ML20202H679

Provides Commission W/Staff Preliminary Insights Gained Primarily from Review of First 24 Licensee Individual Plant Exam of External Events (IPEEE) Submittals.Interim Rept Encl

ML20202H679
Person / Time
Issue date:	01/20/1998
From:	Callan L NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO)
To:	Diaz N, Dicus G, Shirley Ann Jackson, Mcgaffigan E, The Chairman NRC COMMISSION (OCM)
Shared Package
ML20202H665	List: ML20202H661 ML20202H679
References
NUDOCS 9802200441
Download: ML20202H679 (124)

Text

____-_- _ - _-_ - __ - -

L

, t{ j a UNITED STATES

,- p NUCLEAR REGULATORY COMMISSION WASHINGTON, D.c. 300eHOM

- ....+ Janaur,yy 20, 1998

- MEMORANDUM TO: Chairman Jackson

- Commisaloner Dieus .

Commissioner Diaz Commissioner McGnWenn

FROM:

'6it2M)

L. Joseph Caiig-wuw Director for Operations

SUBJECT:

, PRELIMINARY IPEEE INSIGHTS REPORT- i J

- The purpose of this memorandum is to provide the Commission with the stafs preliminary insights gained primarily from the review of the first 24 licensees' Individual Plant Examination -

of External Events (IPEEE) submittals. The staN has also completed preliminary reviews of an

. additional 17 submittels, and perspectives from these reviews are also factored into tho' attached report, The staff has regularly provided the Commission with the status and progress of the iPEEE program. The most recent updates were provided in the following two. Commission papers:

(1) SECY-97132 dated June 23,1997, which was the annual" Status of the Integration Plan for

- Closure of Severe Accident issues and the Status of Severe Accident Research," and, more recently, (2) SECY-97-234 dated October 14,1997, which was the " Quarterly Status for the Probabilistic Risk Assessment implementation Plan." In Attachment 7 to SECY-97 234, the

- staff provided the Commission with a summary of preliminary perspectives that have been obtained from the reviews of the first 24 IPEEE submittals. The staff also committed, in SECY 97-234, to provide the Commission an interim report on preliminary perspectives from :

reviews of the first 24 IPEEE licensee submittals which orovides additional background and -

detalled discussion on the IPEEE program as well as plant-specific information from the reviews of the IPEEE submittals that were not included in Attachment 7 to SECY-97-234.

The staff has now developed the interim report, a copy of which is attached to this paper. The primary objective of this draft report is to provide preliminary perspectives gained from the first 24 IPEEE submittal reviews. These preliminary perspectives include: (a) an assessment of the licensees' overall effectiveness in meeting the IPEEE objectives, (b) summaries of findings and plant improvements reported in the IPEEEs, and (c) additional perspectives related to CONTACT: Alan Rubin, RES (301) 415-6776 I

9002200441 990203 EB PDR ORO poR

-__j

i The Commissioners 2 individual extemal events and the strengths and weaknesses of the licensee ( oubmittals with regard to their success in achieving the IPEEE objectives. The NRC staff has thus far considered only a subset of all the IPEEE submittals that will be reviewed, and the staff has not yet finalized the technical reviews of every submittal included in this report (i.e., some require more detailed reviews in certain areas).- in addition, the perspectives documented in this report have been gleaned, for the most part, from ' submittal-only" reviews of the 24 lPEEEs; in other words, the perspectives have not generally benefitted from review of the detailed supporting documentation that licensees have been requested to maintain.-

Based on the reviews completed to date, the following overall preliminary conclusions can be drawn regarding the IPEEE program. -These conclusions are preliminary for the reasons discussed above.

One of the major preliminary findings of the IPEEE progrum is that seismic and fire events have been found to be important contributors to core damage frequency (CDF) for a majority of plants. In fact, the CDF contribution from seismic or fire events can, in some cases, approach

'(or even exceed) that from intomal events (e.g., seismic CDF at Haddam Neck, and fire CDF at

- Quad Cities). Core damage frequency estimates varied over several orders of magnitude. For-example, fire CDFs were reported to range from less than 1x104 to 5.3x104 per reactor year -

- (RY), while seismic CDFs were reported to range from 2.2x104 to 2.2x10d/RY.

The dominant contributors to seismic CDF most commonly reported by licensees include -

seismic-induced loss of offsite power, failures of electrical and control panels, failures of block walls, and spatial interactions. The ranking of dominant contributors has consistently been reported as being insensitive to the use of different seismic hazard curves.

The dominant fire risk areas most commonly reported by licensees include the main control room, cable spreading room, and switchgear rooms. Other frequently reported areas include all or selected parts of the turbine hall, battery and DC equipment rooms, and diesel generator rooms.

. A seismic vulnerability at Haddam Neck and a fire vulnerability at Quad C'+ies were reported by I

the licensees.- The licensee of Quad Cities has implemented an interim altemate shutdown method involving the use of an independent back-up power supply for both units to reduce the -

L fire CDF from 5.3x104 to 7x1P/RY, and currently is in the process of evaluating long term options for further reducing the fire risk potential. The licensee of Haddam Neck made some improvements to its plant to reduce the seismic vulnerability, but decided to permanently shut down the plant.-

- Many plants have reported some seismic and fire-related plant improvements as a result of the IPEEE effort. A few also reported improvements in the high wind and flood areas. These '

improvements have taken the form of changes to existing procedures, development of new procedures, or plant modifications.

Specific strengths and weaknesses of each IPEEE submittal have been identifMxi as a result of tha NRC'c technical review process. Som submittals have been found to contain weaknesses

p. -

! l0

.o The Commissioners '3 or deficiencies in one or more areas of their analyses (i.e., seismic, fire, high winds, or flood

.F events). The deficiencies have arisen sometimes from inadequate documentation of key -

jaspects of the analyses and/or apparent mistakes or oversights with the potential to -

- fundamentally impact the IPEEE results. Ir.teractions with the licensees, mainly through

, requests for additional information (RAI), have been conducted to obtain clarification of specific d

points in the submittal which were either_ unclear or of questionable basis, These RAls have

generally been limited to items considered to be of sufficient importance that the insights or findings of the IPEEE, or the reviewers' understanding of those findings and insights, might be

, - significantly impacted by the licensee's response. Based on the results of these reviews, a preliminary conclusion is that most submittals have met the intent of the IPEEE program.

However, some submittels need additional review because of either unusual results (i.e.,_ .

extremely high or low CDFs) that were reported and/or the response was not adequate for the .

' staff to conclude that the intent of the IPEEE program was met.-

o Consistent with the PRA Implementation Plan as well as the RES Operating Plan, the staff will 4 issue in mid 1999 (after all the submittals have been received and reviewed) a draft final report

for public comment on perspectives resulting from the IPEEE program. In addition, the staff will assess these perspectives and tAe any appropriate followup a::tions as necessary with the

, licensees. After receipt of public comments, the final IPEEE insights report will be issued in December 1999.

[ . Unless informed otherwise, we will send tne attached report to the Public Document Room ten ~

days after the date of this memorandum, i

Attachment:

interim report on preliminary IPEEE perspectives E cc:

l SECY-OGC OCA OPA-

C101 CFO

[

t J

y ,-w 1 w -,-r---- w , ... , -- , , - ~ - - . . - - - - - . - - - _ _ - _ _ - - . -

s ,

NUREG-XXXX PRELIMINARY PERSPE TIVES GAINED FROM INITIAL INDIVIDUAL PLANT EXAMINATION OF EXTERNAL EVENTS (IPEEE) SUBMITTAL REVIEWS December 17,1997 1

l

,e w t

j- ,

l ' ABSTRACT I

On June 28, igg 1, the US Nuclear Regulatory Commission (NRC) lasued Supplement 4 to Generic Letter (GL) 88-20, " Individual Plant Examination of Extemal Events (IPEEE) for Severe Accident

!- - Vulnerabilities,10 CFR 50,54(f)," and NUREG-1407, " Procedure and Submittal Guidance for the

In'; eldual Plant Examination of Extemal Events (IPEEE) for Severe Accident Vulnerabilities."

i Specifically, the NRC requested that each licensee perform an IPEEE to identify and report to the l

NRC all plant specific vulnerabilities to severe z,ccidents caused by extemal events. The extemal events to be considered in the IPEEE included seismic events; intomal fires; and high winds, floods, and other extemal initiating events (HFOs) involving accidents related to transportation or nearby facilities, l The objective of the IPEEE submittal review is to ascertain whether the licensee's IPEEE process is capable of identifying severe accident vulnerabilities to such stomal events, and implementing I

cost-effective safety improvements to either eliminate or educe the impact of these vulnerabilities.

However, the review does not attempt to validate or verify the results of a licensee's IPEEE.

The primary purpose of this report is to present the preliminary perspectives gained from the reviews of the first 24 iPEEE submittels. In the seismic area, licensees have gained insights from performing seismic walkdowns of their plants. Of the first 24 submittals, only 1 (Haddam Nock) reported a seismic vulnerability; however, a majority of the submittals reported the licensees' implementation of some plant modifications or changes to plant procedures as a result of the i seismic IPEEE analyses.

l

[ ln this intomal fires area, most licensees spent substantial efforts in their fire assessments, including

' extensive plant walkdowns. None of the first 24 submittals identified a fire vulnerability; however, some licensees have implemented plant modifications or procedural changes as a result of the fire i~ - IPEEE analyses. (The reader should note that a fire vulnerability was identifed at the Quad Cities

plant, which was not among the first 24 IPEEE submittals reviewed, in that instance, the licensee j has taken an interim measure, involving an altemate shutdown procedure of using an independent .

l -. back-up power supply to reduce the potential fire risk, and is currently evaluating long-term options

! to further reduce the risk. The staff is currently reviewing the Quad Cities IPEEE submittal.)

!, In the HFO areas, the level of analysis has varied widely from plant to plant. Many licensees

- assessed their plants using simplistic screening methods, while others conducted detailed analyses.

E  :

None of the first 24 submittals identified a vulnerability; however, some licensees have implemented i'

plant modifications or procedural changes as a result of the HFO IPEEE analyses.

p

This report discusses the preliminary perspectives gained from the IPEEE reviews including both L - IPEEE methodology and plant facilities and procedures, as well as the types of plant modifications that licensees have reported,

[

ii

l i

2 1

i i

4, -.

+ d l i

TABLE OF CONTENTS p A BST RACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 11 l . EXECUTIVE

SUMMARY

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii ACKNOWLE DGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . xv i ABBR EVI ATION S . . . . . . . . . . . . . . . . . . . -. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi -

G LOSSARY . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii i

1 _ l NTRODUCTI ON . . , , . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1,1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 1

. 1.2 - Overview of the Technical Review Process for IPEEEs . . . . . . . . . . . . . . . . . . . . . 2

, ' 1.3 Objectives of the IPEEE Perspectives Program . . . . . . . . , . . . . . . . . . . . . . . . 2 - '

, _ 1.4 Scope, Limitations, and General Comments Pertaining to Findings of the IPEEE~

j Perspectives Program - . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.5 Report objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . , 4 '

1.6 Report Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , , , . 4 1

2 - PRELIMINARY CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 i _ 2.1 Overall Effectivaness in Meeting The Intent of GL 88-20 And Achieving The IPEEE Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2 - Plant improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . g 2.3 ' Summary of Major IPEEE Results . . . . . . . , , , . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.4 . Additional Perspectives and Observation ._. . . . .-. . . . . . . , . . . . . . . . . . . . . . . . 11

. 2.4.1 Seismic lPEEEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.4.2 Fire l PE E Es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 -

2.4.3 H F O I PE E Es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.5 Uses of this Report and the IPEEE Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3- SEISMIC IPEEE PERSPECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.1 Overview ' . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.2 - Impact of the Seismic IPEEE Program on Plant Safety . . . . . . . . . . . . . . . . . . . . 20

3.3 Seismic PRA Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.3.1 ' Summary of Quantitative Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

. 3.3.2_. Summary of Qualitative Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.3.3 - Containment Performance Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.3.4 Implications of Different PRA Methodologies . . . . . ................. 26 3.4 Seismic Margin Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.4.1 Summary of Quantitative Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.4.2 - Summary of Qualitative Findings . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 28 3.4.3 Containment Performance Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . 2g 3.4.4 - Implications of Different Margin Methodologies . . . . . . . . . . . . . . . . . . . . . 29 3.5L Other Evaluation Perspectives - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 1 3.5.1 Relay Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 T:

' 3.5.2 Soils Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.5.3 Non-Seismic Failures and Human Actions . . . . . . . . . . . . . . . . . . . . . . . , . 31 3.5.4 Soismio Fire and Seismic-Flood Evaluation . . . . . . . . . . . . . . . . . . . . . . . 31 j 3.5.5 Generic and Unresolved Safety issues . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Ih

'l vm ,, e a - , - - . -- - - - , - - -

a .

3.6 Perepectives Regarding Seismic PRA Versus SMA . . . . , . . . . . . . . . . , , . . . . . 33.

3.7 Consistency of Perspectives Among Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 3.8 Findings that Require Further Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 4 FIRE IPE E E PERSPECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 i 4.1 Ov erview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 5 4.2 Impact of the Fire IPEEE Program on Plant Safety . . . . . . . . . . . . . . . . . . . . . . . 66 4.3 Perspectives Pertaining to Overall Methodology . . . . . . . , , . . . . . . . . . . . . . . . 66 4.4 Walkdown Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 4.5 Dominant Risk Oontributors . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . 67 4.6 Vulnerpilities and Plant improvements . , ..........,,...............68 4.7 Cable Routing Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 4.8 Threshold Value for Screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 4.9 Fire Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 4.9.1 Fire 4.9.2 Fire Detection Damage Modeling and. Suppression . . . . . . . . . . . . . .. .. .. .. ... .. .. .. ... .. .. .. .. ... .. .. .. .. ... '.72. . . . . . . . . . . 71 4.9.3 Electrical Cabinet Fire Propagation , . . . . . . . . . . . . . , . . . . . . . . . . . . . . 72 4.9.4 Inter Compartmental Fire Propagation . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.10 Containment Performance Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.11 Human Action Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 4.12 Generic issues and Unresolved Safety issues . . . . . . . . . . . . . . . . . . . . . . . . . . 75 4.12.1 G I-5 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 5 4.12.2 U SI A-4 5 . . . . . . . . . . . . . . . . . . , ..........................75 4.12.3- Fire Risk Scoping Study issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.13 Consistency of Perspectives Among Similar Piants . . . . . . . . . . . . . . . . . . . . . , , 77 5 HFO lPEEE PERSPECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 5.1 Ove rview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3 5.2 Impact of the HFO IPEEE Program on Plant Safety . . . . . . . . . . . . . . . . . . . . . . 83 5.3 H ig h Wind s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 5.3.1 Quantitative Perspectives . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . 84 5.3.2 Qualitative Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . 84

. 5.4 Extemal Floods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 5.4.1 Quantita'Jve Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . 84 5.4.2 Qualitative Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5.5 Accidents involving Transportation or Nearby Facilities ............. , ... 85 5.6 Other HFO Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5.6.1 Quantitative Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5.6.2 Qualitative Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5.7 Walkdown Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 5.8 Outliers and Plant improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 5.9 Containment Performance Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 5.10 Human Action Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 5.11 Generic issues and Unresolved Safety issues . . . . . . . . . . . . . . . . . . . . . . . . . . 86 5.1.! Generic Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 6 RE F E R E NC E S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 iv l

. _ . __ _ _._ _-_.._ _ _ _._~- _ _ _ _ . _ .. . . _ . _ _ _

LIST OF TABLES Table 11 Basic Characteristics of Plants included in this Study . . . . . . . . . . . . . . . . . . . . . . 6 Table 3.1 Seismic Review Categories and Evaluation Approaches for Plants included in this Study ..................-.,..................................... 36 Table 3.2 - Seismic Core Damage Frequency and Plant Capacity Results from Seismic PRAs 37 i Table 3.3 Dominant Risk Contributors Reported in Seismic PRAs . . . . . . . . . . . . . . . . . . . 38 '

Table 3.4 Summary of Seismic Anomalies, Outliers, Vulnerabilities, Housekeeping Concems, and Plant improvements identified in SPRA IPEEEs . . . . . . . . . . . . . . . . . . . . . . 40 4'

Table 3.5 Summary of Seismic Containment Performance Findings from SPRA IPEEEs . . 43

. Table 3.6 _ - Plant Capacity Results from Seismic Margin Assessments . . . . . . . . . . . . . . . . . 45 Table 3.7- Summary of Selsmic Anomalies, Controlling Outliers,- and Plant improvements identified in SMA IPEEEs , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Table 3,8 Summary of Seismic Containment Perfomence Findings from SMA IPEEEs . . , ,49 Table 3.9 Summary of Findings from Relay Evaluatior a seismic IPEEEs . . . . . . . . . . . . , . 51 Table 3.10 Soll Characteristics and Summary of Finding , of Soil Fai!ure Evaluation in Seismic IPEEEs

..............................................................54 Table 3.11 Summary of the Treatment of Non Seismic Failures and Human Actions in Seismic IPEEEs

..............................................................57 Table 3.12 Summary of Findings of Seismic Fire Interaction and Seismic-Flood Evaluations in IPEEEs

..............................................................60 Table 3.13 Summary of Characteristics of Flux' Mapping Systems and of Seismic IPEEE

' Findings Related to Gl.131 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , 63 Table 4,1. . Plant-Specific Core Damage Frequencies Attributable to Fire Events . . . . . . . . . 79 Table 5.1 Methodologies and Results Associated with HFO IPEEEs . . . . . . . . . . . . . . . . . 88 V

~ ,

EXECUTIVE

SUMMARY

Introduction On November 23,1988, the U.S. Nuclear Regulatory Commission (NRC) issued Generic Letter (GL) 88-20,

• Individual Plant Examination for Severe Accident Vulnerabilities,10 CFR 50.54(f)," to licensees of nuclear power plants. Specifically, GL 88-20 requested that the licensees ' perform a systematic evaluation of existing plants to identify any plant specific vulnerabilities to setere accidents and report the results to the Commission.' Gs. 88-20 also outlined the objectives and overall logisiics of the Individual Plant Examination (IPE) program, which solely addresses intemally initiated events (including intemal flooding).

On June 28,1991, the NRC issued Supplement 4 to GL 88-20, ' Individual Plant Examination of Extemal Events (IPEEE) for Severe Accident Vulnerabilities,10 CFR 50.54(f)." That supplement described the objectives and overalllogistics of the Individual Plant Examination of Extemal Events (IPEEE) program, which addresses extemally initiated events. In particular, the extemal events considered in the IPEEE program include seismic events; intemal fires; and high winds, floods, and other extemal initiating events (HFOs) involving accidents related to transportation and nearby facilities. The Commission formulated both the IPE and IPEEE programs in response to the NRC's 3

" Policy Statement on Severe Accidents Regarding Future Designs and Existing Plants", issued on August 8,1985 (Federal Register,50FR32138). In particular, these programs were intended as a means for licensees to identify potential vulnerabilities to severe accidents, and to conceive cost-effective improvements to ensure that plants do not pose any undue risk to public health and safety.

Along with Supplement 4 to GL 88-20, the NRC issued NUREG-1407, " Procedure and Subm! ital Guidance for the Individual Plant Damination of Extemal Events (IPEEE) for Severe Accident Vulnerabilities," dated June 28,1991. Additionally, on September 8,1995, the NRC issued Supplement 5 to GL 88-20, which notified licensees of modifications to the recommended scope of the seismic analysis portion of the IPEEE for certain plants.

The purpose of this report is to document the preliminary perspectives gained frnm technical reviews of the first 24 IPEEE submittals'. These preliminary perspectives primari'y include: (a) an assessment of the overall effectiveness in meeting the IPEEE objectives, (b) summaries of findings ,

and plant improvements reported in the IPEEEs, and (c) additional perspectives related to individual extemal events and the strengths and weaknesses of the licensees'submittals with regard to their success in achieving the IPEEE objectives. The NRC staff has thus far considered only a subset of IPEEE submittals, and the staff has not yet finalized the technical reviews of every submittal included in this report, in addition, the perspectives documented in this report have been gleaned, for the most part, from " submittal-on!y" reviews of the 24 IPEEEs; in other words, the perspectives

' This report was derived from a May 1997 report prepared by Energy Research,Inc (ERI) that presented results from the technical reviews of the first 24 IPEEE submittals. The majority of those reviews were pet'ormed by ERI, under contract to the NRC's Office of Nuclear Regulatory Research (RES). Since then an a iditional 17 IPEEE submittals have been reviewed by Sandia National Laboratories (SNL), Brookhaven National Laboratory (BNL) and RES staff in the fire, seismic, and HFO areas, respectively; and preliminary Technical Evaluation Reports (TERs) and have been completed for these submittals. One significant additional finding from these 17 reviews was the identification of a fire vulnerability in the turbine building st Quad Cities (discussed in this report). There are also somo differences in the strengths arvi weaknesses from submittal to submittal. Despite these differences, with the exception of the fire vulnerability at Quad Cities, the results from the additional 17 reviews generally support the overall preliminary perspectives obtained from the first 24 reviews discussed in the Executive Summary and Chapter 2 of this report.

. e i

. have not generally benefitted from review of the detailed supporting documentation that NUREG-1407 requests licensees to maintain.

The perspectives documented in this report are somewhst general for the following reasons: (a)

IPEEEs are intended to yield predominantly qualitative perspectives, rather than more prescriptive quantitative findings; (b) IPEEEs address several different types of initiators of varying importance (for a given plant) and, therefore, require the implementation of different methods of analyses offering varying levels of detail and accuracy; and (c) even for a given type of extemal initiator, the procedures and methods used by the various licensees to conduct their IPEEEs have also varied considerably.-

i in addition to an overall summary of IPEEE perspectives, this repirt discusses perspectives specific to the seismic, fire, and HFO areas of the IPEEEs, including dctal ad descriptions of the findings, as l well as the strengins and weaknesses of the submittals.

Overall Effectivenefs in Meetina the Intent of GL 88-20 and Achievina the tPEEF Obha*ives 4 Coi.;' stent with the intent of GL 88-20, the primary goal of the IPEEE program has been for licensees to " identify plant specihc vulnerebilities to severe accidents that could be Mxed with low-cost improvements." More specifically, Supplement 4 to GL 88-20 identified the following four supporting IPEEE objectives for each licensee to achieve:

1. Develop an appreciation of severe accident behavior.

2. Understand the most likely severe accident sequences that could occur at the licensee's plant under full-pour operating conditions.

3.

Gain a qualitative understanding of the overalllikelihood of core damage and fission product releases.

4 Reduce, if necessary, the overalllikelihood of core ciamage and radioactive material releases by modifying, where appropriate, hardware and procedures that would help prevent or mitigate severe accidents.

Based on the reviews conducted to date, the IPEEE program generally appears to have been succeaafulin meeting the overallintent of GL 88-20; however, the degree of success achieved by licensees'IPEEEs has varied conalderably, depending strongly on the methods and -

assumptions employed by the IPEEE analysts.

The following paragraphs summarize the overall effectiveness of licensees' IPEEEs in achieving each of the four identified IPEEE objectives.

Objectin 1: Apprecinhon of Severe Accident Behavior The u .f's review of the first 24 iPEEE submitta;d suggests that the IPEEE program has increased licensees' overall appreciation of severe acc! dent behavior attributed to ll sxternal events at their plants. As requested in NUREG '407, each licensee has performed evaluations of seismic events, intemal fires, and HFO events. These evaluations have ri.sessed the potential for extemally initiated severe accidents, and plant-specific behavior in response to potential severe accidents.

vii

For the most part, licensees have been involved in both the management and execution of their IPEEEs. Licensees have sponsored training of their personnelin specialized facets of IPEEE analysis (e.g., seismic IPEEE training), to develop or enhance their appreciation of severe accident issues and of relevant plant t'ehavior, in accordance with the request of NUREG-1407, licensees have undertaken peer reviews of their IPEEEs.

Objective 2: Understanding of the Most Likely Severe Accident Sequences For the most part, licensees have gained a qualitative understanding of the most likelt severe accident sequences that may occur as a result of external events. This understanding has been evidenced by the fact that licensees have identified (at least qualitatively) the relative risk significance of the various extemal events. Moreover, for each extemal event that was not screened out, licensees have identified the important initiators, as well as critical plant components, operator actions, and plant areas, and have acquired an understanding of their effect on plant systems.

Consistent with the guidance of NUREG-1407, licensees' empasis in conducting IPEEEs has been on obtaining a qualitative (as opposed to quantitative) understanding. As expected, therefore, the IPEEEs do not generally convey a definitive ranking of the risk significance of severe accident sequences or their dominant risk contributors. Rather, by means of systern modeling and screening analysis, licensees have obtained a greater awareness of severe accident sequences and an improved sense of which are the most important sequences.

Objective 3: Qualitative Understanding of the Likelihood of Core Damage and Fission Product Releases By means of IPEEEs, licensees have generally been able to ascertain.whether the risk of core damage associated 4

with each external initiator is comparatively negligible (i.e.,

falling below the 10 per reactor-year screening threshold), low, moderate, or high. In some cases, this understanding arcae through direct quaritification of core damage frequency (CDF); in other cases, the understanding resulted from increased knowledge of the given hazard in conjunction (where necessary) with an assessment of the plant's ability to withstand that hazard.

Each IPEEE submittal reported the findings of a q=,litative evaluation of containment performance in response to seismic events. Some IPEEE submittals also reported a quantitative estimate of the frequency of seismically induced large releases. For intemal fire events, NUREG-1407 requests a containment analysis only if there are containment failure modes that differ significantly from those identified in the IPE; for HFO events, NUREGo407 does not specifically request an assessment of containment performance. Licensees have generally followed the NUREG-1407 guidance in these regards.

Although in many cases licensees have reported numerical risk eatimates (for CDF or frequency of significant radiological releases), it is important to note that the accuracy of such estimttes is often limited because of simplifying assumptions and approximate procedures employed in the analyses. Hence, the results serve only as general indicators of risk level, and they should not be viewed as being well established.

viii

Ob}ective 4: Modincations to Reduce the Likelihood of Core Demoge and Fission Product Releaset Overall, as a result of the IPEEE program, licensees have implemented or proposed I plant modifloations that have had a beneflolal effect on plant safety with respect to i external events. Such modifications have taken the form of hardware changes, procedural l changes, and implementation of severe accident management guidelines. Consistent with

)

the qualttative nature of the IPEEE program, it is not usually possible to numerically deduce

• the risk reductions achieved t>y these modifications. However, some licensees have employed probabilistic risk assesstro it (PRA) in their IPEEEs as a means of determining whether plant modifications are warranted relative to a cost benefit rationale.

- Therefore, in considering the observations and perspectives derived from the review of the first 24 iPE12 submiltals, the staff has drawn the following preliminary conclusions regarding the IPEEE program. Licensees have expended signifloant effort in developing their IPEEEs,' As a result,  ;

they have acquired relevant knowledge ooneerning their plants, and have taken meaningful steps to improve plant safety. On that basis, the staff concludes that the IPEEE program has aohleved a signifloant degree of suooess. Information from the IPEEEs may potential'y be useful in supporting a variety of risk informed regulatory act'vities; however, depending upon the specific application, mors specific and detailed reviews are needed for that puroose, flant improvements Of the 24 iPEEE submittals mviewed, a majority of licensees have implemented or proposed plant improvements to address some concerns identified through the IPEEE program. These improvements have enhanood the plants resistance to external events that might result in severe sooidents. A few licent e s have proposed no improvements to enhance plant capability with respect to important initiato.s. In most cases, however, this was because those licensees had already implemented relevant phnt improvements befcre the IPEEE program began; in developing GL 88 20, the NRC expected that licensees could accomplish signifloant plant improvements at low cost. In keeping with that expectation, the majority of the enhancements proposed or implemented by lloonsees have been of relatively low cost, compared to major hardware or design modifloations However, there have been a few instances in which the potentist for significa.it enhancement was identif,sd in an IPEEE or a prior evaluation (e.g., an IPE or investigation of an existing safety issue), in those instances, the licensees have implemented or proposed certain improvements relevant to the IPEEE program that appear to have involved significant cost..

It is important to note that most licensees have not applied the term " vulnerabilities"in describing the plant conditions for which improvements were proposed or implemented. This occurred when the licensee did not beliete a designation of " vulnerability" was justified. (This reason applied, for.

example, when the licensee did not judge the condition to be sufficiently severe, or the condition did not meet specific criteria explicitly used by the licensee). Thus, it is often difficult to determine the degree to which the plant improvements have succeeded in reducing the identified risk.

Plant improvements related to seismic events have generally $aken the form of various hardware fixes, maintenance actions, and enhancements to maintenance procedures.

Hardware fixes have included such activities as anchoring equipment, bolting cabinets together,

" improving existing anchorage or supports, installing missing fasteners and bolts, installing spacers on battery racks, eliminating potential interaction concems, and replacing vulnerable relays.

Maintenance actions have locluded removing corrosion on equipment anchorages, and applying in

I corrosion protection. Enhancements to maintenance procedures (primarily seismic housekeeping) have also included provisions for preper storage of ladders, tools, gas cylindem, etc., and for proper

- parking of cranes and chain hoists. Similar types of improvements have been implemented with respect to seismic fire interaction concoms.

Lloonoces have proposed or implemented plant improvements related to fire and HFO events.

For instance, licensees have planned certain improvements to fire protection systems, including hardware modifications and enhancements to, or v velopment of, fire response procedures.

Additionally, improvements have often taken the A 6 of severs accident management guidelines that address specific accident sc6narios related to imemal fires, potential effects of wind-induced missiles, and extemal flooding. Imolertentation of Some of the severe accident management guidelines has led to the acquitition of ternport.ry o* podable equipment (pumps, diesel 011 tanker trucks, etc.). One HFO IPEEE reponed the strengthening of the stacks of two adjacent fossil fuel unMs to reduce the high wind risk, and refurbishment of a flood well to reduce flood risk.

In some cases, the IPEEEs also referenced plant improvements that had been proposed or implemented before the IPEEE program began, since those improvements resulted in a beneficial effect on plant safety in the face of seismic, fire, and/or HFO events. For example, at one plant, the addition of diesel generators was idemified as a plant improvement in the IPE, and was correspondingly reported in the IPEEE since it reduced the risk of station blackout for seismic, fire, and HFO events.

Summarv of Major IPEEE Results One of the major pidiminary findings of the IPEEE program is that seismic and fire events have been found to be important contributors to CDF for a majority of plants, in fact, CDF contribution from seismlo or fire events can, in some cases, approach (or even exceed) that from internal events (i.e., seismic CDF at Haddam Neck and fire CDF at Quad Cities). Coro damage frequency estimates varied over several order of magnitude. For example, fire CDFs were reported to range

• rom less than 1x10* to 5.3x10~8 por reactor year (RY), while seismic CDFs were reported to range trom 2.2x104 to 2.2x10d/RY.

The dominent risk contributors to seismic CDFs most commonly reported by licensees include seismic induoed loss of offsite power, failures of electrical and control panels, failures of block walls, and spatialIntera?lons. The ranking of dominant contributors has consistently been reported as being insensitive to the use of different seismic hasard curves.

The dominant fire risk areas most commonly reported by lleensees include the main control room, table spreading room, and switchgear rooms, Other frequently reported areas include all or selected pans of the turbine hall, battery and DC equipment rcoms, and diesel generator rooms.

A seismio vulnerability at Haddam Neck and a fire vulnerability at Quad Cities were reported by the licensees. The licensees for these two plants have used the criterion recommended by the former Nuclear Menegement and Resources Council (presently the Nue, lear Energy Institute) to define a

• vulnerability"(e.g., CDF exceeds 1x10d/ RY). The licensee of Quad Cities has implemented an interim altomate shutdown method involving the use of an independent backap d

power supply for both units to reduce the fire CDF from 5.3x108 to 7x10 /RY, and currently is in the process of evaluating long term options for further reducing the fire risk potential. -The licensee of Haddam Neck made some improvements to its plant to reduce the seismic vulnerability, but decided to permanently shut down the plant.

x

, e o i 1

i i

j For most of the first 24 lPEEE submittels reviewed, licensees have not provided a consistent

- definition of vulnerability. In many cases, no definition of vulnerability was proposed, and licensees L simply stated that no vulnerabilities were found. 1 Many plants have reported some solamic and fire related pla/ *neprovement as a result of the

, IPEEE offert. A few also reported improvements in the high wind and flood areas. These ,

! Improvements take the form of chances to existing procedures, development of new procedures, or i plant modifications.

i 3 I

3 Specific strengths and weaknesses of each IPEEE submittal have been identified as a result of the NRC's technicai review process. Some submittels have bewt found to contain weaknesses or <

i deficiencies in one or more areas of their analyses (i.e., seismic, fire, HFO events). The i i deficiencies have arisen sometimes from inadequate documentation of key aspects of the analyses

~

and/or apparent mistakes or oversights with the potential to fundamentally impact the iPEEE results, i Interactions with the licensees, mainly through the process of requests for additional information '

j (RAI), have been conducted to obtain clarification of specific points in the submittel which were >

either uncisar or of questionable basis. These RAls have generally been limited to items considered j- to be of sufficient importance that the insights or findings of the IPEEE, or the reviewers'

understanding of those findings and insights, might be significantly impacted by the licensee

! response. Based on the results of these reviews, a preliminary conclusion is that most .

submittels have met the intent of the IPEEE, However, some submittels need additional review +

4

. because of either unusual results (i.e., extremely high or low CDFs) that were reported and/or the response was not adequate for the staff to conclude that the Intent of the IPEEE was met.

s Additional Perspectives and Observations l

A number of other important perspectives have been derived from the NRC's review of the IPEEE j submittels. The following paragraphs summartae these key observations separately for the seismic, .

