NUREG/CR-5499, Informs That Final Rept, Rates of Initiating Events at Us Nuclear Power Plants: 1987-1995, NUREG/CR-5499 Completed. Objective of Study,To Update Initiating Event Frequency Estimates Based on Operating Experience from Event Repts

Informs That Final Rept, Rates of Initiating Events at Us Nuclear Power Plants: 1987-1995, NUREG/CR-5499 Completed. Objective of Study,To Update Initiating Event Frequency Estimates Based on Operating Experience from Event Repts

ML20202B059
Person / Time
Issue date:	01/25/1999
From:	Rossi C NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:	Holahan G, Silber J, Strosnider J NRC (Affiliation Not Assigned)
References
RTR-NUREG-CR-5499 NUDOCS 9901280355
Download: ML20202B059 (12)

Text

_ _ _ _ - _ _ _ _ - - _ - - _ - - _ - - - _ - - - - - - -

s

} p*,

49 \4 UNITED STATES

~

! j NUCLEAR REGULATORY COMMISSION f WASHINGTON, D.C. 20555-0001 January 25, 1999

% *sse*f MEMORANDUM TO: Jacqueline E. Silber, Director, DISP:NRR Gary M. Holahan, Director, DSSA:NRR

}

Jack R. Strosnider, Director, DE:NRR David B. Matthews, Director, DRPM:NRR Suzanne C. Black, Acting Director, DRCH:NRR Michael E. Mayfield, Acting Director, DET:RES Thomas L. King, Director, DST:RES John W. Craig, Director, DRA:RES Wayne D. Lanning, Director, DRS:RGN-l

' ,t A. Randolph Blough, Director, DRP:RGN-l Bruce S. Mallett, Director, DRS:RGN-Il j Loren R. Plisco, Director, DRP:RGN-ll John A. Grobe, Director, DRS:RGN-Ill Geoffrey E. Grant, Director, DRP:RGN-Ill Arthur T. Howell, Director, DRS:RGN-IV Kenneth E. Brockman, Director, DRP:RGN-IV '

FROM: Charles E. Rossi, DireC l Safety Programs Division Office for Analysis and Evaluation of Operational Data

SUBJECT:

ISSUE OF FINAL REPORT- RATES OF INITIATING EVENTS AT U.S. NUCLEAR POWER PLANTS: 1987-1995 (NUREG/CR-5499) r The final report, " Rates of initiating Events at U.S. Nuclear Power Plants: 1987-1995" NUREG/CR-5499, has been completed. The report will be distributed to you by NRC Publications. The objective of this study was to update the initiating event frequency estimates based on operating experience from licensee event reports during the time period 1987 through 1995. For rare events such as loss-of-coolant accidents (LOCAs), data beyond the 1987-1995 period was used. The report also includes a comparison with initiating event frequency estimates published in probabilistic risk assessments (PRAs) and individual plant examinations (IPEs), an evaluation of the most significant trends, and an evaluation of dominant contributors /

to risk-significant initiators.

/

A draft of this report was provided to the Office of Nuclear Reactor Regulation (NRR), the Office of Nuclear Regulatory Research (RES), the regions, and industry and public interest V

organizations for peer review and comment. The summary of resolutions to comments is

attached.

h CONTACTS: Steve Mays, RRAB, SPD, AEOD (301-415-7496)

T D4R 7,_;

a gi fyp Y Don Marksberry, RRAB, SPD, AEOD (301- 415-6378) y 79 - C/8 gp J 9901280355 990125 %

-54 PDR Y ll b d O'Sl

l Multiple Addressees 2 l

A number of comments from internal reviews of similar reports indicated a difficulty in obtaining qualitative insights and related detailed information from the reports, especially for inspection-related activities. To help the staff to better identify and relate this detailed information to various risk-important reculatory applications, we have provided a foreword to the report that provides directions to the .elevant quantitative and qualitative information contained in the report. The foreword also indicates the appropriate type of engineering review of this information needed for application on a plant-specific basis. However, it should be recognited that this report is not a handbook for inspection. It is a source of information that can be used 1 to focus both generic and plant-specific risk-informed activities involving unplanned automatic and manual reactor trips. Cognizant RRAB staff will be available to consult with users of this report.

