees will be compiling and maintaining comprehensive equipment failure and unavailability information at the site as part of the maintenance rule program. The NRC will have access to this data. This is adequate and avoids additional layers of recordkeeping and data management.

A second alternative is to make use of the licensee event report (LER) rule. If additional detail on the risk significance of the equipment failure or unavailability is necessary, 10 CFR 50.73 could be revised to reflect that.

It should be noted that the current LER rule requires reporting significant events to the NRC, including risk-significant failures and unavailabilities. It is this LER data that the NRC currently uses for the accident precursor analysis and plant performance indicator programs.

The rule is requesting more extensive and detailed information than is currently required by the Maintenance Rule.

The focus of the Maintenance Rule is at the function level, while this proposed rule focuses at the train level.

Monitoring at the function level is more effective, and reflective of how the plants are operated.

The Maintenance Rule requires the tracking of unavailability and reliability on those systems determined to be risk significant by the utility. Maintenance Rule unavailability is the time the system is unable to perform its function.

It is important to note that the definition in the proposed rule is different than the definition used by the Maintenance Rule. These definitions should be the same to avoid confusion.

The Maintenance Rule also requires the tracking of reliability, in the form of Maintenance Preventable Functional Failures. The proposed rule is asking for the number of failures to start per train, along with the date and mode of each failure.

A periodic assessment is required by the Maintenance Rule for each refueling cycle. This report is not required to be sent to the NRC, but simply kept available for NRC audits and inspections. The proposed rule would require an annual

U. S. Nuclear Regulatory Commission June 10, 1996 Page 4 report be submitted to the NRC; should this rule be promulgated the requirement should reflect reporting on a refueling-cycle basis, to be consistent with the Maintenance Rule.

To collect and evaluate all of the data proposed would place a great burden on Duke Power. It would require extensive reengineering of processes and computer systems. This, of course, translates into a significant impact on our resources.

The rationale for the rule implies high precision and a static and bounding characteristic for reliability and availability data.

Sensitivity studies using NUREG-1150 models and data have shown that the reliability or unavailability value can vary from an original estimate or a reference value by several factors, and in some cases an order of magnitude, without significantly affecting analyses. The proposed rule is striving for a precision which is inconsistent with the technology.

The NRC envisions verifying that the reliability of safety systems, derived from the failure-actual demand data, is consistent with that calculated from estimated demands.

This is impractical, since the failures of safety systems are rare and actual demands are infrequent. A typical system that is estimated to be 99.9% reliable would require hundreds of years of actual-demand failure data before a statistically meaningful comparison could be made.

Conclusion The proposed rule will result in significant costs and resource commitments which are not commensurate with the benefit. There are other data collection mechanisms which would provide an adequate source of data.

It is recommended that the NRC perform sensitivity studies, using NUREG-1150 models and data, to identify what "data" is truly risk-significant. Because a system is risk-significant does not mean that data on component demands and operating hours need to be precisely determined. If this effort identifies additional information to which the NRC does not have access, a better option would be to consider revising

U. S. Nuclear Regulatory Commission June 10, 1996 Page 5 the LER rule to require more reliability/availability data equipment fails to operate as designed.

Very truly yours, M,~.rf'~

M. S. Tuckman xc: Mr. David E. Labarge, Project Manager U. S. Nuclear Regulatory Commission Mail Stop 0-14 H25 Washington, D. C. 20555-0001 Mr. Peter S. Tam, Project Manager U. S. Nuclear Regulatory Commission Mail Stop 0-14 H25 Washington, D. C. 20555-0001 Mr. Victor Nerses, Project Manager U. S. Nuclear Regulatory Commission Mail Stop 0-14 H25 Washington, D. C. 20555-0001 Mr. Thomas E. Tipton Nuclear Energy Institute 1776 I Street, NW - Suite 400 Washington, D. C. 20006-3708

Duquesne Li~ Company Beaver Valley Power Station P.O. Box 4 OOCKETE Shippingport, PA 15077* 0004 US RC

'96 JUN 17 A9 :23 SUSHIL C. JAIN (412) 393*5512 Division Vice President Fax (412) 643*8069 Nuclear Services Nuclear Power Division June 10 9 0 ET R 1'

, IN ICE U.S. Nuclear Regulatory Commission R Attention: Docketing and Service Branch Washington, DC 20555-0001 DOCKET RULE_ I 5 0 p ~~ 5~\SJ .:If:~

0

Subject:

10 CFR Part 50 Proposed Rule, " R * ~ ~ b ~ t ; ;~d,.A~ilability Information for Risk-Significant Systems and Equipment" and the Associated Regulatory Analysis Duquesne Light Company (DLC) is responsible for the operation of Beaver Valley Power Station Units 1 and 2. DLC has reviewed the 10 CFR Part 50 proposed rule, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment," which was published in the Federal Register on February 12, 1996 (61 FR 5318) and the associated Regulatory Analysis. DLC hereby submits the attached comments.

DLC concurs with the comments being provided by the Nuclear Energy Institute (NEI). In particular, the proposed data reporting and recordkeeping is not needed by the Nuclear Regulatory Commission (NRC) for assessing compliance with current regulatory requirements. Also, DLC and NEI previously submitted comments on the supporting statement for the proposed rule to the Office of Management and Budget (0MB). DLC would like to reiterate that the estimate of licensee burden provided by the NRC in the 0MB statement was inaccurate. The proposed rule would impose a substantial burden on the nuclear industry.

Thank you for the opportunity to comment on this issue. If you have any questions on this submittal, please contact Mr. Roy K. Brosi, Manager, Nuclear Safety Department, ( 412) 3 93-5 210.

Sincerely, Sushil C. Jain --

0ELIVER ING OUAL I TY ENERGY

Beaver Valley Power Station, Unit No. 1 and No. 2 10 CFR Part 50 Proposed Rule and the Associated Regulatory Analysis Page2 c: Information and Records Management Branch (T-6-F33)

U. S. Nuclear Regulatory Commission Washington, DC 20555-001 Desk Officer, Office of Information and Regulatory Affairs NEOB-10202 (3150-0011)

Office of Management and Budget Washington, DC 20503

ATTACHMENT 1 Comments on 10 CFR Part 50 Proposed Rule, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment" and the Associated Regulatory Analysis

• Based on the discussion under "Prior Efforts" on page 5319 of the Federal Register notice, the Nuclear Regulatory Commission (NRC) appears to be proceeding with rulemaking because of a perception of industry inaction in this area. It should be noted that the Institute of Nuclear Power Operations (INPO) Safety System Performance Indicators (SSPI) Program was expanded from 3 to 4 systems, with expanded data collection. It is Duquesne Light Company's (DLC) position that a new reporting system could be avoided by modifying the INPO SSPI Program to include an agreed upon minimal list of systems and data to be collected.

• DLC notes that there are other industry initiatives that would seem to preclude the need for the proposed rule. For example, the Westinghouse Owners Group (WOG) is actively pursuing development of a risk-based Inservice Testing (1ST) Program for pumps and valves. As part of this effort, risk significance data is being collected from various WOG members to determine where testing should be strengthened or relaxed.

WOG is also in the process of "benchmarking" member plant Probabilistic Risk Assessment (PRA) models in an effort to make data sharing more meaningful.

• One of the NRC justifications for the proposed rule is that it would provide a source of "current industry-wide reliability and availability information for some of the systems and equipment within the scope of the Maintenance Rule." This information could then be used to enhance the goal-setting process described in the Maintenance Rule. DLC agrees that if the proposed rule is implemented, it would make sense to use this information in the goal setting process. However, as stated by the NRC, this is an enhancement, and it should not be used to justify new regulations.

• Note 3, at the bottom of page 5322 of the Federal Register, appears to indicate a less restrictive application of the Maintenance Rule than is required by NUMARC 93-01 guidance. This may account for the NRC's perception that approximately 30 of 110 plants are collecting a minimum amount, if any, of performance data.

(Continued)

Comments on 10 CFR Part 50 Proposed Rule and the Associated Regulatory Analysis Page 2

• On page 5324 of the Federal Register, "The Commission requests public comments on whether the proposed rule should contain such a sunset provision, and if so, the period of time after which the rule should automatically expire." If the sunset provision would require further rulemaking to expand the scope of data collection, DLC would favor such a provision. DLC has no comment on the expiration time.

• On page 5324 of the Federal Register, "The Commission requests public comments on whether the proposed rule should exempt plants that have announced (or will announce) plans to discontinue operation within a short time (e.g. two years)."

DLC considers such a "grandfathering" clause to be appropriate since it is clearly unwarranted to require compliance from plants who plan to discontinue operations within two years. Furthermore, it seems this data may be of little value to the rest of the industry.

• Page 5325 of the Federal Register contains the following series of questions for which the NRC is seeking public comment.

Question:

Is the proposed collection of infmmation necessruy for the proper performance of the functions of the NRC, and does the information have practical utility?

Response

For reasons already cited, DLC does not consider the proposed rule to be necessruy for the NRC to achieve regulatory improvements through risk-based regulation.

While DLC does agree that the additional data may have marginal practical utility, the cost of achieving this marginal benefit is unreasonable. The Maintenance Rule should be given time to be implemented and evaluated to determine its expected contributions before new rulemaking is proposed.

Question Is the estimate of burden accurate?

Response

DLC's comments on the 0MB Suppmting Statement were submitted to the Office of Management and Budget in a letter dated March 14, 1996, and are included as Attachment 2.

(Continued)

Comments on 10 CFR Part 50 Proposed Rule, and the Associated Regulatory Analysis Page 3 Question:

Is there a way to enhance the quality, utility and clarity of the information to be collected?

Response

DLC suggests that the information be collected less on an industry-wide basis and more in line with the peer groups that NEI has established for Maintenance Rule implementation. It is believed that the utility of the information would be improved and a standardized group of systems and data to be collected could be established.

This is also supportive of the benchmarking effort unde1way by the Westinghouse Owners Group.

Question:

How can the burden of the collection of information be minimized including by using automated collection techniques?

Response

DLC has no response at this time.

• The scheduled January 1, 1997, implementation date for the proposed rule is not practical. Page 5-1 of the Regulatory Analysis Draft provides milestones that would require publishing the final rule and Regulatory Guide in October 1996. It appears from the milestones that the NRC may have difficulty meeting this schedule. In addition, the proposed rule would be implemented at a time when Maintenance Rule implementation and refinement will be competing for utility resources.

• The proposed rule and draft Regulatory Analysis express concern for scheduling outages of trains at power that may put the plant in an unacceptably high-risk situation. The NRC proposes reporting instances of two or more trains of equipment, from the same or different systems, being concurrently unavailable.

This is precisely what the Maintenance Rule seeks to control under paragraph a.3 of the rule. DLC believes the NRC concern for managing risk is adequately addressed under the Maintenance Rule; therefore, the NRC should reconsider the need for this rule based on this justification.

(Continued)

Comments on 10 CFR Part 50 Proposed Rule, and the Associated Regulatory Analysis Page4

• Another NRC concern stated in the background for the proposed rule is that individual plant PRAs be improved and updated with plant specific data. The Maintenance Rule Program, as implemented at Beaver Valley Power Station, will provide for collection and feedback to the PRA for each unit. DLC believes the NRC concern for improving the quality of PRAs will be adequately addressed by the Maintenance Rule.

• Finally, in an NEI (Joe F. Calvin) letter to the NRC (James M. Taylor) dated June 7, 1995, NEI expressed the opinion that the Regulatory Analysis was insufficient to support the proposed rule. NEI also questioned the NRC staffs determination that the Backfit Rule does not apply. DLC agrees with this comment and believes it is still valid.

The NEI letter stated, "The industry's revised Safety System Performance Indicators (SSPI) provide for the compilation of selected industry performance data." Furthermore, "Providing SSPI data to the NRC on a plant-by-plant basis, but without attribution to each particular plant, would support generic licensing actions as well as monitoring industry-wide trends and performance."

DLC supports this position. The data the NRC seeks is not available in many cases.

It is not a matter of collecting pre-existing data and creating a report. Most utilities, including Beaver Valley Power Station, will have to begin collecting the majority of this data. If data collection changes must be made, the SSPI Program should be modified to include the minimum set of data for the minimum number of systems.

ATTACHMENT 2 DLC Comments on 0MB Supporting Statement

1. The burden estimates made in the Office of Management and Budget (0MB)

Supporting Statement are dependent on the following assumption made on page 3-2 of the Draft Regulatory Analysis:

"For the purpose of this regulatory analysis, it was assumed that 80 plants would already be collecting similar data and 30 would not."

Duquesne Light Company (DLC) believes that the Nuclear Regulatory Commission (NRC) may be overestimating the availability and reliability data that utilities are collecting in support of the Maintenance Rule. While 80, or more, of 110 plants are collecting various forms of availability and reliability data, it is believed that the data is primarily failure oriented, not success oriented as is the information requested by the proposed rule. DLC believes that few plants are planning to collect data similar to that requested by the proposed rule.

If the quoted assumption is inaccurate, then the burden estimates for implementation in the "Recordkeeping Burden Table" would be underestimated.

2. The implementation burden estimate of 175 hours0.00203 days <br />0.0486 hours <br />2.893519e-4 weeks <br />6.65875e-5 months <br /> to "Develop data collection program" also does not seem to be realistic considering that program development will also include: review of the proposed rule, review of a new Regulatory Guide, possible review of a new Nuclear Energy Institute (NEI) guideline and development or revision of administrative procedures to implement the new regulation.

3. Page 3-10 of the Draft Regulatmy Analysis indicates the NRC expense for conducting "about two workshops" on proposed rule implementation. The burden estimates apparently do not include the utility expenses that would be incurred in supporting these workshops.

4. The Recordkeeping Burden Table indicates a significant "Annual (recurring)"

burden estimate difference between the 80 plants and the 30 plants. DLC questions why a burden difference should exist on a recurring basis once all plants have developed their programs.

5. Paragraph 12 on page 4 of the 0MB Supporting Statement describes the implementation burden as being "annualized over three years." Given the current schedule of implementing the proposed rule on January 1, 1997, it would seem that the entire implementation burden would be incurred in 1996 in order to begin data collection by the scheduled date.

(Continued)

DLC Comments on 0MB Supporting Statement Page 2

6. Having done a preliminary review of all the information provided, DLC suggests that the best way to minimize the regulatory burden for the NRC and utilities is to postpone issuance of the proposed rule and concentrate on enhancing the Institute of Nuclear Power Operations (INPO) Safety System Performance Indicators (SSPI)

Program to meet the NRC needs. The current SSPI systems could be supplemented from the "Basic Systems List" contained on page 5324 of the Federal Register, dated February 12, 1996. Modifications to the INPO SSPI Program should also be minimized. If this approach is taken, the NRC and utility burden for collecting improved industry-wide data on the performance of risk-significant systems will be minimized.

Station Support Department DOCKETED USNRC PECO NUCLEAR PECO Energy Company Nuclear Group Headquarters A UNIT OF PECO [,'J[R(,Y 965 Chesterbrook Boulevard

• 95 JUN 14 P3 :OS Wayne, PA 19087-5691 OFFICE OF SEC ,ETARY DOCK ET 11f3 & SERVI CE BRANCH June 11, 1996 Docketing and Services Branch *- .......

U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 oocmNtUEAPR pROPOSEDRULE~--115'a----

Subject:

comments Concerning 10CFR50 Proposed Rule, "Reporting Reliability and Availability Information

( b\ fR-6~ \-S \'\

for Risk-Significant Systems and Equipment" (61 FR5318)

Gentlemen:

This letter is being submitted in response to the NRC's request for comments concerning the 10CFR50, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment," published in the Federal Register (i.e., 61FR5318, dated February 12, 1996).

PECO Energy appreciates the opportunity to comment on this Proposed Rule which would require licensees of commercial nuclear power reactors to report plant-specific summary reliability and availability data for risk-significant systems and equipment to the NRC. This Proposed Rule would also require that licensees maintain on site, and make available for NRC inspection, records and documentation that provide the basis for the summary data reported to the NRC. PECO Energy considers that this Proposed Rule may be premature, and recommends that NRC reconsider promulgating as a final rule. In addition, we fully support the Nuclear Energy lnstitute's (NEl's) position and comments with respect to this Proposed Rule.

If you have any questions, please do not hesitate to contact us.

Very truly yours, t.~~::r*P.

• Director - Licensing

,\cknowfed ed b --iuN 1 9 I99l ..

g y card ..............................t"II"

U.S. NUCLEAR REGULATORY COMMISSIOl'-

OOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Document StatisticS Postmark Date ~.LI4-->-=- - - - -

Coples Received_.1--- - - - - -

.~ Copies Reproduced -:::::,l.._ _ __

Spec1al OistributionCW,~ -~ ,

DOCKETED USHRC

• 96 JUN 14 P3 :06 Tennessee Val ley Authori ty, 1101 Market Street. Chattanooga, Tennessee 37402-280 1 June 11, 1996 U.S. Nuclear Regulatory Commission ATTN: Docketing and Service Branch Washington, DC 20555-0001 Gentlemen:

COMMENTS ON THE PROPOSED NEW REGULATION (10 CFR 50.76) - REPORTING RELIABILITY AND AVAILABILITY INFORMATION FOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT We have reviewed the subject proposed regulation, and our specific comments are provided in the enclosure to this letter. TVA supports the Nuclear Energy Institute's position which concludes that the proposed rule is not needed for assessing compliance with any current regulatory requirements and would impose a significant burden on the industry beyond current data collection. We believe that the proposed rule is premature and should not be promulgated.

If you have any questions regarding our comments, please contact Mike Hellums of my staff at (423) 751-2695.

General Manager Nuclear Assurance and Licensing Enclosure cc (Enclosure):

Nuclear Energy Institute ATTN: Mr. Tony Pietrangelo 1776 I Street, NW, Suite 400 Washington, DC 20006-3708 4cknowt9dged by card ~~

• • • • • • • • • • • • • • • • • • • . ***~'-.......

~

U.S. NUCLEAR REGULATORY COMMISSIOt--

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Doc\Jnent Statistics Po~ Dale -----1~---------

Co?tS Received_-1--- - - - -

,\d<fl Coples Reproduced . .: : . t.- ---

Speola OlstnbJtion ~ilW.wl!SZ:!:l.._...t::::!:..,..

ENCLOSURE COMMENTS ON PROPOSED REGULATION 10 CFR 50.76 The fundamental application for the Rule is to improve NRC's ability to make risk-effective regulatory decisions consistent with the Commission's policy statement on the use of probabilistic risk assessments (refer to August 16, 1995, 60 CFR 42622) which we clearly support. However, all NRC justifications for the Rule, relative to supporting risk-based regulation, can be met through existing requirements for data reporting contained within the Maintenance Rule or the Institute of Nuclear Power Operations (INPO) Safety System Performance Indicator (SSPI) reporting program. Additional and redundant equipment performance data reporting is a significant burden to the licensee without apparent compensatory benefit.

TVA has spent between $2.5 million and $5.0 million per site to develop and enhance their Individual Plant Examination model to more accurately reflect the design and operation of the units and to incorporate site-specific performance data as opposed to generic industry data. Data being provided under the Maintenance Rule supports TVA's Probabilistic Risk Assessment (PRA) updates. Additionally, the redundant reporting of the Rule requires far more detail than the Maintenance Rule. It is estimated that an additional 4,300 annual man-hours per unit will be required for TVA to comply with the Rule above what is being expended for compliance with the Maintenance Rule and the INPO SSPI reporting requirements. It is also anticipated that none of these additional man-hours will provide value add to plant availability, reliability, or safety.

The Rule would require data in a format which supports classical PRA models which utilize "demand reliability" or "run reliability" (e.g.,

failure per demand or failure per hours of operation following successful starts). Most existing PRA techniques take individual components and model them into functional equipment groups in an attempt to represent the safety function important to mitigating core damage. The data (e.g.,

unavailability and demand or run reliability) is then used to calculate the probability that the equipment in a functional failure group will be unavailable to mitigate a transient possibly resulting in core damage.

The modeling of equipment reliability to determine function reliability is not required. The affects of reliability on core damage can be modeled in terms of "functional failure rate." The industry monitors reliability of its Maintenance Rule defined functions predominantly through the use of functional failure rate (e.g., failures per operating cycle).

The Maintenance Rule monitors the performance of the function itself based on the performance of its components; that is, data is only taken for the loss of the function and not all successful equipment operations.

This approach reduces the resource burden considerably, since l

alternatively, 90-99 percent of the data retrieval process for "demand or run reliability" is to obtain nonfailure data which has no corrective actions associated with it. The Maintenance Rule directly couples failures with corrective actions and trends the success of the corrective actions. This approach produces direct and immediate response to affect improving reliability as opposed to producing a statistical data base which would require years to trend.

The use of the Rule to "enhance licensee"s capabilities to implement the evaluation and goal-setting activities required by the Maintenance Rule by providing licensees with access to current industry-wide reliability and availability information for some of the systems and equipment within the scope of the Maintenance Rule" is not required. This is being accomplished through the use of the industry Operating Experience Programs, required by NUREG-0737, and INPO SSPI and Nuclear Plant Reliability Data System (NPRDS) data bases. It can also be supported through the use of the individual utility Maintenance Rule data base.

If the Rule becomes law, it is recommended that the need for continued NPRDS reporting be evaluated.

2

Performance Technology PPGtSfIED PO. Box 51663 , Knoxville, Tennessee 37950-1663 Phone : (423) 588-1444 , Fax (423) 58JJso~§ C

'96 JUN 13 P3 :41 June 9, 1996 DOCKET NlllBERM ._____

U. S. Nuclear Regulatory Commission PROPOSED RULE m 59 .;

ATTN: Docketing and Service Branch Washington, DC 20555-0001

. (Cs,\Fe5.5\~") @

\1 Attached are my comments on the proposed rule 10 CFR Part 50.76, Reporting Reliability and Availability Information for Risk-significant Systems and Equipment.

Sincerely,

.SJ Bob Christie AcJmowtMllad vu~

by_... "'J!-Jrf f l 1996" ,

\iCU U **" ...""'"""...."W"llncr "When you measure performance realistically, it improves ."

7 U..:, NUCLEAR REGULATORY COMMISSIOf'.

DOCKETING & SERVICE SECTIOH OFFICE Of THE SECRETARY Of THE COMMISSION Doainent Sldllm Postmark Dale :Mnc £l -£18w ?:Urn ~

Coples Recelved,......:~------

,\d<fl Co?9S Reprodtad ..;:~:::..----

Speola llstrbJtion ()CD I .ro , ""'3>.J.ri;,I 33,;;s;,~

Performance Technology PO. Box 51663, Knoxville, Tennessee 37950-1663 Phone: (423) 588-1444 , Fax (423) 584-3043 COMMENTS OF BOB CHRISTIE Proposed rule 10 CFR Part 50.76, Reporting Reliability and Availability Information for Risk-significant Systems and Equipment.

1. The success of the proposed information data base will depend on how much feedback is allowed by the use of the information in the proposed data base. The best information data bases are those in which the collectors of the information are also the main users of the information. The main users of the proposed information should be the system engineers at the nuclear units.

In order to allow feedback for utility people to use the proposed information data base, the cycle ofreporting must be at least monthly. As time goes on and the nuclear industry gets better at reporting data, the reporting cycle should be shortened. The ultimate goal should be to have continuous, on line feedback.

The annual report as proposed in the rule will not be adequate to provide the necessary feedback. Information that is a year old at the nuclear units is not useful to personnel at the nuclear units and should not be used by anyone in the decision making process at the nuclear units.

2. The objective of the proposed information data base is to help the nuclear utilities maintain adequate protection of the health and safety of the public while allowing regulations to become more cost effective. The proposed information data base will be cost effective by itself if it can replace some of the existing reliability reporting programs such as: Nuclear Plant Reliability Data System, Safety System Performance Indicator program, emergency diesel generator reporting for the Station Blackout Rule, emergency diesel generator reporting per Regulatory Guide 1.108, and the LER system.

The proposed information data base can become the single reporting system for risk significant systems and equipment if the reporting process is done correctly.

The utility system engineers should be involved in the process of developing the proposed reporting system. The data in the proposed information data base should become the data used for monitoring the performance of their risk significant systems.

"When you measure performance realistically, it improves ."

Performance Technology PO. Box 51663 , Knoxvi lle , Tennessee 37950-1663 Phone: (423) 588-1444, Fax (423) 584-3043

3. The proposed reporting system should allow for the individual characteristics of the nuclear unit. It is not the responsibility of the individual system engineers to make their nuclear unit's systems and equipment just like every other nuclear unit's systems and equipment just to report data. Each system engineer should have models and data that are applicable to the systems and equipment at their individual nuclear unit. If it is necessary to make comparisons between systems and equipment, it should be the responsibility of the central data processing people to make comparisons on a "apple to apple" basis. This can be

• accomplished if the system engineers document the assumptions for each system at their individual nuclear unit and forward the documentation to the central data processing people.

4. The proposed rule is written as if all risk significant systems and equipment are standby systems which are not running during normal operation. Some risk significant systems are "dual purpose" systems. An example is the Makeup and Purification System at some Pressurized Water Reactors. These systems are running during normal plant operation and can be used for high pressure injection during emergencies. Such systems may have two models for measuring performance: a model for normal operation and a model for emergencies. If the system engineers are involved in the process of developing the proposed information data base and the reporting system is structured for use by the system engineers, such "dual purpose" systems and equipment will report the appropriate data.

2 "When you measure performance realistically, it improves."

DOCKETED USNRC DEf ONRUE via: (1) INTERNET 101 Federal Street (2) US EXPRESS MAIL 17th Floor

'96 JUN 13 A10 :49 Boston, Massachusetts 02110 Phone: 617/261-2400 FAX: 617/261-2449 OFFICt or s:::CR ETARY DOCKETING f ... '.°_;t" RVICE BR.4N'..H June 10, 1996 DLNRC 9601-01/SM DOCKET NlMBER Secretary United States Nuclear Regulatory Commission PROPOSED RULE PR 50 Washington, DC 20555 (let fle..5-3t 'B)

ATTN: Docketing and Services Branch Re: Reporting Reliability and Availability Information for Risk-significant Systems and Equipment, 10 CFR 50.76, (61 FR 5318-5326, dated February 12, 1996)

Sir:

In this letter, we offer comments concerning the subject proposed rule.

This rule is an important and overdue step for the agency in fulfilling its statutory responsibilities in the modern era. The public and the industry have every reason to anticipate significant benefits.

For many years, industry's agents (INPO and NEI) opposed initiatives of this type. Such opposition is misguided or misinformed. The public can expect to enjoy safe, lower-cost nuclear-fired electricity from the more open regulatory environment this rule promotes. This rule will also significantly benefit the stockholders who must compete with other generating options. In both cases, the benefits of statistically-based management -- methods proven effective in sustaining routine production of extraordinarily high quality products in competitive industries.

Presently, we cannot anticipate how many power companies will support NRC's proposal. Analysis of proposed rules is often delegated to subordinates with immediate imperatives and familiar with the issue's details. Unfortunately, such individuals' experience is often quite narrowly defined to nuclear regulation. It is unusual for such professionals to have worked in a competitive environment outside the nuclear

U.S. NUCLEAR REGU LA.TOfW COMMISSIO~

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Docunent StatisticS Pol>tnltk Date ~ ~ ~~ - - - - - -

Copies Received. _ _..,___ _ __

.\dd'l Coples Reprod~ ~ - - --

Special Distr!Wtion~ ~ =~ ~~ -

~S

DLNRC 9601-01/SM, dated June 10, 1996 industry -- especially, those involving technological processes and the highest of quality standards demanded by competitive markets. Legal analysis is of even less relevance and value to this issue. And, while some engineers claim expertise in risk modeling, too few are knowledgeable or professionally trained in statistics -- an area of applied mathematics essential to understanding the span and scope of the proposed rule. Such limitations can be a significant handicap in properly weighing the direction of NRC's proposal.

Potentially lost from such limited regulatory analysis is the practical impact of an "open systems" approach to performance data -- the consequential shift from how complex regulatory details processes can be made, to how well complex systems actually perform and what is demonstrated to most affect performance. In effect, this rule seeks the triumph of the product over the style.

Perhaps experienced executives are the only people with the requisite qualifications to consider the policy implications. However, with far too many fires to fight these days, few are likely to have the time to properly consider whether component data need be reported to NRC, as this rule will probably be described. As some might sum up: "There is no need for the NRC to look into this kind of detail. After all, that's why we created INPO more than 15 years ago."

Only a handful of innovative executives have taken the time to consider and understand the significance of the paradigm shift NRC proposes.

Times have changed, and power companies have less and less in common with each other. The good performers have little patience for the underachievers -- and see no shame in their own accomplishments and competitive advantage. In many industries, quality metrics are a competitive weapon, not something to hide. But to ensure confidence in the data reported, it must be in the custody and to the standards set by a disinterested party. Only NRC is qualified to serve that role.

