ML20206B249

Seismic Margin Review of the Maine Yankee Atomic Power Station.Volume 1.Summary Report

ML20206B249
Person / Time
Site:	Maine Yankee
Issue date:	03/31/1987
From:	Cummings G, Robert Murray, Prassinos P LAWRENCE LIVERMORE NATIONAL LABORATORY
To:	NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
References
CON-FIN-A-0461, CON-FIN-A-461 NUREG-CR-4826, NUREG-CR-4826-V01, NUREG-CR-4826-V1, UCID-20948, NUDOCS 8704090044
Download: ML20206B249 (190)

Text

,  !

v:

l NUREG/CR-4826 UCID-20948 l Vol.1 l l

i Seismic Margin Review of the Maine Yankee Atomic Power Station l

l Summary Report i

Prepared by P. G. Prassinos, R. C. Murray, G. E. Cummings Lawrence Livermore National Laboratory Prepared for l U.S. Nuclear Regulatory j Commission l

8704090044 870331 POR. ROOCl( OSo0030T p PbR.

NOTICE This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or assumes any legal liability of re-sponsibility for any third party's use, or the results of such use, of any information, apparatus, product or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights.

NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited in N RC publications will be available from one of the following sources:

1. The NRC Public Document Room,1717 H Street, N.W.

' Washington, DC 20655

2. The Superintendent of Documents, U.S. Government Printing Of fice, Post Office Box 37082, Washington, DC 20013 7082

3. The National Technical Information Service, Springfield, VA 22161

, Although the listing that follows represents the majority of dos aments cited in NRC publications, it is not intended to be exhaustive.

Referenced documents available for inspection and copying for a fee from the NRC Public Docu-ment Room include NRC correspondence and internal NRC memorands; NRC Office of Inspection and Enforcement bulletins, circulars, information notices, inspection and investigation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers; and applicant and licensee documents and correspondence.

The following documents in the NUREG series are available for purchase from the GPO Sales Program: formal NRC staff and contractor reports, NRCsponsored conference proceedings, and NRC booklets and brochures. Also available are Regulatory Guides, NRC regulations in the Code of Federal Regulations, and Nuclear Regulatory Commission Issuances.

Documents available from the National Technical Information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nucteer Regulatory Commission.

Documents available from public and special technical libraries include all open literature items, such as books, joumal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.

Documer:ts such as theses, dissertations, foreign reports and translations, and non-NRC conference proceedings are available for purchase from the organization sponsoring the publication cited.

Single copies of NRC draft reports are available free,to the extent of supply, upon written request to the Division of Technical Information and Document Control, U.S. Nuclear Regulatory Com-mission, Washington, DC 20555.

Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, and are available there for reference use by the public. Codes and standards are usually copyrighted and may be purchased from the originating orjanization or, if they are American National Standards, from the American National Standards Institute,1430 Broadway, New York, NY 10018.

-s- -i

- . -- y. - --- -- u- * --e-- - - e 1m-

NUREG/CR-4826 UCID-20948 Vol.1 Seismic Margin Review of the Maine Yankee Atomic Power Station Summary Report Manuscript Completed: March 1967 1

Date Published: March 1987 Prepared by P. G. Prassinos, R. C. Murray, G. E. Cummings Lawrence Livermore National Laboratory 7000 East Avenue Livermore, CA 94550 Prepared for Division of Engineering Safety Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20666 )

NHC FIN A0461 1

0

l 1

ABSTRACT This Sumary Report is the first of three volumes for the Seismic Margin Review of the Maine Yankee Atomic Power Station. Volume 2 is the Systems Analysis of the first trial seismic margin review. Volume 3 documents the results of the fragility screening for the review. The three volumes demonstrate how the seismic margin review guidance (NUREG/CR-4482) of the Nuclear Regulatory Comission (NRC) Seismic Design Margins Program can be applied.

The overall objectives of the trial review are to assess the seismic margins of a particular pressurized water reactor, and to test the adequacy of this review approach, quantification techniques, and guidelines for performing'the review. Results from the trial review will be used to revise the seismic margin methodology and guidelines so that the NRC and industry can readily apply them to assess the inherent quantitative seismic capacity of nuclear power plants.

- iii - j 1

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

7 i

TABLE OF CONTENTS VOLUME-1.

SUMMARY

REPORT P*9' ABST RA CT ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 1 LIST OF TABLES......................................................... xi LIST OF FIGURES........................................................xii P RE FA CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x i i i EXECUTIVE

SUMMARY

...................................................... xv-

1. I NT RODU CT I ON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1 1.1 Organization of the Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4 j 1.2 Scope of the Seismic Margin Review of Maine Yankee...........1-4 1

2. SE ISMIC MARGI NS APPR0ACH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1

2.1 Sy s t ems Scr een i n g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2 J 2.2 Fragility Screening.......................................... 2-4

. 2.3 Approach for Performing Seismic Margin Reviews............... 2-6

!. 2.4 Trial Guidelines for Seismic Margin Reviews.................. 2-7 ,

! 3. RESULTS OF SEISMIC MARGIN REVIEW. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1 4

2 3.1 O ver al l Res ul t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2 3.2 Sy s t em s R e s u l t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 -13 3.3 Frag il ity Res ul ts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17 .

4. INS I GHTS AND LESSONS LE ARNED. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1 I

4.1 Approa c h an d Method ology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1 1 4.2 G u i d e l i n e s . . . . . . . . . . . . . . . . . . . . . . . ; . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 -7 4

4.3 Effort to Perform the Review................................. 4-15 4.4 Findings and Their Resolution................................ 4-16 1

4j 5. CONCLUSIONS....................................................... 5-1 j 5.1 Conclusions Regarding the Methodology, Guidel in es , and I ns ights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3 ,

5.2 Conclusion Regarding the Maine Yankee Licensing Issue........ 5-4 4 6. RECOMMENDATIONS TO IMPROVE METHODOLOGY............................ 6-1

)

! REFERENCES .........................................................R-1 APPENDICES 1

A Expert Panel Correspondence...................................A-1 B Peer Revi ew Group Corres pondence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .B-1 C U t i l i ty C ommen t s . . . . . . . . . . . . . . . . . . . . . . . .~ . . . . . . . . . . . . . . . . . . . . . . C-1 i

-v-. .

I 2 e

, ..=, y ,,. - -.,-,p.m - . -.,-,.___ -_s.___,_.r ,,.m _

,n,__ _,m, y. , _. ,. , ,..m.m. .--y. . _ _ . , .....--,.c-_.-- . - - . -

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

-F 1

L i

.-TABLE OF CONTENTS h VOLUME 2. SYSTEMS ANALYSIS P_ag

!- ABSTRACT.............................................................. 111

- LIST OF TABLES.........................................................vii i LIST OF FIGURES........................................................ 1x

1. INTRODUCTION ..................................................... 1-1 l-I '2. METHODOLOGY ..................................................... 2-1 i 2.1 Gen er a l Appr oach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1

. 2.2 Step 2 - Ini tial Systems Rev1ew. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1  :

2.2.1 Gather System Information............................ 2-1 2.2.2 Classify Front-Line Systems.......................... 2-1 2.2.3 . Identify Group A Front-Line Components............... 2-2 l

j 2.2.4 Cl as s i fy Su ppor t Sys tems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2 l 2.2.5 Identify Group A Support System Components........... 2-2 ,

~

2.2.6 Identify Plan t-Unique Featur es . . . . . . . . . . . . . . . . . . . . . . . 2-3 5

2.2.7 Prepare for First Plant Walkdown..................... 2-3 2.3 . Step 4 - Firs t Plant Wa1 kdown. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3 l 2.3.1 Peer Rev iew Group Meeti ng . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3 l 2.3.2 P l an t Wa l kd own . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 -3

! 2.3.3 Plant Staff Discussions.............................. 2-4 1

2.4 Step 5 - Systems Modeling........'............................ 2-4 .

2.4.1 Review Event Trees................................... 2-4 2.4.2 Develop Fa ul t Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5 2.4.3 Devel op D at a Ba s e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7 t

2.4.4 Determin e Sys tem Cut Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7 2.5 Step 6 - Second Plant Wa1 kdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8 1 2.6 Step 7 - Determine Minimal Cut Sets . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8 4

2.6.1 Finalize Event Trees................................. 2-8 4 2.6.2 Fin al i ze Fa ul t Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 ! 2.6.3 Link Fault Trees..................................... 2-8 I 2.6.4 Determine Preliminary Boolean - Equation. . . . . . . . . . . . . . . 2-9 j 2.6.5 Determine Final Boolean Equation..................... 2-9

3. S Y S TE MS A NAL Y S I S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 1 i 3.1 Sys tem Identi f icati on. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1 3.1.1 Group A Systems...................................... 3-1 3.1.2 Systems Removed from Group A......................... 3-1 3.2 Sys tems An alys i s of Front-Lin e Sys tems . . . . . . . . . . . . . . . . . . . . . . . 3-3 l 3.2.1 . Auxi l iary Feedwat er Sys tem. . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3

+

3.2.2 High Pressure Safety Injection System.. .. . .. .. ... ... . 3-5

, 3.2.3 Power-Operated Relief Valves j (Feed and B1eed)..................................... 3-8 .

3.3 Sys tems Analysi s of Support Systems. . . . . . . . . . . . . . . . . . . . . . . . . . 3-9  !

3.3.1 Component Cooling Water Systems.....;................ 3-9 I - vi.-

f i

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

1 I l TABLE OF CONTENTS (Cont.)

VOLUME 2. SYSTEMS ANALYSIS 4

P_ age 3.3.2 Serv ice Water Sy st em. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11 '

2 3.3.3 Electric Power System................................ 3-13 '

3.3.4 Actuation Systems.................................... 3 i 3.4 Probability Calculations..................................... 3-16 a- - 4. ACCIDENT SEQUENCE ASSESSMENT ~ AND RESULTS. . . . . . . . . . . . . . . . . . . . . . . . . . 4-1 Trees.................................................. 4 i 4.1 Event i 4.1.1 No LOCA Case......................................... 4-1 d

4.1.2 Small LOCA Case...................................... 4-2 J

4.2 Core Damage Sequences........................................ 4-3 4.2.1 T LPt ................................................ 4-3 l 4.2.2 TLD.................................................4-3 j 4.2.3 S D.................................................. 4-4 1 4.2.4 S LP 2 ***********************************************'4~4 '

4.2.5 S LD................................................. 4-4 ,

4.3 Boolean Equations for No LOCA and LOCA Cases................. 4-4 4.3.1 No LOCA Case......................................... 4-5 4.3.2 Small LOCA Case...................................... 4-6 4.4 Plant-Level Boolean Equation................................. 4-6 i

5. ENGINEERING AND METHODOLOGY INSIGHTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1 5.1 Engineering and Operational Insights......................... 5-1 i 5.2 Insights on the Methodology and Execution.................... 5-2 5.2.1 System Calssification and Screening Guidelines....... 5-2 5.2.2 Preparation for Walkdowns and Documentation.......... 5-4 5.2.3 Systems Analysis and Pruning Proces s. . . . . . . . . . . . . . . . . 5-4

, 5.2.4 Minimal Cut . Set Evaluation Process . . . . . . . . . . . . . . . . . . . 5-5 5.2.5 Schedule and Resources............................... 5-6 l

l REFERENCES ......................................................R-1 APPENDIX Identifiers and A Symbols.......................................A-1 8 Auxil iary Feedwater Syst em. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-1 4

C High Pressure Safety Injection System..............'...........C-1 D Primary Pressure Relief System................................D-1 E Primary Component Cool ing Water System. . . . . . . . . . . . . . . . . . . . . . . . E-1 F Secondary Component Cooling Water System......................F-1 G Service Water System..........................................G-1  !

H Electric Power System..........'...............................H-1 i

'I A c t u a t i o n Sy s t ems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I - 1 i

i i

i.

i - vii -

i i

i t

...,,_-.,.r._.-- . . . _ , _ _ _ , _ _ _ _ . - - _ _ _ . . _ . , . . . -, _, ._,,, -___ _ . , _ _ _ , - _ - .

k 4

i1 i TABLE OF' CONTENTS' i

4 VOLUME 3. FRA6ILITY ANALYSIS

+.

Pag i

1 A85 T RA C T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 1

LIST 0F TA8LES;........................................................vii

! LIST OF FI6URES........................................................ 1x

1. I NT RODUCT I ON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 1 7

i -1.1 Background................................................... 1-1

1;2 Objective of the-Study.....;..;.............................. 1-2 j '1.3 Scope o f t he St udy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 1.4 Organizati on of the Report. . . . . ; . . . . . . . . . . . . . . . . . . . . . . . . . . . ; . 1-3 l 2. REV I EW EA RTHQUA KE LEVE L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1 i 2.1 Review of Available Seismic Hazard Studies . . . . . . . . . . . . . . . . . . . 2-1  :'

l 2.2 Implication of the Selected Review Earthcpake Level.......... 2-1

{

2.2.1 Screen i ng of C omponents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1

2.2.2 Estimation of Seismic Capacities . . . . . . . . . . . . . . . . . . . . . . 2-5 ,

) 2.3 Different Methods for Specifying Review Earthquake Level..... 2-5

) 3. DESCRIPTION OF PLANT STRUCTURES, SYSTEMS AND COMP 0NENTS' . .......... 3-1

3.1 Maine Yankee Plant / Structures and Systems.................... 3-1 l 3.1.1 St r u c t ur es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 1 s 3.1.2 Sy s t ems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 1

{ 3.2 Sei smi c Des i gn Criteri a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2 3.3 Avail abil ity of Plant Des ign Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2 l 4. REV IEW 0F P LANT INFORMATI ON AND WALKD0WN. . . . . . . . . . . . . . . . . . . . . . . . . . 4-1 2- . ,

4.1 Ini tial Screening of Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1 .

4.2 Review of Design-Analysis and Seismic Reevaluation Reports... 4-4 4.3 P l a n t W a 1 k d ow n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 -6 j 4.3.1 Identification of Target Areas for First Walkdown..... 4-6 j

4.3.2 Wal kd own P roced ures . . . . . . . . . . . . . . . . . . . . . . ; . . . . . . . . . . . . 4-2 2  :

4.3.2.1 Wa l kdown Te am . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 -2 2 l' 4.3.2.2 Procedures for Structures and Equipment...... 4-23' 4.3.3 Wal kdown Doc umentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-29

4.3.4 W a l kd own Res u l t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 -3 0

! 4.3.4.1 Wal kdown F ind i ngs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 -30

! 4.3.4.2 Plant Uni que Featur es . . . . . . . . . . . . . . . . . . . . . . . . 4-56

!: 4.4 Maine Yankee Component Modifications . . . . . . . . . . . . . . . . . . . . . . . . . 4-56 i

4

t. 5. EVALUATION OF SEISMIC CAPACITIES OF COMP 0NENTS AND PLANT.......... 5-1 i

5.1 Review of Stru ctural Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1 5.2 Simplified Analysis and Use of Screening Tools............... 5-2 j 5.3 Second Wa1kdown.............................................. 5-4

!. 5.4 HCLP F Ca pa c i ty o f C ompo nen t s . . . . . . . . . . . .'. . . . . . . . . . . . . . . . . . . . . 5-4

) - viii -

, - - _ - , ___.m.w-,_.. m%..._--.-~_-__._.-.,_.-_. . ,.,,,_, _ - r . _ ,..-~ - - . - - -,- _.,m ..

1 TABLE OF CONTENTS (Cont.)~

.l VOLUME 3. FRAGILITY ANALYSIS P_ge 1

5.4.1 - Fragil ity Analysi s Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4  !

5.4.1.1 Methodology.................................. 5-4' ,

5.4.1.2 Steel St ru ctur es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-23 '

5.4.1.3 B l oc k Wa 1 1 s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 -30 5.4.1.4 Flat Bottom Storage Tanks.................... 5-32 5.4.1.5 Inverter..................................... 5-40 5.4.1.6 Diesel Fuel Oil Day Tank Fragility '

Evaluation................................... 5-46 5.4.1.7 Containment Spray Fans....................... 5-52 5.4.2 C D FM Me t h od . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 -5 3 5.4.2.1 Circulating Water Pumphouse.................. 5-54 5.4.2.2 B l oc k Wa l l S8 35-3. . . . . . . . . . . . . . . . . . . . . . . . . . . 5 -54 5.4.2.3 Refueling Water Storage Tank................. 5-55 t 5.5 H CLP F C a pa c i ty o f _ P 1 a n t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 -5 6 5.5.1 Accident Sequences.................................... 5 5.5.2 Probab il i sti c Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 -59 5.5.3 Deterministic Method..............'.................... 5-67,

~

5.5.4 Sen s i t i v i ty Stud i es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 -68

6. CO M NTS ON SEISMIC MARGIN REVIEW METH000 LOGY..................... 6-1 6.1 Selection of Review Earthquake Level . . . . . . . . . . . . . . . . . . . . . . . . . 6-1 6.2 Use of Screening Guidel ines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2 6.2.1 Ex t ent o f Rev i ew Needed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2 6.2.2 Additional Screening Guidelines. . . . . . . . . . . . . . . . . . . . . . . 6-2  !

6.2.3 Difficulties of Reviewing Certain Components.......... 6-3 6.2.3.1 Reactor Internals and CEDM. . . . . . . . . . . . . . . . . . . 6-3 6.2.3.2 Instrumentation within Containment.... . ..... . 6-3 '

6.3 Availability of Qualific ation Data. . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4 6.4 Walkdown Procedures.......................................... 6-5 i 6.5 Guidance on C0FM Capacity Calculation Procedures.. .. .... . ... . 6-6 6.6 Staf fing Requ irements and Schedul e. . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1 6.6.1 Staf f i ng Requ i r ements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-7 -

6.6.2 Sc h ed u l e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 -8 6.7 Applicability of Methodology to Other Plants......... . .. .... . 6-8

7. RE SU LT S AND CONC LUS I ON S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 -1 ,

7.1 Screened Out Systems and Components.......................... 7-1 7.2 HCLPF Capacities of Screened-in Components................... 7-2 7.3 HC LPF Ca pa c ity o f P 1 ant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-32 7.4 Identification of Low Capacity Components.................... 7-32 7.5 C on c l u s i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 -3 2 RE FE RE N C E S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R- 1 t

t

- ix -  ;

'^^'e , Me gu- w--ee - m i-wt e-- -W w + ke,4%.-c9 e,9m -g wap -er.w g-uy ge

TABLE OF CONTENTS (Cont.)

VOLUME 3. FRAGILITY ANALYSIS j

P,.a lg APPENDIX A Maine Yankee Atomic Power Station Arrangement Drawings.............A-1

.x.

4-

l LIST OF TABLES 1 I

Page Table 3-1 Component Sei smic Fragil ity Parameters . . . . . . . . . . . . . . . . . . . 3-3 Table 3-2 Probabilities for Nonseismic Failures.................... 3-4 Table 3-3 Summary of Plant Level HCLPF Capacities.................. 3-9 Table 3-4 Front-line System vs Support System Dependency Matrix.... 3-14 Table 3-5 Support System vs Support System Dependency Matrix....... 3-15 Table 3-6 Minimal Cut Sets for Important Systems. . . . . . . . . . . . . . . . . . . 3-18 Table 3-7 Maine Yankee Structures and Block Walls.................. 3-21 Table 3-8 Maine Yankea Equipment List.............................. 3-23 Table 4-1 Cost Breakdown for Seismic Margins Trail Plant Review.... 4-11 i

4 t

l

- xi -

y. - , -- -- - -.- ,--,- , - - ,, .-- ,, , ,. - . ,  % , --- ,

l 1

l l

l l

l l

LIST OF FIGURES Page Figure 2-1 Graphic rpresentation of the' screening operations

-(Figure 2-6 from NUREG/CR-4482).......................... 2-6 Figure 2-2 Exampl e systemati c event trees . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Figure 2-3 Exampl e of a simplified f ault tree. . . . . . . . . . . . . . . . . . . . . . . 2-11 Figure 2-4 Fault tree for failure of RCS integrity.................. 2-13 Figure 2-5 Exampl e fragility curves for a s tructure. . . . . . . . . . . . . . . . . 2-16 Figure 3-1 Fragility curves for small LOCA core damage.............. 3-7 Figure 3-2 Seismic fragility curves for NO-LOCA core damage......... 3-8 1

t:

1 2

a d

i

- xii -

q

PR2 FACE The seismic margin review of Maine Yankee Atomic Power Plant was performed for the U.S. Nuclear Regulatory Commissinn as part . of a research program on quantification of seismic margins at nuclear power plants. The approach, methodology (NUREG/CR-4334), and guidelines (NUREG/CR-4482), were developed by the Expert Panel on the Quantification of Set:mic Margins and its technical-support personnel. Maine Yankee and Yankee Atomic provided necessary data and information to this margin review. The results were reviewed by a Peer Review Group and the NRC Seismic Design Margins Working Group. This report is a collective effort and. presents the results of the seismic margins review of the Maine Yankee plant.