Notable observations and finoings of the study include the following:

• Combined initiating events frequencies for all initiators are lower than the frequencies used in NUREG-1150, Severe Accident Risks: An Assessment for Five U.S. Nuclear Power Plants, and IPEs by a factor of five and four, respectively. j e General transients constitute 77% of allinitiating events. Events that pose a more severe challenge to the plant's mitigation systems (nongeneral transients) constitute the remaining 23%.

• Over the nine-year span considered by this report, either a decreasing or constant time trend was observed for all categories of events. A decreasing trend was identified in i over two/ thirds of the more risk-significant (with respect to core damage frequency) categories that had sufficient data for trending analysis. The overallinitiating event frequency decreased by a factor of two to three during the nine-year span. Most risk-significant (with respect to core damage frequency) initiators frequencies (such as total loss of feedwater flow, loss of instrument or control air, inadvertent closure of all main steam isolation valves, and total loss of cendenser heat sink for BWRs) decreased at a faster rate than the overall initiating event frequency.

• Loss-of-coolant accident frequencies are lower than those used in NUREG-1150 and industry-wide IPEs. For LOC.A initiators, the frequencies were evaluated using data and ,

information prior to 1987 due to their relative low frequency and the corresponding l sparsenes of data. No pipe break LOCA events were fouM in the U.S. operating i experience. For small pipe break LOCA frequency, the estm.ute from WASH-1400, Reactor Safety Study, was updated using U.S. reactor experience. For medium and large pipe break LOCAs, frequency estimates were calculated by using the frequency of leaks or through-wall cracks that have occurred which challenge the piping integrity. I Further, conservative estimates were used for the probability of break given a leak (based on a technical review of information on fracture mechanics, data on high energy pipe failures and cracks, and assessment of pipe break frequencies estimated by others since WASH-1400).

Multiple Addressees 3 These findings are discussed in more detail in the report. Graphical and tabular displays, along with specific discussions, are included. Categorical listings of the reactor trip events used in the ,

analyses are included in an appendix in the report. For a perspective on the implications of l these initiating event frequencies on overall plant risk, it is necessary to also consider other l factors such as system and component reliabilities and common-cause failure probabilities. l The paper " Indications of U.S. Nuclear Industry Trends from the Risk-based Analysis of i Operating Experience," presented at PSAM4 provides some perspective on the implications of the findings of this report with respect to overall risk. l Based on these findings and conclusions, the following can be used to improve risk-informed regulatory activities:

o Based on knowledge from the operating experience and the need to provide updated frequencies for NRC PRA programs, the task to update pipe break LOCA frequency estimates was included as an objective of this report. The goal of this effort is to refine the original estimates based on operating experience and current knowledge of pipe break mechanisms. It is recognized that the approach in this report will result in ,

reduction of unnecessary conservatism in LOCA frequency estimates. However, the result is still conservative. We recommend that RES conduct analyses to develop best estimates of pipe break LOCA frequencies and their associated uncertainties. This evaluation should consider current operating, surveillance, and maintenance experience at U.S. nuclear power plants along with advances in the science of crack initiation and growth from various mechanisms. These best estimates of pipe break LOCA frequencies and their associated uncertainties would benefit future PRA analyses and provide a bases for re-evaluating priorities of open LOCA-related generic safety issues. j l

e NUREG-1489, A Review of NRC Staff Uses of Probabilistic Risk Assessment, provides i guidance on PRA use in risk-informed regulatory activities, such as issue screening and analysis. This guidance recommends the use of up-to-date PRA information, such as ,

operational exponence and data.. Future updates to Standardized Plant Analysis Risk models should consider, as appropriate, incorporation of the updated frequencies from this report to make these models more realistic. In addition, the staff should consider, as appropriate, use of these updated frequences in regulatory analyses of the risk '

significance of issues under Regulatory Guide 1.174, An Approach for Using Probabilistic Risk Assessments in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis; NUREGICR-0058, Regulatory Analysis Guidelines of the U.S.