In any case, we encourage NRC to stay the course and help move the industry toward a more competitive and performance-based approach -- one founded on actual statistics and proven mathematical methods.

The agency can count on at least a few licensees for quiet support of that goal.

Comment Summary Our comments are summarized below:

2

DLNRC 9601-01/SM, dated June 10, 1996 (1) The Need for a Consistent Statistical Basis NRC correctly notes the success of key initiatives in the current regulatory program (such areas as risk-based inspections, maintenance rule, and aging) depends on a consistent, high quality set of operational data.

(2) The Importance of Public Access and Independence Public access to a consistent set of operational data is essential to public confidence in agency actions -- particularly at this time with serious cracks appearing in current approaches.

(3) Only NRC Can Serve Role If performance-based standards are to supplement and displace traditional, inefficient, document-based methods, such standards must be rooted in a consistent, statistical basis. Well-intentioned industry agents, such as INPO, lack the independence and accountability to serve this role much longer -- particularly in the current era of power company competition. Only NRC brings the requisite independence and public access essential to this important goal.

We also address certain process and technical details of the rule with the goal of ensuring data quality and efficiency. Our comments include responses to specific NRC questions concerning the potential impact of information collection practices discussed in the proposed rule.

We begin our comments where NRC begins: regulatory background.

Background

Current Requirements: A Focus on Documentation Rather Than Performance For an industry so committed to the Holy Grail of self-regulation, and so quick to rely on unvalidated estimates of core damage frequency (CDF), we never understood the dearth of standards for systemic reporting and data warehousing of actual performance information. It seems to us prudence dictates a reconciliation of the millions on replacement power and capital costs in the name of CDF reduction .1 It Unless, of course, such estimates are not really used to make real decisions, and are merely regulatory artifacts.

3

DLNRC 9601-01 /SM, dated June 10, 1996 would also seem to follow that industry would want to reconcile the disparate risk estimates among substantially similar plants.

We never understood why this has yet to happen.

Industry's repeated failure to reconcile performance to risk forecast is punctuated by nuclear's rapidly deteriorating competitive position in contemporary power generation markets. Whereas in the past, operational uncertainties due to compliance problems might be attributed to "rising standards" and neatly swept under the ratepayers rug , customers now have options. Clearly, there must be fundamental change in core processes and institutions if many members of the current generation of operating reactors are to remain economically relevant.

The most important change needed at this time is in the vernacular of compliance programs -- the "lost language" of plant construction work packages as some sacred writ of "design bases." Like overturned stones at a ruin site, these engineering documents are missing essential pieces required of their regulatory role for a very simple reason -- they never existed in the first place. These documents were never intended to serve as "design bases" -- rather, they were intended to provide the minimum information necessary to procure, fabricate, assemble, and construct a power plant in an era predating computers and imaging systems.

Even if design bases restoration projects can do for nuclear plant documentation what artists did for the ceiling of the Sistine Chapel, there are many theoretical and practical problems with attempting to sustain a "zero defect engineering-document approach to safety management. Such attempts would constitute a positive feedback system -- and all such systems are structurally unstable, if not explosively so.

Consider the system dynamic. Most design bases "reconstructions" tend increase the document and cross-indexing count. With each new document and cross-index, the system's information entropy grows.

With greater entropy comes increased likelihood of transcription error, inconsistencies, and gaps. Error feeds back on error. If the solution to such problems is document rework, still more documents are produced with still more cross-indexes. Unchecked, the positive feedback loop closes, and system is set to detonate with a time constant varying from weeks to years.

We see introduction of statistical methods as an attempt to check the natural tendency for systems to "weed" -- supplementing the unstable dynamics of design documents with the simpler data structure of the measurable record.

4

DLNRC 9601-01/SM, dated June 10, 1996 By standardizing, emphasizing, and publicizing performance data as data structure primitives, NRG's proposed rule seeks to enhance and modernize the language and measure of safety performance.

Ironically, it seems the industry is the reactionary, rather than the innovator. Industry position papers, public statements, and legal documents seem aimed at further locking-in the existing choking structure of document management as the coin of regulation.

Unfortunately, the practical impact of continuing current practices is that the licensee's management performance is measured by how well the documents hang together in design engineering files, less so, the material, maintenance, and operating practices at the real plant.

- There certainly are issues requiring a detailed engineering treatment as technology advances, standards evolve, and operational experience reveal weaknesses in a design approach. For example, in certain areas (e.g., fire protection, plant materials, electrical system setpoint control, and pipe codes) a standards-based approach is essential.

There are other motivations for increasing the regulatory ratio of statistics to documents. The traditional focus on engineering design for a mature technology does nothing for issues not conveniently defined in engineering and construction work packages. In these areas, statistical analysis of performance to functional goals must play the deciding or reconciliation role.

There is no other alternative. Neither NRG or the industry can much afford equating safety with how e meticulously "engineering-monks" copy over moldy, design "scrolls" and texts -- incomplete as they are, and written in a medieval period some three decades ago. Even the quickest monks can't hand copy data into "design basis documentation" that quickly. And, even if they do, so what?

To our view, no amount of FSAR rebaselining, and the extraordinary resources (NRG and licensee) committed to such epic productions, has ever been statistically correlated to positively impact on safety performance. More generally, no statistical evidence has ever been developed relating reactor safety and the quality of regulatory scribes and their design basis scrolls. At the extreme, some document rebaselining efforts even appear more ceremonial, if not ritualistic, than substantive -- a technological (and financial) holocaust offering to the icon of document control. We think both agency and industry can do better.

Then, too, there are issues where the most timely design documents are clearly irrelevant to safety. And, as NRG notes, current methods often fall short in information resolution.

5

DLNRC 9601-01/SM , dated June 10, 1996 Consider, for example, human error rates. Industrial, military, and transportation segments recognize the human prominence in accident statistics. The National Transportation Safety Board (NTSB) has long held that upwards of half of all accidents can be attributed to pilot error or result of human action. Yet, operator error or inattention does not neatly map to timely engineering files or even system unavailability statistics.

Certainly many licensees apparently act on the belief operators are never a factor in CDF estimates -- a position totally at odds with NTSB. Operator error has been totally AWOL from the bulk of CDF estimates we've seen published in IPEs, PRAs, or the nuclear literature. In the nuclear industry, failures obediently follow the imperatives of the PRA data structure: either coincidental and random, or exogenous. In the word of PRAs and human factors programs, operators never err -- (1) they abuse or don't abuse substances, and (2) they act or fail to act, always in rather naive ways.

Consider the evidence in contrast to the naive nuclear view of many PRA analysts. Anecdotally, human error figures prominently in nuclear and industrial incidents such as Three Mile Island, the 1975 Browns Ferry fire, the Chernobylsk accident, the Exxon Valdez grounding, and the Bhopal run-away chemical transient. 2* We posit one common variable for these and other serious incidents is the time of day: they all began quickly and "snow-balled" out of control in the "post-midnight-6 AM" period.

It would seem, therefore, error rates are not uniformly distributed across any 24-hour period, and do not map well into the Boolean logic used in PRA models. Focusing on the former, it seems reasonable to ask why.

The very data-rich biomedical literature suggests a physiological dynamic for event concentrations in the early morning hours: the circadian cycle.3 This hormonal cycle has been extensively documented to predict both high and low periods of alertness in the course of a day. Not surprising, a human's "low alertness" period corresponds to precisely the early morning hours when things always most often appear to go wrong.

The literature goes further, demonstrating that rotating shift scheduling practices, such as commonly used in nuclear operations, amplifies the disruptive effects of circadian cycle. Such amplification manifests itself 2

The Valdez certainly has its alcohol dimension. But, other deck officers exercised poor or inadequate judgment in lieu of the skipper's role.

3 For a very readable introduction of circadian cycles, see Schwartz, W. J., "Internal Timekeeping," Science &

Medicine, V. 3, N. 3, May/June 1996, pp. 44-53.

6

DLNRC 9601-01 /SM, dated June 10, 1996 in lower levels of alertness, microsleep episodes, and higher than normal error rates involving routine and emergency tasks.

The late US Office of Technology Assessment inventoried problems related to the circadian cycle, and the role of rotating shift-worker scheduling practices. A section examined practices at nuclear plants. 4 OTA reiterated linkages reported in the literature between rotating shift scheduling practices and tendencies towards substance abuse for self-treatment of physiologically-triggered depression, in a vain attempt to promote sleep during day-light hours, and to offset drowsiness on the backshift. 5 While the nuclear industry sees no difference between operator performance at 2 AM from performance at 2 PM, NASA does. Space shuttle mission schedules are explicitly designed around the crew's circadian cycle. Prior to launch, NASA even aligns crew alertness levels to mission schedules through "phototherapies" aimed at reducing the potential for error on the mission backshifts.

Some years ago, we performed a Fourier analysis of the circadian cycle in conjunction with researchers at Boston's Brigham and Women's Hospital. Suspecting evidence of a circadian cycles effect on reactor event frequencies, we extended the analysis to statistically test the "time of day" hypotheses using LER data.

Our results detected evidence of the familiar "early morning" clustering experienced in other industries. 6 Reactor trips and ESFAS due solely to human error (as reported by the licensee herself) are more likely to occur in the early morning hours than at any other time of the day. Time of day seems to matter as a predicator of transient frequency.

This simple anecdote highlights several points -- the limits of PRAs and work packages as a language of safety, the importance of statistical analysis, and the need to provide public access to sufficient information to address issues not captured in plant designs.

Industry Organizations Can't Create and Manage Essential Data Structures Industry systems, such as the Nuclear Plant Reliability Data System (NPRDS), and organizations such as INPO, track component engineering characteristics and failures. However, as NRC notes, such sources fail to monitor all data elements (e.g., maintenance unavailability) essential to measuring system 4

US Congress, Office of Technology Assessment, Biological Rhythms: Implications for the Worker, OTA-BA-463, Washington: US Government Printing Office, September 1991.

5 pp. 96-97, 102, 165-166, op. cit.

7

---

~----.

DLNRC 9601-01/SM, dated June 10, 1996 functionality. Maintenance monitoring requirements in 10 CFR 50.65, " Requirements for Monitoring the Effectiveness of Maintenance at Nuclear Power Plants," also fall short in this area.

There are many explanations tor why self-regulation by industry organizations has practical limits. We believe these reasons include the reality that the managing organizations (e.g., NEI and INPO) do not have an operating license or fiduciary responsibility . Consequently, the entity, its officers and representatives cannot be held accountable tor errors and omissions in data, or licensee failure to consistently implement an industry commitment.

Whatever convenience "one-stop" regulatory shopping offers the agency in policy discussions, there is not much in the way of "warranty" on the product these organizations should ever be expected to deliver.

Essential functions of safety management will never be accomplished by organizations in downtown Washington or an Atlanta suburb -- such functions can only be performed by the licensee.

To paraphrase contemporary market campaigns for financial services, NRC cannot escape the reality that safety management is essentially done the "old fashioned way" -- one license at a time.

And, licensees do not have a structure or standard tor consistent treatment of statistical data used in the regulatory process. Not surprisingly, GIGO.

The penalty for relying on industry's casual approach to structure is driven home by the Individual Plant Evaluation (IPE) experience.

Years of discussions with industry's agents set the expectation that IPE would be a simple, "easy to use" process that would quickly and efficiently identify risk outliers for modification. With great fanfare, Generic Letter 88-20 and its supplements hit the streets.

Unfortunately, no one bothered to standardize the data definitions, statistical methods, and data structures used in the PRA models. Almost a decade and many million of dollars later, IPE results are all over the map. Seemingly identical plants lack any glimpse of mathematical essentials of related topologies.

The mathematical fixed points are hidden from view, if they exist at all.

6 Maloney, S., "Evidence of Circadian and Extended Shift Effects on Reactor Transient Frequency,"

Transactions of the American Nuclear Society, V. 65, (1992) pp. 512-513.

8

DLNRC 9601-01/SM, dated June 10, 1996 So, what happened? Is the wide dispersion in IPE results indicative of a pattern, or a lack of pattern? Is the problem in the method, or the data? Do IPE results meet a precision criterion? Which IPEs are good, which ones are partly good, and which ones are worthless?

No one knows.

Because the data foundations were never properly established and consistently enforced, the IPE structure was erected on the poor footing of an undrained data swamp. For all the extraordinary resources expended by all parties, !PE appears to be just the latest analytical disaster.

Some may try to dress-up the IPEs as instructive and useful. We disagree.

If usefulness and confidence is any measure, we look to the power companies' plant managers and NRC's resident inspectors who live safety responsibilities, on the line, 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> each day, 7 days each week. We are unaware of any plant manager familiar with his plant's IPE results or who routinely checks its results and models for guidance or "insights." We are certainly not aware of any rebaselining efforts aimed at reconciling IPE assumptions to detailed design features at the real plants.

If regulatory documents are suspect, then the even sparser and more artificial IPE assumptions can't be much better.

Fortunately, IPE's wreckage is not likely to be around very long before the tide comes in and washes this latest sand castle out to sea. But, the IPE experience can still be instructive. The lesson we draw from IPE's failure is that data structure design and preparation must precede analysis -- not the other way around.

Industry had its shot at accomplishing this task, and failed to execute. Perhaps it's time NRG steps in to set and enforce consistent standards.

Industry Resistance to Consistently Define the Underlying Data For reasons that continue to escape us, industry's agents or processes seemed to stonewall publication of data standards and results that would benefit the licensees. As Pogo might say, we have met the enemy and it is us.

9

DLNRC 9601-01/SM, dated June 10, 1996 About 5 years ago, we watched NRC staff participate on an INPO-established review group to enhance the NPRDS. The result was a reasonable first step. Quite familiar with industry's knee-jerk fears, however, we were not surprised the group's recommendations went nowhere.

About 2 years later, NUMARC/INPO proposed to adopt the Safety System Performance Indicator (SSPI) as a surrogate for reliability data. This proposal seemed too little, too late. Even though NUMARC offered the ability to extract actual reliability and availability data, the data structures and primitives were to remain hidden. In any case, the issue was moot. As NRC notes in the NOPR, " ... no voluntary system of providing data resulted from these discussions." Once again, despite good intentions, voluntary processes are only as good as the warranty -- which is not much.

In fact, as we previously stated, industry's agents do not have an operating license and are not accountable to stockholders. In matters of public confidence, essential information, and regulatory requirements, there really is no place for voluntary actions , and processes providing mere brief glimpses of basic performance information.

In the NOPR, NRC explains the reason for the stonewall: reliability data, if publicly available, would be subject to misuse. If, indeed, this is the issue, the horse left that barn a long time ago.

For years, unvalidated estimates of reactor risk, core damage frequency, and radiological dose from hypothetical accidents flowed like a river into countless reactor dockets. Such estimates were used to explain away deviations from operating licenses and design standards, and justify continued operation when an LCO was indicated. Based on our expert professional opinion in these matters, there was a substantial amount of chaff cluttering the screen -- enough statistical errors among enough licensee filings over the years to make INPO/NEl's concerns ring hollow.

If INPO/NEl's concerns are to preclude misuse and misunderstanding by everyone, there's every incentive for NRC to address the issue head-on, rather than hide the data in a closet. Certainly, public airing is one benefit from the proposed rule and its related regulatory guide. NRC management would also relieve the industry's agents from the impossible task of being "always on the point -- for the very snipers INPO/NEI fear are out there ready to bushwhack them as soon as plant data goes public.

Truly proprietary component performance data, if there really is any, does not have to be a problem, either.

Policies are in place limiting access to information that satisfy strict tests related to trade secrets, privileged or confidential commercial or financial information. Of course, if such information is truly proprietary, licensees should already be providing special internal controls and security as specified in federal and 10

DLNRC 9601-01/SM, dated June 10, 1996 state laws governing intellectual property and confidentiality. Otherwise, the requested protection would be a canard.

We agree with NRC on this one. The data the agency seeks under this proposed rule does not appear to qualify for withholding.

We also argue the agency's purposes -- as well as licensee interests, notwithstanding the views of INPO and NEI -- are best served by transferring control from industry organizations to the public domain. In particular, reliability data and attendant statistical methods used as input to risk-based regulatory decisions must be clear and accessible to the public.

Move to Risk-Based Regulation Both industry and the agency are interested in using PRAs as a tool for power reactor regulation. We have our own views about PRA's role, and believe a statistical approach should displace the probabilistic, wherever data is available.

In any case, we agree with NRC that any initiative involving probabilistic or statistical analysis requires enhancement in the quality and consistency of equipment and human reliability data.

NRC also correctly suggests information may have existed which, if collected, recognized, and acted upon might have led to preventive actions. We agree access to such information would sharpen NRC's oversight capability, and ensure its actions are more focused on where the problems might be. In other words, rather than necessarily acting solely on judgment to shut down a unit and require a regulatory rebaselining, a statistically-based approach would more efficiently focus resources on the more significant problems. At least, our experience in statistical process control (SPC) techniques in industries demanding quality in product seems to support this expectation.

We suspect many executive committees would rather see their regulatory grade rise or fall on measurable terms than on a "beauty contest" concerning the work of their scribes and whether inconsistencies can be found among the many volumes of regulatory scrolls.

As for the importance of relating process to data, industry's silence is deafening. Many speak of the importance of risk-based regulation, but when it comes to validating results, the methods and data seem to slip just out of reach.

11

DLNRC 9601-01/SM, dated June 10, 1996 Clearly, NRG is the innovator in this area. Staff speaks in terms of process rather than content: the need for a framework, the importance of procedures, the role of credible models -- the essential ingredients of a process. Surely, the details need to be worked out, but it is encouraging to see safety regulation broaden from lawyer hair-splitting and "regulatory bases" towards results that can be measured and validated.

If NRG is to avoid repeating the IPE debacle on a grander scale, it must fix a firmer data foundation for risk-based management. We see this proposed rule as an important step in that direction.

Generic Issue Resolution We agree with the agency's proposal to use risk estimates in: (1) prioritizing safety issues, (2) deciding whether new requirements or staff positions to address these issues are warranted, and (3) deciding whether proposed new requirements or staff positions should be implemented.

However, we advise a more balanced approach be used in the area of PRA. In our view, PRA results are not a performance metric; they are merely a "risk forecast -- a contemporary statement about known accident sequences, and a structured estimate concerning unknown risks. We believe strongly in the need to routinely refine forecasts, and feedback statistical data and methods. Otherwise, even with better quality component and system data, PRAs will continue to be fabricated with a very murky footing in reality.

A frame of reference is essential to any model. Fortunately, reactor accidents are rare events. We would like to make them rarer. Hence, to validate a PRA, we must go back to a statistical treatment of system performance data.

Certainly, validating component data used in a PRA is an important ingredient in refining a risk forecast.

But, it is also essential the PRA "computer be correct. Statistical analysis of system availability data against published probabilistic models should go a long distance to establishing consistency and confidence in the current unvalidated results. We assume NRG sees the same goal and already has it on its "target list."

With such validation of incipient events and system reliability models, rebaselining and recalibration of PRA results can proceed. We are confident statistically-based indicators of plant performance and trends can replace some of those currently used to identify plants for increased or decreased regulatory attention.

Accident Sequence Precursor (ASP) and Event Analysis We have no special views of the importance of the ASP and Event Analysis Program. For reasons already discussed, it seems obvious to us a statistical process and statistical quality method would more efficiently 12

DLNRC 9601-01/SM, dated June 10, 1996 validate plant-specific ASP models used for determining the risk-significance of operational events and whether further actions are warranted.

Risk-Based Inspections It has never been clear to us how risk-based inspections would work absent hard, consistent, publicly available data.

Lacking statistical validation, any inspection topics and timings specified by a PAA are suspect. PAA models map closed sets to closed sets. What does one do about issues not explicitly addressed in that model (e.g., the human error question)? What about things that do not neatly map to the model (e.g.,

accident sequences not envisioned)? How does one know the model is reasonably accurate?

If PRAs are to be used for these purposes, the models and data must satisfy a performance criterion rooted in reality . Consequently, IF industry gets its wish to use PAA-models for risk-based inspection, THEN the data must be used in statistical process control terms.

In reviewing NRC's discussion about risk-based inspections, it also seems plain to us SPC methods commonly used outside the nuclear industry may be more appropriate than the probabilistic approach used in PAA. Remember: PAA was developed some 20-30 years ago precisely because there was no data, at that time.

There sure is a lot of data right now.

fillirul Probably one of the more important issues this rule may help address is component aging.

In the 1960s, there was no technical basis to conclude nuclear reactors would see the 40-year licensed life with the bulk of original components intact. In fact, despite the press releases of the period, anything beyond 20 years was clearly known to be a real push, even then.

History has borne out the technical literature of that period. Regardless of what people knew and when they knew it, it is evident few steam generators will see even 20 years service without a major refurbishment or replacement. Electrical cable insulation and jacket materials degrade at rates driven by ambient temperature, chemical exposure, and local stresses -- suggesting turnover every decade or so.

Most utilities now employ statistics to predict the likelihood of premature catastrophic pipe failure -- and base their capital programs on those results.

13

DLNRC 9601-01/SM, dated June 10, 1996 Consequently, except tor the carefully-worded industry position papers, it requires some effort to keep a straight face in claiming any "current plant will make 40 years as-built. We are encouraged the agency, at least, isn't playing this game, and is interested in statistically forecasting component degradation and the risk-significance of aging.

The safety benefits are self-evident. From a stockholder and ratepayer perspective, we've waited too long for this data tor all to see in honestly weighing the business case, as well.

Risk-Based Technical Specification and In-Service Testing The statistical bases tor component reliability would provide benefits in technical specifications, LCOs, and surveillance.

About a decade ago, we used such methods on industry's behalf in developing NUMARC 87-00 to address the station blackout rule. Among other things, fairly rudimentary statistical analysis of operational data we performed a decade ago showed no benefit from routine, cold, "fast-starts" of emergency diesel generators. Accordingly, licensees committed to NUMARC to halt such practices some 1O years ago, and tile for changes to technical specifications. There is no reason for any plant to be performing tests in this manner today.

It was not a PRA, but a standard statistical model that made this case.

Such examples demonstrate the benefits of statistically-based approaches using public data in halting arbitrary surveillance practices whose actual performance demonstrate minimal or no benefits. It would also provide the statistical bases for streamlining ANSI codes and standards, as some societies are currently doing.

We agree with NRC, and fail to see any reason not to implement this rule.

NRC Maintenance Rule Frankly, if the maintenance rule, 10 CFR 50.65, is ever to be more than simply another documentation program, the tenets of SPC/SQC must be introduced into operational practice. SPC/SQC begins with the data. We concur with NRC's objectives in this area.

14

DLNRC 9601-01 /SM, dated June 10, 1996 The Proposed Rule The proposed rule would require holders of operating licenses for nuclear power reactors to report reliability and availability data for certain risk-significant systems and equipment.

We have no particular complaint with NRG's proposed data structure. Our experience with data structures related to performance measurement is that they tend to interact with operational goals in a self-reinforcing way. Wherever the agency starts, we expect revisions over time and convergence to a more efficient process.

We make 3 specific suggestions to assist both NRG and industry in this important endeavor:

(1) time of day should be introduced as a parameter (2) statistical methods used in calculations need to be described (3) key statistical measures and uncertainties should be reported with the requested information.

Time of Day Should be Introduced as a Parameter Previously, we noted a circadian cycle may be implicated in reactor trip and ESFAS statistics. Our example highlights the role of "time of day" can play in isolating factors not identified in subjective engineering analysis.

In the particular case of the hypothesized circadian linkage, time of day may also provide a statistical relationship between equipment performance and human performance. From experience we suspect "time of day" and date are useful in isolating temporal associations and trends -- the first step in identifying performance issues not initially suspected or assumed in event scenarios.

Statistical Methods We previously commented on the naive misuse of data and statistical applications in reliability statistics and PRAs on the dockets. We view such errors to be comparable to running, say, a RELAP code without proper validation .

Since NRG and licensees expect to use statistical methods to support regulatory compliance and policy issues, we suggest any statistical methods be subject to the same controls as applied to engineering and 15

DLNRC 9601-01 /SM, dated June 10, 1996 design calculations. Otherwise, rather than improving data quality, NRC will simply get more of the correct range and quality in data analysis practices often employed by industry.

Key Statistical Measures and Uncertainties Should Be Reported Key statistical measures related to each data vector and matrix should be provided at the same time, using the methods already described. Thus, a measure of quality and variation can be assigned to each parameter reported -- and analyzed accordingly. Otherwise, the "bad data" will drive out the good data.

Response to Specific NRC Questions

1. ts the proposed collection of information necessary for the proper performance of the functions of the NRG, and does the information have practical utility?

For reasons discussed above, we believe the proposed collection of information is essential to NRC statutory responsibilities. Its collection is long overdue. We also believe the information has practical value to business decisions.

2. ts the estimate of burden accurate?,

3. ts there a way to enhance the quality, utility, and clarity of the information to be collected?

We suspect NRC's burden estimate is a reasonable first approximation. However, we offer some additional considerations on the issue of burdens.

Organizations make processes as efficient as incentives motivate and disincentives deter. Consider the case of a typical "rebaselining effort associated with an extended regulatory shutdown. As a practical matter, it sometimes seems the performance metric of such an effort is the quantity of information assembled and the money spent, with more of both being better. Hence, the longer and more expensive the effort, the more people involved, the more paper assembled, "clearly," the better the remedy --

beyond a certain point, the "sand castle" of cross-indexed "design bases" actually constructed becomes secondary and of declining relevance with time.

16

DLNRC 9601-01/SM, dated June 10, 1996 Such data collection efforts as regulatory shutdown projects are the lower bound of resolution and the upper bound of inefficiency. Still, they set the context for all other regulatory data collection processes.

More generally, regulatory data collection "fills out" to the effort budgeted -- reminiscent of all cost-benefit ratios and Nash equilibria. From such dynamics, processes will always cost more than estimated for data quality and quantity less than requested.

Consequently, we can assure NRC that any first-order estimate of data quality and process costs will be off, but still within acceptable ranges.

Now, if the goal is efficiency, an entirely different business model is needed -- one with clear data collection processes directly related to product and incentives.

Consider the experience in competitive industries, such as electronics assembly, wafer fabrication, and semiconductor manufacturing. Rather than a source of cashflow, costs incurred to ensure product quality directly impact the bottom line. Only in the nuclear power industry is regulatory cost incentivized to be inefficient. The reason: the cost recovery formula.

For example, Motorola's quality control efforts related to operation of, say, its multibillion dollar Austin MOSS-11 tab facility, directly impacts cash margins. However, these costs are less than the revenue erosion associated with an inadequate product. Therefore, Motorola has a direct financial incentive to maintain quality to performance standards -- and every incentive to do so as efficiently as possible.

- To a Motorola or an Intel, it's not how pretty the documents and records are -- it's how well the company can capture performance in the form of quality statistics of the product itself. For chip manufacturers, there is no incentive to employ engineers to lay in gold-leaf on carefully copied technical scrolls -- or any attempt to randomly select parameters to check -- or do the classic 100% verification approach. Proven inference techniques serve this function more accurately (due to sampling randomization) and at lower cost.

Not surprisingly, in competitive industries, processes are measured, quantified, automated and streamlined to deliver the requisite results at least cost. A common saying in the semiconductor industry is "cost walks on 2 feet." Hence, the fewer engineers individually tailoring systems, the more the automation of data collection, analysis and process control -- the cheaper AND better the product.

We think of safety as a product attribute no differently that material purity or any other product standard.

17

DLNRC 9601-01 /SM, dated June 10, 1996 So, for the agency to reduce the burden and enhance the quality, market incentives should accompany the data standardization NRC proposes to establish for the industry.

While this topic may be outside NRC's traditional jurisdiction , we generally suggest the sooner such data collection activities are removed from the rate base, the faster and greater the incentives for licensees to stop philosophizing about how things are done or how many people can be brought onsite -- and the more efficient and safely processes are run.

4. How can the burden of the collection of information be minimized including by using automated collection techniques?

Until market forces fully impact the industry, the agency must create market incentives as best it can through requirements , specifications, and process standardization.

To this end, we suggest the regulatory guide be more specific concerning data structures, formatting, and transmissions mechanism.

We further suggest specifying automated data collection processes be used from utility client-server networks and operating floor data collection sites. We also suggest imaging, video, and Web-based technologies be employed wherever possible.

- Finally, we recommend the agency consult with true experts in this field. Not people a chapter or two into the "book," or last year's PRA experts recycling themselves with a new proposal. Rather, those with production experience in other industries and applications handling comparable problems using mathematical methods.

We will provide additional comments separately in reviewing the draft regulatory guide.

Conclusions This is an important issue for the agency and the industry. How it is implemented could have significant impacts on both reactor safety, and the industry's structure in an increasingly competitive market.