AUTHORS OF THE REPORT AND ANALYSIS TEAM U Volume 1. Summary Report Lawrence Livermore National Laboratory P. G. Prassinos R. C. Murray, Project Manager G. E. Cummings Volume 2. Systems Analysis Energy Incorporated D. L. Moore, Systems Analysis Team Leader 3

D. M. Jones M. D. Quilici J. Young Volume 3. Fragility Analysis l

EQE Incorporated j M. K. Ravindra, Fragility Analysis Team Leader M. J. Griffin G. S. Hardy P. S. Hashimoto U.S. NUCLEAR REGULATORY COMMISSION D. J. Guzy, Project Leader A. J. Murphy, Section Leader, Engineering Branch J. E. Richardson, Chief, Engineering Branch, Division of Engineering Safety t

- xiii -

, l f

r NRC Seismic Design Margins Working Group N. Anderson, Cochairman J. R,ichardson, Cochairman .

G. Bagchi L.'Beratan d' Chen N. Chokshi C. Grimes

/ D.' Guts P. K. Niyogi L. Reiter UTILITY Maine Yankee Atomic Power Company S'. V. Evans 7 G. D. Whittier Yankee Atomic Electric Company P. L. Anderson W. E. Henries J. T. McCumber W. J. Metevia R. P. Kennedy, Consultant REVIEWERS Peer Review Group R. J. Budnitz (Chairman), Future Resources Associates

- M. P. Bohn, Sandia National Laboratories J. W. Reed, Jack R. Benjamin & Associates, Inc.

J. Thomas, Duke Power Company L. A. Wyllie, H. J. Degenkolb & Associates Expert Panel on the Quantification of Seismic Margins R. . A Budnitz (Chairman), Future Resources Associates P. J. Amico, Applied Risk Technology Corporation 3 .- C. A.' Cornell, CAC, Inc. l J. W, Reed, Jack R. Benjamin & Associates, Inc.

M. Shinozuka, Columbia University I

- xiv -

l EXECUTIVE

SUMMARY

The U.S. Nuclear Regulatory Commission is sponsoring a research program to l develop and demonstrate a method for assessing safety margins at nuclear power  !

plants with respect to seismic events. This program is called the Seismic Margins Program and was started in 1984 at the Lawrence Livermore National l Laboratory, [Cummings et al.,1984]. The need for such research results from I the changing perception of the seismic; hazard in certain localities which could require reassessment of the adequacy of seismic design at some power plant. sites. It is generally accepted that nuclear power plants are capable of withstanding earthquake motion substantially greater than their design basis, but methods are needed to systematically demonstrate this. An " Expert Panel on the Quantification of Seismic Margins" was formed to develop such a margins assessment method [Budnitz et al., 1985] and [Prassines et al., 1986),

and this report discusses the application of-that method to the Maine Yankee Atomic Power Station, a Combustion Engineering, three-loop pressurized water reactor located near Wiscasset, Maine.

The margin review process involves the screening of components based on their importance to safety and their seismic capacity. The products of the review are High Confidence of Low Probability of Failure (HCLPF) capacities for components, accident sequences and the plant. Mathematically, the HCLPF can be thought of as an ' estimate of the 5% failure probability point with 95%

confidence. These HCLPF capacities, measured in terms of peak ground acceleration (pga), are compared to the peak ground acceleration predicted for the earthquake against which the plant is to be assessed, called the review level earthquake. This review level earthquake is chosen at some level above the design basis (safe shutdown earthquake, SSE) and a favorable comparison with the HCLPF capacity of the plant indicates that there is high confidence of a low probability of failure (core damage). The extent to which this plant HCLPF capacity is above the SSE level is a measure of the seismic margin of the plant.

Systems analysis is used to determine those plaat systems and components (including structures) that are important contributors to plant seismic safety and thus allow focusing of effort on components requiring a margin review. By studying previous seismic probabilistic risk assessments (PRA) of pressurized water reactors (PWR), it was found that only systems and components needed to assure reactor subtriticality and early emergency core-coolant injection needed to be considtred. Therefore, only systems and components related to these two functions (with a fr axceptions) were considered at Maine Yankee.

Capacitis -of generic sets of components were estimated and given in NUREG/CR-4334 based on estimated capacities in. previous PRA's, experience data gained from studying earthquake effects on industrial facilities and engineering analysis. These generic capacities were used to screen out from further consideration those components identified as important by systems analysis if the generic capacity was found to be higher than the review earthquake level for Maine Yankee. This review level earthquake was established by the NRC as

- xv -

4 "" - -e - , +y

9 I

0.30g with a 50th . percentile Newmark Hall spectra, as defined' in NUREG/CR-

'0098. To make sure that the generic capacity of all important cnmponents

_ .could be properly predicted, by using the tables found in NUREG/CR-4334, these

, components _ underwent review and inspection before being screened out. In '

! -addition, any potential system interaction _or plant-unique features discovered were added to the list of review items.

- The components remaining after the systems and fragility screenings, plus the systems interaction and plant-unique global features, were then subjected to a-

margins quantification. Prior to this quantification, _each remaining i component was thoroughly inspected .and studied, and systems models were developed to describe the possible seismic-initiated accident behavior of the plant. The quantification 'was accomplished by. calculating the HCLPF capacities for each of these components using structural / mechanical analyses t

- and then analyzing the minimal cut sets derived from the systems analysis

using the rules of Boolean algebra and Discrete Probability Distribution methods to arrive at accident sequence, and plant level HCLPF capacities.

Component HCLPF capacities were calculated for the important components remaining after screening, using the Fragility Analysis .(FA) method. This method requires estimating the' median failure capacity for the component, and i its random and modeling uncertainties. . Assuming a lognormal failure probability distribution, the HCLPF (5% failure with 95% confidence) capacity '

can be calculated. Checks on several key components were made using the-Conservative Deterministic Failure Method (CDFM) which uses a deterministic,

- more design-oriented method of. calculating component HCLPF's.: Further work is i

~

underway to cross-check the two methods. Random, test and maintenance, and human error failure modes were also included in-the analysis.

MARGINS REVIEW OF MAINE YANKEE The seismic margins review of Maine Yankee was -an eight-step process which involved Maine Yankee, Yankee Atomic Electric, NRC, LLNL as project manager, and fragility and system analysis teams (EQE Inc. and . Energy Incorporated, respectively). The first step was to establish the review level earthquake (0.30g). Steps 2 and 3 involved information gathering and preliminary i analysis by the two teams leading to step 4, a plant inspection (called a walkdown). In step 5, following the plant inspection, event and fault trees were constructed including those components not screened out. In step 6, a second visit was made to the plant to recheck the components remaining which

. might require further analysis. In step 7 minimal cut sets leading to core L damage were determined. In step 8, the component- HCLPF capacities were finalized and.- HCLPF capacities for accident sequences and the ~ plant

calculated. This entire effort took about 3.0 man-years of_ effort.

- For Maine Yankee, two important accident sequences were identified. Both are initiated by a seismically induced loss of offsite power (LOSP) assumed to always occur at the review earthquake level. In one sequence there is a small-loss-of-coolant accident (LOCA of-3/8 in. to '2 in. diameter equivalent ~-area) assumed to occur _ because of seismically ~ induced pipe breakage. HCLPF capacities for components'which might cause other types of small LOCAs -(pump

seal or power-operated relief valve _LOCAs) were sufficiently -high so they i could be screened out. .The other accident sequence assumed no small LOCA.

- xvi -

l n*

4

. ,. _ , , , , , _ , _ ,_,,,_4 _ , , , , , _ , , , _ _ ,, __, _ , , , , , ,__ ,, _

e +-v- ---w

, ,-,----.,..,-e

r -

.i l

l The small LOCA accident sequence involved seismic failures only and resulted in a plant HCLPF of 0.219 Many plausible arguments, including the component screening table of NUREG/CR-4334, indicate that a small_LOCA could be screened out for the review level earthquake considered. Since the analysts involved in this review could not get inside the Maine Yankee containment to inspect the small primary system piping, they chose not to screen small LOCA out. If they had, this sequence would not be considered and the plant HCLPF would be above 0.30g.

The small LOCA accident sequence was composed of three singleton cut sets with the dominant contributor being failure of the Refueling Water Storage Tank (RWST) which had a HCLPF of 0.21g. Other singletons in the sequence had HCLPF's greater than 0.30g.

The second accident sequence, LOSP with no small LOCA, involved no singletons but a number of doubletons, some combining seismic and nonseismic (random, test and maintenance, human error) failures. The most important doubleton is the Demineralized Water Storage Tank (DWST, HCLPF - 0.179) and the Circulating Water Pump House (HCLPF > 0.30g).

Nonseismic failures were found not to be important contributors. They made no contribution to the overall plant level HCLPF capacity. The most important

, nonseismic failure found was a common cause failure of the Auxiliary Feedwater Sysgem, caused by steam binding (median unavailability per demand of 1.2 x 10- ).

It should be noted that during the review process, important components were found for which a low HCLPF would result or insufficient data was available to determine the HCLPF. These were the lead-antimony station batteries, the station service transformers, a block wall near HVAC equipment, parts of the PCCW and SCCW air conditioning heat exchangers, and anchorage of the diesel generator day tanks. To reduce the uncertainty in their capacities, these components are being replaced or upgraded and the results of this review are based on the upgraded configurations.

INSIGHTS AND LISSONS LEARNED In addition to the results already described, the Maine Yankee seismic margins review provided some lessons applicable to future reviews. The event / fault tree methods used provided a complete description of dominant contributors and considered all important systems. The fragility screening table (NUREG/CR-4334) needs to be strengthened and more guidance given on how to select a review level earthquake. Also, more guidance is needed on how to use and compare the two methods used to determine component capacity (FA and CDFM) and 4

how to combine seismic and nonseismic failures.

The systems analysis effort should start early so that component screening and plant inspections can be done efficiently. Information in the plant Final Safety Analysis Report can be effectively used for this effort. Plant

~

inspections (walkdowns) need to be carefully planned, taking into account '

auxiliary systems such as the HVAC and actuation / control system, as well as

- xvii -

1 l

l

'l

.I

- - , = - . . - - . _ , . - - _ . .. -

l important systems and components identified in the systems analysis. These.

walkdowns are essential to successful margin reviews. Specifically, they permit accurate data collection and allow identification of potential low .

capacity components.

More guidance is needed on the consideration of the seismically initiated small LOCA. This initiating event turned out to be particularly important at Maine Yankee. It may be impossible to review and inspect all the small' primary pressure boundary piping, and other screening methods need to be developed, e.g., inspection of similar piping outside of containment.

Finally, there is a question about maintenance of hot shutdown for a specified period of time after an earthquake, e.g., 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />. The current methodology addresses the attainment of hot shutdown but not necessarily its maintenance. Review of. previous seismic PRAs indicates that once hot shutdown is attained, the probability of maintaining it is large. Therefore,- this issue may be of less importance.

- xviii -

s P

1

~

VOLUME-1. SUMARY REPORT

- CHAPTER 1

1. INTRODUCTION In 1984, the U.S. Nuclear Regulatory Commission (NRC) initiated the Seismic-Design - Margins Program (SMP) to . address regulatory . needs and a changing

.. perception of the seismic hazard. The NRC . formed the " Expert - Panel on the -

Quantification of Seismic Margins" and charged it to work closely with an.in-house NRC staff, " Working Group on Seismic Design Margins," to provide -

technical guidance on the assessment of seismic margins. _The overall goal of the SW is the development of a methodology and guidelines that can be readily-used by the NRC and industry for assessing the. inherent quantitative seismic capacity of nuclear power plants [Cummings et al.,1984].

The development of a soundly ' based, efficient and effective method for the assessment of - how much- margin actually exists in .important components, systems, and the plant will serve to minimize the . impact. of- changing regulatory requirements and licensing actions as the - estimates of seismic hazards change. In addition, a seismic margins assessment can provide a basis' for confidence in .the capacity of nuclear power plants and this methodology -

can be applied when questions arise about their seismic capacity.

The most important regulatory need anc' the focus of the seismic margins effort is stated as follows:

"There is a need to understand how much seismic margin exists at nuclear power plants. This seismic margin is to be expressed in terms of how much larger must an earthquake be above then safe shutdown earthquake (SSE) before it compromises the safety.- of the plant."

The Expert Panel and its technical support personnel studied. the available information on the quantification of seismic capacity ~ of nuclear power plants and other industrial facilities. The results of several seismic lprobabilistic risk assessments of nuclear power plants were reviewed along with the behavior l of industrial facilities during earthquakes. These studies were used .to develop a margin ~ review approach that involves both the. screening of components based on their importance in' preventing seismic core melt and their inherent seismic capacity.

The seismic margins review approach has been documented in the report, "An Approach to the. Quantification of Seismic. Margins in Nuclear Power Plants"

[Budnitz et al., 1985]. This document formed the basis for the development of guidelines for performing ~ seismic margin reviews. .These guidelines are given in "Reconsnendation to the Nuclear Regulatory Consnission on Trial Guidelines for Seismic Margin Reviews of Nuclear Power Plants, Draft. Report for. Comment"

[Prassinos et al., 1986].

1 1

1-1

1e 4.

The performance of a trial plant review is needed to verify and . test the review methodology and guidelines. For the trial review, the NRC negotiated with and selected the Maine Yankee plant. Once the trial plant was selected, the two analysis teams (Systems Team and Fragility Team) were chosen. .These .) '

teams were selected based on their technical approach and team. composition.

In addition, a Peer Review Group was selected and charters were. developed for

! both this group and the Expert Panel ~. The seismic margin review was then

- organized allowing for participation by the plant owner and its representative (Yankee Atomic Electric Company), the NRC Working Group- on Seismic Design l

Margins, and the appropriate NRC program managers.

The purpose of this report is to summarize the results.of this first trial seismic margin review. The detailed results from this review can be found.in Volumes 2 and 3 of this report. Volume 2, Systems Analysis, documents the results of the systems screening and analyses portion of the review, while Volume 3, Fragility Analysis, gives the component fragility screening and analyses results. -

The overall objectives of this trial review are to assess the seismic margins

of a pressurized water reactor, the Maine Yankee Atomic Power Station', and to

test the adequacy of the seismic margin review approach and quantification techniques, and the guidelines for performing these reviews. There are four 4

i related objectives of this study:

o To demonstrate the use of the Expert Panel's approach (NUREG/CR-4334) and guidelines (NUREG/CR-4482) for seismic margin reviews.

o To provide a basis for revising and . upgrading - the approach and-  ;

guidelines.

o To provide a benchmark for possible future seismic margin reviews, including an understanding of the level of effort in performing a

, seismic margin review.

o To provide an assessment of the plant's

capability to withstand a -

specific' earthquake level greater-than the SSE.

I The results from this trial review are intended to be used to revise the

! seismic margin methodology and guidelines so they are more' prescriptive and ready for general use by the NRC and industry.

l J

The role of the Expert Panel during the Maine Yankee study was to review the

[ procedures being used at an early stage of the process and to assure that the methods and techniques being employed are. consistent with the Panel's guidance and are relevant for performing a seismic margins review.' When ' the . trial-plant review is completed, the Panel will study the results, examine how the review was implemented, . and evaluate the overall effort. The Expert Panel will be expected to appraise the overall usefulness of. the seismic margin review approach and identify any limitations whether previously recognized or

, not. Appendix A defines the role of the Expert Panel .for this review and.

contains minutes of the panel's telephone conference call.

1-2 y y- w- y-p..g,-s , _ , . -

y

h A Peer Review . Group was selected and . chartered to review ' the technica'l

-- adequacy .of. this study including participation in the plant walkdowns.- The objective of the Peer Review Group is to assure that the seismic margin review -

i is executed in a fully competent and professional manner, uses appropriate i l

methods, and follows the guidance established in NUREG/CR-4334 and NUREG/CR-4482. At the conclusion of the Maine Yankee review, the Peer Review Group -l

• provided its best judgment with. regard to both the review procedures and the  !

' technical competence of the reviews, based on its collective expert opinion J 4

[ Anderson, 1986]. Peer Review' Group correspondence including its charter, meeting minutes and sumary report on this trial review are gjven in Appendix

B.

Appendix C contains comments received from the utility.

1.1 Organization of the Report The remainder of this section gives a discussion of the scope of the effort.

Chapter 2 presents a brief overview of the seismic margin review methodology

. and the guidelines as applied to this review. -Chapter 3 provides the overall ,

results .of the Maine Yankee review, including a summary of the significant l fragility and systems results. Insights and lessons learned are given in 1 Chapter 4 followed by the conclusions in Chapter 5. Recommendations from this study are given in Chapter 6.

1.2 Scope of the Seismic Margin Review of Maine Yankee 4

i The Maine Yankee Atomic. Power Station is a' Combustion' Engineering (CE) three-2 loop pressurized water reactor (PWR) located approximately 3.9 miles south of Wiscasset, Maine. The architect engineer -for. the- plant was Stone and Webster-Engineering Corporation. The Maine Yankee plant _ started comercial operation i in 1972. Its present net electrical power output is 825 megawatts electric j (2630 megawatts thermal). A brief description of the plant- configuration, including structures and systems, is given in Volumes 2 and 3. A detailed j description of the Maine Yankee plant is given in the FSAR [ Maine Yankee, FSAR].

I All seismic " Class 1" structures and components of the - plant which are

'important to nuclear safety, and could affect the. health and safety of the public, are designed based on a minimum horizontal ground acceleration of i 0.05g and a safe shutdown earthquake (SSE) horizontal acceleration of 0.lg

[MaineYankee,FSAR]. Damping at these acceleration levels is 2 percent and 5

percent, respectively. Vertical acceleration is taken as two-thirds of the horizontal acceleration and is considered to .act simultaneously with .each horizontal component.

The occurrence of two seismic events in the vicinity of the plant, one-in 1979 and the other in 1982, prompted Maine Yankee to upgrade the capability of the

plant to withstand a potential seismic event in excess of the original design-basis event. Based on these upgrades and the inherent design capacity of the 4 plant, Maine Yankee concluded that the plant ' structures, systems, ~and l

4 1-3 1

?

. l

\

, . - _ . - - _ ~ . _ . , . - _ . _. _ ._ __ __ . _ . . __ _ . . _ J

components had sufficient strength to withstand a seismic event of at least ,

0.2g with a Regulatory Guide 1.60 [USNRC,1973] spectrum and still shut down without danger to the public health and safety [Miraglia, 1986].

To assure the NRC that the plant could withstand earthquake motion greater than the design basis, the utility agreed to participate in the trial seismic margin review of the Maine Yankee plant. For this review, it was agreed the seismic margin review earthquake level would be 0.3g with a 50th percentile Newmark Hall Spectra defined in NUREG/CR-0098. Magnitude and duration '

criteria specified in NUREG/CR-4334 apply [Guzy,1986], [Crutchfield,1986].

The margins concept requires the HCLPF capacity to be associated with a defined response spectrum and a specified nonexceedance probability.

The HCLPF capacity used for screening as well as the calculated .HCLPF l capacities for particular components not initially screened out and the fina' plant level HCLPF capacity are considered to be valid provided ground motion from any earthquake does not exceed the review earthquake level spectrum for more than 16% of the spectral frequencies within the range of interest. The review earthquake level spectrum is a spectral shape defined by the 50%

exceedance spectrum specified in NUREG/CR-0098 and anchored at 0.3g pga for the initial screening. The seismic margin for the components and plant is referenced to this spectrum but anchored to the pga corresponding to the HCLPF capacity.

l This definition of spectra used to determine a HCLPF capacity does not in any way refer to the probability of occurrence of an earthquake. It is no more than an arbitrary spectrum used to define the HCLPF capacity that recognizes the dependency of a component capacity on the frequency content of the spectrum and not just the pga. l The results of the seismic margin study are interpreted as follows. The HCLPF capacity of the structures, equipment and plant are conditional on the actual site-specific spectrum not exceeding the target spectrum; exceedance is defined as the event when 16 percent of the spectral ordinates exceed the-target spectrum over the frequency range of interest. It is assumed that the spectrum peak-to-peak and earthquake direction variabilities are removed from the hazard analysis leading to the selection of the review earthquake. The review earthquake is specified by the same spectrum in two horizontal directions and 2/3 of the horizontal spectrum in the vertical direction. It is also assumed that the review earthquake level is specified as the higher of the response spesctra from the two orthogonal horizontal directions.

i 1-4 o

9 s ,._m,--.r - , - = , - - - - y- - ,v- - - - , -

,-i -----,n- -.--~ ,. , .