Nuclear Regulatory Commission; and other risk-informed activities. 1 Attachment As stated cc w/att.:

A. Thadani, RES i S. Colles, NRR R. Zwnmerman, NRR B. Sheron, NRR B. Boger, NRR E. Adensam, NRR .

J l

l j, . .

Multiple Addressees 4 MEMORANDUM DATED: Janudry 25,.1999 l

SUBJECT:

ISSUE OF FINAL REPORT- RATES OF INITIATING EVENTS AT U.S. NUCLEAR l POWER PLANTS: 1987-1995 (NUREG/CR-5499) i Distribution w/att,:

l Tign:centerD JLarkins, ACRS JStc'z, NRR LLund, RES t RRAB R/F MMarkley, ACRS GParry, NRR PWilson, NRR l SPD R/F ARubin, RES LMarsh, NRR DCoe, NRR l Public _ JFlack, RES JTrapp, RI

- REmrit, RES MRubin, NRR TShediosky, RI _

FCongel MCurmingham, RES AEl-Bassioni, NRR WRogers, Ril l KRaglin '

,,MS JCalvo, NRR RBernhard, Ril  !

DHickman Widrouin, RES RJenkins, NRR MParker, Rlli

! JRosenthal HVanderMolen, RES CPoslusny, NRR SBurgess, Rlli l MKnapp, DEDE ESullivan, NRR RBarrett, NRR WJones, RIV i KWichman, NRR JMitchell, OEDO JShackelford, RIV l

i DOCUMENT NAME: H:\ DON \lE-LTR.WPD 1

l

1. -

l 4

To receive a copy of the document, Indicate in the box "C" wolattach/enci $"E" wIsttac no co py OFFICE RRAS/RRAB E RRAS/RRAB E RRAB E SPD NAME DMarksber[ SMays b PBaranoNk CERo g yf()L DATE 17 / 2 1 /98 /> /d >/98 A/ A'A/98 / //9 /97 j OFFICIAL RECORD COPY -

1 R

i i

. \

Attachment 1 RESOLUTION OF COMMENTS FOR DRAFT REPORT RATES OF INITIATING EVENTS AT U.S. NUCLEAR POWER PLANTS: 1987-1995 Peer review comments on the draft report titled, " Rates of initiating Events at U.S. Nuclear Power Plants: 1987-1995," April 1998, were received from 14 organizations internal and extemal to NRC. The resulting changes to the report based on comment resolutions are summarized below. Detailed responses to all peer review comments are available for review and can be obtained from Don Marksberry in AEOD. Peer review comments were received from the fo!!owing organizations:

• Notes and discussion from EDO (J. Mitchell and D. Marksberry) on May 21,1998 '

• E-mail from NRR/DSSA (G. Perry to D. Marksberry) dated May 29,1998 e Memorandum from Region IV (A. T. Howell 111 to C. E. Rossi) dated June 3,1998

• E-mail from Region 111 (S. D. Burgoss to D. Marksberry) dated June 4,1998 1

• E-Mail from AEOD/SPD/RAB (H. Ornstein to D. Marksberry) dated June 15,1998 e Memorandum from NRR/DE/EELB (J. Calvo to J. F. Stolz) dated June 18,1998 e Note from NRR/DSSA/SPSB (M. Rubin to D. Marksberry) dated June 24,1998

• Memorandum from RES/DET (L. C. Shao to C. E. Rossi) dated June 29,1998 e Facsimile from B. J. Garrick, Consultant, and D. J. Wakefield, PLG, Inc., dated l June 30,1998 e Memorandum from NRR/DE/EMCB (E. J. Sullivart to P. W. Baranowsky) dated June 30,1998 e Memorandum from RES/ DST (T. L. King to C. E. Rossi) dated July 2,1998

• Letter from Westinghouse Owners Group (L. A. W'alsh to NRC) dated July 21,1998 e Letter from Nuclear Energy Institute (A. R. Pietrangelo to C. E. Rossi) dated August 21,1998 l

• Letter from Duke Energy Corporation (M.S. Tuckman to NRC dated l September 14,1998 f