1 fire, and HFO areas of the IPEEE program.

8eismic Events In reviewing the first 24 seismic IPEEEs, the staff observed the following key preliminary perspectives:

• Plant opecific improvements have been implemented at many plants. A seismic walkdown was performed for each plani; in most cases, the walkdown identified conditions pertaining to anchorages, interactions, maintenance, and/or housekeeping that required further investigation and often resulted in plant specific improvements being implementsd.

• The seismic IPEEE program has improved licensees' appreciation of the potential for and effects of relay chatter, At many plants, low-ruggedness relays have been identified to e limited extent (i.e., a small number of such relays has been discovered), in some cases, low ruggedness relays have been replaced; however, in most cases, relay chatter was screened out on the basis of a consequence assessment. In many cases, chatter was deemed acceptable because it was determined to be recoverable. '

e Seismic IPEEE studies of containment performance have improved appreciation of the potential for failure of containment cooling and isolation (including effects of relay chatter).

< in a few cases, concems or improvements have been identified with respect to containment cooling and isolation, in general, safety systems for maintaining the containment

.-a,nem~rm- .~,.r,n_ _-,_--_ar.,~ ,,-,-m--m- .,,m.w w e e--w. m n ,rr. N- , r vwwww U

integrity have been found to be rugged, and the seismio capabilities of these systems are typloally controlled by the seismio capability of the support systems.

• Some IPEEEs used simplifloations in systems analyses, unsubstantiated assumptions regarding human error rates, and simplified screening fregilities (e.g., using surrogate elements), lo some oasks, these have obsoured findings pertaining to dominant seismic risk contributors and produoed unrealistle (high or low) CDF estimatos.

• Among the IPEEE submittels reviewed, licensees have not employed a consistent spectral shape (characterizing the seismic demand) in analyzing fragilities and seismic capacities for components and plants. Hence, it would be misleading to o6mpare capaolties among PRA studies or for PRA versus Seismio Margin Assessment studies.

• The spectral shapes employed in seismlo PRAs for many plants in the eastern United states have not oloarty demonstrated plant seismio margins beyond the design basis.

in seismic PRA studies, licensees have used different hazard curves (e.g.,1993 Lawrence Livermore National Laboratory [LLNL),1989 LLNL,1989 EPRI, or licensee sponsored studies). Hence, it is diffloult to achieve a meaningful comparison of seismic CDFs across plants.

• The logic models for seismic IPEEE PRAs have been derived by modifying IPE logic models.

In some instances, the seismic IPEEEs do not provide adequate justification for screening out somo potentially important initiators (e.g., loss of cooient accidents (LOCAs), ,

steamline/foedwater line breaks, failure of reactor ir.temals, steam generator tube ruptures (SGTRs]) from the logic models.

Seismic fire and seismic flood evaluations, conducted as part of the IPEEE by means of piar,t walkdowns, have generally enhanced licensees' appreciat6on of the potential for seismically induced fires and the potential for and effects of inadvertent actuation of fire suppression systems. The most consistent strong points of these evaluations appear to be the treatment of inadvertent actuation of fire suppression systems and identifloation of potentialintereotion conooms. However, the level of effort and treatment of seismically induced fires and floods varied significantly among the IPEEE submittals.

Fire Events in reviewing the first 24 fire IPEEEs, the staff observed the following key perspectivos:

. Overall, licensees have addressed almost all safety-related areas within their plants. Many licensees have undertaken extensive plant walkdowns to verify existing data and to collect additional information for fire risk analysis.

e. - A majority of licensees have apparently gained a qualitative understanding of the overall likelihood of core damage. CDF has been used in all submittels as a screening measure and to establish the importance of various fire scenarios. Lloonsees have typically conducted their fire analyses to a point at which they could convince themselves that the risk is acooptable.

• The results of IPEEE fire analyses confirmed the risk reduction achieved from the implementation of NRC fire protection requirements (e.g., Appendix R requirements). None

of the fire scenarios identified by licensees fail a minimal cutset of equipment leading to core damage, in other words, additional failures, somewhat independent of the fire, must occur for core damage to result.

The staff noted several weaknesses in applying the methodologies and data in some of the fire analyses.- These weaknesses affect the robustness and completeness of those submittaia, as follows:

Virtually all of the 24 submittels provided some assessment of room to room fire effects. However, for many of the submittels, the staff could not ascertain whether these assessments were property conducted. Past PRA results indicate that this can represent a serious problem for some (albeit a minority) of the plants. The possibility of barrier failule, which may have a significant probability of occurrence, has not been included in most analyses, in addition, active fire barriers have generally not been modeled property. (The significanoe of active fire barriors is a function of plant layout and separation of redundant trains.)

Several licensees have screened out important compartments (in terms of 3

combinations of systems that can feil) solely on the basis of fire frenuency, without sufficiently reviewing the potential equiprnent and instrumentation dainage possibilities.

In several cases, licensees have argued that, since a specific compartment contains qualified cables and no other equipment, the possibility of a fire in that compartment is very unlikely and therefore can be screened out from further analysis. Given the large uncertainties in fire occurrence frequencies for such compartments, the practice of earty screening has not allowed licensees to gain a clear appreciation of potential-accident sequences that can be attributed to fires in those compartments.

Operator actions in response to the effects of fire on systems has rarely been modeled in detail.

Several submittals have used questionable methods, procedures, or data for fire -

damage modeling.

Several submittels have used the Nuclear Safety Analysis Center (NSAC)/181 and/or the EPRI Fire PRA Implementation Guide documents, for which some optimistic guidelines and data have been identified.

HFO Events in reviewing the first 24 HFO IPEEEs, the staff observed the following key perspectives:

• The high winds and external floods portions of the IPEEE program have h&d some impact on improving safety at certain plant sites. For some plants, licensees have gained a greater appreciation for the potential risk impact of high winds /tomadoes and extemal flooding Some licensees have proposed or implemented plant improvements including procedural enhancements, severe accident management guidelines, and hardware installation. Procedural enhancements include sandbagging, closing or welding doors,-

i hooking up pumps, and creating new circuits to reduce the risk from flooding. Hardware i improvements include modifications to enhance flood protection at entry pathways as well as I

equipment (such as portable water pumps) to enhance flood protection, among other 4

examples. Some submittals also noted that hardware changes undertaken in response to j

_. L.... - - - - - --i -~- = - - - - -

the licensees' IPE analyses (for example, the addition of diesel generators) have reduced or eliminated the risk associated with HFO events.

• Accidents related to transportation and nearby facilities have been screened out in all I 24 iPEEEs reviewed to date.

1

• Even though flood hazards were screened out at some plants, the staff noted that a flood <

level just a few inches (or less) below the failure-incipient level might have an annual rate of )

occurrence of one to two orders of magnitude greater than the hazard for the failure-incipient i level. Given the large uncertainties in site specific flood hazard curves, screening may have  ;

been premature in some cases.

Many submittals simply used the IPE conditional core damage probability (CCDP), given loss of af'4& >wer and loss of service water without modeling the specific significant impacts of high d% e Goods. As a result, these submittals may underestimate the CDF for such eveis, xiv

ACKNOWLEDGMENTS This document was derived from a report,

• Individual Plant Examination of Extemal Events (IPEEE)

Program: Preliminary insights on Results, Plant improvements, and Imp;ementation Strengths and Weaknesses," dated May 19g7, prepared by Energy Research, Inc. (ERI), which presented the results from the technical reviews of the first 24 iPEEE submittels. The majority of those reviews were performed by ERI, under contract to the NRC's Office of Nuclear Regulatory Research (RES).

The following individuals contributed to the reviews of licensees' IPEEE submittels:

D. A. didwell ERIN Engineering (Formerly of PLO, Inc.)

R. J. Budnitz Future Resources Associates,Inc.

H. Esmaili Energy Research, Inc.

K. N. Fleming ERIN Engineering (Formerty of PLG, Inc.)

M. V. Frank Safety Factor Associates, Inc.

M. Kazartans Kazarians and Associates

. M. Khatib-Rahbar Energy Research, Inc. (Principal investigator)

A. S. Kuritzky Energy Research, Inc.

J. A. Lambright Lambright Technical Associates M. Modarros University of Maryland A. Mosloh University of Maryland

R. T. Sewell Energy Research, Inc.

S. Sholly Bets Corporation Intemational R. Vijaykumar Energy Research, Inc.

W. Womer Safety Assessment Consulting The following NRC staff members provided technical guidance and support for the IPEEE reviews:

G. Bagchi Office of Nuclear Reactor Regulation A,' J. Busiik Office of Nuclear Regulatory Research -

. T.Y. Chang Office of Nuclear Regulatory Research J.T.Chen office of Nuclear Regulatory Research N. C. Chokshi Office of Nuclear Regulatory Research E. Connell Office of Nuclear Reactor Regulation

. M. Cunningham  : Office of Nuclear Regulatory Research M. Drouin _ Office of Nuclear Regulatory Research D. Jong Office of Nuclear Reactor Regulation R. Komasiewicz Office of Nuclear Regulatory Rosaarch R.' Rothman Office of Nuclear Reactor Regulation A. Rubin Office of Nuclear Regulatory Research -

H. VanderMolen Office of Nuclear Regulatory Research Tee' ' *l evaluation reports were reviewed by a senior review board (SRB), consisting of the NRC s' irs listed above, as well as the following contractors:

. .Bohn Sandia National Laboratories

8. P. Nowlen Sandia National Laboratories.

SRB members also participated as reviewers of ERI's report, along with the following members:

R.'J. Budnitz Future Repources Associates,Inc.

' J. R. Lehnee . Brookhaven National Laboratory M. K. Ravindra - EQE Intemational, Inc.

t .

ABBREVIATIONS ADV - Automatic Depressurization Valve AFW Auxiliary Foodwater AOP- Abnormal Operating Procedures ATWS Anticipated Transients Without Scram BWR- Boiling Water Reactor CCDP Conditional Core Damage Probability CCW- Component Cooling Water CDF Core Damage Frequency CDFM Conservative Deterministic Failure Margin CFR Code of Federal Regulations -

CST Condensate Storage Tank DG Diesel Generator DHR Decay Heat Removal EFWST Emergency Feedwater Service Tank EPRI Electric Power Research Institute ERCW Essential Raw Cooling Water ERI Energy Research, Inc.

E88W Essential Station Service Water ESW Emergency Service Water EUS Eastem United States i- FCIA Fire Colnpartment Interaction Analysis FIVE Fire Induced Vulnerability Evaluation Method FPS Fire Protection System FSAR_ Final Safety Analysis Report GERS Generic Equipment Ruggedness Spectrum

.Gl Generic issue GlP Generic implementation Procedure GL Generic Letter G8I Generic Safety issues HCLPF High Confidence of Low Probability of Failure (Capacity)

HEP Human Error Probability HFO High Winds, Floods, and Other Extemal (Initiating Events)

HPCI H6gh Pressure Coolant injection HPC8 High Pressure Core Spray

HVAC. - Heating, Ventilation, and Air Conditioning IPE Individual Plant Examination IPEEE Individual Plant Examination of Extemal Events

- IRT independent Review Team 18LOCA interfacing System Loss of Coolant Accidsnt LLNL Lewrence Livermore National Laboratory LOCA Loss-of-Coolant Accident LOSP ' Loss of Off site Power

-LTSP Long-Term Beismic Program MCC Motor ControlCenter MFW Main Feedwater

.MOV Motor-operated Valve

- M81V Main Steam isolation Valve MSRP Multiple System Responses Program NFPA Nations Fire Protection Association "i

NRC U.S. Nuclear Regulatory Commission ,

NSAC iduelear Safety Analysis Center '

NSSS Nuclear Steam Supply System NUMARC Nuclear Management and Resources Council OL Operating License PGA Peak Ground Acceleration PMP Probable Maximum Precipitation PORV Pressure-Operated Relief Valve PRA Probabilistic Risk Assessment PSA Probabilistic Safety Assessment PSF Performance Shaping Factor PWR Pressurized Water Reactor RAI Request for AdditionalInformation RCIC Reactor Core Injection Cooling RCS Reactor Coolant System '

RG Regulatory Guido RHR Residual Heat Removal RHRSW Residual Heat Removal Service Water RLE Review Level Earthquake -

RWST - Refueling Water Storage Tank SBO Station Blackout SEP Systematic Evaluation Program SER Staff Evaluation Report SGTR Steam Generator Tube Rupture SISIP Seismically Induced Systems Interaction Program SMA Seismic Margin Assessment SME Seismic Margin Earthquake SMM Seismic Margin Methodology SNL Sandia National Laboratory SPRA Seismic Probabilistic Risk Assessment SPSA Seismic Probabilistic Safety Assessment SQUG Seismic Qualification Utilities Group SRB Senior Review Board SRP Standard Review Plan SRT Seismic Review Team SRV Safety Relief Valve SSE Safe Shutdown Earthquake SSEL Safe Shutdown Equipment List TER Technical Evaluation Report UHS Uniform Hazard Spectrum USl Unresolved Safety issue WUS Westem United States xvii 1

Glossary Active fire barrier - a fire border element that must be physically repositioned from its normal configuration to an altemate configuration in order to provide its protective function. Examples include ventilation system fire dampers and normally open fire doors.

Anomaly - an observed plant condition that deviates rom normal variations with an unknown risk significance Antielpeted transient without scram (ATWS)- an anticipated transient event not socompanied by an automatic reactor trip Appendla R fire area - an area, as defined in the Appendix R analysis, sufficiently bounded by fire barriers that will withstand the fire hazards within the fire area and, as necessary, to protect important equipment within a fire area from a fire outside the area. A fire area must be mtde up of fire barriers having at least a 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> fire rating or equivalent, with openings in the barriers provided with fire doors, fire dampers and fire penetration seal assemblies having a fire resistance rating at least equivalent to the barrier in which it is installed.

Appendix R requirements - fire protection requirements specified in Appendix R to 10 CFR 50,

• Fire Protection Program for Nuclear Power Facilities Operating Prior to January 1,1979"

" Bad actor" relay - a low ruggedness relay, as defined for U8l A-46/ Seismic Quanlification Utilities Group (SQUG)

- Barrier failure - the breaching of a fire barrier by a fire or other causes results in the spread of fire or fire damage to the protected side of the barrier element Sounding analysis - an analysis that uses conservative assumptions for model parameters to '

l obtain an upper bound estimate Category 1 structure - a structure designed to withstand a safe shutdown earthquake Charleston earthquake issue - the issue came about as a result of a U.S. Geological Survey letter in 1982 that pointed out the possibility that large, damaging earthquakes have some likelihood of occurring at locations that had not been considered in past licensing decisions. As a result of work carried out by the NRC and EPRI to resolve the Charleston earthquake issue, eight plants at five Eastem U.S. sites with the likelihood that large, damaging earthquakes beyond the licensing bases may occur and require further assessments Common cause failure - a single event that adversely affects two or more components at the  !

same time (also referred to as common mode failure)

COMPSRN - a computer code as detailed in NUREG/CR-4566,"COMPBRN lli - A Computer Code for Modeling Compartment Fires

• Component - an element of plant hardware designed to provide a particular function (for system modeling purposes, a component is at the lowest luvel uf detail in the representation of plant hardware in the models) xvui

Condellenal oore damage probability (CCDP)- a probability of reaching core damage given specific conditions (for example, a fire has failed all equipment in an area)

Conservative deterministie failure margin (CDFM)- an estimate of the high confidence of low-probability of failure (HCLPF) capacity of components obtained based on procedure recommended by EPRI Containment performance - a measure of the responsa of nuclear plant containments to severe occident challenges (containment performance is typically represented by the conditional containment failure probability)

Containment failure modes - descriptions used to classify the type of containment failure, such as isolation failure, bypass failure, and earty or late failure Control system intereotion - the potential effects of fire on the ability to achieve safe shutdown from either the control room or the remote shutdown panel; for example, a fire may damage the common circuits or cables shared by the control room and the remote shutdown panel; control

. system interaction is identified as Generic Safety issue 147 Coro damage - uncovery and heatup of the reactor core as a result of a loss of core cooling to the point where prolonged clad oxidation and fuel damage is anticipated .

Coro damage frequeney (CDF)- the frequency, per reaclor year, of an accident leading to core damage Cross sono analysis - the analysis of a potential fire scenario involving fire propagation between adjacent fire zones

'  : Dependoney - requirement extemal to a system, structure, or component (SSC) and upon which -

the SSC's function depends Design basis event - any of the events specified in the nuclear power plant's safety analysis that are used to establish acceptable performance for safety related functions (events include anticipated transients, design basis accidents, extemal events, and natural phenomena)

Dominant contributor- an accident class that has a major impact on the total core damage frequency, or a containment failure mechanism that has a malcr impact on the total radionuclide release frequency l- Eastern U.S. seismielty issue - formerly the Charleston earthquake issue -

Event tree - an inductive logic model that begins with an accident initiator or condition and progresses through a series of branches that represent possible system performance, human actions, or phenomena that yield either a safe, stable state or an undesirable one, such as core damage or containment failure Eaternal event - an event initiated outside the plant systems that can affect the operability of plant systems (examples include earthquakes, tomados, and floods and fires from sources outside the

_ plant);

s

External flood - a flood initiated outside the plant that can affect the operability of plant systems Fault tree - a deductive logic model used to identify fauhs required to lead to an undesirable event (fault tree analysis begins with an undesired top event and attempts to identify the sub-events that are necessary to cause the top event; fault tree analysis contrasts with . failure modes and effects analysis, which is a bottom up approach)

Foodend blood - a method to provide primary coolant makeup to a PWR and at the same time open the pressuriser PORVs or safety valves to blood off the primary fluid and romave the reactor heat Fire area - a physical area bounded on all sides by rated fire barriers (See Appendix R fire area)

Fire barrfor - a physical construct intended to limit or prevent the spread of fire and fire effects to the unexposed or protected side of the construct. Fire barriers mey include structural elements

. such as walls and floor / ceiling elements; seals for openings in these structural elements such as doors, penetration seals, and dampers; and localized systems for the protection of cable trays conduits or other localiasd equipment.

• Fire cornpartment -in fire analysis, a space bounded by non combustible barriers where heat and products of combustion from a fire within the enclosure will be substantially confined Fire compartment interaction analysis (FCIA) -- a step in the fire-induce vulnerability evaluation (FIVE) methodology for considering qualitatively the potential for fire spread between compartments and the consequences of such an event on plant shutdown Fire damage modeling - modeling of all the necessary fire damage sequences (f6re scenarios and fire induced sequences)

Fire 4nduce vulnerability evaluation (FIVE) - a quantitative screening technique developed under the guidance of the Severe Accident Working Group of the Nuclear Management and Resources Council (NUMARC) and the industry's experts to address the fire portion of the IPEEE Fire PRA methodology - a PRA methodology to estimate the core damage frequencies due to fire events Fire sones - a subsection of a fire area that is not fully bounded by rated barriers but which is expected to substantially contain the effects of a fire including the spread of fire and the spread of fire produds in sufficient quantities so as not to threaten equipment outside the zone.

Focused scope - a term used in NUREG 1407 to designate the seismic IPEEE scope for specified nuclear power plants, and includes a detailed walkdown of the safe shutdown equipment list (SSEL), evaluation of low ruggedness relays, scrooning of structures and SSEL items for a 0.3g PGA Review Level Earthquake, calculat/Mreporting of high-confidence of low-probability of failure (HCLPF) capacities for the weaker elements, and the plant HCLPF capacity Fragility - the conditional probability that a component or system would fail for hazard intensities less than or equal to a specified value (for example, some ground motion or response parameter)

Free fleid peak ground aseeleretlen (PGA)- peak acceleration of the ground in a free field (without a structure) during a seismic event Full esope - a full scope seismic IPEEE goes beyond a focused scope selsmic IPEEE and includes relay chatter evaluation and more effort in evaluating soil failure modes and the number of high confidence of low-probability of failure (HCLPF) calculations Functional internetlen - the potential effects of one component or system on another because of the functional dependency between them Generic Letter 88 20 - a generic letter issued by the U.S. Nuclear Regulatory Commission on November 23,1988, which requested that U.S. nuclear utilities submit an Individual Plant Examination for severe accident vulnerobilities for each licensed nuclear power plant Generlo implementation procedures (GIP)- the screening guidance given in the Generic Implementation Procedure for Seismic Verificat6on of Nuclear Power Plant Equipment; GIP was developed under the sponsorship of Seismic Qualification Utility Group Generic lasue (GI)- an abbreviated name for generic safety issue (G81); accoMng to NUREG-0933, 'A Pr6oritization of Generic Safety issues,' a GSI is a safety concem that rNy affect the riesign, construction, or operation of all, several, or a class cf nuclear power plants and may hPve the potential for safety improvements and promulgation of new or revised requirements or guidance Ground motion response - the intensity of g,round shaking in the free field as a function of period or frequency Hasard - a source of risk (e.g., combustible material, high pressure piping, chemical solution, radionuclide)

Hasard surve - A monotonically decreasing curve indicating the frequency per unit time of an extemal event of a specified severity or greater occurring at a specific site; most often used for earthquakes and high winds; only one parameter is used to describe the event severity

- High wind, flood, and other external events (HPO)- the extemal events examined in an IPEEE excluding seismic and fire events are HFO events, namely, high wind /tomado, extemal flood, and transportation and iwar facility accidents High confidence of low probability of failure (HCLPF)- the earthquake acceleration level at which there is a g5% confidence that the chance of a specific structure or component failure is less than 5%, when the structure or component is subjected to an earthquake of that magnitude Hot short- an electric cable failure mode resulting from a fire that involves shorts between electrical conductors without a simultaneous short to ground or open circuit cWition. Such a fault might, for example, simulate the closing of a control switch, cause errors in an a_trument reading or result in the application of power to an unpowered circuit.

Human error rate - a measure of the likelihood that the operator will fall to initiate the correct, required, or specified action or response needed to allow the continuous or correct function of an item of equipment; human error rate and human error probability are used interchangeably w

Human error probability (HitP)- the probability that the operator will fail to initiate the correct, required, or specified action or response needed to allow the continuous or correct function of an item of equipment in core flux mapping system - a system used in a PWR to measure the magnitude and distribution of neutron flux in the reactor core Individual plant examination of external events (IPEEE)- Generic Letter 88 20, Supplement 4, requested U.S. nuclear utilities to perform an evaluation to identify any plant specific vulnerabilities to severs accidents initiated by extemal events during full power operation Individual plant examination (IPE)- Generic Letter 88 20 requested U.S. nuclear utilities to perform an evaluation to identify any plant specific vulnerabilities to severe accidents initiated by I

intemal events during full power operation interfacing system loss of coolar.1 accident (ISLOCA)- a loss of coolant accident resulting from the Intrusion of primary system fluid into a secondary system Internal events - accident initiators originating in a nuclear power plant and, in cnmbination with safety system failures and/or operator errors leading to core damage accident sequences internal fire - a fire initiated anywhere within the plant boundaries including both within plant structures and buildings and contiguous outdoor areas such as the electrical switchyard and transformer areas Level 2 probabilistic safety assessment (PSA) - evaluation of containrr? 1t response to severe accident challenges and quantification of the mechanisms, amounts, and probabilities of subsequent radioactive material releases from the containment Long terrn seismic program (LTSP)- a long term seismic monitoring program implemented at the Diablo Canyon nuclear plant Loss of coolant accidents (LOCA)- an accident caused by a break in the reactor coolant system pressure boundary Low flame spread cable - a cable that is certified to pass the flame spread portions of the IEEE-383 qualification test standard Low-ruggedness relay -a relay or relay type device which has the potential to change state or to chauer under a relativer/ low intensity seismic event Minimal cut set - A cut set is a combination of a set of events (e.g., initiating event and component failures) that, if they occur, will result in an undesirable condition (such as the onset of core damage or containment failure). A minimal cut set is a necessary and sufficient combination of the set of events that would result in the undesirable condition Mission time - the time period that a system or component is required to be operable in order to cairy out its mission; (for example, a containment spray mission time of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> implies that containment sprays are required to be operable for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> in order to prevent containment failure from occurring within that period) xxii

. i Muhlple system responses program (MSRP)- NUREG/CR 5420,

• Multiple System Responses Program-Identification of Concems Related to a Number of Specific Regulatory lasues*

National fire protection association (NFPA) standards - standards endorsed by NFPA for ,' ire protection; for example, ' Standard for the Installation of Sprinkler Systems

• i NEl 9104 " Severe Accident Closure Guidelines" - guidelines proposed by Nuclear Management and Resources Council (NUMARC) for identi8ying vulnerabilities to severe accidents N14C seismic margin assessment (8MA) methodology - the NRC SMA methodology uses event tree / fault tree modeling instead of success paths used in the EPRI SMA methodology Outlier- a component that cannet be screened out because a condition is encountered in a seismic walkdown or document review, that violates one or more key criteria of standard screening tables Passive component - a component or part of a component which performs its intended function without any moving parts or changes in state (e.g., tanks, piping runs, valve bodies, ductwork, etc.)

Passive fire barrier- a fire barrier that provides its protective function while in its normal orientation without any need to reposition. Examples include walls and normally closed fire doors Peak ground acceleration (PGA)- in an IPEEE, same as the free field peak ground acceleration Performance shaping factor (PSF)- an influence on the performance of an operator (undertying PSFs is the idea that the human error rates for a set of specified actions can be derived by investigating how a small set of PSFs influence the likelihood of success or failure of the operators; PSFs include such considerations as training, experience, availability and quality of a procedure, stress, interdependence among operators, environment, and timing)

Plant - a general term used to refer to a nuclear power facility (for example, plant could be used to refer to a single unit or a multi unit site)

Plant logic model- a mathematical rsoresentation that simulates the behavior of the plant following an initiating event. It is used to delineate sequences of events that,lf not prevented, could result in a core damage state and to quantify the sequence frequencies. A plant logic model typically involve the development of an event tree and its associated fault trees.

Plant-level capacity- The capacity of a plant to resist the effects of a hazard. A plant level capacity for a seismic hazard is typically represented by the lowest HCLPF value of the components and structures in the plant (e.g., in the most rugged success path in an EPRI SMA).

Probability risk analysis / assessment (PRA)- for a nuclear power plant, an analytical process that qu:.:.t'fies the potential risk associated with the design, operation and maintenance of a plant to the health and safety of the public; the risk evaluation involves three sequential parts or " Levels" Probable maximum precipitation (PMP)- the probable maximum rainfall as stated in Generic Letter 89 22 xxiii

a s i f

t Quellflod emble - a cable that is certified to meet all the requirements of the IEEE.383 standard i (including both the flame spread and the LOCA exposure test protocols)

{

v Random failure - a failure event whose occurrence is represented by probability distribution. l Typically, in IPEEEs, the term is used to refer to failure events not caused by the extemal event being analyzed i

Rosetor coolant pump (RCP) seal LOCA - a loss of coolant accident (LOCA) resulting from a l failure of a reactor coolant pump seal  !

Rated fire barrier- a fire barrier with a fire endurance rating established consistent with ASTM l testing standards.

Roseter year - a period of the reactor operation that accounts for the downtime during a calender l year i Roeovery action - en operator action intended to briing failed or unavailable equipment back to operable status Reduced scope - a reduced scope seismic margin method which emphaslaes the walkdown that - I l la accepted by the NRC for performing the seismic portum of the IPEEE for sites with low seismic hazard ,

I Relay ohatter - the changing of states (for example, closed to open) for electric relays, conductors, and switches during a seismic event i Remote shutdown panel (F;SP) - a panel of instrumentation and control located outside the l control room for shutting down the reactor in the event that the control room is unavailable (for example, a fire in the control room); note that for seine plants multiple panels in diverse locations ,

may collectively represent the plant remote shutdown panel i Request for additional information (RAl) - questions sent to a licensee from the NRC to seek additional information not available in the IPEEE submittal Review level earthquake (RLE)- the specific earthquake level at which the high confidence of -

. low probability of failure (HCLPF) seismic review is being conducted l Roof pending - the accumulation of rain water on the roof of a structure c Safe shutdown earthquake (SSE) - the design basis earthquake defined for a nuclear power i plant por Appendix A to 10 CFR Part 100 Safe shutdown earthquake (SSE) spectrum - the ground response spectrum associated with a safe shutdown earthquake Safe shutdown equipment list (SSEL)-- the list of equipment required for safe shutdown following

. a safe shutdown earthquake Safe shutdown model- modeling of systems and components needed to bring a plant to safe shutdown g ,

Scaling factor- a factor used to estimate seismic margin earthquake (SME) demand (in structure f

response spectra) from (scaling) previously performed design or reevaluation building response analyses Screening analysis - an analysis used to narrow the list of all components of a plant to a smaller list which needs further review (for example, the 6 mponents with known high confidence of low-probability of failure (HCLPF) values above the review level are screened out)

Seismic capacity - the ability of a component to sustain a seismic impact measured in terms of the seismic capacity levol (e.g., peak ground accelerstion) below which the component continues to perform its function Seismic demand - a seismic demand refers to a ground response spectrum (for example, the safe shutdown earthquake (SSE) ground response spectrum can be considered as a seismic i

demand) l Selsmic-fire interaction - the effects of a seismic event causing an occurrence of a fire or the degradation of fire protection features l

Seismic flood evaluation - the evaluation of a flood scenario resulting from a seismic event Seismic hazard - any physical phenomenon (e.g., ground shaking, ground failure) associated with an earthquake that may produce adverse effects Seismic margin - the margin to accommodate earthquake ground motion levels well above the safe shutdown earthquake (SSE) ground motion level Seismic margin assessment (SMA)- methodology developed for assessing seismic capacities of nuclear power plants. Two seismic margin methodologies (NRC and EPRI), as described in NUREG 1407, were developed for performing such assessment of seismic margins Seismic margin earthquake (SME)- the specific earthquake level at which the high-confidence of low probability of failure (HCLPF) seismic review is being reviewed Seismic margin methodology (SMM)- See SMA above Seismic PRA (SPRA) methodology - a PRA methodology to estimate the core damage frequencies due to seismic events Seismic Quantification Utilities Group (SQUG) - a group of utilities joined together to gather experience data on component behavior in past earthquakes Senior review board (SRB)- a panel of IPEEE reviewers consisting of the NRC staff and NRC's consultants from nationallaboratories who are experts on extemal events Sensitivity analysis - an analysis in which one or more input parameters to a model are varied in order to observe their effects on the model predictions Settlement - downward movement of a building or other facility due to compaction of earth beneath its foundation xxv

Severe accident - an accident that goes beyond the design basis of the plant and usually involves extensive core damage Severe accident management - strategies and guidance developed for incorporation into the emergency response procedures of a plant to prevent or mitigate events during a severe accident Soll liquefaction - sollis changed to a " fluid like" state during a seismic event Soll etructure interaction (SSI)- the potential effects of soll foundation on its structure during a seismic event Spatial Interaction - the potential spatial interfering effects on multiple systems, structures, and components (SSCs) because of a common environmental hazard (e.g., seismic event)

Spectral shape - the shape of a response spectrum associated with a seismic ground motion l Spurious actuation - an actuation of a component or system by a spurious signal l

Standard Review Plan (SRP)- the NRC's 1975 SRP containing the criteria of plant / facility design and operation Station blackout (SBO)- an accident sequence initiated by loss of all offsite power with failure of onsite emergency AC power (diesel generators)

Steam generator tube rupture (SGTR)- an accident involving a rupture of at least one steam generator tube Step 2 review- a review of Tier 2 IPEEE documentation located in a plant Step 1 review - a review by the NRC and/or its contractor of the licensee's IPEEE submittal and associated documentation; a request for additionalinformation (RAl) may be sent to the licensee based on this review Submittal only review- a review of an IPEEE hased on information in the IPEEE submittal and additional information obtained from the licensee's responses to requests for additional informatbn (RAls)

Success path - a set of components whose operability and survivability are required to bring a plant to a stable condition (either hot or cold shutdown) and maintain that condition for at least 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> Surrogata element - a representative element used in the seismic PRA (SPRA) to account for the effects of the components that are screened out during the walkdown and screening phase of the SPRA. The use of a surrogate element results in the representation of the failures of several components by the failure of a single surrogate element j

Systematic Evaluation Program (SEP)- an NRC program for examining a number of old plants for regulatory issues xxvi

Technical evaluation report (TER) - a report prepared by NRC contractors or staff detailing the technical review findings of an IPEEE Tier 2 documentation - supporting IPEEE documentation (for example, notebooks and detailed calculations) located on a plant site or in a corporate office Tomado missile - a flying object produced by a tomado Translent - a change in the reactor coolant system condition (temperature and/or pressure) that would result in a reactor trip and the need of core heat removal. Transients can be caused by events related to the balance of plant (e.g., turbine trip, loss of feedwater) or events associated with plant support systems (e g., loss of service water, loss of AC bus)

Transient fuel- Combustible materials which are not part of plant systems, structures, and components (SSCs). Transient fuels typically are associated with maintenance or plant modifications, or the temporary accumulation of materials within the plant Transportction and near facHity accidents - transportation accidents involve moving vehicles (i.e., planes, ships, barges, trucks, and railroad cars) near the plant causing explosion (overpressure and missiles), impact with the plant structures / components, release of hazardous material, and formation of a traveling vapor cloud with potential for ignition / explosion; near facility accidents involve accidents of industrial facilities near the plant with release of hazardous material, and accidents of rupture of a pipeline carrying a hazardous gas or liquid under pressure Uncertainty analysis -the quantification of the uncertaint%s in the PRA estimate that results from uncertainties in the PRA models and their input variables Uniform hazard spectrum (UHS)- a response spectrum which is made by connecting independently predicted spectral values for several frequency points at a given annual exceedance probability (usually expressed by a retum period), for a constant percentile (e.g.,50th percentile for median spectra), and for a given damping value (typically 5%)

Unit - refers to a single nuclear power reactor with its associated systems and components; most nuclear power plant sites have one or more units; at multi-unit sites, some support systems can be shared between units Unresolved safety issue (USI)- according to NUREG 0933, "A Prioritization of Genene Safety issues," a USl is defined as a matter affecting a number of nuclear power plants that poses important questions conceming the adequacy of existing safety requirements for which a final resolution has not yet been developed and that involves conditions not likely to be acceptable over the lifetime of the plants affected USl A 17 - Unresolved Safety issue (USI) A 17, ' Systems interaction in Nu.,iear Power Plants' USl A45 - Unresolved Safety issue (USI) A-45, ' Shutdown Decay Heat Removal Requirements

• USl A46 - Unresolved Safety issue (USI) A 46, " Verification of Seismic Adequacy of Equipment in Operating Plants," assesses the seismic ruggedness of safety related equipment to withstand a safe shutdown earthquake in those plants with construction permit applications docketed before about 1972 xxvu U

Walkdown -Inspection of local areas in a nuclear power plant where systems and components are physically located in order to verify the location of the equipment, assess its operating status, and ascertain any environmental effects or system interaction effects on the equipment which could occur during accident conditions

~

xxviii l

-1'- INTRODUCTION

- 1,1 Backaround On November 23,1988, the U.S. Nuclear Regulatory Commission (NRC) issued Generic Letter (GL) 88 20, Individual Plant Examination for Severs Accident Vulnerabilities,10 CFR 50.54(f),*

to l6censees of nucisar power plants. Spelfical y, GL 88 20 requested that the licensees

• pertbem a systematic evaluation of ex!) ting plants to identity any plant specHic vulnerebilities to

- severe accMents and report the resuMs to the Commission.* GL 88 20 also outlined the objectives and ovetall logistics of the Individual Plant Examination (IPE) program, which solely addresses intomally initiated events (including intomal flooding).-

On June 28,1991, the NRC issued Supplement 4 to GL 88 20,

• Individual Plant Examination of -

Extemal Events (IPEEE) for Severe Accident Vulnerabilit6es,10 CFR 50.54(f)." That supplement described the objectives and overall logistics of the Individual Plant Examination of Extemal Events (IPEEE) program, which addresses extemally initiated events. In particular, the extemal events considered in the IPEEE program include seismic events; intomal fires; and high winds, floods, and other extemal initiating events (HFOs) involving socidents related to transportation and neart>y facilities. The Commission formulated both the IPE and IPEEE programs in

. response to the NRC's ' Policy Statement on Severs Accidents Regarding Future Designs and Existing Plants", issued on August 8,1985 (Federal Register, 50FR32138). In particular, these programs were intended as a means for licensees to identify potential vulnerabilities to severe accidents, and to conceive cost effective improvements to ensure that plants do not pose any

' undue risk to public health and safety. The impetus for this policy statement arose from perspectives developed from earty probabilistic risk assessments (PRAs)c This impetus was further emphaslaed by the general finding that systematic examinations have been beneficial in identifying plant specific vult erapilities to severs accidents that could be fixed with low cost

=

improvements.