~ . - - . .. - - - . .. - - . .

. CHAPTER 2

2. SEISMIC MARGINS APPR0A'CH Insights. gained from the results of 'seven published 'probabilistic risk:

assessments (PRAs) ~ were used in the development of a screening approach that combined systems insights and fragility. information- to simplify the margin i

review process. This approach is directed at reviewing a specific-plant at a selected earthquake acceleration level greater than the SSE.

A general definition of " seismic margin" is stated below:

! " Seismic margin is expressed in terms of the earthquake motion level that compromises plant safety specifically leading to melting of the. reactor core. In.this context, margin needs to be-defined for 3

the whole plant. The margin concept also can be extended to any.  ;

particular function or component."

The adopted measure of margin is the earthquake level for which there Jis' a High Confidence, Low Probability of_ Failure (HCLPF). The HCLPF.is a conserva-tive representation of capacity and in simple terms corresponds to the earthquake level at which it is extremely unlikely that core damage will-

occur. Generally, the median capacity is at least a factor of 2 greater than

< the HCLPF capacity and thus nof proverbial " cliff" or sudden. failure is t expected to occur immediately upon exceeding the HCLPF capacity. From the l

i mathematical perspective of a failure probability distribution on capacity' (fragility), the HCLPF capacity is approximately equal to a 95 percent confidence (probability) of not exceeding about a 5 percent probability of failure. HCLPF capacity for specific types of components are derived from a combination of engineering data, either test data or data from real earthquake j experience, and engineering analysis.

) 2.1 Systems Screening i

l The review of the available seismic PRAs indicated some key trends and i insights useful in developing seismic margin review criteria. Chapter 4 of i [Budnitz et al., 1985] discusses the details of the screening approach. These trends and insights could only be obtained for pressurized water. reactors (PWRs), for which there were seven available PRAs.

1 The Expert Panel review indicated that systems screening must be performed at

the functional level due to the diversity of plant designs and system '

configurations. This led to considering the general plant safety functions normally performed in PWRs:

? 1. Reactor Subcriticality.

2. Normal Cooldown.

3. Early Emergency-Core Cooling (injection).

4. Late Emergency Core Cooling (recirculation).

5. Containment Heat Removal.

6. Early Containment Overpressure Protection (injection).

i 7. Late Containment Overpressure Protection (recirculation).

l J

2-1 t

a

+m-. ,.- -~ . - - -n-,a.., , = - - - - N n ,... ,..,,,,.,...m,.,- , w, .,a ,-,-,.ewu

b , . Examination of PRA results indicated that the= dominant plant damage states to

[ core melt for seismic events are generally early core melt with early containment failure. .In addition, these plant damage states generally involve core melt induced by failure .of the first three functions listed above followed by loss of containment integrity early in the-accident progression.

i. This insight led to the division of plant safety function into two groups: l Group A - Functions.1-3.-

Group B - Functions 4-7.

The systems screening approach is based.on this insight. The dominant plant damage states are caused -by failure of the Group A functions and these plant i damage states are also characterized by : functional faib e of the Group B functions. Therefore, we consider core damage to occur whenever there is a failure of the systems that provide the initial shutdown of the nuclear reaction and cooling of the reactor core. These failures are followed - by failure of the systems- that provide the Group B functions so as to preclude the mitigation of consequences by_ providing containment protection and cooling.

The normal cooldown function is not considered in our screening criteria because a loss of offsite power is assumed to occur in all seismic PP.As as a result of " low capacity" switchyard components.

i The relationship between the Group A and Group B functions indicated that they

, are highly coupled so that the combined failure of Group A and success of' Group B (or the combined success of Group ' A and failure of Group L B) is l

virtually precluded since their weakest links 'are coupled. This insight

indicates that only those systems and components needed to perform the Group A function must be considered in a seismic margins review.

i .

i The Expert Panel also discovered that seismic core melt can occur due to the

failure of specific unique features. This implies that it will always be

necessary to perform some kind of plant walkdown to look for.~ any unique

features requiring a margin review. It must be emphasized.that only major, d unique features are of concern. Examples of these are the presence of a dam upstream of a plant or a free-standing stack that could damage surrounding buildings if it collapsed.

! The insights gained from our systems review of the PRAs and the development of

a systems screening criteria to simplify the margin review process has

! resulted in the following conclusions:

f 1. It is possible only to come to conclusions regarding the . relative

! importance of plant systems and safety function for PWRs for which six plants were studied (one by.two different methods). No function / systemic conclusions can be made about boiling water reactors (BWRs) without examination of additional PRAs.

l 2. For PWRs, it is possible to categorize plant safety functions as j belonging to one of two . groups, . one of which is important to the assessment of seismic margins and one.of which is not.

2-2 l

~ . - . .- - - . - - - . .

.5 I 3. The important group involves only two plant functions that must be considered for estimating seismic margin. These two functions are shutting down the nuclear chain reaction and providing cooling to the reactor -core in the time period immediately following the ' seismic L

event.

4. It is possible to reasonably estimate the seismic margin of the plant by performing a margins study only involving the analysis of the plant systems end structures which are-required in order to perform those two safety functions.

Using the systems screening criteria and combining it with fragility insights,

!' we - can establish functional / systemic guidelines for margin reviews. These~

guidelines show that it is- possible to perform a reasonable seismic margin review by concentrating on those functions (and associated systems) .which are-required in the .early part of the seismic event and eliminating from the review those functions which are not required until later. Further, depending

) . on the level of earthquake for which it is desirable to define a margin, certain initiating events would not have to be considered (e.g., large LOCAs). This reduces the level of effort and scope of the analysis.'

Event trees need only be constructed up to the point of determining whether or a

i not there is an early core damage. Fault trees would only have to .be

{ constructed for those front-line systems (and their support systems) that j appear on these abbreviated event trees. By combining a nonseismic failure probability and seismic fragility screening criteria with these sys_tems models, it would only be necessary to include those components which have not been removed during the screening process.

2.2 Fragility Screening

The available fragility information that was reviewed and ~ assessed is based I primarily on the detailed analysis of nuclear power plants performed for PRAs. This PRA information was supplemented by recent systematic investi-gations of historic earthquakes. The information includes past earthquake performance data for eight classes of equipment obtained. by the Seismic l Qualification Utility Group (SQUG), and reviewed by the Senior Seismic Review

] and Advisory Panel (SSRAP). Work is ongoing to document the historic

! earthquake performance and qualification data for additional components such as piping, valve operators, penetrations, diesel generators, battery racks, and electrical equipment. This work is being conducted by SQUG, the ElectHc Power Research Institute, and the American Society of Civil Engineers Dynamic

!. Analysis Committee [ASCE,1986]. Additionally, the collective knowledge of

the Expert Panel members with actual earthquake experience -at nuclear and nonnuclear industrial facilities, perfonnance test data, and other' analyses were included in the assessment.

These available sources of fragility information were used to arrive at  !

conclusions about which components should be assessed from a seismic capacity -;

standpoint. In making statements about the need for capacity assessments for

each component, three ranges, stated in peak ground acceleration (pga), were t

l i 2-3 l i

i

-_lI

used: (1) less than 0.3g, (2) 0.3g to 0.5g, and (3) greater than 0.5g. Each type of nuclear power plant component was - assessed to have a generic HCLPF capacity within one of these ranges. This resulted in an extensive table of components indicating at what earthquake level each component will require a margin review or be removed from the review process. This table is given in

[Budnitz et al., 1985).

This categorization of components is based on the available information, and 1 the fragility screening resulting from the use of this table should only be performed with consideration of the caveats, limitations, and assumptions presented in Chapter 5 of [Budnitz et al., 1985].

During the process of assessing the HCLPF capacities for the various nuclear power plant components, an extensive fragility information base was developed from the available seismic PRAs. This fragility information base is available on a diskette for use on personal computers, and is documented in [ Campbell et al.,1985].

2.3 Approach for Performing Seismic Margin Reviews The combined insights gained on plant functions and component fragilities were used to develop an outline of an approach for performing seismic margin reviews. The review approach consists of eight steps, the first of which is the selection of an earthquake review level for which it is desirable to demonstrate margin. The eight steps developed in [Prassinos et al., 1986] are outlined later in this section.

It is important to point out the assumptions and limitations of the approach.

1. The systems screening part of the approach presently applies only to PWRs.

2. The review approach focuses on earthquakes that could occur in the eastern part of the U.S., specifically east of the Rocky Mountains.

3. The assessment of component HCLPF_ capacities is limited to earth-quakes of less than a magnitude of about 6.5, which are characterized by 3 to 5 strong motion cycles with a total duration of 10 to 15 seconds.

4. The effects of undiscovered design and construction errors are not covered.

l 5. Possible vulnerabilities in hydraulic systems associated with sensors and pneumatic systems are not fully covered.

6. Electrical and control systems are incompletely ccvered because unrecoverable relay chatter and breaker trip is not adequately treated at this time.

7. Evaluation of the effect of wear and aging on equipment function is not fully covered.

2-4 I

e -- -

I' i

8. Possible adverse human responses caused by earthquake-induced stress are not explicitly covered.

The first three limitations listed above are based on the data from PRAs and industrial facilities that were used in the development of this approach and present true limitations on the methodology but not on the Maine Yankee l review. Some of the remaining items represent limitation on our knowledge of how to adequately address these issues, while others require considerable effort and are beyond the scope of this analysis.

2.4 Trial Guidelines for Seismic Margin Reviews The objective of the seismic margin review guidelines is to provide guidance for determining whether a plant can resist with high confidence a specified earthquake level greater than the SSE. To accomplish this objective, analyses are performed on components, systems, and the plant to determine what the HCLPF capacity is so that it can be compared to the specified earthquake level. Plant failure is defined as the onset of core damage.

A flow chart of the margin review process is shown in Figure 2-1 This process involves the screening of components based on their importance to plant safety and their seismic capacity. Inspection of Figure 2-1 indicates that Steps 2, 5, and 7 are primarily concerned with plant safety functions and systems, and are performed by a team of systems analysts. Steps 3, 6, and 8 are mainly concerned with capacity assessment and are performed by a team of fragility analysts. Step 4 is performed by both teams of analysts. The entire process requires close cooperation and interaction between the two teams of analysts and the utility.

The initial step in the review proccss is the selection of the margin review earthquake level. The margin review earthquake level selected for the Maine Yankee review was discussed in Section 1.2.

In Step 2, plant information gathering, review, and analysis is performed to determine those plant systems and components (structures and equipment) that are important contributors to plant safety and thus allow focusing of the effort on the components requiring a margin review. Also performed during Step 2 is an identification of the relevant seismic initiating events and the development of preliminary event trees that describe the systematic behavior of the plant following these initiating events.

The team of systems analysts review the plant information and determine those front-line systems that perform the two functions important to plant safety, {

reactor subtriticality and early emergency core cooling. Examples of these  !

systems are the reactor scram, emergency boration, high pressure safety l injection, and auxiliary feedwater systems. The support systems to these j front-line systems are then determined. Examples of the support systems j include electrical power, cooling, actuation, and control.

2-5 l

l

Start

.T.im6 axis 1 Select an earthquake review level Gather information .;..., '.**. / Gather information on on systems and sort *.. the plant. Determine which Group A functions. 7, *;. .. . . .* . * .

• broad classes or groups of Interaction components have HCLPF Use information . . . , . . . +----* /3/

on Table 2.3 and ,. ., . . - /// / values greater than the review Ref.1. . .: ;

• .* *. *. // // level. Possibly identify-. plant-unique features.

.. .. First plant walkdown:

. . , . , - Concentrate on identification of problems.

. .,4 .

Emphasize systems interaction. Confirm

, " . . . * . .. , applicability of screening tools. Complete Identification of plant-unique features.

v KEY: Task is performed by: '

, , ;l,;;; , . Revision of systems

• *' '

• relationships established Systems analyst .

. . . . .t 5.; ; t ;

In Step 2. Develop fault g ,

. ; *.*,*t*.*,*.*.

trees and event trees.

Fragility analyst u

13oth Second plant walkdown:

Primarily fragility analyst for checks.

6 Collect specific data (size and other physical characteristics) or components requiring detailed analysis.

u

, , . . . .5.* . .*

,, ; "[ /

Determine minimal ** ,. **.* ; . . ; . *.* Finalize HCLPF value for cut sets for end- ...7*.*.* 8 components in final cut sets point core melt. * ' ; *.,, * ,. ,*

.f* f, .' . (components not screened out).

. /

o Margin assessment complete Figure 2-1 Graphic representation of the screening operations.

2-6 t

s

-,-m ,--- - -

- - - - . , - _ _ , , - . - _ . , _ + y- -y . ,.-

P Once these systems have been determined.. the components that make up these front-line and . support systems are _ listed. 1This list is- shared with the fragility analysts and these components become the focus of the first_ plant walkdown.

Using the identified front-line systems, preliminary systemic event trees.are

-_ developed for seismic initiating events. An example of an event tree is shown in Figure 2-2.- An event tree is a logic diagram that is used to determine the sequence . of successes- and failures. of the systems - of : concern following a

-postulated-initiating event.-

An illustration of an accident sequence is shown in' Figure 2-2 by the heavy dark line on the tree. This sequence assumes the_ initiation f event' (IE) has occurred followed by the failure of sys1, the failure of sys2, the success.of~ .

sys3, and the failure of sys4. - ' This accident sequence _ is represented by the _-

Boolean expression:

AS1 - (IE) * (sys1) * (sys2) * (sys3) * (sys4), .(1) where the sysiiem identifier (sys1, sys2,. etc.) with the bar_ indicate system success and the others indicate system failure. This equation reads "the initiating event must occur and sysl must fail and sys2_ must fail and sys3 must succeed and sys4 must fail" for AS1 to occur. The multiplication is the mathematical representation of the logical "and" operator where each event must occur for the occurrence of the outcome.

Based on the systems that fail and. succeed in each accident sequence Boolean expression, it is determined whether that sequence of events leads to core damage (CD). If core damage is postulated to occur, we then need to determine the occurrence of each event within the accident sequence to determine the sequence outcome. Since we have assumed the initiating event always. occurs given the occurrence of the margin review earthquake, the occurrence of the initiating event is taken as a certainty (probability - 1.0).

Normally, system success is quantified by one minus the' probability of system failure. However, where system failure _ is small, the occurrence of each:

system success is assumed to be unity (probability - 1.0). This' assumption is made for a seismic margin review at. a review earthquake level _ of 0.3g. For-higher review levels, system success may need to be considered.

A Boolean expression for the occurrence of each system failure is determihed by. fault tree analysis. Fault trees are used to determine the combinations of -

front-line and support system component -failures' that._ lead to system-failure.- For the Maine Yankee analysis, the component failures - are represented by their fragility curves and nonseismic failure probabilities.

Each accident sequence Boolean expression is then quantified by replacing each system failure with its Boolean expression and solving the-resulting logical expression for the sequence HCLPF capacity.-

2-7

O t ti SYS4 Status SYS1 SYS2 SYS3 Success n

II Failure OK OK OK CD CD = Core damage CD t

Figure 2-2 Example systemic event trees 2-8

A more detailed discussion of systems analysis and event tree development is given in the IREP Procedures Guide [Carlson, 1983] and the PRA Procedures Guide-[USNRC,1983],

For the Maine Yankee review, the initiating event was considered a seismic induced loss-of-offsite power (LOSP). An event tree was constructed that considered the systems that perform the reactor subcriticality and early emergency core cooling functions.

In addition, the event tree considered the integrity of the reactor coolant systems (RCS). This event considered whether a loss-of-coolant accident (LOCA) would occur as a result of the earthquake along with the LOSP. Only a small LOCA was considered to occur. A large LOCA was not considered because large RCS piping and their supports have generic capacities above the review earthquake level and were screened out. The break size of the small LOCA was considered to have an area equivalent to a 3/8-to-2-in.-diameter pipe and requires the operation of the high pressure injection system for mitigation.

The event trees developed in this step were preliminary. They were revised following the first plant walkdown and finalized after the second walkdown.

If, however, the first walkdown uncovers reasons for including other systems, initiating events, or components within the analysis, these will be included.

Concurrently, in Step 3, knowledge gathered from the plant and knowledge of the inherent capacity of components is used to sort the components developed

< in Step 2 into two groups, those with a generic HCLPF capacity larger than the review earthquake level and those that have smaller HCLPF capacity.

For the Maine Yankee review, the culmination of steps 2 and 3 resulted in the identification of structures, block walls, equipment, and areas of the plant that need to be inspected and reviewed. In addition, the first walkdown was planned including organizing the walkdown teams, developing procedures for the review of the various components, developing data sheets for recording the findings, and making arrangements with the plant for the necessary health physics counting, badging and training.

A first plant walkdown is performed in Step 4. This walkdown is performed to inspect the plant and confirm that the plant's configuration is such that the rules developed for doing a margins assessment are -applicable and that components can be screened out based on the generic evaluation. Assuming this is the case, the appropriate components are eliminated from further consideration. During this walkdown, any system interactions, system dependencies, and plant unique features will be identified along with confirmation of the accuracy of the system descriptions and configurations.

i During the first walkdown of the Maine Yankee plant, the analysis team members formed into groups and inspected components and plant areas that were identified during the previous steps. The Peer Review Group and NRC personnel

. also formeo into groups to walk down the plant. The walkdowns were performed with various levels of detail depending on the requirements of a particular group. Arrangements were made to have team meetings at the beginning and end of each day. Meetings were also arranged with knowledgeable plant personnel to discuss details about the plant and the review.

2-9 l

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

O l

i.

l

. Following the completion of Step 4, many of the components identified as c belonging to the two important plant safety functions were screened out based on the inspection and their : generic HCLPF capacities being larger than 0.3g.

In addition, plant information that .was gathered- during this review was used to revise the plant models and perform a conservative evaluation of, the remaining component HCLPF capacities. Those components that were evaluated to

' have a HCLPF capacity larger than 0.3g were also screened out.

During Step 5, the information and understanding of the operation of the plant ~

following Step 4 are used to review and revise, if necessary, the event trees developed in Step 3. Fault trees are developed during this step for systems

that perform the two safety functions (subcriticality and early core cooling)..

A system fault tree is a logic diagram that models the various parallel and sequential combinations of faults that will . result in the occurrence of the-

- predefined undesired top event. For the seismic margins review,-the faults

~ are associated with component seismic failure capacities, human error, and-other pertinent nonseismic failure events which can lead to system failure following an earthquake. A fault tree ~ thus - depicts the logical interrelationships of basic events that lead to the occurrence of' the top event, system failure.

A fault tree is a diagram of " gates" which serve .to permit' or inhibit the passage of fault logic up - the tree. The gates show the relationships of' events needed for. the occurrence of a. " higher" event. .

The " higher" event .is the " output" of the gate; the " lower" events are the " inputs" to the gate.

The type of gate denotes the relationship between the input events required for the output event.

The two basic gates in a fault tree are the "0R" gate (+),. and the "AND" gate (*). The OR gate indicates that the occurrence of the output event will ~

result from the occurrence of any of the input events. The output failure of an OR gate occurs if any of its inputs fail. The AND gate indicates.that the output event will only occur if all of the input events occur.

An example of a simplified fault tree is shown in' Figure 2-3. This figure shows that the system (SYS1) will fail if both components-(COMP 1 and COMP 2) fail. COMP 1 fails if either basic event (BE1 or BE2) occur, and COMP 2 fails if both basic events (BE3 and BE4) occur. The basic: event represents the failure modes of individual components. ,

The Boolean expression for the system failure, SYS1, is: _

SYS1 - COMP 1

• COMP 2, (2) where the mathematical representation of the logical AND operator is multiplication.

The logical expressions for the component' failures, COMP 1 and COMP 2, are:

i COMP 1 - BE1 + BE2, (3) 2-10 1

- , - - , . . . , . . , -,.--.,.r-, - - -

-~~r- . # - . --m y,- .y .m-m . . , ,,n m_ , ,,,-.,____,m.. . , ,

SYS1 I

and l l

COMP 1 COMP 2 )

i i

or and r%

l BE1 BE2 BE3 BE4 SYS1 = COMP 1

• COMP 2  !

COMP 1 = BE1 + BE2 COMP 2 = BE3

• BE4 I SYS1 = (BE1 + BE2) * (BE3

• BE4) -

SYS1 = BE1

• BE3

• BE4

+ B E2

• B E3

• B E4 Figure 2-3 Example of a simplified fault tree 2-11 --

COMP 2 - BE3

• BE4, (4) where the mathematical representation of the OR operator is addition. ,

The system Boolean expression in terms of the basic events-is then derived by replacing the component failures with their respective' logical expressions.