18

DLNRC 9601-01/SM, dated June 10, 1996 We recognize traditional institutions and regulatory concepts are ill-equipped to independently analyze and implement the bold new direction proposed by the NRC. No doubt, there will be some who lack the experience and the professional training to envision a world most other industries lived in for many years.

Unfortunately, the market will not wait for a cushy introduction and re-education of a generation of managers and engineers. The game is already on. For this reason, we encourage the agency to aggressively implement the standards, structures and methods essential to ensure safe and efficient reactor operations in the current era. We have every confidence this industry will be up to the task. With the proper incentives, it may even learn to do so more efficiently.

We appreciate the opportunity to provide these comments, and wish the best to both NRC and the industry in the difficult days to come.

19

Omaha Public Power District DOCKETED U NRC 444 South 16th Street Mall *95 JUN 13 A9 :OS Omaha NE 68102-2247 June 11, 1996 LIC-96-0069 U. S. Nuclear Regulatory Commission DOCKET NLUBERPR '

PROPOSED RULE . .

ATTN: Docketing and Service Branch Washington, D.C. 20555-0001 (~\A2-.5~\~) @

References:

1. Docket No . 50-285

2. Federal Register, Vol. 61, No. 29, dated February 12, 1996, (61 FR 5318)

3. Draft Final Report, Pilot Study on Implementation of Proposed Reliability Data Rule Reporting Guidelines

4. Draft Regulatory Guide DG-1046, Guidelines for Reporting Reliability and Availability InformaUon for Risk-Significant Systems and Equipment in Nuclear Power Plants

SUBJECT:

Comments on Notice of Proposed Rulemaking, 10CFRS0.76, Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment The Omaha Public Power District (OPPD) has reviewed Reference 2 regarding the proposed rulemaking on reporting reliability and availability information for risk-significant systems and equipment. Specifically, the proposed rulemaking requires reporting of detailed information on demands, failures, unavail-abilities and operation associated with basic systems and those systems/equipment determined to be risk-significant by the utility. The purpose of this letter is to express OPPD's continued concerns with regard to this proposed rulemaking.

In August of 1995, OPPD volunteered for participation in an NRC sponsored Pilot Study conducted by Scientific Applications International Corporation (SAIC) to identify the problems, issues and concerns associated with implementation of the regulatory guidance document associated with this rulemaking. OPPD supplied comments which were included in the Draft Final Report (Reference 3). Review of this rulemaking and the associated Regulatory Guide indicates that OPPD's comments which were included in the Pi 1ot Study were not given adequate consideration in the development process. In addition, to our knowledge, the SAIC final report has yet to be made public, which would substantially improve the knowledge of other utilities regarding this issue.

45.5124 Employment with Equal Opportunity

U-.S, NtiCLE,.::* ., . _*~i-..

OOC ETING SER  :'fW'M OfflCE Of THE SECRET~A'1 OF THE COMMISSION Do<unent Statlatlct Poltmafk Da taRece

U. S. Nuclear Regulatory Commission LIC-96-0069 Page 2 OPPD concurs with the comments previously submitted by the Nuclear Energy Institute (NEI) in their March 12, 1996 letter regarding the significant burden which the rulemaking will place on the industry if the rule is implemented in its present form.

The Maintenance Rule (10CFRS0.65), which is effective July 10, 1996, is designed to improve plant perfonnance through monitoring of equipment performance against established perfonnance criteria. There is some concern that the Data Rule will increase the focus on the accuracy of the data being reported, rather than on the use of the data to improve plant performance. The establishment of a specific list of equipment for which data is reported will result in an increased focus on these systems rather than on the poor performers identified by the Maintenance Rule processes.

OPPD understands that the NRC perceives a need for the gathering of industry-wide perfonnance data to support risk-based regulation. However, the NRC has not yet made it clear to the industry what the NRC plans to do with the data collected under the proposed Data Rule. It is inappropriate for the NRC and utility resources to study data without a clear published objective. There is the danger that raw data, manipulated by the NRC or being publicly available to other agencies, could provide erroneous conclusions with respect to safety, causing utilities to apply resources to address such issues and potentially taking attention away from real safety issues.

As the proposed rulemaking does not eliminate existing data gathering requirements but rather proposes additional data gathering burdens, the proposed Data Rule is not cost beneficial/effective. Because of OPPD's methods chosen for Probabilistic Risk Assessment (PRA) and the Maintenance Rule, the Data Rule would provide an additional unnecessary reporting requirement. The NRC has seriously underestimated the number of labor-hours required for infonnation gathering under the Data Rule. Also, the definitions provided in the proposed Data Rule for trains, failures, demands and other items (as defined in the draft Regulatory Guide) are different than the already developed and accepted terminology of the Maintenance Rule. This apparent lack of coordination between the Maintenance Rule and the proposed Data Rule demonstrates that the Data Rule is premature.

As a result, the overall impact on Maintenance Rule implementation efforts would be negative in that the proposed requirements would hamper programs currently in place by requiring significant and costly changes in established data gathering techniques.

The Maintenance Rule currently requires that data be kept to determine the placement of component groups or trains into one of two categories:

1) Category (a)(l), structure, system or component (SSC) that does not meet its prescribed performance criteria, or

2) Category (a) (2), SSC that does meet its prescribed performance criteria and is subject to nonnal preventive maintenance monitoring.

U. S. Nuclear Regulatory Commission LIC-96-0069 Page 3 OPPD considers the FCS plant-specific implementation of the Maintenance Rule and living PRA program sufficient for nearterm and future risk-based regulation applications. If the NRC expects all plants to perform data collection as OPPD does, there a re more efficient ways other than the Data Rule to make these expectations clear.

OPPD has the following specific comments with regard to the proposed rulemaking:

lOC FR50. 76 ( b) (1) ( i )

Each licensee wi 11 submit an annual report containing "the number of demands, the number of fa;lures to start assoc;ated w;th such demands, and the dates of such fa;lures, character;zed according to the ident;fication of the train affected, the type of demand (test, inadvertent/spurious, or actual need), and the plant mode at the time of the demand {operat;ng or shutdown); 11 OPPD's Comments:

Early in the implementation process for the Maintenance Rule there was extensive discussion among the affected utilities regarding the collection of demand information to support the SSC's performance assessment and 11 11 associated PRA applications. Development efforts in this area found that collection of this data was extremely difficult and not cost beneficial.

Specific demand data as required by the proposed Data Rule is not current-ly being gathered by many operating facilities. Since this additional information would have to be gathered by Control Room personnel, there is much concern that the data collection efforts required would distract the Operations personnel from their primary focus, which is the safe operation of the plant. During the performance of the Data Rule Pilot Study, it was found that the PRA could be adapted to achieve the same risk assessment objective using failure to start on demand and unavailability information currently recorded in support of Maintenance Rule objectives.

It is OPPD's position that the proposed rulemaking should require reporting of data only on "demand failures" as already collected in support of the Maintenance Rule and avoid additional burden on control room personnel.

10CFR50. 76 (b) (1) (iii)

Each licensee will submit an annual report containing "the number of hours equipment is unavailable, characterized according to the ;dentification of the train affected, the plant mode at the time equipment is unavailable (operating or shutdown), characterization of the unavailable period

{planned, unplanned, or support system unavailable), and, if due to a support system being unavailable, i den ti fi cation of the support system; 11

U. S. Nuclear Regulatory Commission LIC-96-0069 Page 4 OPPD's Comments:

This section of the proposed rulemaking requires the collection and characterization of unavailability information during both operating and shutdown time periods. The proposed rule does not take into account the fact that certain plant systems that are risk-significant during operation are not risk-significant during shutdown. Likewise, systems such as Shutdown Cooling (or Residual Heat Removal) which are not risk-significant at power are indeed risk-significant during shutdown.

In addition to the above comments, OPPD takes serious issue with the artificial designation of system trains proposed by the draft Regulatory Guide. During the Data Rule Pilot Study, it was determined that functional groups were a much more efficient way to monitor equipment than artificially produced trains. Use of functional grouping would allow for comparisons among different plants. The train definitions proposed in the current draft Regulatory Guide will not yield useful data for a specific plant or allow comparisons between plants.

In conclusion, the Data Rule, as proposed, is premature, costly and not sufficiently justified. The cost estimates for compliance with the Data Rule are inaccurate and without sound adequate technical basis. OPPD encourages the NRC to carefully reconsider the promulgation of the Data Rule. Furthermore, OPPD encourages the NRC to look closer at industry initiatives, such as the Institute of Nuclear Power Operations' (INPO) proposed functional database system, as alternatives to the Data Rule. This system could supply the data needed by the industry and the NRC for risk-based regulation.

If you should have any questions, please contact me.

Sincerely,

~.1P~:=c-=====---

T. L. Patterson Division Manager Nuclear Operations TLP/dll c: Winston & Strawn L. J. Callan, NRC Regional Administrator, Region IV L. R. Wharton, NRC Project Manager W. C. Walker, NRC Senior Resident Inspector A. R. Pietrangelo, Nuclear Energy Institute Document Control Desk

Telephone (508) 779-6711 YANKEE ATOMIC ELECTRIC COMPANY TWX 710-380-7619

~:a-~

\..;W'ANKE1/

---

* une 11, 1996 DOCKETING& YC 96-002 SERVICE BRANCH SECY-NRC John C. Hoyle DOCKET NlllBERPR 0-Secretary of the Commission PROPOSED RULE..!..!!-.. .5 :. ~ - - -

• U.S. Nuclear Regulatory Commission ( lo\ fR..5~\i)

Washington, D.C. 20555-0001 Attention: Docketing and Service Branch

Subject:

Proposed Rule: Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment (61 FR 5318, dated February 12, 1996)

Dear Mr. Hoyle:

Yankee Atomic Electric Company (Yankee) submits this comment letter in response to the subject NRC Proposed Rule. Yankee is the owner of the Yankee Nuclear Power Station and provides engineering and licensing services to the nuclear power plants in the New England Region. Yankee is also submitting these comments on behalf of Northeast Utilities.

In general, Yankee opposes the promulgation of the subject proposed rule. We endorse the comments submitted by the Nuclear Energy Institute in opposition to the proposed rule, and offer the additional remarks contained herein. We also intend to submit comments in response to the draft Regulatory Guide DG-1046 that relates to this proposed rule.

Scope of Proposed Rule The subject proposed rule will require that licensees for commercial nuclear power reactors report plant-specific summaries of reliability and availability data for selected systems and equipment to NRC. This proposed rule will also require that records and documentation of each occurrence of a demand, failure, or unavailable period that provide the basis for the summary data reported to the NRC be maintained onsite and be made available for NRC inspection for five years. Proposed Section 50. 76 establishes an implementation schedule whereby licensees will be required to begin reporting the summary data, compiled at least each calendar quarter, for the calendar year 1997 with the first report submitted by January 31, 1998, and annually thereafter.

U.S. NUCLEAR REGULATORY COMMlSSIOt-.

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Document Statistics Postmm Dale ~ lo ~~e~0 Coples Received_....__ _ _ _ __

.\dd'1 Copies Reproducoo ____ _ __

Speclal Disbibution C()O;.J"tS'"> >--::a;g ~

Letter, YAEC to U.S. Nuclear Regulatory Commission June 11, 1996 Page2 Stated Basis for Proposed Rule The stated basis and rationale for the additional reporting requirements proposed in this rule is that it "is necessary to substantially improve the NRC's ability to make risk-effective regulatory decisions consistent with the Commission's policy statement on the use of probabilistic risk assessments (PRA's)" (60 FR 42622, August 16, 1995). NRC also asserted that such additional reporting will assist the agency in becoming more efficient by focusing its regulatory program on those issues of greatest risk significance and reducing regulatory burdens on licensees. The data required by the proposed rule would be used in "generic issue resolution, developing quantitative indicators that can assist in assessing plant safety performance, performing risk-based inspections, and pursuing modifications to specific plants and basic regulations and guidelines." The stated intent of the proposed rule is also to improve NRC's oversight oflicensees' implementation of the maintenance rule.

Comments We strongly oppose promulgation of the subject proposed rule for the following reasons:

1) The additional requirements proposed in this rulemaking were never intended as necessary for implementation and oversight when NRC established the PRA Policy Statement and promulgated the Maintenance Rule. Accordingly, the additional recordkeeping and reporting are not needed by NRC relative to compliance with any current regulation applicable to reactor licensees. Furthermore, the desire for such data or mere assertion that the information would improve the NRC's oversight oflicensee implementation of the Maintenance Rule are insufficient reasons and bases for promulgating such additional requirements.

2) The NRC has not prepared a backfit analysis documenting that there is substantial additional public health and safety protection provided by the proposed regulation as required under the provisions of the backfit rule (10 CFR 50.109). The NRC backfit process is defined on the principle that new positions and requirements are to be reviewed for backfitting considerations and should meet the standards of the backfit rule.

The proposed rule correctly states that the maintenance rule does not contain reporting requirements. Thus, if the proposed data collection/reporting requirements are considered to be a necessary part of a maintenance rule implementation, then the subject proposed rule represents a backfit. As such, the new rule must demonstrate appropriate cost-benefit with a backfit analysis.

Letter, YAEC to U.S. Nuclear Regulatory Commission June 11, 1996 Page3 The rule also states that the new data are needed so that NRC can monitor the licensee's on-line maintenance activities. Under the maintenance rule, the licensee is responsible for providing and justifying an appropriate (and documented) balance of reliability and availability under the maintenance rule. If the licensee is now taxed with new and prescriptive data reporting requirements as called for by the subject proposed rule, the new rule must be justified by a backfit analysis.

We believe that the standards of the backfit rule are appropriate and should be applied for this rulemaking. Furthermore, the information reporting requirements imposed by the rule should also be justified by an evaluation against criteria analogous to the provisions in 10 CFR 50.54(£), that is, by demonstrating that the burden of reporting is justified in view of the potential safety benefits to be obtained from the information reported.

3) There are currently no regulatory requirements for licensees to conduct PRA' s.

Accordingly, it is premature for NRC to impose reporting and recordkeeping requirements applicable to analytical tools that are not currently required by regulation, but rather are associated with future risk-based initiatives and regulations under consideration. Therefore, any such additional requirements should be developed in an integrated manner with the future risk-based initiatives in order to assess all associated costs and benefits accurately.

The supplemental information to the proposed rule states that the data are needed to support NRC's application of PRA methods to regulatory issues. However, the proposed rule fails to recognize that: (a) all plants have already performed PRA analyses (i.e.,

individual plant examinations) without the need for the requested data, and (b) NRC and its contractors have also performed detailed PRA analysis of licensees' plants without the requested data. The rule does not identify any specific and quantifiable benefit that such data would add to the already large amount of reliability/availability data that exists in the PRA literature.

The rule also fails to place the role of such data within the overall context of PRA limitations. There will always be uncertainty in equipment failure data as well as other aspects of PRA (e.g., completeness of models, human reliability). Plant-specific IPEs have generally shown plant-specific differences (in design and operation) and analyst-specific assumptions (regarding scope, methods, dependencies, etc.) to be the dominant reasons for plant-to-plant variability in PRA results. Plant-specific differences in equipment failure rates have not been shown to be particularly important in the plant-to-plant variability of PRA results. Thus, the new data will not allow NRC to perform more meaningful risk-based comparisons among plants as is stated in the proposed rule.

Letter, YAEC to U.S. Nuclear Regulatory Commission June 11, 1996 Page4 The proposed rule also provides no evidence that the new data collection/reporting will significantly increase the accuracy (i.e., reduce the uncertainty) of such equipment reliability/availability data, or even that more accurate failure rates would significantly affect results within the overall context of PRA uncertainties. Thus, the rule does not show how the new data will help the NRC or industry to improve the use of PRA in risk-based regulatory initiatives.

4) There is a huge burden placed on licensees by the added requirements which are excessive in scope and unreasonable in cost.

The NRC Maintenance Rule already places a large burden on licensees for the collection and analysis of reliability/availability data. The proposed data collection is an additional burden because it would require the licensee to collect data according to the NRC's new desire (under the proposed rule) in addition to the licensees need (under the Maintenance Rule). Neither the scope nor the type/format of data are necessarily the same under the two rules. This creates both confusion and burden.

NRC has assumed in its supporting rule statement that, in connection with implementation of the Maintenance Rule, 80 to 110 licensees are collecting, or plan to collect, reliability and availability data similar to that called for by the subject proposed rule. There is no basis to this assumption and, in fact, New England reactor licensees typically collect only reliability and availability data that is within the scope of their implementation of the Maintenance Rule. Therefore, not all of the data for all of the systems proposed to be covered by this rule are necessarily collected now, or are planned to be collected by each licensee.

Thus, the proposed data will not necessarily help the industry implement the Maintenance Rule. The industry has taken the initiative to collect and maintain the data needed to implement the Maintenance rule, both on a plant-specific basis and on a generic (INPO) basis. The rule claims that it would "enhance licensees' capabilities to implement the evaluation and goal-setting activities required by the maintenance rule." The supplemental information to the proposed rule fails to demonstrate how the new data collection and reporting will accomplish such enhancement. In fact, the additional burden created by such a rule will actually detract from licensees ability to implement cost-effective and safety-optimized maintenance rule programs.

Conclusion We urge the Commission not to adopt the rule as proposed on the basis that 1) the requirements were never intended as necessary by NRC for implementation and oversight of current regulatory

Letter, YAEC to U.S. Nuclear Regulatory Commission June 11, 1996 Page5 requirements; 2) no backfit analysis has been conducted to support the proposed rule; 3) the proposed rule is premature; 4) no specific and quantifiable benefit has been identified relative to the large amount of reliability/availability data that already exists in the PRA literature; and 5) the proposed requirements impose an excessive and unreasonable burden on licensees.

In closing, we wish to take this opportunity to iterate that Yankee is indeed an advocate for moving forward toward enhancement of nuclear safety and reliability through risk-based and performance-based regulation. In our efforts to achieve that vision, we agree with the Nuclear Energy Institute that "additional data needs can best be determined and fulfilled in step with the e pilot activities in NRC's PR.A Implementation Plan as well as other risk- and performance-based regulatory initiatives."

Sincerely, YANKEE ATOMIC ELECTRIC COMPANY eM. Grant nager, Regulatory and Industry Affairs c: M. Fairtile, NRC, NRR J. White, NRC, Region I

WINSTON & STR 35 WEST WACKER DRIVE 1400 L STREET, N.W. 6, RUE DU CIRQUE CHICAGO, ILLINOIS 60601-9703 WASHINGTON , D.C. 20005-3502 '96 JU 12 A9 :27 75008 PARIS, FRANCE 200 PARK AVENUE SULAYMANIYAH CENTER (202) 371-5700 NEW YORK, NY 10166-4193 RIYADH 11495, SAUDI ARABIA FACSIM IL E (202 3

) 71-5950 OF FI CE or -.:: CR[ Lo.RY DOC KET li~G 1 *:[RVICE 43, RUE DU RHONE BHA~1 CH 1204 GENEVA, SWITZERLAND June 11, 1996 [)()C1(ET NtllBERpn PROPOSED RULE so -

VIA MESSENGER (b\ FR.S~\~J @

U.S. Nuclear Regulatory Commission One White Flint North 11555 Rockville Pike Rockville, MD 20852-2738 Re: Comments On Reporting Reliability And Availability Information For Risk-Significant Systems and Equipment; Proposed Rule (61 Fed. Reg. 5318; February 12, 1996)

ATTN

• Docketing and Service Branch On February 12, 1996, the U.S. Nuclear Regulatory Commission (NRC) issued the above-captioned proposed rule for public comment. Provided below are the comments of the Nuclear Utility Backfitting and Reform Group (NU BARG) .1 1 NUBARG appreciates the opportunity to comment on this important rulemaking.

I. Introduction and Summary The proposed new regulations (to be added at 10 C.F.R. § 50.76) would require nuclear power reactor licensees to compile specific information on the reliability and availability of certain risk-significant systems and components and submit annual reports containing that information to the NRC. Licensees would also be required to maintain on site, and make available for NRC inspection, records and documentation that provide the basis for the data reported to the NRC. The proposed requirements would go into effect on January 1, 1997, with a licensee's first report to NRC due for submittal by January 31, 1998.

1/

NUBARG is a consortium of 16 nuclear utilities formed in the early 1980s, which participated actively in the development of the NRC's backfitting rule in 1985, and which has closely monitored the NRC's application of the rule since that time.

"'JU~- 1-9 199f "cknowledged by card .................uMt* ~*~

U.S. NUCLEAR REGULATORY COMMISSIOt--.

DOCKETING & SERVICE SECTION OFFICE Of THE SECRETARY OF THE COMMISSION Docunerrt Stattstics p.-oate 0...u.:n~.,._

Copes Ree81ved _.._ _ _ _ _ __

.\Ml Coples Reprodtad . .~  ::=--- - -

Speola l)strtootiOO C);D *.:.e,_ J ~ t

WINSTON & STRAWN U.S. Nuclear Regulatory Commission June 11, 1996 Page 2 Of specific interest to NUBARG, and the focus of our comments below, is the NRC's position in the "Backfit Analysis" portion of the proposed rule that "the backfitting rule, 10 CFR 50.109, does not apply ... because these amendments do not involve any provisions which would impose backfits as defined in 10 CFR 50.109(a)(1)." 61 EfilLReg_._ at 5325, col. 3. Instead, the NRC has chosen to limit itself to the preparation of a draft Regulatory Analysis (RA), dated December 19, 1995, to examine the costs and benefits of the proposed rule and various alternatives considered by the Commission. 21 In sum, and as detailed below, NUBARG's comments on the proposed rule are as follows:

The language of the Backfitting Rule and the record underlying the rule do not support the NRC's determination that the Backfitting Rule, in general, does not apply to information collection and reporting requirements such as those in the proposed rule.

Licensees will need to develop new procedures, provide training, make organizational changes and install plant modifications as part of their programs to implement the information collection and reporting requirements of the proposed rule and, therefore, a backfitting analysis is required.

It would be inappropriate for the NRC to attempt to compare the process of generic rulemaking to produce new requirements for information collection and reporting to the process of issuing information requests under 10 C.F .R. § 50.54(f), in an effort to support a determination that the Backfitting Rule does not apply to the proposed rule.

2/

The NRC states that the RA constitutes a disciplined process for examining II projected benefits/impacts of information collection and reporting 11 requirements such as the proposed rule. lbi.d... The NRC further states that, because it has prepared the RA, the objective underlying the Commission's Backfitting Rule (La._, "that regulatory impacts are assessed under established criteria in a disciplined process") is being met for the proposed rule. lbi.d...

WINSTON & STRAWN U.S. Nuclear Regulatory Commission June 11, 1 996 Page 3 II. Discussion A. The language of the Backfitting Rule and the record underlying the rule do not support the NRC's determination that the Backfitting Rule, in general, does not apply to information collection and reporting requirements such as those in the proposed rule.

The scope of the NRC Backfitting Rule, 10 C.F.R. § 50.109, is broadly defined by § 50.109(a)( 1) to include "modification of or addition to systems, structures and components, ... or the procedures or organization required to design, construct or operate a facility ... any of which may result from a new or amended provision in the Commission rules .... " Accordingly, the Backfitting Rule applies if the NRC's proposed rule on reliability and availability information collection and reporting may result in such "modifications." The Backfitting Rule, on its face, does not exclude from its reach information collection and reporting requirements such as those in the proposed rule. Moreover, in adopting the Backfitting Rule the Commission never expressed the position that information collection and reporting requirements are not backfits within the scope of the Backfitting Rule. The NRC now appears to have reinterpreted the Backfitting Rule to exclude information collection and reporting rules, regardless of their effects on licensees.

- In the statement of considerations to the 1985 Backfitting Rule, the Commission emphasized that the policy of the Backfitting Rule is to address new regulatory burdens on licensees "regardless of their source." 50 Eed... Be.g_._ 38,097, 38101, col. 1 (September 20, 1985). The Commission explained that it had adopted a broad definition of "backfitting" in the rule to address not only the "cause" of the backfit (source and type of requirement) but also the "effect" on licensees. 50 Eed...

Rag_. at 38,101, cols. 2-3. In so doing, the Commission explained that it had chosen:

[t]o adopt a management process not only for the promulgation of regulations as backfit instruments, but also for the lower tier staff review and inspection process known to result in reactor plant changes.

50 .EerL Be.g_._ at 38,101, col. 2. Such changes, the Commission stated, include changes to "plant design, construction, operation, organization and training." lb.id..

A review of the Backfitting Rule and its regulatory history reveals no indication that the Commission intended that new reporting and recordkeeping requirements be exempt from the disciplined and objective standards of the rule. Had the Commission

WINSTON & STRAWN U.S. Nuclear Regulatory Commission June 11, 1996 Page 4 intended to exclude new reporting and recordkeeping requirements from the Backfitting Rule, it easily could have said so.

If changes to design, operation, procedures, organization and training "may result from" the information collection and reporting requirements of the proposed rule, the Commission must make the determination required under 10 C.F.R.

§ 50.109(a)(3) that, based upon the analysis described in § 50.109(c), "there is a substantial increase in the overall protection of the public health and safety ... to be derived from the backfit and that the direct and indirect costs of implementation ..

. are justified in view of this increased protection." The draft RA prepared for the proposed rule does not make a determination that a substantial increase in overall safety will result. Thus, the RA cannot substitute for the analysis described in § 50.109(c).

By choosing not to perform a backfitting analysis, the NRC may be overlooking the existence of potential "modifications" and the likely "effects" of proposed modifications that trigger the applicability of the Backfitting Rule. As noted below, a review of the proposed rule and supporting NRC documentation indicates that licensees will need to develop new procedures in response to the new requirements. The proposed rule also contemplates training and likely will result in changes to a licensee's organization and some plant modifications as well.

B. Licensees will need to develop new procedures, provide training, make organizational changes and install plant modifications as part of their programs to implement the information collection and reporting requirements of the proposed rule and, therefore, a backfitting analysis is required.

NUBARG's review of the proposed rule and NRC supporting documents confirms that licensees will need to develop procedures implementing the new information collection and reporting requirements. The NRC states (RA at 3-10) that it expects to conduct initial inspections of "licensee programs ... " and the NRC's draft Regulatory Guide DG-1046 (April 26, 1996, for comment) illustrates (Fig. 1 at

p. 6) that such programs are comprised of numerous elements, activities and inputs.

As a practical matter, licensees will need to develop procedures and related instructions to guide personnel and ensure that an adequate program exists to satisfy the new NRC requirements.

WINSTON & STRAWN U.S. Nuclear Regulatory Commission June 11, 1996 Page 5 New procedures will obviously be required as a direct result of the proposed rule. The proposed information collection and reporting requirements pertain to risk-significant systems and equipment (see proposed § 50. 76(b)(3)) and, therefore, would be generally subject to a licensee's quality assurance program, in accordance with 10 C.F.R. Part 50, Appendix B. Specifically, Criterion V

("instructions, procedures and drawings") of Appendix B requires activities to be prescribed by appropriate instructions, drawings or "procedures." It would not be prudent for licensees to attempt to implement the new information collection and reporting requirements without the aid of procedures. Figure 1 of the draft regulatory guide illustrates that the implementation of the proposed rule is a multi-faceted activity involving multiple inputs and opportunities for process flaws.

The NRC also contemplates the need for training on the software and data bases required by the proposed rule (see RA at p. 3-10). As noted above, in adopting a broad definition for backfitting in 1985, the Commission explicitly mentioned "training" as an important type of change that results from NRC Staff activities including promulgation of regulations and inspections.

The effects of the proposed rule are not limited to procedures and training. Page 3-11 of the RA for the proposed rule indicates that the NRC has assumed a one-time cost of $10,000 for each licensee that includes "space allocation, computer purchase, and any associated hardware." Because these items are required for a safety-related activity, Criterion V and other related Appendix B provisions may apply.

The proposed rule will also require licensees to assign additional personnel, to perform the work involved with data collection and reporting and to make organizational changes, such as the addition or modification of supervisory personnel to ensure compliance with the new requirements. For example, the NRC indicates (RA at p. 3-12) that a licensee's "project management" of the information collection and reporting activities will be necessary under the proposed rule. It is clearly a possibility that licensees will need to provide additional supervisory personnel and related changes to their organizations to respond to the proposed rule.

In addition, plant hardware modifications likely will be necessary for many plants. For example, some licensees may need to install run-time meters on equipment within the scope of the proposed rule as a means to acquire the data to be reported to the NRC.

WINSTON & STRAWN U.S. Nuclear Regulatory Commission June 11, 1996 Page 6 C. It would be inappropriate for the NRC to attempt to compare the process of generic rulemaking to produce new requirements for information collection and reporting to the process of issuing information requests under 10 C.F.R. § 50.54(f), in an effort to support a determination that the Backfitting Rule does not apply to the proposed rule.