Along with Supplement 4 to' GL 88 20, the NRC losued NUREG 1407, " Procedure and Submittel Guidance for the Individual Plant Examination of Extemal Events (IPEEE) for Severe Accident Vulnerabilities," dated June 28,1991. In NUREG 1407 (Chen et al.,1991), the NRC provided .

guidelines for conducting IPEEEs. Specifically, the guidance pertained to evaluations concoming the following extemal initiators: seismic events; infomal fires; and high winds, floods, and other

- extemal events (HFOs) involving accidents related to transportation or nearby facilities.

Subsequent to the publication of NUREG 1407, Supplement 5 to GL 88 20 (USNRC,1995) was

--Issued on September 8,1995, to notify licensees of modifloations to the recommended scope of the seismic analysis portion of the IPEEE for certain plant sites in the eastem United States .

i (EU8).

Consistent with the intent of GL 88 20, the primary goal of the IPEEE program has been for lhoensees to *\ identity plant spool 6c vulnerabilities to severe accidents that couM be Mxed with

- low cost improvements.* More specifically, Supplement 4 to GL 88 20 identified the following

four supporting IPEEE objectives for each licensee to achieve:

ic Develop an appreciation of severe accident behavior. -

2. Understand the most likely severe accident sequences that could occur at the licensee's plant under full-power operating conditions.

3; Ga!n a qualitative understanding of the overalllikelihood of core damage and fission 1

- product releases.1 i

, , f I

4. Reduce, if necessary, the overall likelihood of core damage and radioactive material i

, releases by modifying, where appropriate, hardware and procedures that would help prevent or mitigste severe accidents. 3 i

1.2 Overview of the Technical Review Process for IPEEEs  !

The priraary objective of the NRC's technical review process is to ascertain the extent to which  !

the licensees' IPEEE submittals have achieved the intent of GL 88 20, satisfied the four principal  !

1 lPEEE objectives listed above, and followed the recommended guidance in NUREG 1407. l However, the reviews are not intended to validate or vertfy the results of a licensee's IPEEE.

, As originally conceived, the review process comprises a "Stop 1" review of each submittal, with follow-on

• Step 2" reviews of individual submittels on an "as needed" basis. The Stop 1 reviews

,. consider only the submittalitself. This means that none of the undertying or supporting (second tier) documents are examined. Step 1 reviews also include interactions with licensees in the form of Requests for Additional Information (RAls), conference calls, or public meetings. The -

objective of these interactions is to obtain clarification regarding specific points in a given submittal which are either uncient or of questionable basis. These RAls are generally limited to items considered to be of sufficient importance that the insights or findings of the IPEEE, or the reviewers' understanding of those findings and insights, might be significantly impacted by the ,

licensee's responses.

l if,'at the end of the Step 1 review, the reviewers cannot conclude that a given submittal has met i I the intent of the IPEEE process, or if the submittel reported unusual results (i.e., extremely high a j or low core damage frequencies [CDFs) or a high confidence of low probability of failure i l

[HCLPF)), a Step 2 review might be undertaken. A Step 2 review typically includes further i l licensee interactions (i.e., review of supporting second tier documents, a plant visit, interviews with plant personnel, and a plant walkdown) to resolve identified conoems.  ;

The NRC has also convened a Senior Review Board (SRB) to oversee the technical aspects of the review process. The SRB includes NRC staff members and contractors who are experts in the field of general risk assessment and the specific areas addressed by the IPEEE analyses ,

(seismic events, intomal fires, and HFOs) The SRB members also perform abbreviated reviews of each IPEEE submittal and hold regular meetings at which the reviewers with primary responsibility for a given plant submittal present their findings, insights, and recommendations.

The SRB then comments on the completeness of the review, whether the reviewers' technical findings are of sufficient importance to warrant an RAI, and v; Nether tha submittal meets the  ;

iPEEE intent. The SRB participates in alllevels of Step 1 and Step 2 reviews.

, To date, the NRC has received 65 IPEEE submittels. Amoe.g the 49 t currently in various L

stuges of review, with the first 24 having completed Step t reviews. ',N najority of these 24 reviews were performed by Energy Research, Inc. (ERI), uredor contract to the NRC's office of L Nuclear Regulatory Research (RES).

1.3 Ohlectives of the IPEEE Pernosotives Pronram +

' i,1 addition to performing technical reviews of the IPEEE submittals, the NRC has instituted an IPEEE Perspectives Program to extract perspectives from the submittals. The objective of this program is to provide perspectives in the following areas:

l 4

2

4 l

e description of the overaillPEEE process, fmdings, and impacts of the major areas of

} evaluation for extemal inhistors (seismic events, intomal fires, and HFOs)

>

• overview of plant improvements related to the IPEEE program, with a description of their beneficialimpact on reactor safety

• identification and assessment of the impacts of site specific hazards, plant specific I

design and operational features, and modeling and screening assumptions that affect the 1 i- understanding of a plant's severe accident behavior and containment performance

• - description of the overall strengths and weaknesses in the implementation of evaluation i

methodologies, including the implications of assumptions consistently made in IPEEEs

• summary of the extent to which the licensees have met the intent of Supplement 4 to GL 88 20 4

3 This report presents the preliminary perspectives gleaned through this program.

1.4 Boose. Limitations. and General Comments Portainina to Findinns of the IPEEE Peramentives Pronram 4

To date, IPEEE studies have been limited to the consideration of plant behavior under full power operating conditions, and the results have been influenced by a wide spectrum of factors. These ,

factors naturally include the basic intomal plant characteristics, such as plant type (boiling water reactor [BWR) or pressurized water reactor [PWR)), plant layout, and so forth. Other related factors include proximity to sources of earthquakes, high winds, floods, and other natural -  :

hazards; proximity to sources of man made hazards, such as transportation routes and industrial '

l facilities; structural design of buildings and equipment; and exposure of buildings and equipment.

Additionally, the IPEEE submPtals myiewed to date used various sources of information, such as ,

use of seismic hazard curves derived from different sources (e.g., Lawrence Livermore National ,

Laboratory 1993 [LLNL); LLNL 1989; Electric Power Research Institute (EPRI) 1989; and site-specific studies), or applied simplified conr.orvative methods in some studies while others used more realistic approaches. These inconsistencies make it difficult to draw plant to-plant l comparisons of analysis results. Comparisons of IPEEE results among plants and among the various types of extemal hazards are also limited because of variations in the quality of submittels. Hence, the staff made no attempt in this report to compare IPEEE findings with IPE results, or to compare IPEEE results among the various categories of extemal initiators. For the most part, then, discussions in this report are kept distinct for seismic events, intomal fires, and HFO initiators.

3

e l

It should also be emphaslaed that the perspectives documented in this report must be considered preliminary, because this study encompassed only 24 IPEEE submittels' and certain aspects of the reviews (Step 2 reviews of certain plants) are stillin progress. (Table 1.1 lists the L 24 plants ir,cluded in this report.)

1.5 Rasort Ohlectives The purpose of this report is to document the preliminary perspectives gained from technical I

reviews of the first 24 lPEEE submittels. These preliminary perspectives primarily include: (a) en assessment of the overall effectiveness in meeting the IPEEE objectives, (b) summaries of flndings and plant improvements reported in the IPEEEs, and (c) additional perspectives related l - to individual extemal events and the strengths and weaknesses of the licensees'submittels with L regard to their success in achieving the IPEEE objectives. The NRC staff has thus far considered only a subset of IPEEE submittels, and the staff has not yet finalized the technical reviews of every submittalincluded in this report. in addition, the perspectives documented in this report have been gleaned, for the most part, from " submittal only" reviews of the 24 IPEEEs; in other words, the perspectives have not generally benefitted from review of the detailed supporting documentation that NUREG 1407 requests licensees to maintain.

The perspectives documented in this report are somewhat general for the following reasons: (a)

IPEEEs are intended to yield predominantly qualitative perspectives, rather than more prescriptive quantitative findings; (b) IPEEEs address several different types of initiators of varying importance (for a given plant) and, therefore, require the implementation of different methods of analyses offering varying levels of de' ail and accuracy; and (c) even for a given type of extemal initiator, the procedures and methods used by the various licensees to conduct their IPEEEs have also varied considerably, in addition to an overall summary of IPEEE perspectives, this report discusses perspectives specific to the seismic, fire, and HFO areas of the IPEEEs, including detailed descriptions of the

. . findings, as well as the strengths and weaknesses of the submittals, in particular, the qualitative perspectives addressed in this report include a summary of the licensees' findings pertaining to their investigations of severe accident issues, including identification of plant improvements. in addition, the qualitative perspectives include a summary of the staff's observations conoeming the validity of licensees' methodologies and findings, as well as assessments of the consistency and potential usefulness (or limitations) of the IPEEE .

resultsc By centrast, the quantitative results include licensees' estimates of CDFs, release (or

, containment failure) frequency, and plant capability.

• This report was derived from a May ige 7 report prepared by Energy Research, Inc. (ERl) that presented results from the technical reviews of the first 24 IPEEE submittals. The majortly of those reviews were performed by ERI, under contract to the NRC's Office of Nuclear Regulatory Research (RES) Since then an addllional 17 IPEEE submittels have been reviewed by Sandia National Laboratories (SNL), Brookhaven National Laboratory (BNL) and RES staff in the fire, seismic, and HFO areas, respecWvely; and preliminary Technical Evalue60n Reports (TERs) have been completed for these submittals. One significant addluonal finding from these 17 reviews was the identfication of a fire vulnerability in the turbine building et Quod Citos (discupeed in this report). There are also some difference in the strengths and weaknesses from submittal to submittal. Despite these difference, with the excep#on of the fire vulnerability at Quod Cilies, the results from the additional 17 reviews generally support the overall preliminary persperctives obtainert from the first 24 reviews discussed in the Execu#ve Summary and Chapter 2 of this report.

4

j . .

i j 1.g Smoort Ornanlaation in developing this report, the staff sought to address each distinct, significant topic considered in NUREG 1407, including seismic events, fires, and HFOs, as well as the relevant generic safety issues (GSis) and unresolved safety issues (USts).

Section 2 presents conclusions drawn from this study, where an assessment is made of the extent to which the licensees have achieved the IPEEE objectives and met the principal intent of Supplement 4 to GL 88 20.

Section 3 of this report discusses the perspectives derived from the seismic portion of the IPEEE

- submittals, and includes comments regarding licensees' seismic probabilistic risk assessments .

(PRAs) and selsmic margin assessments (SMAs), it also discusses information provided in '

seismic IPEEE submi+tals relevant to specific GSfs and USIs.

Section 4 of this report discusses the perspectives derived from the fire portion of the IPEEE submittals, and includes comments regarding licensees' fire PRAs and fire induced vulnerability evaluation (FIVE) studies, it also discusses fire-related findings concoming specific GSis and <

Usts, as well as issues arising from the fire risk scoping study conducte by Sandia National Laboratories (SNL).

Section 5 presents findings derived from the HFO portion of the IPEEE submittals. Each major category of HFO initiator is discussed, including high winds and tomadoes, extemal floods, and accidents related to transportation or neart>y facilities, it also discusses HFO related findings concoming specific G81s and USls.

Sections 3 through 5 each provide summaries of applicable walkdown findings, human action perspectives, containment performance perspectives, plant improvements, generic versus plant-specific perspectives, as well n cbservations of specific strengths and weaknesses relevant to the evaluation of each particular type of extemalinitiator.

Section 6 lists the references cited throughout the report.

Table 1.1 Beslo Charsoteristlos of Plants included in this study Plant Name Loce90n Plant Type Containment Stortop Year Brunsw6ck Unns 1 and 2 Southeast North Carohne G E. BWR 4 Merk l 1975,1977 Cotswey Single Unit Cabewey County, Mesourt W PWR 4-L Lar9e-Dry 1985 Catawba Unas 1 and 2 York County, South Cerohne 1985,1986 W PWR 4 L los Condoneer Comanche Peek Unas 1 and 2 40 mi SW of Fort Worth, Tomas W PWR 4 L Large-Dry 1993 Cook Unas 1 and 2 Near Bridgmen, Michigen W PWR 4 L lee _ 1975,1978-Condoneer Diablo Canyon . Unas 1 and 2 Near Sen Luis Obepo, CA W PWR 4-L Large-Dry 1985,1986 L

Fort Calhoun 86ngle Una Near Omaha, Nebraska CE PWR 2 L Large Dry 1973 Heddam Neck Single Unt Haddom, Connectcut W PWR 4 L Large Dry 1968 Kewounee Single Unt Kowounce County,Woconsin W PWR 2 L Large-Dry 1974 LaSalle Unna 1 and 2 $$ mlSW of Ch6cago,Illnois G E. BWR 5 Mark Il 1984 Limonck Unns 1 and 2 Near Pottstown, Penneyhenis G E. BWR 4 Merk Il 1996,1990 McGuire Unna 1 and 2 Near Charlotte, North Carohne W PWR 4 L loe 1981,1984 Condoneer Mdistone ' Unt 3 SE Connec9eut WPWR4L Substmoesh 1986 Nme Mile Po!nt Unt2 Near Oswego, NewYork G.E. BWR 5 Mark Il 1988 Palmedes Single Unn Near South Haven, Mehigen CE PWR 2 L Large Dry 1971 Pugnm Single Unn Plymouth County, MA G.E. BWR 3 Mark-l 1972 Po6nt Beech Unas 1 and 2 Manitowoc County, Woconsin W PWR 2 L Large Dry 1970,1972 Robinson UnN2 Northoest South Carohne W PWR 3 L Large-Dry 1971 St Lucie Units 1 and 2 Hutchinson leland, Florida CE PWR 2-L Large Dry 1976,1983 Seabrook Single Unt Seebrook, New Hampshire W PWR 4-L Lorpo-Dry 1990.

Sequoyoh Unas 1 and 2 Near Chattanooga, TN W P W R 4-L les 1981,1982 Condonner South Texas Proj. Unas 1 and 2 89 miSW of Houston, Tomas W PWR 4-L Large-Dry 1988,1989 Susquehenne Units 1 and 2 Lute w County, Pennsylvanie G E. BWR 4 Mark-Il 1985 Turkov Point Units 3 and 4 n==me Bav Florida W PWR 3-L Laran-Dtv 1972 1973 6

a i

)

I 2 - PRELIMINARY CONCLUSION 8 '

l This section summarizes the preliminary perspectives and conclusions derived from the technical

!- reviews of the first 24 lPEEE submittels. Sections 3 through $ summarine the staffs evaluation 1

3 findings derived from the reviewing of licensees' IPEEE submittals with regard to the process and results of seismic, intomal fire, and HFO extemal events. Among these, the principal  ;

perspectives of this report include:- '

description of the overall IPEEE process, findings, and impacts of the major elements of evaluations performed for seismic events, infomal fires, and HFO extemal initiators e

overview of plant improvements that have been made as a result of the IPEEE program, i and a description of their beneficialimpact on reactor safety e

identification and assessment of the impacts of site specific hazards, plant specific design and operational features, and modeling and screening assumptions that adect the understanding of a plant's severe accident behavior and containment performance e

description of the overall strengths and weaknesses in the implementation of evaluation methodologies, including the assumptions made in IPEEEs e

summary of the extent to which the licensees have met the intent of Supplement 4 to GL 88 20 The following subsections pressat the preliminary observations and conclusions drawn from this study.

2.1 Overall Effectiveness in Meetina the intent of GL 88-20 and Achievina the IPEEE Ohlectives Consistent with the intent of GL 88 20, the primary goal of the IPEEE program has been for -

tioensees to

• identify plant speciMc vulnerabilities to severe accidents that could be Rxed with low cost improvements." More specifically, Supplement 4 to GL 88 20 identified the following four supporting IPEEE objectives for each licensee to me.hieve:

1. Develop an appreciation of severe accident behavior.

2. Understand the most likely severs accident sequences that could occur at the licensee's plant under full-power operating conditions.

3. _ Gain a qualitative understanding of the overalllikelihood of core damage and fission product releases.

4._ Reduce, if necessary, the overall likelihood of core damage and radioactive material releases by modifying, where appropriate, hardware and procedures that would help prevent or mitigate severe accidents.

Based on the reviews conducted to date, the IPEEE program generally appears to have been successfulin meeting the overallintent of GL 88 20. Licensees have expended significant effort in their IPEEEs, have acquirdd relevant knowledge concoming their plants, and have taken 7

i $

l Specific steps to improve plant safety. Thus, the IPEEE program can be considered generally  ?

successful, even though the degree of success achieved by licensees'iPEEEs has varied considerably, depending strongly on the methods and assumptions employed by the IPEEE analysts.

The following paragraphs summarize the overall effectiveness of licensees'iPEEEs in achieving i

cach of the four identified IPEEE objectives.

Objective 1: Appreciation of Severe Accident Behavior The staff's review of the 24 IPEEE submittals suggests that the IPEEE program has increased licensees' overall appreciation of sovere accident behavior attributed to extemal

, eunts at their plants. As requested in NUREG 1407, each licer.see has performed evaluations of seismic events,intemal fires, and HFO events. These evaluations have assessed the potential for extemally initiated severe accidents, and plant specific behavior in responding to potential severe accidents.

For the most part, licensees have been involved in both the management and execution of their IPEEEs. Licensees have sponsored training of their personnelin specialized facets of IPEEE analysis (e.g., seismic IPEEE training), to develop or enhance their appreciation of severe accident issues and of relevant plant behavior, in acccrdance w!th the request of NUREG 1407, licensees have undertaken peer reviews of their IPEEEs.

Objective 2: Understanding of the Most Liksly Severe Accident Sequences For the most part, licensees have gained a qualitative understanding of the most likely severe accident sequences that may occur as a result of extemal events. This understanding has been evidenced by the fact that licensees have identified (at least qualitatively) the relative risk significance of the various extemal events. Moreover, for each extemal event that was not screened out, licensees have identified the important initiators, as well as critical plant components, operator actions, and plant areas, and have acquired an understanding of their effect on plant systems.

Consistent with the guidance of NUREG 1407, licensees' emphasis in conducting IPEEEs has been on obtaining a qualitative (as opposed to quantitative) understai, ding. As expected, therefore, the IPEEEs do not generally convey a definitive ranking of the risk significance of severe accident sequences or their dominant risk contributors. Rather, by means of system modeling and screening analysis, licensees have. obtained a greater awareness of severe accident sequences and an improved sense of which are the most important sequences.

Ob}ective 3: Qualitative Understanding of the Likelihood of Core Damage and Fission Product Releases By means of IPEEEs, licensees have generally been able to ascertain whether the risk of core damage4associated with each extemal initiator is comparatively negligible (i.e., falling below the 10 per reactor year screening threshold), low, moderate, or high, in some cases, this understanding arose through direct quantification of core damage frequency (CDF); in other cases, the understanding resulted 1,om increased knowledge of the given

hazard in conjunction (where necessary) with an assessment of the plant's ability to withstand that hazard.

Each IPEEE submittal reported the findings of a qualitative evaluation of containment performance in response to seismic events. Some IPEEE submittals also reported a qucntitative estimate of the frequency of seismica:ly induced large releases. For intemal fire events, NUREG-1407 requests a containment analysis only if there are containment failure modes that differ significantly from those identified in the IPE; for HFO events, NUREG-1407 does not specifically request an assessment of containment performance.

Licensees have generally followed the NUREG-1407 guidance in these regards.

Although in many cases licensees have reported numerical risk estimates (for CDF c frequency of significant raaiological releases), it is important to note that the accurry "

such estimates is often limited because of simplifying assumptions and approximate procedures employed in the analyses. Hence, the results serve only as generalindicators of risk level, ano they should not be viewed as being well established.

Objective 4: Modifications to Reduce the Ukelihood of Core Damage and Fission Product Releases Overall, as a result of the IPEEE program, licensees have implemented or proposed plant E modifications that have had a ber.eficial effect on plant safety with respect to extemal events. Such modif;;.ations have taken the form of hardware changes, procedural changes, and implementation of severe accident management guidelines. Consistent with the qualitative nature of the iPEEE program, it is not usually possible to numerically deduce the risk reductions achieved by these modifications. However, some licensees have employed probabilistic risk assessment (PRA) in their IPEEEs as a means of determining whether plant modifications are warranted relative to a cost-benefit rationale.

Therefore, in considering the observation and perspectives derived from the review of the first 24 IPEEE submittals, the staff draw the following preliminary conclusions regarding the IPEEE program. Ll.censees have expended significant effort in developing their IPEEEs. As a result, they have acquired relevant knowledge conceming their plants, and have taken meaningful steps to improve plant safety. On that basis, the staff concludes that the IPEEE program has achieved a significant degree of success. Information from the IPEEEs may potentially be usefulin supporfng a variety of risk-informed regulatory activities; however, depending upon the specific application, more specific and detailed reviews are needed for that purpose.

2.2 Plant Improvements Of the 24 IPEEE submittals reviewed, a majority of licensees (e.g.,18 of 24 in the fire area) have proposed or implemented plant improvements to address some concems identified through the IPEEE program. These improvements have enhanced their plants resistance to severe accidents thei might result from extemal events. A few licensees have proposed no improvemena to enbarice plant capability with respect to important initiators. In most cases, hcvaver, this was because those licensees had already implementeri relevant plant improvements before the IPEEE program began.

In developing GL 88-20, the NRC expected that licansees could accomplish significant plant improvements at low cost. In keeping with that expectation, the majotity of the enhancements 9

proposed or implemented by licensees have been of relatively low cost, compared to major

- hardwara or design modifications. However, there have been a few instances in which the potential for significant enhancement was identified in an IPEEE or a prior evaluation (e.g., an IPE or investigation of an existing safety issue), it Nse instances, the licensees have -

Implemented or proposed certain improvement red int to the IPERE program that appear to .

- have involved significant cost.

It is important to note that most licensees have not applied the term vulnerability" in describing the plant conditions for which improvements were proposed or implemented, This occurred

' when the licensee did not believe a designation of " vulnerability" was justifedi (This reason applied, for example, when the licensee did not judge the condition to be sufficiently severe, or the condition did not meet specific criteria explicitly used by the licensee). Thus, it is often difficult to determine the degree to which the plant improvements have succeeded in reducing -

the identified risk.

Plant improvements relater! to seismic events have generally taken the form of various hardware -

fixes, maintenance actions, and enhancements to ;naintenance procedures. Hardware fixes have included such activities as anchoring equipment, bolting cabinets together, improving existing anchorage or supports, installing missing fasteners and bolts, installing spacers on battery racks, eliminating potential interaction concems, and replacing vulnerable relays.

Maintenance actions have included removing corrosion on equipment anchorages, and applying corrosion protection. Enhancements to maintenance procedures (primarily seismic housekeeping) have also included provisions for proper storage of ladders, tools, gas cylinders, etc., and for proper parking of cranes and chain hoists. Similar types of improvements have been implemented with respect to seismic. fire interaction concems.

Licensees have proposed or implemented plant improvements related to rire and HFO events.

For instance, licensees have planned certain improvements to fire protection systems, including hardware modifications and enhancements to, or 6velopment of, fire-response procedures.

Additionally, improvements have often taken the form of severe accident management guidelines that address specific accident scenarios related to intomal fires, potential effects of wind-induced missiles, and extemal flooding. Implementation of some of the severs accident management guidelines has led to the acquisition of temporary or portable equipment (pumps, diesel oil tanker

. trucks, etc.). One HFO IPEEE reported the strengthening of the stacks of two adjacent fossil-fueled units to reduce the high wind risk, and refurbishment of a flood wall to reduce flood risk.

In some cases, the IPEEEs also referenced plant improvements that were proposed or

!mplemented before the IPEEE program began, since those improvements resulted in a beneficial effect on paat safety for seismic, fire, and/or HFO events. For example, at one plant, the addition of diesel generators was identified as a plant improvement in the 'PE, and was correspondingly reported in the IPEEE since it reduced the risk of station blackout for seismic, fire, and HFO events.

' 2.3 - Summary of MaiorIPEEE Results One of the major preliminary findings of the IPEEE pregram is that seismic and fire events have been found to be important cont:ibutors to CDF for a majority of plants. In fact, CDF contribution from seismic or fire events can, in some cases, approach (or even exceed) that from intomal events (i.e., seismic CDF at Haddam Neck and fire CDF at Quad Cities). Core damage frequency estimates varied over several order of magnitude. For example, fire CDFs were 10

reported to range from less than 1x10* to 5.3x10'8 per reactor year (RY), while seismic CDFs were reported to range from 2.2x10 4 to 2.2x10"/RY.

The dominant risk contributor, to s91smic CDFs most commonly reported by licensees include seismic induced loss of offsite pov.or, failures of electrical and control panels, failures of block walls, and spatialinteractions. The ranking of dominant contributors has consistently been reported as being insensitive to the use of different seismic hazard curves.

The dominant fire risk areas most commonly reported by licensees include the main control room, cable spreading room, and switchgear rooms. Other frequently report _ed areas include all or selected parts of the turbine t ;.1, battery and DC equipment rooms, diesel generator rooms, ,

areas associated with component cooling water, and various cable routing areas.

A selsmic vulnerability at Haddam Neck and a fire vulnerabilltv at Quad Cities were reported by the licensees. The licensees for these two plants have used tne criterion recommended by Nuclear Management and Resources Council to define a " vulnerability" (e.g., CDF exceeds 1x10'

'/RY). The licensee of Quad Cities has implemented an interim altemate shutdown method i

involving the use of an independent back up power supply for both units to reduce the fire CDF from 5.3x10'8 to 7x10"/RY, and currently is in the process of evaluating long term options for j further reducing the fire risk potential. The licensee of Haddam Neck made some improvements to its plant to reduce the seismic vulnerability, but decided to permanently shut down the plant.

l

' - For most of the first 24 IPEEE submittals reviewed, licensees have not provided a consistent definition of vulnerability, in many cases, ao definition of vulnerability was proposed, and licensees simply stated that no vulnerabilities were found.

Many plants have reported some seismic and fire-related plant improvement as a result of the iPEM effort. A few also reported improvements in the high wind and flood areas. These improvements take the form of changes to existing procedures, development of new procedures, or plant modifications.

The staff identified certain strengths and weaknesses in the submittal as a result of the NRC's technical review. Specifically, the staff found some submittals to contain weaknesses or deficiencies in one or more arest of their analyses (i.e., seismic, fire, HFO events). The . -

deficiencies have arisen sometimes from inadequate docementation of key aspects of the analyses and/or overty optimistic assumptions or oversights with the potential to fundamentally impact the IPEEE results. The staff interacted with the licensees, mainly through the process of request for additional information (RAl), to obtain clarification of specific points in the submittal which were either unclear or of questionable basis. These RAls have generally been limited to

' items considered to be of sufficient importance that the insights or findings of the IPEEE, or the reviewers' understanM of those findings and insights, might be significantly impacted by the licensee response. Based on the results of these reviews, the staff drew a preliminary conclusion that most submittals have met the intent of the IPEEE. However, some submittals need additional review because either the licensees had reported unusual results (i.e., extremely high or low CDFs) and/or the response was not adequate for the staff to conclude that the licensees'submittals met the intent of the IPEEE.

11

2.4 Add ltienal Perspectives and ObserY.AUSDA A number of important perspectives have been derived from the NRC's review of the IPEEE submittals. The following subsections vummarize these key observations separately for the seismic, fire, and HFO areas of the IPEEE rogram.

2.4.1 Seismic IPEEEs 2.4.1.1 Lessons Leamed and Recommendations The principal lessons loamed from these IPEEEs include the following:

A seismic walkdown was performed for each plant, in most cases, the walkdown identified conditions pertaining to anchorages, interactions, maintenance, and/or housekeeping that required further investigation. As a result, plant-specific fixes have been implemented at many plants.

The seismic IPEEE program has improved licensees' appreciation of the potential and effects of relay chatter. At many plants, low-ruggedness relays have been identified to a limited degree. In a few cases, low-ruggedness relays have been replaced; however, in l

most other cases, relay chatter ws: semened out on the basis of a consequence and recovery assessment. In many cases, the licensees reported that relay chatter is deemed accep'able because the function of the Effected re!ay is recoverable.

Seismic IPEEE studies of containment performance have improved appreciation of the potential for failure of containment cooling and isolation (including effects of relay chatter). In a few cases, containment related concems or improvements have been identified with respect to containment cooling and isolation, in general, safety systems for maintaining the 'ontainment integrity have been found to be rugged, and the seismic capabilities of thesu systems are typical!y controlled by the seismic capability of the support systems.

Some IPEEEs employed simplifications in systems analyses, unsubstantiated assumptions regarding human error rates, and use of simplified screening fragilities, in certain cases, these deficiencies have obscured findings pertaining to dominant seismic risk contributors and produced unrealistic (unusually high or low) CDF estimates.

Among the IPEEE submittals reviewed, licensees have not employed a consistent spectral shape (characterizing the seitmic demand) in analyzing seismic fragilities and margin capacities for components and plants. Hence. a direct comparison of the seismic capacities among PRA studies or between PRA sid SMA studies would not be meaningful or may even lead to misleading conclusions.

The spectral shapes employed in seismic PRAs for many eastem United States ( EUS) plants have not clearly demonstrated plant seismic margin beyond the design basis (i.e.,

safe shutdown earthquake [SSE)).

In seismic PRA studies, licensees used different hazard curves (e.g.,1993 LLNL,1989 LLNL,1989 EPRI, and site-specific results) from plam to plant. Hence, it is difficult to obtain a meaningful comparison of seismic CDFs across plants. Nonetheless, the 12 I

ranking of dominant contribuinrs has consistently been reported in seismic IPEEEs as being insensitive to the choice of seismic hazard curves.