The system Boolean becomes:

SYS1 - (bel + BE2) * (BE3

• BE4) . (5)

- or after expansion,

,  : SYS1 - (BE1

• BE3

• BE4)

[ + (BE2

• BE3

• BE4). (6)-

Each term in expression. (6), separated by the addition operator, is called a

" cut set." A minimal cut set represents the smallest number.of events that will cause system failure if all the events fail.

4 The above simplified example is intended to show the basic concepts of fault trees and systems analysis. Normally, the analysis of fault trees is very complex and requires the use of computers. Logical and Boolean expressions

, are usually very large and can contain thousands of terms that need to be reduced using Boolean algebra.

A more detailed discussion of fault- tree analysis is given in the Fault Tree Handbook [USNRC, 1981] along with [Carlson, 1983].

For the Maine Yankee review, fault trees were developed for the front-line and support system that perform the two important safety functions. For the RCS integrity event, a fault tree was developed that consisted of an OR gate containing three inputs, as shown in Figure 2-4. These inputs are small pipe ruptures, a failure in one of the power-operated relief valves (PORV) causing i

it to remain open, and a failure of a main coolant pump seal causing a small LOCA. During the review, it was possible to screen out the PORY failures and I the pump seal LOCA because their components were estimated to have HCLPF-

! capacities greater than the review earthquake level.

The small pipe ruptures, however, could not be screened out. This was due to radioactivity concerns, because components within the Maine Yankee containment structure were unaccessible for review during the. walkdowns.- In particular, i_nstrument impulse lines that form part of the RCS pressure boundary could riot be inspected or reviewed. Therefore, these lines could not be screened out and were assumed to be a source of a small LOCA at the Maine Yankee plant.

This prompted the development of two event trees. . One that considered a small l

! LOCA concurrent with the LOSP and the other that considered the:LOSP with no LOCA.

T

l

~

2-12 i

___.~m,,,,,.,,,_,,,,.,,,_,m._7_,.-,,.,7., . , ,

e-- - -- , =-r-r- - , , , ,_,c m.. . . - _ _ , _ _ , _

i l

4 l

f Failure of l RCS integrity (small LOCA)

/

l

-)

l i

, Small Stuck Primary coolant pipe open pump

, break PORV seal LOCA l l

Figure 2-4 Fault tree for failure of RCS integrity.

2-13 -

'At the completion of Step 5, the systems fault trees were:" pruned" by removing those components that had been screened out in the previous steps. Care was taken when pruning the fault trees so that the paths from the remaining lower ,

level components were left intact. These lower level components represented the possible; failures for the systems under consideration.

A final . plant walkdown' is performed in Step 6. This- walkdown is used to obtain additional specific information for determining the HCLPF capacities of-

-. .the components that remain'in the analysis. In addition, the systems models

! are verified for accuracy and any additional information needed to complete

, them is obtained.

During the second walkdown of the Maine Yankee plant, the fragility analysts i collected detailed information about those components for which a complete i

evaluation was necessary. This information included detailed measurements of l

anchorages, equipment supports, and structural details along with gaining an.

j understanding of the general physical condition of equipment, their. mechanical i layouts including the peripheral components, and collecting other information needed to assess capacities.

The systems analysts collected information needed to finalize their systems models. This included discussions with knowledgeable plant operations and maintenance personnel. The information collected included details concerning scheduled test and maintenance frequencies and durations for the components, '

, and collecting plant data on equipment failures and repair times.

During Step 7, the systems models developed in Steps 3 an'd 5, and finalized in i Step 6, are analyzed to determine the Boolean. expressions for the seismic-l induced core damage accident sequences. This step involves the analysis of the event trees to determine the accident sequences that lead to seismic core  :

i damage and the analysis of the fault trees to determine the Boolean expression

1. for each system failure, i The development of accident sequence Boolean expressions ' follow in. the same manner as the development of the systems Boolean expressions. For the j

accident sequence, the systems identifiers are replaced with their respective Boolean expressions. The sequence expression is reduced and the minimal cut sets are determined. The fragility curves and nonseismic failures for the l basic events are then used to quantify the accident sequences.

A plant level Boolean exp'ression can be derived by logically combining all the core damage accident sequence Boolean expressions. The initiating events for- l these plant level accident . sequences are assumed -mutually exclusive. l l'

Therefore, the occurrence probability for each initiating event has. to be determined before the accident sequences can be combined..

i

This probability can be compared to the fraction of the time one initiating

, event occurs with respect to the occurrence of all the accident initiators.

Since we consider a fraction for the occurrence of each accident sequence, we multiply accident sequences by a factor called a split fraction. The sum of all the split fractions is 1 since we assume that an initiating event always

occurs. For example, - during the SSMtP study of Zion, the small. LOCA

,' 2-14 i

9 l

I T

e J

initiating event for a'0.3g earthquake was de'termined to occur about 1% of the time.

The plant; Boolean expression is then derived by multiplying each ' accident .

sequence by the appropriate split fraction and logically combining all the sequences. For the Maine Yankee review, two plant leveli. accident sequences

-were eventually developed. One for a LOSP initiator concurrent with small 4 LOCA and the other for a LOSP initiator with no LOCA.

The final step in the margin. review process, Step 8, is to calculate the HCLPF capacities for the important low-capacity. components, important systems,

{ accident . sequences,. and the plant. The HCLPF. capacities are finalized for those - components that appear in the single, double, and some low-capacity triple member cut set.of the. Boolean expression derived from the above systems analyses. These HCLPF calculations required detailed structural / mechanical L analyses based on information gained in - the previous steps. The fragility l' curves for the components .are then used to quantify the Boolean expressions

, for the system failures, accident sequences and the plant.

i l There are two methods available to calculate . the HCLPF capacity of

} components: the conservative deterministic failure . method (CDFM) and the

fragility analysis method (FA). For this trial review, the fragility method was used to calculate the HCLPF capacities for these components. This method.

'! was employed because the fragility analysis team has a detailed understanding of its application and use. In addition, the fragility method also allows the j inclusion of nonseismic failures into the overall plant HCLPF and accident-

! sequence HCLPF calculations.

j For the FA method, a component's fragility is represented by a simple model .

using three parameters: median capacity A and logarithmic- standard '

deviations So and Sn representing, respectively,,, randomness in the capacity and uncertainty i# the median value. Using a double lognormal model,

! fragility curves like the one shown in Figure 2 are developed. The median, 8,and 8 are estimated using design-analysis information, test data, earthquakv exper ence data, and engineering judgment. The median capacity may j be estimated as a product of an overall median safety factor times the SSE,

where the overall safety factor is a product of a number of factors

representing the conservatisms at different stages of analysis and design.

i When the scaling of response is not appropriate (e.g., soil sites), the median

! capacity is evaluated using median structural and equipment response

! parameters, median material properties and ductility factors, median capacity

, predictions, and realistic structural modeling and methods of analysis.

! The HCLPF capacity is expressed'using this fragility model as:

, HCLPF - A ,exp [-1.64 (SR + OU I3

• a 1 The FA method allows the combination of nonseismic and seismic failures for i the determination of plant HCLPF capacities. A more detailed explanation of i calculating component HCLPF capacities can be found in [ Kennedy and Ravindra,

! 1984].

1 l l 2-15 ,I l

! i I

. _ - . _ , . . . . _ , . _ ..,_m -

_ _ _ , _ , , , . . _,_m.. _.,-,,.-,-..,,,.m..

f 1.0 , i i '

95% confidence curve 2 -

2 0.8 -

Median fragility

_To Curve o

$ 0.6 -

=.O j 0.5 - - - - - - - - - - - - - - - - - - 0.90g o i

c. l -

0,4 _

m e

o l

= l 5c I 0.2 - I S HCLPF  ! 5% confidence I curve 0.05 l 0 -- - -I -} ' I I I I O 0.170.30.4 0.8 1.2 1.6 2.0 Peak ground acceleration (g)

Figure 2-5 Example fragility curves for a structure.

i 2-16 8

+-

--, . , , . - , , - , , , p-- , _ , - - _ _ _ _ . . , - - - . , - - _ , -, - - - - g

The unavailability of the components (failure per demand) is determined by combining its random failure probability with its unavailability due to normal test and maintenance. Equipment unavailability also included both planned and unplanned maintenance and repair. The unavailability due to random failures considers the time between normal test and maintenance, or between scheduled plant outages. A component's unavailability due to human error is considered as a separate event.

The component failures are combined following the rules of" Boolean algebra and a . discrete probability distribution (DPD) numerical procedure described by

[Kaplan,1981]. For this purpose, the component fragilities are discretized into a family of fragility curves. Each fragility curve is assumed to be completely described by the median capacity and the value of SR . The fragility curve is not truncated in either the lower or upper tail.

The " union" operation is performed for the first two components to obtain the composite fragility curves which are in turn combined with the next component fragility curves with either " union" or ' " intersection" operation in the Boolean expression. After each operation, the resulting fragility curves are condensed to keep tho computation manageable.

The details of the analysis are given in Volumes 2 and 3.

l l

l 2-17 h

CHAPTER 3

3. RESULTS OF SEISMIC MARGIN REVIEW An objective of this trial seismic margins review is to assess the capability of the Maine Yankee plant to withstand a specified earthquake level greater than the SSE. -The results of this review consist of:

o Boolean expressions for each seismic-induced core-damage accident sequence, o The dominant component for plant seismic safety, o An assessment of the HCLPF capacities for important components, accident sequences, and the plant.

o Insight into the seismic behavior of the plant systems required to fulfill the safety functions of subcriticality and early emergency core cooling injection.

The HCLPF capacities for the important components are logically combined as indicated by the Boolean expressions to estimate the HCLPF capacity for each core-damage accident sequence. Each of 'hese accident-sequence HCLPF capacities represents a plant HCLPF capacity for the particular initiating event and plant systems response.

A plant level Boolean expression can be derived by logically combining the accident sequences after they have been multiplied by their respective split fraction. An overall plant HCLPF capacity can then be determined from the plant level Boolean expression.

Section 3.1 presents a sumary of the overall result of the Maine Yankee review. A sumary of the systems results, including a discussion of the accident sequences, is given in Section 3.2. Details of the systems analysis are in Volume 2. A summary of the component capacity assessment is given in Section 3.3. Details of the component capacity assessment are in Volume 3.

3.1 Overall Results Following the review of plant information, a list of the components that make up the front-line and support systems required to perform the plant safety functions was developed. This list is given in Volume .3 and provided the basis for the remainder of the review. After the plant walkdowns and subsequent analyses by the systems and fragilities analysts, the list of components was reduced by screening out those components that had HCLPF values greater than 0.3g. For the remaining components, HCLPF capacities were calculated. The list of components for which a HCLPF capacity was calculated is given in Section 3.3.

3-1

r These remaining components were used in the development of 'the ' event trees and fault trees for the seismic-induced core-damage ~ accident sequences as-  :

described in Volume 2. The event trees. and fault trees were analyzed to I

determine Boolean expressions for each accident sequence that could lead to -

core damage. The component failures that are significant to these Booleans are given in Tables 3-1 and 3-2. Table 3-1 gives the seismic induced failures y along with the fragility parameters used to quantify their HCLPF capacities.

Table 3-2 gives:the nonseismic failures and their unavailabilities. Note that 4

the component items and nonseismic failure events are numbered consecutively; the missing numbers represent:the items that were screened out in the final pruning of the event and. fault trees.

One component has been ' upgraded following the review during the .two - plant walkdowns. This component (number 4,- Table 1) is a General Electric (GE) i station service transformer (4160 V to 480 V) that supplies power to pumps and-components needed to perform the seismic safety functions.

This transformer was originally installed with a " floating" bus bar to limit the amount of noise on the system. . Several years after its installation, GE performed seismic qualification testing on these types of transformers and made modifications that essentially resulted in securing the " floating" bus -

i bar and greatly increasing seismic capacity. Preliminary estimates of the transformer HCLPF capacity were approximately 0.lg. More detailed '

calculations could have been performed, however, . the utility decided to upgrade this component ' during its March 1987 : outage. An analysis ~ of; this planned upgrade has shown an increase in the estimated HCLPF capacity for. the i transformer to 0.30g.

I Another component appearing in Table 3-1 with a HCLPF less than 0.3g is Number 20, the seismic failure of the primary water storage ' tank (PWST). This tank provides an alternate supply of water to the auxiliary feedwater system. The 4

PWST has .a HCLPF of 0.279 This tank does not appear in either of the core damage Boolean expressions.

l 1

There is one type of component, not liste'd. in -Tables 3-1 or 3-2, for which '

there was insufficient data to determine the HCLPF capacity. This component.

is the lead-antimony type batteries used at the Maine Yankee plant. The four

4) have passed their electrical sets of Station Batteries (1, 2, 3, qualification testing. However, there is no seismic qualification or test 1

data available on these types of batteries to estimate the seismic capacity of-

< the aged lead-antimony plates within the battery casings.

Systems analysis indicated that the important batteries to plant seismic safety are Station Batteries 1 and 3. Maine Yankee had intended to change out~

i one of the station batteries during their plant refueling outage scheduled for

, March 1987. This analysis prompted them to replace both Station Batteries 1 and 3.- The remaining Station Batteries 2 and 4 have been scheduled for replacement during the 1988 outage.

1 3-2 e

i f

4 y r_-., --%e,--, . - - _ -----.--__y+ y- - - , - - ,,.-r - , , . , . _ . , . - . , _ . . . - , , _ , .-m-. , -

e ,e--., - , , -v - +-

s Table 3-1 Component' seismic _ fragility parameters.

HCLPF Item No. Item A,(g) S O U

Capacity (g)

R

-4 Transformers 0.84 0.30 0.32 0.30 (X507,X508) 7 RWST 0.45 0.20 0.25 0.21-(TK-4) 8 OWST 0.36 0.20 0.26 0.17 20, Circulating Water 0.69 0.24 0.27. 0.30 a Pumphouse 21* PWST 0.57 0.20 0.26 '0.27

_(TK-16) 1 i

• HCLPF less than 0.3g, but does not appear in the plant Boolean expressions.

f K

4 3-3 e

+  !

i I

-Table 3-2 Probabilities for nonseismic failures.

Median Unavailability- Error Item No. Description- (per demand) Factor

• 10 Operator Failure to Close PCC 8.0E-02 2 Isol. Valves 11 Random Failure of DG-1B 4.2E-02 5 12 Random Failure of DG-1A 4.2E-02 5 13 Operator Failure to Place AFW 1.5E-01 2 Pump Train B in Service Locally 14 Nonseismic Common Cause 1.6E-03 5 Failure of DGS 15 Nonseismic Common Cause 1.2E-04 5 Failure of AFW 16 Operator Failure to Refill DG 8.0E-03 3 Fuel Tanks by Opening Valve or Running P-33A,B 17 Operator Failure to Place AFW 4.0E-02 3 Pump Train B in Service from MCR 22 Random Failure of the Turbine 3.0E-02 5 Driven Aux Feedwater Pump Error factor equals (95% Confidence Value/ Median Value).

2.

l 3-4 l I

e

,- -, - ~ , ,, , , .- .-,

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

f1 .

' Seismic qualification and test data on the new batteries indicates. that they - &

have a HCLPF capacity' greater than the review earthquake level. Subsequently, .

the -station batteries were screened 1out ~and eliminated from further' consideration. .The new station batteries and racks should be inspected for

. proper installation according to Expert Panel guidance after they are in

- place.

The . capacity assessment of the new batteries being installed at the Maine Yankee plant ' assumes they possess the qualities of new equipment and are installed correctly. - However, as with the analysis of ~ all components during this seismic margin review, the effects of aging beyond this " snapshot" of the.

plant are not considered.

- The analysis of the two event trees resulted in two Boolean expressions that.

lead to seismic-induced core damage.- One of these expressions is the. logical combination of three accident sequences that were initiated by the seismic induced LOSP concurrent with a small LOCA. The other expression is the logical combination of two accident sequences initiated by the seismic-induced LOSP without a small LOCA. These two Boolean expressions are given below:

j Small LOCA Core Damage

- (SL) [4 + 7 + 20].

f No LOCA Core Damage 2 - (LOSP) [(4 + 20) * (8 + 13 + 15 + 17 + 22)

+ 8 * (14 + 16) + 15

• 7],

where the numbers in the expressions correspond to the failure of the components given in Tables 3-1 and 3-2. Entries with designators 4, 7, 8, 20, are seismic-induced failures and given.in Table 3-1. Entries with designators ,

from 10-17 and 22 are nonseismic failures and given in Table 3-2. The missing numbers in Tables 3-1 and 3-2 are component sereened- out during the final screening. The terms SL and LOSP represent the small LOCA and loss-of offsite power initiating' events, respectively. In the above . expression, _ the "+" ,

, notation denotes probabilistic - addition (union) and._the '"*" denotes  !

probabilistic multiplication (intersection). l l

The small LOCA initiating event could. be determined using the guidance given i by the Expert Panel. This guidance indicates that for this review earthquake j level, a walkdown of sample piping system should be, conducted along with  !

ir.spection of piping between buildings to look for'possible problems such as i I

n ' weak spots, unanchored attached equipment, ~ stiff sections between flexible i

termination points, and brittle connections.

8 During the walkdown of the Maine Yankee plant, none of these problems were identified and piping was found not to be a problem. Therefore, the small LOCA initiating event could be screened out- thus eliminating the small LOCA.'

d Boolean expression from the seismic margins review. However, the analysis 1

1 3-5 i

I i

s

_j l

teams . felt that eliminating the small LOCA' from consideration!was_not conservative. The possibility of a small LOCA should not be screened out because the containment, which contains many small RCS pipes, could not be 1 inspected and reviewed as previously discussed.'.

1 Impulse line failures were assumed to be the source of'a small LOCA at Maine

Yankee. ' This conservative assumption was required due to the tremendous

' number of. hours which would be required to walk down each of these impulse -

lines and. assess potential system interaction problems. These lines originate from the primary pressure boundary inside containment (i.e.,- RPV, steam generator, pressurizer, primary coolant loop piping, etc.) and are field

- routed to instrument racks inside containment. The amount of work required to j demonstrate the - seismic margin in each of < these lines plus the fact that

- walkdown of . these lines would have to take place during a plant -outage ~

' necessitated the assumption of a small LOCA as an initiating event.

Inspection of the small LOCA core-damage Boolean expression indicates that the dominant components are the three singletons 4, _7,-- and 20. The singleton

- component-with the lowest HCLPF capacity is the refueling water storage tank -

(RWST,- number 7) with a HCLPF capacity of 0.21g. Failure of this tank results in no coolant being available for reactor vessel injection following a small LOCA.

4 The other singleton components have HCLPF capacities of 0.30g. The capacity -

of the transformer- (number 4) is estimated based on the proposed upgraded 4 ,

2 condition. The other singleton is the circulating water- pumphouse (Number 20).

For the no LOCA core-damage Boolean expression, the: occurrence of the LOSP initiating event is assumed a certainty -(probability - 1.0)- due to the' low capacity of switchyard components. Therefore, the LOSP initiating event could be removed from the no LOCA Boolean expression.: ,

The HCLPF capacity for the no LOCA core damage Boolean expression is estimated- ,

1 to be greater than 0.30g. The higher capacity for this sequence, compared to t

, the small LOCA sequence, is due to the absence of singletons and 'no low -

2 capacity doubletons in the expression. Although the DWST, i.e., component 8, with a HCLPF capacity of 0.17g appears in this sequence, its failure has to occur simultaneously with one of the higher capacity components, i.e.. . the transformer or the circulating water pumphouse.

~

The most important nonseismic failure is number 15, a conunon cause failure'of the auxiliary feedwater system caused by steam binding. This failure results

.! in the inability to cool down the reactor _ coolant systems using the steam generators. This nonseismic failure is the most important because it appears i in a majority of the doubleton cut sets.  !

Figures 3-1 and 3-2 show plots of the small LOCA and LOSP core-damage i fragility curves, respectively, in which the family of fragility curves is l reduced to the 5%, 50%, and 95% confidence levels.

l 3-6 3

, f

_. - . , _ _ _ _ -. _ - _ _ . - . _ . . _ . _ - - _ .____. _. . . _ . _ . , . , _ .-.- -.. ~__.l

4

, Conf.

Ei 0.05 7 '

5 0.8 -

0.5 0 --

'I 0.95 o

E 0.6 - -

2i E

o Ei. 0.4 - -

3 C

.9 5c 0.2 - -

O

, o l I I O

O 0.2 0.4 0.6 0.8 1.0 Peak ground acceleration (g) l l

Figure 3-1 Fragility curves for small LOCA core v.smage 3-7

f 9

1 1.0 i , ,

$ Conf.