1. Most of the comments resulted in revisions in the form of clarifications, corrections, and expanded discussions. Thc .ollowing changes are incorporated in the final report:

• The summary of major findings in the Executive Summary was revised to state that all initiating events with sufficient data for trending analysis (10 or more events) displayed either a decreasing or a constant time trend. Two/ thirds of these categories and headings displayed a decreasing trend. (Resolution of l

comment from RES/ DST) i e The des to executive summary Tables ES-1 and ES-2 were revised to state that j the frequency estimates are per critical year. (Resolution of comment from

} Westinghouse Owners Group) 4 1 1

4 I

• The industry average criticality factor (75%) that was used to convert frequencies from units of reactor calender year to per critical year was used throughout the report. This factor is based on the 1987-1995 experience of U.S. plants.

(Resolution of comment from B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

2

• Section 2.2 that defines initial plant fault (IPF) and functional impact (FI) event categories groups was revised to clarify the differences between the two groups and how the groups are used in the study. In addition, the revised Section 2.2 provides more informative definitions of the IPF and F1 cvent groups.

(Resolution of comments from NRR/DSSA, Westinghouse Owners Group, and B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.) j l

• The term " update" was changed to " revised" in discussions that do not refer to Bayesian updating (for example--to bring the frequencies for the occurrenge of initiating events up-to-date with more current experience). Section 3.0 was revised to refer to those event categories where Bayesian updating was applied j to estimato frequencies. (Resolution of comment from B. J. Garrick, Consultant, l D. J. Wakefield, PLG, Inc.)  !

• Section 3.2.1, Tables 3-1 and D-12, and the tables in Appendix J were revised to clarify the operating experience used to calculate rare event frequencies.

(Resolution of comments from NRR/DE and Westinghouse Owners Group)

• Sections 3.2.1, Frequencies of Initiating Events, and 3.2.2, Investigation of l Possible Trends, were revised to state that the frequency of a category that showed a statistically significant trend was based on the end point of the trend line (i.e.,1995, the last year of the study). That year was selected as the one that reflects the most recent industry experience during the time period of this study (i.e.,1995). In addition, a statement was added in the Foreword of the report to recommend that a review of recent expe%nce in the licensee event reports will determine whether performance has undergone any significant change since the last year of the study. (Resolution of comments from RES/ DST, EDO, and Region Ill)

• Section 3.2.1, Frequencies of initiating Events, was revised to discuss the plant-specific nature of certain initiating event categories. (Resolution of comments from NRR/DSSA and B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

• A short discussion was added to Section 3.9.1, Frequencies of Initiating Events, and the introduction to the loss-of-coolant (LOCA) appendix (J) that describes the unit of measure used to calculate LOCA pipe break frequencies. In addition, the title of tables in Appendix J were revised to include LOCA frequency units (i.e., per calender year). (NRR/DSSA, Westinghouse Owners Group) 2  ;

e in the case where an event category had no or very few event occurrences, the single rate model was used to calcu! ate the mean frequency. As explained in Appendix E, the mean of the distribution for this model is (n +1/2)/t, where n is )

the number of events and tis the total time period of the operating experiences l in critical years. This discussion was added to Section 3.2.1, Frequencies of  :

Initiating Events. (Westinghouse Owners Group) 1 e Tables 3-4 and 3-5 in Section 3.3.2, Comparison to NUREG-1150 and NUREG/CR 3862, were revised to include fire frequencies from NUREG-1150.

In addition, the discussion on the comparisons of fire frequencies estimated from the 1987-1995 experience and NUREG-1150 was revised. No initiating event frequencies for loss of service water were referenced in NUREG-1150.