The logic mode's for seismic PRAs have been derived by modifying IPE logic models, in some instances, the seismic IPEEEs do not provide adequate justification for screening out certain important hitiators (e.g., LOCAs, steamline/feedwater line breaks, failure of reactor intemals, steam generator tube ruptures) from the logic models.

For seismic IPEEEs that used EPRs seismic margins methodology, most addressed the impacts of random failures and operator errors in a judgmental manner, In a few submittals, licensees used the screening criterion as recommended in NUREG-1407 to address these issues, With regard to seismically induced fires and floods, the staff noted the following observations:

Seismic-fire and seismic-flood evaluations conducted as part of the IPEEEs, by means of plant walkdowns, t. ave generally enhanced liceasees' appreciation of the potential for seismically induced fires, as well as the potential and effects of inadvertent actuation of fire suppression systems. The most consistent strong points appear to be the treatment of inadvertent actuation of fire suppression systems and identification of potentialinteraction concems. However, the level of effort and treatment of seismicclly induced fires and floods varied significantly among the IPEEE submittals, in addition, few IPEEEs discussed such potential failures from the perspective of loss of fire-suppression capability coincident with an earthquake. With respect to seismic-flood evaluation, only a few IPEEEs have discussed the potential for seismic failure of nonsafety tanks and piping.

In gt.neral, licensees closely coordinated their USl A-46, " Verification of Seismic Adequacy of Equipment in Operating Plants," and IPEEE seismic evaluations, including the implementation of related plant improvements, in order to better Lnderstand and enhance the potential future uses of certain seismic IPEEE results, the staff will consider the following:

Conduct a study to assess the impact of surrogate element modeling on CDF results and dominant contributors, e

improve the methods for assessing seismic impacts, on human error probabilities and relay chatter recovery actions.

A::sess the significance that the spectral shapes employed in certain EUS plants have not clearly demonstrated plant seismic margin beyond the design basis.

2.4.1.2 Generic Safety issues The funu fag issues have been addressed in seismic IPEEE submittals:

USI A-45, " Shutdown Decay Heat Removal Requirements" GI-131, " Potential Interactions involving the Flux Mapping System for Westinghouse Plants" 13

_ _ _ _ . . _ -~

The submittels have generally evaluated potential vulnerabilities related to decay heat removal '

(DHR) systems and, thus, the IPEEE progmm has served to resolve seismic concems related to USl A 45.

Wdh regard to Gl 131, all applicable plants completed upgrades or installed specific measures to address this issue, in some submittals, licensees assessed the adequacy of the flux mapping i _ system against the review level earthquake (RLE). In one case, the licensee resolved a unique

-interaction hazard by means of the IPEEE, Thus, the applicable IPEEEs have served to resolve the Gl-131 for those plants.

In addition, certain information providwd in the IPEEEs has effectiveY resolved specific seismic issues related to GSI 156, " Systematic Evuiustion Program (SEP)," and GSI-172, " Multiple

, System Responses Program (MSKI)".

2.4.2 Fire IPEEEs .

2.4.2.1 Lessons Leamed Key observations obtained from the reviews of the first 24 fire IPEEEs include the following:

Overall, licensees have expended considerable effort in their fire IPEEEs. In many cases, licensees have generated extensive databases and computerized models. Many have also undertaken extensive plant welkoowns to verify the existing data and to collect additional information for fire risk analysis. Thus, the staff concludes that, overall, __

licensees have addressed most safety-related plant areas, have been diligent in applying the selected methodologies, and have considered the possibility of spurious actuation caused by a cable fire.

In conducting the fire IPEEEs, licensees employed Fire induced Vulnerability Evaluation Method (FIVE), PRA, or a combination of the two methodologies.' For fire initiation data,-

licensees used the generic fire frequencies derived from overall industry experience.

Licensees gleaned cable and equipment location data primarily from the fire hazard analyses conducted for compliance with the requirements of Appendix R to 10 CFR Part

50. Licensees have s!so included additional cables, that were not included in the  :

Apperidix R analysis, to the list of elements for the plant response model. (The plant response modelis typically a simplified version of the intomal events model developed for the IPE.)

A majority of licensees have apparently gained a qualitative understanding of the overall likelihood of core damage. Licensees have used CDF as a measure to screen and to establish the importance level of various fire scenarios in all submittals. Licensees have typically conducted their fire analyses down to a point at which they could convince themselves that the risk is acceptable.

.- The results of the IPEEE fire analyses confirmed the reduction of risk that resulted from the implementation of the NRC's fire protection requirements (e.g., Appendix R requirements). None of the fire scenarios identified by licensees fail a minimum cutset of equipment leading to core damage, in other words, additional failures, somewhat independent of the fire, must occur for core damage to be realized.

14 i

Despite the demonstrated diligence of the licensees, the staff noted several weaknesses in-applying the methods and data in some of fire analyses which affect the robustness and

. completeness of certain submittals. These weaknesses are noted below:

Licensees used generic values for suppression system reliability without providing a basis l for their selection in terms of suppression system design features and adherence to National Fire Protection Association (NFPA) standards, Virtually all submittais provided some assessment of room-to room fire effects. However, most of these assessments were deficient in modeling the active fire barriers.

Several licensees screened out important compartments solely on the basis of the frequency of fire occurrence in the given compartment, without reviewing the potential equipment and instrumentation dam

• At certain plants, licensees' emergency safe shutdown procedures involve a deliberate self induced station blackout in the event of a fire to preclude the occurrence of spurious operation of equioi $nt resulting from hot shorts. However, none of these licensees assessed the ris'. Jact associated with this shutdown procedure in their submittals.

Licensees carely rlAeled the operator recovery actions in response to the effects of fire on systerr a in detail Most submittals appear to indicate that licensees used operator recovery error values directly from their IPE analyses without correcting the values using performance shaping factors affected by the fire.

CMain submittals indicated that the licensees used optimistic guidelines and data described in NEAC/181 (19g3) or EPRI's Fire PRA Implementation Guide.

2.4.2.2. Generic Safety lasues

- The fire-related generic and unresolved safety issues include GI 57 (safety significance of inadvertent and advertent actuation of fire protection systems), USI A-45 (decay heat removal requirements), and those identified in the Fire Risk Scoping Study (FRSS).

Most of the submittals reported that licensees performed walkdowns to ensure that Fire Protection Systems (FPSs) would not fall on safety-related components. Licensees also investigated the potential failure of safety equipment because of spray from water-based FPSs.

With regard to USI A-45, most fire IPEEEs included a risk assessment that yielded a CDF for fire-induced loss of decay heat removal scenarios. Other fire IPEEEs explicitly reviewed the IPE models associated with the loss of decay heat removal to determine the effect of fire on decay heat removal unavailability.

Wdh regard to FRSS issues, most submittats provided a typical response stating that circuit i

isolation, remote location, and procedures preclude control system interactions. However, the submittals provided little information concoming the verification process to ensure that control j system interactions do not represent a safety concem.

Most submittals provided little information conceming the potential for smoke buildup to hinder the effectiveness of manual fire-fighting or misdirect suppression efforts because manual fire suppression was not credited. A few licensees did consider electronic equipment damage or 15

degradation from secondary (non-thermal) fire environmental effects (for which there is very little empirical data), either in the submittals or in response to RAls, but provide little detail.

To investigated seismic-fire interactions, licensees typically performed plant walkdowns.

However, the investigation reported in most submittals was limited to the consideraticn of flammable liquids or gases as seismic-fire sources.

Many fire IPEEEs also included a walkdown to inspect passive fire barriers. In all cases, licensees assumed passive fire barriers to be 100% reliable. Most submittals also assumed that active fire barriers were 100% reliable.

2.4.3 EO mgggs On the basis of review the first 24 IPEEE HFO submittals, the staff concluded that most licensees have met the principal objectives of the IPEEE program, and that the HFO areas of the IPEEE program have had some impact on improving plant safety. For some plants, the IPEEE program gave licensees a greater appreciation of the potential risk impact of high winds /tomadoes and extemal flooding (including dam breaks), in addition, some licensees have proposed and/or implemented plant improvements, including procedural enhancements, hardware installation, and severe accident management guidelines. Procedural enhancements included sandbagging closing or welding doors, hooking up pumps, and creating new circuits to reduce the risk from flooding. In addition, two submittals reported that the licensees are considering development of severe accident management guidance to reduce the risk of high winds. Hardware improvements have included plugging flood entry pathways and providing portable water pumps. Moreover, some submittals noted that hardware changes undertaken in response to their IPE analyses (e.g., adding diesel generators) have also reduced or eliminated the risk from HFO events.

A number of important observations and perspectives have resulted from IPEEE HFO analyses.

These can be summarized as follows:

A few submittals reported CDFs associated with high winds and extemal flooding, which the licensees determined using either a PRA or bounding analysis These results indicated that risks associated with high winds and extemal floods for those plants located along rivers are of concem for a few plants.

In all 24 submittals reviewed, licensees' evaluations effectively screened out accidents  ;

involving transportation and nearby facilities.

Licensees have addressed GI-103, " Design for Probable Maximum Precipitation (PMP)"

and USl A-45.

One plant reported that plant-unique hazards, such as lightning and snow and ice loads, result in non-negligible risk.

At a few plants, even though the licensees have screened out flood hazards, the staff observed that a flood level just a few inches below the failure-incipient level might have an annual rate of occurrence of one to two orders of magnitude greater than the hazard for the failurtr-incipient level. Given the large uncertainties in site-specific flood hazaro estimates, licensees may have been premature in their screening.

16

A few submittals reported that the licensees considered potential failures of upstream dams, leading to flooding at their sites,

+

For many submittals, licensees simply used the IPE conditional core damage probability, given loss of offsite power (LOOP) and loss of service water, without modeling the specific significant impacts of high winds or floods. As a result, these licensees have sometimes underestimated the total CDF for such events.

2.5 Uses of this Report and the IPEEE Find!nas l

The reader should note that the staff focused its review primarily on the licensees' ability to extimine their plants for severe accident vulnerabilities. Although the staff explored certain aspects of the IPEEE in more detail than others, the review is not intended to validate the accuracy of the licensees' detailed IPEEE findings (or quantitative estimates). Information from IPEEEs may potentially be usefulin supporting a variety of risk-informed regulatory activities, however, dependir g upon the specific objective 0, more specific and detailed reviews will be needed for such uses.

l l

17

3- SEISMIC IPEEE PERSPECTIVES 3.1 Overview This section summarizes the key qualitative and quantitative findings derived from the staff's technical reviews of the seismic portions of the first 24 IPEEE submittels. In particular, the staff focused its review on assessing licensees' compliance with technical guidance for conducting --

seismic IPEEEs, documented in Section 3 of NUREG-1407 and in Supplement 5 to GL 88-20.

NUREG 1407 separates plant sites into the following seismic catego,ias:

• Plant Sites East of the Rocky Mountains

1. Reduced scope

2. 0.3g Focused scope

3. 0.3g Full scope

. 4. Committed to perform a seismic PRA

. Plant Sites in the Westem United States

6. 0.3g (Full scope)

6. 0.5g (Full scope) -

l 7. Seismic PRA These seismic categories have been established primarily on the basis of relative comparisons of plant to-plant seismic hazard. As highlighted in NUREG-1407, the common denominator for each of inese seismic evaluations is a well-conducted, detailed walkdown. The scope of the L analysis varies among these seismic categories, with higher-hazard sites designated for more extensive investigation. The following paragraphs briefly describe the applicability, scope, and procedures for each of these seismic categories:

• Reduced scope: This evaluation,- appropriate for the lowest hazard sites in the eastem United States (EUS), involves a detailed plant walkdown with outliers evaluated against the plant's design basis earthquake. The equipment list for the selected success paths is developed in accordance with the guidance described in EPRI's seismic margin assessment (SMA). No evaluation of soil failures is required. '

Focused scoce: This evaluation, appropriate for intermediate-hazard EUS sites, also involves a detailed plant walkdown. In this case, however, outliers are evaluated against a review level earthquake (RLE), defined by an earthquake with a peak ground acceleration (PGA) of 0.3g and a median spectral shape in accordance with NUREG/CR-0098. The equipment list, expanded to include containment performance in preventing large early releases, may be developed in accordance with the guidance described in either NRC's or EPRI's SMA. An evaluation of low-ruggedness re'ays is required, as is an evaluation of the effects of soll failures. Seismic capacities for thosa components / outliers expected to control plant capacity and the plant's high-confidenes of I8

o low probability of failure (HCLPF)* capacity are o'>tained . as the principal results of the -

SMA, Full scopei This evaluation is appropriats for the higher hazard EUS sites and some Westem U.S. (WUS) sites, it involves a ostalled olant walkdown with outliers evaluated .

against a review level earthquake (RLE) of 0.3g or 0.5g PGA and a median NUREG/CR-00g8 spectral shape. As in the focused scope evnluation, the equipment list is expanded to include containment performance in prov enting lame early releatc;, and may be developed in accordance with the guidance described in either NRC's or EPRl's SMA. A l

- full scope, detailed relay chatter evaluation is required, as is a full evaluation of soil l failures. In addition, seismic capacities are assessed for allidentified outliers, and a plant level HCf.PF capacity is obtained.

Seistric PRA: This evaluation includes a Level 1 seismic PRA with specified methodological enhancements in plant walkdowns, relay chatter evaluation, and analysis of liquefaction and other pctential soil failure modes, as appropriate. In addition, a seismic PRA evaluation should also include a qualitative assessment of containment performance (as in a full-scope analysis) or an analysis of the sequenos involving containment, containment functions, and containment systems with seismic failure modes <

or timing that are significantly different from those found in the IPE.

Table 3.1 identifies the NUREG-1407 seismic IPEEE category designated for each of the 24 plants included in this study. These designations define a minimum recommended level of evaluation, as described in NUREG-140", for aach p: nt site. Thus, a seismic PRA is acceptable for plants in any of the review categories, whereas, s ;tsmic margin methodology (SMM) is an acceptable approach only for plants in review categories 1,2, 3, 5, and 6. In addition, NUREG-1407 indicates that a licensee may propose an altemative seismic evaluation approach, which the NRC will consider with regard to acceptability for IPEEE purposes, I- Table 3.1 also identifies the seismic IPEEE categories and the methods that were actually implemented for each of the 24 plants considered in this report. Among the 24 plants,13 plants were evaluated with a seismic PRA,10 plants were evaluated with an SMA, and 1 was evaluated with both a seismic PRA and an SMA.

The findings yielded from a seismic PRA (SPRA) are notably different from that of an SMA. An EPRI SMA is principally a deterministic approach, whereas an NRC SMA may involve a determinictic or probabilistic assessment of component capacities, and a SPRA involves a full probabilistic treatment. The principal products of a SPRA consist of a probabilistic plant-level

[ capacity (i.e., fragility curve), an estimate of seismic CDF, and a list of dominant contributors to CDF; whereas the principal products from an SMA consist of a list of component capacities and an estimate of the HCLPF capacity of the plant. Hence, the discussion of seismic perspectives in this section distinguishes licensees' IPEEEs between these two types of evaluation methodologies.

Nevertheless, the evaluation of some aspects of the seismic IPEEE proceeds in the same manner for either a SPRA or a SMA. For example, the assessments of soil liquefaction potential

'The HCLPF capacity is a measure of the seismic margiti a plant possesses beyond its design basis. Explicitly,lt is the ground motion for which one has 95% confidence that the failure probability does not exceed 5%.

t 19 i-

(and other soll failure modes), relay citatter potential and its consequences, seismic-fire interactions, and other issues are essentially the same for both evaluation approaches.

Consequently, this report does not discuss separately the perspectives derived from the evaluation of such issues using SPRA and SMA approaches.

By contrast, this report discusses the peespectives derived from containment performance assessments separately for SPRA and SMA approaches. It is important to note, though, that, among the 14 SPRAs, only a few licensees chose to perform a quantitative evaluation of seismic containment performance (Level 2 PRA); while other licensees elected to perform a qualitative evaluation of seismic containment performance, in all EPRI SMA cases, the licensees performed qualitative assessments of seismic containment performance. One plant implemented an NRC SMA, together with a quantitative evaluation of seismic containment performance.

in addition to summarizing the qualitative and quantitative findings, this section draws comparisons among the seismic IPEEE results. Specifically, the comparisons were made for SPRA versus SMA methodologies, and to highlight similarities in findings among various plants.

An overall summary of the seismic IPEEE perspectives is provided in Section 2.2.

3.2 Impact of the Seismic IPEEE Pronram on Plant dafety The seitmic IPEEE program has had a notable impact on improving plant safety. Although only one IPEEE submittal reported seismic vulnerabilities, most submittals reported a substantial number of seismic " anomalies,"" outliers," or other concems. In the conte::t of this report, the term " anomaly" refers to a component or system of unknown seismic capacity. The term

" outlier"is used in this report to designate a component that cannot be screened out because a condition was encountered, in a seismic walkdown or documentation review, that violates one or more key SMM screening criteria. However, in many cases, it is not possib!e to make a clear distinction between an anomaly and an outlier. The terminology "open issue" has also been used in submittals as a general characterization denoting anomalies, outliers, or potential vulnerabilities.

b resolve open issues identified in IPEEEs, licensees have implemented or proposed a large r' umber of plant improvements with respect to hardware installations, procedural enhancements, maintenance and housekeeping actions, and severe accident management guidelines.

Haruware improvements have typically taken the form of anchorage improvements, elimination or minimization of the impact of potential adverre *ysical interactions, and replacement of vulnerable relays. Maintenance actions hav; suded the removal of corrosion on equipment anchorages, and application of corrosion protection. Enhancements to maintenance procedures (seismic housekeeping) have included provisions for proper storage of ladders, toon ess cylinders, and so forth, as well as proper parking of cranes and chain hoists, in a tw cases, equipment has been replaced. In some cases, licensees have undertaken changes to ensure that the installed condition matches the configuration assumed for licensing. In addition, a significant number of hardware improvements reported in the seismic IPEEEs have overlapped with improvements made for the USl A-46 Program, and a few have overlapped with improvements made as a result of the IPE Program.

20

3.3 Seismic PRA Perspectives A seismic PRA produces both quantitative and qualitative findings, as discussed in Sections 3.3.1 through 3.3.4. The key quantitative results summarized in this report include component fragility parameters, estimates of plant seismic capacities, and values of computed seismic CDF.

By contrast, the principal qualitative results include walkdown findings, dominant CDF contributors, and identified outliers and plant improvements.

3.3.1 Summary of Quantitative Findings As discussed below, the following unalytical elements are necessary for the quantification of seismic CDF in an IPEEE:

e seismic hazard results 3 e seismic fragility results for components (equipment and structures)

] .

seismic plant logic models numerical analysis to quantify accident sequence frequencies Seismic Hazard Results in general, licensees have employed one or two of the following seismic hazard results in seismic IPEEEs:

1989 EPRi seismic hazard results 1989 LLNL seismic hazard results 1993 (revised) LLNL seismic hazard results e

site-specific seismic hazard study results LLNL and EPRI hazard results are available only for plant sites east of the Rocky Mountains; however, EPRI hazard osults are not available for all EUS plants. Licensees have employed site specific seismic results for WUS plants and for those EUS plants for which EPRI hazard results were not available or for which the licenree chose to conduct a site specific study. For many submittals, licensees performed sensitivity studies to examine effects related to seismic hazards. In particular, these studies focused on the impacts of the type of hazard results (EPRI, LLNL, or site-specific) on seismic CDF and the ranking of dominant contnbutors, as well as the impact of the truncation of ground motion.

Seismic hazard results are used in two principal, related ways in seismic PRAs: (1) to develop seismic initiating event frequencies based on a seismic hazard curve, or (where plant / sequence.

level fragility curves are developed) to convolve conditional accident sequence probabilities with the hazard curve to obtain an annual frequency of seismically induced CDF; and (2) to characterize the ground motion response spectrum for use as input into component seismic fragility calculations. As shown in Table 3.2, licensees have used a variety of hazard results (or a combination of hazard results) to cliculate CDFs in IPEEEs involving seismic PRA. In many cases, the PRA spectral shape used in evaluating fragilities, and the hazard curve used for quantifying CDF were not derivod from a consistent set of hazard results. For example, the uniform hazard spectrum (UHS) derived from the 1989 LLNL hazard analysis has typically been used to define the SPRA spectral shape, whereas the seismic hazard curve derived from the 1989 EPRI or 1993 LLNL hazard analyses has typically been used to quantify CDF. Also, some submittals reported that licensees have used a spectral shape initially developed for a different plant.

21

.o . .

Seismic Frogility Evaluations Compoent seismic fragility results represent a significant product of IPEEEs. The 24 submittals reviewed have reported a voluminous quantity of relevant results. In addition, licensees have employed diverse methods in performing fragility calculations, representing the preferences of the analysts and the practical constraints imposed on the level of analytical detail. In general,'

SPRA IPEEEs have liberally employed various forms of simplified fragility analysis, in contrast to the predominance of detailed, conventional fragility analyses in past seismic PRAs. In some cases, the use of simplified fragilities has obscured findings related to dominant contributors to seismic CDF This has occurred because the dominant risk contributors have sometimes been derived based on the simplified fragilities, which are generally conservative.

in estimatir.g the seismic CDF, some licensees have developed plant level fragility curves in their IPEEE submittals. !n such instances, the licensees have frequently also reported plant-level HCLPF capacities. In other SPRA submittals, however, the licensees did not provide plant-level fragility curves and, in some of these cases, approximations to the plant level fragility and HCLPF may be inferred from the results presented in the submittal. Since reporting of the plant HCLPF -

capacity is optional in a seismic PRA IPEEE (Appendix C to NUREG 1407), such results are not -

always available from IPEEE submittals based on a SPRA.

. Seismic Plant Logic Models .

In general, licensees have developed plant logic models for SPRAs directly from the IPE event trees and fault trees for loss of offsite power (LOOP), small loss-of-coolant accidents (LOCAs),

and general transients, with modifications to include seismic structural failures, failures of i passive equipment (tanks, piping, ductwork, etc.), and to incorporate specific seismic failure correlations. In some cases, licensees also included cedain seismic failures (usually building failures) as additional initiating events. - Typically, licensees have developed a seismic event tree  ;

to map the seismic initiating events (i.e., occurrence of an earthquake, or occurrence of a -

particular level of ground motion) to the relevant IPE initiators (e.g., general transients, LOCAs, or anticipated transients without scram [ATWS]) ' In some cases, licensees' plant seismic models also include an initiating event category unique to seismic events; some of these include seismically induced intomal floods, extemal flooding attributable to seismically induced dam >

breaks, and seismically induced fires.-

. Licensees generally modified their IPE fault trees to include unique common-cause effects related to earthquakes, including failure dependencies and the effects of passive component..

failures. In most cases, the mission times for the seismic analyses have remained the same as those used for the IPEs.- In addition, seismic plant logic models have generally incitded recovery from seismically induced relay chatter.

Numencel Analysis in general, licensees have accomplished their numerical evaluations of seismic CDF, sequence frequencies, and so forth by directly quantifying event-tree sequences or by developing relovant

. fragility curves that were subsequently convolved with the given seismic hazard curve.

22

3.3,1,1-- Plant Seismic Capacity Results Table 3.2 lists the seismic CDF and plard HCLPF values repoited in SPRA IPEEEs. For plant sites east of the Rocky Mountains, the plant HCLPF results ranged from less than 0.05g to 0.50g of peak ground acceleration (PGA). For the only WUS plant included in this report, the licensee estimated an HCLPF valea of 0.67g PGA. The finding of unusually low HCLPF capacity (0.05g) is noteworthy; a more detailed staff review is needed in order to determine whether such unusually low ceipacity value relates to actual plant weaknesses, or primarily arises as a consequence of the assumptions and approximations used in the IPEEE analysis. Overall, the CDF findings highlight the value of performing plant-specific seismic studies to discover the degree and source of variations in seismic capacity estimates, in addition to the PGA value, a spectral shape is also needed to define a plant's HCLPF capacity.

As shown in Table 3.2, seismic capacity results developed from seismic PRAs are most often  ;

associated with a spectral shape derived from a uniform hazard spectrum._ in other cares, '

. licensees have used the median NUREG/CR 0098 spectral shape for evaluating seismic i capacity.

A noteworthy insight revealed from reviewing the seismic PRA portions of IPEEEs for some EUS i

plants is that, within the frequency range important to response (e.g.,2 to 15 Hz), the plant safe.

l-

' shutdown earthquake (SSE) spectrum (seismic demand) exceeds the plant HCLPF spectrum (seismic capacity) for some plants, in order to clearly state that a plant has adequate seismic margin, the HCLPF spectrum should envelop the SSE spectrum over all vibration frequencies relevant (J plant response. (The staff review should further explore this aspect as recommended I in Section 3.8 of this report.)

3.3.1.2 Core Damage Frequency Results Table 3.2point-estimate Specifically, also listsseismic the seismic CDF resultsCDF results presented obtained in this table from range from about licensees' 2x10 SPRA IPEE

, per reactor-year (ry) to 2.3='0*/ry. The broad variability in these results cannot be attributed to the selection of seismic hazards (e.g., LLNL versus EPRI), since the extreme seismic CDF -  !

values were obtained using the same hazard results (EPRI,1989). Rather, the broad variabil:ty probably results from a combination of many factors (6., eismic plant capacities, differences in seismic hazards, and inconsistencies in methods and ar, , cal assumptions),

it is also interesting to note that most SPRAs evaluated the seismic CDF using both EPRI and LLNL seismic hazard data. Only one SPRA evaluated the seismic CDF based solely on LLNL hazard data, in addition, only one case used the original LLNL seismic hazard curve (LLNL, 1989) to quantify the seismic CDF.

3.3.2 Summary of Qualitative Findings

' Qualitative findings of a seismic PRA are those that cannot be described by numerical results, although a quantitative analysis may often be necessary to derive the qualitative perspectives.

' For instance, qualitative findings of a SPRA conducted in an IP8EEE relate to the physical conditions and characteristics of a given plant, as well as the c Tiponents identified as most important to seismic risk, and the list of outliers and related plant improvements developed by the licensee.

23

3.3.2.1 Walkdown Perspectives o All SPRAs conducted in the first 24 IPEEEs involved a seismic walkdown by trained cr qualified personnel. In many cases, licensee personnel received Seismic Qualification Utilities Group (SQUG) training for the USl A-46 program, as well as the " add-on" IPEEE training.

Most of the SPRAs included in this report referenced use of EPRI NP-6041 SL (1991) or EPRI NP-6041 (1987) procedures for performing the seismic screening and walkdowns. All IPEEE submittals for USI A-46 plants also referenced the use of the walkdown procedures and criteria described in the Generic Implementation Procedures (GlP) (SQUG,1992).

Based on the results of walkdowns, licensees have reported outliers and anomalous conditions at many plants pertaining to the following general issues:

adequacy of equipment anchorage

] *

• functional adequacy of equipment

• physicalinteractions e

seismic maintenance and housekeeping Of the 14 seismic PRAs considered in this report,8 reported outliers or anomalies pertaining to at least one of these issues. Because of the overlapping scope of USl A-46 and the seismic IPEEE, many of the USl A-46 plants have reported plant improvements related to USl A-46 in the IPEEEs.

Some IPEEEs have assumed an upgraded plant condition, reflecting USl A-46 improvements, even though such improvements had not actually been implemented. Conversely, for some plants reporting no IPEEE related improvements, relevant improvements had already been made os part of previous seismic studies or as part of the IPE program.

3.3.2.2 Dominant Contributors to Computed Core Damage Frequencies NUREG-1407 requests that licensees report the dominant functional and/or systemic sequences leading to core damage. In most instances, the dominant CDF contributors (seismic failures, random failures, and operator errors) are also reported in IPEEEs. Table 3.3 summarizes the most significant CDF contributors reported in the SPRA IPEEEs included in this study. As this table indicates, the following contributors (listed roughly in order of decreasing frequency of occurrence in Table 3.3) have been reported to be of significance to seismic CDF resuhs:

Seismic Failures Most Freauentiv Observed: Offsite power Electrical and Block walls Physicalinteractions mntrol panels Frecuentiv Observed Surrogate element Building failures Switchgear Cable trays Fuel oil tanks Tran= formers Pumps Refueling water storage tanks (RWSTs)/

Condensate storage tanks Observed' Switchgear chatter Ice condenser AFW Pipe Support MFW Heaters Containment fans Battery racks inverters Battery chargers Accumulators Bus undervoltage relay MCCs Electncal buses Surge tanks Control rod drive Lead centers AFW storage tank Chiller Diesel generator (DG) oil cooler bolts 24 l

Random Failures Most Freauentir Observed Diesel generators Freauentiv .. .

Ohnerved Rehofvolves Audary feedwater(AFW) pumps Operator Action Errors Most Freauente .

Ohnerved Align for AFW flow Freauentiv Ohnerved Ineste coobng/rocircuisbon Ohnerved Reduce component coonng water (CCW) host loads ' Cross-tie units Shutdown

• rom remote control panel Roset relay Station biscotout (SBO) diesel procatJre For PWRs, it is not surprising that diesel generators and alignment for AFW flow are the most frequently observed rar'.1om failures and h>Jman errors, respectively. Most seismic scenarios are accompanied by a LOOP, with the associs ted dominant core damage sequences involving common-cause failure of diesel generators coupled with failure to align and initiate the steam-driven AFW pump, it is of interest to note that none of the SPRAs reported that the list of dominant contributors is

. significantly altered by the hazard curve used for seismic CDF quantification. That is, the.

dominant contributors are substantially the same regardless of which hazard curve (LLNL or EPRI) is used. Additionally, the SPPA IPEEEs revealed only isolated minor changes in the -

ranking of dominant risk contributors using different hazard curves.

3.3.2.3 - Outliers, Vulnerabilities, and Plant improvements -

Table 3.4 lists the outliers identified in the SPRA IPEEEs, and indicates whether or not the

' licensee has clcssified an outlier as a vulnerability. This table also describes the plant -

improvements, if any, ur.dertaken to reduce the impact of each outlier. .

. As a result of the SPRA IPEEEs, licensees have implemented a number of maintenance actions and minor fixes. In addition, licensees have undertaken certain mo,e significant plant changes,

based on analyses and resolution strategies implemented by the licensee. Some of the reported plant improvements reduce seismic CDF, while others have simply been undertaken to ensure proper plant maintenance and/or configuration. Licensees have generally not proposed major Jdesign-related changes in response to IPEEE findings.

- In the IPEEEs reviewed for this report, licensees have not provided a consistent definition of

" vulnerability." in most instances, licensees did not propose any definition of vulnerability in the IPEEE submittals; and licensees simply stated that no vulnerabilities were found. in one

~ instance, however, a submittal referred to a significant number of plant anomalies as being vulnerabilities, in addition, a couple of instances reflect the licensees' application of the guidelines proposed by the Nuclear Management and Resources Council (NUMARC) for use in

' vulnerability assessment.

25

3.3.3 , Containment Performance Perspectives Although not required in NUREG-1407, a few of the IPEEE reports indicates that licensees have implemented a quantitative assessment of seismic containment performance. The relevant NUREG 1407 guidance for SPRA quantitative containment performance analysis focuses on the assessment of containment failure modes significantly different from those encountered in the IPE. In some instances, the quantitative results are presented as frequencies of small and larv releases; in other cases, they are presented in the form of frequencies of small and large containment failures, in addition, some SPRA IPEEEs have also reported a containment HCLPF capacity.

- Table 3.5 presents both qualitative and quantitative findings associated with containment

- performance, it should be noted that licensees have obtained the most useful results when both quantitative and qualitative assessments were made. (Some SPRAs have reported findings of both types of assessments.) Section 3.4.3 describes additional perspectives gleaned from l_ qualitative containment performance assessments.

Notably, the containment perfonnance perspectives gleaned from SPRAs for some BWRs have been observed to be dependent on an assumption that there is a significant likelihood of recovery-in the long term (i.e., beyond 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />).' However, licensees did not attempt to link the related seismically induced equipment failures to the ability to repair the failures in the time available before containment failure occurs.

3.3.4 Implicatior:s of Different PRA Methodologies All of the SPRAs have generally followed the conventional seismic PRA methodology, as described in NUREG/CR 2300 and NUREG-1150. However, certain licensees implemented a significant variation on this methodology (i.e., using a surrogate element in SPRAs).- Table 3.2 f

indicates those IPEEEs for which licensees employed the surrogate element.

Tim basis and approach for surrogate element modeling is discussed in detail by Reed et al.

(EPRI, igg 3) and Reed and Kennedy (1994). The overall concept of the surrogate element is to account (albeit approximately) for the effects of components that are screened out during the walkdown and screening phase of a SPRA. Hence, the failure of a single surrogate element represents the potential failures of several components that might normally be excluded from a SPRA model. Use of the surrogate element helps to ensure that the SPRA does not overlook a potentially significant portion of the seismic CDF.

Use of a surrogate element represents 0 ad seismic PRA practice when screening is performed at a sufficiently high threshold, when the e.apacity of the surrogate element is assessed to be consistent with the screening threshold, and when the surrogate element is appropriately included in the plant logic model. Otherwise, the usefulness of the approach and the validity of SPRA findings are compromised.

if the surrogate element is used to represent a low screening threshold, such that few SPRA components have fragilities lower than the fragility of the surrogate element, dominant risk contributors will be masked, and the ranking of the dominant sequences may be misleading.