5 0.8 -

0.05 _

~ 0.50 a 0.95 _

E 0.6 -

3a

.O 2

Q. 0.4 - -

, c S

l

@c 0.2 - -

o O

I I I 4

0 O 0.2 0.4 0.6 0.8 1.0 Peak ground acceleration (g)

Figure 3-2 Seismic Fragility Curves for No-LOCA Core Damage.

3-8 G

- , - . . . - - - - - - - . , - - - - , - . . ,-.r . , - - , .

.,#ww--- - . . , .- - , , - - c.- - - -,--- -r , - ~ -- -r-- - .-c------+--- - . . . -

. ~ , . . , . - - - - - . . . -- -

w 4

v To account for the fact that we could not-~ quantify the small LOCA initiating

. event, we can develop an overall plant core-damage Boolean expression by .

logically combining the two Boolean expressions accounting for the split -

fraction p between the small LOCA and LOSP initiating events. This overall plant core-damage Boolean expression is given below: ,

i Core Damage

- p [small LOCA core damage] +

(1 - p) [no LOCA core damage]

In this case, the split fraction accounts for the fraction of the time a small LOCA will occur along with the LOSP event.

l

! When' evaluating the HCLPF capacity for the small LOCA Boolean expression, j there must be a consideration for the occurrence of the small LOCA initiating event. If the piping, whose failure will cause the small LOCA, has a HCLPF 4- capacity lower than all the singleton components in this Boolean expression, i the -lowest cepacity component (the RWST, number .7) will dominate the. small

!~ LOCA core damage HCLPF capacity. If, however, the piping has a HCLPF capacity l larger than the lowest capacity singleton, then the piping will dominate the small LOCA HCLPF capacity.

Overall Plant Core. Damage HCLPF Capacity As indicated above, to account for the two Boolean expressions in the overall-i plant core-damage HCLPF capacity, a sensitivity calculation can be performed

! accounting for a variation in the split fraction 'between the two accident-i sequence initiating events (small LOCA and no LOCA). For different assumed I split fraction values, the overall plant core-damage HCLPF capacities were j obtained as shown in Table 3-3.

I i The conclusion regarding the dominance of RWST failure in the HCLPF capacity J estimation (displayed in the small LOCA accident sequence) is a function of l the split fraction assumed. If the . plant HCLPF capacity needs to be l

increased, it is not necessary to concentrate only on RWST. A walkdown and review of small impulse lines within the containment may be performed to estimate their fragilities in order.to assign a realistic HCLPF capacity or split fraction. By this procedure, the plant HCLPF capacity may. be shown ,to .

be higher without the necessity of any upgrading of the components, f

Effect of Nonseismic Failures t

The overall plant HCLPF capacity was calculated using the Boolean expressions

• i for the core-damage accident sequences which contained both seismic (Table.3-J

1) and nonseismic (Table 3-2) failures. Since this is a seismic margin review F and the interest is only in the seismic capacity of the plant, one may choose to ignore the nonseismic failures in calculating the overall plant HCLPF In the small LOCA Boolean expression, there are no significant

, capacity.

j- nonseismic failures. The effect of not including the nonseismic failures on i the HCLPF capacity of the no LOCA Boolean had no effect on the plant HCLPF.

4 3-9

]

i 3

4

%, ,,.-m ~ - - ~ - - , r-w-..~. , ,,7,%%-w . ,., ,.e p-,-.,-,oeww.m.--,-w.w E. ..p.,.--m--- , . --

1

\

Table 3-3 ' Summary of plant;1evel HCLPF capacities. I 1

Case Description' HCLPF Capacity (g) l 1 Small LOCA .0.21 1

- Independent Seismic Failures i:

4 2 Small LOCA .- 0. 21~

i' 4

f

- ' Dependent Seismic Failures j 3 No LOCA $0.30 Independent Seismic Failures with

Nonseismic Failures l

4 No LOCA 2 0.30 il

- Independent Seismic Failures-

< without Nonseismic Failures 5 Core Damage Split Fraction p - 0.01 2 0.30 p - 0.10 0.28

. p - 0.50 0.23 6 Small LOCA

- RWST 0.26 Reduced Fluid Levei s

3-10 i

. . , -. ,. _ .._.- - ,,-.m. .. - . . . . . . . . _ _ . . - . . , , , . . _ . . . . . _ , - - ,

Correlation Between Seismic Failures

' The above calculations were per. formed assuming perfect. independence between seismic failures of different components; i.e., the seismic capacities are I assumed to be ' statistically independent both -in randomness and uncertainty.

This is a realistic assumption because -the-components involved in the core-

~

damage Boolean expression are yard tanks (RWST, and DWST), transformer and -

circulating water pumphouse. These are dissimilar items, their locations in the plant are different, and their dynamic characteristics are different. -)

Hence, correlation in the seismic responses and capacities .of these components is judged to be minimal.

It is realistic, however, to expect some correlation inithe component failures.. The. assumption of perfect dependence in both uncertainty and; randomness is an extreme case. = Assumption . of: perfect dependence in . the uncertainties of different component fragilities means that the median ground acceleration capacity of a component is known if the median ground acceleration capacity of another component is known. Since uncertainty arises from an insufficient understanding of structural material properties, approximate modeling of the structure, inaccuracies in the representation of -

mass and stiffness, . and the use of engineering judgment - in lieu of plant specific data, it is expected that all components wil1 ~ be affected to some degree by these uncertainties. Therefore, some probabilistic dependence between component median capacities may be expected.

Dependence in the randomness arises from a common earthquake generating the responses in different components and common structural / material properties.

Assumption of dependence in the randomness means 'that if the fragility (conditional probability of failure) of a component for a given peak ground acceleration is known, the probability of failure of the other components is somewhat modified by that knowledge.

The plant level fragility, and therefore, the HCLPF capacity depends on the degree of dependence in randomness and uncertainty between the component failures. However, the degree of dependence is difficult to estimate. One approach is to bound the core-damage HCLPF capacities by assuming perfect dependence as opposed to the case of perfect independence.' This calculation was performed for the small LOCA core-damage Boolean expression given above.

The small LOCA core-damage HCLPF ^ capacity is estimated to be 0.21g, i.e.,

governed by the capacity of RWST. If the seismic failures were assumed to be perfectly dependent and the nonseismic failures were assumed to be perfectly.

independent. The plant level HCLPF is estimated to be 0.21g.

When the Boolean expression is dominated by singletons, the assumption of perfect independence is more severe than the assumption of perfect dependence between failures if the fragilities are approximately equal; if there is a single . component with a very low capacity compared to the rest of the-components in the Boolean expression consisting of singletons, both assumptions give about the same plant level HCLPF capacity.

The question of dependence between failures is important when there are similar components experiencing common seismic excitation. The case in point 3-11

is the yard tanks (i.e., RWST, DWST, and PWST). By reviewing the cut sets, it was found that a tripleton cut set could lead to core damage. It is DWST

• PWST

• RWST.

Although high dependence between the failures of tanks is possible and their seismic capacities are similar, the cut set should have .a HCLPF capacity of 0.27g.

Deterministic Method This approach is based on the assumption that the HCLPF capacities of components are true lower bound values. The HCLPF capacity of t'ne plant is obtained directly by studying the Boolean expressions for small LOCA and no LOCA:

Small LOCA Core Damage

- 4 + 7 + 20 No LOCA Core Damage

- (4 + 20) * (8 + 15 + 17 + 22) 4

+ 8 * (14 + 16) + 7

• 15.

For calculating the plant level HCLPF capacity in this method, the nonseismic failures are ignored. Also, the cut sets that includes low probability nonseismic failures are also omitted from this estimation. .Therefore, the-simplified no LOCA Boolean expression becomes:

No LOCA Core Damage

- (4 + 20)

• 8.

In the deterministic method, the HCLPF capacity of a "doubleton" cut set is represented by the higher of the two component HCLPF capacities. The HCLPF capacity of a "tripleton" cut set is represented by the highest of the three component HCLPF capacities. The HCLPF capacity of a union of singleton cut sets is estimated to be the lowest of all the component HCLPF capacities.

Using this procedure, the HCLPF capacities of the core damage accident '

sequences are:

HCLPF Capacity for Small LOCA Core Damage

- min [0.30, 0.21, 0.30]

- 0.21g.

HCLPF Capacity for No LOCA Core Damage

- max [ min (0.30, 0.30), 0.17]

3-12 6

. , _ ., . - . _ . . -- _ _ ._.m . _. _,.-. -

- max [0.30, 0.17]

- 0.30g.

Note that the' failure of the DWST, with a HCLPF capacity of 0.17g, is not governing the plant level HCLPF capacity because the DWST has to fail simultaneously with one of two nonseismic failures to lead C ***9 Since these failures have median probabilities of 6 x 10-3 and 8 x 10-I, respectively, it is appropriate to ignore the DWST failure in the plant level HCLPF capacity calculation.

Sensitivity Studies Two sensitivity studies were performed to assess the effect of certain assumptions on the plant HCLPF capacity. These sensitivity studies addressed the following:

o Effect of shearwall stiffness reduction on structure response and equipment seismic input.

o Reduction of RWST fluid level.

For the first sensitivity, on-going scale model testing at Los Alamos National Laboratory being sponsored by the NRC has indicated potential reductions in the stiffness of concrete shear walls of up to a factor of 4, from elastically calculated values due to cracking. This would imply a reduction in the elastic structure frequencies of up to 50%. These variations were considered at the suggestion of the Peer Review Group, however the research results are still preliminary.

For no LOCA, the 4160/480-V transformer is the dominant contributor to the plant HCLPF capacity. The transformer has a 12-13 Hz fundamental frequency which corresponds to a spectral acceleration on the downward slope of the floor response spectra. A reduction of the building frequencies would result in a reduction of the seismic input to the transformer with a corresponding increase in its HCLPF capacity. As a further check on the effect of the building frequency shift, a similar evaluation was made on each of the components within the final Boolean expression. The potential building frequency reduction did not lower the HCLPF capacity for any of these components. Thus, for the purpose of the Maine Yankee seismic margin study, the shear wall stiffness reduction and resulting building frequency shift does not affect the plant' seismic margin.

For the second sensitivity study, the effect of reducing the fluid level within the RWST was investigated. - A reduction of the fluid level to an arbritary height of 33 ft, which is approximately equal to 10% reduction from the current level of 37 ft, was assumed. This change leads to a reduction in the effective fluid weight, mass, centroid height, and overall tank seismic loads. The RWST HCLPF capacity is increased to 0.28g with a corresponding increase of the small LOCA HCLPF capacity to 0.26g.

I l

3-13 1

- , =. . -_ - - - . ,-. .

n k'

R 3.2 Systems Results h

4 L A ' review of. the ' Maine-Yankee plant information resulted in identifying five i front-line 1 systems 1 that are required tolfulfill the safety functions ;of.  ;

reactor subcriticality and early ECC injection. These. systems are:

) o The ' reactor protection system -(RPS) including the control rod ' drive -

4 .

. system used to shutdown the nuclear chain reaction in the core and the core internals through which .the rods pass.

j o' The boric acid transfer system used to provide emergency boration to the reactor. system and also shut .down the- nuclear- chain reaction in j the core.

j; o The high pressure safety injection. (HPSI) system : used to . supply .

i coolant to the primary system.

I oi The auxiliary feedwater system used.to cool' down and depressurize the primary system through the steam generators.,

I o The pressurizer power-operated relief valves .(PORV) used in a feed and

! . bleed mode to depressurize the primary system.

The support systems to these front .line systems lare shown in matrix form in'

{- Table 3-4. The support system versus support system matrix is shown in Table t 3-5. . The components - that make up the front-line and support systems are identified and listed in Volume 3.

! During the review of the plant and as a result of discussions with Combustion -

! Engineering (CE, the NSSS supplier), it was determined that the components

.l that make up the RPS have.-HCLPF capacities ' larger than the review level

earthquake level based on data and calculations performed by CE. These

- components included the control rod. drive mechanisms, core internals,:and the l RPS actuation systems. Subsequently, the reactor subcriticality function was j screened out and eliminated from further consideration. Screening : out: the i reactor subcriticality function also eliminates the need for consideration ~of 4

the boric acid transfer system.

Systemic event trees were developed for (1) seismic-induced loss-of-offsite power (LOSP) concurrent with a small LOCA and (2) LOSP only initiating l events. Fault trees were developed for the front-line and support systems >

! using the identified seismic components that were not initially screened i out. The fault trees incorporated nonseismic failures including random  ;

i failure, coninon cause failure, and operator- error.- A- discussion of the success criteria, assumptions, and bases ' for the ' fault trees is given in Volume 2.

! When modeling the primary component cooling system (PCCS), the support system-

!- which provides cooling to one train of required front-line equipment,' .it was.

l discovered that alarge portion of the system cools nonessential shutdown

, equipment which is not automatically isolated following an earthquake.

i i 3-14

-_.,. ..._,.._ .,.--. _ ,.; ~ - _ ,. _ , _ ,...,_ ., _ .- ,__ _ _..- _ __._,.,_ ..; .-...,.. -, d

[a Table 3-4 Front-line system vs. support' system dependency matrix.-

SUPPORT SYSTEM 5

^

AC Power DC Power CCW SIAS IA FRONT-LINE Bus 5 Bus 6 DC-1 DC-3 Train Train SYSTEMS Bus 7 Bus 8 DC-2 DC-4 PCC- SCC A B TK-25.

HPSI/ P-14A RECIRC to RCS' .X X X X-P-14B to RCS- X X X X FN-44A X X

'FN-44B X X AFW P-25A X X P-25B X P-25C X X PROVs PR-S-14 X PR-S-15 X NOTE: To determine the front-line system dependencies on the support systems, locate the front-line component in the first column and read across the row to find the support system dependencies.

CCW - component cooling water SIAS - safety injection actuation system IA - instrument air 3-15 l

Table 3-5 Support system vs. support system dependency matrix.

% S.*RC N**?!

J!aev ses _ Jeof_ht. 878' 'l!!L*=5 85 IF5!.f=1 ._ CEV M _5885 qgac _ _i,ag 5HPP088 ST5IEN & 6 7 8 8 2 3 4 IA 24 8 2 3 4 NC SCC A 8 faa tas 38 3r Its AC 4 teet 5 3 I I I Pe=er tes 6 I I 3 3 400v 7 I I I I I I I hs 8 I 3

120W 8 2 3 vital Ses 3 3 4 I I I I Diesel IA a 3 seeerster le I I I DC lt5v I E 5 5 Pomer Ses  ! I 8 3 3 8 I I I ca 4 I

w PCC 3 I I I CCW O SCC I I I SM5 5 3 A I 5845 Cheneet 3 I I

C O I Ac t . 515-4 3 515-0 3 uvaC fn-20a B

1. 5-200 I IN-3t B is-32 I note: le detereise the support sFstes dependescles se other support systems, locate the support systne le the first celues. and read across the row to deterelse dependeetles se the other support systems. )

l d

I Therefore, leakage in these components, if not isolated, could compromise the entire system. Much of this equipment, located in the PAB and containment, .

was noi. inspected during the walkdowns due to accessibility. This consideration prompted the development of postearthquake procedural guidance requiring the isolation of nonessential portions of the PCCS in the event of a major earthquake and combined plant trip. The nonessential portions of the system can be isolated using valves which are remotely operated ;from the control room and have HCLPF capacity greater than the review earthquake level. This isolation of nonessential components has no impact on seismic safety. Based on this procedure change, the only remaining failure to be considered is the operator error associated with not following the procedure and closing the valves (component No. 10 in Table 3-2).

The event tree analysis resulted in five accident sequences that led to seismic-induced core damage.- Three of these are small LOCA sequences and two are no-LOCA sequences. The three LOCA sequences are designated S2D, S2LD, and S2LP2, and given below:

520 - HPSI S2LD - HPSI

• AFW S2LP2 - PPS-LOCA

• AFW, where HPSI, AFW and PPS-LOCA represent the high pressure safety injection system, the auxiliary feedwater system, and the plant pressure protection system operating for a LOCA condition, respectively.

For the seismic small LOCA, core damage will result if the HPSI system fails (520), or if both the HPSI and the AFW systems fail (S2LD), or if both the AFW system and power-operated relief valves (PORV) fail (S2LP2). For the LOCA case, both PORVs must fail since only one PORV is required to operate within approximately 30 minutes for feed and bleed. The block valve on each pressurizer relief line is included with its respective PORV.

The combined LOCA Boolean expression is the logical sunination of these three accident sequences.

LOCA Core Damage - S2D + S2LD + S2LP2 - S2D + S2LP2 The two no-LOCA accident sequences are designated TILD and TILP1, and are given below:

TILD - HPSI

• AFW TILP1 - AFW

• PPS, where the PPS represents the plant pressure protection system operating for the no-LOCA case.

For the no-LOCA case, core damage will result at Maine. Yankee if both the AFW-and HPSI systems fail (T1LD) or if both the AFW and one PORY fail (TILP1).

Based on PRA results, for the no-LOCA case, only one PORY must fail since both PORVs must open within approximately 30 minutes for feed and bleed.

I 3-17

__ _~ ._ . . _

P 5 ^ ,

i The systems . failures were determined i by - fault tree analysis. The' systems . l fault trees were developed and analyzed following an-11-step process. A brief g description of the 11 steps is given in Volume 2.

The analysis of.the fault trees gives the minimal cut sets that lead to system

~

failure. .The . single-failure '(singleton) and double-failure -(doubleton) cut sets for the systems' given in the accident sequences' are shown in Table. 3-6.

.The cut sets shown contain either . seismic-only failures or a combination of seismic and nonseismic' failures. ,

l .The HPSI system contains three seismic singletons and no seismic doubletons

.that lead to system failure. The AFW ' system. contains. no seismic singletons

!. and 10 doubletons. The PPS system contains. two singletons and no doubletons R i operating for both the small LOCA (PPS-LOCA) and no-LOCA (PPS) sequences.

I 3.3 : Fragility Results-

, The safety system components-identified by~the_ systems analysts were reviewed.

and subjected to three different types of screening.

! 1. Those components with 'a generic HCLPF value greater . than the review .

earthquake level, as given in Table 2-1 in [Prassinos et -al.,1986],

l were flagged for. inspection during the first plant walkdown.- The first' '

!' ~ plant walkdown confirmed that these components could be screened ~out based on their generic HCLPF values and the use of the screening table. Not all components with a generic HCLPF value greater than the

l review earthquake level were screened out after the first plant i walkdown.

j 2. The remaining components were examined following. a. detailed walkdown i review of their, seismic capacity during the second plant walkdown. .-If i the seismic capacity was ' judged to be greater than the- review j earthquake level, they were screened out.

3. The components remaining after the two screenings listed above were screened based on - a calculated HCLPF capacity. If their . calculated. i 4 HCLPF capacity was greater than the review earthquake level, they were

, screened out. Data needed to calculate the component HCLPF capacities .

were collected during both plant walkdowns, but specifically during the

second plant walkdown.

I j Both plant walkdowns were also performed to verify the configuration of the plant systems and to look for systems interaction and any plant unique

features..

1 . i l The components remaining after the first screening and those requiring support l

[ systems and operator action to perform the safety functions were used in the l

' development of the system fault trees. As 'the' subsequent screenings were '

performed, the system fault trees were pruned to reflect the elimination of j these components.

! ~

3-18 4 4

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

~

Table 3-6 Minimal cut sets for important systems.

High Pressure Safety Injection System (HPSI)

Singletons 1. 4160 V to 480 V station service transformer.

2. Refueling water storage tank (RWST).

3. Circulating water-pumphouse.

Doubletons 4. No seismic double failures.

} ' Auxiliary Feedwater System (AFW) i

. Singletons 1. No seismic single failures 1

Doubletons 2. Primary water storage tank (PWST), and demineralized water j storage tank (DWST).

3. 4160 V to 480 V station service transformer, and the DWST.

4. Circulating water pumphouse and the DWST.

5. The DWST and human error, failure to open AFW pump trains A i & B PWST isolation valves.

6. 4160 V to 480 V station service transformer, and human error, failure to place AFW pump train B (P-258) in service from the control room.

7. 4160 V to 480 V station service transformer, and random l failure of AFW turbine driven planp P-25B. ,

8. Circulating water pumphouse and random failure of AFW turbine driven pump P-25B.

9. Circulating water pumphouse and human error, failure to place AFW pump B (P-25B) in service from the control room.

s i

1 3-19 i

l l

Table 3-6 Minimal Cut sets for important systems. (Cont.)