(Resolution of comments from Region ill and Westinghouse Owners Group) . i e Since the operating experience that was used to estimate initiating event frequencies based on the 1987-1995 experience and NUREG/CR-3862,  ;

Development of Transient initiating Event Frequencies for Use in Probabilistic 1 Risk Assessments, included new and decommissioned plants, a comparison of the total number of plants in the two operating experience populations would not be meaningful. However, a comparison of the average critical year per plant- 4 type (i.e., PWR, BWR) was added to Section 3.3.2, Comparison to NUREG- '

1150 and NUREG/CR-3862. (Resolution of comment from Westinghouse  :

Ownero Group) l e Section 3.3.3, Loss of Offsite Power--Comparison'to NUREG-1032, was revised to reterence NUREG/CR-5496, Evaluation of Loss of Offsite Power Events at Nuclear Power Plants: 1980-1996, that was recently issued. (Resolution of comment from NRR/DSSA) l

• The data used in NUREG-1150 and the ATWS Rulemaking was based on the

, operating experience at the time of the studies. The ciscussions in the Executive l Summary and Section 3.4, Comparison to the ATWS Rule, were revised to reflect this fact. (Resolution of comments from NRR/DSSA/SPSB, EDO, and j AEOD/RAB)

• The ATWS probabilities that were used in the ATWS Rulemaking comparison (Section 3.4) were based on not only failure probability of the reactor protection system (1.2E-5 per demand) , but also included the failure probabilities of ATWS mitigation systems (i.e., auxiliary feedwater system, high pressure injection systems). The frequency estimates in Table 3-6 are based on the limiting set of transients for ATWS as defined in EPRI NP 2230 (Table 4-2), A7WS: A

\

Reappraisal, PartIII, Frequency of Anticipated Transients. Section 3.4 was revised to add this clarification. in addition, a foot note was added to Table 3-6 to clarify the meaning of the frequency. (Hesolution of comments from EDO, Westinghouse Owners Group, and B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

a

3 s

4

m . _ _ _. _ __ _ _._ ____ _ _ _ . _ _ _ . _ _

i

• The distinction of a steam generator tube rupture as a " rare" event was removed in summary of major finding in the introduction discussion in Section 4, Engineering Analysis of Results. (Resolution of comment from Region lil) e Section 4.2, Industry-Wide Trends, was revised to correct the typographical error in the time trend formula and to revise the discussion about the formula to add clarification. (Resolution of comments from Westinghouse Owners Group and B.

J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

• An introduction to Section 4.3, Plant-Type and Plant-Specific Evaluations, was added to explain how between-plant variation of an, initiating event category was detected. Figures showing " water fall" plots and tables presenting plant-specific frequencies provide the quantitative results. (Resolution of comment from B. J.

Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

• Section 4.3.2, Between-Plant Variations, was revised to explain the reason why two of the four plants in Table 4-2 with an "XX"in the Loss of Condenser Vacuum category (L2) column do not have a "XX"in the Loss of Condenser Heat Sink heading (L) column. (Resolution of comments from Region lli and EDO)

• The LOCA discussion in Section 4.4.1, LOCA, was limited to pipe breaks. A section was added to discuss the four very small LOCA events found in the

1987-1995 operating experience. (Resolution of comment from B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

i e Section 4.4.2 (now 4.4.3), Steam Generator Tube Rupture, was revised to l' include the results of a sensitivity calculation that used total U.S. operating experience (1969-1997) to determine if a trend is evident at the end of 1997.

(Pesolution of comment from NRR/DE) j

• Section 4.4.4 (now 4.4.7), ATWS, was revised to include a brief discussion on the three ATWS precursor events where control rods failed to fully insert at Browns Ferry 3 (1980), South Texas 1 (1995), and Wolf Creek (1996).

(Resolution of comment from EDO)

• The Manual Reactor Trip subsection, under Section 4.5.2, Dominant Contributors to Risk-Significant Events, was revised to include high level insights on the root causes of manual reactor trips that were classified under the initial plant fault category OR6, Manual Reactor Trip. (Resolution of comment from ,

EDO) i L

• The definitions for small, medium and large pipe break LOCAs were revised in l Appendices A and J to reflect the break sizes used in NUREG-1150 analyses.

(Resolution of comments from NRR/DE, RES/DET, and Duke Energy Corp.)