The proposed rule does not explain the NRC's rationale for determining not to apply the Backfitting Rule. The NRC Staff apparently has developed a position that, for purposes of the Backfitting Rule, new reporting requirements are similar to information requests under 10 C.F.R. § 50.54(f). 31 NU BARG will briefly address the NRC's assertion as part of its comments on the proposed rule, in view of the absence of other articulated reasons from the NRC in support of its determination not to apply the Backfitting Rule. 41 The Commission has long recognized that even new information requests issued under 10 C.F.R. § 50.54(f) can sometimes represent a backfit. In the 1 985 statement of considerations the Commission expressed particular concern for the burden imposed by "extensive information requests." 50 Ee.d.._ B.e..g.... at 38,102, col. 2. There appears to be no question that the burden imposed by the proposed rule is extensive. In this case, the Nuclear Energy Institute has estimated (letter to NRC dated March 12, 1996, p. 2) that the annual recurring industrywide burden of the information collection and reporting requirements

.3/

Footnote 11 of Revision 6 of the Charter of the Committee to Review Generic Requirements, as approved by the Commission on March 22, 1 996 (SECY-96-032), contains the statement that NRC reporting requirements, such as those contained in 10 C.F.R. 50.72 and 10 C.F.R. 50.73 ... , are more akin to the information requests covered under 10 CFR 50.54(f) than they are to modifications covered und~r the backfit rule ....

41 To comply with the mandate of the Administrative Procedure Act, the NRC would need to supply a more complete rationale for its decision not to apply the Backfitting Rule. Regardless, if a backfitting analysis of the proposed rule is required by the Backfitting Rule, the NRC has not followed its own regulations in this case.

WINSTON & STRAWN U.S. Nuclear Regulatory Commission June 11, 1 996 Page 7 proposed by NRC is approximately $30 to $105 million. 51 Further, in cases where it is not clear whether the proposed action meets the definition of backfitting in § 50. 109{a) { 1), the Commission has directed that a backfitting analysis should be performed. 50 Eed... B.e..g_._ at 38,102, col. 2.

Unlike new information collection and reporting requirements resulting from NRC rulemaking, information requests pursuant to § 50.54{f) are not a proper method for imposing new requirements to modify plant systems, structures, components, procedures or organizations. The primary purpose of § 50.54{f) information requests is to provide the NRC with a means to obtain information to verify licensee compliance with the current licensing basis or determine that a license should be modified, suspended or revoked. Licensees are required to respond, but may choose to submit information responsive to the NRC's request or provide a basis for not undertaking the action requested.

Taken to its logical extreme, the NRC's assertion that information collection and reporting requirements are more akin to § 50.54{f) information requests may lead to anomalous results. For example, would the NRC assert that significant substantive revisions of 10 C.F.R. §§ 50.72 or 50.73 or 10 C.F.R.

Part 21 would not constitute backfitting? Such an assertion would signal, in NUBARG's view, a significant change from past NRC precedent. In connection with the Commission final rule promulgated in 1 991 that requires licensees to report emergency response data, a backfitting analysis was performed (56 Ee.d..

B.e..g_._ 40,178). The emergency response data rule was closely analogous to the proposed rule. For Backfitting Rule scoping purposes, there is no practical difference between that rule and the instant proposed rule. A more reasoned approach would be for the NRC to examine in each case the potential for information collection and reporting requirements to result, for example, in licensee modifications to the procedures, training, hardware or organization required to operate a facility.

5./

If the proposed rule would have an annual effect of $100 million or more, it would be a "significant regulatory action" within the meaning of Executive Order 12866 of September 30, 1993 (58 Eed... B.e..g_._ 51,735; October 4, 1993) and the NRC would be required to conduct a more stringent analysis of the proposed rule's projected costs and benefits to be submitted for approval by the Office of Management and Budget.

WINSTON & STRAWN U.S. Nuclear Regulatory Commission June 11, 1996 Page 8 In short, the evaluation which § 50.54{f) requires the NRC Staff to perform, Le...., a demonstration that the burden of reporting is justified in view of the potential safety benefits to be obtained from the information reported

{evaluation not required for information sought to verify compliance by the licensee), is limited only to such requests and does not supersede the requirements placed on NRC to justify its actions that fall within the scope of the Backfitting Rule.

Ill. Conclusion The NRC view that, in general, information collection and reporting requirements such as those contained in the proposed rule are not subject to the Backfitting Rule is unsupported and, in our view, is incorrect as a matter of law. In any case, that view is not applicable in this case, because the proposed rule will result demonstrably in "modifications" as defined in the Backfitting Rule. A backfitting analysis in accordance with 10 C.F.R. § 50.109 is therefore required, with the requisite findings that the proposed requirements will result in a "substantial increase" in safety and be cost-justified.

Sincerely,

~c.~r Nicholas S. Reynolds Daniel F. Stenger Robert E. Helfrich Counsel to Nuclear Utility Backfitting and Reform Group

DOCK[1T0

• . Detroit Fermi 2 6400 1-lc rtr- D1~1e H19111vay USNRt ::. **

Edison Nev,po1: M1(:.**:1q,1c1 (313) 586-40(;0.

413166 Nuclear Operations

'96 JU 12 A8 :17 OF FI r c~ 3U' RETARY DOC KETl.:G 1 \E.PV !CE Bf~ AN1, H June 11, 1996 NRC-96-0068 oocm Nt111E111111

• U .S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, D. C. 20555

References:

1) Fermi 2 PACroSEO RULE..;.ffl..::.:....-.:s...,..-

(Co\F~55\~J NRC Docket No. 50-341 NRC License No. NPF --43

2) Proposed Rule: "Reporting Reliability and Availability Information for Risk-Significant Systemsand Equipmen :*

published in the Federal Register dated February 12. 19 6 (61FR5318)

3) Draft Regulatory Guide DG*l046 dated April 1996, "Guidelines for Reponing Reliability and Availability Information for Risk-Significant Systems and Equipmen in Nuclear Power Plants"

Subject:

Detroit Edison Comments on Proposed Rule: Reporting Reliability and Availability Information for Risk-Signific t S st E ui ment Detroit Edison is pleased to have an opportunity to comment on the N1,1clear Regulatory Commission's (NRC) proposed rule. "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment," publis din the Reference 2 Federal Register.

In summary, we believe the proposed rule is unnecessary and should not be promulgated by the NRC. Our view is based on th~ following points:

JUN 1*

• Ackn0wl811ged by cant

U.S. NUCLEAR REGULATOOY COMMISSIOt--.

OOCKETING & SERVICE SECTK)N OFFICE OF THE SECRETARY Of THE COMMISSION Doa.merrt Statilb Postnark Date - =-,u.-..~ ~ - -

Coples Reostved_.,__ _ _ _ __

,\ckfl Copies Reprodwid ~ - - - -

Speola Dtsb1tlution (:a;).,.,,,~. :::,,1) 5

JLJN-11-1996 17:25 FERMI 2 DET EDISON 313 586 4208 P.02 USNRC June 11, 1996 NRC-96-0068 Page2

• The additional data reporting and recording keeping is not needed by the NR for assessing compliance with any current regulatory requirement.

• The estimate ofindustl)' burden pro~ded in the supp~rting statement to_ 0 ~ is maccurat*~. The rule would actually tmpose an excesstve burden on the md ry.

• The potential total impact of the proposed rule is uncertain because definitio s of terms are not clear and would be subject to interpretation.

• The data required by IO CFR 50.76 represents a substantial increase in data collection needs over and above the Maintenance Rule.

With regard to the first point above, the reporting and record keeping of rellabilii and availabil~ty information is not necessary for the NRC to oversee the

• implementatfon of current regulatory requirements by reactor licensees. There atf1e already hundreds of different reporting and record keeping requirements that ful ll this purpose (ref NUREG-1460, Guide to NRC Reporting and Record keeping Requirements). As stated in the Federal Register notice, the NRC believes << ... th information ii; necessary to substantially improve the NRC' s ability to make risk

• effective decisions consists with the Commission's policy statement on the use o probabilistic risk assessments (PRA)." It should be noted, however, that there is no regulatory requirement for licensees to conduct PRAs or maintain PRA models.

Thus, the proposed rule has no statutory basis and would require record keeping and reporting information for use in an analytical tool that is not required by regulati n.

Its promulgation, in and of itself: does not result in any improvement to public h th and safety.

We believe that data reporting and record keeping requirements should be direct y associated with regulatory activities that are necessacy for the NRC to fulfill its I statutory mission to protect public health and safety or that provide substantial additional protection under the provisions of the bacldit rule. Furthermore, we believe it is premature for the NRC to require additional data reporting and reco d j

keeping in advance of future risk-based regulations or regulatory activities. Any additional da1:a reporting or record keeping requirements should be integrated an m step with the risk-based regulatory initiative under consideration so that the associated costs and benefits can be assessed accurately.

Our second p,omt above addresses the estimate of industry burden imposed by th proposed rule. In the supporting statement provided to 0MB. it is assumed that 80 of 11 O licenses are collecting, or plan to collect; similar reliability and availability information. A basis for this assumption is not provided.

JUN-11-1996 17:25 FEF"M I 2 DET ED I S0t'1 313 586 4208 P.03 USNRC June 11, 1996 NRC.96w0068 Page3 Based on our discussions with other utility personnel at Maintenance Rule conferences, few if any, licensees collect the information as described in the prop rule. There are two major differences from what data is typically collected and hat the proposed rule would require. These major differences involve the collection d reporting of demand information and infonnation on the conCWTent unavailabili two or more :systems or trains.

Regarding demand infonnation, the proposed rule calls for the collection and reporting of actual demand infonnation, whereas the industry typically estimates e number of demands on equipment. By requiring actual demand information that s equivalent to the information collected for actual failures. the burden of data 1 collection is dramatically increased. For example, for equipment that is tested oq a monthly basis, the infonnation associated with each of the twelve demands in a y~ar would be collected and reported under the proposed rule. Today, if one of th§ twelve demands resulted in a failure, the only equivalent infonnation that would e collected would be on the demand that resulted in the failure. Thus, for this ex pie, the proposed rule would increase the burden of data collection by greater than a factor often, just on test demand information alone. Given that the reliability ofrsk-significant equipment in the industry is generally well above 90%, and that the proposed rule calls for information on both failures and successes on all types of~

demands (test, inadvertent, or actual need), the actual burden of record keeping d reporting on demand infonnation is substantially greater than the estimate provid to O:MB.

The second key difference involves the collection of the number of hours when more trains from the same or different systems are concurrently unavailable, as be required under the proposed rule. Current industry practice focuses on monit ring the number of hours that a single train is unavailable, not on the overlap of hours when two or more trains are unavailable. This requirement would be particular!

burdensome during refueling outages, when several plant systems are removed service for maintenance.  !

Regarding the third and fourth points above, the NRC has.prelimmarily chosen alset of basic systems for which reliability data (i.e., demands, failures to start, etc.)~ be reported for all plants which have them. Basic Systems for Fermi are seen in Ta*e 1.

I

JUN-11-1996 17=26 FERM I 2 DET ED I ~;[ltj 313 586 4208 P.04 USNRC June 11, 1996 NRC-96-0068 Page4 Table 1 Reactor core RCIC Isolation Feedwater HPCI YES coolant m*ection Reactor RPS YES Function Protection onl Low pressune LPCI YES Start and n coolant injection CS* YES* Function onl Emergency EDGs YES Start and !run*

ower

• CS,._ Core Spray is not risk significant at Fermi 2 from Maintenance Rule I perspective. Data collected for annual report per Technical Specification 6.9.1. ~.c only for outages during the time when the system required to be operable per 1 Technical Specifications.

I Monitoring of Maintenance Rule performance criteria at Fenni 2 is currently as teen in Table l _ The monito;iflg established is ~otally _co?~sten~ with PSA _as~~mptio s.

For examplet RPS reqwes an extremely high reliability. Smee the reliability mu be very high, monitoring functional failures is sufficient to determine if the system ~oes not meet this performance criteria since one functional failure would cause the st5tem not to obtain its very high reliability performance requirement. This philosophy '

applies to other "high reliability" systems. If there is one functional failure~ then li t has failed to meet its performance criteria. Keeping track of demands will add no v;Jue_

Further, Cor,e Spray is not risk significant and therefore does not merit detailed [

Maintenance Rule monitoring performed for other more risk significant systems.

Therefore, for some basic systems increased monitoring burden would be neces to accommodate Data Rule monitoring.

The basic syi;tems listed in Table 1 are not sufficient by themselves per the prop sed rule. Additi<>nal systems and equipment to be monitared will depend on plant-s ecific features. These additional systems would be selected based on plant-specific P studies. Pot*~ntial systems which may fall in the scope of the Data Rule are seen in Table 2.

JUN-11-19%, 1 I

• C. I FERMI 2 DET EDISON 313 585 4208 P.05 USNRC June 11, 1996 NRC-96-0068 Page 5 Table 2

.ms:.: . :. *. . :::o:: \SYSDM : :AVAILABitJTY'. DATA< rs~*-:"-~"\"';:i)A.TA:!tSTaiVRUN-) ** .

Feedwater SBFW YES Functional failures 1 coolant injection Reactivity SLCS YES YES control Decay Heat RHR YES YES Removal Service Water GSW* NO* System functional failur( s

• For the GSW system, redundancy is qualitatively monitored in lieu of quantitative availability measures i I

As can be seen. substantial increases in monitoring requirements would be needeI to accommodat e the Data Rule. The Data Rule also requires plants to report relia ility 1

and availability for certain risk significant systems and equipment. This applies t the event mitigating systems and equipment which could have si:IP)ificant effect on ri k in terms of avoiding core damage accidents or preserving containment integrity. I Clarification regarding scoping of systems with "significant effect on risk" is req4ired.

This is very subjective and would be left to the opinion of an inspector unless fu~er definition is provided. In summary, the concern is that the small list of Data Rulf systems in Table 2 could be expanded through inspection actions to include all 1 potentially risk-significant systems due to an ambiguity in the Data Rule language.

I I

Detroit Edison also has specific comments discussed below.

I (61FR5320 and 5326): This comment pertains to the following excerpted j statements: ..... licensee(s) might schedule train outages fur maintenanc, at certain times, such that risks are substantially increased over what would~

exp~cted based on random outages. This situa:tion. would not be indicated y current reporting requirements, or even by simply reporting train unavail *ty, but it could be indicated by the concurrent unavailability of two or more

• as would be reported under the proposed rule." These statements imply Data Rule is necessary to control such occurrences. but in met, describe ri management efforts required to properly implement the Maintenance Rule. The PSA model already handles the effects of multiple systems out of service.

(61FR5321): With regard to the statement, "For example, an individual pant may ha:ve an atypical reliability problem with a specific risk-significant syst

JUN-11-1996 17:27 FERMI 2 DET EDISON 313 586 4208 P.06 USNRC June 11, 1996 NRC-96-0068 Page6 I

and thereby warrant additional attention." This seems to duplicate the inte1t of the Maintenance Rule. Systems that do not meet their perfonnance criteria would normally be classified as (a)(l) and an appropriate get well plan established. I I

(61FR5321): With regard to the statement "It is anticipated that licensees will request a number of relaxations in surveillance intervals and allowed outage times ..... " This example of benefit under Risk Based Technical Specificatiqns implies allowances for relaxation of allowed outage times (AOT). It is DetfOit Edison's concern that if the NRC judges individual changes on a '"risk-neuttaJ" of basis, that any relief of this nature will be precluded, eliminating any value the rule for such applications.

{61FR5322): With regard to the statement, "The NRC would use the hour~

when any two or more trains from the same or different systems are 1 concurrently unavailable to monitor how well licensees are managing the risk associated with such maintenance the NRC seems to be intruding into Maintenance Rule space. This again is covered by the Maintenance Rule a4d on-line maintenance control which is under heavy scrutiny by the NRC wit4out the Maintenance Rule The interplay between two rules is confusing and could well lead to reporting difficulties.

(61FR5322): Under "Licensee Implementation" it is implied that plant spelific data on reliability would play significant role in OOS decision making. Th1 major component for such decision making is to set up risk models and a framework for obtaining risk impact of OOS, independent of the Data Rul~.

Better data would only reduce the uncertainty in the quantified result, a i secondary benefit.

( 61FR5323 ): The task of data taking including the setting of criteria such ~

what constitutes a failure as defined under the section Failure" can be v~

complex. The proposed Regulatory Guide DG-1046 goes into this in som~

detail. For example, note the statement under Section 5: 'COegradation in equipment performance that do not satisfy operability requirements for desilgn basis accidents but would not prevent the accomplishment of a risk-signiti. ,-

• function are generally not reportable as failures under this rule." Addressin

• rt each su;ch detail can well be time-consuming and open to the threat of~

disagre1mient on the criteria chosen on the one hand or over conservative d~ta results nn the other. This problem is always there1 but now it will be under! the weight ,of a rule. Specific examples of the complexity are actually cited in I[>G-1046 in Sections 5 and 5.3 when they discuss desired start times and recov,ries.

JUN-11-1996 17:28 FERMI 2 DET EDISON 313 586 4208 P.07 USNRC June 11, 1996 NRC-96-0068 Page 7 (61FR5324): The scope of the rule is very nebulous and could lead to a lot f effort in establishing and could result in a large number of systems/trains re ative to say the systems of interest to INPO. Recall the effort that went into the Maintenance Rule scope. To minimize this effort, a conservative approach would likely be taken (all risk significant systems/trains) le.ading to a large number of systems (instead of the 7 to 10 estimated by DG-1046) and a ,

corresponding large effort to collect the data. Moreover, when the rule ~s of

..systems and equipment," DG-1046 implies that data at the component l~l is likely tc, be required, leading to significant increase in effort. See Appen

• E of DG~l046, including example 6. See also section (b)(3) of the rule on p. ,

• 5326Fll.

(61FR5326): The requirements under section (b)(l)(iv) that include the cat se and effect of each failure and the proposed form titled "Component Failure Records" (Appendix F to DG-1046) that specifies a description of the corrective actions appears to extend beyond reliability data and correspond4igly adds to the effort. The rule appears to be codifying corrective action progrjams, which should not be the intent of the Data Rule. I Given the concerns discussed above as well as several ambipties in the infonna~*on required by the rule versus current industry practice, we believe burden imposed y the rule would outweigh its potential benefits. The estimate of burden provided the supporting statement to 0MB is fundamentally flawed and inaccurate. The burddn of the record keeping and reporting imposed on licensees would be 2 to 7 times of 1hat estimated on an annual/recurring basis. These estimates do not include any cost~

associated with the installation and maintenance of equipment such as total run-hpur meters, building an infrastructure to archive and retrieve the data, training requir~d for those obtlining the data, nor the cost of sending individuals to industry I .

conferences which are sure to come if this rule is promulgated.  :

In summary, we believe the proposed rule is unnecessary and should not be promulgated by the NRC.

JUN-11-1996 17=28 FERMI 2 DET EDISON 313 586 4208 F'. o::::

USNRC June 11, 1996 NRC-96-0068 Pages If you have questions on our comments, please contact Mr. Robert Newkirk at (313) 586-4211. i Sincerely,

~

cc: H. J. Miller M.P. Phillips D. V_ Pfokett A. V,:gel Supervisor, Electric Operators. Michigan Public Service Commission - J. R. Padgett TOTAL F'.08

DOCKETED US 'RC NUCLEAR ENERGY INSTITUTE

'% JUN 11 P2 :5 7 Thomas E. Tipton VICE PR ESI D EN T Q P[RA T!ONS & EN G INEER I NG June 11, 1996 Docketing and Service Branch DOCKET NlltBERPR PROPOSEORULE.......,.......,iilli,o_,_

U.S. Nuclear Regulatory Commission (lo\FR.-55 \~J Washington, D.C. 20555-0001 These comments are submitted by the Nuclear Energy Institute (NEI), 1 on behalf of the nuclear energy industry, in response to the request for comments on the Nuclear Regulatory Commission's (NRC) proposed rule, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment," published in the Federal Register dated February 12, 1996 (ref. 61 Fed. Reg. 5318). This submittal is intended to supplement our comments on the proposed rule that were provided in our letter dated March 12, 1996, to NRC's Information and Records Management Branch.

The industry and the NRC have held numerous discussions over the past several years regarding the desire for greater sharing of plant performance data. The industry recognizes that data are important to moving forward toward the long-term vision discussed in our industry white paper, NEI 96-04, Enhancing Nuclear Safety and Reliability Through Risk-Based and Performance-Based Regulation. The industry also recognizes that data are important to NRC's PRA Implementation Plan. For these reasons, the industry is committed to working with the NRC to address these data needs in an effective and efficient manner. However, this same commitment provides the rationale for why the industry must vigorously oppose the promulgation of the proposed rule.

We believe the proposed rule will be neither effective nor efficient in supporting the data needs of the industry or the NRC. The fundamental problem is that one cannot predict in advance with any confidence the scope, level of detail, or precision of data required to support any future risk-based or performance-based regulatory 1

NEI is the organization responsible for establishing unified nuclear industry policy on matters affecting the nuclear energy industry, including the regulatory aspects of generic operational and technical issues. NEI's members include all utilities licensed to operate commercial nuclear power plants in the United States, nuclear plant designers, major architecU engineering firms, fuel fabrication facilities, materials licensees, and other organizations and individuals involved in the nuclear energy industry. NEI is the successor organization to the Nuclear Management and Resources Council (NUMARC).

I 776 I S T REET , NW SU I H '100 WA S H I NCION. DC 2000 6-3708 PH ONE 202 7 39 8 1 07 fAX 202 785 1898

1.S. NUCLEAR REGULATORY COMMISSIO~

DOCKETING & SERVICE SECTION OFFICE Of THE SECRETARY OF THE COMMISSION DoQ.mef11 StatisticS Postmark Date3/4 Copes Received \

Ce--2' *.ell- ==

Coples Reprodi£8d :,2, _

Special Otstr\00ti0r1 .C&Q*ePD ~

U.S. Nuclear Regulatory Commission June 11, 1996 Page 2 application. And the determination of what data are required (and at what levels of detail and precision) will be highly application-specific. Thus, we believe much of the data captured under the proposed rule will be superfluous in that they are unlikely to support specific needs that have yet to be determined. Enclosure 1 provides additional detailed comments that underscore our concern.

Another fundamental problem with the rule is that the data required are duplicative of other data sources that the NRC has available to support its activities. The burden on the industry is already formidable with regard to recordkeeping and reporting of data. The proposed rule does not eliminate any of the existing data requirements. Thus, the additional burden imposed by the proposed rule, which in our previous submittal we stated is two to seven times higher than the estimate provided to the Office of Management and Budget, is excessive. Enclosure 2 provides additional detailed comments on NRC uses of and reasons for requiring these data.

We also highlight our concern with the draft regulatory analysis performed for this rulemaking. Many of the assumptions made in this analysis are false and only undermine the validity of the analysis. The most egregious example is the assumption that the recordkeeping and reporting of data required by the rule will result in a one percent increase in the availability of all plants. Improving plant availability by simply collecting and reporting additional data has no basis. Clearly in this case, the regulatory analysis is fundamentally flawed and does not support the promulgation of the proposed rule. Enclosure 3 provides additional detailed comments on the regulatory analysis.

We strongly disagree with the conclusion that the proposed rule does not involve any provisions which would impose back.fits as defined in 10 CFR 50.109(a)(l). The rule would require the development of procedures by licensees to collect, record and report the data, and also would likely involve hardware modifications, such as run-time meters and computers. In light of our review of the regulatory analysis discussed above, the statement that the underlying objective of the Back.fit Rule was met by the "disciplined process" used to assess the regulatory impacts of the proposed rule is not defensible.

In closing, we reiterate our strong belief that the proposed rule is premature and should not be promulgated by the NRC. The industry remains committed, however, to working with the NRC to ensure that data needs are fulfilled in an effective and efficient manner. As a starting point, we support greater sharing of the reliability and availability data that are already collected in the industry on a voluntary basis.

U.S. Nuclear Regulatory Commission June 11, 1996 Page 3 Additional data needs can best be determined and fulfilled in step with the pilot activities in NRC's PRA Implementation Plan as well as other risk- and performance-based regulatory initiatives.

ARP/jes Enclosures

ENCLOSURE I COMMENTS ON THE PROPOSED DATA RULE REQUIREMENTS It is recognized that data is an essential element of risk-informed and performance-based regulation. However, the fundamental issue missed in the regulatory analysis and background information is that the data needed to support specific regulatory initiatives is specific to that initiative and cannot be pre-defined. This is pointed out in numerous industry documents including the PSA Applications Guide and NEI 96-04, the industry white paper on risk-based and performance-based regulation. A key technical point is made repeatedly in the PSA Applications Guide: the data (and models) required for a PSA application are application specific. The measurement of cause and effect in terms of risk is dependent upon the specific use of the PSA. In some cases only a very limited element of the PSA model is even needed. In other applications, the entire model may be required and should be appropriately benchmarked. In NEI's white paper, the performance-based shutdown rule example identifies some of the specific data needs for shutdown performance monitoring. Depending upon the performance-based initiative, the data needs will be different. The proposed rule seems to have ignored this fact and attempts to define "generic" requirements. In doing so (further proving this point), it ignores the largest existing data source in the industry, the Maintenance Rule performance data, and defines a new set of requirements.

As an example, in the risk-based IST initiatives currently underway in the industry, low safety significance components may be subject to different IST requirements. These components, since they are less safety significant, would likely fall outside the reporting requirements of this rule. Thus, this rule does not provide assurances that data collection to evaluate the impact of reduced IST will occur. Data collection would be performed regardless of the rule to assess the impact of different testing requirements.

The following sections detail other concerns with the proposed rule with regard to data coherency, precision, PSA model validity, and the risk significance of data variations.

DATA COHERENCY Data coherency has long plagued efforts to define broad-based data collection efforts. The proposed data rule is no exception. In the discussion for the need for data, it is repeatedly implied (and in some cases explicitly_ stated) that the data will be useful for plant specific risk assessments. This position reflects a misunderstanding of the relationship of data and PSA models. The linkage between reliability/availability data and the PSA model is plant and model specific.

That is, the specific data required to support a PSA calculation for a specific plant is 1

dependent upon the plant specific design of the systems and the architecture of the models of those systems. That is, not only are system designs different from plant to plant, but within the range of accepted PSA technology, there are different approaches taken to defining component boundaries, level of detail and other modeling features. This creates a point of conflict in the NRC's data needs.

Barring standardized plant designs, plant specific data is not comparable, but generic data is not as useful in analyzing plant specific performance. This problem has been well known to the industry for many years, based on the experience gained in the SSPI program. This can be illuminated through the simple example below.

The service water system (or equivalent) would be considered a "risk significant" system in most, if not all, PWRs due to its function as the heat sink for many key plant systems. The difficulty in comparing plant specific data can be clearly seen if the service water systems for two four loop Westinghouse PWRs are compared:

• The service water system at Plant "A:', Auxiliary Saltwater, is comprised of two pumps per unit for a total of four pumps. A single pump is capable of providing cooling needs for a unit. A cross-tie exists which allows the systems to be cross-tied between units.

• On the other hand, Plant "B", another four loop Westinghouse PWR, has a service water system, Essential Raw Cooling Water, which is comprised of four pumps per unit, for a total of eight pumps. With only Unit 1 operating at this site, all eight pumps are available to support the needs of the single unit. The ERCW trains within a unit are generally operated separately; however, the capability exists to cross-tie headers within the unit for maintenance or emergency operations.

The unavailability of these systems at these two plants are vastly different and cannot be reconciled in generically reported data. For example, it is much less significant for Plant "B" to have a single ERCW pump out of service for corrective maintenance than Plant "A:' due to level of redundancy in the system.

Consequently, the manner in which maintenance is performed (i.e., around the clock vs. day shift only) might vary and, as a result, the unavailabilities might vary significantly.

Similar problems exist with reconciling data between different plants for systems such as CCW, RBCCW, electric power distribution systems, even auxiliary feedwater and safety injection systems.

2

Thus, the data coherency issue presents the following dilemma:

The data could be collected and reported on a plant specific basis (on the basis it is meaningful to that plant), but comparison of any values between these two sites or across a group of plants would be meaningless. Alternatively, collection and reporting of generically defined data inhibits direct application of the data in plant specific uses.

DATA PRECISION One of the difficulties in analyzing data from reliable systems (like those in nuclear power plants) is the relative paucity of failure data. By definition, reliable systems do not fail often. A paucity of failure data makes data analysis subject to significant uncertainties and, therefore, imprecise. Various statistical methods can be employed to mathematically combine data (such as Bayesian techniques), but the underlying problem presented by small statistical samples is not solved by these methods, it is only worked around.