Auxiliary Feedwater System (AFW) (Cont.)-

Doubletons 10. DWST and human error, failure to refill diesel generator (Cont.) tanks by opening valves and/or ranning auxiliary fuel pumps.

11. DWST and nonseismic common cause failure' of the diesel generators.

Plant Pressure Protection System (PPS)

(small LOCA and no-LOCA)

Singletons 1. Circulating water pumphouse.

2. 4160 V to 480 V station service transformer.

Doubletons 3. No seismic double failures, t

3-20' i

,., e xe,- - , - - - - ,. , - -, . - - . . , - . .n---

N The screening processes resulted in a reduced list' of components for which a HCLPF capacity .was calculated. This reduced list has been divided into two lists,' one consisting of structures and block walls, and the other consisting-of equipment. These two lists are given in Tables 3-7 and 3-8, respectively, along with the calculated HCLPF capacities.

i i

i f

i 4

t 4

I 6

6 4

3-21 r-1 i

4 w- + --r-m,----w-s-...,~.,-<-->mm4 ---.-9 ~- e . _= y -,.e r -r -,w T-+=-- -

w----- -r -:---~s- - - +T--m. = -'e e e,--~s--*"-ew-e-

4 Table 3-7 Maine Yankee Structures.and Block Walls HCLPF (g)

Structure Construction Capacity t

!; Circulating Water Pumphouse, . Structural steel framing and 0.3 steel portion above El 21'-0" diagonal bracing, concrete slab Service Building, El 39'-0"~ Structural steel framing and 0.38

+

-floor diagonal bracing, concrete slab ,

~

Main Steam Valve House, Structural steel framing and >0.3 interior steel structure diagonal bracing, metal grating t'

j Group A Components HCLPF(g)'

Capacity j Wall ID No. and Lifelines C 0.5-1 Pressurizer instrumentation >0.3 C 20-1 (PT-104, LT-106) and tubing

, SB 21-17 Main control board (aux feedwater >0.3-i system panels) l SB 21-18 Main control board (aux feedwater system >0.3 panels), aux logic panels 1

i SB 35-1 Battery groups 3 and 4, safety-related >0.3' i cable trays i

l SB 35-2 Battery groups-3 and 4, safety-related >0.3

]

cable trays

$ SB 35-3 Battery groups 3 and 4, safety-related' >0.3 j cable trays i

SB 35-4 Battery groups 3 and 4 >0.3

{ SB 35-7 PCC Surge line, PCC temperature controller >0.3 4

h i

1

! i j

-i 3-22 j l

4 1

,J , . - , , - - . . . . - - - - , , . . - , -,._,-.,--.-.--.~i--,--.... m, .,- , . -

-m_.., , - , ,.._ , ,v.., ...-,.--e._ ,e--,wm, - -.e,

TABLE 3-7 Maine Yankee structures and block walls (cont.)

Group A Components HCLPF (g)

Wall ID No. and Lifelines Capacity SB 45-1 125V DC distribution cabinets 1 to 4, 0.3 battery group 2, inverters #1 and #2, Bus 8 SB 45-2 Battery group 1, MCC 8A, 480 Y emergency >0.3 switchgear SB 45-3 Battery group 1 ' >0.3 VE 21-1, 2 Containment spray pumphouse Fans 44A >0.3 and 448, filter VE 21-3, 4 SCC line to penetration coolers >0.3 4

4 4

5 J

t 3-23 l

i Table 3-8 -Maine Yankee equipment list for margins review.

}

Building and HCLPF(g)

Equipment Item System Elevation Capacity TANKS

! Refueling Cavity Water HPSI Yd. + 20' O.21 l

Storage Tank TK .

! Primary Component PCC -SB + 61' >0.3 Cooling Surge Tank TK-5 Primary Water Storage AFW Yd. + 20' O.27 Tank TK-16

) Demineralized Water AFW Yd. + 20' O.17 j Storage Tank TK-21

! Spray Chemical Addition HPSI Yd. + 20' >0.3 J

TK-54

! Secondary Component SCC SB + 70 ' >0.3 Cooling Surge Tank TK-59

. Emergency Diesel Day 0F AB + 21' O.43 Tank TK-62A, TK-62B Diesel Compressed Air DG AB + 21' >0.3 Tanks TK-76A1, TK-76A2, l TK-76A3, TK-76B1, ,

i 1 TK-76B2, TK-76B3 Diesel Starting Air DG AB + 21' >0.3 i Receivers TK-76A4, TK-76AS, TK-76A6, TK-76B4, TK-76B5, ,

. TK-76B6 i

i PUMPS 3

Service Water Pump SW CW + 7 ' >0.3 P-29A, P'298, P29C, P-290 Containment Spray Pumps CS CS + 14' >0.3

!. P-61A, P-61B i

3-24 8

l f

I

, ,. ~ , _ . _ _ , _ . . . . . . , - - . _ , , . . , , - - - - , - . . . . _ . . ~ _ _ _ _ _ _ . _ _ - - - - . . - , . . . , , - ..._....--,.-.--,..._._..__._-n

Table 3-8 Maine Yankee equipment list for margins review. (Cont.)

Building and HCLPF (g)

Equipment Item System Elevation Capacity HEAT EXCHANGERS Residual Heat Removal PCC CS + 14' >0.3 Heat Exchanger E-3A*

Residual Heat Removal SCC CS + 14' >0.3 Heat Exchanger E-3B*

Primary Component PCC TB + 21' >0.3 Cooling Heat Exchangers

< E-4A, E-4B Secondary Component SCC TB + 21' >0.3 Cooling Heat Exchangers E-5A, E-5B Fuel Pool Heat Exchanger

• PCC TB + 31' >0.3 E-25 CEDM Coolers

• PCC RC + 46' >0.3 E-53-1, E-53-2, E-53-3, E-53-4 Reactor Containment

• PCC RC + 46' >0.3 Air Recirculation Coolers E-54-1, E-54-2, E-54-3, E-54-4, E-54-5, E-54-6 Charging Pump Seal

• SCC PAB + 11' >0.3

) Leakage Cooler E-92A Charging Pump Seal

• PCC PAB + 11' >0.3 Leakage cooler E-928 i

• Component whose failure may breach critical system pressure boundary.

3-25 1

4 J - v A Table 3-8 Maine Yankee equipment list for margins review. (Cont.)

Building and HCLPF (g)

Equipment Item System Elevation Capacity ELECTRICAL DISTRIBUTION SYSTEMS 4160-Y Emergency Buses Elec SB + 46' >0.3 Bus 5, 6, 7, 8 480-V Emergency Motor Elec SB + 46' >0.3 Control Center MCC-7A, 8A 480-V Emergency Motor Elec RMC + 21' >0.3 Control Center MCC-78, 8B 480-V Emergency Motor Elec CS + 20' >0.3 Control Center MCC-781, 8B1 i

Station Battery No. 1, 2 Elec SB + 46' >0.3 New Lead Calcium Batteries Station Battery No. 3, 4 Elec SB + 35' >0.3 New Lead Calcium Batteries Battery Chargers BC-1, BC-2, Elec SB + 46' >0.3 i BC-3, BC-4 Inverters INVR-1, INVR-2, Elec SB + 46' O.82 INVR-3, INVR-4 Station Service Transformer Elec SB + 46' O.30 X-507, X-608 (located adjacent to Bus 7 & 8)

Diesel Generator Control Elec AB + 22' >0.3 Panel 1A, 1B Main Control Board Elec SB + 21' >0.3 120-V AC Vital Bus 1-4 Electrical Control Board Elec SB + 21' >0.3 '

OG-1A & 18 Start 1 & 2 Circuits and Control Power

3-26 1

Table 3-8 Maine Yankee equipment list for margins review. (Cont.)

Building and HCLPF (g)

Equipment Item System Elevation Capacity ELECTRICAL DISTRIBUTION SYSTEMS (Cont.)

Auxiliary Logic Cabinets Elec' SB + 21' >0.3 ESF Auxiliary Panels A & B Elec SB + 21' >0.3 Air Condition Control Elec SB + 21' >0.3 Panel ACCP Safety Parameter Display Elec SB + 21' >0.3 System Cabinets j HVAC Computer Room Air Conditioner

• SCC SB + 39' O.38 AC-1A ,

]

Computer Room Air Conditioner

• PCC SB + 39' O.38 AC-1B Lab Air Conditioner AC-2* SCC Unknown 0.37 Containment Spray Fan

• SCC CS + 20' O.3 FN-44A, FN-44B -

VALVES Power-Operated Relief PORV RC + 66' >0.3 Valve PR-S-14, PR-S-15

Power-Operated Block PORV RC + 64' >0.3 Valve MOV PR-M-16, PR-M-17 .

• Component whose failure may breach critical system pressure boundary.

3-27 t

i

CHAPTER 4

4. INSIGHTS AND LESSONS LEARNED During the trial seismic margin review of the Maine Yankee plant, several lessons were learned about the various aspects of performing the review. In addition, many insights have been gained about the entire process of performing seismic margin reviews of nuclear power plants.

Insights and lessons learned have been gained on:

o Seismic margins approach and methodology including the quantification techniques.

o The guidance for actually conducting seismic margins reviews including plant walkdowns, o The level of effort needed to perform such a review.

o The identification and resolution of significant issues concerning the operation and licensing of the plant under review.

The purpose of this section is to highlight these insights and lessons learned.

4.1 Approach and Methodology (NUREG/CR-4334)

This section presents a brief discussion of the insights and lessons learned that pertain to the seismic margins approach and methodology. We gained a better understanding of the areas listed below:

1. The seismic margins review methodology should be able to be applied by knowledgeable engineers in the areas of earthquake engineering, component fragility analysis or systems analysis. They need to have some familiarity with these technologies.

Guidance on performing seismic margin reviews needs to be prescrip-tive and descriptive enough to guide knowledgeable engineers so that an accurate determination of the plant's seismic margin can be made without allowing over-conservatism into the analysis. ,

An independent review body should also be used to verify the application and accuracy of the review.

2. The fragility screening table needs to be strengthened. The actual meaning (representation) of the cut-offs between the table columns needs further explanation. In addition, more guidance in applying the table needs to be given.

3. More guidance is needed on selecting, establishing, and defining the review earthquake level (Step 1 of the review).

4-1 l

]

4y E V y There are -a number of waysiof specifying this review -earthquake level:

Uniform hazard spectrum at a. specified ~ annual probability of .

~

o-exceedance and confidence, o Site specific response spectrun evaluated for 'the specific ~

earthquake magnitude and epicentral distance ranges.

o Standard spectrum such as the NUREG/CR-0098.-

o_ Use of peak acceleration or mean acceleration.-

o Peak ground acceleration without_specifying any spectrum.

Knowledge of the target spectrum allows the -analysts' to make a more definitive statement about the seismic capacity -of the plant. The.

review earthquake must be completely specified -prior to any plant . -

review,.walkdowns or analysis.

4. The seismic margin quantification techniques need to be better.

defined. There are two competing . techniques for calculating the fragility of components, CDFM and FA. For this review, the FA '

technique was employed. However, the C0FM . technique would be more easily applied .if more prescriptive .and better defined.- More guidance is needed to use the CDFM technique. It is difficult to use as presently defined.

More guidance is needed on the evaluation of -plant level HCLPF-capacity including the use of nonseismic : failures and their significance to the results. More guidance is needed on methods for consideration of correlation between component failures.

5. More generic information is,needed-with respect to the safety function of reactor subcriticality. This function can be performed by either the insertion of the control rods, or' by boron injection.

Although the control rod drive mechanisms are screened out in Table 5-1 of NUREG/CR-4334 for review earthquake -levels - of 0.3g,; the reactor internals are not screened out based on insufficient information. This meant that both the reactor internals and .the boron injection system components were included in:the information gathering process before and during the' first walkdown. The boron injection system is fairly complex, with numerous components that are -

not initially screened out. Appreciable resources were expended in gathering information concerning the - boron injection system. This i, was unnecessary after the seismic capacity review of the reactor. ,

internals. It would probably be more efficient in future seismic i margin reviews if initial effort is placed- in verifying that the -]

reactor internals have high capacity, and only look at alternate l means of subcriticality if this 'is not the case. If necessary, this examination of alternate systems could be accomplished during .the  ;

second walkdown. j 4-2 i

. - - _ = . . . - - . . -- --

l 6. Guidance in NUREG/CR-4334 and 4482, based on evaluation of ' previous 1

PRAs' for PWR plants, states that the emergency core cooling (early) a function is - . included. in Group A,2 while the emergency core cooling j (late) function is in Group Not-A,'and therefore screened out of the

{ analysis. This screening is conditional.on not; finding any extremely

{ gross plant-specific differences.

f 4 The systems analysis team therefore included the initial switchover -

. phase from emergency core cooling injection (early) to emergency core cooling recirculation.(late) as-a' screening verification step in.the first plant walkdown. While the guidelines . are ambiguous, this.

i

!' screening verification included long-term area cooling -for_ the recirculation systems. It was determined that the containment spray pump area cooling fans FN-44A and B, and a block wall near the fans, j VE-21-1, could not be screened out based on the first 'walkdown.-

+ Based on the plant Boolean equation for small LOCA, both.of these items were single failures resulting in core damage in the long term.-

! The utility . will make changes to these items - to increase - their i capacity so that they do not impact overall plant capacity. Based on these findings, guidance in NUREG/CR-4334 and.4482 should be revised

to insure that- potential failures such as these are explicitly i evaluated during a seismic margins review.

1 '

I

7. For the no-LOCA case, the emergency core cooling (early) function is '

, defined in NUREG/CR-4334 and 4482 as achievement of residual heat removal. The AFW or EFW system at Maine Yankee or other PWR plants i will achieve this balance within the first hour. For most PWR

!. plants, irrecoverable failure of the.. emergency ac power system

, (station blackout) will not prevent the turbine-driven AFW train from performing early residual heat removal, and therefore satisfy the

!- emergency core cooling (early) function. However, in the longer term

' without ac power, the station batteries would be depleted, resulting in loss of instrumentation and AFW control power. Core damage could occur if de power is not restored,'and. manual control of the turbine-

! driven AFW train or other feedwater source fails. . Based on the

guidelines, this long-term failure of AFW was screened out of the

' analysis. If it were assumed that battery depletion' and ' loss of L L instrumentation and control power results in loss of AFW and other feedwater sources, then the Boolean expression for the no-LOCA core damage case would be dominated by the seismic failures that result in f station blackout:

1 i o Failure of the SCC and PCC heat exchangers E-5A and 48.

o Failure of the station service transformers X-507 and 608.

o Failure of the DG day tanks TK-62A and B.

o Rupture of SCC and PCC' because of chiller heat exchanger

failure for the air conditioners AC-1A, IB, and 2.

Structural failure of the circulating water pumphouse o

failing the SWS.

l i 4-3 i

i 1

' y!_ '

c Explicit guidance on the treatment of these long-term battery depletion sequences would be helpful.

8. The' Expert Panel reports have not explicitly discussed the seismic capacities of steel structures. At Maine Yankee, the steel i

1. structures have capacities in excess of 0.30g. At the same time, they could not be screened out based on a walkdown review because of '

unusual connection details and structural arrangement. More guidance is needed on the treatment of steel structures.

9. The [vailability of in-structure response spectra generated by current techniques can reduce the amount- of effort necessary to quantify component fragilities since modification or regeneration of ,

rerponses was not necessary. Accurate floor spectra may not always -

be available for older plants. In such cases, it may be necessary to develop new dynamic models and perform dynamic analysis to define the seismic input to equipment. This will increase the. amount of time and funding necessary to perform seismic margin studies.

/

10. Detailed information on the Maine Yankee block walls was available.

- This data was very useful in conducting the trial plant review since

it provided locations of all block walls in the plant and identified any safety-related components that could be affected by their failure. In particular, this latter set of information was valuable since it permitted quick identification of several Group A block -

walls.

1

11. From our review of Maine Yankee, we identified an additional plant l

unique feature. This unique feature is cast- iron service water piping. This cast iron piping at Maine Yankee was adequately supported and judged to have an acceptable capacity. l 4.2 Guidelines (NUREG/CR-4482)

This section provides a brief discussion of the lessons learned and insights gained concerning the guidelines for conducting seismic margins reviews.

1. Insights were gained about the planning and conduct of the physical plant walkdowns from our review of Maine Yankee. ,

l The most essential part of a plant walkdown is. preparation and '

i

, planning. All walkdown participants should help organize and plan

the specifics of the plant visit which are dependent on the nature of the walkdown and the specific plant under review. . The objectives of the walkdown should be clearly understood and arrangements for

1. -security, radiological safety, training, and plant operation should be considered.

Walkdown groups consisting of three to five people .with background i and expertise in various engineering disciplines should utilize pre-developed, inspection criteria and forms to collect information and data pertaining to the objectives of the plant. review. The groups i

l 4

4-4

< l

.J s

4 4

g g -,- - ,, -- -- A. w w --w- - . - - --- e,.,r+ , - - - - - , - . --n . , . , - - , ~--,n.-- , - - - - -

,., - ,n , - - - ~ ~ , ,

h l

1 4

should be' organized with respect to the information and- data that needs to be collected. Each group should contain a utility person 7

familiar with the plant layout, and the'. location and operation of the systems and components under review.

A clear understanding of what is poing to be reviewed and inspected l is also essential to the walkdown. 4.11st of the systems, components L

t and plant areas that need to be visited during the actual walkdown i should ' be developed. An itinerary will;also facilitate the actual  !

j plant walkdown. (

1 During the plant visit, meetings should - be arranged each day to organize the groups and prepare for the walkdowns. These meetings i

should allow time for oiscussion of findings at the ~end of each day and the preparation for the next day's effort.

Walkdowns can be performed in many ways, from a quick "walkby" of._the plant to a thorough " crawl through," depending on. the needs of each group, and the information and data collection requirements.- The types of walkdowns and their progressions should be planned 'in advance. Tools for the collection and recording of the information 4

j

' and data will also help facilitate the walkdown.

4

2. The identification of the front-line system components was relatively l

1 straightforward because of the detailed nature of the available information, and the small number of components. However, identification of support system components was more difficult. ,This

is because these systems are generally more complex, have many more '

4 components and branches, and are not generally documented as well.

In addition, the references concerning interfaces between the front-line systems and support systems are often ambiguous. Finally, the actual physical ' nature and location of some items, .such as distribution cabinets and panels, is not shown _on pl. ant drawings or

{

j documentation. Based on this ' experience, there are two recommenda-tions. First, when reviewing the plant information, emphasize the -

interfaces with support systems- such as ac . and de power, cooling _

water systems, HVAC systems, and instrument air systems. Document 4

the ambiguities for~ 1ater clarification. - Second, plan to spend considerable effort tracing down these support' system components during the first walkdown, and be prepared to make substantial revisions to the component list.

l 3. Another insight concerns documentation and information' transfer between the Systems Analysis Team and the Fragilities Team. Many of  !

3 f

the components identified by the Systems Analysis Team for HCLPF screening or evaluation were selected because of the potentia 1' for.

j component rupture to cause flow diversion and consequent system failure. The component itself was not needed to . fulfill a _ safety function, but the integrity of the component pressure boundary had to 4

be assured for overall system success. The common example,was heat  ;

exchangers for nonessential equipment .whose seismic rupture would i fail a necessary cooling water system. Since it can make a 4-5  ;

\

l-f . .

r i

difference' to the HCLPF assessment, the systems team aust make a clear differentiation between components that. are required to

. function for system success, and components that are required.only to maintain pressure boundary integrity.

4. There are a number of insights concerning the systems analysis and - I the fault tree pruning process which may be helpful to future seismic

margin reviews.

a .- The procedure proposed to isolate the PCC lines and components inside containment, and the automatic rupture isolation system on the SCC greatly reduced the amount of components that had to be considered for HCLPF evaluation.. Both the _ systems analysis and

.' fragilities analysis efforts _ would have been . larger, and '

eventually a containment walkdown might have been necessary.

b. Early evaluation and screening out of potential recovery actions and alternate systems, such as the small positive displacement pump for core cooling injection, reduced the number of components I that required systems and fragility evaluations.

c. Being able to define all the components on one skid as one supercomponent, such as the DGs, reduces the systems analysis -

effort, but the evaluation for the fragility team may not be 1

reduced.

As the fault trees are developed, it is useful to keep a list of -

~

d.

i the failure nodes which should be considered for. each

component. For example, the pump failure modes include fail to

. start, fail to run, and test or maintenance outage, as well as seismic failure, but the pump appears only once on -the initial.