4 i

4

l l

l

• The definition of Very Small Leak (G1) and Primary Leak (OG9) in Appendix A 1 was revised to clarify the distinction between the two categories. Primary Leak l category includes leaks less than 10 gpm. (Resolution of comment from I Westinghouse Owners Group) l e The LOCA category names and various discussions in the report were revised to  !

clarify the distinction between various small LOCA categories. In particular, l

category G3 was renamed Small Pipe Break LOCA. (Resolution of comment from NRR/DSSA) e Table B-1 in Appendix B, Category Cross-lkeference Tables to Previous Studies, was modified to exclude comparisons to LOCA categories, since NUREG/CR- I 3862 did not consider LOCAs (only leaks). (Resolution of comment from Westinghouse Owners Group) 1 e Appendix J on LOCA estimates was revised to include a table of plants that were l used to estimate world-wide experience for PWR LOCA frequency estimates, (Resolution of comment from NRR/DE) e Appendix J was revised to clarify the bases for the conservative conditional ,

break probabilities and the intergranular stress corrosion cracking (IGSCC) I improvement factor used to estimate the pipe break LOCA frequencies. In l addition, the appendix will provide comparisons of these adjustment factors with those estimated from the results of various analyses performed by Battelle, Lawrence Livermore National Laboratory, Pacific Nonhwest National Laboratory, and Westinghouse using probabilistic fracture mechanics computer codes, such as PRAISE, pcPRAISE, PROLBB, and SRRA. (Resolution of comments from RES/DET, NRR/DE, and Westinghouse Owners Group) e Appendix J was revised to include a discussion on other potential LOCA mechanisms to primary pressure boundary piping and components that could dominate pipe break LOCA frequencies. (Resolution of comment from l I

RES/ DST) e An appendix was added to discuss the classification of stuck open safety / relief l valve (SRV) events into initial plant fault and functional impact categories. In addition, the definition of stuck open SRV (category G2) in Appendix A was revised to provide guidance for the classification of a stuck open SRV event in the initial plant fault category (G2). (Resolution of comments from NRR/DSSA and NRR/DSSA/SPSB)

2. The resolution of several comments required re-analysis of initiating event frequencies for some categories, and the addition of analyses and evaluations. One significant comment resulted in the evaluation of the operating experience of new plants that went commercial during the 1987-1995 time period and its effect on the industry average and plant-specific frequency estimates. The following changes are incorporated in the final report:

l l

b

t*. -

l l C

• A sensitivity study was performed to evaluate the effects of a learning period for new plants that went commercial during the 1987-1995 time period. Analyses on trending and between-plant variations were performed using data without the first four months of commercial operation. Section 3.5, Effect of Learning Period at New Plants, and Appendix G, Results Based on Data after Learning Period, including Plant-Specific Results and Time Trend, presents the results of trending analyses with and without the learning period. The plant-specific results for  !

those initiating events categories with between-plant variations are based on  !

I experience after the fourth month of commercial operations. A discussion of the methodology used to detect the learning period and results of trending analysis were added to the report. (Resolution of comments from RES/ DST and Westinghouse Owners Group)

• Tables 3-2 and 3-3 in Section 3.3.1, Comparison to IPE/PRAs, were revised to  !

l l include only the total loss of service water frequency (not the combined total and l partial loss of service water frequency) in the comparisons to average IPE values i for BWRs and PWRs. !!is recognized that the definition of initiating events  !

associated with the loss of service water vary between IPEs. However, the comparisons show that the frequency based on total U.S. operating experience (no events) fall within the range of IPE values. The comparisons with IPE averages provide a reasonable sanity check. (Resolution of comment from B. J.

Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

• Three additionalloss of offsite power (LOSP) events were added to the

, functionalimpact LOSP category (B1). A subsequent re-analysis shows no l statistically significant trend of LOSP events. This result is consistent with that presented in the recent AEOD study on LOSP events (NUREG/CR-5496). The NUREG/CR-5496 report also provides an analysis and evaluation of recovery times during LOSP events that occur during power and shutdown operations.