This problem has been grappled with by the industry over the past several years during the development and implementation of the Maintenance Rule. In developing performance monitoring programs and performance criteria for specific systems under the Maintenance Rule, utilities have been faced with the challenge of monitoring SSC reliability and availability on a plant specific basis. Recognizing the paucity of reliability data, most utilities have opted for monitoring the number of functional failures (either all functional failures or MPFFs) experienced in a given time period.

Take, for example, a PWR safety injection (SI) system involving two redundant, independent trains. Most plants would monitor the number of failures of a train of SI, based on maintenance preventable functional failures (MPFFs) rather than a true reliability calculation. The reason for this is simple: inadequate data.

Insights into system or train reliability can be inferred using a number of methods that require significantly less resources than exacting data collection methods to support quantitative reliability determinations.

A typical safety injection system may only be "demanded" (i.e., exercised in a manner which generates meaningful failure data) a few times a quarter. Thus, over a two year fuel cycle, it would not be uncommon to see less than 20 total "demands" per train or 40 for both trains. For a typical, reliable Safety Injection System the mean number of demands between failures would be on the order of 100 to 300. Consequently, you would expect that it would be necessary to accumulate eight, two year fuel cycles (16 years at 20 demands per year) to get an adequate amount of data to determine plant specific reliabilities with a significant level of confidence.

3

This is why most licensees have elected to monitor the number of failures of risk significant standby systems rather than quantitative estimates of system reliability as part of the Maintenance Rule. Furthermore, this reinforces the SSPI approach to tracking unreliability based on estimated demands (or operating hours) rather that a precise tracking of actual demands (or hours). In this regard, the proposed rule implies that more precision will be expected than can be statistically justified.

PSA MODEL VALIDITY The background information on the proposed rule implies in numerous areas that the NRC intends to utilize this data reported under the proposed rule in PSA analyses. As described above, the relationship between data and a PSA model is model specific and cannot be defined on a generic basis except at the lowest level (e.g., component level). Furthermore, in order to yield appropriate insights, PSA models must be plant specific. Therefore, the data needs for PSA analyses are plant specific. This raises an additional issue related to the use of data in PSA models:

PSA model validity.

The most comprehensive, plant specific PSA models reside at utilities. These were generally developed for the purpose of responding to Generic Letter 88-20 and, in many cases have been updated and improved since submittal. The maintenance and updating of a PSA model to reflect plant specific design features, procedures and data requires a significant resource commitment. The comparison of the results between the NUREG-1150 plants and the associated PSAs performed for the IPE shows the level of plant specific insight gained by involving plant personnel. In almost all cases, the plant specific IPEs found significantly different dominant contributors than the NUREG-1150 studies found and, in one case, a new plant vulnerability to flooding was found in the IPE.

This raises the question of how PSAs used by the NRC will be validated and the overall quality of the PSA will be assured. For example, the ASP PSA models described in Appendix A of the draft Regulatory Analysis are referred to as a source of "plant specific" PSAs. As described in Appendix A, these PSAs cannot be adequate to support the level of PSA analysis required to achieve meaningful insights. Specifically, the ASP models only contain four or five initiating events. A typical PSA includes 10 to 20 internally initiated events alone, depending upon plant design and level of resolution required. Additionally, the four or five initiators treated in these PSAs do not include support system initiators, a dominant contributor in some plant specific PSAs. Furthermore, these models do not include models of any support systems, except AC pow:er. Here again, the IPE process found numerous examples where support system failures are dominant contributors to CDF.

4

Consequently, the reliance on these simplified, unreviewed, unverified PSA models creates significant uncertainty in the results and may cause additional effort on the part of licensees and the NRC in resolving erroneous information generated by these models. This concern underscores our conclusion that the proposed rule is premature, and that data needs should be integrated and in step with other elements of risk- and performance-based regulatory activities.

RISK SIGNIFICANCE OF DATA VARIATIONS One of the underlying themes of the background information for the proposed rule is that variations in component unreliability and unavailability could have significant impact on plant risk measures. In order to investigate this premise, two NUREG-1150 PRAs were used to evaluate the risk impact of significant variations in reliability and availability data and how that varies from plant to plant. These PRAs were selected because they were both Westinghouse PWRs and they were conducted and documented in a similar manner.

As a first step, the core damage frequency results were reviewed in terms of the importance measures to determine which train reliabilities and availabilities were identified as risk significant. The list of importance measures contained in Appendix E of NUREG/CR-4550 was reviewed to identify all basic events which involved:

• Component Failure To Start (Pumps & EDGs)

• Component Failure To Run (Pumps & EDGs)

• Component In Test and Maintenance (Pumps, EDGs, and MOVs)

This review identified the following system trains for each plant (list in order of decreasing Fussell-Vesely importance):

PWRA PWRB Emergency Diesel Emergency Diesel Generators Generators (2) (3)

Turbine Driven AFW Pump Turbine Driven AFW Pump (1)

(1)

Motor Driven AFW Pump (2) Motor Driven AFW Pump (2)

LPI Pumps (2) LPI Pumps (2)

Low Pressure Recirc Fans Charging Pump Cooling Pumps (2)

Low Pressure Recirc MOVs HPI Pump (1) -

The top four of these system trains were investigated further. The other items were not tracked because a comparison could not be made between plants and because 5

they were less risk significant than the top four. Based on the importance measures provided in NUREG/CR-4550, sensitivity evaluations were performed to evaluate the relative risk increase due to an increase in unavailability/unreliability.

It was assumed that an increase in unreliability or unavailability of a factor of 10 would be considered significant. The results are shown graphically in Figure 1 in terms of the ratio of the core damage frequencies (CDF) and in Table 1 in terms of the change in base unreliability/unavailability and the percentage increase in CDF.

This data shows that even a factor of 10 increase in unavailability/unreliability has a very limited impact on CDF for all but a few key plant system/trains. In fact, changes to the EDGs show the largest impact, but if their unreliability went up by a factor of 10, it would exceed 20%. Unreliabilities of this magnitude would clearly be identified by existing programs, including the Maintenance Rule. The next largest risk increase is observed for the turbine driven AFW pump. Again in this case, an increase in unavailability/unreliability of a factor of 10 would result in values in excess of 10% and the risk increases would be 10 to 36%, depending upon the plant and parameter. In the case of the LPI pumps, the observed risk increases are 5% or less in all cases, even for a factor of 10 increase. Keeping in mind that these system/trains were the most risk significant, it is very unclear how much benefit the NRC would glean from the significant effort involved in collecting and evaluating the data required under the proposed rule. It should also be noted that the systems identified as most risk significant in these NUREG-1150 studies are already monitored and reported on under the industry SSPI program.

Again, the above underscores our point that the data collected under the proposed rule does not add significant value beyond that provided through existing data sources.

6

TABLE I BASE UNAVAILABILITY/UNRELIABILITY VALUES AND ASSOCIATED INCREASES PWRA PWRB BASE BASE CDF BASE BASE CDF COMPONENTtrRAIN/SYSTEM VALUE XlO INCREASE VALUE Xl0 INCREASE EDG 1 FAILS TO START 3.0E-02 0.30 76% 2.2E-02 0.22 224%

EDG 2 FAILS TO START 3.0E-02 0.30 76% 2.2E-02 0.22 119%

EDG 1 IN MAINTENANCE 6.0E-03 0.06 13% 6.0E-03 0.06 50%

EDG 2 IN MAINTENANCE 6.0E-03 0.06 13% 6.0E-03 0.06 28%

TURBINE DRIVEN AFW PUMP FAILS TO 3.0E-02 0.30 36% 1.lE-02 0.11 18%

START TURBINE DRIVEN AFW PUMP IN 1.0E-02 0.10 11% 1.0E-02 0.10 16%

MAINTENANCE MOTOR DRIVEN AFW PUMP 1 FAILS TO 3.0E-03 0.03 1% 6.3E-03 0.06 3%

START MOTOR DRIVEN AFW PUMP 2 FAILS TO 3.0E-03 0.03 5% 6.3E-03 0.06 3%

START MOTOR PRIVEN AFW PUMP 1 IN 2.0E-03 0.02 1% 2.0E-03 0.02 1%

MAINTENANCE MOTOR DRIVEN AFW PUMP 2 IN 2.0E-03 0.02 3% 2.0E-03 0.02 1%

MAINTENANCE LPI PUMP 1 IN MAINTENANCE 2.0E-03 0.02 3% 2.0E-03 0.02 1%

LPI PUMP 2 IN MAINTENANCE 2.0E-03 0.02 3% 2.0E-03 0.02 1%

LPI PUMP 1 FAILS TO START 3.0E-03 0.03 5% 3.0E-03 0.03 2%

LPI PUMP 2 FAILS TO START 3.0E-03 0.03 5% 3.0E-03 0.03 2%

7

RELATIVE INCREASE IN CDF I\J c..i CJ1 I\J CJ1 c..i CJ1 C')

-o EOG 1 IN

~ i:

.., "tJ MAINTENANCE~""""""""""""'""

)>

EOG2IN en ~

MAINTENANCE ct> en TURBINE DRIVEN  :::l 0 AFWPUMPF~LSTO ~,.,.......i -n z START m

=0 TURBINE DRIVEN ..,C: -n AFWPUMPIN MAINTENANCE ct) ~

MOTOR DRIVEN AFW

o en PUMP 1 F~LS TO ,...

m " -n START e_ en c: m cc-*

MOTOR DRIVEN AFW PUMP 2 F~LS TO

l z C:

START ~ ~ ciJ MOTOR DRIVEN AFW m :::! -1, PUMP1 IN =<

m -

~~

MAINTENANCE MOTOR DRIVEN AFW PUMP2IN 0

-n MAINTENANCE 0 LPIPUMP 1 IN ...,

-n m MAINTENANCE m

n en ~

LPIPUMP2 IN 0 ~

MAINTENANCE m

.., en LPI PUMP 1 F~LS TO

~ :;i:::1

~ am

-1,

-I

i

::1 START 0 i:

o:l  ::i> en LPI PUMP 2 F~LS TO START

ENCLOSURE 2 COMMENTS ON NRC USES OF DATA NRC USES COMMENT Generic Issue The staff states that they need data to support risk Resolution estimates associated with generic issue resolution.

However, generic issues, by definition, are generic.

Therefore, plant specific reliability and availability data would not be necessary. Existing data sources (e.g.,

Maintenance Rule, SSPI) provide the NRC with performance data that may be used for generic issue resolution.

Indicators of Plant The Maintenance Rule provides the NRC with a Performance regulatory mechanism to identify plants with performance issues. The establishment of performance monitoring programs and goal setting provide indication of poor performers. AEOD currently collects and analyzes data on plant performance and provides reports to the Commission on various results.

Accident Sequence The current ASP program relies upon "representative" Precursor (ASP) and PSAs to evaluate the significance of industry events.

Event Analysis These models are not plant specific. The generic data utilized in these models is more than adequate to assess the general implications of events. The data reported under the proposed generic rule would not (and could not) necessarily match these PSAs (see comments -

Data Coherency). Therefore, the use of the data gathered under this rule would not significantly enhance the ASP program. If the plant specific implications of a given event need to be evaluated, the NRC can simply request the licensee for a PSA-based evaluation.

1

COMMENTS ON NRC USES OF DATA NRC USES COMMENT Risk-Based Inspections Risk-Based Inspections rely upon the insights from plant specific PSA to identify plant specific risk significant issues. The sensitivity analysis contained in Enclosure 1 demonstrates, for most equipment included in the proposed rule, that the PSA results are relatively insensitive to changes in data. Therefore, decisions about risk-based inspections are not likely to be affected by the data reported.

Aging The assessment of aging impacts requires a significant level of understanding of not only the data, but also the plant specific factors which influence component performance (e.g., maintenance practices, operating practices, environmental conditions, etc.). While the data requested under this rule could provide raw information on component failures and causes, these data do not include sufficient information to glean insights into aging related factors.

Risk-Based Technical Some Risk-Based Technical Specification changes Specifications utilize information on component failure rates and repair times. However, in general, the approach used in these analyses is to identify the desired change and evaluate the risk sensitivity of the proposed change.

The NRC's justification (pg. 3-7 of Reg Analysis) implies that the AOT/STI is derived from the data. Rather, the proposed AOT/STI change is investigated and shown to be risk insignificant. As a result, the level of precision required is not as high. In addition, as technical specification changes are requested by licensees, the plant specific data required to support a specific change could be developed and reported in a manner which is much more cost effective than having a rule to report all data all the time.

2

COMMENTS ON NRC USES OF DATA NRC USES COMMENT Inservice Testing Changes to Inservice Testing strategies may be detected in reliability and availability data. However, current regulations are very prescriptive in the strategies allowed. While the impact of decreasing the test frequency of low safety significance components can be assessed in advance by making assumptions about the reliability and availability of those components, the actual performance of those components can only be demonstrated through the collection of performance data. Since the proposed rule focuses on risk significant or high safety significance equipment, it is unlikely to capture the data of interest for assessing changes to the IST program. Also, without knowing what changes have been (or will be) made to testing strategies, it will not be possible to determine the cause and effect relationship between IST changes and system train reliability.

NRC Maintenance Rule Over the past several years utilities have invested significant resources developing internal data collection and evaluation systems. These systems are tailored to the plant specific systems, plant specific risk characteristics and plant specific information management systems. Imposing data reporting requirements under this rule would add an additional burden on all utilities. The proposed rule assumes that only 30 of the 110 plants would be significantly impacted. There is no basis for this assumption. Very few utilities currently collect all the data required under this proposed rule.

3

COMMENTS ON NRC USES OF DATA NRC USES COMMENT NRC Monitoring This item seems to be focused on the monitoring of Maintenance Rule implementation. There are many other means for the NRC to monitor licensee performance under the Maintenance Rule such as obtaining the plant specific performance data reports and reviewing the results of the required periodic assessments. This information is readily available to resident inspectors and is the most appropriate data for a specific site .

COMMENTS ON NRC REASONS FOR REQUIRING THE DATA NRC REASONS COMMENT No Current Systems This overlooks the various existing data collection and For Systematically reporting systems already in place in the industry. The Reporting Data Safety System Performance Indicator (SSPI) program includes much of the data requested under this rule.

This system has recently been upgraded to include most of the essential data elements and covers most of the same key systems identified in the proposed rule.

Furthermore, utilities are currently collecting failure and unavailability data on a plant specific basis for the maintenance rule. While this data is not currently reported, it represents a significant body of data and effort currently underway in the industry.

Improved Ability To In addition to LERs, other industry initiatives tracking Identify Plants & overall plant performance, such as the SSPis and the Systems At Risk Maintenance Rule include performance criteria which will serve to identify and improve outliers across the industry. These tools could be used to provide the NRC with ample data for identifying plants and systems that may be outliers.

4

COMMENTS ON NRC REASONS FOR REQUIRING THE DATA NRC REASONS COMMENT Licensees Might The maintenance rule requires licensees to evaluate the Schedule Multiple impact of removing equipment from service. Therefore, Equipment Outages beginning in July 1996, all plants will have systems to ensure that the safety impact of multiple equipment outages is understood and managed effectively.

Scrutable Plant The maintenance rule will collect a significant amount Specific & Generic of scrutable data on a plant specific basis for risk Data Are Required To significant systems, structures and components. Up Move To Risk-Based until this year, the NRC has its own program for the Regulation development of generic component databases by INEL.

These efforts were recently discontinued at the direction of the NRC. While data is an important element in risk-informed regulation, it is premature to attempt to define the essential data needs for this purpose.

Currently, the only risk-based regulatory improvement in place is the changes to containment leakage testing (Appendix J) requirements. Data related to Appendix J are not included under the currently proposed data rule. Other initiatives, such as risk-based IST and ISI, Graded QA and others may have entirely different data needs than those identified in the proposed rule.

Data Is Needed To Risk based regulatory applications are most Improve Efficiency & significantly influenced by the quality of the PSA Effectiveness Of NRC models used and the expertise of the analyst applying Regulatory the PSA tool. Data is generally a second order effect Applications when considered in most applications. Even NUREG-1150, the most comprehensive PSA effort ever undertaken by the NRC, relied heavily on generic data.

Furthermore, the level of effort to collect, analyze, interpret and utilize plant specific data may actually decrease the efficiency with which the NRC can evaluate regulatory aspects of risk. In addition, based on past experience, the interpretation and use of this data in NRC PSA models will undoubtedly increase the effort on the part of licensees by requiring more licensee review and interaction.

5

COMMENTS ON NRC REASONS FOR REQUIRING THE DATA NRC REASONS COMMENT Data Can Be Analyzed In addition to LERs, numerous higher level To Estimate Overall performance indicators are currently reported to the Industry Performance NRC .

& To Identify Outliers 6

ENCLOSURE 3 COMMENTS ON THE REGULATORY ANALYSIS FOR THE PROPOSED DATA RULE The Regulatory Analysis for the proposed Data Rule is inadequate to justify the significant burden on the industry. The major inadequacies fall into two categories:

• Insufficient Options Were Evaluated

• Cost & Benefits Are Generally Without Basis The following summary provides an overview of these deficiencies.

OPTIONS CONSIDERED The Regulatory Analysis basically identifies only four options for implementing a data collection/reporting system:

• No Action

• Alternative A - Collect By Rule

• Alternative B - Use NRC Resources To Collect Data

• Alternative C - Voluntary Submission of Data By Industry The first is to continue as is and is considered the base case. The fourth option, Alternative C, is not considered due to the similarities in the benefits to Alternative A and the lack of agreement with the industry to pursue this option. It is our understanding that Alternative C is still a viable option and is still being pursued in parallel with this rulemaking. On the surface these options seem reasonable.

However, there are many approaches to defining what and how data is to be reported which are not considered. That is, the Regulatory Analysis predefines the exact data required and evaluates the assumed costs and benefits of that data.

Alternatively, the Regulatory Analysis could have analyzed the costs and benefits of requiring reporting of some of the data already gathered and analyzed by the industry.

This is particularly true for Alternative A. In this case the Regulatory Analysis predetermines the data and evaluates the assumed costs and benefits. Other, more cost effective options are possible. For example, a significant amount of effort is already applied by licensees to the collection and reporting of data under the SSPI program and Maintenance Rule. Two specific alternatives not evaluated in the regulatory analysis include:

• Collect All Data Currently Collected By Industry By Rule

• Collect All Maintenance Rule Data By Rule 1

These options may have had much less impact on licensees and would have provided the NRC yielded benefits consistent with those cited in the Regulatory Analysis (at least qualitatively).

BASIS FOR COSTS AND BENEFITS Generally, the costs and benefits cited in the Regulatory Analysis are without adequate technical basis. In many cases, the benefits cited are from an ill defined future projection of how the regulatory process might be improved in a risk-based regulatory environment, but these benefits are not directly attributable to the Data Rule. For example, one of the major benefits cited is the increase in plant availability due to risk-based operations and on-line maintenance. There is nothing in the Data Rule which increases a utility's capability to perform on-line maintenance. The Maintenance Rule already covers this and the Data Rule is neither necessary or beneficial to a utility's capability to shorten outages by performing on-line maintenance. Additionally, the staff has taken credit for risk-based technical specification improvements which are not directly associated with the Data Rule. If the staff desires credit for risk-based regulation, then that regulatory change must be included in the regulation being evaluated in the Regulatory Analysis. Otherwise, every risk-based initiative could take credit for all benefits expected from all risk-based initiatives.

The second area of difficulty in the regulatory analysis is the basis for the specific estimates of costs and benefits. Many of these estimates are derived from assumptions which have little basis or correlation to the proposed rule. In order to identify these assumptions and the impacts they have on the results of the regulatory analysis, several sensitivity studies were performed on the cost and benefits associated with the proposed rule. The key assumptions investigated and the sensitivities evaluated are described below.

Characterization of Base Case Results The cost and benefit calculations are summarized in Tables 3-1 and 3-2 of the Regulatory Analysis (pg. 3-15, 3-16). Since the preferred alternative is Alternative A - Collect Data By Rule, this case is considered the Base Case (Table 3-1). The quantified costs and benefits associated with this alternative can be summarized as follows:

2

Table 2 TYPE TITLE DESCRIPTION IMPACT COST NRC Implementation Develop Reg. Guide, -$0.?M Rulemaking, Software Industry Identify and Develop Data -$4M Implementation Systems Industry Reporting Data Collection, Analysis -$84M and Annual Reports BENEFIT NRC Operations Elimination of Staff Due To +$8.5M More Efficient Use of Staff Industry Availability 1% Increase In Availability +$750M For All Plants Tech Spec Changes Risk-Based Technical +$43M Specification Improvements These costs and benefits are shown in Figures 2 and 3, respectively. Figure 4 provides a side-by-side comparison of the costs and benefits. Several things are apparent from a review of this data:

• The costs are primarily borne by the industry. The NRC actually achieves a

$8.5M benefit from the rule with only a $700,000 investment. According to NRC estimates, for the industry, the costs are significantly larger, $84M.

Industry estimates of implementation costs are significantly higher than the NRC's.

• The major cost component in the Regulatory Analysis is the assumed 1%

availability improvement ($750M benefit). All other costs and benefits are less than 10% of this one item.

• The NRC's benefits include reduction of staff effort in reviewing technical specification changes while the industry benefits are based on the NRC approving more risk-based technical specifications. *Thus, the regulatory analysis has double counted.

3

Based on a review of the Regulatory Analysis and its assumptions, the following sensitivity cases were analyzed and presented in Figures 5A and 5B. Figure 5A presents the results on a cost scale which shows all the cases comparably.

However, in order to overcome the huge contribution of the single assumption regarding improved plant availability, Figure 5B was provided to allow interpretation of the other cases. The cases are describe below:

Base Case - This case represents the Alternative A cost and benefit assessment described in the Regulatory Analysis and summarized in Table 2 above and Table 3-1 of the Regulatory Analysis.

No Credit for Availability Improvement - As noted above, the largest cost element in the Regulatory Analysis is the benefit accrued from an assumed 1% improvement in plant availability for all plants. This benefit is based on the expected reduction in outage durations gained from utilities performing more on-line maintenance. While it is true that on-line maintenance and outage reduction can increase plant availability, there is no link to the requirements of this rule and, therefore, no benefit should be accrued in the regulatory analysis.

Current regulation, including the Maintenance Rule, allow for on-line maintenance and this rule does not in any way increase the availability of plants.

In this sensitivity study, the availability improvement benefit was eliminated from the Regulatory Analysis and the benefits requantified.

The result shows that when this credit is removed, the costs are nearly a factor of two larger than the benefits for this case ($89M vs. $5 lM).

Revised Estimate of Industry Reporting - The Regulatory Analysis assumes that 80 plants are currently gathering essentially all the data necessary to report to the NRC and that only 30 plants would be required to gather additional data. Industry estimates, based on a sampling of utilities and feedback received at our recent workshop on Maintenance Rule implementation, indicate that the majority of plants would be required to collect a significant amount of additional data (e.g., concurrent train unavailabilities, run times, separation by type of demand).

For the purposes of this sensitivity study, it was assumed that the ratio of plants was reversed. That is, 80 plants would be required to significantly alter their existing programs and g~th~r additional data and only 30 plants are only marginally impacted. This case was requantified using NRC estimates for implementation and reporting costs in addition to the corrections identified in the previous sensitivity study. In this case, the results show that the costs are over 2.5 times 4

greater than the benefits, showing the rule to be even less cost beneficial ($128M vs. $5 lM). It should be noted that, if industry cost estimates for implementation and reporting were used, the computed cost would be substantially larger.

Elimination of Event Investigations - One of the major NRC benefits cited in the Regulatory Analysis is the reduction in staff effort related to investigation of industry events when this data is available. These estimates fail to account for several significant factors: (1) the effort required to perform the implied risk analyses used to eliminate event investigations, (2) the costs associated with developing and maintaining the plant specific PSA models used for these assessments, (3) the costs to industry in responding to NRC requests for information for these studies and in resolving PSA model issues raised in these studies, and (4) the fact that many insights can be gleaned just from understanding the existing PSA studies and results and therefore, this data is not essential to accomplishing this reduction in staff.

While some cost savings could occur, some cost increases will occur.

Thus, in this sensitivity case the benefit due to event investigations was eliminated, in addition to the changes identified in the cases above. The results of this sensitivity show that the costs are nearly three times greater than the benefits, showing the rule to be even less cost beneficial ($128M vs. $45M).

Elimination of Tech Spec Credit - The Regulatory Analysis credits the data rule with allowing the NRC to reduce staff requirements associated with evaluating technical specification changes if this rule is approved. At the same time, the Regulatory Analysis claims credit for processing more risk-based technical specification changes. This amounts to double counting the benefits. In addition, it is arguable the degree to which the data reporting requirements of the proposed rule are essential (and therefore creditable).

In this sensitivity case, the double counting was eliminated by deleting the benefit to industry for risk based tech spec improvements. This reflects the proposed reduction in NRC staff available to review and process requests. This case was requantified using this revised assumption and those cited in the above cases. The results of this sensitivity show that the costs are over a factor of sixty times greater than the benefits, showing the rule to be even le~s cpst beneficial

($128M vs. $2M).

These evaluations of the costs and benefits assumed call into question the validity of the Regulatory Analysis.

5

Figure 2 CONTRIBUTORS TO COSTS OF NRC DATA RULE D Industry Reporting 80

• Industry Implementation mNRC Implementation 70

- 60

ii:

~

w

_J 50

<(

2 I-w 40

"'0::

w ll.

30 20 10 o-----------

COSTS 6

Figure 3 CONTRIBUTORS TO BENEFITS OF NRC's PROPOSED DATA RULE 900 Tech Spec Changes 800 II Industry Availability e NRC Operation 700

~

~

600 w

> 500

..J

<(

z I-w 400 w

0:::

c..

300 C 200 100 -

BENEFITS 7

Figure 4 CONTRIBUTORS TO COSTS & BENEFITS OF NRC PROPOSED DATA RULE 900 Ell Tech Spec Changes 800 Industry Availability NRC Operation 700 - Im Industry Reporting Industry Implementation ml NRC Implementation

~

~

600 w

J 500

.J

<(

I-z w 400

"'0::w a.

300 -

200 100 o+-------

COSTS BENEFITS 8

Figure SA COMPARISON OF CONSEQUENCES OF NRC's PROPOSED DATA RULE

$900 I ffi.lCOSTS BENEFITS I

$800

$700

~ $600 w

..J

<(

$500 I-z $400 w

en w

0:: $300

• CL

$200

$100

$0 Base Case No Credit For Revised Elimination of Elimination of Availability Estimate of Event Tech Spec Improvement Industry Investigations Credit Reporting CASE 9

Figure 58 COMPARISON OF CONSEQUENCES OF NRC's PROPOSED DATA RULE

$80tM

$200

$180

$160 i" $140 --- ----- --- --- - ~----- ------ ------------ ------~-------- ----

~

w

, $120 *

..J

,ct

> $100

• t-z w $80 *

(/J w

0::

a. $60 *

$4Q. -

$20 *

$0 Base Case No Credit For Revised Elimination of Elimination of Availability Estimate of Event Tech Spec Improvement Industry Investigations Credit Reporting 10

DOCKETED IJfl Northern States Power Company US RC 1

Prairie Island Nuclear Generating Plant 1717 Wakonade Dr. East

'% JUN 11 P2 :32 Welch, Minnesota 55089 June 11, 1996 U.S. Nuclear Regulatory Commission DOCKET NLUBEAntt ~

Washington, DC 20555-0001 ATTN: Docketing and Service Branch PROPOSED RULE rn so

(<c:.lF~55\~'5 ~

~

Comments - Proposed NRC Rule Title 10 Section 50.76 The following comments and recommendations are being submitted in regard to proposed Nuclear Regulatory Commission (NRC) rules Title 10 section 50.76, published in the Federal Register on February 12, 1996 (Volume 61, No. 29, pages 5318-5326):

Summary We understand the Nuclear Regulatory Commission staffs concern for being able to compare the risk of all U.S. plants on an equivalent basis. We do not believe the proposed rule is the appropriate method of accomplishing that end. We consider the proposed regulation a significant increase in documentation and recordkeeping beyond current requirements with little additional value added beyond data submitted. We believe alternative methods are available as outlined below.

Docketed Information The Commission staff currently has docketed individual plant examination (IPE) reports for operating plants. Those docketed reports contain system and component reliability data, albeit not complete. The NRC already requires licensees to keep the data and analyses available for inspection of the PSAs and any application for which the licensee uses the PSA when dealing with the NRC. Thus, there appears to be no need for additional regulation.

Minimizing Costs The Commission staff states the intent to have data collected under this rule is essentially the same as would be required for monitoring reliability or availability for other purposes, e.g.

maintenance rule. In the case of our plant, we have established target out of service (00S) times as performance criteria based on consideration of - the individual plant examination input values, the desire to limit public risk, and sensitivity studies to determine the effect of out of service times. Reliability is based on the Maintenance Preventable Functional Failure as described in the NUMARC 93-01. This document was endorsed in the NRC Regulatory Guide for the Maintenance Rule.