, fault tree. This information is needed--later for the.

quantification process.

e. In the initial trees, it is necessary to include all the components that require support systems, including those components such as motor-operated or air-operated valves that will likely be screened out later because of their high HCLPF and low nonseismic unavailability. 0therwise, .if they are pruned

, from the initial trees, their dependency on support systems may_

be overlooked in the rest of the analysis. Also, since physical

interactions between the component and structures, such as block walls or restraints, must be checked,-it is better to include the-component in the initial fault trees,

f. In this project, those components and structures ' that were assigned a HCLPF of ">0.3g" either because they fit. the generic screening guidelines, or- because a conservative fragility

calculation was performed, were pruned from the . fault trees before the Boolean equations were. developed. The analysis team felt that their actual HCLPF would be well in excess of 0.3g, and that they would not . affect the plant seismic capacity 4-6 i

. .. .. -. , .V  ;-. . - . , . - --, - -. - - - - - - . . - - - - - . .i

l calculation. This pruning process .resulted in very manageable fault. trees, relatively small Boolean equations, and satisfactory calculations for the plant HCLPF. .It 'also enabled easier.

communication, understanding, and insight into the results than if less pruning had been performed, but provided more information on seismic margin than a more severely pruned Boolean equation.

The balance appears to be satisfactory.

Although a few seismic interactions, such as the possible impacts g.

of potential seismically induced fires, were not evaluated the potential for threaded firewater piping . ruptures ' to . damage equipment was reviewed. The DG control . panels and distribution panels were located under firewater piping that could have considerable lateral movement. Upon investigation,.however, it was determined that the piping was dry, and two signals would be The probability of inadvertent -

required to fill the piping.

actuation was therefore negligible. The PCC and SCC pumps were also located under sprinkler nozzles',.but.the motor housings for these pumps were designed to prevent -water from entering. -

' Therefore, ruptures of firewater piping was not considered to impact seismic capacity.

5. Minimal cut sets were developed . at four stages in this . project.

Front-line system level cut sets using partially pruned fault trees, including their support systems, were developed 1just before .the second walkdown to provide some guidance to the fragility team.

These pointed out some potentially important system minimal cut sets. Although plant or sequence level cut sets could have been of additional assistance, because the fault trees were still fairly large, the number of minimal cut ~ sets would have been large as well. The additional effort to develop plant level cut sets before

. the second walkdown is not judged to be an effective allocation of resources, but the effort to develop system cut sets is effective.

6. The seismic systems models of the plant (event trees and fault trees) i should be developed using-best-estimate success criteria based on the 3

FSAR and any relevant experience data. , The FSAR analyses, in a  :

sense, represents "high confidence" because of the regulations upon 1

which they are based. A best-estimate realistic success criteria )

i beyond the FSAR should be used if there are data and/or calculations- l to support their.use. .

7. Due to radioactivity concerns, critical components inside the Maine Yankee containment are accessible for review-only at the time of a plant outage. Fortunately, Maine Yankee has maintained a relatively complete data file on components within the contMnment.

Difficulties in reviewing plant components will = be more severe for plants which do not maintain organized documentation on the 4 components within containment.

~

More guidance is needed on the consideration of the seismic . small LOCA-initiating event. . It may be impossible to ~ review and inspect J

l 4-7 i

, , , , , , , _ ,%,_, y __ ,. ,,..

e 4

e the reactor coolant piping. (primary pressure boundary) throughout the plant, in particular within the containment, in order to screen out this possible accident initiator. For the Maine Yankee review, the small LOCA-accident sequence was explicitly considered in the systems analysis and was the controlling event sequence with regard to plant capacity.

8. During the walkdowns, plant and utility personnel having expertise or specialized knowledge in ' . particular fields were interviewed.

Specific areas of expertise included:

o Structural / mechanical, o Firewater systems.

l o Electrical.

o HVAC.

4 o Instrumentation and control.

o Control room personnel.

o Block walls.

The insight gained from discussions with knowledgeable individuals proved to be very useful .in assessing the overall state-of the plant and resolving any particular seismic issues.

9. Walkdown data sheets, while not noted in the review guidelines, were 4

' developed prior to the first walkdown. These data sheets served as a' checklist of items to review during the walkdown and facilitated the gathering of information necessary for HCLPF evaluation. While no

' set format need be established for future margin reviews, the review t guidelines should indicate that the use.of data sheets is encouraged.

10. Major distribution systems such as cable trays, piping, and ducting extend throughout the plant, there local configurations- and support details can vary. Detailed walkdowns of these systems can be time-consuming. More guidance in'the seismic margins methodology on the level of walkdown for these systems would be useful.

11. In the course of this study, five example components;(i.e., refueling water storage tank, steel structure,- diesel day.' tank', inverter and block wall) were analyzed using the two candidate methods (i.e., CDFM j and FA). Our experience was that several judgmental decisions had to c be made in arriving at the parameters of the CDFM method. In each case we were not sure whether we met the intent of the method, i.e.,

conservative estimation of the capacity, yet more liberal than the SRP requirements; in some cases, we may have been overly conservative.

. as was pointed out by the Peer Review Group'. The difficulties arise 4

because of two factors:

0 The CDFM method has not been fully defined for all l structures and equipment items.

l

! o The parameters of the CDFM method such as damping, material-strength, static capacity equations, system ductility, and ,

l 4-8

1 l

methods for floor spectra generation are not explicitly-specified; even where they are specified they may be overly conservative.- Also, the appropriate conservatism in the selection of the CDFM parameters needs to be determined using calibration methods.

We recomend that such a comparison study be performed.

12. This plant review examined a PWR on a rock site with a review earthquake level of 0.30g. The methodology, the review guidelines, and - the staffing requirements have not been verified for other conditions including BWRs, soil - conditions, and a higher review earthquake.

4.3 Effort to Perform the Review An objective of the trial seismic margin review is to gain an understanding of the level of effort, and thus cost, of performing seismic margins reviews of nuclear power plants. The trial review involved effort by these personnel:

the Systems Analyst Team, the Fragility Analyst Team, the utility or plant operator, and the management of the overall review.

NUREG/CR-4482 gives estimated staffing requirements for a seismic margin review; for a plant founded on rock with a review earthquake level of 0.30g pga, the estimate to perform the review is 2.5 staff-years. The actual effort expended in the present study is about 25% 'more than the Panel's estimate since this was the first review conducted.

Both analyst teams have divided their efforts into specific tasks. The tasks and the level of effort expended by both teams is given in Table 4-1. The Systems Analyst Team spent 1532 man-hours performir.g the review, and the-Fragility Analyst Team spent 4176 man-hours. The combined effort was approximately 35 man-months.

A breakdown of the utility effort is given in Appendix C. They spent 1060 man-hours for providing data, answering question, organizing walkdowns, and reviewing results.

4.4 Findings and Their Resolution A number of issues became apparent during the Maine-Yankee review. Primarily, these were findings concerning the seismic capacity of various components'at the plant. While many of these issues were discussed in the result section (Section 3.0)', this section is intended to briefly discuss all the seismic margins related findings and their resolution concerning the Maine Yankee -

review. A listing of each issue and brief discussion of the resolution are given below:  ;

1. Several types of analysis techniques can be employed to calculate the seismic capacity of free-standing tanks. For the Maine Yankee RWST these techniques resulted in the HCLPF capacity being greater than 0.20g. The lowest of these values is shown in Table 3-2.

i 4-9

The Maine . Yankee DWST, which is not a singleton to'the seismic safety of the plant, was found to have a HCLPF capacity of 0.17g.

J 4-10 )

J k,

l s

Table 4-1 Cost breakdown items for seismic margins trial plant review.

Actual Actual Fragilities Team Systems Team Review Items Staff-Hr Staff-Hr 1.0 Collect ~Information on Design 1.1 First Round Information 120 20 1.2 Additional Specific Information 80. 34 1.3 Visit to Utility /AE/NSSS Vendor 96 ' 8 2.0 Review of Plant Information 2.1 Review of Review Earthquake Level 40 --

2.2 Initial Systems Review --

60 2.3 Identify Components for -

Group A Functions --

36 2.4 Perform Initial Screening of Components 176 50 2.5 Review of Design-Analysis and Seismic Reevaluation Reports 216 --

3.0 Plant Walkdowns 3.1 Identify Target Areas for First

. Walkdown 240 50-3.2 Perform First Walkdown 184 84 i~ 3.3 Conduct Simplified Analysis 200 36 3.4 First Walkdown Documentatien 320 34 3.5 Perform Second Walkdown 200 26 -t 4.0 Systems Modeling 4.1 Develop Event and Fault Trees -- 539 4.2 Derive Accident Sequences -- 70 4.3 Develop Boolean Expressions and Minimal Cut Sets --

50 l

4-11 ,

1

i i

Table 4-1 Cost breakdown items -for seismic margins

! trial plant review (Cont.).

Actual .

Actual-Fragilities Team Systems Team Review Items -Staff-Hr Staff-Hr 5.0 Seismic Margin Evaluation 5.1 HCLPF Capacity of Components i 5.1.1 CDFM Method .

280 --

5.1.2 Fragility Analysis Method 800 --

5.2 HCLPF Capacity of Plant 160 10 6.0 Reporting 6.1 Internal Review 180 22 6.2 Letter and Final Reports 480 253 7.0 Meetings 7.1 Project Teams 224 104 ~

7.2 Peer Review Group 80 22 7.3 NRC/ACRS/ Expert Panel 100 24 4176 1532

.i 4-12 .

9

N 4

2. There is.no data or experience on the-seismic behavior of aged lead-antimony station batteries at the Maine Yankee plant. Therefore,.

i there was no way to estimate their seismic capacity and a HCLPF value.

Maine Y'ankee will have , these station. batteries replaced with lead--

~

calcium batteries _ during the 'next two refueling outages. Station Batteries 1 and 3 will be replaced during the next refueling outage .-

(March 1987) . and Station Batteries' 2.and '4 will be replaced during

- the 1988 outage. The HCLPF capacity of the station batteries was evaluated for the replacement units.

c.

3. Older General Electric station service transformers (4160 Y to 480 V) '

were not tested or qualified for seismic loads- and their seismic capacity is low. This component became the most dominant contributor to the seismic capacity of the plant prior to the upgrade.

GE had recent seismic testing and qualification performed 'on its equipment transformer. This testing.resulted in modifications. . Base i

anchorage modifications are scheduled .for the Maine Yankee transformers.

/

4. Anchorages on the containment spray pump area fans FN-44A and B will be strengthened. Failure of these fans could have led to long term heat-up and failure of the containment spray pumps, with subsequent I failure of high pressure safety recirculation if recovery actions

+ were not effective.

i 5. Block wall VE 21-1 will be strengthened to prevent- its potential collapse from failing the containment spray area fans FN-44A and B discussed above.

6. The anchorages of the chillers for the computer room air conditioners AC-1A and B and the laboratory air conditioner AC-2 will' be strengthened. Failure of the heat exchangers on these chillers could 4

have failed the pressure boundary integrity of_ the SCC and PCC, and l resulted in core damage upon loss of component. cooling water.

7. A procedure is being developed to isolate nonessential PCC lines ana l

heat exchangers following a large earthquake and receipt of a low level indication in the PCC surge tank. Although the PCC system was '

designed to seismic standards, the project teen could not verify the

capacity of all components. Isolating the PCC lines provides .

assurance that any potential'small leakage of the pressure boundary will not fail the entire PCC system.

8. An unanchored monitor in the main control room panel was anchored. -i Its impact on other components and the seismic capacity of the plant therefore did not have to be evaluated.

9. The emergency lights in the control room and throughout the plant

> were strapped and anchored.

4-13 l.

10. A missing bolt on the anchorage of. some level transmitters for the RWST was replaced.

11. Loose pressurized gas cylinders, a welding machine, and some heavy parts .near the containment spray pump area fans were moved or tied securely..

12. Additional anchorage was added to both diesel generator day tanks. .

1 4-14

l l

l l

CHAPTER 5

'5 . CONCLUSIONS The trial seismic margins review of the Maine Yankee plant was conducted with the concerted effort ' of -all parties involved. The analysis teams worked together closely and followed the guidance on performing the review .to estimate the overall plant HCLPF and the HCLPF capacities for.the accident sequences that lead to seismic-induced core damage. The Maine Yankee utility and Yankee Atomic Electric Company provided invaluable assistance in -

performing the review. Without their efforts, this review would have been~

much more difficult. The Expert Panel provided an initial review of the-

. approach early in the project. The Peer Review Group provided guidance and a critical examination of the process and interim results at each stage of- the review. The NRC assisted in the definition of the scope of the review and licensing issues involved.

The conclusions for this trial review are divided into two sections:

conclusions concerning the application of the seismic margins methodology, and guidelines including the numerical results and insights gained from this study are given in Section 5.1. The conclusions regarding Maine Yankee's licensing issue are given in Section 5.2.

i This seismic margin review has been performed with the following assumptions and limitations:

o The review earthquake level- was specified by the NRC 'as the '

NUREG/CR-0098 median spectrum anchored to 0.3g.

o The structural models and the' in-structure response spectra generated by Maine Yankee have been judged to be adequate for the purposes of this margin review.

o Since the Analysis Team could not perform the walkdown inside the containment, the seismic capacity of - - components inside the containment could not be determined. We could not confirm the

- absence of potential system interaction effects that may make the impulse lines inside the containment vulnerable to earthquakes and lead to a small LOCA.

t o In keeping with the Expert Panel's -philosophy, the screening of components was performed using conservative procedures. For the screened-in components, the seismic capacities have been calculated using conservative methods. In all cases, the factors contributing to the seismic margin and their variabilities are identified and i quantified using: procedures normally used within the state-of-the-art.

o The HCLPF capacity of the plant has been determined based on the seismic capacities of components in their existing or proposed' modified conditions. Maine Yankee has proposed that certain modifications or replacements would be made for station batteries, 5-1

, transformer internal core / coil assembly anchorage,- vibration-isolation supports for containment spray fans and air- conditioners, anchorage of diesel day tank, and block ~ wall near the containment spray fans.

o The results of this J seismic margins review . represents the best i estimate- analysis of the components and the plant following the proposed modification as a result of this review. No effort was made to account for the effects of future aging.

~

5.1 Conclusions ReGarding the Methodology, Guidelines, and Insights The conclusions from the trial seismic margin review include:

o The plant HCLPF capacity was determined'to be 0.21g. This capacity is dominated by the smal1~LOCA-accident sequence with the RWST.being the dominant component. There was no effect ' on the plant HCLPF capacity when - we considered the dependence between. component 1

failures.

An' arbitrary 10% reduction in RWST fluid level results in an increase in HCLPF capacity to 0.26g.

Assuming an arbritrary ten percent (10%) probability of occurrence of f the small LOCA-initiating event as compared to the occurrence of all the other possible initiating events in the plant, HCLPF capacity-increases to 0.289 o We found that careful plant walkdowns are essential to successful seismic margin reviews.

! o Insights and lessons learned (discussed.in Chapter 4) concerning:

- The selection of the review earthquake level.

The methodology and use of the review guidelines.

Component qualification data.

Plant walkdown procedures.

' - Guidance on the CDFM HCLPF calculation procedure.

i i- o The maintenance of hot shutdown following a seismic-induced initiator was considered by performing a thorough walkdown review and analysis of the components needed to perform this function. This led to upgrading of Fans FN-44A, B and the adjacent block wall.

o Important components and failure modes discovered during this review are:

i l

~:

5-2

.J

! .y 1

l *) ,

-,-u- ---a _,, __ __ -- -

r - - , -a w ,wm,, -.--my, , - - , q. m ,,. .q,., e ,wy,q,a., . ,,, 9 y q

Station service transformers (4160 Y to 480 V) require a review of their internal configuration.

Lead-antimony station batteries may fail due to the failure of the plates within the battery casing. No data are available.to estimate a HCLPF capacity for lead-antimony batteries.

- Consideration must be given to the location and possible systems interaction from threaded fire water piping, o Consideration must be given to modifications and upgrades identified during the seismic margins review process.

o The small LOCA-initiating event had to be considered for this review because of the difficulty in performing a walkdown inside the containment building.

o The walkdown review of closed-loop component-cooling systems may require considerable effort.

o The components that affect the reactor subcriticality function, in particular the control element drive mechanisms and the reactor internals, need to be considered early in the review.

5.2 Conclusion Regarding the Maine Yankee Licensing Issue The Maine Yankee plant was found to be clean and well-maintained. There has been a concerted effort to seismically upgrade the plant consistent with newer plants. The Maine Yankee utility and its contractor, Yankee Atomic Electric Company, were very cooperative with this effort and responded on their own initiative to increase the seismic margin of the plant.

A number of components were identified as having a low seismic capacity. A letter indicating that Maine Yankee will upgrade these components by the end of the next two refueling outages in March 1987 and 1988 is included in Appendix C. These components are listed below:

o Important station service transformers (4160 V to 480 V) o A block wall near the HVAC equipment (Fan 44A & B) needed to cool the containment spray pump enclosure. This enclosure houses the long-term cooling equipment.

o The PCC and SCC heat exchangers for both the computer ' room air conditioning compressors (chillers).

o Station batteries 1 and 3 will be replaced during the March 1987 outage. Batteries 2 and 4 will be replaced in the 1988 outage, o Upgrading the anchorage of both diesel generator day tanks.

5-3 m

u,--m a- +w+---

/

The plant HCLPF capacity after the planned upgrades was estimated - to be 0.21g. This HCLPF capacity is . governed by the RWST and represents a conservative estimate of the seismic capacity of the plant. . That is, given an earthquake producing this ground acceleration and specified spectral shape, there is high confidence (95%) that there is a low probability _of core damage (occurring only approximately 5% of the. time).

It is recomended that an inspection and review of Maine Yankee be performed after the March 1987 outage to assure that the proposed ~ modifications- have been carried out and that their HCLPF capacity calculated in the present study are still applicable.

4 4

5-4 1

l 1

l l

l l

CHAPTER 6

6. RECOMMENDATIONS TO IWROVE METHODOLOGY During the development of the seismic margins review methodology. and guidelines, and as a result of performing this first trial seismic margins. ,

review, a number of reconsnendations can be made for further analysis and l research that will improve the applicability and usabil;ty of this method for future reviews. These reconsnendations are listed below:

o Need to revise and clarify the seismic margins methodology and guidelines based on lessons learned so they are more usable, o Need to make a comparison of this methodology and its application with a similar effort being performed by EPRI.

o Need a comparison study that addresses the CDFM and FA methods for calculating HCLPF capacities.

o Need a study. of the available methodologies for calculating the capacity of older tanks, o Need a trial seismic margins review of a BWR plant to test enhancement of the methodology presently underway that will address this type of plant.

o Need testing of aged batteries to understand their seismic capacity. Additionally, to understand the internal plate failure mode for batteries in general.

6-1

- - . .. -- . -- - ~ . .

1 '

j' REFERENCES American1 Society of Civil Engineers (ASCE), "The Effects of Earthquakes on Power .and Industrial Facilities- and Implications for Nuclear Power . Plant '

i

- Design," Draft UCRL-93320 (1986). To be published. by .the ASCE.

Memorandum.. from.. Newton . Anderson,' .NRR Co-Chairman. Seismic Margins. Working -

Group, to Mr. James Richardson, RES Co-Chairman Seismic Margins Working Group, L " Comments on the Seismic Margins Program and the Trial Plant Review," July 7,.

1986.

Budnitz,- R. J. , .P. J. Amico, C. A. Cornell, W. J. Hall. R. P. Kennedy, J.' W.

Reed, and M. Shinozuka, "An Approach to the Quantification of Seismic Margins.

in Nuclear Power Plants," NUREG/CR-4334, UCID-20444 (August 1985).

! Campbell, R. D., M. K. Ravindra, A.' Bhatia, and R. -C. Murray,," Compilation .of c Fragility Information from . Available Probabilistic Assessments'," Lawrence.

Livermore National Laboratory, Livermore, CA, UCID-20571 (September 1985).

Carlson, D. D., " Interim : Reliability Evaluation Program Procedure l Guide,"

i NUREG/CR-2728, SAND 82-1100 (January 1983).

Memorandum from Dennis M. Crutchfield, Assistant Director .for. Technical Support . Division of PWR Licensing-8, to Patrick M. Sears, Project Manager, Project Directorate No. 8, Division of PWR Licensing-B, "Recomended Spectrum for Maine Yankee Seismic Design Program,":May 7,1986.

i' Cumings, G. E., J. J. Johnson, and R. J.- Budnitz, "NRC Seismic Design. Margins Program Plan," UCID-20247 (October-1984).-

Memorandum from: D. Guzy, MSEB,- DET, to NRC Seismic Design Margins -Working Group, " Maine Yankee Seismic Margins Review," May 12, 1986.

Kennedy, R. P. and M. K. Ravindra, " Seismic Fragilities for Nuclear . Power -

Plant Risk Studies," Nuclear Engineering and Design, Vol. 79, No.1, pp. 47-68 (May 1984).