Conversely, if the surrogate element is applied with respect to a high screening threshold, but is not sufficiently introduced within the plant logic model, the possibility exists that a small fraction of the seismic CDF will be missed. Therefore, the use of the perspectives derived from some of 26 l

e

e -c

- tho' SPRA IPEEEs using surrogate element approach needs to be treated with care. Table 3.3 I identifies the plants where the surrogate element is listed as a dominant risk contributor. I To date, the staff has not undertaken a detailed investigation concoming the implications of using the surrogate element, or provided guidelines concoming its use (particularly with respect to -

sensitivities in plant logic modeling). However, in most circumstances, if failure of the surrogate.

1 element is modeled as leading to core damage, and if the surrogate element is found to be a

- minor contributor to selsmic CDF, its use is probably reasonable.

3.4 3elsmic Marnin Perspectives A seismic margin assessment produces both quantitative and qualitative findings. The key .

= quantitative results summarized in this report include component HCLPF values and estimates of plant HCLPF capacities; by contrast, the principal qualitative results inc,i,de walkdown findings, identification of components that control plant HCLPF capacity, and a description of outliers and plant improvements.

3.4.1 - Summary of Quantitative Findings To quantify a plant's seismic margin for the IPEEE, licensees must consider the following principal elements in the analysis:

- success path development (EPRI SMA) or plant logic analysis (NRC SMA)

+ . seismic screening and outlier identification evaluation of component HCLPF capacities e

assessment of plant HCLPF capacity Success Path Development or Plant Logic Analysis In an EPRI SMA, two attemative success paths are developed to take the plant to a stable shutdown condition and maintain that condition for at least 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />. ~ At least one of the attemative success paths must involve mitigation of the effects of a small LOCA. Tim principal functions considered when developing the succass paths include reactivity control, reactor pressure control, reactor coolant inventory control, and decay heat removal. .The product of the

-- success path development is a safe shutdown equipment list (SSEL) that identifies the components to be evaluated, as well as the random failures and operator actions that could impact the integrity of the success paths. In some isolated EPRI SMA IPEEE submittals, licensees have not entirely followed the criteria for success path development, or the applicability of the criteria themselves is somewhat unclear.

l

Seismic Screening and OutfierIdentincation i

Once a licensee has generated an equipment list, the components are evaluated against L screening criteria. Such screening typically involves a documentation review and seismic walkdown (if the component is accessible). Outliers are identified as any component that does not satisfy screening criteria or a check of anchorage capacity.

Evaluation of Component HCLPF Capr@s HCLPF capacities are calculated fu ionents identified as outliers. In an EPRI SMA, a conservative deterministic approach, t ,wn as the conservative deterministic failure margin 27

(CDFM) methodology, is typically used to assess HCLPF capacities. In an NRC SMA, licensees may employ either the CDFM methodology or a probabilistic fragility analysis.

Component HCLPF results represent a significant product of the IPEEEs. The 24 submittals reviewed for this report have reported a voluminous quantity of relevant results. in general, the staff has observed a somewhat greater level of consistency in assumptions and procedures implemented for HCLPF calculations in SMAs (in comparison to fragility calculations in SPRAs).

For instance, the majority of SMA IPEEE submMtals have used the median NUREG/CR-0098 spectral shape, as opposed to site specific spectral shapes, to characterize ground-motion input.

Assessment of Plant HCLPF Capacity in an EPRI SMA, the plant level capacity is assessed as the lowest component HCLPF capacity in the most rugged success path. By contrast, in an NRC SMA, the plant-level HCLPF capacity

] is evaluated by inspecting a core-damage Boolean expression, or is determined directly from a plant-level fragility curve.

A reduced scope evaluation is performed using input from the plant's seismic design basis (SSE spectra); hence, a reduced-scope evaluation does not convey the degree of seismic margin, as I would a full SMA. Other forms of deterministic review (full-scope or focused scope) do yield an estimate of plant HCLPF capacity.

R 3.4.1.1 Plant Seismic Capacity Results Table 3.6 lists plant level HCLPF results reported in the SMA IPEEEs included in this study. All have been performed for plants located east of the Rocky Mountains and the plant HCLPF -

capacities vaned from 0.21g to 0.50g.

All HCLPF values presented in Table 3.6 have been derived based on a NUREGICR-0098 spectral shape for rock or soil (depending on the site conditions at the plant). It is noteworthy that, for SMA IPEEEs, the plant HCLFF spectrum (seismic capacity) generally exceeds the plant SSE spectrum (seismic demand) for the given plants. Hence, the SMA studies for these plants have been effective in demonstrating a level of seismic margin.

3.4.2 Summary of Qualitative Findings Qualitative findings of an SMA are those that cannot be described with numerical results, although a quantitative analysis may often be necessary to derive the qualitative perspectives.

Forinstance, qualitative findings of an SMA relate to the choice of success paths and playsical conditions and characteristics of a plant, as well as the component (s) that control the plant's HCLPF capacity, and the list of outliers and related plant improvements developed by a licensee.

\

3.4.2.1 Walkdown Perspectives All SMA IPEEEs involved a seismic walkdown by trained or qualified personnel, in most cases, licensee personnel received SQUG training for the USl A-46 program, as well as the " add-on" IPEE : training.

With the exception of the evaluations for St. Lucie and Turkey Point, the seismic margin assessments included in this report referenced the use of EPRI NP-6041 or EPRI 6041-SL walkdown and screening procedures. The evaluations for St. Lucie and Turkey Point themselves 28

also have some (albeit limited) similarities with these procedures. For USl A-46 plants, the IPEEEs typically also referenced the use of GlP procedures and criteria.

The walkdown perspectives obtained from the SMA studies are very similar to those already documented in Section 3.3.2.1 for SPRAs. This similarity exists because NUREG 1407 recommends that licensees follow the same walkdown procedures.

3.4.2.2 Plant HCLPF Capacity, Outliers, Vulnerabilities, and Plant Improvements Table 3.7 lists the lowest HCLPF capacities for components, including all components that do not exceed the plant's review level earthquake (RLE), and those components that control the overall plant-level HCLPFs. This table illustrates that low ruggedness relays, components anchorages, sollliquefaction, interaction concems, and tank failures are some items that have controlled plant HCLPF capacities.

Table 3.7 also provides a general description of anomalies and outliers identified in SMA IPEEEs, together with a description of the plant improvements, if any, undertaken to reduce their impact. All of the SMA IPEEEs have resulted in some degree of plant improvements or actions in response to identified outliers or anomalous conditions.

3.4.3 Containment Performance Perspectives SMA IPEEEs have ger.erally implemented a qualitative assessment of containment performance involving screening and/or walkdown examination of the following items:

containment struct # integrity containment penetrations, hatches, and seals e

containment cooling systems Licensees have not reponed my anomalous conditions with respect to containment structural integrity. However, in a few instances, they have identified outliers pertaining to containment penetrations and containment cooling.

Table 3.8 summarizes the findings derived from SMA IPEEEs with respect to seismic containment performance. For most IPEEEs, licensees have not encountered any significant findings. In isolated instances, however, licensees have estimated and provided the containment HCLPF capacity, in one submittal, the licensees noted a concem with a potent llriicrfacing systems LOCA (ISLOCA) occurring inside the containment, and considered the possibility of implementing a relevant severe accident management guideline.

3.4.4 Imp!ications of Different Margin Methodologies As previously discussed, the two different SMA approaches include the NRC and EPRI methodologies. Of the SMA IPEEEs included in this report, only one used the NRC SMA approach. However, for that study, component capacities were assessed based on the CDFM methodology outlined in EPRI NP-6041 SL. Hence, the study differed notably from an EPRI SMA only in the manner of plant logic analysis. (Event trees and fault trees were constructed, rather than a set of success paths.) To assess overall plant HCLPF, the NRC SMA study evaluated the HCLPF of each core damage cuiset as the HCLPF of the strongest element in the cutset; then, the plant HCLPF was approximated by the cutset having the minimum HCLPF. The NRC SMA study used the same spectral shape as that used for an EPRI SMA study.

29

c .

3

Hence, the principalinsight to be gained by a comparison of application of the two SMAs is that,

.

n many areas, they can yield substantially similar findings. However, the plant logic modeling approach used in an NRC SMA method allows enhanced flexibility in developing perspectives

concoming risk contributors (including random and operator failures, in addition to seismic l failures). Perhaps the most significant difference in the NRC and EPRI SMA methodologies is

- the necessity of a more qualitative treatment of random and operator failures in the latter

[

L approach.

3.5 other Evaluation Perspectives l

In addition to the aforementioned aspects of SPRA and SMA findings, an IPEEE may also yield l - findings from assessments of relay chatter, soil failures, non seismic failures and human actions,

seismic-fire and seismic flood evaluations, and generic and unresolved safety issues.

Perspectives derived from IPEEE assessments of these areas are discussed in the following j- subsections.

1 3.5.1 Relay Evaluation -

l NUREG 1407 describes the recommended procedures for relay evaluation, depending on the

!' scope of the seismic evaluation and whether or not the plant is an USI A-46 plant.. Relay

evaluation for USI A-46 plants has revealed " bad actor" relays at a significant number of plants.

However, bad actor relays have been encountered in selected success paths in only a few of l these cases.

l- . -

l When bad actor relays have been encountered, they have often been found to exist only in alarm j - circuitry, they have been assessed as having neg'igible consequence, or the licensees have e

assumed that operator actions will provide for effective reset, in only isolated instances, L therefore, have licensees actually proposed replacing relays based on the analysis of IPEEE only systems. Table 3.9 summarizes the plant specific findings of IPEEE relay evaluations, and

[. describes licensee actions (if any) taken in response to these findings.

l 3.5.2. Soils Evaluation i Table 3.10 describes the site soil characteristics for the 24 plants included in this report. For j - those soil site plants identified as beyond the redaced-scope seismic category, most IPEEE 4 submittals addressed the issue of soils failure evaluation, including liquefaction potential and the

! - potential and effects of slope instability, settlements / displacements, and stresses in buried

! pipingJ A few of the IPEEE submittals referred to the modified seism;c IPEEE guidelines described in Supplement 5 to GL 88-20 and, thus, the licensees did not perform a soils t . evaluation.

.- Table 3.'10 also summarizes the findings of soil failure evaluations for the 24 iPEEE submittals 1

reviewed.1ln one case, the evaluation indicated tnat liquefaction is likely to occur below the RLE, a but above the SSE. In another case, the liquefaction potential has been identified as a potential

. concem._ Seismic slope instability at the RLE was also indicated for two plants; however, the i magnitude of slope deformations has been assesse.. As minor. In addition, the submittals generally reported that seismically induced differentia soil settlement has only minor impacts.

• Theseey r la s are " low-ruggedness" relays, as defined for USl A-46/ SQUG.

]

2 i

j 30

}

d

-*e- e ryW@- +- - - , - - - --

- The principalinsight from these evaluations is that soil failures are potential concems at some plants, but the potential for such failures may be difficult to fully rectify in a cost-effective manner.

Despite the probability that soll improvements may not be cost-effective, the IPEEE findings regarding soil failures provide useful information concoming the expected plant response to i seismically induced severe accidents for these plants.

3.5.3 - Non Seismic Failures and Human Actions All of the first 24 IPEEE submittals provid .a treatment or discussion of non-seismic failures and human actions. For SPRAs, these ei.. .

' introduced in seismic event tree and fault-tree models, which reflect the plant logics co ted for intomal events. However, among the IPEEEs reviewed, the seismic impacts on o - -

error rates were modeled in a highly variable fashion. in some instances, licensees devt  ? Jmplified operator error fragilities. In other instances, licensees applied debatable scalu,;;;i tors (in relation to the importance of the human action) on intamal event error rates or other factors. When operator error fragilities were applied, they often masked the seismic failures that dominate seismic CDF.

With regard to the treatment of human actions, toe SMA IPEEEs yielded a relatively consistent finding that licensees have made little attempt to evaluate seismic impacts on operator error rates. Also, there were only a few cases in which licensees applied screening criteria with respect to random failure rates'and human error rates. Most frequently, the SMA submittals

' simply reported an attempt to rely on those success paths that are most familiar to plant operators and that utilize the most reliable equipment.

. Table 3.11 summarizes the findings of seismic IPEEEs pertaining to non-seismic failures and human actions, as well as the treatment applied in response of those failures, in addition to the observations discussed above, this table indicates that most submittals considered oper5 tor actions required to reset relays.

3.5.4 Seismic Fire and Seismic-Flood Evaluation All of the first 24 IPEEE submittals reported that the licensees have examined seismic-fire interaction issues, including seismically initiated fires, as well as seismic actuation and '

degradation of fire suppression systems.

Licensees typically investigated seismic-fire interactions by means of plant walkdowns. In most submittals, the investigation was limited to the consideration of flammable or combustible liquids or gases as seismic-fire sources. These gases orliquids are almost always stored in nonsafety areas or exterior to the plant: Therefore, it is not surprising that most submittals reported that seismic-fire interactions are an insignificant safety concem.

The most consistent strengths of the seismic fire evaluations appear to involve the treatment of inadvertent actuation of fire suppression systems and the identification of potentialinteraction concems, in many of the IPEEE submittals, seismic-fire interaction evaluations revealed significant findings and, in a number of instances, resulted in plant improvements. Table 3.12 summarizes the findings from the IPEEE evaluations of seismic-fire interaction issues. Some of the relevant improvements include component anchorages and waterproofing, replacement of sight glass tubes, and implementation of procedures to properly secure transient fire-protection equipment.

31 1

& O 3.5.5 Generic and Unresolved Safety lasues i

Most of the seismic IPEEE submittals reviewed for this report have addressed many of the-following generic and unresolved safety issues (GSis & USIs):

• USl A-46, " Verification of Seismic Adequacy of Equipment in Operating Plants"

• USI A-40, " Seismic Design Criteria"

• USl A 17, " Seismic interactions in Nuclear Power Plants" USl A-45," Shutdown Decay Heat Removal Requirements" GI 131, " Potential Systems interactions involving the Moveble in-Core Flux Mapping System in Westinghouse Plants" Eastem U.S. seismicity issue (i.e., Charleston earthquake issue)

- In addition, the seismic IPEEE submittals provided information related to the following generic

safetyissues (GSis)

• GSI 156, " Systematic Evaluation Program (SEP)"

- GSI-172, " Multiple System Responses Program (MSRP)"

USI A-46 is being resolved separately from the seismic IPEEEs and, hence, is not extensively 1 addressed in the IPEEE submittal reviews. USls A-40 and A 17 are subsumed as part of the licensee's USI A-46 program. By contrast, as indicated in NUREG-1407, the eastem U.S.

-seismicity issue is reso'ved with the submittal of a satisfactory IPEEE. Thus, the following subsections focus on discussions related to USI A-45, Gl.131, and GSis 156 and 172.

~ 3.5.5.1 USI A-45 Whether a licensee uscs a SPRA or an SMA for the seismic IPEEE, the capability of decay heat removal (DHR) functions is modeled. Thus, any findings encountered in the IPEEE with respect -

to seismic capability of DHR functions are also applicable to USl A-45. In other words, for seismic events, USl A-45 perspectives are a subset of the IPEEE perspectives. Consequently, the IPEEE submittals have generally re-iterated those seismic IPEEE findings pertaining to DHR capability as the resolution of USl A 45. This approach is valid; however, any weaknesses in the seismic IPEEE with respect to the assessment of DHR functions will be mirrored by weaknesses -

in the treatment of USI A-45. Licensees have identified some significant seismic ou' liers related -

to DHR functions in their IPEEEs; and these outliers are listed in Tables 3.4 and 3.7.

3.5.5.2 GI-131 GI-131 applies only to Westinghouse plants that have a movable flux mapping system (see Table

- 3.13).' For most GI-131 plants, the licensees had previously addressed this issue through specific plant improvements, and the issue was considered resolved by those licensees. In some instances, the lionsees undertook a walkdown to verify the installation of a previous improvement as part of the seismic IPEEE. In addition, a few submittels reported that licensees

- evaluated the capability of the llux mapping system for RLE loads. Hardwars improvements related to Gl-131 have not been made as a result of the seismic IPEEE. However, in one case, the licensee implemented an administrative procedure to help eliminate the potential for an interaction hazard involving an overhead chain hoist.

Overall, licensees have undertaken appropriate actions to address GI-131.

32 I

3.5.5.3 GSI 156 Regarding GSI 156, the seismic IPEEE submittals of applicable plants provided relevant information conceming some or all of the following seismic-related issues:

settlement of foundations and buried equipment a

dam integrity and site flooding

• design codes, criteria, and load combinations a

seismic design of structures, systems, and components The NRC and relevant guidelines have not explicitly requested that licensees discuss the resolution of GSI 156 in their IPEEE submittals. Nevertheless, the staff has used the relevant information found in the licensees'submittals to assess the resolution of the related issui,s listed above for each plant.

3.5.5.4 GSI 172 Regarding GSi-172, the seismic IPEEE submittals provided relevant information conceming some or all of the following seismic-related issues: '

seismically inducea spatial and functional interactions e seismically induced fires seismically induced fire suppression system actuation

- seismically induced flooding a

seismically induced relay chatter evaluation of earthquake magnitudes greater than the safe shutdown earthquake effects of hydrogen line ruptures The NRC and relevant guidelines have not explicitly requested that licensees discuss the resolution of GSI-172 in their IPEEE submittals. Nevertheless, the staff has used the relevant information found in the licensees'submittals to assess the resolution of the related issues listed above for each plant.

3.6 Perspectives Renardino Seismic PRA Versus SMA An important distinction between perspectives drawn from SPRAs and those derived from SMAs pertains to the effectiveness of demonstrating the plant seismic margin. This distinction arises because NUREG-1407 recommends the use of the 1989 LLNL uniform hazard spectrum (UHS) ,

to describe the input motion characteristics for use in SPRAs, but it recommends the NUREG/CR-0098 spectrum for SMAs. Consequently, the HCLPF capacities determined using the two approaches are not directly comparable. In addition, because the UHS shape for EUS plants is generaliy less severe than the NUREGICR-0098 spectrum, SPRAs would ganerally yield higher HCLPF capacities. Interestingly, however, this has not been observ6d to be the case in the first 24 IPEEE submittals. In fact, the HCLPF spectra deterrnined in SPRAs for EUS plants have generally not exceeded even the plant's SSE spectrum over important frequency ranges.

3,7 Consistency of Perspectives Amona Plants in reviewing the first 24 IPEEE submittals, the staff observed certain similarities in the perspectives gleaned for the various p, ants. These similarities appear to be primarily attributed to the following factnrs:

33

e .

e similar plant characteristics- s

. ' same plant licensee e

same contractor / analyst .

e a combination of the preceding factors .

Interestingly, the greatest consistency in the findings has been observed where the IPEEEs involved the same plant licensee and/or the same contractor / analyst. This observation implies the predominant effect of methcdological approach and assumptions on IPEEE results.

3.8 Findinna that Raouire Further inv==6Me*!en The seismic IPEEE program has revealed certain preliminary generic perspectives, as well as new perspectives that had not been observed in previous seismic evaluations of US nuclear plants. The following paragraphs briefly discuss the relevant preliminary findings.  ;

in this report, generic findings are defined as those fmquently observed among plants, wheress -

plant-unique findings are those that are limited perh.ps to a single plant. Clearly, the seismic IPEEE results presented in this report have revealed both generic and plant-unique perspectives.

For example, one pk. unique finding involves the condition of a vulnerable socket weld in AFW piping that was encountered at one plant.~ loentification of such plant-unique anomalies has been a significant objective of the IPEEE program. However, the collective, generic findings o' IPEEEs are also important in providing perspectives about problematic conditions that rtay exist in some plants.

The results of the seismic IPEEEs included in this report indicate that the following conditions have commonly been encountered among some plants:

Seismic failures that dominate seismic CDF (e.g., electrical and control panels, block walls, interactions, buildings, switchgear, cable trays, fuel oil tanks, transformers, and Pumps)

B' ndom failures that dominate seismic CDF - (e g., diesel generators, relief valves, and A/W pumps) -

Operator failures that dominate seismic CDF (e.g., failures involving operator actions to maintain AFW flow and initiate cooling and recirculation)

Because bad actor relays have been identified in several seismic IPEEEs, the existence of bad actor relays is confirmed as a potential generic insight. However, IPEEE submittals have yielded the generic insight that chatter of vulnerable relays in selected IPEEE success paths generally do not have adverse consequences.

- Seismic fire interaction evaluations have also yielded a common finding that suppression .

- equipment (e.g., tanks, bottles, and extinguishers) must be better anchored or restrained, and that the operation of fire ; umps may be compromised because of the failure of the fuel oil supply or relay chatter effects. These findings can also be classified as potentially generic, in reviewing the first 24 IPEEE submittals, the staff also identified the following new

- perspectives that were not fully revealed from past seismic evaluation studies of U.S. commercial nuclear power plants:

34

i A notable fraction of EUS plants present a significant risk of core damage from seismic

, events, exhibiting seismic CDF values nur 1x10"/ry and higher, regardless of whether the licensees use EPRI or LLNL seismic hazard results to quantify the seismic CDF.

Even after identified plant improvements have been made, some plants have an HCLPF capacity spectrum that is less than the SSE demand over important vibration frequency l

l. ranges.

Cable trays have been identified as dominant risk contributors at a number of plants.

For some of these perspectives, it is difficult to ascertain, without more detailed review, whether they relate to actual plant weaknesses or are primarily consequences of the assumptions and approximations used in the IPEEE analyses.

35 i

Table 3,1 Seismic Review Categories and Evaluation Approaches for Plants included ir. this Study Plant Name Seismic Review Seismic IPEEE Evaluation Approach Category (Bin)

Brunswick 2 (0.3g Focused scope) Focused-scope EPRI SMA Callaway 2 (0.3g Focused scope) Focused-scope EPRI SMA Catawba 2 (0.3g Focused scope) Existing Seismic PRA Comanche Peak 1 Reduced-scope EPRI SMA Cook 2 (0.3g (Reduced scope)

Focused . Seismic scope) PRA Diablo Canyon 6 (Seismic PRA) Existing Seismic PRA Fort Calhoun 2 (0.3g Focused scope) Focused-scope NRC SMA (using surrogate element at 0.5g) -

Haddam Neck 2 (0.3g Focused scope) Seismic PRA (using surrogate niement et 0.3g)

Kewaunee 2 (0.3g Focused scope) Seismic PRA (using surrogate elemerst at 0.30)

LaSalle 2 (0.3g Focused scope) Existing Simplified Seismic PRA (SSMRP)

Limerick - 2 (0.3g Focused scope) Reduced-scope EPRI SMA McGuire 2 (0.3g Focu*ed scope) Existing Seismic PRA Millstone-3 2 (0.3g Focused scope) Existing Seismic PRA Nine Mile Point-2 2 (0.3g Focused scope) SPRA & Focused EPRI SMA (using surrogate element at 0.5g)

Palisades 2 (0.3g Focused scope) Semmic PRA (using surrogate element at 0.5g)

Pilgrim 4 (Seismic PRA) Seismic PRA (using surrogate element at 0.5g)

. Point Beach 2 (0.3g Focused scope) Seismic PRA (using surrogate element at 0.3g)

Robinson-2 3 (0.3g Full scope) Full-scope EPRI SMA St. Lucie 1 (Reduced scope) Site-specific approach

• Seabrook - 4 (Seismic PRA) Existing Seismic PRA Sequoyah 3 (0.3g Full scope) Full-scope EPRI SMA South Texas Proj. 1 (Reduced scope) Existing Seismic PRA Susquehanna 2 (0.3g Focused scope) Focused-scope EPRI SMA Turkey Point l 1 (Reduced scope) Site-specific approach

• i i

l 36

Table 3.2 Selsmic Core Dama9e Frequency and Plant Capacity Results from Selsmic PRAs Plant Name Mean Seismic CDF HCLPF' Spectral Surrogate (EPRI or Other) (LLNL) (g) Shape Element 7 Catawba 4 '

1.6 =10 - -

Sequoyah 5 No 4 '

Cook 3.2 =10 1 =104/ry' 025'8 1989 LLflL h No Diablo Canyon

• 4.2 = 10' - 0.C7'M LTSP S".e Specific k No Haddam Neck d 2.3 =10 ry 1.5=10 /ry' < 0.05' '" 1089 EPRlh Yes Kewaunee 1.1 = 10' ry 1.3 = 104/ry' 023" 1989 LLNL h

Yes LaSalle 7.6 = 10'7/ry - - Not Specified' 4 No McGwre 1.1 =10 /ry -

4 NUREG/CR 0098' No Millstone 3 025' Site Specific' 9.1 =10 fry' -

No Nine Mile Point 2 2.5=10' /ry 1.2=104/ry' O.50 (24 hr) NUREG/CR 00988 Yes

1. 023 (72 hr)

Palisades -

8 9=10 0.229 1993 LLNL h

4 Yes Pilgrim 5.8=10 /ry 9.4 = 10' 025' h 4 1989 LLNL Yes Point Beach 1.4 =10 /ty 1.3 = 10' O.16' 1 A89 LLNL" d Yes Dombrook 6.1 = 10 1.3=10 /ry - Site Specific' No 1.2 =10'[

South Texas 1.9 = 10' 22=104/ry' -

River Bend No Project Hazard Spectra

--v .,

• Unless otherwise indicated, the HCLPF capacity applies to peak gem f*d acceleration (PGA) b *r>* denotes ' reactor year" c * ' indicates that the retut* was not reported d 84th percentile site-specific spectrum for the Sequoyah plant site e Based on a site specific hazard curve developed by licensee's consultant f Infened approximate value as reported in TER. (Licensee did not report.)

g Interred estimates of lent fragibty parameters, based on a lognormal destribution of capacity, are:

A.=0 4Bg (PGA),and =027, where denotes median-median capacity, and be denotes composite logarithmic standard vlation in capa h 10,000-yr median uniform hazard spectrum (UHS) l This HCLPF value applies to SWdamped spectral acceleration averaged over the frequency range of I 3.0 to 8.5 Hz; the correspondin0 PGA HCLPF value Lt approximately 0.67g.

interred estimates of plant fregility parameters, based on a lognormal distnbution of capacity, are:

Am=1.31g (PGA), and be=029. A value of A m=3.06g is estiranted for SSdamped spectral acceleration averaged over the frequency ran0e of 3.0 to 8.5 Hz.

k Spectral shape developed from the long term sehmic program (LTSP) l Based on 1993 LLNL seismic hazard results.

m inferred plant median capacityis A =0? 6g 7,3A) n This HCLPF value hcludes the effe, cts of non-seismic failures and human errors; the HCLPF value without these effectsis 026q(PGA) o The TER indicates that this LDF value is about a factor of 2 too low, p SWdamped median spectral shape from NUREG/CR 0098 (Newmark and Hall 1978).

q This hcl.PF value includes the effects of non seismic failures and human errors,; the HCLPF value wtthout non seismk: failures, but with human orrors, is 027g (PGA) r This HCLPF value includes the effects of non-seismic failures anu human errors; the HCLPF value without these effects is 0.329 (PGA) s This HCLPF value includes W Sffects of non-seismic failures and human errors; the HCLPF value without these effectsis 025g (PGA) n \

Table 3.3 Dominant R6sk Contributors Repofted h Selsmic PRAs Catewtie Cook D6eWa Canyon Heddam Neck Salgg; Offene Power Offene Power Offene Power A; W Pye DG Denery Chorgers Audhery puleding 230 kV Treneformer Staten Mein foodwater (MFW)

DO 0a Tenha Diook Wats 4 kV Swmohseer (Chenar) Hesters AC Swuchgear 280 VDC Penets DG Control Penel Cont. Air Recirc. Fans Irworters RPS Penew Benery Bank AC and DC Penste los Condenset (Turtune Bulkhng Podsetsi)

(4 kV SwitcW (Cable itsfs)

SaDklk Donel Generators Turtune.prhen AFW Pump Deeel Generators Preaturtset telety tehof volves (SRVs)(Rectose)

Geerster; Reduce Component Cochng Wolor (CCW) Host Leeds Cross-te Unas 1 and 2 SwechCortsen SumpRecere Nowounee LaSalle McGuire M416stono4 Agigg: offene Power Offene Power offene Power offene Power Surropste Element Condeneste Storege Tank 120 VDC Doesi Generator ON Ceciers (anchor hohe)

Roof Deephroom(Control Dieg)

Weg Foohng (EDG Dide)

Sheer Wen (ESF Dide)

Pumphouse Sinhng (Soil)

RaDeus- Deee4 0enerators DeeeWorators Done10enerators AFW eystem Qantater; Swech CST to SW for AFW None Ahen SW to Pond Nine Mee Point Pebeados Pilgrtm Point gesch laimm; surrogene Elemord Dessi Fire Pmp Dey Terce Motor control contors (MCCs) Cetdo Treve Nerogen AW : . Dessi Fire Pmp Control PenelBus Sunegene Element Offene Power Stehen trenetwar Penets 4 kVTransformere Main Steam leoishon Velves CCW Pumps 400V Lead Centers (l*llys),interacten Reenduel Heat Remoni (RHR) Block Wees DJ Fuel Os Tank Purros Rue Undervotege Reley SW Pumps CCW Surge Tenha sanok Wan Certrol Red

• structurel Failures CSTs,interechon Bandagt; Dweet Generators AFW Pump Automshc Cm .

Velves (ADVs)

Ehegger indhete Ormo Through Cochng 180 Doeel Procedure Shutdown from R vnote Penol inlhete AFW Mehe4sp Reest Sto-Reisted Reley Ahon SW to AFW Suchon ArW risw Conpos Inn. Suppreewon Pool CorW 38

Table 3.3 Dominent Risk contr6butors Reported in Seismic PRAs (Continued)

Seabrook South Tease Pre $ set

$dgik; Oftees Power Offene Power 416 kV Swachgeer (chenor) Doesi Fuel 08 Day Tenho nWST 416 kV Swnch0eer theesi Generaler. Large Cheer.Tenho CCW Surge Tank AFW Storage Tank Elsetnoel CatWneto (Wworters, chargers)

Rendum ameraser: nemt relay.

39 a -

Table 3,4 Summary of Sels nic Anomalies, Outliers,VulnerabilHies, Housekeeping Concerns, and Plant improvements identified in SPRA IPEEEs 71 ant Walkdown Type (s) of Description of Plantimpb monts Name Screening Findings Findl.ics Level Catawba Structures' Outlers

• Reactor Building and Fires were made to three 2 Sg median Reactor Building ContalnmentInternal minor spebalinteracbon Eoulprned Containment Structures could not concerns, and were 2 Og mesan internal Structures be screened out. deemed not to be rak 6nomalies

• Walkdown identified significant.