The stated intent by the NRC staff is to "gather and report data without significant additional cost.* The current regulation, as proposed, imposes a significant additional recordkeeping burden without a significant benefit beyond that provided by the accepted method specified in the NUMARC guidance. At one facility, plant staff estimated that the proposed data rule would require about 7500 person-hours during the first year and about 2500 person-hours per year thereafter.

AclcnowtedgGd by earr1.. -?W l l 1995 ~

- *....-.,..,n,.,.,

'J Z IJC* EAR t1ECi,: :_,., ,vn* COt.1 ,,SSIOI\

OOCI< :TING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Dorunent Statistics Postmark Date """¥-.~._____ _ __

Coples Received--.1._ _ _ _ _ __

,\dd'l Coples Reproduced -=~"'--- - - -

Special Distrtbution~ Q<<ailllCs~ :lJ:j~-

Specific Aspects of the Proposed Regulation

{b)(1) Annual report compiled on the basis of calendar quarters Probabilistic safety assessment reliabilities can be related on an annual basis. Quarterly, or more frequent breakdowns do not appear to add significant worth to PSA calculations.

(i) Number of demands Demands of risk-significant equipment are often based on the surveillance testing required by Technical Specification requirements. Spurious and actual demands of equipment are treated equivalently in a PSA analyses. Licensee event reporting of such demands is often required. There is insufficient variation, or impact on risk, to warrant tracking demands in the detailed manner specified on a quarterly or annual basis.

• The regulation requires taking data for operating and shutdown cases, irrespective of whether the component has any risk significance at shutdown condition. In addition, regarding the regulatory guide there is a significant expansion of the data gathering to a point we believe is unwarranted.

Specifically, we are concerned with the urequiremeni- to take data for 3 separate shutdown states. Typically, many plants are operating with capacities above 80%. If one assumes equivalent risk from operating and shutdown conditions, it would appear nominally 8 to 12 times the data must be gathered for shutdown compared to operating states. This would appear to be an unnecessary burden.

Average annual demands is considered in the PSA and is often predictable. If reporting is required, only deviations from the norm need to be considered to give focus on unusual conditions.

(ii) Hours of operation following successful start Risk significant equipment is often run infrequently, but often for predictable times. We believe that deviation reporting is a more appropriate mechanism for tracking unusual conditions to ensure appropriate focus is made.

(iii) Number of hours unavailable We recommend the approach described under Alternative to Proposed Regulation be used.

(iv) Failure record We recommend the approach described under Alternative to Proposed Regulation be used.

(v) Concurrent unavailability This requirement, as written, does not appear to require any consideration of whether the system is risk significant under the condition being evaluated.

Clarification of the requirement should be made.

Alternative to Proposed Regulation We believe that the following approach would provide the NRC staff the information that they need to conduct benchmark Probabilistic Safety Analyses (PSA) for each plant. This approach would have minimal impact on the plant and NRC staffing.

1. For the currently effective IPE, have each licensee submit the events, related probabilities, system and component reliabilities assumed.

2. On an annual basis, have the licensees provide - the target OOS performance criteria times used for the risk-significant systems, cases where target levels are exceeded, changes to risk-significant systems, changes in requantified PSA data (if applicable),

failure reports for those cases where Nuclear Plant Reliability Database System (NPRDS) reports were not required. It is recognized that the NRC currently has access to failure reports submitted under NPRDS.

3. Data may be submitted electronically, e.g. by electronic mail using Microsoft Word, Excel, and Access documents as appropriate.

In essence, this approach would ensure that the NRC can perform their own benchmark analyses for each plant. Failure reporting would ensure that the NRC is aware of failures of risk significant systems and components.

Basis for Recommendation Sensitivity studies show that compliance with the target levels will ensure plant risk is kept within acceptable limits. Tracking Maintenance Rule Functional Failures places the appropriate emphasis on reducing failure rates, albeit in a qualitative, rather than quantitative, manner.

The NRC is proposing a Rule that will apply to all plants, regardless of each plant's level of implementation or plans of implementing PSA and risk-based initiatives. Some plants are pursuing PSA applications to reduce regulatory burden (e.g. Graded QA, Graded ISI/IST, and Risk-Based Tech. Specs}. However, the Rule assumes that all plants will eventually do this, which may not be the case. As a result, the NRC is potentially creating a burden on many, if not most, licensees. The decision to use PSA-based initiatives depends on plant economics and each licensee's management and business philosophy.

The data numbers, ( demands, failures, unavailabilities, etc.) are not what really tends to drive plant risk. What really drives plant risk are the plant configuration issues as:

• how many trains of redundancy for a given function exist to mitigate an event,

• whether there are diverse means of providing the same safety function (diesel and motor driven pumps vs. just motor driven, Containment spray and Containment Fan Coil Units (CFCUs) to provide containment pressure control vs. just one or the other), etc.

While it is true that each plant has a few unique equipment failure or maintenance unavailability frequencies that may affect plant risk, experience shows this would be a much smaller subset of data numbers than that being proposed under the Rule.

We consider day-to-day risk management provides the major benefit by keeping closer track of individual component/system/train failures and unavailabilities. However, that concern is already being more than adequately handled through each plant's NRC required actions to address configuration-specific risk under the Maintenance Rule and voluntary shutdown risk management activities.

With respect to shutdown reliability data reporting, plants without shutdown PSAs will likely not be using plant-specific shutdown PRA risk analyses in their discussions with the NRC. Thus,

there should be no need then to produce the extensive sets of data being interpreted in the regulatory guide.

Nuclear Energy Institute (NEI) Position We concur with the June 11, 1996 NEI letter that "the proposed rule is premature and should not be promulgated* at this time. We believe alternative methods are available to meet the NRC's data needs and request that our suggestions offered above be considered.

Request regarding Rule implementation and Basis for request If the NRC does decide to implement this rule as written, we request that implementation be delayed until January 1, 1998 with the first report submitted by January 31, 1999.

Experience in meeting the Generic Letter 88-01 and Maintenance Rule over the past 6 years has involved considerable data gathering. Maintenance Rule implementation has involved some aspects of the requirements (e.g. expert committee, PSA reviews). While we believe we have a respectable computer database system, we recognize that considerable changes must be made and that appropriate reviews as suggested by the regulatory guide will take longer than the current 6 months allotted by the NRC in their implementation plan.

Respectfully,

~~J.__ f's.<..--~~-

Joseph Alan Gonyeau ~

~

Senior Nuclear Consultant cc: R.O Anderson (NSP)

S.D. Floyd (NEI)

M.A. Wadley (NSP)

W.J. Hill (NSP)

OOCKE ED Attorney General U._ RC Betty D. Montgomery

.96 JUN 11 P2 :55 June 10, 1996 Via Federal Express DOCKET NlllERPR 5e>

U.S. Nuclear Regulatory Commission Washington, D.C. 20555-0001 ATTN: Docketing and Service Branch PROPOSED RULE

( b\ F'R. 65\~J Re: Federal Register Notice, Vol. 61, No. 29, Page 5318, 10CFRS0, RIN 31250-AF33 (February 12, 19%).

Dear Sir /Madame:

Enclosed please find an original and one copy of the Comments of the Ohio Utility Radiological Safety Board to be considered in the above-captioned matter.

Please acknowledge receipt by date-stamping the extra copy and returning it to me in the enclosed self-addressed, stamped envelope. Thank you for your attention to this matter.

Respectfully submitted, BETIY D. MONTGOMERY Attorney General of Ohio

~ J YL~lk STEVEN T. NOURSE Assistant Attorney General Public Utilities Section 180 East Broad Street Columbus, OH 43215-3793 (614) 466-4396 FAX: (614) 644-8764 STN/kk Enc.

cc: Dennis Allison "JUN f 9 1996 .

~edged by cardff............ llllllllNN . . . ..

State Office Tower/ 30 East Broad Street/ Columbus, Ohio 43215-3428 An Equal Opportunity Employer

@ Printed on Recycled Paper

U.S. NUC1. C ; * ** *_,,,LA..i OR GO MISS DOCK!: 11NG & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Document Statistia!

Postmark Date ~ -~"iG--- - - -

Coples Reailved_.__ _ _ _ _ __

.\dd'l Copes Reprodt£ed ~ - - - -

Special DislrttxJtion OJ;)*z¥ic ,---;:, ~ r

~l-~~ - - - - - - - -

BEFORE UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, DC 20555-0001 Reporting Reliability And Federal Register Notice, Availability Information For Vol.61, No. 29, Page 5318, Risk Significant Systems 10CFR50, RIN 31250-AF33, And Equipment (February 12, 1996)

COMMENTS OF THE OHIO UTILITY RADIOLOGICAL SAFETY BOARD I. INTRODUCTION These comments are submitted on behalf of the Utility Radiological Safety Board of Ohio (URSB), which consists of representatives from six state agencies: The Ohio Departments of Health, Agriculture and Commerce, The

- Ohio Emergency Management and Environmental Protection Agencies and The Public Utilities Commission of Ohio. The URSB was created in order to develop a comprehensive policy for the state regarding nuclear power safety, to promote safe, reliable and economical power, and to recommend policies and practices that promote safety, performance, emergency preparedness, and public health standards that are designed to meet Ohio's needs. The comments that follow are submitted in response to issue number one listed in the Notice of Proposed Rulemaking ("NPRM"). 61 Fed. Reg. at 5325.

II. DISCUSSION

1. The URSB supports the intent of the NPRM. However, the URSB would not support requiring licensees to resubmit data to the

NRC if the data is already provided under existing NRC rules. The URSB is in favor of the NRC Staff having complete access to all reliability and availability data on nuclear power plant systems and equipment, as collected by individual plants and their equipment suppliers and contractors. Such access to information will help the NRC Staff determine the future direction for their regulatory efforts, particularly now that the NRC is rightsizing operations. However, if such data is proprietary, then the NRC should make reasonable efforts to maintain the information it is given as proprietary, as permitted by law. Doing otherwise would jeopardize the relationships between vendors and owners, and increase the difficulty in owners maintaining and replacing equipment as it ages.

2. Since the 1980's, nuclear power plant operators have created a Nuclear Plant Reliability Data System (NPRDS) to which the NRC already has access. In addition, many industry associations, such as the Institute of Electrical and Electronics Engineers have collected data on nuclear power plant components, some of which predates equipment use at nuclear plants. Any requirement to put all of this accumulated data into a new format should have a demonstrable safety benefit improvement. If Institute for Nuclear Power Operations (INPO) agrees to upgrade the NPRDS to incorporate data required for probabilistic risk assessment, and make this available to the NRC, then a new data base system should not be needed.

3. Currently, Licensee Event Reports are submitted to the NRC by each plant as incidents arise. In addition, INPO produces Significant Operating Experience Reports (SOERs) and Significant Event Reports (SERs) to provide its members with lessons learned from the operating 2

experiences of its members. The URSB believes that the NRC should have access to these reports, with the appropriate proprietary safeguards, in order that the NRC can better focus its regulatory activities in a cost effective manner.

4. Therefore the URSB encourages the NRC to continue negotiating in earnest with INPO to make whatever improvements may be necessary to bring the NPRDS up to current requirements for probabilistic risk assessment determinations.

In closing, the URSB thanks the NRC for this opportunity to comment on the proposed rule.

Respectfully submitted, BETTY D. MONTGOMERY Attorney General of Ohio DUANE W. LUCKEY Section Chief STEVEN T. NOURSE Assistant Attorney General Public Utilities Section 180 East Broad Street Columbus, OH 43215-3793 (614) 466-4396 FAX: (614) 644-8764 3

BEFORE UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, DC 20555-0001 Reporting Reliability And Federal Register Notice, Availability Information For Vol.61, No. 29, Page 5318, Risk Significant Systems 10CFR50, RIN 31250-AF33, And Equipment (February 12, 1996)

COMMENTS OF THE OHIO UTILITY RADIOLOGICAL SAFETY BOARD I. INTRODUCTION These comments are submitted on behalf of the Utility Radiological Safety Board of Ohio (URSB), which consists of representatives from six state agencies: The Ohio Departments of Health, Agriculture and Commerce, The Ohio Emergency Management and Environmental Protection Agencies and The Public Utilities Commission of Ohio. The URSB was created in order to develop a comprehensive policy for the state regarding nuclear power safety, to promote safe, reliable and economical power, and to recommend policies and practices that promote safety, performance, emergency preparedness, and public health standards that are designed to meet Ohio's needs. The comments that follow are submitted in response to issue number one listed in the Notice of Proposed Rulemaking ("NPRM"). 61 Fed. Reg. at 5325.

IL DISCUSSION

1. The URSB supports the intent of the NPRM. However, the URSB would not support requiring licensees to resubmit data to the

NRC if the data is already provided under existing NRC rules. The URSB is in favor of the NRC Staff having complete access to all reliability and availability data on nuclear power plant systems and equipment, as collected by individual plants and their equipment suppliers and contractors. Such access to information will help the NRC Staff determine the future direction for their regulatory efforts, particularly now that the NRC is rightsizing operations. However, if such data is proprietary, then the NRC should make reasonable efforts to maintain the information it is given as proprietary, as permitted by law. Doing otherwise would jeopardize the relationships between vendors and owners, and increase the difficulty in owners maintaining and replacing equipment as it ages.

2. Since the 1980's, nuclear power plant operators have created a Nuclear Plant Reliability Data System (NPRDS) to which the NRC already has access. In addition, many industry associations, such as the Institute of Electrical and Electronics Engineers have collected data on nuclear power plant components, some of which predates equipment use at nuclear plants. Any requirement to put all of this accumulated data into a new format should have a demonstrable safety benefit improvement. If Institute for Nuclear Power Operations (INPO) agrees to upgrade the NPRDS to incorporate data required for probabilistic risk assessment, and make this available to the NRC, then a new data base system should not be needed.

3. Currently, Licensee Event Reports are submitted to the NRC by each plant as incidents arise. In addition, INPO produces Significant Operating Experience Reports (SOERs) and Significant Event Reports (SERs) to provide its members with lessons learned from the operating 2

experiences of its members. The URSB believes that the NRC should have access to these reports, with the appropriate proprietary safeguards, in order that the NRC can better focus its regulatory activities in a cost effective manner.

4. Therefore the URSB encourages the NRC to continue negotiating in earnest with INPO to make whatever improvements may be necessary to bring the NPRDS up to current requirements for probabilistic risk assessment determinations.

In closing, the URSB thanks the NRC for this opportunity to comment on the proposed rule.

Respectfully submitted, BETTY D. MONTGOMERY Attorney General of Ohio DUANE W. LUCKEY Section Chief STEVEN T. NOURSE Assistant Attorney General Public Utilities Section 180 East Broad Street Columbus, OH 43215-3793 (614) 466-4396 FAX: (614) 644-8764 3

DOCKET tUABER pn PROPOSED RULE n so _ DOCK ETED (b\rR.6 ~ ~J (!)

June 8, 1996 USN C US NRC Washington, DC 20555-0001 *96 JUN 11 P2 :56 ATTN.: Docketing and Service Branch OFFIC E o:: SECRETARY COMMENTS ON THE RULE RIN 3150-AF33 OOC KET l,'G & ..;crr*1CE REPORTING RELIABILITY AND AVAILABILITY INFORMA'B:F© *,:EOR RISK-SIGNIFICANT SYSTEMS AND EQUIPMENT The objective of proposed Rule is to obtain statistics on demands, failures, and unavailability periods of components for PRA-based decision making. The Rule and the Regulatory Guide 1042 require the industry to provide conditioned "summary" data with an unreasonable discretional power to select the data base and conditioning routines. The Rule does not provide any mechanism for the public's access to the data base, and the proposed mechanism for the NRC access to the data base is the on-site visits by NRC personnel with a "currently valid 0MB control number." The language of the Rule is intentionally obscure as manifested by the following passage:

Public Protection Notification The NRC may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid 0MB control number.

The Rule is an attempt to conceal the vital information from the public in blatant and criminal circumvention of the FOIA. The correct way of gathering information for PRA applications is processing the pool of all information obtained electronically in the real-time mode. The most economical method appears to be the installation of black boxes onsite (no burden to the industry). The second best method would be daily recording on CD at a cost of less than

$100/year. The present draft Rule and Regulatory Guide are unconstitutional and can not be published.

I

/J

) ,

Pfj--1-iJ..A Jnr. Zinovy V. Reytblatt fll{/4,l( 61 Myrtle Avenue Bridgeport, CT 06604 Tel.: (203)-366-7490 FAF: ( 2 0 3) - 3 6 6-7490 e-mail: zvr@aol.com

U.S. NUCLEAR REGULATORY COMMISSIOt--

OOCKETING & SERVICE SECTION OFFICE OF THE SECRET ARY OF THE COMMISSION Document Statistics Postmark Date ~  !:!4_;=+-~ ---~

Co~s Received __,_ \ _ _ _ _ _ __

,\dd1 Copies Reproduced -" .5:....-----

Speclal Distrlbu1ion ~

• mmo ;:B?C2 ,

Florida Power & Light Company, P.O. Box 14000, Juno Beach, FL 33408-0420 DOCKETED USt.!RC

'96 JU 11 P2 :56 JUN 6 1996 L-96-137 U. S. Nuclear Regulatory Commission DOCKET NUMBER PR Washington, DC 20555-0001 Attn: Docketing and Service Branch PROPOSED RULE § o (lo\r~S6\'8) G)

Subject:

Proposed Rule: Reliability and Availability Information for Risk-Significant Systems and Equipment 61 Fed. Reg. 5318 Request for Comments On February 12, 1996, the Nuclear Regulatory Commission (NRC) published a notice for public comment regarding the issuance of a proposed rule: "Reliability and Availability Information for Risk-Significant Systems and Equipment." These comments are submitted on behalf of Florida Power & Light Company (FPL), a licensed operator of two nuclear power plant units in Dade County, Florida and two units in St. Lucie County, Florida.

The Nuclear Energy Institute (NEI) and Nuclear Utility Backfitting and Reform Group (NUBARG) will be issuing comments on the subject proposed rulemaking. FPL endorses the NEI and NUBARG comments and recommendations.

FPL appreciates the opportunity to submit these comments.

Very truly yours,

~ .6'(\. ? ~ ~ .

~6\ W. H. Bohlke Vice President Nuclear Engineering WHB/msd an FPL Group company

U.S. NUCLEN, RE:GULATORY COMM!SSIOfl.

DOCKETING & SERVICE sec TION OFFICE OF THE SE~FJET ;,.RY OF THE COMMIS~iON Document Statistics Postmffl Date -"'l\~ -'-=-- - - - -

Co?Qs Received~ ~- -- - - -

.,dd'1 Copies Reproduced ~ ----

Special Distribution<::XQ. ; or, . ~~

JUNE 6, 1996 PAGE 1 DOCKETED Steven E. Farkas tJS RC 4901 Henican Place I 50 Metairie, LA 70003 (fo\Fe.53 \~ '96 JUN 10 P2 :53 U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 (9

OF FH~t CL SP' RETAPY Docketing and Service Branch: OOC KF_ T" * ' r~~ 1 iCf R.,N( I The following are comments on the proposed 50. 76, Reporting Re iability and Availability Information for Risk-significant Systems and Equipment. As advertised, this rule is a "means for the Commission to maintain, and in some cases improve, safety while reducing impacts on licensees as well as NRC resource expenditures, by focusing regulatory requirements and activities on the most risk-significant areas." The proposed rule seems to serve two other purposes. One, create a public database that unscrupulous people could draw unwarranted conclusions about the capital assets of a nuclear power plant owner. And two, a low cost way for the NRC to populate its Accident Sequence Precursor Event Analysis models by transferring those costs to the power plant licensees.

Threat to fiduciary duties Simply stated, it would be contrary to utility shareholders' interests to put data, identified by plant, in the public domain in a way that is "scrutable and accessible to the public."

The rule itself says that the data reported to the NRC may give "a plant specific indication of low reliability or availability for several systems [that] might indicate a programmatic problem with its attendant risk, ... " Later the proposal says: "it is expected that reliability and availability information for those systems and equipment will be well suited for identifying plants and systems at increased risk for significant events."

If the NRC or anyone else were to freely access this data and publish incorrect or unreviewed conclusions, there would be a direct negative effect on the financial health of the electric utility owner. Providing this commercially sensitive data would violate numerous fiduciary duties that the electric utility has to its shareholders. Disguising the plant name with Plant A, Plant B, etc., defeats the intent of a scrutable and accessible database. The proper location of such a database is in the INPO organization so that it is scrutable and accessible to the practitioners for their models.

Another hidden tax The NRC has spent licensee user fees to create its unvalidated Accident Sequence Precursor (ASP) and Event Analysis. Now it wants to "improve the plant-specific ASP models which the NRC uses to compute conditional core damage probability for determining the risk-significance of operational events." Anyone who has participated in model development has faced the daunting cost of gathering plant specific failure rate data. The NRC intends to pass this cost off onto the utility ratepayers and shareholders.

The rule would require the utility to gather data on risk significant safety functions that "could have a significant effect on risk" without defining risk. Core damage frequency is one measure of risk. However, the odds of a core damage event depend on the current status of the power plant, i.e., what is currently in-service and out-of-service for repair.

The point is, components that do not on average have high significance in terms of core damage frequency, could suddenly become very important on that particular day. So, even though the rule says that "data ... collected ... would be required for monitoring

U.S. NUCLEAR REGULATORY COMMIS81u~

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Do:ument Statistics Postmark Oat .~ ~.....,,.,- - - - -

Copies Received. _ _ _ _ _ __

Add'I Copies Reproduced ~ ~ - - -

Special DistributionC'Q?

• ae---. 1 3?~.>

JUNE 6, 1996 PAGE 2 reliability and/or availability ... for compliance with the [already mandated] maintenance rule" the scope must be much broader. A real plant specific data base would capture the configuration specific risk of normal operations and maintenance by collecting data on all components in the model. Should the NRC build such a real database, they would put upward pressure on power plant licensee fees to pay for resources needed to evaluate and disseminate the data collected. Those licensee fees are paid for by ratepayers and shareholders.

Conclusion The rule creates an unfunded mandate on the industry and its customers to pay for a database that violates the fiduciary duties of the licensees. The rule demonstrates no benefit to those customers or the public in general. No other commercial industry would tolerate the intrusive nature of this proposal. The current rules (such as 50.73 --

licensee event reports, and 50.65 -- maintenance rule) provide the NRC with more than

• enough information to peiform its industry oversight function .

Steven E. Farkas cc: The Honorable Robert Livingston U.S. House of Representatives

DOCKETED

.James A. Perry Consultant 200-1 Irene DrivJJfJ!tll&i;nsville, NY 13027 315-638-8350 FAX 315-638-5437

.% JUN - 6 P2 :2 0 May 31, 1996 Mr. John C. Hoyle Secretary U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 DOCKET NlUBER PROPOSED RULE 59 -

ATTENTION: Chief, Docketing and Services Branch ( \o\ rR-6~\-&) @

Dear Mr. Hoyle:

These comments are hereby submitted in response to the request for comments on the Nuclear Regulatory Commission's (NRC) proposed rule, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment," published in the Federal Register dated February 12, 1996 (ref. 61 Fed. Reg. 5318).

The proposed rule should not be promulgated since:

• it has not been justified,

• it is not needed by the NRC for oversight of current regulatory requirements, and

• the cost to the industry to implement the rule as proposed would be excessive.

The data reporting and retention requirements should be justified by being directly applicable to regulatory activities that are needed for the NRC to fulfill its mission to protect the health and safety of the public or that provide necessary additional protection under the backfit rule, 10CFR 50.109. The NRC has not prepared a backfit analysis for this proposed rule. In addition it is premature to require additional data reporting and retention requirements in advance of future risk-based regulations. It is suggested that the data reporting and retention requirements should be an integral part of the risk-based regulatory initiative under consideration so that all the associated costs and benefits can be determined and accurately assessed.

The proposed rule is not needed for the NRC to determine reactor licensees compliance with current regulatory requirements. For example NUREG-1460, Guide to NRC Reporting and Recordkeeping Requirements, already identifies the many different reporting and recordkeeping requirements that fulfill this purpose.

The cost to implement the proposed rule is considered excessive, in that it calls for the collection and reporting of actual demand information. This dramatically increases the burden of data collection and reporting. Reactor plant licensees typically estimate the number of demands on equipment. It is strongly suggested that since the reliability of risk-significant equipment in the industry is generally above 90% that any proposed rule limit the collection and reporting of data to information on those demands that result in failures.

JUN 1 9 1996-

!\cknowledged by card ..".....'"..."..""'" 'tH*

U.S. NUCl.tAH REl3ULAiO

• Y COMMlSSlOt--.

OOCKETING & SERVICE SECTION OFFICE Of THE SECRETARY OF THE COMMISSION DoCtJnent Statistics Postmark Date JJ~~- - - - -

Co~s Rea1ived____,.___ _ _ _ __

,\d<f1 Copes Repr'Xluced .; :, ,<_- - - -

Specla DtstrlbJtiOfl~,ocm I~

Secretary, US Nuclear Regulatory Commission May 31, 1996 Page 2 It is also suggested that consideration be given to not requiring any new reports to be submitted to the NRC. Have the licensees maintain the required information on site and available for NRC inspection. Submittals of failure data to NRC would be limited to the existing regulatory requirements.

In summary, the proposed rule should not be promulgated for the reasons stated above.

In addition, I strongly support the comments submitted by the Nuclear Energy Institute (NEI) to the Secretary, US Nuclear Regulatory Commission on this proposed rule.

Sincerely,

~ . {7, ~

11/);mes A Perry 1

Duquesne Lig1t Company Beaver Valey Power Station DOC t{ ET E0 P.O. eox 4 Shippingport, PA 15077* 0004 us~JRc f'i

• 96 APR -9 AlO :06 SUSHIL C. ~IN (412) 393-5512 Division Vice Prnldent Nuclear Services OF FICaaflch,.,[~,qfi.9.9~ y Fax (412) 643-8069 Nuclear Power Division OOCK ETf~-JG '/:RV ICE Desk Officer SR NCH Office of Information and Regulatoiy Affairs NEOB-10202 (3150-0011)

DOCKE1' NUMBER PR Office of Management and Budget PROPOSED RULE sol Washington, DC 20503 (Co\Fe.5~\~J 0

Subject:

0MB Supporting Statement for Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment, Proposed Rule Duquesne Light Company (DLC) is responsible for the operation of Beaver Valley Power Station Units 1 and 2. DLC has reviewed the supporting statement provided by the Nuclear Regulatoiy Commission (NRC) to the Office of Management and Budget (0MB) on the above referenced proposed rule and hereby submits the attached comments.

Thank you for the opportunity to comment on this issue. If you have any questions on this submittal, please contact Mr. R. G. Williams, Supervisor, Surveillance and Assessment, (412) 393-7554.

Sincerely,

~~

Sushil C. Jain c: -WQ~*ilt Reeeffl§ fflRlt~PBNINIF~)-,-

U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 DELIVERING kknowJBdgQCJ by ca QUALITY ENERGY

U.S. NUCLEA, RE'.;uLA,ORY COMMISSIOl'-

DOCKETING & SERVICE SECTION OFFIC OF lHE SECRETARY OF THE COMMISSION Doccmeiit " anstics Postmark Date~~,O t- S..o-QR)eh-Cop!es Recs:, d--si--- - - - - -

Add'l Copies Rsw _,.. ~J ... ...;;=--- -- -

Speclal Distribut;on Qrl),,.,;,..... ~ ,

1 ATTACHMENT DLC Comments on 0MB Supporting Statement

1. The burden estimates made in the Office of Management and Budget (0MB)

Supporting Statement are dependent on the following assumption made on page 3-2 of the Draft Regulatory Analysis:

"For the purpose of this regulatory analysis, it was assumed that 80 plants would already be collecting similar data and 30 would not."

Duquesne Light Company (DLC) believes that the Nuclear Regulatory Commission (NRC) may be overestimating the availability and reliability data that utilities are collecting in support of the Maintenance Rule. While 80, or more, of 110 plants are collecting various forms of availability and reliability data, it is believed that the data is primarily failure oriented, not success oriented as is the information requested by the proposed rule. DLC believes that few plants are planning to collect data similar to that requested by the proposed rule.

If the quoted assumption is inaccurate, then the burden estimates for implementation in the "Record.keeping Burden Table" would be underestimated.

2. The implementation burden estimate of 175 hours0.00203 days <br />0.0486 hours <br />2.893519e-4 weeks <br />6.65875e-5 months <br /> to "Develop data collection program" also does not seem to be realistic considering that program development will also include: review of the proposed rule, review of a new Regulatory Guide, possible review of a new Nuclear Energy Institute (NEI) guideline and development or revision of administrative procedures to implement the new regulation.