(Resolution of comments from Westinghouse Owners Group and B. J. Garrick, i Consultant, D. J. Wakefield, PLG, Inc.)

l l

• A new LOCA category, Reactor Coolant Pump Seal LOCA in PWRs was added l to the report. Section 4.4.5 was revised to reference an AEOD report, NUREG/CR-6582, Assessment of Pressurized Water Reactor Primary System Leaks, and to discuss reactor coolant pump seal leaks caused by failures in the seal cooling water system. The data included in the leak report can be used to develop reactor coolant pump seal failure probability. (Resolution of comments j from RES/ DST, Westinghouse Owners Group, and B. J. Garrick, Consultant, D.

! J. Wakefield, PLG, Inc.)

• Section 4.5.3 was revised to provide additional general insights on conditional occurrences of risk-significant evsnts after the reactor trip initiator. (Table D-13 in Appendix D provides a matrix that maps the subsequent functional impact events to the initial plant fault categories.) (Resolution of comment from NRR/DSSA) 6

i l

3. The following events were reclassified and incorporated in the re-analysis of initiating i event frequencies for functional impact and initial plant fault categories:

• The North Anna Unit 1 steam generator tube rupture event (LER 338/89-005) was classified as a Very Small LOCA/ Leak event (category G1) due to a leak rate less than 100 gpm. (Resolution of comment from NRR/DE)

• The Oconee (LER 270/97/01) through-wall crack event was not used in the medium pipe break LOCA frequency estirpate because the effective break size was limited by a thermal sleeve with a 1.5 inch diameter. (Resolution of comment from Duke Energy Corp.)

• LER 287/91-008 was reclassified to a very small LOCMeak. LER 269/94-002 was removed from the fire category. LER 269/93-008 was reclassified as a Turbine Trip (OR5). LERs 287/92-003 and 287/94-002 were reclassified as a loss of the l&C bus (OC4). (Resolution of comment from Duke Energy Corp.) {

• The initial plant fault classification for LER 348/91-008 was reclassified as a Spurious Reactor Trip (OR8). The initial plant fault classification for LER 424/88-025 was reclassified under the Reactivity Control category (OR3).

(Westinghouse Owners Group) l

4. The implementation of several comments would require significant re-analysis, evaluations and data rev!ews, which would require effort (and funding) beyond the original scope of this stuj y. The recommendations summarized below will be considered for implementation in the update to this report.

• The identification of demands on safety / relief valves (SRVs) to estimate l conditional probabilities of SRVs failing open. (Resolution of comment from l NRR/DSSA/SPSB)

• An evaluation of the causes in reactor trip reduction since NUREG/CR-3862 and NUREG-1032 analyses. (Note: Two AEOD studies on the industry l effectiveness in reducing scrams, and scrams occurring during maintenance and surveillance activities are nearing completion.) (Resolution of comments from RES/ DST, NRR/DSSA, and B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc)

• An evaluation of initiating events frequencies during shutdown operations. l (Note: Other sources of insights into shutdown risks are the Accident Sequence l Precursor Program (NUREG/CR-4674, series) and NUREG-1449, Shutdown and Low-Power Operation at CommercialNuclear Power Plants in the United States, 1993. (Resolution of comment from Region IV) l

• The classification and analysis of reactor trip events between 1983 and 1987.

(Resolution of comment from Westinghouse Owners Group)

\

7

o.

• The collection of initiating event experience involving failures of ventilation systems. (Resolution of comment from B. J. Garrick, Consultant, D. J.

Wakefield, PLG, Inc.)

a e Comparisons of results based on 1987-1995 experience to other initiating event frequency databases, such as the PLG database. (Resolution of comment from B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

l

• The collection of main steam isolation valves failures to close. (Note: The Sequence Coding and Search System can be used to identify these events.)

(Resolution of comment from NRR/DSSA/SPSB) i l

e The analysis and evaluation of trends in time based on low-power license date.

(Note: A sensitivity analysis was performed to evaluate the effects of the early plant experience for plants that went commercial during the 1987-1995 time period. See Section 3.5, Effect of Learning Period at New Plants.) (Resolution of comment from B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.)

• The analysis and evaluation of reactor trips at low power operations. (Resolution of comment from B. J. Garrick, Consultant, D. J. Wakefield, PLG, Inc.) ,

l l

l 8