Minor spahal minor spatial

• DG battery rock interacton interaction concems modAcatons concerns

• Instrument relocated

• VaNo replaced

, (Table 3 3 of IPEEE)

Cook Not repoited Several anomales Block walls, poor fire- Three des 6gn-related and housekeer vJ extingusher imaravements concems mountings; inters . bon @ support of w,.t fire protection i. .,n,unt tack, halon pilot knes;interacbon bottles, and emerge.sey of fluorescent bghts in sonnee water (ESW) controf room; piping) and 13 mesing/ broken housekeeping-te'sted fixes ancht. capes on some (rt,ilacing of tightening MCCs; queebonable nuts / bolts of clamps, rust support of a 17. ton protecton, etc )were CO2 tank; potenbalfor made.

earthquake-induced hydrazine spill Diablo Not reported None None No IPEEE plant Canyon improvements. Earber programs'

. LTSP imptovements

• DCPRA-based improvements

• Ongoing 6mprovs monts Haddam 0 3g HCt.PF,0.8g identmed Numerous condrbons, Numerous meaningful plant Neck ePectral vulneratwhbes, inclueng poor improvements have been acceleraton includinu 30 seemic anchorage / support, proposed (Table 7.1 1 of rok outhors and 8 interaction concems, IPEEE submrttel) seemic. fire rok housekeeping outhors concems, and relay chatter Kewaunee 0 3g HCLPF; 0.8g Open Mesmg fasteners on Resolution of USl A-46 spectral issues /anomahes; DG excitabon and concoms has led to some accelerabon no vulnerabilibes or control cabinets; poor equipment enhancements, rek signrhcant anchorage of station one procedural concems service transformers; implementabon, an potenbalinteracton of administrabve control, and relay racks; other several housekeeping equipment imp ovemehis (Submrttal anchorages concoms; Table 3-4) mercoid swttehos LaSalle Not reported Outher; anomales CST was found to be None, but the submittal not reported an outher notes that some plant improvements have been made since 1985 40 1

Plant Walkdown Type (s) of Description of Plant improvements Name screening Findings Findings l

Level McGuire Structures Momaht)

• Walkdown identifieJ e Spacers installed on DG 2 Sg median Minor spetal 6 spetalinteracton bettenes/ rocks Ecutoment inte' action concerns, two e Gretnginmmed near

? Og median concerns and equipment steam-vent vakes maintenance mounting / support

• MCCs bolted together concerns concoms, and one

• Guideknes developed for maintenance concern movable equipment

• Panel modifed to clear 8-l In ppe i

e Arc barners tightenedin main control boards

• Grout instehed below toddle support of CCW Hx

• Mesing bolts instened on surge tank

• Corrosion on anchor bolts of AFW/ CST cleaned and i

bolts recosted l

(Table 3 3 of IPEEE)

Millstone. Not reported Outhers; anomales Desel generator oil None for IPEEE; however, 3 not reported cooler bolts identifed desel generator oilcooler se an outher bohs were previously replaced with stronger bolts Nine Mile 0 5g HCLPF,1.2g Anomehos/open Three concoms were Rack ovwr a motor Point.2 spectral issues ened potentalfor operated veNe (MOV)was accelerabon overhead rock to secured, rail stops were impact an MOV; installed to prevent potenbalinteraction of movement of host host assembhos assembhos on electncal mounted on electnc cabinets catxnots; and fire water piping in control building Palisades 0 3g HCLPF and Outhors and Fifty two (52) None 0 Sg HCLPF Anomales condebons were encountered, including instances of poor anchorage, unqualifed (and uneneWed) block walls, and interacton coneems Pilgrim 1 Og median (EPRI Outhors and Vanous concems Strffening of SBO desel NP 6041 SL SMA anomehos were identified in mumer support fix a Column 2 USl A 46/iPEEE seemic interaction hazard screening critena) evaluebons, but not due to potenbalfailure of a fully reportedin IPEEE main transformer bushing submittal and adjacent bghtning arrestor; and fix potenbal weaknees of fncbon-chp restraints connechng A8 but to its concrete foundebon 41 L

Plant Walkdown Type (s) of Description of Plant improvements Name Screening Findings Findings Level Point 0 3g HCLPF; and Outers Vanous concerns Flx anchorage deficiencies Beach 0$gHCLPF we's 6dentrfied in USl on cable trays and some A-46 evaluation, and equipment (for USl A-46);

apply to IPEEE. resolve concems Weaknessee in associated wtth RWST and CST Westinghouse ModellTH encountered (and rolsys (for USl A-46); and modeledin SPRA add two diesel generators fragility anatyees) and their support systenis (for IPE)

Seabrook 2 Og median Anomalies Some minor None. (Fire and IPE otmervabons improvements 12ve been subsequently resolved ctted )

es setsfactory South Not reported Not reported Not reported None reported Texas Project l

1 1

l 42

. _ _ _ _.. -. _ _ _ _ _ _ _ . . _ . _ _ . _ . _ - . _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ . _ . . ~ . _ . _ _ _

Table 3.5 Summary of Seismic Containment Performance Findings %m SPRA IPEEEs Plant Name Containment Quantitative Qualitative Plant Type Findings Findings improvements Catawba Ice Condenser, None

• Reactor building, None Pressure containmentintemal Suppression Type, structures ed not i with Steel Pnmary screen out Containment and a

• Cabinets, panel Reinforced boards, and MCCs Concrete SNeld (for antainment Building isolation system)ed not screen out

• No fragihty anaWes for ke condenser Cook lee Condenser Direct / structural No signMcant None containment failure anomalies were caed leads to 1% of seesmic CDF; frequencies of earty friesses otherwee not quientrfied Diablo Large dry type, Leme earty release None No related IPEEE Canyon steel-kned treauency. Improvements reinforced- 3% of seemic CDF concrete Small omrtv release freauenev:

16% of seemic CDF Haddam Large-dry type None Vulnerabilrbes found Licensee's Neck in Adams fWter units resolution to these and CAR fans (CAR ltemsis unclear system), essel fire pump battenes and desel fire pump fuel o* tank (CS system),

and exhaust penetration P3g(CP system)

Kewaunee Large dry type; Mean freousney of No sign 6 cant None free-standing steel containment failure- anomahes were cRed containment 6.2 104/ry (EPRI vessel, surrounded hazard)

' " ' 'C Pentenment HCLPF f,arae earty failures) building, wtth an annular space in. 0 3g between the two LaSalle Mark 11, with None reported None rersoried None inerted, pnmary containment of post tensioned re:nforced concrete with steel -

kner; secondary containment is the reinforced concrete reactor buil6ng 43

Plant Name Containment Quantitative Qualitative Plant Type Findings Findings improvements McGuire ice Condenser None None (No fragiltty None anaWe was performed forice condenser)

Mllistone 3 Not desenbod Not reported Containment None rectreulatng system heat exchangers are outiers Nine Mlle Mark 11 Credrbng operator No signthcant None Point 2 actions to close valves anomebos wera cited outsado containment, less than 2% of the CDF is associated wtth early containment failure ortypass Palisados Large-dry type, Dominant sommic No signthcant None pre-stressed, post- containment failure anomales were etted tensioned mode was found to be reinforced relocaton of core debns concrete structure to the auxiliary building, hned with a 1/4 having a frequency of inch carbon steel toyer 2.3104/ry Pilgrim Mark t pressure ELtauenev of eartv No docussion None suppression type containment faIure- provided 1.610 4/ry(EPRImean hazard),

4 3 2=10 hy(1993 LLNL mean harard)

Point Beach Large-dry type, Freauenev of earht tatoe None, although the None wtth pre 4 tressed, release quantitattve anaPyse post ensioned 1.3= 10 sj ,y indicates that the rein' . ted automate cont rete structure (The submittel robes on containmentisolabon manual containment functon hh inleton to reduce the.

by a factor of 10 ) sesmic capatulity Seabrook Large-dry type None documented No signrhcant None. (However, anomabes were cited relevantIPE improvements have been cited )

South Texas Large-dry type, None None reported None, afthough Project with a steoLhned, earner (as part of the post tensioned Level 2 PSA effort) reinforced- enhancements were concrete structure made to selected contaihment isolation vakes 44

f , a Table 3.6 Plant Capacity Results from Selsmic Margin Assessments Plant Name Format of SMA HCLPF' Spectral Shape (9)

Brunswick Focusea scope EPRI SMA =0 3 NUREG/CR-0098"Medan Soil Callowey 6 Focused-scope EPRI SMA >0.3 NUREG/CR-0096 Moden,So#

Comanche Peak Reduced-scope EPRI - Plant SSE (R G. 1.60),0.12g. Rock

- 8 6 Fort Calhoun Focused-scope NRC SMA 0 25 NUREG/CR 0098 Moden, Soll Limonck Reduced-scope EPRI -c Plant SSE (Newmark),0.15g, Rock Nine Mile Point Focused-scope EPRI SMA 0.50(24 hr) NUREG/CR U006"Moden, Rock 0.23 (72 hr)

Robinson FuB-scope EPRI SMA 6 0.28' NUREG/CR-0098 Medan, Sol St Lucio Sne-erecffic -

' Plant SSE,0.10g," Structural Fui Sequoyah Fus-scope EPRI SMA 0.27 NUREG/CR-0098

Susquehanna Focused-scope EPRI SMA 0.218 NUREG/CR 0098" Medan (Rock and SoH)

Turkey Point Site-spec #fic - Plant SSE (Housner),0.15g, Rock

" Unless otherwoe anecated, the HCLPF capacity apphes to peak ground acceleraton (PGA) b SWdemped medan spectral shape from NUREG/CR-0098 (Newmark and Hal,1978) c * "ind. cates that no result was reported d Controlled by hquefaction; without hquefaction, the plant HCLPF is reported as 0.27g e Assumes plantimprovements f So#-etructure-interacbon (SSI) aneWe uses rock toectrum se exensbon, such that the minimum froo4 eld PGA at the top of the souis 0 3g g Inferred based on TER of beenue's IPEEE; Deensee reports a plant HCLPF of 0.30g 4$

o e Table 3.7 Suinmary of Selsmic Anomalies, Conuoliing Outliers, and Plant Improvements identmed in SMA IPEEEs Plant Name Walkdown Anomalies and Cont'olling HCLPF Plant Screening Outliers Cutlieis Capacity (g) Improvements Level Brunswick 0 3g HCLPF Joint USl A. Several All outhors Being made 'Jnder 46/IPEEE seemic outhers were have HCLPF USI A-46 evaluabon effort ident$ed, but capactbos resolubon; hasidentMed a none had exceeding 0.3g number of calculated housekeeping, HCLPF Note: the IPEEE maintenance, and findings assume capactbes interacbon USl A-46 less than concerns, and improvements 0.3g equipment outhers (which are stil to y ,,og Calleway 0 3g HCLPF 21 anomahen / Outhors had Allouthors

• Remounted open lasues were calculated have HCLPF hand-held identMed, some HCLPF capacibes exbnguishers outhers were capactbes exceeding 0 3g

• Trtmmed floor identMed exceeding grebng 0.3g

• MCCs bolted to walls

• Mening sheer pins insta#ed on AFW pump

• Procedures and signs for storage of transent equipment I

• Procedure for l sec'Jnng chain l hosts Comanche SSE Some minor NoSSL Not apphcable Follow-up actons Peak anomehos and outhors w,re to resolve:

maintenance idenbiled . U.1 anchored connems were notyplant identAed equipment near safety equipment in contrelroom

• Insumcient clearance i between an MCC and cable tray sUDDort _

46

Plant Name Walkdown Anomalies and Controllirig HCLPF Plant Screening Outilers Outliers Capacity (g) Improvements Level Fort Calhousa 0.3g HCLPF Several ousers Relays 001g Replacement of were identhed MCCs 005g bad actor relays; (anchor ) 010g Imtrovement of SermeBidJ 010g MCC anchorages; Fire Pumps o tog raw water system TurtWne BPJg o g70 be-in to the 0179 f sie storage WS 0.170 tank (EFWST),

CST 0.24g and others MCCs 0.25g (anchor.) 0.25g Lkluvection 0.77g Transformer 0.29g MCCs (anchor.)

RWS pump Limerick SSE (but the Some maintenance None Not Evaluated Tracking of 0 3g HCLPF and housekeeping identined housekeepmg screening anomaW wre and maintenance tables were obsers issues essentally used)

Nine Mile 0 Sg HCLPF - See Table 3 4 - Nittopen 0.23g - See Table 3.4 -

Point 2 bottloo HFA Model 0 45g 154 relay Robinson 2 0.3g hcl.PF 33 issues / MOV 0.28g Concerns for 32 anomalies related RHR 750 components were to interactons, addressed by MOV 0.28g maintenance, of RHR-751 housekooping were actons; identrhed,47 enhancements for components were 34 components identrhed as outers required reports or modencetons; 16 lesuesinvoMng electncal raceways involved maintenance of modrncatons; manyof these concerns are being resoNed under USl A-46.

Note thel?EEE assumes USl A-46 Improvements (which are stdl to be resobed) 47

0 o f

Plant Name Walkdown Anomalies and Cont olling HCLPF Plant Screening Outliers Outliers Capacity (g) Improvements Level St. Lucle SSE UnL1: None Severalsignrhcant 11 anchorage reported improvements to concerns, low having anchorages, capacity of CCW HCLPF maintenance surge tank capacityless actons; and platform; and three than SSE implementaton of intermeton a strict concerns housekeepng Policy Qnd two interaction concerns;four maintenance issues Sequoyah 03gHCLPF A design-related RHR Heat 0.27g Replacement of deficiency, four Exchangers MCC anchorages; i

anomalous upgrade of RHR condtbons, as well heat exchanger l as outhors were anchorages; and identrhed corrective changes to ehminate interactions Susquehanna 0 3g HCLPF Some anomakes interactons: Troiieys removed and maintenance from swttchgear concems were cabne Uwo HPCI pump 0.21g noted, numerous anomahes and escherp outhors were three identrhed housekeepng 0.21g concen,s are Suppression being tracked )

poolinlet valve 0.250 Automate transfer 0.26g switch MCC Turkey Point SSE 26 anchorage CST 0.11g Plant accons, hupport concems, analyses, or 12 6nteracton enhancements harardo, two RWST 0.11g were undertaken functonal to resolve all concerns, and D'ese! 08 0.21g outhers as part of some seesmic Tunk USI A-46 housekeepmg neues were Nott The capacites are idenM for the upgraded condebon 48

. c Table 3.8 Summary of Seismic Containment Performance Findings from SMA IPEEEs  !

Plant Name Containment Walkdown HCLPF Capacity IPEEE Plant ,

Type Findings / Outliers (g) Improvements '

Brunswick Mark-l No outhors or HCLPF capocrty None. (However, anomahes were against largo-earty may be effected by reported failureis atleast USl A-46 plant 0 3D improvements)

Calloway Large-dry type, No outhers or HCLPF capacity None with steel 4ned, anomshes were against earty fGifures post tensioned reported is at least 0.3g reinforced concrete Comanche Peak Large-dry type, No oumers or Not opphetble None with steel 4ned anomahes were reinforced reported concrete structure Fort Calhoun Large-dry type Not reported, a Not reported Noimprovements quantrtstrve Level-2 were made annWes was specifically to performed, address indicabng that the containment condtbonal performance probab6ltty of large-earty release, gNon seemic core damage, is about 1%

Limerick Mark-Il No oumers or Not reported, None anomahes were although all reported components essentistly screened at 0 3g HCLPF Nine Mile Point 2 Mark-Il - See Table 3.5 - A HCLPF capacity - See Table 3.5 -

against targe-earty release was not reported Robinson.2 Large-dry type, of Potentialinterfacing Reported HCLPF Potential severe prestressed systems LOCA capacity against accident concrete wtth a (ISLOCA)inside large-earty failure of management steelher containment due to atleast 0 3g guidehne to mrbgate MOV failures. the ISLOCA (Evaluabon of concem.

containmentfan coolers are 308 to be reported )

St. Lucie Steel vessel No evaluabon was No evalusbon was None currounded by a conducted conducted reinforced-concrete biologeal shield, with an annular space in-(%

between Sequoyah ice Condenser Walkdown reveakd HCLPF capacity None no anomshes or against large-early ou$ers failure of atleast 03o 49

Plant Name Containment Walkdown HCLPF Capacity IPEEE Plant Type Findinga / Outfiers (g) Imptovements Susquehanna Mark-fl None (No insuffcent None comprehensive evaluabon to walkdown of determine containment safeguards; ony pipingNaives and containment structure were considered )

Turkey Point Large-dry type, of No evaluebon was No evaluAon was None steel-Sned po'et. conducted conducted tensioned reinforced-concrete l

I 50 l

.. .~ - . . _-. - . - _ _ - _ _ . - ._ .- . - - - . - . . - - . . - - . - _ - . - . _

Table 3.9 Summary of Findings from Relay Evaluation in Selsmic IPEEEs Plant Name Treatment Low Ruggedness Safety Related Plant Relays identified implications improvements Brunswick USl A-46 relay Severblidentrhed, The IPEEE only Concerns are belng evaluebon, four in IPEC5-ony relays were found addressed under expanded to circuitry (for acceptable based USI A46 IPEEE-only containment on conuquence circurtry performance) review; others are being addressed in USl A-46 Callaway Documentabon. Some low- Roley chatter was None based evalumbon ruggedness relays determined to be to identifylow- wereidenbfed acceptable wtth ruggedness relays respect to safe and determine shutdown of the consequences of plant cha2er; spot-check of relay instaflebens Catawba Low-ruggedness One,in a diesel Modeledin seemic None evaluabon; generator PRA relay chatter and maintenance and recovery acbons lostng circult modeled in SPRA Comanche Peak None, and none None, not applicable Not applicable None required (non-USl A-46 reduced-scope plant):

Cook USl A 46 relay A number of low. Being addreewEn Licensee plans to evaluabon, ruggedness relays USl A-46 raplace low-expanded to were identfied, none ruggedness relays IPEEE-only in IPEEE only affectng safety circuttry circuitry equipment Diablo Canyon Relay evaluabonin Nonein IPEEE Modeled in wismic None LTSP PRA Fort Calhoun USl A-46 relay Sm low-ruggedness The low-capacity The om low-evaluabon, relays in diosal relays were ruggedness relays expanded to generator lock-out assessed as hmibng are being repirace as IPEEE-only circuitry were the plant HCLPF part of USl A 46 circuitry 6dentthed; no capacity to 0.01g resolu%n IPEEE-only low-ruggedness relays were found Haddam Neck US; A-46 relay Severalinstallabons Addressed in USI A. The submittal states waluabon, ofWestnghouse 46 and SPRA mcdel that relay chatteris a expanded to COM 5 relays, nok outlier to be IPEEE-only mercoid relaysin resolved, changes circultry actuabon circurtry for to abnormal fire protecbon operabng systems procedures (AOPs) have been proposed Kewaunee USl A-46 relay 12 (Wesbnghouse Being addroceedin Low-ruggedness evaluebon, SC) low-rugg@ess USl A46 reisys are to be expanded to relays identrhed, replaced or circuitry IPEEE-only none in IPEEE-onty re worked circuftry circuitry 51

Plant Name Treatment Low Ruggedness Safety Related Plant Relays identtied impl6 cations imgovements LaSalle Not documented None reported Not documented None toported Limerick Relay evaluation Five chatter pro..e The telsys were None followng EPRI NP. relays were evaluated and found 6041 SL 6denWed to be acceptable guidehnes McGuire Low-ruggedness Low ruggedness Low ruggedness None evaluabon; relays found in alarm relays effect alarm relay chatter and circuitry circultry only; other recovery actons relay chatter effects modeled in SPRA are modeledin the seesmic PRA Millstone-3 Potent >alty Not documentedin Modeled in SPRA The beensee vulnerable relays IPEEE submntal updated AOPs to were identified; report enhance recovery relay chatter and from earthquake.

recovery actons induced relay modeled in SPRA chatter Nine Mlle Point 2 Detaded relay All role)s screened Based on a relay None evalusbon at 0.5g out et 0 So, except screening and HCLPF one (HFA ModeL consequence 154)which was assessment, the determined to have beensee concludes a HCLPF capacity relay chatter will not of 0 45g bmt the plant HCLPF to be below 0 So Palisades USI A-46 relay A number of kw- Being addressed in Concems are being evaluaton, ruggedness relays USl A-46; SPRA dispostboned under e.panded to were iden%ed, none modehng of relay USl A46 IPEEE only in IPEEE-only chatteris unclear circuitry circuary Pilgrim USI A-46 relay Not specihed Being addressedin Concerns are being evaluston, use of USl A-46; SPRA addressed under relay generic modehng of relay USl A-46 equipment chatter assumes ruggedness USl A46 resolubon spectrum (GERS),

ard SPRA rewow Point Beach USI A-46 relay A number of Being addressedin Concems are being evaluabon, WestnghouseITH USl A-46; no SPRA addressed under expanded to relays modehng of relay USl A46 oartally address chatter

• IPEEE ony cireuary Robinson 2 USl A46 relay No low ruggedness All relays were found None evaluebon, releys wwre acceptable based expanded to identified effechng on capacity partaPy address the SSEL screening ansor IPEEE-only consequence cereuttrv assessment 52

Plant Name Treatment Low Ruggedness Safety Related Plant Relays identmed implications improvements St Lucie No evaluation Not reported Ocensee concluded None required for Unn 2 there were no (non-USl A-46, deletehous effects of reduced-scope chatter of bad actor plent) For Unit 1, rolsys USl A 46 evaluebon seart hed for bed actor releys, ver6ed mountnge ofreleys Seabrook A redey evaluebon Not specmed,10 Modo 6 edin the None was conducted to types oflower SPRA identfylower. capseny relays were capacty releys and identmed (submittal 6dentify operator Table 34) recovery actons Sequoyah Full rolay chetter Severallow. Consequence None evaluebon, ruggedness rolsys analyse indicates no including capocay were identMed, none effects on SSEL l

I screening and of which were consequence determined to cause assessment mattuncbon of SSEL equipment South Tomas None, and none None, not applicable Modeledin SPRA None Project required (non-USl A-46 reduced-scope plent);

however, relay chatter is modeled in SPRA Susquehanna identd.cabon of Fourlocatons of Effects of chetter None low-ruggedntSe low ruggednees deemed acceptable relays; walkdown toisys were verMcaton; 6dentM evalvebon of chatter effects Turkey Point USl A 46 Not reported Ucensee concluded None evaluebon there were no searched for bad deletenous effects of actor relays, chatter of bad actor venfied mountnge relays of relen 53

Table 3.10 Soll Characteristics and Summary of Findings of Soll Failure Evaluation in Selsmic IPEEEs Plant Name Soll/ Foundation SSIorSoll SollLlquefaction Other Soll Failure Characteristics Response Modes Analysis Stunswick 76 ft sou,50 ft of Results of design No IPEEE No IPEEE structuralM over 26 ft SSI aneWs, scaled evaluabon evaluation densesands and mo$fied for frequency shlM Callaway Some structures FLUSH finite No IPEEE Capability of buned founded on rock; element SSI aneWe evaluebon; however, pping (between others are founded on for power block the bconsee power block and structuralN or structures concludes the fdl other structures) stabuged backM ma'enals are not was determined to hewng a depth of suscepbble to exceed the RLE anywhere from 10 flto bquefacton

$4 ft over bedrock Catawba Cateaery I Structures' None (deemed No concems No concerns Rock or concrete n*9'pbie) 1 6 den W d iden W d M extending to rock Some components founded on, or buned -

in, soll Comanche Rock site Not reported No IPEEE No IPEEE Peak predominantly evalumbon required evaluabon reoulted Cook Solshe; a slope SSI marpn fa4 ors No concerns Sol pressure failure (apptox. 2.1) bounds developed idenWed found to dominate the plant site to the containment east building fragdety; No other concems idenWd Diablo Canyon Rock site; some None desenbod No concerns No concoms components founded idenWd idenWed, effects on, or buned in sol modeled in freDelfbes Fort Calhoun 65 75 ft of sandy soll Sol spnngs in Uquefacbon Solfailures are over t,edrock; lumped-mass model HCLPFa0 25g for domintled by .

structures are soloutside the bquefacton supported on ppe vicinity of Categoryi piles structures; conter>is capacRy of desel fuelolstorage tanks and raw water system pipinD Haddam Neck Predominantty a rock SSI conducted for No concems No concems she' the new DG and

, new DG and idenWed iden%ed switchgear buildings swttchgear buBdings are founded on shalow sol, desel fuel ot tanks and piping are buned Mewounee Cley-sand sol depos#ts Elaste half-spnngs Assessed as tong Screened out based to a depth of 76 fL used to modelsoa very unkke9 on high seismic behavior capacity LaSalle No information No information No informabon No informabon prowded provided provided provided

$4

q 4 I Plant Name SolVFoundation SSIor Soll Soll L6quefection Other Soll Failure Characteristics Response Modes Analysis Limerick Rock site None reported None None McGuire f&t2orv l Structures- None (deemed No concerns No concems Rock or concrete negligible) identMed identMed j fdl extending to I rock Some components founded on soittackfdl Millstone.3 So# site (specMc Not desenbod No concerne includedin fragilrty charactersbes were identmed for power evaluabons not described); beach block; solladjetent and glacialoutwash to SW pumphouse sands edjacent to SW assumed to fall, but pumphouse determined not to impair functon Nine Mile Rock site None No concerns None Point 2 identmed Palisades 150160 ft of soa New 30 nonhnear No concems Screened out sol (dense fann sands, SSI anaWes; also 6dentMed deplacements and over ve / dense fine SHAKE computer settlements sands, over P 'd tilly code used for clay and 9th . bil) ground response over shale bedrock anaWes Pilgrim 30 to 50 feet of heavity New 3D SSI No concems Soll settlements and compacted fdl aneWes identMed foundation rocking matenals above 30 to of CST were 50 feet of very dense modeledin fragNy gleciel outwash calculebone depasas undertain by bedrock Point Beach 100 ft of soa (stiff to New 3D nonnnear Assessed as bemg Sou settlements and very-stiff glacial SSI anaWes very unlikely deplacements deposits)over affectng fractured dolomite components were bedrock screened out Robinson 2 Very deep (460 ft) nod Now, mulbple SSI Some data points Other soll-related site, dense below 50 ft aneWss uomo indicated fadures, such as depth; some structures CLASSIconducted bquefacbon at embankment failure are supported on piles for fWe Class-l isolated locabon, and Weve4nduced to a depth of 50 ft, the structures which the bconnee strains in buned circulabng water intake concluded was piping, were structure e founded at acceptable. considered and 50-ft depth (However, because concluded not to be a low magnitude signMeant was usedin the lquefacton anaWis, bquefacbon remains as an issue requinng further review )

St Lucie Cateoorv i structures- Sol modeled using No aneWas required No anaWie required founded on Category-l translabonal and (reduced-scope (reduced-scope fd, underisin by rotebonalspnngs plant) plant) cemented sands and I sandy nmestones l 55

Ffont Name Soll/ Foundation SSIor Soll Soll Liquefaction Other Soll Failure Characteristics Response Modes Analysis Seabrook Sommie Cateaorvi None desenbed No concerns No concems

$tructures ident6ed ident6ed Rock or concrete flu to rock Sequoyah Rock sne New, probabiletic Assessed as haang CompacbonI predominenty; some evaluebons of low tuscepbbihty settioment and structures are founded response, including failures due to slope on shahow sol (clays SSI effects, using instabihty were and tilts over shale ebck models considered, but not bedrock) assessed as being important South Texas Very deep soll depooft SSI finne elemeit No analysis required No oneWe required Project anaWie (reduce $ scope (reduce & scope plant) plant) __

Susquehanna AA Category-l SSI modelused for No concems No concerns structures are founded ESSW pumphouse, ident$ed ident6ed wtth on rock, except the flexible-base model set 0ements, essenbet stebon used for reactor instabikty, sidng, or serwce water (ESSW) building fixe &bese, detorbon of buned pmphouse and the modelused for pipe sprey pond other structistes Turkey Point Rock arte None described No anaWe ruquired No analyse required predominantly (reduced scope (reduce & scope olent) olant) _

$6

l Table 3.11 Summary of the Treatment of Non Seismic Failures and Human Actions in Seismic IPEEEs Plant Name Treatment in Screening Criteria impacts of Ground Human Actions Systems Motion on Human Unique to Selsmic Modeling Error Probabilities Events (HEPs)

Stunswick Success path Random failures having None modoied Chatter recovery selection, location probot@y exceeding and timing of 0 001 and ecstingin a required operator s,gnihcant cutset were actions were screened 6n; no reported screening was performed for human octions, although the success patlis were chosen to maximtre and operator familiarty and redundancy Callaway Success path Random or 6i uman None modeled Not documented selecbon failures having probatetty exceeding 0.r)01,if faGute impacts munip6e trains or systems, or 0 01,if failure impacts onY a single trmin and system, were screened in Catawba IPE model Not applicable None Relay chatter recovery Comanche Peak Success path Qualitstrve screening- None None reported selecbon success paths involve dual-train systems and actions are famihar to operators Cook IPE model Not apJeable None documented None documented Diablo Canyon IPE model with Not appbcable HEPs increased Relay chatter unique seemic based on spectral recovery and others lmpactsintroduced ere.eierabon in seem6c event tree Fort Calhoun IPE model Not applia,able HEPsincreased Not reported based on spectral accelersbon Haddam Neck IPE model Not appbcable HEPsincreased for Response to seemic events; seemic failures of I

operator fingiiny upstream dams curves were dn' eloped Kewounee IPE model Not applicable HEPs increased for None documented seemic events, as based on simpirhed operator error fragilebes t.aSalk IPE model Not apphcable None documented None documented 8

$7 l

a .

Plant Name Treatment in Screening CrHerla impacts of Ground Human Actions Systems Motion on Human Unique to Seismic Modeling Etrur Probabilities Events (HEPs)

Limerick Success path No screening, success None None reported selection paths were chosen considenng redundancy and operator famlhartty McGuire IPE model Not applicable None Relay chatter recovery Millstone 3 IPE model Not oppheath None documented Relay chatter recovery Nine Mlle Point.2 IPE model None Not speemed Not specMed Polisades IPE .tiodel Not opphcable HEPs increased for None documented sesmic evens, as based on simplMed operator error fragilibes that account forlocabon and timing of netions Pilgrim IPE model Not opphcable HEPsincrew for Relay chatter solamic events, as recovery based on simplMed operator error fragiltbes that secount forlocabon of setons Point Beach IPE model, with Not appkcable HEPs increased for Some acbons unique sesmic sesmic events, as mooeled, but none effects modeledin based on simpl6ed related to relay the entry sesmic operator error chatter recovery event tree fragileos that account forlocabon of actions Robinson 2 Success path Random failures having None modo 6ed Chatter recovery selecton;locabon probatAty exceeding and tming of 0 001 and exstngin a required operator signMcant cutset were actons were screened in; no reported screening was per*ormed for human actons, etthough the success paths were chosen to maximtre and operator famliarity and redondency 8t Lucie Operetng None None modeled None considered procedures were reviewed in developing success paths 58

-_-- . _ . _ _ . - - - - _. = - - _ .=- _ - . . - _ _ _ _ - .- -. .. .._..-. - -

e e Plant Name Treatment in Screening criteria impacts of Ground Human Actions Systems Motion on Human Unique to Seismic Modeling Error Probabilities Events (HEPs)

Seabrook IPE model Not applicable HEPs for relay r7elay chatter recoveryincreased recovery for sesmic events, as based on simpirhed operator error frecilfbes Sequoyah Success path No screening, success None mooeied Relay chatter selecton paths were chosen recovery considenng redundancy and operator familiertty South Texas IPE model Not applicable Not specthed Reisy chatter Project recovery Susquehanna Success path No se'eening; failure None modeled None reported selecton probabiltbos are reported as being 1

consstent wrth

, nereensng values used in the Maine Yankee SMA; HPCI and RCIC have a high comtuned failure probability of 0 0024 per demand; manualstartng of residual heat removal service water (RHRSW) pumps is a key seton Turkey Point Operatng None None modeled None considered procedures were rowewedin developing success oaths

$9

lable 3.12 Summary of Findings of Selsmic-Fire Interaction and Selsmic Flood Evaluations in IPEEEs Plant Name Evaluation Approach Seismic-Fire Seismic Flood Related Plant Observations and Observations and improvements Outliers Outliers Brunswick Walkdown for sesmic- Potentialinteractone Potential concems Procedure to secure fire and seemic-flood invohnno fire water with overhead water CO2 cytanders when '

concerns piping, as well as lines and CST were notin use; the  !

mobile / cart mounted ultimately screened submittal also cites  !

CO, cyhnders out several past improvements made to enhance Are protecton system seemic capabiltty Callowey Walkdown for seemic. No concems identfied Relay chatter effects None fire and seemic-flood on fire pumps, and concerns sprinkler head breakage, could lead tolocr kred floodng; but it was determined not to effect SSEL equipment Catawba Walkdown for sesmic- None None None f:re and seemic-hood concems Comanche Wall.down for seemic. None None None Peak fire and seemic flood concerne Cook Walkdown for sesmic- Potential breakage of Same as for None fire and seismic-flood glass fusesin p66ot seemle nre contms knes; subsequently screened out because no poter talwas identified for spnnkler head breaks D6ablo Canyon Walkdown for sommic. None None Addressed earberin fire and seemic-flood LTSP and concerns Seemicallyinduced SystemsIntersebon ProDram (SISIP)

Fort Calhoun Walkdown for seemic. Vanous concems Low seemic Fuel od tank to be fare and seemic-flood identrhed in turtune capacity of adequately concems buil&ng;in the intake shutdown heat anchored, a sight buildng, a fuel of tank exchangers; plass tubeis to be suppfying fire water floo&ng ofjuncbon replaced, anchorage pumps has low boxes in Room 23; of storage cabinet, capacity (HCLPF of extemal Aoodng adsbonalanchor about 0 05g); due to seemic dem bolts on shutdown break heat exchangers; waterproofing of juncbon boxes; extemalfloAng addreeted by severe accident management Du6dence 60

. e a

Plant Name Evaluation Approach Seismic Fire Seismic-Flood Related Plant Observations and Observations and improvements Outliers Outliers Heddam Neck Walkdcwn for temmic. Eight vulnerabilfbes or None reported A number of esues fire and seemic flood rnk outliers were have been resolved 4

concems; SPRA identifed or proposed for modeling of flooding 4

renolubon (See due to dem failure Tatde 7.1 1 of IPEEE submmel)

, Kewaunee Walkdown for seemic. Potentaldamage to Some as for None fire and seemic-flood fire water capatnitty sommic fire concerns and spnnklers/ lines; mercoid fire pump jockey owitches and Cardor prosaure switches LaSalle None documented None None None Limerick Walkdown for sesmic- Sight glass tubes on No addfbonal None fire and seismic-flood lube oil ma'e-up concerns tanks do not have isolabon vetves; mercoid owttches in two Are protecbon systems These concerns were decided not to be sianl6 cant McGuire Walkdown for seemic- None None None fire and seemic-flood coneems Mill 6 tone 3 None documented None None None Nine Mile Walkdown for sesmic- None None None Point 2 fire and sesmichod coneems Palisades Walkdown and SPRA Hydrogen piping Seemic int 4 iced None modeling of aesmic- through turbine flooding in the fire and seemic-flood bullding is not turbine building and concoms sesmically designed snreenhouse were and paowed through 6 dent:6ed for SPRA block walls and cable modeling; circulating trays which pose a water pipe falures in ruptyre hazard there screenhouse were exet n number of later screened out unanchored flammable bquid

• *:;;, cabnets throughodthe turbine buildint ,

61

e e Plant Name Evaluation Approach Seismic-Fire seismic-Flood Related Plant Observations and Observations and improvements Outliers Outilers Pilgrim Walkoown for sesmic. Trucklockin turtune interacton potenbal None, but the fits and seemic-flood building contains between CST 105B bcensee stated that concerns hydrogen and lube 08 and cryogenic conalderabon ppirig runs, and a nitrogen storage should be given to hydrogen control tank, modeled as isolabon of staton; owltchgear leading tolose of combushble sources room *B' also CST as water foMvir- m containslengths of source for HPCI and earthquake pping which contain RCIC lube oil Point Beach Walkdown for sesmic. None RWST could fail None fire and seemic-flood and dsable RHR

, coneems pumps Robinson 2 Walkdown for seemic. Some tesues None reported None reported fire and seismic 4ood pertaining to panel concerns interacbons and poorty anchored electrical cabinets were identrhed St. Lucie Documentation revew, None None None no dscussion of a sesmic fire walkdown Seabrook Walkdown for sommic. None None None fire and seemic4ood concerns sequoyah Walkdown for seemic- Potentialof four bght Potenbalfor None fire and seemic4ood transformersin the sprinider head concems auxihary building to breakage, but not in impact SSEL relttod the vicinRy of SSEL cebles. (Tranaformers equipment were subsequently assessed as having a HCLPF capacity of 0370)

South Temas None documented None None None Project Susquehanna Walkdown for teamic- Fire pumps in non. Non-seemicely Nor.e Are and sommic4ood seemicahy designed deeigned fire water concems structure; CO2 supply system. The tank a not seemicaRy submrttal notes that anchored, Batten a the potenbal for for fire pumps do not inadvenent have spacers; actumbon of fire unenchored smal water system is low.

electncal cabinets Turkey Point Documentabon revew, None None None no docussion of a saemie4re walkderwn 62 l

I

! Table 3,13 Summary of Characteristics of Flux Mapping Systems and of Selsmic IPEEE Findings Related to 01131 Plant Name GI-131 Previous Upgrade IPEEE Findings Appincable to l Related Plant? IPEEE Plant Improvements Brunswlex No Not opphenble Not apphcable Not apphcable Callaway Yes in 1987, the hol&down None; flux mappng No addtbonal assemNy of the flux system was inacesasible improvements for mappng system was due to radioactrve IPEEE upgraded byincreasing exposure concerns; no the site and strength of anaWit for beyond bohs and plates of the design base assemNy Catawba Yes Restraints added dunng No anaWe for beyond. No addebone!

construebon dessgn bass improvements for i

IPEEE Comanche Yes None; however,in None None l Peak promousheensing spabal interaction program actrvtbes,it was determined that the flux mapping system was designed and constructed l

to preclude interactH>ns at SSE loads Cook Yes Hold-down straps 0 32 g HCLPF capacity No addtbonal attached to the top of the assessed based on improvements for cart were redesigned and walkdown and review of IPEEE mo@fied, a lower lateral the modified configurabon restraint to the flun mappng cart was installed at en elevabon Just above the seal table Diablo Canyon Yes SISIP-related No anaWe for beyond. No addtbonal i idificabons to improve design base improvements for 1 sammic structural IPEEE anegittyof theframe

_ essemNes Fort Calhoun No Not opphcab!e Not spokeable Not opphenble Haddam Neck Not erectly, None; flux mapping cartis Walkdown venfed None a nce flux already bolted to the adequacy of configurabon mappng cartis platform not movaNe Kowaunce Not directly, None, lateral reestance of Walkdown found the' a Adminstrative since flux flux mapping system has chain hoist above the flux comrolwas mappng cartis already been determined mappng cart might implemented to not movable to be adequate interact wtth the ten-path bet:er secure assemb9 chain host LaSalle No Not appheabie Not applicable Not applicable Limerick No Not apphenble Not apphenbie Not appheable McGuire Yes Prewous seemic anaWes No anaWe for beyon$ No addrbonal havein&catedinstalled design bass improvements for restraints are adenvate IPEEE 63

. __~

o .