3. Page 3-10 of the Draft Regulatory Analysis indicates the NRc* expense for conducting "about two workshops" on proposed rule implementation. The burden estimates apparently do not include the utility expenses that would be incurred in supporting these workshops.

4. The Record.keeping Burden Table indicates a significant "Annual (recurring)"

burden estimate difference between the 80 plants and the 30 plants. DLC questions why a burden difference should exist on a recurring basis once all plants have developed their programs.

5. Paragraph 12 on page 4 of the 0MB Supporting Statement describes the implementation burden as being "annualized over three years." Given the current schedule of implementing the proposed rule on January 1, 1997, it would seem that the entire implementation burden would be incurred in 1996 in order to begin data collection by the scheduled date.

• ii Attachment (Continued)

DLC Comments on 0MB Supporting Statement Page 2

6. Having done a preliminary review of all the information provided, DLC suggests that the best way to minimize the regulatory burden for the NRC and utilities is to postpone issuance of the proposed rule and concentrate on enhancing the Institute of Nuclear Power Operations (INPO) Safety System Performance Indicators (SSPD Program to meet the NRC needs. The current SSPI systems could be supplemented from the "Basic Systems List" contained on page 5324 of the Federal Register, dated February 12, 1996. Modifications to the INPO SSPI Program should also be minimiud. If this approach is taken, the NRC and utility burden for collecting improved industry-wide data on the perfonnance of risk-significant systems will be minimized.

March 25, 1996 1996 APR -5 PM 3: 13 NOTE TO: Dennis Allison, AEOD FROM: Beth~.~

SUBJECT:

COMMENTS RECEIVED FROM DUQUESNE LIGHT COMPANY We just received these comments from Duquesne Light Company. In the case of the NEI comments, I told you that you could respond to their comments in the final rule, and in the 0MB clearance package that will probably have to be prepared for the final rule. However, in this case, these comments were sent to 0MB with a copy to NRC. It is extremely doubtful that 0MB will issue any approval for the Reliability Rule without NRC satisfactorily responding to these comments.

Therefore, please be prepared to respond to each issue in this letter when 0MB calls and requests either a written or oral response.

ROUTING AND TRANSMITTAL SUP 10: ~*me, office symbol, room number, Initials Date lldin1, Aceney/Post)

1. £/?7//e 2.

I.

4.

!Ad1on FIie Note and Return IAooroval For Clearance Per Conversation

~ Requested For Correction Prepare Reply Circulate For Your Information See Me r.omment Investigate S11nature l'.n,vdinatlon Justifv REMARKS

/~~.f~ c0/17/1?-"/?~

.c?/7 /,/~ /A~/3-/7/#/h,,

ve/lkchb'o O.f/'ecT (fr~ /e/Q'h~?--

lfH/4 C o/r=-~ S-.J'/~ i /2#. ~ /P?1 ~A:,L

~W-4/1/4/ fr ///e ~ _R/7~ f ~

.'1cy gt?~//' .Re- /4c,/4'~ .,4 /,6z_

~c#f/ r?X- ~ ~~r A'k-DO NOT use this form as a RECORD of approvals, concurrences, disposals, clear!'nces, and similar actions FROM: (Name, or,. symbol, Agency/Post) Room No.-Bld1.

fik__ AM>ne No.

IICMl-102 OPTIONAL FORM 41 (Rev. 7-76)

Prncrtbed by GM

• U.S. GPO: 1990 - 262-080 FPMR (41 CFR) 101-11.206

Clean Water ~ DOC ETEO US RC Fund of North Carolina DOCKET PUIBER P ~ RULE,..:..::::..,;5:.0

( Co I Fe~\"SJ

---

*-'96 APR -1 A8 :16 ATTN: Docketing & Service Branch*

US Nuclear Regulatory Commission Washington DC 20555-0001 The Clean Water Fund of North Carolina offers the following comments on the proposed reliability reporting rule, as noticed Monday, 12 February 1996, at 61 FR 5318ff The Fund considers that documentation, of potential continuing reliability problems, is especially important. The Fund notes that multiple failures of equipment (including not only the simultaneous failure of duplicate trains but also the simultaneous failure of independent systems) ~e not exceptionally uncommon. The Fund considers that information, provided by licensees to NRC under the proposed rule, should not be eligible .

for withholding under claims of privilege or confidentiality. In particular, the Fund strongly supports the concept that such information ~hould be "scrutable and accessible to the public." However, the Fund remains concerned about the NRC's potential abandonment. of the defense-in-depth" philosophy in favor of so-called "risk-based"*

regulation. The Fund also remains concerned that "generic issue resolution" or voluntary

• licensee action (based, for example, on regulatory guides) may not provide the public with an adequate margin of protection from particular problems.

Carl Rupert, Ph.D.

Raleigh Re~ar.ch Director

• The Clean Water Fund of North Carolina PO Box 1008 Raleigh NC 27602 Raleigh Office: P.O.Box 1008, Raleigh, NC 27 602

• 919-832-7491 A heville Office: 29 1/2 Page Ave., Asheville, l~C 28801

• 704-251-1291

U.S. NUCLEAR REGU TOAY COMMISSIOt,.

OOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Document Statistics Po8bnm Date ~~:::t1'l.- - - --

Copes Rece Adcfl Copies R SpeclaDisbi

~

OOC KE TEO USt'R" NUCLEAR ENERGY INSTITUTE

.96 HAR 19 P2 :20 William H. Rasin c.r r-i]. R~ T;t RY S £ r-JtO ~ VIC E PPl ~ I0fr-Jl.

OF FICE ll CHN ,CA, , , ,cu,A, mv DOCK ET Ir!*~-" Sl:R ,jCf:

March 12, 1996 8HA!lC11 Information and Records Management Branch (T-6F33)

DOCKET NlMBERPR U.S. Nuclear Regulatory Commission PROPOSED RULE so Washington, DC 20555-0001 (lo\F"e.Sa \~J @

These comments are submitted by the Nuclear Energy Institute (NEI) 1, on behalf of the nuclear energy industry, in response to the request for comments on the Nuclear Regulatory Commission's (NRC) proposed rule, "Reporting Reliability and Availability Information for Risk-Significant Systems and Equipment," published in the Federal Register dated February 12, 1996 (ref. 61 Fed. Reg. 5318).

In summary, we believe the proposed rule is premature and should not be promulgated by the NRC. Our view is based on the following points:

• The additional data reportmg and recordkeeping is not needed by the NRC for assessing compliance with any current regulatory requirement.

• The estimate of industry burden provided in the supporting statement to 0MB is grossly inaccurate. The rule would actually impose an excessive burden on industry.

With regard to the first point above, the reporting and recordkeeping of reliability and availability information is not necessary for the NRC to oversee the implementation of current regulatory requirements by reactor licensees. There are already hundreds of different reporting and recordkeeping requirements that fulfill this purpose (ref. NUREG-1460, Guide to NRC Reporting and Recordkeeping Requirements). As stated in the Federal Register notice, the NRC believes" ... the information is necessary to substantially improve the NRC's ability to make risk-effective decisions consistent with the Commission's policy statement on the use of probabilistic risk assessments (PRA)." It should be noted, however, that there is no regulatory requirement for licensees to conduct PRAs today. Thus, the proposed rule has no statutory basis and would require recordkeeping and reporting 1

NEI is the organization responsible for establishing unified nuclear industry policy on matters affecting the nuclear energy industry, including the regulatory aspects of generic operational and technical issues. NEl's members include all utilities licensed to operate commercial nuclear power plants in the United States, nuclear plant designers, major architect/

engineering firms, fuel fabrication facilities, materials licensees, and other organizations and individuals involved in the nuclear energy industry. NEI is the successor organization to the Nuclear Management and Resources Council (NUMARC).

Acknowledged by card .. ~~~..~..  !..!~ . *J 1 776 I STRE ET . NW SUI TE 400 WASHING TO N. DC 20006-37CE PH O N E 202739 . 8088 FAX 202 785 I 898

U.S. NUCLEAR REGULATORY COMMISSIOt-.

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Docunent Statisb Postmm Date:i\W: .Q, ~3=>--t>g..

Coples Received_.___ _ _ _ __

Add'! Coples Reprodixed """'"-- - - -

Special DislritxstionCQ;;) ,a,;,.., ~ 1

J U.S. Nuclear Regulatory Commission March 12, 1996 Page 2 information for use in an analytical tool that is not required by regulation. Its promulgation, in and of itself, does not result in any improvement to public health and safety.

We believe that data reporting and recordkeeping requirements should be directly associated with regulatory activities that are necessary for the NRC to fulfill its statutory mission to protect public health and safety or that provide substantial additional protection under the provisions of the backfit rule, 10 CFR 50.109. The NRC has not prepared a backfit analysis for the proposed rule. Furthermore, we believe it is premature for the NRC to require additional data reporting and recordkeeping in advance of future risk-based regulations or regulatory activities.

Any additional data reporting or recordkeeping requirements should be integrated and in step with the risk-based regulatory initiative under consideration so that the associated costs and benefits can be assessed accurately.

Our second point above addresses the estimate of industry burden imposed by the proposed rule. In the supporting statement provided to 0MB, it is assumed that 80 of 110 licensees are collecting, or plan to collect, similar reliability and availability information. A basis for this assumption is not provided.

In reality, few, if any, licensees collect the information as described in the proposed rule. There are two major differences from what licensees typically collect and what the proposed rule would require. These major differences involve the collection and reporting of demand information and information on the concurrent unavailability of two or more systems or trains.

- Regarding demand information, the proposed rule calls for the collection and reporting of actual demand information, whereas the industry typically estimates the number of demands on equipment. By requiring actual demand information that is equivalent to the information collected for actual failures, the burden of data collection is dramatically increased. For example, for equipment that is tested on a monthly basis, the information associated with each of the twelve demands in a year would be collected and reported under the proposed rule. Today, if one of those twelve demands resulted in a failure, the only equivalent information that would be collected would be on the demand that resulted in the failure. Thus, for this example, the proposed rule would increase the burden of data collection by greater than a factor of ten, just on test demand information alone. Given that the reliability of risk-significant equipment in the industry is generally well above 90%,

and that the proposed rule calls for information on both failures and successes on all types of demands (test, inadvertent, or actual need), the actual burden of recordkeeping and reporting on demand information is substantially greater than the estimate provided to 0MB.

U.S. Nuclear Regulatory Commission March 12, 1996 Page 3 The second key difference involves the collection of the number of hours when two or more trains from the same or different systems are concurrently unavailable, as would be required under the proposed rule. Current industry practice focuses on monitoring the number of hours that a single train is unavailable, not on the overlap of hours when two or more trains are unavailable. This requirement would be particularly burdensome during refueling outages, when several plant systems are removed from service for maintenance. Again, the supporting statement to 0MB has greatly underestimated this burden by using an inaccurate assumption.

Given the two major differences discussed above as well as several less significant differences in the information required by the rule with current industry practice, we believe the estimate of burden provided in the supporting statement to 0MB is fundamentally flawed and grossly inaccurate. In our estimation, the burden of the recordkeeping and reporting imposed on licensees would be 2 to 7 times of that estimated on an annual/recurring basis. In monetary terms, this would equate to a total annual recurring industrywide burden of approximately $30 to $105 million.

The annual cost range is broad as some licensees may choose to automate the data collection while others may choose to have a manual system. These estimates do not include any costs associated with the installation and maintenance of equipment such as total run-how; meters.

In closing, we reiterate our belief that the proposed rule is premature, that additional data reporting is not needed by NRC for oversight of current regulatory requirements, that the burden imposed on the industry would be excessive, and therefore, that the proposed rule should not be promulgated.

- Sincerely,

~~_:_

William H. Rasin ARP/jes

DOCKETED USHRC

.96 MAR -7 P1 :08 5- -R TARY OFFICfTOtfG i \ERVICE DOCK BRANCh DOCKET NU.4BER PR PROPOSED RULE .5?

(~I Ae.55\"B.) (J) mrr r:r 1996.:.__

Acknowledged by card ............. A"""""..,,,

U.S. NUCLEAR tb..ivU-1 TORY COMMISSIO~

DOCKETING & SERVICE SECTION OFFICE OF THE SECRETARY OF THE COMMISSION Document Statistics Postmark Date -=-~...w....,' - - - - - -

Copies Received_.___ _ _ _ __

Add'l Copies Reproduced ~ - - - -

Special Distribution Cm.~ 1

~

DOCK E. TED us~ RC [7590-01-P]

NUCLEAR REGULATORY COMMISSION '96 FEB - 7 P3 :09 10 CFR PART 50 DOCKET NlNBERPR ----

RIN 3150-AF33 PROPOSED RULE 50 (lo\~R..5~\'a)

Reporting Reliability and Availability Information for Risk-significant Systems and Equipment AGENCY: Nuclear Regulatory Commission.

ACTION: Proposed rule.

SUMMARY

The Nuclear Regulatory Commission (NRC) is proposing to amend its regulations to require that licensees for commercial nuclear power reactors report plant-specific summary reliability and availability data for risk-significant systems and equipment 1 to the NRC. The proposed rule would also require licensees to maintain on site, and to make available for NRC inspection, records and documentation that provide the basis for the summary data reported to the NRC. The systems and equipment for which data would be In relation to this proposed rule, the term equipment is intended to apply to an ensemble of components treated as a single entity for certain probabilistic risk assessments (PRAs) where a system or train treatment would not be appropriate.

2 provided are a subset of the systems and equipment within the scope of the maintenance rule.

The Commission has determined that reporting of reliability and availability information is necessary to substantially improve the NRC's ability to make risk-effective regulatory decisions consistent with the Commission's policy statement on the use of probabilistic risk assessments (PRAs) (August 16, 1995; 60 FR 42622). This would assist the NRC in improving -

its oversight capabilities with respect to public health and safety and becoming more efficient by focusing its regulatory program on those issues of greatest risk significance and reducing unnecessary regulatory burdens on licensees. The Commission would use the data that would be required by the proposed rule in generic issue resolution, developing quantitative indicators that can assist in assessing plant safety performance, performing risk-based inspections, and pursuing modifications to specific plants and basic regulations and guidelines. Furthermore, this information would improve the NRC's oversight of licensees' implementation of the maintenance rule. It would also enhance licensees' capabilities to implement the evaluation and goal-setting activities required by the maintenance rule by providing licensees with access to current industry-wide reliability and availability information for some of the systems and equipment within the scope of the maintenance rule.

DATES: Comments regarding any aspect of the proposed rule are due to the

~ \ \) \9C:, ~

Commission by {insert date 12~aflcf' publication iR th~ Federa-l Register). Comments received after that date will be considered if it is

3 practical to do so, but the Commission can give no assurance of consideration for late comments. The Commission intends that this expiration date will be at least 30 days after publication of an associated draft regulatory guide for *.

public comment.

In addition, comments regarding the collection of information, including the burden estimate and suggestions for reducing the burden, should be submitted to the Office of Management and Budget (0MB), and to the NRC, by (insert date 30 days after publication in the Federal Register). For further information see the discussion below under the heading Paperwork Reduction Act Statement.

ADDRESSES: Mail written comments to: U.S. Nuclear Regulatory Commission, Washington, DC 20555-0001, ATTN.: Docketing and Service Branch. Deliver written comments to the NRC at One White Flint North, 11555 Rockville Pike, Rockville, MD, between 7:30 am and 4:15 pm on Federal workdays.

Send comments regarding the collection of information, including the burden estimate and suggestions for reducing the burden, to: (1) Desk Officer, Office of Information and Regulatory Affairs, NEOB-10202 (3150-0011),

Office of Management and Budget, Washington, DC 20503, and (2) Information and Records Management Branch {T-6F33), U.S. Nuclear Regulatory Commission, Washington, DC 20555-0001. For further information see the discussion below under the heading Paperwork Reduction Act Statement.

4 Copies of the draft regulatory analysis, the supporting statement submitted to the 0MB, and comments received may be examined, and/or copied for a fee, at:

The NRC Public Document Room, 2120 L Street NW. (Lower Level), Washington, DC. ,

FOR FURTHER INFORMATION CONTACT: Dennis Allison, Office for Analysis and Evaluation of Operational Data, U.S. Nuclear Regulatory Commission, Washington, DC 20555-0001, Telephone (301) 415-6835.

SUPPLEMENTARY INFORMATION:

BACKGROUND Current Requirements.

There are no existing requirements to systematically report reliability and availability information; nor is there an industry-wide database to provide such information.

Current reporting requirements in 10 CFR 50.72, "Immediate notification" and 10 CFR 50.73, "Licensee event report system," require the submittal of extensive descriptive information on selected plant and system level events.

The Nuclear Plant Reliability Data System, a data base that industry supports and the Institute for Nuclear Power Operations (INPO) maintains, provides data on component engineering characteristics and failures. Neither of these sources includes all the data elements (i.e., number of demands on a system, number of hours of operation, and information on maintenance unavailability)

5 that are needed to determine the reliability and availability of systems and equipment. Maintenance effectiveness monitoring requirements in 10 CFR 50.65, "Requirements for Monitoring the Effectiveness of Maintenance at Nuclear Power -*

Plants", also do not contain reporting requirements.

In recent years, plants have performed Individual Plant Evaluations (IPEs), as requested in Generic Letter 88-20 and its supplements, and submitted the results to the NRC. These submittals provide measures of risk such as core damage frequency, dominant accident sequences, and containment release category information. While system and component reliability data have been collected as part of some utility IPEs, this information is typically not included in the IPE submittals to the NRC.

Prior Efforts.

In late 1991 and through 1992, the NRC staff participated on an INPO-established NRC/industry review group to make recommendations for changes to the Nuclear Plant Reliability Data System (NPRDS). The group's final recommendations to INPO to collect PRA-related reliability and availability data would have provided most of NRC's data needs. However, INPO took no action on these recommendations.

During 1992 and 1993, the NRC staff continued through correspondence and meetings to outline the particular data needed and to seek INPO's assistance in obtaining the data. In a December 1993 meeting with NUMARC (now the Nuclear Energy Institute (NEI)), INPO representatives suggested their Safety

6 System Performance Indicator (SSPI) as a surrogate for reliability data. They proposed expanding the indicator to additional systems and indicated that data elements could be modified to compute actual reliability and availability data. Although general agreements were reached with INPO on which systems and components and what types of data elements are appropriate for risk-related applications and maintenance effectiveness monitoring, no voluntary system of providing data resulted from these discussions. In the fall of 1994, the NRC staff began work on this rulemaking action. In June 1995, NEI proposed to discuss a voluntary approach of providing reliability and availability data to the NRC based on SSPI data. The NRC staff will continue to work with industry on voluntary submittal of reliability data, under a program that will meet the needs of all parties, while at the same time proceeding to obtain public comment on this proposed rule.

Industry representatives have expressed concern that reliability data, if publicly available, would be subject to misuse. In certain circumstances it is permissible for the NRC to withhold information from public disclosure.

For example, pursuant to 10 CFR 2.790(b)(l), a licensee may propose that a document be withheld from public disclosure on the grounds that it contains trade secrets or privileged or confidential commercial or financial information. However, the data that would be reported under this proposed rule would not appear to qualify for withholding. Reliability data used as input to risk-based regulatory decisions should be scrutable and accessible to the µublic. The Commission's PRA policy statement indicates that appropriate supporting data for PRA analyses that support regulatory decisions should be publicly available. Similarly, the Commission's draft report on public

7 responsiveness (March 31, 1995; 60 FR 16685) indicates that the policy of the NRC is to make information available to the public relating to its health and safety mission, consistent with its legal obligations to protect information and its deliberative and investigative processes. Commenters who believe that there is information subject to a proper 10 CFR 2.790(b)(l) withholding determination requested by the proposed rule should provide a specific justification for such belief.

Move to Risk-Based Regulation.

For severai years the Commission has been working towards increased use of PRAs in power reactor regulation. In its policy statement on the use of PRAs, the Commission has indicated that the use of PRA technology should be increased in all regulatory matters to the extent supported by the state-of-the-art in terms of methods and data, and this implies that the collection of

• equipment and human reliability data should be enhanced. Implementation of these policies would improve the regulatory process through (1) improved risk-effective safety decision making, (2) more efficient use of agency resources, and (3) reduction in unnecessary burdens on licensees. These improvements would enhance both efficiency and safety.

The data reported under this proposed rule would improve the NRC's oversight capability with respect to public health and safety by focusing the NRC's regulatory programs in a risk-effective manner. Generally, the NRC's ability to identify plants and systems at increased risk for significant events and, thus, to take appropriate action would be substantially improved.

8 For example, a generic indication of low reliability or availability for a system might indicate a technical problem, with its attendant risk, that may warrant generic action. Similarly, a plant-specific indication of low reliability or availability for several systems w.ight indicate a programmatic problem, with its attendant risk, and may warrant plant-specific action.

It has been noted that prior to some significant events (such as the scram failure at Salem and the accident at Three Mile Island) there was previously existing information (such as challenge data and reliability data for scram breakers and power operated relief valves) which, if collected, recognized, and acted upon might have led to preventive actions. Accordingly, it is expected that reliability and availability information for selected risk-significant systems would improve the NRC's oversight capability with respect to public health and safety - i.e., the ability to maintain or enhance safety by identifying and reviewing indications of increased risk and, if appropriate, taking generic or plant-specific action.

Such problems could be subtle in nature. For instance, licensee(s) might schedule train outages for maintenance at certain times, such that risks are substantially increased over what would be expected based on random outages. This situation would not be indicated by current reporting requirements, or even by simply reporting train unavailability, but it could be indicated by the concurrent unavailability of two or more trains, as would be reported under the proposed rule. Additional examples discussed below describe further specific uses of the data that would help to enhance safety.

9 In order to move towards risk-based regulation and the increased use of PRA information, the NRC needs scrutable, plant-specific and generic reliability and availability information. The framework for an overall move towards risk-based regulation involves the development of a regulatory process. This process includes operational procedures and decision criteria that require credible PRA methods, models, and data. This framework would provide for predictable, consistent, and objective risk-based regulatory decision making. The data that would be reported under this rule represent one of the needed elements. In addition, these data are needed to improve the efficiency and effectiveness of NRC regulatory applications that employ a risk-based perspective in advance of defining the entire framework.

Generally, plant-specific information is needed because there can be wide plant-to-plant variations in the design, importance, reliability and availability of particular systems and equipment. It is necessary to identify similar equipment in various plants so that the data can be properly grouped and analyzed to estimate overall industry performance and plant-specific performance and to identify outliers (good or bad). 2 2

For many of the systems involved, plant specific demand and failure data will be sparse, at least initially. Until data have been collected for some time, it will be necessary to use data from similar equipment, applications, and environments at several plants in order to obtain practical estimates of reliability and uncertainty. Even when sufficient plant-specific data exist to estimate plant performance, comparison to industry or group averages is often desirable.

10 Some examples of how reliability and availability information would be used to improve current NRC regulatory applications that consider risk in the decision process are discussed below. One of the examples involves the need ,

for information to support generic regulatory actions - i.e., generic issue resolution and its associated rulemaking or regulatory guide revision.

Another example involves the need for information to determine whether further NRC action is needed at specific plants - i.e., indicators of plant performance. Some involve a mixture of plant specific and generic elements.

For example, analyzing an event at a given plant could lead to a plant-specific action such as a special inspection and/or to a generic action such as a bulletin or generic letter.

Generic Issue Resolution. The NRC currently uses risk estimates in:

(1) prioriti~ing safety issues, (2) deciding whether new requirements or staff positions to address these issues are warranted, and (3) deciding whether proposed new requirements or staff positions should be implemented. Knowing the current, updated reliability and availability of key systems would, in

• some cases, lead to a better understanding of the risk in these areas and, thus, to more risk-effective decisions. This should both enhance public protection and reduce unnecessary regulatory burdens. Generic data would usually suffice for this purpose; however, in some cases the data would need

~

to be divided to account for specific classes or groups of plants.

Indicators of Plant Performance. PRA models with plant-specific reliability and availability data would be used to develop indicators of plant performance and trends in plant performance which are more closely related to

11 risk than those currently in use. These new indicators would replace some of those currently in use and thereby enhance NRC's ability to make risk-effective decisions with regard to identifying plants for increased or decreased regulatory attention. For example, it is important to detect situations where an individual plant may be having reliability or availability problems with multiple systems.

Accident Sequence Precursor (ASP) and Event Analysis. Plant-specific, train-level reliability and unavailability data would be used to improve the plant-specific ASP models which the NRC uses to compute conditional core damage probability for determining the risk-significance of operational events. In addition, dates and causes of equipment failures would be used to identify common cause failures and to compute common cause failure rates for input to these models. Improving these methods would enhance the staff's ability to make risk-effective decisions about which events warrant further

• inspections or investigations and/or generic actions such as bulletins and generic letters. Plant-specific data are needed to better understand an event and calculate the associated conditional core damage probability. It is also useful to identify systems that have the most influence on the results. Then the risk associated with the potential for similar events at other plants, which may be known to have low reliability for the key systems, can be considered in determining whether further actions are warranted.

Risk-based Inspections. Current and upd~ted system reliability, availability and failure data in a generic and plant-specific risk-based context would be used to enhance the staff's ability to plan inspections

12 focused on the most risk-significant plant systems, components, and operations. While generic data would be used in developing risk-based inspection guides and a framework for inspections, plant-specific data would be used to focus and optimize inspection activities at specific plants. For example, an individual plant may have an atypical reliability problem with a specific risk-significant system and thereby warrant additional attention. In addition, special studies can be conducted to determine the root cause of reliability problems by comparing the characteristics of plants that have these problems with those that do not .

.8.gj_ng. Equipment reliability data would help identify equipment that is being degraded by aging and define the extent and the risk-significance of aging problems.

Another class of examples involves the need for information to evaluate anticipated cost beneficial licensing actions, where the rationale is that risk permits reductions in previous margins of safety or less prescriptive requirements without adverse impact on overall safety. The NRC is actively pursuing a variety of modifications to the basic regulations and guidelines that govern the operation of commercial nuclear power reactors. These modifications are characterized by allowing individual licensees to utilize insights from plant-specific risk evaluations to reduce or remove current requirements that are found to have low risk-significance. Current regulatory requirements under consideration for risk-based modification include those

13 prescribing quality assurance, in-service inspection, in-service testing, and surveillance testing. It is anticipated that a significant number of additional requests will be received that rely upon risk-based arguments.

These changes could adversely affect the level of safety achieved by the plants if the risk evaluations are flawed or the changes are improperly executed or the changes involve synergistic effects that are not covered by the risk models or captured by historical data. Current, plant-specific

• reliability and availability data would help the NRC monitor the licensees' programs to maintain safety while reducing regulatory burdens. Relaxation of undue regulatory burdens then can proceed with confidence that there will be appropriate feedback to assure that the level of safety is not being degraded.

Some examples are discussed below.

Risk-based Technical Specification. Technical Specification requirements specify surveillance intervals and allowed outage times for

• safety equipment for the various modes of plant operation. It is anticipated that licensees will request a number of relaxations in surveillance intervals and allowed outage times. Current, plant-specific reliability and availability data would help the NRC monitor performance for the systems and equipment subject to the proposed rule. Thus, proposed relaxations of surveillance intervals and allowed outage times for such systems could be evaluated more effectively based on past performance and on confidence that there would be appropriate feedback to ensure that performance is not being degraded. In addition, failure rates from actual demands will be used to verify that failure rates estimated from testing are approximately the same.

14 Inservice Testing. Inservice testing requirements, which are based on the provisions of the American Society of Mechanical Engineers Boiler and Pressure Vessel Code (ASME Code}, measure the functional characteristics of equipment performance, such as pump flow, in order to detect degradation. The ASME and licensee owners' groups are working toward establishing risk-based frequencies for inservice testing, based on plant-specific risk ranking methodologies. Changes in testing frequency can affect reliability in many ways. For example, less frequent valve testing might lead to an increase in the demand failure rate because the valve actuating mechanism tends to bind or freeze after extended periods of idleness. However, using plant-specific demand failure and unavailability data, proposed changes can be more effectively evaluated based on the risk-significance and performance of plant systems and based on confidence that there will be appropriate feedback to assure that the level of safety is not being degraded.

NRC Maintenance Rule.

The maintenance rule, 10 CFR 50.65, was issued on July 10, 1991 (56 FR 31306). The reliability and availability information that would be required by the proposed reporting rule would improve the NRC's oversight of licensees' implementation of the maintenance rule. It would also enhance licensee's capabilities to implement the evaluation and goal-setting activities required by the maintenance rule by providing licensees with access to current industry-wide reliability and availability information for some of the systems and equipment within the scope of the maintenance rule.

15 NRC Monitoring. As discussed above, current plant-specific data can provide feedback on the effectiveness of licensee programs, including maintenance programs. Accordingly, these data would improve the NRC's monitoring ability by providing risk-based measures of the effectiveness of individual licensee maintenance programs and the overall effectiveness of the maintenance rule.