Plant Name ' Gl.131 Previous U grade IPEEE Findings Related APPilcable to IPEEE Plant Plant? improvements Mi:Istone 3 Yes A mod 4::AJ:P.ivas None desenbod No addmonal implemented to limit improvements for relatrve deplacement IPEEE between the flux mapping equipment and the seal tatdo Nine Mile No Not apphcable Not apphcotdo Not opphcable Point.2 Palisades No Not appheatde Not opphenble Not opphenble Pilgrim No Not applicatde Not appheab6e Not applicable Point Beach Not directly, None IPEEE submrttal notes None since flux that the flux mappng mapping cart le system is idenbcalto not movable Kowaunee's, which was found to be adequate Robinson-2 Yos Four hold-down restraints Seemic revow team No addeonal were fabncated of steel determine the flux improvements for angle, welded to the cart, mappng system to be IPEEE and bolted to the adequate for RLE loads.

structure St. Lucie No Not appheatde Not appheatde Not apphenble Seabrook Yes installation of hold-down No anatyne for beyond- No addeonal bolts for the flux mappng design base improvements for cart, procedures IPEEE implemented to ensure the cart is bolted after use Sequoyah Yes Restraints have been Seismic review team No IPEEE installed on the flux determine the flux improvements mapping cart mapoing system to be adequate for RLE loads South Tetas Yes None desertbed IPEEE submrttal notes None Project that there are no vulnerabilmes or rak outhers associated with this moue Susquehanna No Not apphenble Not apphcable Not applicable Turkey Point Yes in 1980, lateral restraint No evaluation None was added to the movable support assemblyof theflux mappng system, and was evaluated as being adeounte _

64

, c>

4 FIRE IPEEE PERSPECTIVES 4.1 Overview All 24 IPEEE submittals listed 'n Table 1.1 addressed intemal fire events. In the reviews of '

IPEEE fire analyses, the staff primarily considered the submittals and releted information provided by licensees in response to questions raised by the review team. Specifically, the reviewed submittals dealt with PWRs and BWRs of different vintages, different nuclear steam supply system (NSSS) vendors, different architect engineers, and different power levels.

Table 4.1 summarizes the first 24 fire IPEEE analyses, including the plant specific CDFs reported by licensees. No fire vulnerabilities were reported in these IPEEE submittals. (The reader should note, however, that a fire vulnerability was reported at the Quad Cities plant, which was not among the first 24 fire IPEEE eubmittals reviewed for this report.) However, about half of the first 24 submittals indicated plant improvements that involved changes to existing procedures, development of new procedures, and/or plant modifications.

Overall, licensees have expended considerable cffort in performing the fire analysis portions of their IPEEEs. In many cases, licensees generated extensive databases and computerized models of plant behavior, and most licensees undertook extensive plant walkdowns. Overall, licensees have addressed the majority of the safety-related plant areas, in so doing, they have expended considerable effort in applying the selected methodologies, and have included the possibility of spurious actuation in equipment failure mode considerations. In addition, most '

iicensees have adopted aither Fire Induced Vulnerability Evaluation (FlVE) (EPRI), PRA

[NUREG/CR 2300j or a hybrid of th* two methodologies. In virtually all cases, licensees used generic fire occurrence databases, and have not updated the databases to reflect plant specific expcrience in their analyses, in general, the IPEEE submittals identified fire scenarios in varying levels of detail. The most conservative fire scenarios used in most screening analyses assumed that, given a fire in a a compartment, the entire contents of that compartment wnuld be lost. Scenario refinements include suppression of the fire before critical damage occurs, and considerabon of fires that are localized to an electrical panel, motor, or control panel with or without potential spread to adjacent combustibles, in all cases, the information collected for complance with NRC fire protection requirements (App 3ndix R to 10CFR Part 100) has been used to establish the contents of a compartment in tnrms of cables and equipment critical to plant safety, in addition, to assess the probability of core damage esulting from a fire event, licensees used the intemal events model developed for the IPE for the majority of submittals. Licensees used existing PRA models for this purpose in a few cases. Therefore, available information from the fire hazard analysis conducted for compilance with Appendix R is augmented with addit %al cable routing data to enhance the use of the IPE intemal events model.

Several plants have multiple reactors, For such plants, licensees typically conducted a focused c fire analysis of one of the units, and attempted to identify differences (if any) that may influence

'A the results for the remaining unit (s). In almost all cases, desphe the differences, licensees concluded that the results of oi.e unit also apply to the other unit (s). In some cases, multi-unit plants have shared compartments or systems and, in one such case, the licensee chose to analyze each unit separateiy.

65

e e s

4.2 Impact of the Fire IPEEE Pronram on Plant Safety Although none of'the first 24 IPEEE submittels reported any fire vulnerabilities, about one third reported that licensees have propsed or implemented fire-related plant improvements with respect to procedural enhancements, severe accident management guidelines, and hardware installation. Examples of procedural enhancements include improving fire response procedures and revising procedures to reduce transient comoustibles in critical areas. Examples of hardware improvements include instal!ing additional sprinkler heads and fire supnssion systems, rerouting critical cables, !solating the circuits of the auxiliary feedwater system from non vital switchgear, installing an electrical cross-tie between units, and providing water proof cabinets or splash guards for non-water proof cab' nots. These efforts have led to reduced residual risk and improved plant safety.

(The reader should note that a fire vulnerability was identified at the Quad Cities plant, which was not among the first 24 IPEEE submittals reviewed for this report. In that l'istance, the licensee has taken an interir7 measure, involving an altemate shutdown procedure of using an independent back up power suoply to reduce the potential fire risk, and is currently evaluating long term options to further reduce the risk. The staff is currently reviewing the Quad Cities IPEEE submittal.)

4.3 Pernsectives. Pertaininn to _Overall#ethodelemy Most submittals are based on the EPRl's FIVE method (EPRI,1992), fire PRA (i.e., NUREGCR-2300 [NRC,1983) and Kazarians et al. [1985)), or a hybrid of FIVE and fire PRA - A few licensees used other methods to deal with specific issues in the fire analysis which required

- additional Interactions with the NRC through RAls to determina inoir adequacy for meeting the IPEEE objectives.

A number of FIVE-based analyses, especially those considered as hybrid FIVE/PRA analyses, omitted or modified certain aspects of the FIVE guidelines. Common areas where licensees altered the FIVE guidelines include the analyses of fire detection and suppression timing, fire compartment interaction, manual fire fighting, plant recovery, and data input for fire growth / propagation.

- Similarty, a number of PRA-based submittals employed certain aspects of the FIVE methodology to simplify the analysis. This practice typically included aspects of the fire area screening, fire formulations and data for fire occurrence and damage probability assessment, and generic fire risk scoping study issue assessments. Fundamentally, there are many similarities between the FIVE and PRA methodologies, especially in the context of the area screening analyses FIVE prescribes the evolution for introducing details into the probabilistic model. By contrast, the PRA approach gives the analyst complete freedom. For example, FIVE prescribes that the failure probability of redundant /altemative shutdown paths must be introduced into the fire scenario frequency evaluation immediately after fire initiation frequencies are e=tablished, in the PRA approach, however, the analyst may postpone this step to a stage after gaining information regarding the relative location of cables and other critical equipment within the fire zone to ensure that a small fire cannot cause critical damage.

For fire occurrence frequencies, FIVE provides a generic data set and a method for applying and partitioning the data for specific fire zones. PRA analysts have often used data and methods that c

e

are similar to those provided in the FIVE handbook, and many have directly applied the FIVE methodology for partitioning the frequencies.

The FIVE methodology provides a set of pre formulated worksheets and look up tables for fire propagation aneiysis, in some cases, the PRA analysts have used the computer program COMPBRN (UCLA-ENG 9016 [Ho et al.,1990]) for this purpose. In the case of the hybrid methodology, licensees have often used FIVE formulations for area screening and for fire propagation analysis.

4.4 Walkdown Persoectives

- All 24 IPEEE submittals reviewed for this report have indicated that licensees conducted A walkdowns to support their fire analyses. From most submittals, the staff inferred that licensees have undertaken extensive walkdowns with the participation of licensee fire protection personnel, risk analysts, and consultants. Yhe IPEEE submittals vary considerably in the level of detail-provided in describing the walkdown scope and procedures For those submittals which provide some detail on walkdown scope and procedures, almost all closely follow the procedures recommended by FIVE, which include the following:

• - evaluatica ofignitie sources review of fire protection system locations (including evaluation of the potential for seismic-fire interactions) revi2w of fire penetration integrity and condition of fire boundaries, including doors and dampers -

raview of th:r ' ire compartment interaction analysis, which includes the position of combustl% near zone boundaries and the possibility and path for smoke propagation =

• i review to .,nsure that hydroge or other fh nmable or liquid storage vessels would not be a significant source of seismially induce i . as -

4.5 Dominant Risk Contributors '

The IPE5E submittals reviewed present the dominant risk contributors in terms of the location of -

1 the fire, and the equipment and/or systems affected by the fire. As shown in Table 4.1, the submittals often identified the control room as the most significant fire risk contributor. (The reader should note, however, that at the Quad Cities site, the licensee reported that the top fwe accident sequences, which contributed about 40% of the total fire CDF, all occurred in the turbine building and involved potential cil fires.) The dominant sequence is typically a fire in a vital control panel that leads to control room evacuation in combination with the operators failing to successfully shut down the plant using D.e remote shutdown systems. Approximately half of the submittals identified the cable spreading and high-voltage switchgear rooms as significant dominant fire risk contributors. Cable spreading rooms contain almost the same set of instrumentatiori and control (l&C) circuits as those in the control room. - Consequently, one would expect the cable spreading rooms to be identified as risk sigreificant for as many cases as the control room is ldentified as a dominant risk contributor. Differences can be attributed to the fact that several plants have a control room served by multiple cable spreading rooms that are separated by fire walls. In such cases, a cable spreading room may only contain one train of the vital circuits and, thus, the fire events in those rooms are less significant contributors. Operator 67

actions are another key element of fire scenarios associated with fires in the control room or cable spreading room. This appect is further discussed in Section 4.9 of this report.

Several submittals reported that fires in areas for service water and component cooling can also be important fire risk contributors. in addition, the submittals have identified various areas in the auxiliary / reactor building and the turbine builolng as fire risk contributors. However, a fire a" acting the turbine building v>as found to be significant in only a few cases. Most other turbine building assessments attribute the dcminant firo scenarios to a compartment or a localized area that is part of the turbine building.

Regarding dominant accident sequences, the subm.aals vary considerably in the level of information provided. Licensees generally used the intemal events model of either toe IPE or an existing PRA to determine the core damage frequency contributions of various fire scenarios.

_ Often, licensees used only a portion of the intemal events model, either because of a lack of

_ sufficient information on cable routing, or to simplify the core damage analysis. In a few cases, licensees explicitly considered the possibility of a LOCA from a fire, but most often concluded that a fire-induced LOCA is not possible. In the case of PWRs, a large majority of the submittals reported that licensees have taken credit for the possibility of feed and-bleed. Also, the staff inferred that almost all PWR IPEEEs included the possibility of a fire-induced reactor coolant pump seal LOCA.

4.6 Vulnerabilities and Plant imorovements Approximately half of the first 24 submittals reviewed indicated that licensees proposed of implemented modifications to plant equipment or procedures as a result of their IPEEE fire analyses. The following paragraphs briefly discuss the fire CDFs, fire vulnerabilities, and plant improvements.

Com Damaae Freauency All 24 IPEEE submittals used CDF as a measure to screen out, or establish the importance of various fire compartments. Table 4.1 shows that the fire-related CDFs for those submittals (not including Quad Cities) ranged between 1x10* and 2x10 per reactor-year (RY), (The Quad Cities submittal reported that the total fire CDF was about S.3x10' /RY.) Based on these results, one can condude the following:

Despite licensees' stated compliance with the NRC's fire protection requirements, fire remains a significant contributor to CDF for a number of plants.

There is significant plant-to-plant variability in terms of the reported overall fire risk, especially when the specific contributors are considered. The vintage and layout of a given plant contribute to a certain extent; however, most of this variability can be attributed to differences in methods and assumptions employed in the annlysis.

Fire is a significant core dimage contributor because the estimated risk is comparable to many intemal initating events in addition, fire may impact the operator's ability to recover from the fire event. (Section 4.9 presents a more detailed discussion conceming operator actions.)

The plant-to-plant variability with 4 i to fire-related risk can be attributed to two issues, including the inherent variation in plant layout and the variation in licensee's underlying assumptions and application of fire risk methodc!ogies. Section 4.5 addresses issues related to 68

plant layout. The issues related to underlying assumptions include considering the cable spre. ding room to be free of transient fuels, screening out cable chases that include a large number of cables from various systems / trains, and using optimistic parameter values for fire propagation or suppression modeling.

Fire VvInerabilities

] None of the first 24 IPEEE submittals identified a fire vulnerability; in ract, most did not provide a definition of

• vulnerability". Two submittals, Brunswick and Fort Calhoun, used the criterion described in NEl 91-04, " Severe Accident Issue Closure Guidelines * (1994), to define vulnerabilities. (The reader should note that the Quad Cities IPEEE identified a fire vulnerability, and the licenste used NEl's criteria to define " vulnerabilities

• in its evaluation.)

Plant Imorovements Although, none of the first 24 IPEEE submittals reported any fire vulnerabilities, about half of those submittats indicated that the licensees have made plant improvements including changes to existing procedures, development of new procedures, and plant modifications (see Table 4.1).

(As noted earlier, the licensee identified a fire vulnerability at the Quad Cities plant site.)

4.7 Cable Routina information M

Cable routing information is perhaps one of the most important elements of a fire risk analysis for a nuclear power plant. Errors in this aspect of the evaluation can severdy limit tha validity and conclusions of the entire analysis. For newer plants, this informa' ion is typically available in the form of computerized databases. For older plants, the retrieval of this information may be very time consuming. The routing of a select set (a large number) of cables is messary for fire risk analysis.

Almost invariably, the subrnitials stated that the licensees' IPEEE analyses considered the cable routing information, established as part of the fire hazard analysis conducted for compliance with the NRC's fire protection requirements. This practice is acceptable, but the following two issues must be taken into consideration:

How does the safe-shutt own model of the fire hazard analysis for compliarv with the NRC's fire protection requirements compare with the IPE intemal events i ~/

Has the licensee's IPEEE model considered the possible occurrence of an initia6ag event other than a reactor trip?

The safe shutdown model of fire hazard analysis for compliance with the NRC's fire protection requiremenis is predicated on the assumption that a reactor trip has successfully occurred. This assumption is reasonable because, given the design features of reactor protection systems for PWRs and BWRs, the possibility of an ATWS event as a direct consequence of a fire is very unlikely.

The main goal of the safe-shutdown model for compliance with the NRC's fire protection requirements is to deterministically demonstraie available paths for safe shutdown. Therefore, such a model does not address the probability of occurrence of a chain of events (especially those that may include human a n ons characterized by high failure probabilities), and the 69

resulting effect on containment-related functions. This can lead to certain differences bet veen the IPE and safe shutdown systems and components.

Very few of the first 24 IPEEE submittals articulated how the licensees identified and addressed

. the differences discussed above, it is unclear from the submittals if the licensees property and consistently assumed that those components for which cable routing information is not available are in the failed position (especially in tise worst possible failure mode). If licensees failed to make such an assumption, the core damage results can be optimistic.

A few submittals also explicitly addressed the possibility of a fire causing an initiating event other than a reactor trip. In several cases, the information provided by the licensees indirectly implied that the licensees considered an extensive list of initiating events. For example, almost all PWR submittels reported that the licensecs considered, but did not explicitly treat, the possibility of a reactor coolant pump seal LOCA. Often, the safe shutdown analysis conducted for compliance with the NRC's fire protection requirements does not include the cables associated with the initiating events addressed as part of the IPE model.

Proper modeling of the failure modes of control and instrumentation circuits is another important aspect of the fire analysis. A cable fauure under fire conditions may cause a combination of

" shorts" among various wires of a circuit. One set of shorts may lead to spurious actuation or damage of equipment in such a way that further recovery of the ft lure may not be possible. The submittals typically did not explicitly discuss how the licensees considered cable failures, and

especially spurious actuations.

Regarding compliance with the NRC's fire protection requirements, the FIVE methodology (pages 1 through 3) states that: "It is important, however, to analyze the plant 'as is'." - it further states that, for IPEEE purposes, the "as built" conditions must be considered. Although most submittals indicated that the licensees have performed some review of cable routing in addition to that '

taken from the fire hazard analysis conducted for compliance with the NRC's fire protection requirements, almost none indicated that the licensees had verified existing cable routing or other fire-related information. Also, almost none of the submittals discuss exemptions or

- deviations from the NRC's fire protection requirements (Appendix R to 10 CFR Part 50).

4.8 Threshold Value for Screenina All 24 fire IPEEE submittals reviewed included at least one screening step to reduce the number of compartments and fire scenarios requiring detailed anelysis. For this purpose, licensees used a variety of procedures, both quantitative and qualitative (e.g., screening out compartments that do not contain any important equipment and cables, as discussed in FIVE). The most common quantitative procedure involves comparison against a threshold CDF value. If the CDF of a fire scenario is above the threshold value, the corresponding scenario is subjected to further, more detailed analysis.

d value typically employed by a large number of licen s, and recommended by The FIVE, is10 threshp/ry; however, a few submittals indicated the 4 use of 10'p/ry or 10 /ry as the value. Table 4.1 lists the thrpshold4values employed by the licensees. (The effects of using a low threshold value (i.e.,10' or 10 /ry), would require a more detailed review.)

l 70

. o 4.9 - Fire Modelina Licensees have employed a variety of approaches and data sources to model fire damage, detection, and suppression. This section discusses the analytical approaches employed and weaknesses found in modeling (1) fire detection and suppression, (2) fire damage, (3) electrical cabinet fire propagation, and (4) inter-compartmental fire propagation. In reviewing the IPEEE submittels, the staff occasionally ider tified weaknesses or optimistic assumptions that would significantly impact the resLJts and conclusions of the IPEEE. In such instances, the NRC sent RAls to request that licensees assess the potential impact related to these assumptions.

4.9,1 Fire Detsetion and Suppression.

All licensees maintain fire brigades ccmposed of at least flee members, with one member from 3 the operations' staff as required by Appendix R. The brigades include personnel knowledgeabls in nuclear power plant systems and trained in fire fighting techniques. The brigades typically practics fire fighting skills in simulated fires and respond to simulated fire emergencies within the plant. Fire data indicate that, in most situations,- fire brigades successfully extinguish or control fires within 30 minutes or less (Lambright et al.,1994).

Fire models (either FIVE look up tables or COMPBRN) typically predict damage to critical components ln 10 minutes or less, or not at all. Therefore, in many cases, fire damage will occur before the fire brigade arrives at the scene.

Many of the IPEEE submittals did not credit manual suppression of fires (except in the case of control room fires). While this is understandable under circumstances where the brigade cannot -

respond before safety related equipment experiences critical damage, the degree of conservatism is unknown, On the other hand, when licensees neglect to model fire brigade.

actions, they may not be able to gain insights related to the brigade's potential to unintentionally damage equipment (in the course of manual suppression activities) that would not otherwisa be damaged by a fire. For those submittals that assumed that fires would damage all components within the area, secondary damage attributable to the fire brigade's suppression activities is already implicitly included. However, modeling fire brigade actions may be optimistic for some -

cases.

It is important to note that, in most IPEEEs, licensees used a simple model for fire suppression :

(automatic and manual actions combined). Specifically, the fire occurrence frequency is

- multiplied by the failure probability of the suppression system, and the failure probability is often gleaned from either FIVE or other sources. However, as a conservative first step, licensees assume that upon fire occurrence, the entire compartment is affected and all of the cables and equipment within are failed. The multiplication by a suppression failure probability implies the

- additional assumption that no critical damage may occur if fire suppression is successful. .This -

may represent an optimistic assumption without knowledge of the layout of cables and equipment in the compartment. For example, this multiplication process may be optimistic if a critical set of cables and equipment are in close proximity to one another, and on top of a likely ignition source.

In addition, several licensees have used generic values for supprest ion system reliability without providing a basis for their selection in terms of their appression system design features and adherence to the standards of the National Fire Protection Association (NFPA).

7:

1

4.9.2 : Fire Damage Modeling Fire damage modeling is one aspect of the fire analysis that determines the severity of a fire risk.

Most submittals utilized FIVE look up tables or the COMBPRN code to model fire damage from fixed and transient combustible sources. A few licensees conservatively assumed that fTe would damage all components within the given compartment where the fire originated and, therefore, did not model fire propagation.- Both the FIVE look up tables and COMBPRN code must be used with caution; otherwise, physically unrealistic fire damage scenarios may result. This is -

especially true with regard to fires that might spread beyond the ignition source.

In review;ng the first 3 MEEE submittals, the staff found that some licensees used a variety of optimistic, or otherwise inappropriate, assumptions:

• - inedequate consideration of all transient combustible and fixed combustible fire sources (e.g., cable qualifications have been used as the basis for screening compartments) e physically unrealistic fire damage modeling (e.g., event trees used for fire propagation and suppression modeling that underestimate the risk) e temperature damage thresholds and heat loss factors (e.g., a value of 0.85 has been used in place of 0.7)

These factors may lead to inappropriate screening of fire areas, underestimation of CDF  ;

contributions from fire areas that survive screening, and (in some cases) overestimation o' the i risk significance of non-critical fire areas (thus masking those areas most critical to fire risk).

4.g.3 : Electrical Cabinet Fire Propagation Fires originating in electrical cabinets are potentially important to fire-related risk because of the co-location of these cabinets with electrical cables. For some panels, the contents of the -

cabinets may be important by themselves.- Plant-specific details of electrical cabinets are potentially very important to risk, yet (as discussed in Section 4.3) the IPEEE fire submittals were very short on reporting such details.

Two important examples exist in the NRC sponsored fire-related PRA literature. At one plant, penetrations where cables exited the top of the switchgear cabinets were not adequately sealed, thereby allowing an exit patnway for the chimney effect. This situation allowed the fire to be postulated as propagating up and out of the switchgear cabinet, leading the licensee to conclude that a switchgear fire was capable of damaging overhead cables. At another plant, control cables in the control room were arrayed across the top of the cabinets with open tops. Again, this led the analysts to postulate the propagation of cabinet fires to the overhead cables. j Notwithstanding the prior evaluations and operating experience, at least four fire IPEEE submittals excluded electrical cabinets as credible fire sources. Similarty, some submittals used optimistically low heat release rates in modeling electrical cabinet fires, and some submittals failed to adequately consider inter-cabinet propagation of electrical cabinet fires.

- ControlRoom Fire Modelina '

Most submittals used well astablished and validated methods for control room fire modeling, which resulted in CDF estimatas typical of what have been reported in past fire PRAs. However, 72

a few IPEEE submittals reported very low control room fire CDFs. In those instances, low CDFs or screening of the control room resulted from the licensees' use of optimistic assumptions. For example, licensees assumed a 15-minute period for requiring control room abandonment, thereby postulating that smoke forced abandonment of the control room would occur only if multiple cabinet fires were involved. This use of optimistic assumptions regarding the time available to suppress control room fires before the smoke and heat would iorce abandonment of the control room led 6 of the 24 submittals to screen out the control room as a potential fire risk contributor.

ElectricalIndonendence of Remote ShtMan Locations from ControlRooms Almost all of the fire IPEEEs checked for electrical indapendence between the control room (or '

sometimes cable spreading room) and remete shutdown panels, but did not report checking for.

subtle interactions (as identified in the Fire Risk Scoping Study). Therefore, some IPEEEs may have either explicitly or implicitly (by the lack of modeling of any such interaction) assumed that ntrol room or cable spreading no roomsuch fire CDFsinteractions repotted in the exist. This submittals is potentially sie often above 10' optimistic since p/ry, assuming the room / remote shutdown interactions. If such interactions exist and occur with high likelihood during a control room fire, the cit,d CDF contributions could rise markedly, and become potentially dominant.

In addition, at certain plants, licensees' emergency safe-shutdown procedures require their operators to deliberately put the plant in a self induced station blackout in the event of a fire in order to preclude the occurrence of spurious operation of equipment resulting from hot si. orts.

However, none of these licensees reported the results oiiheir assessments with regard to the risk impact associated with this shutdown procedure in these submittalsi  ;

4.g.4 -Inter-Companmental Fire Propagation

+

inter-compartmental fim propagation has the potential to damage cables and equipment of

. multiple safety trains. The IPEEE submittals reviewed have treated this aspect of fire anaiysis with varying degrees of detail and sophistication, in the majority of cases, the submittals did not reveal much information. Many submittals quoted the Fire Compartment Interaction Analysis (FCIA) of FIVE as the methodology' employed for inter-compartmental fire analysis, and did not provUe any additional information. Only a few submittals provided tabulations of compartment groups that may be affected by the same fire.

An important aspect of this issue is the method by which licensees define compartments. Aside from the upper floers of a typical Reactor Building in a BWR, a large majority of the -

compartments in nuclear power plants are defined by fire barriers that are rated to contain the '

effects of a fire for 1 to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />. However, practically all of the compartments have passages to other compartments via such elements as doors and heating, ventilation, and air conditioning

- (HVAC) ducts.- An important propagation mechar: Ism for the effects of a fire is the escape of a

- hot gas layer through these openings.

Practically all of the IPEEE submittals reviewed ignored the following potential fire propagation scenarios:

failure of an active fire barrier (e.g., self-closing doors and fire damoers) to close, and propagation of hot gases to adjacent compartments 73

, m -

failure of fire barrier integrity as a result of fire-fighting activities (e.g., opening of doors to ,

gain access or route hoses)-

failure of a fire barrier from being overwhelmed by an excessive source (e.g., diesel fuel tank fire, or the walls separating the tarbins building from the rest of the plant being overwhelmed by a lcrge fire in the turbine hall)

In cartain cases, the possibility exists for fire-fighting activities to breach the integnty of the fire bariers. For example, if trains A and B of a safety system are located in two adjacent rooms that are connected by a door, the possibility exists for the fire brigade personnel to pass through the adjacent room in order to enter the room where the fire is, if the door is le*t ope, and the fire continues to bum, it is possible for the effects of the fire to propagate through the open door.

The IPEEE submittals pmvided no evidence initially that licensees have considered such details - '

in establishing possible fire propagation scenarios. 4 4.10 Containment Performance Per=:::t'=

All IPEEE submittals reviewed for this report explicitly addressed the impact of fire on containment performance, in all cases, the licensees used the same model developed for the IPE, and did not identify any failure modes unique to fire occurrence. However, the percentages of total fire CDF associated with different containment failure modes are different from those obtained from intomal events analysis.

All IPEEE submittals include a screening phase where compartments are screened on the basis of CDF. 4

used 10'for this purpose, most submittals used the value of 10 /ry, although a few submittals -

/ry. Since the screening process does not consider containment-related sequences, the potential exists that licensees might inadvertently omit release scenarios that are significant contributors to a major release, even though they have a frequency less than 104 /ry,,

in the case of PWRs, a!most all submittals addressed the possibility of interfacing systems LOCA (ISLOCA). In all cases, the licensees concluded that ISLOCA is not possible because of the design features of the valves (e.g., some valves cannot open against a large pressure differential), because the manual valves are normally closed, or because power circuits are disabled. "

None of the reviewed submittals revealed containment failure modes unique to fires. However, if

' the IPE or intomal events model is truncated on the basis of event sequence frequencies, it is possible for fire to lead to unusual failure combinations that the intamal events analysis would predict as extremely unlikely. None of the reviewed 24 IPEEE submittals reported such failure

- combinations.

Fire inside a containment can only occur in a PWR; however, because of the inert atmosphere, occurrence of a fire inside a BWR containment is practically impossible, in addition, reactor coolant pump fires have occurred at several PWRs; however, all PWR submittals screened containment fires as risk-insignificant. This conclusion is valid for those plants that do not have

- active components within the containment that are needed for safe shutdown. However, this is not the case for some plants, in which case dismissal of the containment could lead to optimistic conclusions.

74

___ _______-.m__._--_-.-

. 4.11 Husn Action Peramentives Human errors, typically those involving recovery actions, showed up as being important in fire IPEEE studies. This finding is in be expected; however, most of the fire IPEEEs provided very little detail to explain how the licensees treated human performance.

,s For many of the fire IPEEE analyses, licensees used humsn reliability models and data for intamal events without accounting for the unique aspects of fires. For recovery actions that take place in the area where the fire occurs, or require operators to pass through the area where the fire occurs, it is inappropriate to use the IPE recovery probabilities directly without providing justification. In a few submittals, licensees used IPE human error probabilities,- and provided sufficient justification for not modifying those probabilities.-

Licen as must exercise extreme caution in using human reliability models intended for intamal events in Essessing extemal event initiated scenariod, such as f,res This is because such -

extemally initiated events are subject to performance shaping factors (PSFs) that are not present in the intemal event scenarios. Many of the fire IPEEEs indicate that control room fires are

'mportant contributors to fire-induced CDFs, yet only two of the fire IPEEEs reviewed to date mentioned increasing human error probabilities to account for the degrading effects of fire and smoke on human reliability. (One of these submittels failed to consider the possible deleterious effects of fire and smoke on an operator's ability to move around the plant in responding to the l- fire.)

The review suggests that licensees' assessments of human performance in fire scenarios must i

generally be regarded as a weak area. The lack of details regarding human performance modeling in the submittals precludes significant reliance on human performance perspectives gleaned from the fire IPEEEs.

4.12 Generic issues and Unresolved Safety issues 4

The following subsections _ discuss the fire-related generic and unresolved safety issues concoming the safety significance of inadvertent and ady"C nt actuation of fire protection systems (FPSs) (GI-57), the fire-related vulnerabilities with respect to the adequacy of decay heat removal (USl A-45), and those issues identified in the Fire Risk Scoping Study (NUREG/CR-50S8 [Lambright et al.,1989]). Almost all of the reviewed submittels followed the guidance offered by FIVE to treat the Fire Risk Scoping Study issues.

4.12.1 GI-57 i Gi-57, " Effects of Fire Protection System Actuation on Safety-Related Equipment," addresses the Issue of inadvertent FPS actuation, as well as actuation in the presence of a fire with the potential for fire suppressant agent damage to co-located safety-related equipment. The GI-57 study identified tne following potentially risk-significant perspectives:

Suppressant agent diversion from critical plant fire areas can result from a seismic event.

Mercoid or other seismically weak slave relay chatter in a seismic event in automatic FPS controllers can lead to either failure or interlocked tripping of important safety-related components.

75 l

e, e f

For automatic FPSs actuated soloty by r,moke detectors, inadvertent actuation can occur

cause of either dust raised in a seismic event or because of smoke or steam (from a -

broken pipe) propagation from anothar t're area.

Inadvertent actuation of diesel generator room CO2 FPSs can fall the co-located diesel

- generator (s) as a result of oxygen starvation.

Most of the submitte'3 reported that the licensees performed walkdowns to address certain Gl.

57 issues (e.g., to ensure that FPSs would not fall on safety-related components and to assess the potential for spray from water based FPSs failing safety equipment). Almost all submittals i stated either that the assessment of safety-related component damage attributable to the inadvertent actuation of CO, or Halon FPSs was beyond the scope of the ana;ysis or was l

- considered to be insignificant to safety. '

4.12.2 USl A-45 >

USI A-45, " Shutdown Decay Heat Removal Requirements," has the objective of determining whether or not the decay heat removal function at operating pets is adequate, and if related cost beneficial improvements could be identified. Most fire IPEEEs have included a risk assessment yielding a CDF for fire-induced loss of DHR scenarios. Other fire IPEEEs have

' explicitly reviewed the IPE models to determine the effect of fire on DHR unavailability, Most submittals considered the following factors:

the compartments that contain DHR systems and their support systems e

the potential for firo-induced damage of DHR systems in these compartments the scenarios in each of the identified compartments that include DHR systems .

. the fraction of fire-induced CDF associated with loss of DHR equipment, or a similar measure of the importance of the effect of fire on DHR equipment 4.12.3 Fire Risk Scoping Study lasues The Fire Risk Scoping Study identified six fire-related risk issues which had not previously been addressed in a fire risk context:

. control systems interractions e

seismic-fire interactions e

manual fire fighting effectiveness (including smoke control)

.. total environment suipment survival (including the spurious operation of FPSs) e adequacy of fire barrier qualification methods e

adequacy of analytical tools for fire modeling.

4.12.3.1 Control System Interactions Control system interactions identified in the Fire Risk Scoping Study have also been classifed as GSI-147. Licensees' typical responses to this issue stated that circuit isolation, remote location, and procedures preclude control system interactions. Alsc, most licensees stated that they considered the potential for " hot short" LOCAs. However, most submittals provided little information concoming the verification process to ensure that control system interactions are not 7

a safety concem. Moreover, a few submittals failed to consider the possibility of spurious actuation of components.

76

. e

,4.12.3.2 Smoke Control and Manual Fire Fighting Effectiveness / Total Environment Equipment Survival The impact of smoke control on manual fire-fighting effectiveness, one issue identified in the Fire Risk Scoping Study, is also classified as GSI 148. That issue specifically addresses the concem that the potential buildup of smoke will hamper the manual efforts of the fire brigade to promptiy and effectively suppress fires. Most submittals provided little information concoming the potGcl for smoke to hinder manual fire fighting effectiveness or misdirect suppression efforts.

- Moreover, the pool of empirical data is extremely limited with respect to other smoke related concems (such as potential damage of electronic equipment, or degradation from secondary, non thermal fire environmental effects), nonetheless, a few licensees did consider these concems.

4.12.3.3 Seismic-Fire Interactions Section 3.5.4 of this report presented a detailed discussion of seismic-fire interactions. That discussion applies to both seismic and fire-related perspectives gleaned from t'le IPEEEs.

4.12.3,4 Barrier Reliability '

Many IPEEE fire risk investigations involved a plant walkdown to inspect passive fire barriers, in all cases, these barriers were assumed to be 100% reliable. Active fire barriers were also as. ..wd to be 100% reliable in most submittals. Those submittals tha'. considered active fire barrier failures stated that the resulting spread of fire to other plant areas is an insignificant risk contributor.-

4.12.3.5 Analytical Tools for Fire Propagation Analysis l

All submittals considered the adequacy of analytical tools for fire modeling to be beyond the scope of the analysis.-

4.13 Consistency of Perspectives Amona Similar Plants it is difficult to draw meaningful comparisons of perspectives among similar plants, given the differences in methodological assumptions and the level of reporting detail. However, the staff expected that differences should also exist based on factors such as plant age, NSSS fype, and the architect-engineer who designed and constructed the plant.

Despite the difficulties mentioned above, the staff crew the following perspectives from the 24 IPEEE submittals reviewed fo; this report:

For most plants, the critbal fire areas include 15e control room, cable spreading room, and electrical switchgear room (s).

Even though the ranking of criticald areas is simili.r, the reported fire CDFs span a wide range from about 10* to 2.2x10 per reactor year.

- None of the submittals reported that licensees have found the Fire Risi: Dcoping Study issues to be important risk contributors.

77

c. ,

• None of the submittals reported that licensees have found multi-compartment fire scenarios to be important risk contributors.

. Almost all submittals reported that operator actions are critical to the reduction of fire risk. 1 4

4 4

4 1

78

Table 4.1 Plant-Specific Core Damage Frequencies Attributable to Fire Events r

Threshold Core i

for Damage Vulnera. Plant Significant Fire Plant Methodology Screening Frequency bilities improvemenM Areas Brunswick FIVE + PRA 3.4=10* None To be determined Controlfocm and with cable spreadng for CDF>scenap/ry 10 toom Causway FIVE 89 = 10* None- None Control room and CDj/ry 10 two ESF switchgear rooms Catawba PRA CDj 4.7=10* None None Control room, cable 10 hy spreadng room, and component cooling room Comanche PRA + FIVE 2.1 = 10* None None Controlroom Peak Cook PRA 3.8 = 10* None None Control room, dAs7 CDJ 10 /ry generater rooms, ESW sys'am rooms,4kV ewitchgear rooms, en MCC toom, a battery room, a general area within the auxiliary buildng, and an area within the turbine buAding -

Diablo - PRA 7 2.7 = 10* None None Control room and Canyon cable sproedng room Ft Calhoun PRA CCpP 2.7 = 10* None Procedural - Controiroom and 10- modencebons to east basement of reduce possibihty of the auxiliary interfacing system buRdng, turbine CDJ LOCA and buildng, and an 10 hy implementabon of electrical

• Severe Accident penetrabon area.

Management Guidelines

• i 79

e s i

I Threshold Core for Damage Vulnera. Plant Significant Fire Plant 1 P[egodology Screening Frequency bilities improvements Areas Heddam PRA+OlW.. 61=10* None Develop Neck CD{hy f6 procedures for connecting air cooled desel gener6 tor, for deshng with loss of DC power, for reducing transient combustibles.

InstaA addtbonal spnnMet heads.

Improve fire procedures.

Roroute cables for .>

a charging pump or auxiliary tube oH pump Kewaunee FIVE + PRA g.8=10* None None Auxihary feedwater CD{/ry 10 pump rooms, cable spreading room, I and diesel gelierator

. . . . , . . . . room La Salle RMIEP f None None Main control room, auxiliary equipment roem, and essential switchgear areas Umenck FIVE <10-6 None Reduce transient - 12kV switchgear combustdes. room, stabc improve fire barrier converter room, and procedures. remote shutdown room McGuire PRA #

Dagege 2.3 = 10 None None Control room, cable 10 hy spreading room, vitalinstrumentabon and control area, i and auxihary shutdown panel Mestone-3 PRA Safe 4.g = 10* None None Charging pump and Shutdown component cookng -

Condebons pump area, cable spreading room, and control tocam Nine Mile FIVE + PRA 1=10* None None CD,{hy Control room Point Unit 2 10 d

Paksades FIVE + PRA 7 2=10 None May upgrade fire Main control room, protecton program, cable spreadng and depending on room, turbine the results, may building, spent fuel requentifyIPEEE pool equipment roo.n, and aux.

Bido El 590' I so

Threshold Core for Damage Vulnera- Plant Significant Fire Plant Methodology Screet.ing Frequency bilities improvements Areas Pegnm FlVE + PRA v F 2.2 = 10* None None Main control room, 10 /ty switchgear rooms, vtta! MG set rsom, RBCCW /TBCCW pump rooms, turbine buildng heater boy, and main transformer Pt Beach FIVE 5.1 = 10* None Two addesonal Control room, cable 6esel generators, spreadng room, audary feedwater auxibery feedwater system pump room, gas independence from turbine room, vital non-vital and notwital evwtchgear, and switchgear rooms, controlroom deselGenerator evacuabon rooms, and monitor procedure tank room Rotanoon FIVE + PRA 2.2 = 10* None implement' Severe CgF Control room, 10 /ry Accident bettery room, Management emergency Guidehnes' switchgoer room, and a yard transformer, St Luoe, FlVE 1.9 = 10" None CQF Power cross-be Control room, cable Unni 10 try between units, a spreadng room, roll-up door should and a switchgoer be kept closed, room analyze methods for rsducing core damage frequency from firesin the B switchgear roorn St Lucie, FIVE 1.2 = 10* None CgF Control room, cable Unit 2 10 /ry spreadng room, and a switchgoer room Seabrock PRA CDF 1.2 = 10* None Procedural Controt room, improvements, auxiliary building, expension of a switchgoer rooms, water suppression and turbine buildng system, fire detector in addbonal areas Sequoyah FIVE CF 1.6 = 10* None None 10'g/ry Audary buildng HVAC room, essential raw cooling water (ERCW),125V battery room,6.9kV switchgear room, and Nrbine buildnD 81

Threshold Core for Damage Vulnera- Plant Significant Fire Plant Methodology Screening Frequency bliities improvements Areas South PRA CD 5.1 = 10 None None Controt room

! Texus 2x10'y/ry Projoet Susque- PRA c10* None Splash guero ca a hanna few electncal cabinets, provnions for draining water from cable spreading room Turkey FIVE 10*/ry <2=10** None For the cc' ale Control room, cable l Point Units spreading room, spreading room, )

3 and 4 waterproof cabinets and intake coole:g and inetas dry pepe, water etnn,ture reachon euceression system

• Inferred from the submittal, since the licensee has not provkled a total core damage frequency I

r I2

+

5- ' HFO iPTEE PERSPECTIVES 5.1 1 Qgnadgg This section summarizes the key findings from the first 24 IPEEE submittals, with regard to high winds, floods, and other extemal events (HFOs) involving accidents related to transportation or nearby facil6 ties. It also summarizes the significant review findings de. ;mented in ERl's technical evaluation reports.

. Technical guidance for conducting HFO IPEEEs is documented in Section 5 of NUREG-1407. In particular, NUREG 1407 recommends a progressive screening approach to identify potential HFO related vulnerabilities at US nuclear power plants. This progressive screening approach can be summarized as follows?

e-Demonstrate compliance with the NRC's 1975 Standard Review Plan (SRP).

1 e~

If the 1975 SRP criteria are not met, one or more of the following optional steps should be taken:

Determine if the hazard frequency of the original design is apbly low by demonstrating that the hazard frequency is less than 1.0=10' per year.

Perform a bounding analysis _ This analysis is intended to show that the hazard 4

would not result in CDF above the screening criterion of 1.0=10 per year.

Perform a PRA.

- All 24 IPEEE submittals listed in Table 1.1 address the possibility of HFO occurrences at the '

given V. ants in reviewing the HFO IPEEEs, the staff primarily considered the licensees' submittals and responses to RAls. In some cases, the staff also consulted other documentation, such as a plant's final safety analysis report (FSAR) or a previous PRA.

Table 5.1 provides a summary of the IPEEE HFO analyses. Typically, licensees have screened

"- out the HFO events either by assessing conformance with the 1975 SRP criteria or by using unding

. analyses, quantitative analysis.

the CDF from high winds For those 4plants anj tomadoes where varies from 4 the licensees perbrmed 5 7x10 to 3.7=17 PRAs or b CDF from analysis CDF to flooding be 8.0=10 varies frg/ry from' lightning 4 and to be 6 7=10 /ry from snow and addition, all HFO evaluations reviewed to date screened out accidents involving transportation and nearby facilities. Almost all of the first 24 IPEEE submittels included a walkdown for HFO events, although submittals did not provide detailed descriptions of the walkdown prucedures and results. .Section 5.4 of this report summarizes the resulting HFO lPEEE perspectives.

5.2 ' Imoact of the HFO IP888 Proaram on Plant Safety The HFO IPEEE program has had some impact on improving plant safety. Although most plants

- have not reported significant enhancements associated with HFO events, three licensees have proposed or implemented plant improvements with respect to procedural enhancements, severe accident managsment guidelines, and hardware installation. Procedural enhancements related

- to HFO events include sandbagging, closing or welding doors, hooking up pumps, and creating new electrical circuits to reduce the risk from flooding. Two submittals reported that the a3

> e l

1 licensees are considering the development of severe accident management guidance to reduce the risk of high winds. - Hardware improvements include (among others) modifications to enhance  ;

flood protection at entry pathways, and equipment (such as port ble water pumps) to enhance l flood protection. In addition. 9 f.w submittals noted that hardware changes undertaken in l 1 -'

response to their IPE malyses (for example, the addition of diesel generators) also reduced or eliminated the risk from HFO events. Moreover, for certain plants, the IPEEE program has i resulted in a greater appreciation of the potential risk impact of high winds /tomadoes and '

extemal flooding.

5.3 Hlah Winds i

5.3.1 Quantitative Perspectives ad, seven reported a CDF fmrn high winds and tomadoes ranging Of the from 24 sugmittals 5,7x10 to 3.7x10' revip/ry. In each case, the licensees generally assumed that hu tomadoes, or tomado-generated missiles lead to a loss of offsite power (LOOP) and, in all but one submittal, LOOP was assumed to be irrecoverable.

Typically, the dominant sequence involved a LOOP in combination with random failure of emergency power. Other random failure modes reported to lead to core damage include loss of service water, auxiliary feedwater, feed-and bleed cooling, and high-pressure injection, in addition, one plant reported that wind-generated missiles failed the diesel generators, service i

-water system condensate storage tank, or ventilation system, thereby leading to core damage.

At another plant, the fire water system, domineralized water system, turbine driven auxiliary feed water pump, main steam lines, main feedwater lines, atmospheric relief valves, diesel generator exhaust, turbine building, instrument air system, feedwater/ condensate systems, station service water traveling screens and screen wash pumps, and control room were found to be vulnerable ~

to wind-generated missiles, while the switchyard, domineralized water system,~ firewster system, and turbine building were vulnerable to straight winds. At another plant, the diesel fuel oil transfor pumps and lines were identified as being exposed to tomado-induced missiles.

'n a few of the submittals, the staff noted that the licensees used optimistic assumptions and responses to related RAls led to reappraisal of the CDF. For example, in one submittal, the licensee's treatment of direct winds initially screened out wind speeds from 108 to 125 miles per hour. However, upon subsequent analysis, the licensee found that such wind speeds lead to 4

station blackout, and increase the plant's CDF from 2.0x10 to 8.0x104/ry.

- 5.3.2 - Qualitative Perspectives For plants that did not perform a PRA of high-wind risk, approximately half reported compliance with the 1975 SRP. Most of the remaining plants performed quantitative screening with a -

4 majority reported that bounding CDF estimates were at about 1.0x10 /ry or less. In addition, two submittels reported that the licensees employed qualitative screening to address their high wind analyses.

5.4 External Floods f 5.4.1 Quantitative Perspectives reviewed, four submittals reported CDFs from extemal flooding ranging from Of the to 1.0x10 g42.1x10 submittalg/ry. Floods induced by dam breaks, hurricanes, or intense precipi as

to a LOOP, which t' i licensees assumed to be irrecoverable. However, one licensee reported that, given a LOOP, additional random failures might lead to core damage. The other submittats listed additional flood-related damage, including the loss of intake structure, diesel generator building, auxiliary building, turbine building, and diesel fuel oil transfer pumps.

5.4.2 - Qualitative PerspecWes For plants that did not perform a PRA of extemal f;ooding, approximately half performed quantitative screening while almost all of the rest reported compliance with the 1975 SRP.- One submittal employed qualitative screening to address the extemal flooding analysis. At certain plants, the etaff observed that, even though flood hazards were screened out, a flood level just a few inches (or less) below the failure-incipient level might have an annual rate of occurrence one to two orders of magnitude greater than the hatard for the failure incipient level Given the large uncertainties in site specific flood hazard curves, licensees may have been premature in their screening in those cases. In addition, many submittals considered and screened out potential failures of upstream dams, leading to flooding at the site.

A few licensees propcsed flood-related countermeasures, which the staff considered as highly optimistic. For example, one licensee took credit for sandbagging up to a level of g feet.

- Another submittal credited all equipment below grade in the turbine building and auxiliary building as being capable of submerged operation, 5.5 Accidents involvina Trananortation or Nearbv Fae!MM None of the submittals reviewed to date reported a CDF above the 1.0=104/ry screening threshold from accidents involving transportation or nearby facilities. Slightly more than half of the subriittels reported that the licensees performed quantitative screening or PRA to address such accidents in their risk analysis. Almost all other licensees reported compliance with the 1975 ERP, and the submittals typically did not report walkdown procedures, team composition, or resul'.s.

[ 5.8 Other HFO Eventa 5.6.1 Quantitative Perspectives 4

One submittal reported a CDF contribution of 8.0=10 /ry from lightning and 6.7x104 /ry from snow and ice. Lightning was assumed to cause a LOOP with other random failures required to result in core damage. In the ice and snow analysis, the licensee found that the screen house, service building, and primary auxiliary building did not have roof load capacities much more than the snow load with a 100-year retum interval. Critical equipment failures attributable to roof-collapse with other random failures were required to lead to core damage.

5.6.2 - Qualitative Perspectives NUREG-1407 does not require any explicit ev iluation of HFO events other than high winds, extemal flooding, and accidents involving transportation or nearby facilities. Consequently, most submittals did not report an analysis of "other" HFO events. For the few that did, most screened these events using standardized and recognized screening techniques, in submittals that reported such 4 risk results, almost all "other" HFO events were found to have CDFs much less than 1.0x10 /ry.

85

a e

- 5.7 -- Walkdown Pers7dthti Most of the submittals reviewed to date did not provide information regarding either walkdown findings or team composition.; Two licensees did not conduct a walkdown at all, in many other cases, !icensees employed walkdowns to confirm that no significant changes had occurred since

- the operating license was issued. A few of the submittels provided some details regarding -

walkdown findingsi Specifically, the walkdown findings noted (in regard to flooding events)-

include identification of two conduits for flood entry into critical structures, and the discovery that

~ loads on the roof of a spent fuel cooling pool from roof ponding could potentially exceed the -

roof's design load.

- other susceptible components and structures were found to include the emergency feedwater service tank (EFWST), emergency switchgear room, condensate storage tank (CST), diesel-driven fire pump, fire water system, domineralized water system, turt> ins-driven auxiliary .

- feedwater pump, main steam lines, main feedwater lines, atmospheric relief valves, diesel generator exhaust, turbine building, instrument air system, feedwater/ condensate system, station service water traveling screens and screen wash pumps, control room, and diesel fuel oil transfer pumps.

5.8 Outliers and Plant Imorovements Although most submittals did not identify any HFO-related vulnerabilities, the IPEEEs have resulted in improvements in the form of procedural enhancements, severe accident management

_ guidelines, and enhanced hardware installation. Section 5.2 of this report generally describes -

such improvements, while Table 5.1 lists plant specific HFO related improvements.

8.9 - Containment Performance Persoectives None of the sutsmittals reviewed to date identified any HFO-related perspectives regarding '

containment performance.

P10 Human Action Perspectives Some of the submittals documented operator recovery actions to mitigate the effects of HFO-induced plant transients. in those instances, the important operator recovery actions included -

recovery of offsite power or diesel generators given a tomado er high-wind-induced LOOP, and -

use of sandbagging or installation of stop logs to mitigate an extemal flood.

5.11 Generic issues and Unresolved Safety issues Most submittele provided some discussion concoming generic and unresolved safety issues -

including USl A-45 and GI-103, " Probable Maximum Precipitation (PMP)." In all cases, the licensees concluded that USI A-45 is closed out. With the exception of one submittal, licensees also found that the existing plant design can accommodate roof ponding as a result of intense '

local precipitation (GI 103). For that one submittal, roof ponding only affected the spent fuel pool. Therefore, this issue (Gi-103) was considered closed out in all submittals.

The HFO IPEEE submittals did not explicitly discuss other GSis. Nonetheless, the staff considered that certain information provided in the submittals is considered relevant for

- addressing issues associated with GSI-156, " Systematic Evaluation Program (SEP)," and G31-172, " Multiple System Responses Program (MSRP)" as follows:

s6

GSI-156 (SEP issues)

Dam Integrity and Site Flooding. When applicable to the plant, HFO IPEEE submittals generally discussed the potential for and effects of site flooding as a

' result of independent or combined failures of upstream dams. One submittal also considered the potential for loss of cooling water caused by failure of an onsite dam.  ;

Site streiiAgy and Ability to Withstand Floods. HFO IPEEE submittals generally provided discussion directly relevant to this issue in their assessment of floods.

=

Industrial Hazards. HFO IPEEE submittals generally provided discussion directly relevant to this issue in their assessment of accidents involving transportation or nearby facilities.

Tomado Missiles. HFO IPEEE submittals generally provided discussion directly relevant to this issue in their assessment of high winds /tomadoes.

Severs Weather Effects on Strue.tures. In general, HFO IPEEEs screened out the effects of direct winds and flooding on plant structures. Nonetheless, where applicable, the submittals generally provided relevant information concoming the effects of wind-induced missiles on those structures.

GSI-172 (MSRP issue)

[ -

Effects of Flooding and/or Moisture Intrusion on Nons,afety Related and Satety-i

- Related Equipment. Wdh respect to safety-related equipment, HFO IPEEE submittels generally provided discussion directly relevant to this issue in their -

assessment of floods. However, the submittals generally did not discuss flooding or moisture intrusion effects on nonsafety-related equipment.

5,12 Generic Perspectives CDFs frog tomado-Whigh winds and extemal flooding have been found As shown to range in Table from 5.7x10' 5.)to 2.1=10 /ry Conservative bounding ganalysis

. estimate

' winds and extemal floods have been repoded, and associated risks have been identifed as a concem for six plants (see Table 5.1). By contrast, all evaluations screened out accidents involving transportation and nearby facilities.

s7

___._._........,-_-.._....._..,.._.-...T

Table 5.1 Methodologies and Results Associated with HFO IPEEEs Core Demage Plant - Methodology Frequency Walkdown Results Plant improvements Brunswick Hoh vnnds and, Hoh winds- None reported Hoh winds-thgl5:PRA 4.0 =104 Development of severe Others- accident management Quantitative Floodo: guidance for hi9h winds screening 1.5 =104 is being considered.

Callaway bil: Not applicable None reported None Conformance with the 1975 SRP -

Catawba Hmh winds Tornadoes- None reported None PRA 2.6 =104 Floods-Quantitative screening Others-Conformance with the 1975 SRP and quantitative screening. y i Comanche Hoh winds Tomadoes- Hmh winds- None Peak PRA 3.7 = 10* The fire water system, domineralized water

. Floods and system, turbine 4 riven g 91 TEE- auxiliary feedwater Quantitative pump, main steam screening lines, main feedwater lines, atmospheric relief valves, diesel generator exhaust, turbine building, instrument air system, feedwater/ condensate systems, station service water traveling screens and screen wash pumps, and control room are

, vulnerable to wind-generated missiles; the switchyard, domineralized water system, fire water system, and turbine building are vulnerable to high winds.

88

l l

Core Damage

1. tant Methodology Frequency Walkdown Results Plant improvements _

Cook 6 11: Not Applicable None report 6d None Quantitative screening Diablo Canyon M: PRA M: < t 0* None reported None e Foit Calhoun Moods: PRA RamjFtah: Hioh windy: Dam break:

7.01104 Offsite poweris Modify procedures and Hiah winds and susceptible to any stage four portable others' Periode tornado intensity; the water pumps. Two i Quantitative floodina: EFWSTis susceptible conduits, one into the i screening 3.0 =104 to tomado minelles intake structure and the  !

entering skylights; otherinto the auxiliary emergency switchgear building,were plugged.

room is susceptible to failure from EFWST failure;the CST and diesel-driven fire pump are susceptible to tornado missiles.

Floods Two conduits for flood water entry were discovered.

l Haddam Neck Hioh w nds: Hioh winds: None reported; l

Hoh winds PRA 5.7 x1 D* however, many Prior tomado PRA vulnerabilities are identified the need for Floodina and Floods: noted. an air-cooled diesel others 5.0x104 generator;make Bounding attangements with fuel analysis Llahtnino- oil supplier to deliver 8.0 x1 D* additional fuelwithin 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.

Snow and ice: Floodina-6.7=1D* Ensure that flood door can be installed in eight hours-- procedure and sufficientinspections.

Snow and leg:

Generate snow andice removal procedure.

Kewounee klEQt Not applicable None reported None Quantitative screening LaSalle Hioh winds and Tomadoes: 4 No walkdown was Information not Others 3.7x104 performed available PRA bounding analysis Aircraft crash:

Floods- 5.7 x104 Qualitative screening 89

e .

Cor? Damage Plant Methodology Frequency Walkdown Results Plant improvements Umerick M: Not applicable None reported None Conformance with the 1975 SRP McGuire Hiah winds: Tomadoes: No walkdown was None PRA 1.9 =104 performed I Floods and Pillts:

Quantitative screening  ;

l Millstone Hiah winds and others Qualitative screening Not applicable information not available Information not available Eloods conformance I Nine Mile with the 1975 SRP M: Not applicable None reported None Pt. #2 Conformance with the 1975 SRP Palisades M: Not applicable Fiooding: during a None Quantitative storm the spent-fuel screening cooling pool roof may exceed the design load.

Pilgrim Floods- Not applicable None reported None Conformance with the 1975 SRP Hiah winds and others Quantitative screening Point Beach Hiah winds: list!LELrls[g Hiah winds: Plantimprovement of PRA <104 Winds exceeding 240 adding two new diesel mph fall diesel generators (as Eloods and generator exhaust motivated by the IPE grg: stacks which fail two findings)is cited as Quantitative diesel generators, having a beneficial screening l effect on HFOs.

90

Core Damage Plant Methodology Frequency Walkdown Results Plant improvements Robbson Hioh winds: Hioh winds: Hoh winds- Hioh winds:

Bounding 8.0 = 10' Diesel fuel oiltransfer Development of severe analysis pumps / lines identified accident management Floods. add as being susceptible, guidance for high winds others: (Reported in response is being considered.

Quentstative to RAls.)

screening Plani-SDecific hazard Lake Robinson Dam Seabrook M: Not applicable None reported None Conformance with the 1975 SRP Sequoyah Hah winds Not applicable None repcrted Nono Quantitative '

screening Floods and others' Conformance with the 1975 SRP l

South Texas - Hoh winds: Floods: None reported None Project Conformance 2,1 x104 with the 1975 SRP Elenda: PRA Others-Qualttative screening

~ '~

~~

St. Lucie M: Not applicable None reported None Conformance with the 1975 SRP busquehanna M: Not applicable Nons reported None Conformance with the 1975 i SRP Turkey Point Hioh winds and Not applicable None reported Hiah winds:

flggda Reinforcement of the Quantitative screening Unit 1 and 2 (fossil plant) stacks, enhe'1 cement of the Others " Natural Emergencies

• Conformance procedure, with the 1975 Elenna:

SRP Refurbishment of the flood well.

91

-8 REFERENCES EPRI TR 100370, " Fire-induced Vulnerability Evaluation (FIVE)", Revision 1, Electric Power Research Institute (EPRI), September 1993.

ERl/NRC 97 501, " Review of the EPRI Fire PRA implementation Guide," Final Report, Energy .

Research, Inc., Rockville, MD, August 1997.'

Kazariant., M., N.O. Siu, and G. Apostolakis, " Risk Analysis for Naclear Power Plants:

Methodological Developments and Applications," Risk Analysis, Vol. 5 No.1, March 1985.

~

Lebright, et al., 'A Review of Fire PRA Requentihcation Studies Reported in NSAC/181,"

(unpublished letter report available in the NRC Public Document Room), Sandia National-Laboratories, April 1994.'

NEl 9104, " Severs Accident issue Closure Guidelines," Revision 1, Nuclear Energy institute, December 1994.

- NSAC/181, " Fire PRA Requentification Studies," W, Parkinson, et al., Electric Power Research Institute, Nuclear Safety Analysis Center, March 1993.

l NUREG 1407, " Procedural and Submittal Guidance for the Individual Plant Examination of .

Extemal Events (IPEEE) for Severe Accident Vulnerabilities, Final Report," J. Chen, et al., U.S.

Nuclear Regulatory Commission, June 1991.

NUREG/CR 8143, SAND 93-2440, " Evaluation of Potential Severe Accidents During Low Power -

and Shutdown Operations at Grand Gulf, Unit 1 (Analysis of Core Damage Frequency from

- Intemal Fire Events for Plant Operational State 5 During a Refueling Outage)," Vol. 3, J.

Lambright, et al., Sandia National Laboratories, July 1994.

J

- NUREG/CR-5789, SAND 01-1534, "Evalurtion of Generic issue 57: Effects of Fire Protection

- System Actuation on Safety Related Equipment," J. Lambright, et al., Sandia National Laboratories, December 1992.

NUREG/CR 5088, SAND 88-0177, " Fire Risk Scoping Study: Investigation of Nuclear Power Plant Fire Risk, including Previoush Unaddressed issues," J. Lambright, et al., Sandia National Laboratories, January 1989; NUREGICR-5037, " Fire Environment Determination in the LaSalle Nuclear Power Plant Control Room," J. Usher and J. Boccio, Brookhaven National Laboratory, October 1987, NUREG/CR-4840, SAND 88 3102, " Procedures for the Extemal Event Core Damage Frequency Analyses for NUREG-1150," M. Bohn and J. Lambright, Sandia National Laboratories, November 1990, NUREG/CR-4550, SAND 88-2084, " Analysis of Core Damage Frequency: Peach Bottom Unit 2 Extemal Events," Vol. 4. Rev.1, Part 3, J. Lambright, et al., Sandia National Laboratories, December 1990.

NUREGICR 4550, SAND 86-2084, " Analysis of Core Damage Frequency: Surry Unit 1 Extemal Events," Vol. 3, Rev.1, Part 3, M. Bohn, et al., Sandia National Laboratories, December 1990.

92

' NUREG/CR-4527, "An Experimental Investi9ation of Intemally Ignited Fires in Nuclear Power  ;

Plant Cabinets, Part 11 Room Eff. ts Tests," J. Chavez, and 8. Nowlen, October 1988.

j NUREG/CR 2300, 'PRA Procedures Guide," American Nucisar Society, fr'stitute of Electrical and l

, Electronic Engineers, and U.S. f4ucioar Regulatory Commission, January 1983.

NUREG/CR 1278, SAND 80 0200, " Handbook of Human Reliabist Analysis with Emphasis on .

Nuclear Power Plant Applications, Final Report," A. Swaln and H. Guttmann, Sandia National j

Laboratories, August 1983 (errata noted in A. Swain, Accident Sequence Evaluation Program Human Reliability Analysis Procedure, Sandia National Laboratories, NUREG/CR-4772, l

8AND891998, February 1987).

4 i

SAND 891788, ' Analysis of Cors Damage Frequency Due to Fire at the Savannah River K-Reactor,' J. Lambright, et al., Sandia National Laboratories, January 1991, i

! SAND 891i47,

• Analysis of Core Damage Frequency Due to Extemal Events at the DOE N. '

l- Reactor,' J. Lambright, et al. Sandia National Laboratories, November 1990.

{

UCLA ENG 9018, " COMP 8RN lile: An interactive Computer Code for Fire Risk Analysis, ,

Vincent Ho,8. Chien, and G. Apostolakin, EPRI, UCLA School of Engineering and Applied '

Science, October 1990.  !

j USNRC, Generic Letter 88 20, Supplement 5,

• individual Plant Examination of Extemal Events {

..(IPEEE) for Severs Accident Vulnerabilities," September 8,1995.

i USNRC, Generic Letter 88 20, Supplement 4, " Individual Plant Examination of Extemal Events ,

(IPEEE) for Sevem / :*$ent Vulnerabilities,10 CFR 50.54(f)," June 28,1991, 1 USNRC, Generic Letter 88 20, Supplement 1, " initiation of the individual Plant Examination for Severs Accident Vulnerabilities,10 CFR 50.54(f)," Augost 29,1989.

USNRC, Generic Letter 88 20,'" individual Plant Examination for Severs Accident Vulnerabilities, 1 10 CFR 50 54(f)," November 23,1988. '

- USNRC, " Policy Statement on Severs Reactor Accidents," FedereI Megister, Vol. 50, p. 32138, August 8,1985, t

s i i

h i

4 93 P

,,,,v.-..g ..--.,.,my.~,...,_,,.7m%,. - , . , ,v..n,,m. ., , , . , , , , . . -~ --.c.,%,- .-....,n._,