• In addition, the NRC has expressed concern about the extent to which some reactor licensees are taking systems and equipment out of service for maintenance during plant operation. Although this practice may offer economic benefits by reducing plant downtime, it must be properly managed to assure that safety is not compromised. It should be noted that licensees are required by 10 CFR 50.65(a)(3) to periodically conduct assessments and make adjustments to ensure that the objective of preventing failures through maintenance is appropriately balanced against the objective of minimizing

• unavailability due to monitoring and preventive maintenance. The NRC would use the hours when any two or more trains from the same or different systems are concurrently unavailable to monitor how well licensees are managing the risk associated with such maintenance. As discussed below, under "Licensee Implementation," the data would also enhance licensees' capabilities to make prudent on-line maintenance decisions.

The maintenance rule is also important to license renewal (10 CFR Part 54). Hence, improving the NRC's oversight of the maintenance rule could strengthen one of the bases for the scope of the license renewal rule.

16 Licensee Implementation. In connection with the NRC's PRA policy, the NRC staff has defined the data elements that would improve the evaluation of maintenance and has established that they are the same as those needed to support a transition toward a risk- and performar.ce-based regulatory process.

The NRC believes that the reliability and availability data that would be required by this rule would enhance licensee's capabilities to implement the evaluation and goal-setting activities required by the maintenance rule by providing licensees with access to current industry-wide reliability and availability information for some of the systems and equipment within the scope of the maintenance rule. 3 3

The systems and equipment covered by this proposed rule are a subset of the systems and equipment within the scope of the maintenance rule. The data elements are more extensive than what would be required for compliance with the maintenance rule; however, for the systems covered, these data elements

• would serve to improve implementation of the maintenance rule. To cite one example, under 10 CFR 50.65(a)(2), risk-significant systems may be considered to be subject to an effective preventive maintenance program and, thus, not subject to condition or performance monitoring unless "maintenance preventable" failures occur. However, gathering the reliability and availability information specified in this proposed rule, including data elements such as concurrent outages and the causes of failures, would provide a better picture of a system's performance and the effectiveness of the preventive maintenance program than simply awaiting the occurrence of "maintenance preventable" failures.

17 In some circumstances, the maintenance rule requires licensees to establish performance or condition goals, taking into account industry-wide operating experience where practical. It also requires periodic program evaluations, including considerati~n of unavailability due to monitoring or preventive maintenance, taking industry-wide operating experience into account, where practical. Licensees will need to monitor reliability and availability of risk-significant systems, particularly for the periodic

• program evaluations. 4 For many of the systems involved, plant-specific demand and failure data will be sparse, ~t least initially. However statistical analysis techniques exist that allow a licensee to analyze and evaluate data from similar equipment, applications and environments from other plants, besides the data from their plant. These analyses yield meaningful reliability estimates for the subject plant that can be compared with performance goals. Industry-wide data would also provide a practical source for comparing plant-specific performance with industry operating experience. Although plant-specific 4

NUMARC 93-01, which the NRC has endorsed as describing one acceptable way of meeting the requirements of the NRC's maintenance rule, indicates in Section 12.2.4 that the adjustment for balancing of objectives needs to be done for risk-significant structures, systems, and components (SSCs).

However, for other SSCs it is acceptable to measure operating SSC performance against overall plant performance criteria and standby system performance against specific performance criteria. This is reasonable in that, for systems that are less risk-significant, the expense of a rigorous balancing is not warranted.

18 information is generally available on site, and utilities review licensee event reports and other generic event information, NRC site visits, associated with early efforts to prepare for maintenance rule implementation in 1996, indicate that utilities do not use industry operating experience in a systematic and consistent way for goal setting purposes under the maintenance rule. Based on these considerations, the availability of current, industry-wide reliability and availability data would enhance licensee's capabilities to implement the evaluation and goal-setting activities required by the maintenance rule.

As discussed previously, the NRC has recently found cause for concern about how some reactor licensees handle on-line maintenance. Prudent on-line maintenance decisions depend on a full appreciation of the risk-significance of taking equipment out of service (individually or collectively) and use of plant-specific and generic reliability and availability data would play a significant role in improving such decision making.

Description of Proposed Rule.

The proposed rule would require holders of operating licenses fof nuclear power reactors to report reliability and availability data for certain risk-significant systems and equipment. The proposed reporting requirements would apply to the event-mitigating systems and equipment which have or could have a significant effect on risk in terms of avoiding core damage accidents or preserving containment integrity. Summary information reported to the NRC would be:

19

1. The number of demands, the number of failures to start associated with such demands, and the dates of any such failures, characterized according to the identification of the train affected, the type of demand (test, inadvertent/spurious, or actual need), and the plant mode at the time of the demand (operating or shutdown);

2. The number of hours of operation following each successful start, characterized according to the identification of the train affected and whether or not the operation was terminated because of equipment failure, with the dates of any such failures;

3. The number of hours equipment is unavailable, characterized according to the identification of the train affected, the plant mode at the time equipment is unavailable (operating or shutdown), characterization of the unavailable period (planned, unplanned, or support system unavailable), and, if due to a support system being unavailable, identification of the support system;

4. For each period equipment is unavailable due to component failure(s), a failure record identifying the component(s) and providing the failure date, duration, mode, cause, and effect; and

5. The number of hours when two or more trains from the same or different systems were concurrently unavailable, characterized according to the identification of the trains that were unavailable.

20 The first annual report would identify the systems, trains, and ensembles of components covered by the reporting requirements of the rule; subsequent annual reports would either state that no changes were made subsequent to the previous annual report or describe the changes made.

The summary information would be reported annually and compiled on the basis of calendar quarters, or on a more frequent basis at the option of each individual licensee. Records and documentation of each occurrence of a demand, failure, or unavailable period that provide the basis for the summary data reported to the NRC would be required to be maintained on site and made available for NRC inspection.

In developing these data elements the NRC has, over the past three years, reached a consensus on the minimum data needed to support risk-based applications and enhance implementation of the maintenance rule. During this period NRC staff has also interacted extensively with INPO and NEI in an effort to define the minimum reliability and availability data needed to

• satisfy the needs of both NRC risk-based regulatory applications and industry (licensee) uses of PRA.

The number of demands and the number of successful starts are needed to estimate demand reliability, i.e., the fraction of demands that result in successful starts. (The complement of this fraction provides an estimate of the probability of failure on demand). The actual number of demands and successes, as opposed to the ratio, is needed for purposes such as: (1)

21 providing a measure of confidence in the results and (2) permitting proper combination of data from different plants.

The type of demand is needed to determine whether or not the demand reliability estimated by testing is approximately the same as the demand reliability for actual demands. Sometimes it is not, indicating a need for additional data analysis in making reliability estimates.

The plant mode at the time of a demand is needed to estimate the demand frequency, demand reliability, and unavailability according to plant mode.

These factors, as well as the risk associated with unreliability and unavailability, can be quite different depending on whether the plant is in operation or shut down.

The hours of operation following successful starts are needed to estimate the probability the equipment will function for a specified period of time. This information is needed for systems that must operate for an extended period following an accident to fulfill a risk-significant safety function.

The number of hours that equipment is not available {unavailable hours) is needed to estimate the fraction of time that a train is not available to perform its risk-significant safety function. For some systems this can be an important or dominant contributor to the overall probability of failure to perform the system's safety function. It can be significantly affected by elective maintenance.

22 The type of unavailable hours (planned or unplanned) is needed to effectively utilize these estimates. For example, a high unplanned unavailability may indicate a need for more preventive maintenance; a high planned unavailability may indicate the opposite.

The unavailable hours due to support systems failure or unavailability are needed to properly capture concurrent outages and to eliminate double counting. For example, an Emergency Service Water (ESW) train being unavailable may result in other trains being unavailable as well; however, for purposes of estimating risk in a PRA study, that unavailability should not be counted more than once.

The date of each failure is needed to allow screening for potential common cause failures. Failures that occur closely together in time warrant review to see whether a common cause failure may be involved. Common cause failures may indicate a need for revised maintenance procedures or staggered testing. Common cause failure rates are also needed for PRA models because of

• their importance in system reliability and availability estimates.

Failure cause and failure mode information are needed to support common cause failure analysis as discussed above and to associate the failure with the correct failure mode for input into PRA models.

Quarterly data are needed to conduct first order trending studies to identify areas of emerging concern with regard to overall plant and system

23 performance. More frequent compilation is acceptable at the discretion of each licensee.

An identification of the systems, trains, and ensembles of components subject to the rule is needed because identification of the components within the systems, trains, and ensembles is necessary for proper use and evaluation of the data by the staff and for industry wide generic applications to account

• for physical differences between plants. For example, simplified system diagrams could be marked to show the systems, trains, and ensembles against which the data would be reported.

Retention of records and documentation that provide the bases for the summary data report to the NRC for a period of several years is consistent with maintenance rule applications. For example, monitoring reliability for a few years may be used to determine trends in order to achieve the balance described in 10 CFR 50.65(a)(3) - i.e., the balance between preventing failures through maintenance and minimizing unavailability due to monitoring and preventive maintenance. In addition, on-site data are needed to provide a scrutable basis for regulatory decisions. For example, it is expected to be necessary to review the actual unavailable hours in order to estimate the mean repair times for key components for the purpose of updating the staff's PRA models.

24 Regulatory Guide.

A new regulatory guide will be prepared and issued to provide supplementary guidance. The guide will present an acceptable way to define the systems and equipment subject to the rule and it will provide risk-based definitions of failure as well as train and system boundaries consistent with PRA applications. The format in which data would be provided to the NRC and a suggested format for maintaining on-site documentation and record keeping would be included. In order to reduce costs, use of electronic data submittal will be considered a priority objective in developing and implementing the guide. A draft guide will be published for comment before it is finalized. A public workshop is planned after publication of the draft guide. The comment period for this proposed r~le will not expire until at least 30 days after publication of the draft regulatory guide.

Definitions. The basic definitions used in reporting under§ 50.76 are discussed below; further details will be addressed in the regulatory guide.

• For example, the basic definition of failure is provided here; further details, such as how to handle a case where the operators prematurely terminate system operation following a real demand, will be discussed in the regulatory guide. In particular, the regulatory guide will define risk-significant safety function{s) and failures for systems and equipment covered by this proposed rule.

Demand is an occurrence where a system or train is called upon to perform its risk-significant safety function. A demand may be manual or

25 automatic. It may occur in response to a real need, a test, an error, an equipment malfunction or other spurious causes. For the purposes of r~porting under this rule, the demands of interest are those which are actual demands or-closely simulate actual demands for the train or specific equipment involved.

Failure, for the purpose of reporting under this rule, is an occurrence where a system or train fails to perform its risk-significant safety function.

A failure may occur as a result of a hardware malfunction, a software malfunction, or a human error. Failures to start in response to a demand are reported under paragraph 50.76(b}(l}(i}. Failures to run after a successful start are reported under paragraph 50.76(b)(l){ii).

Unavailability is the probability that a required system or train is not in a condition to perform or is not capable of performing its risk-significant safety function. This may result from failure to start, from failure to run,

• or from intentional or unintentional removal of equipment from service (e.g.,

for maintenance or testing}.

Risk-significant safety function is a safety function that has or could have a significant effect on risk (in terms of avoiding core damage accidents or preserving containment integrity for the purposes of reporting under this proposed rule}.

Reportable systems and equipment are the event-mitigating systems and equipment which have or could have a significant effect on risk in terms of avoiding core damage accidents or preserving containment integrity. The

26 reportable systems and equipment will be determined by each licensee. The regulatory guide will describe acceptable methods for making that determination.

It is expected that the rule will produce a set of basic systems for which reliability data will be reported for all plants that have them.

However, these basic systems are not sufficient by themselves. Additional systems and equipment to be addressed will depend on plant-specific features.

Listed below is the set of basic systems that the Commission is currently considering for identification in the draft regulatory guide.

BASIC PWR SYSTEMS BASIC BWR SYSTEMS Auxiliary feedwater Reactor core isolation cooling or isolation condenser High pressure safety injection Feedwater coolant injection, high pressure coolant injection or high pressure core spray, as appropriate Reactor protection Reactor protection Low pressure safety injection Low pressure coolant injection and low pressure core spray Emergency ac power Emergency ac power

27 As discussed above, the systems and equipment to be included in the scope of the rule would be those event-mitigating systems and equipment that have or could have a significant effect on risk in terms of avoiding core damage accidents or preserving containment integrity. To ensure that this approach is consistent with operating experience, the NRC has considered the systems and equipment that have been substantially involved in significant events in U. S. reactors. These systems were found to fall into the following categories:

1. Basic systems. As indicated above, the NRC expects that these systems would be included in the scope of the rule for all plants. The basic systems on the proposed list have been confirmed to have been substantially involved in significant events.

2. Plant-specific systems. Systems such as service water and component cooling water are risk-significant, but the significance varies widely, depending upon plant-specific designs. It is expected that these systems will be included, as appropriate, based on plant-specific PRA studies. Other systems, such as containment purge, appear infrequently in connection with significant events and are not expected to be risk-significant for any plants.

3. Initiating systems. Systems such as main feedwater and offsite power are primarily considered to be initiators of significant events, rather than mitigation systems. Existing reporting requirements in 10 CFR 50.72 and 10 CFR 50.73 provide enough information to characterize the important initiating systems for the purpose of PRA studies.

28

4. Non-measurable items. Items such as reactor coolant system corrosion are not amenable to meaningful measurement by the methods of this proposed rule.

Based on this review, the systems and equipment to be included in the scope of the rule are considered reasonably consistent with operating experience in terms of involvement in significant events. Accordingly, it is expected that reliability and availability information for those systems and equipment will be well suited for identifying plants and systems at increased risk for significant events.

Minimizing Costs. The NRC intends that the data required to be collected and reported under this proposed rule be essentially the same as would be required for monitoring reliability and/or availability for other purposes, such as monitoring system reliability where that is the option chosen for compliance with the maintenance rule. Thus, it should be practical to gather and report the data without significant additional cost. This will be a priority goal in developing the guidance to be included in the new regulatory guide.

Sunset Provision. As experience is gained with implementing the proposed rule and utilizing the information required to be collected and reported, a reassessment may be necessary or desirable. One way of assuring such a reassessment would be to include a "sunset provision" in the rule, whereby the rule would automatically expire after a specified period of time unless: (i) a condition specified in the rule is fulfilled, or (ii) the

29 Commission engages in a rulemaking which extends the effectiveness of the rule. The Commission requests public comments on whether the proposed rule should contain such a sunset provision, and if so, the period of time after which the rule should automatically expire.

Grandfather Provision. There may be some plants for which, at the time that the proposed rule may be adopted by the Commission as a final rule, licensees have already announced plans to discontinue operation in the near future. Furthermore, licensees may determine in the future to discontinue operation at some plants. In either case, there may be less reason to require collection and reporting of the information contemplated by the proposed rule at such plants and it may be advisable to exempt such plants from the information collection and reporting requirements of the proposed rule (i.e.,

"grandfathering"). The Commission requests public comments on whether the proposed rule should exempt plants that have announced (or will announce) plans to discontinue operation within a short time (e.g., two years).

Conclusion.

As discussed under the subject "Move to Risk-Based Regulation," the information to be collected under the proposed rule is necessary for the development and implementation of risk-based regulatory processes. Risk-based regulatory approaches provide a means for the Commission to maintain, and in some cases improve, safety while reducing impacts on licensees as well as NRC resource expenditures, by focusing regulatory requirements and activities on the most risk-significant areas. In addition, this information would improve

30 the NRC's oversight of licensees' implementation of the maintenance rule. It would also enhance licensee's capabilities to implement the evaluation and goal-setting activities required by the maintenance rule by providing licensees with access to current industry-wide reliability and availability information for some of the risk-significant systems and equipment within the scope of the maintenance rule. The Commission has also prepared a regulatory analysis (see "Regulatory Analysis") which identified alternatives for collecting the information for use by both licensees and the NRC, and evaluated the costs of each viable alternative. Based upon these factors, the Commission believes that the costs of the proposed rule's information collection and reporting requirements are justified in view of the potential safety significance and projected benefits of the information in NRC regulatory activities.

Submission of Comments in Electronic Format Commenters are encouraged to submit, in addition to the original paper

• copy, a copy of their comments in an electronic format on IBM PC DOS-compatible 3.5- or 5.25-inch, double-sided, diskettes. Data files should be provided in WordPerfect 5.0 or 5.1. ASCII code is also acceptable, or if formatted text is required, data files should be submitted in IBM Revisable Format Text Document Content Architecture (RFT/DCA) format.

Environmental Impact: Categorical Exclusion

31 The proposed rule sets forth requirements for the collection, maintenance, and reporting of reliability and availability data for certain risk-significant systems and equipment. The NRC has determined that this proposed rule is the type of action described in categorical exclusion, 10 CFR 51.22(c)(3)(ii). Therefore, neither an environmental impact statement nor an environmental assessment has been prepared for this proposed regulation.

Paperwork Reduction Act Statement This proposed rule amends information collection requirements that are subject to the Paperwork Reduction Act of 1995 (44 U.S.C. 3501 et seq.). This rule has been submitted to 0MB for review and approval of the Paperwork Reduction Act requirements.

The public reporting burden for this collection of information is estimated to average 1375 hours0.0159 days <br />0.382 hours <br />0.00227 weeks <br />5.231875e-4 months <br /> per response (i.e., per commercial nuclear power reactor per year), including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. The Commission is seeking public comment on the potential impact of the collection of information contained in the proposed rule and on the following issues:

1. Is the proposed collection of information necessary for the proper performance of the functions of the NRC, and does the information have practical utility?

32

2. Is the estimate of burden accurate?

3. Is there a way to enhance the quality, utility, and clarity of the information to be collected?

4. How can the burden of the collection of information be minimized including by using automated collection techniques?

Send comments on any aspect of this proposed collection of information, including suggestions for reducing the burden, to the Information and Records Management Branch (T-6-F33), U.S. Nuclear Regulatory Commission, Washington, DC 20555-0001, and to the Desk Officer, Office of Information and Regulatory Affairs, NEOB-10202, (3150-0011), Office of Management and Budget, Washington, DC 20503.

Comments to 0MB on the collections of information or on the above issues should be submitted by (insert date 30 days after publication in the Federal Register). Comments received after this date will be considered if it is practical to do so, but assurance of consideration cannot be given to comments received after this date.

Public Protection Notification The NRC may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid 0MB control number.

33 Regulatory Analysis The Commission has prepared a draft regulatory analysis on this proposed*

regulation. The analysis examines the costs and benefits of the alternatives considered by the Commission. The draft analysis is available for inspection in the NRC Public Document Room, 2120 L Street NW. (Lower Level), Washington, DC. Single copies of the draft analysis may be obtained from: Dennis Allison, Office for Analysis and Evaluation of Operational Data, U.S. Nuclear Regulato-ry Commission, Washington, DC 20555-0001, Telephone (301) 415-6835.

Regulatory Flexibility Certification In accordance with the Regulatory Flexibility Act of 1980 (5 U.S.C. 605 (B)), the Commission certifies that this rule will not, if promulgated, have a significant economic impact on a substantial number of small entities. The proposed rule affects only the licensing and operation of nuclear power plants. The companies that own these plants do not fall within the scope of the definition of "small entities" set forth in the Regulatory Flexibility Act or the size standards adopted by the NRC on April 11, 1995 (60 FR 18344 - 10 CFR 2.810.

Backfit Analysis The proposed rule sets forth requirements for reporting and record keeping. The NRC has determined that the backfit rule, 10 CFR 50.109, does not apply to this proposed rule, and therefore, a backfit analysis is not

34 required for this proposed rule because these amendments do not involve any provisions which would impose backfits as defined in 10 CFR 50.109(a)(l).

However, as discussed above in "Regulatory Analysis," the Commission has prepared a regulatory analysis which summarizes the purpose and intended use of the information proposed to be collected, identifies alternatives for collection and reporting of the proposed information, and identifies the impacts and benefits of the alternatives.

This. regulatory analysis constitutes a disciplined process for evaluating the potential benefits and projected impacts (burdens) of information collection and reporting requirements such as the proposed rule.

The Commission therefore concludes that the objective underlying the Commission's adoption of the Backfit Rule - that regulatory impacts are assessed under established criteria in a disciplined process - is being met for this proposed rule.

List of Subjects in 10 CFR Part 50 Antitrust, Classified information, Criminal penalties, Fire protection, Intergovernmental relations, Nuclear power plants and reactors, Radiation protection, Reactor siting criteria, Reporting and record keeping requirements.

For the reasons set out in the preamble and under the authority of the Atomic Energy Act of 1954, as amended, the Energy Reorganization Act of 1974,

35 as amended, and 5 U.S.C. 553, the NRC is proposing to adopt the following amendments to 10 CFR Part 50.

PART 50 DOMESTIC LICENSING OF PRODUCTION AND UTILIZATION FACILITIES

1. The authority citation for Part 50 continues to read as follows:

AUTHORITY: Sections 102, 103, 104, 105, 161, 182, 183, 186, 189, 68 Stat. 936, 937, 938, 948, 953, 954, 955, 956, as amended, sec. 234, 83 Stat.

1244, as amended (42 U.S.C. 2132, 2133, 2134, 2135, 2201, 2232, 2233, 2236, 2239, 2282); secs. 201, as amended, 202, 206, 88 Stat. 1242, as amended, 1244, 1246 (42 U.S. C. 5841, 5842, 5846).

Section 50.7 also issued under Pub. L.95-601, sec. 10, 92 Stat. 2951 as amended by Pub. L. 102-486, sec. 2902, 106 Stat 3123, (42 U.S.C. 5851).

Section 50.10 also issued under secs. 101, 185, 68 Stat. 936, 955, as amended (42 U.S.C. 2131, 2235); sec. 102, Pub. L.91-190, 83 Stat. 853 (42 U.S.C.

4332). Sections 50.13, and 50.54(dd), and 50.103 also issued under sec. 108, 68 Stat. 939, as amended (42 U.S.C. 2138). Sections 50.23, 50.35, 50.55, and 50.56 also issued under sec. 185, 68 Stat. 955 (42 U.S.C. 2235). Sections 50.33a, 50.55a and Appendix Q also issued under sec. 102, Pub. L.91-190, 83 Stat. 853 (42 U.S.C. 4332). Sections 50.34 and 50.54 also issued under sec. 204, 88 Stat. 1245 (42 U.S.C. 5844). Sections 50.58, 50.91, and 50.92 also issued under Pub. L.97-415, 96 Stat. 2073 (42 U.S.C. 2239). Section 50.78 also issued under sec. 122, 68 Stat. 939 (42 U.S.C. 2152). Sections 50.80 -

36 50.81 also issued under sec. 184, 68 Stat. 954, as amended (42 U.S.C. 2234).

Appendix Falso issued under sec. 187, 68 Stat. 955 (42 U.S.C. 2237).

2. Section 50.8 (b) is revised to read as follows:

§ 50.8 Information collection requirements: Office of Management and Budget approval.

(b) The approved information collection requirements contained in this part appear in§§ 50.30, 50.33, 50.33a, 50.34, 50.34a, 50.35, 50.36, 50.36a, 50.48, 50.49, 50.54, 50.55, 50.55a, 50.59, 50.60, 50.61, 50.63, 50.64, 50.65, 50.71, 50.72, 50.75, 50.76, 50.80, 50.82, 50.90, 50.91, 50.120, and Appendices A, B, E, G, H, I, J, K, M, N, 0, Q, and R.

3. Section 50.76 is added to read as follows:

§ 50.76 Reporting reliability and availability information for risk-significant systems and equipment.

(a) Applicability. This section applies to all holders of operating licenses for commercial nuclear power plants under 10 CFR 50.21b or 50.22 and

37 all holders of combined operating licenses for commercial nuclear power plants under 10 CFR 52.97.

(b) Requirements.

(1) Each licensee shall submit an annual report to the NRC that contains the following information, compiled on the basis of calendar quarters, or on a more frequent basis at the option of each licensee, for systems, trains, and ensembles of components in paragraph (b)(3) of this section:

(i) The number of demands, the number of failures to start associated with such demands, and the dates of such failures, characterized according to the identification of the train affected, the type of demand (test, inadvertent/spurious, or actual need), and the plant mode at the time of the demand (operating or shutdown);

(ii) The number of hours of operation following each successful start, characterized according to the identification of the train affected and whether or not the operation was terminated because of equipment failure, with the dates of any such failures; (iii) The number of hours equipment is unavailable, characterized according to the identification of the train affected, the plant mode at the time equipment is unavailable (operating or shutdown), characterization of the unavailable period (planned, unplanned, or support system unavailable), and,

38 if due to a support system being unavailable, identification of the support system; (iv) For each period equipment is unavailable due to component failure(s), a failure record identifying the component(s) and providing the failure date, duration, mode, cause, and effect; and (v) The number of hours when two or more trains from the same or different systems were concurrently unavailable, characterized according to the identification of the trains that were unavailable.

(2) The initial annual report described in (b)(l) above shall identify the systems, trains, and ensembles of components covered by paragraph (b)(3) below; subsequent annual reports shall either state that no changes were made subsequent to the previous annual report or describe any changes made.

(3) The requirements of paragraphs (b)(l) and (b)(2) of this section

• apply to those event-mitigation systems, and ensembles of components treated as single entities in certain probabilistic risk assessments where a system or train treatment would not be appropriate, which have or could have a significant effect on risk in terms of avoiding core damage accidents or preserving containment integrity.

(4) Each licensee shall maintain records and documentation of each occurrence of a demand, failure, or unavailable period that provide the basis for the data reported in paragraph (b)(l) of this section on site and

. 39 available for NRC inspection for a period of 5 years after the date of the report specified in paragraph (b)(l) of this section.

(c) Implementation. Licensees shall begin collecting the information required by paragraph (b) of this section on January 1, 1997, and shall submit the first report required by paragraph (b)(l) of this section by January 31, 1998. Thereafter, each annual report required py paragraph (b)(l) of this section shall be submitted by January 31 of the following year.

,:i Aci Dated at Rockville, MD, this _er

_ _ day of ; . , : ~ , 1996.

For the Nuclear Regulatory Commission, I

s:~retary of the Commission.

~/

p(f~~

I- i FY xtg'35

/7 L v t ~ ~

~~ j,)ce_, c/4-tlt~ ~~1/U/f h :

~ - t5 ()

{; / 1:::-f< .5 3 1 7

United States Nuclear Regulatory Commission Office of Public Affairs Washington, DC 20555 Phone 301-415-8200 Fax 301-415-2234 Internet:opa@nrc.gov No. 96-36 FOR IMMEDIATE RELEASE (Friday, February 16, 1996)

NRC PROPOSES REPORTING REQUIREMENTS FOR RISK-SIGNIFICANT REACTOR SYSTEMS The Nuclear Regulatory Commission is proposing to amend its regulations to require that operators of commercial nuclear power plants provide NRC with reliability and availability information for selected risk-significant equipment.

The proposed reporting requirements would apply to systems

• and equipment which could significantly affect risk in terms of avoiding core damage accidents or preserving containment integrity. Examples of components affected include reactor protection systems, high-pressure and low-pressure safety injection systems, and emergency power.

The NRC believes the information is needed to improve its oversight of public health and safety and to assist the agency in taking generic and plant-specific regulatory actions. The collection of equipment reliability data would enhance NRC's ability to identify plants and systems at risk for significant events, particularly at a time when it is moving toward increased use of probability risk assessments.

Information required under the proposed rule also would improve the NRC's oversight of licensees' maintenance programs under the agency's maintenance rule issued in 1991.

Data to be reported under the proposed rule would be:

1) Number of demands on a system or component, the number of failures to start associated with such demands, and the dates of any such failures, the type of demand (test, inadvertent/spurious, or actual need), and plant mode at the time (operating or shut down);

2) Number of hours of operation following each successful start of the system or component and whether or not the operation was terminated because of equipment failure, with dates of any failures;

3) Number of hours equipment is unavailable, plant mode at the time, and description of unavailable period (planned, unplanned);

4) Failure record identifying the component and failure date, duration, mode and cause;

5) Number of hours when two or more significant parts from the same or different systems were concurrently unavailable.

The information would be reported annually to the NRC, but compiled by calendar quarter, or on a more frequent basis at the option of each licensee. The first report would be required by January 31, 1998.

Under the rule, licensees also would be required to maintain on site, for NRC inspection, the records and documentation that provide the basis for the information reported to the NRC.

Interested persons are invited to submit written comments on the proposed rule, published in the February 12 edition of the Federal Register, by June 11. The comments should be addressed

'to the Secretary, Nuclear Regulatory Commission, Washington, DC 20555-0001, Attention: Docketing and Service Branch.