Maine Yankee Atomic Power Company, Final Safety - Analysis Report- (FSAR),-

"Section 2.5, Seismology."

Letter from Frank J. Miraglia, Director, PWR Project Directorate No. 8, ,

Division of PWR Licensing-B, to Mr. J. B. Randazza, Executive Vice President, '

Maine Yankee Atomic Power Company, " Maine Yankee Participation in~. Seismic Design Margins. Program," March 31., 1986.

Newmark, N.- M. and W. J. Hall, " Development of Criteria for Review of Selected j Nuclear Power Plants," NUREG/CR-0098 (May 1978).

1 Prassinos, P. G. , M. K. Ravindra, and J; B. Savy, " Recommendations to the

' Nuclear Regulatory Commission on Trial Guidelines for-Seismic Margin Reviews of Nuclear Power Plants," NUREG/CR-4482, UCID-20579 (March 1986).-

R .

I i

i

l l

Oral Communication, Richardson, J. E., Office of Nuclear Regulatory Research,  !

U.S. Nuclear Regulatory Commission, on the clarification of the seismic margin review earthquake level for the Maine Yankee trial review (December 11,'1986).

U.S. Nuclear Regulatory Commission, " Design Response Spectra for Seismic -

Design of Nuclear Power Plants," Regulatory Guide 1.60 (December 1973).

U.S. Nuclear Regulatory Commission, " Fault Tree Handbook," NUREG-0492 (January 1981).

U.S. Nuclear Regulatory Commission, "PRA Procedures Guide," NUREG-2300 (January 1983).

p R-2

APPENDIX A EXPERT PANEL CORRESPONDENCE I

I A-1

I 4

INTERFACE BETWEEN-THE' EXPERT PANEL ON'SEISHIC MARGINS ~

AND THE TRIAL MARGIN REVIEW' 0F MAINE YANKEE

.1 R. J. Budnitz, Chairman

'The NRC's. Expert Panel on Seismic Margins developed the original approach to performing seismic margin reviews (NUREG/CR-4334) and' oversaw the developnent _

of_ interim guidance on how such a' review .should be carried out (NUREG/CR-4482). It is important that the Expert Panel have confidence that the trial

~

margin review is accomplished in a technically sound manner; this confidence will be assured by the Peer Review Group that is overseeing the trial margin - .

study at Maine Yankee.

The Expert Panel's role during the trial review will be very? limited.

Specifically, it is expected that the Expert Panel will review the approacn being taken at an early stage of the Maine Yankee trial review, to assure

itself that the methods and techniques being employed are consistent with the Panel's guidance and are relevant for performing a seismic margin review.

l This will involve a review of procedures prepared for the Maine Yankee review, discussion of these procedures on a conference call set up by the Chairman of 4

the Expert Panel, and a follow-up letter stating the Panel-members' opinions

, concerning the Maine Yankee review.

l After the early interaction, the Expert Panel will not be involved again until the trial review has been completed. At thatL time, the Panel 'will be convened so that it can study the results, examine how the trial review was

~

implemented, interview the study team, and then evaluate the overall effort.

The Expert Panel will be expected to re-examine its interim guidance and to l revise it, if necessary, prior to issuance in final form. The Expert Panel will also be expected to evaluate the overall usefulness of the seismic '

margins review approach, any limitations whether previously recognized or not, and provide any other relevant comments to LLNL and the NRC.~

! A-2 ,

. +

4 0

. - ... - - .- --, _-_-. - - - . - _ - , . . _ , - . . c . - . :. _ , - - , -

FutureReaurcesAmciates,Inc.

i 2000 Center Street - Suite 418 Berkeley. CA 94704 415-526-5111-i 3  ;

11 July 1986 l-j TO: Expert Panel on Seismic Design Margins (P. Amico, C.A. Cornell,

J. Reed, M. Shinozuka) f R.C. Murray, Lawrence Livermore National Laboratory FROM

R.J. Budnitz, Chairman of the Panel REF: Minutes of the Telephone Conference Call Meeting of June 27, 1986 This memorandum consists of the minutes of meeting of June'2'7 3 1986 of the NRC's " Expert Panel on Seismic Design Margins", whose membership consists of the individuals listed above. The ' meeting' was actually a conference telephone' call arranged by Bob Murray of LLNL, who is the NRC/LLNL liaison-for the Panel. In ' attendance' were all five members of the Panel; Murray

and P.G. Prassinos of LLNL; David L. Moore of Energy, Incorporated; and .

j M.K.Ravindra, P. Hashimoto, G. Hardy, and M. Griffin of EQE, Incorporated.-

The conference call began at 10:00 AM and ended at 12:40 PM, Pacific. time.

4 The objective of the meeting was to provide a forum for the Expert Pane!.to discuss progress to date on the trial seismic margin review that is nr ; underway under NRC/LLNL sponsorship using the Maine Yankee plant as the trial plant.

The review has been underway for tli months, and this was an ideal time to obtain Panel comments and input as to the progress being'made. As it turned-out, there were several technical issues that had arisen in the course of

the start-up phase of the trial review which benefitted from input from the Panel.

The ' bottom line'. outcome of the meeting was that the Panel was satisfied with the way the trial review was being undertaken so far, with the exception of a few (important) comments that will be covered in the remainder of these minutes. It is NRC's intention that the Expert Panel be involved next at the 4 stage when the trial review has been completed, at which time the Panel will

study the results, interact with the team of analysts, and then reevaluate

!. the guidance that has been provided in-interim form in the Panel's NUREG publi-cations on this subject, NUREG/CR-4334 and NUREG/CR-4482. (During the course of the conference call, the Panel estimated that about two person-weeks of effort would probably be required of each Panelist at this.later time to carry ,

out the review effectively, and this estimate does not count effort that may l be needed'to revise the interim guidelines.)

1) The meeting began with Bob Murray's discussion of the' progress to date of 1

the analysis team, whose key participants were all on the telephone line:

Pete Prassinos and Dave Moore representing the systems analysis team, and Bob Murray, Ravi Ravindra, and Ravindra's EQE colleagues representing the-fragilities team. Murray discussed the sechedule for the review, whose major.

j A f 1

-=.r- a n-v-- r.., . ..m-- -w, - , , - - , . - , - - < . .- ,~ n , , , - - - , s .--,s , - + . _ , ,----:- , - - , , , . - , .

4 page 2 ---- Minutes of Expert Panel _ Meeting of June 27, 1986

].

upcoming milestone will be the first plant walkdown on July 21-26, at the.

Maine Yankee plant. He discussed the interactions with the plant team, which interactions have been entirely favorable and cooperative so far; and the arrangements made for a ' peer review group' to follow the progress of the trial review. Murray's introductory discussion set the stage for the rest of the conference call discussion.

2) The Panel discussed the fact that the trial margins review at Maine Yankee-is intimately-tied up with~an'NRC licensing action. The Panel expressed regrets that this linkage was necessary, since.the Panel believes that this trial review should be viewed in-an important way as a research~ project whose outcome i

is by no means well understood until it has been accomplished. The Panel expressed a strong desire that NRC find a plant (expected to be a BWR) for the planned second trial review which does not have a licensing action tied to 1

the seismic margin review.. .

3) The Panel discussed the peer review arrangement for the trial review, and agreed.that the set-up was consistent with the guidance provided in its earlier NUREG reports. The peer review group's charter is to assure the technical com-petence of the review, and to report on its findings to NRC and Maine Yankee.

.4) The Panel discussed at length the selection of the ' review level earthquake'

! , (henceforth abbreviated RLEQ) by NRC, which s' election-is at the 0.30 g level, with a spectrum using the 50th percentile amplification factors from NUREG/CR-0096.

A question arose as to how to cope with motions at higher frequencies than those in the more-or-less standard spectrum chosen. The Expert Panel reaffirmed i

its earlier position that a s_eparate research project-is needed to provide in-formation to the Panel before the Panel can develop margin-review guidance on i this issue, and that higher-frequency motions are explicitly not included in the guidance already developed in the earlier NUREG reports. '

5) Regarding the choice of the RLEQ, the Panel agreed that especially for a trial review like the one being undertaken at Maine Yankee, it would be preferable j

if the RLEQ were selected to be at a high enough level that the ' plant HCLPF value' can be affirmatively determined through the review, rather than having

' the review determine only that the plant HCLPF value is "at least as high as the RLEQ 1evel." The choice of 0.30 g for Maine Yankee's RLEQ does not obviously-i meet that criterion, although it may (after the fact) turn out that way. The Panel agreed to reconsider at a later date whether it is necessary to provide e

more detailed guidance on 'how to select the RLEQ'.

. 6) The Expert Panel, after much discussion, agreed that it would be unfortunate if the fact that a given plant's HCLPF turned out to be less than the RLEQ were considered a ' negative' outcome for the review. Considering the conservatisms 1

embedded in the HCLPF idea, a plant's median . capacity would be expected to be

} considerably greater than its plant-level HCLPF value.

i j 7) The Panel discussed the meaning of its categorization in NUREG/CR-4334 and i

NUREG/CR 4482 of fragilities into three groupings, "below 0.30 g", "0.30 to 4

0.50g", and "above 0.50g". In its earlier reports, the Panel pointed out that i

these boundaries were not to be taken as being too precise: specifically, the Panel stated that the 0.30g boundary might just as easily have been a rough range from about 0.25g to about 0.35 9 . While this Panel judgment still stands, A-4 i

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

=

'1

.i

. i 1

. page 3 ---- Minutes.of Expert Panel Meeting of June 27,'1986 the calculated HCLPF values are not to b'e interprated in the same'way....that- . ,

is, if a plant-level HCLPF value is determined by the analysis to be, say, 0.30 :9  !

the Panel believes that it is not correct,to believe that the HCLPF value might just as easily have been anywhere in the range.of, say, about 0.25g to about 0.359 . '

The conservatisms embedded in,the "HCLPF" concept, through the concept of 'high

. confidence' and the concept of ' low probability of failure', should allow the regulatory decision-maker to cope with any technical. decision without the additional

' smearing' of placing.an analyzed HCLPF value in a broad range like that just men-tioned. The Panel was explicit and strong about its insistence on this point.

8) The Expert Panel provided guidance to the margins review team that, even though categories or specific items of equipment are thought to possess quite high capacities compared to the RLEQ it is still necessary to do at least some type of review of these items to confirm that they are not ' outliers'. The cri-teria in the Panel's chapter 5 discussion (NUREG-CR-4334) are considered by the Panel to be ' generally conservative' but not ' absolutely conservative in'every case.' Specifically, every screening decision must be confirmed somehow. '

9) The Panel expressed its strong disappointment that there has been no start to date on the separate research project that was recommended to compare the CDFM (conservative deterministic failure margin) method and the FA (fragility analysis) method for calculating capacity values. The Panel was told that in the Maine Yan-kee review Ravindra's fragility team will do such a comparison on about a half-dozen items. This outcome, in which all of the capacities are to be determined. .

using the FA method and the CDFM approach used only as a ' trial' or ' check', is.

certainly not what the Panel had envisioned when the guidelines were being developed.  :

The trial review approach is not a substitute- in the Panel's opinion, for a sepa-  !

rate study to confirm the adequacy of the CDFM method. The Panel expressed a  ;

strong need to go on record that the CDFM method needs to be confirmed. Although ,

the Panel believes that the CDFM approach is ultimately the preferable method, it also believes that CDFM should not be used and relied on until confirmed. The Panel discussed possible dangers of two kinds: first, the CDFM method might be used prior to its being studied and confirmed; and/or second, it might never be used because it is not a confirmed approach, thereby leaving margins analysts with no choice but to use the much more expensive and less desirable FA n.ethod.

10) The Panel was told that it is very unlikely that Maine Yankee's reactor internals and control rod drives can be reviewed to obtain useful fragility .

information, because they are inaccessible and the vendor. design / test infor-  ;

mation is unavailable. The Panel's guidance was that in-such a case it-is preferable to do a detailed study of the back-up reactivity control sysems such as the borated water tanks and plumbing, rather than to leave the issue open,  ;

even though there is high confidence on a generic basis that the' capacities of the reactor internals and CRDs are above the RLEQ selected for Maine Yankee.

11) There was extensive discussion on the safe-shutdown end-point used by the Panel. The Panel reiterated that it used the traditional PRA approach: for transients a stable state is typically achieving hot shutdown plus holding it for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> (sometimes 36 or 48), while for LOCAs a stable state is usually cold shutdown. The Panel did not explicitly use ' hot shutdown plus 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />'. . ,

The Panel reiterates its finding, based on the extensive PRA literature for PWRs, that it is-very unlikely in a probabilistic sense that the systems being screened in the seismic margin method proposed would have a high HCLPF value while the systems needed for post-hot-shutdown heat removal would have a quite low HCLPF A-5 l

_s z.-a-e ,_ a.-_sau .s ._ r sm - a w_ s. a .

...+q e -

. . . _ . . . , . + n. - .

2

• 4 i

!- page 4 ---- Minutes of Expert Panel Meeting of June 27, 1986;

d.

l l

value. Therefore, based'on.its study of the PRA literature the Panel has concluded'that it is not necessary to do.a detailed margin review that includes j

the systems supporting the function of post-hot-shutdown heat removal. I

' 12) Notwithstanding the above, the Panel reaffirms its earlier guidance that it is necessary to do a modest review of these and other systems to assure that there are not ' plant unique features' in the sense of the discussion in NUREG/

CR-4334 and 4482.

13) The Panel also reaffirms'its earlier approach to non-seismic-induced failures or unavailabilities, which approach is that these are to be combined with-seismic-induced failures using an. approximate probability-based cut-off criterion that. relies on expert judgment (see NUREG/CR-4334). There is no precise numeri-- ,

i cal guidance on how to apply this probabilistic cut-off.

14) Regarding seismic-initiated failures, the Panel was told that it would not.

be possible to enter the Maine. Yankee containment during the upcoming review.

' .Therefore it will not be possible to rule out small LOCAs inside containment caused by. earthquakes like the RLEQ. The Panel agreed with Dave Moore (leader of the systems analysis subcontractor team) that an appropriate approach is to i

do the analysis two ways ---- one way assuming that,a small LOCA inside contain-ment does occur, and the other way assuming that .it does not occur. ' Mitigating systems (injection pumps, etc.) will be studied to gain engineering insights as j to the plant's possible vulnerabilities, if any.

15) The Panel agreed that, at the RLEQ selected, the traditional large LOCA pipe breaks were r.ot a problem, on a generic basis. The Panel was told that Maine Yankee has large loop valves in the primary piping which were placed there in the-j original design to isolateone of the main loops. These might be more~ vulnerable i to the RLEQ than the primary piping itself. The Panel's guidance was to treat these as ' plant unique features' and to study them through design information if

} that was the only available approach.

16; Panel discussed at length the difference between the Panel's and EPRI's '

approaches to the systems analysis. In a nutshell EPRI's approachLis to search 1 for one ' success path' at the RLEQ, while the Panel's approach is to search for i- cut sets representing combinations of failures. The Panel reaffirmed its earlier conclusion that the cut-set approach is preferable since it provides more exten-sive engineering insights, but agreed to revisit this issue after both the EPRI l trial review of Catawba and NRC's trial review of Maine Yankee are completed.

'! 17) Bob Murray reported to the Panel about the excellent cooperation being re-i ceived from the EPRI margin study effort, which is doing a parallel review of

! Catawba. The Panel wishes to encourage the maximum cooperation, since both groups have already learned from each other and this will likely continue to be true.

$ 18) The Panel was informed that the BWR systems study is now underway, which '

should enable the Panel to develop BWR guidance sometime.next fiscal-year. The 4

} Panel was urged by Bob Murray to give thought now to possible nominees of a BWR l- candidate for the second trial margin review.

3 A-6 '

i i

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

=

page 5 ---- Minutes of Expert Panel Meeting of June 27, l'986 1

' In conclusion, the Panel's main finding based on the conference call meeting 1

of June 27 is that the trial margin review at Maine Yankee is proceeding appropriately, and that the Panel's overall guidance is being followed, with the exception of cur comment on the CPFM method (see item 9 above). The  !

Panel is very pleases otherwise with the progress so far.

i Robert J. Budnitz Chairman, Expert Panel i

! )

i

• e 4

f i

i i

i e

i I

b 1 i

A-7

) _ . _ _ . . _ .. . ___ _.- _ _ . . _ _ . _ . - _ - . . _ - _ _ _ _ _ - . . . . _ . - _ . _ _ _ - _ .

rr +

d Applied Risk Technology Corporation  !

P.O. Box 175, Columbia, MD 21045 (301) 992-0525 y  % D Date: 28 July 1986 To: R. C, Mur ay, L-197, Lawrence Livermore National Laboratory From: P.

/ mico, Applied Risk Technology Corporation Subj: P.O. 9225705, Comments on Maine Yankee Trial Plant Review for Seismic Design Margins Program This memo constitutes the final d eli v er abl e for th e subj ec t purcliase order, and is assumed to be sufficient documentation for the tasks performed. As you are aware, I reviewed the procedures developed by the NRC .for the Maine Yankee trial plant review and pa rtci pa te d in the conference ca l l held on 27 June 1986. In Beneral, my comments on the trial plant re view are f ully recorded on the tape recording of that meeting and adequately represented in the minutes of the meeting produced by R. J. Budnitz and dated 11 July 1986. Insofar as this is the case, my comments will not be repeated in this report. However, I would like to make an addition / clarification to my position which was not included in the recording or minutes.

First, regar ding the cursory re view of systems not part of the Group A functions (not-A systems), it should be clearly understood that the pu rp os e of r e vie wing these systems is to identify any plant unique f ea tu re s which indicate that these systems are significantly more fragile with respect to earthquake than the G ro up-A s ys tems. Under no circunstances is this to be construed as demonstrating that these not-A systems have a HCLPF greater than the RLEQ. By way of il lus t r a tio n , if a tank is located such that its rupture would cause failure of only not-A systems, this would be included in the re view. Similarly, if for some reason there is similar equipment in both Group-A and not-A systems, but the not-A e qu ip me nt has anchorages which are noticably inf erior to the Group-A anchorages (easily recognizable during a walkthrough), this would be - in c lud ed in the re v iew.

However, if the above deficiencies existed for both the Group-A and not- A sy stems, the ef f ect on the not-A sy stems would not be co nsid er ed. Again, the purpose of the re view is to determine only that the susce ptibilities o f the not- A s ystems to seismic events are Benerally of the same order as the susceptibilities of the Group-A systems, not whether they meet the HCLPF requirement.

The re view is very cursory in natu re, and anything which cannot be clearly identified visually during the initial plant walkthrough is not to be pursued further.

O A-8 C

i

page 2 - Comments on Maine Yankee-Trial Plant' Review l g Second, I wish to strongly express my disappointment 'with the i

decision that the expert panel would not constitute the. peer re view team, as was originally in tend ed. 'In my opinion, the expert panel as a whole is much better suited to oversee'the implementation o f. its re view ~ method and- to assess its affectiveness and the required modifications than the group which has been selected as a peer re view' team. I' wish to. ma ke ' clea r that I will not " rubber stamp" the conclusions of the peer' review team with regard to their assessment of the trial. review process

-and guidelines, -or blindly appro ve their suggested changes to the

methodology based solely on ' their opinion of what changes!should be made. However, it will obviously be extremely difficult ' f or .

me to formulate an informed opinion-on the validity of. their peer

. review comments while not having any involvement in the trial plant . re view or its peer re view process. .The resolution. to this-dilemma is not clear. In the ex tr eme cas e, it may be necessary for me to dissent .from .the peer re view team's conclu sion s j regarding the utility of the trial review guidelines.. ,

l.

i j i 1

l 4 l l

i l

1 h

4 A-9 l

l . , ,

[.

g .. hi8u/k'N

. 'o UNITED STATES i

! o NUCLEAR REGULATORY COMMISSION E,

.j waswiNcToN. D. C. 20SSS

• s.,

• / auus an Dr. R. Murray LLNL (L-197)

P.O. Box 808 Livermore, CA 94550

References:

1. 7/11/86 letter from R. Budnitz " Minutes of the Telephone Conference Call Meeting of June 27, 1986".

2. 6/18/86 memo from D. Guzy " Seismic Design Margins Program".

Dear Dr. Murray:

Reference 1 (enclosed) discusses the review of the Maine Yankee seismic margins review procedures by the Expert Panel on Seismic Design Margins. The Panel also discussed topics brought up by the NRC staff at our June 10, 1986 meeting held here at the Nicholson Lane Building. I would like to have clarification of the following items from Reference 1.

General: In Reference 2, I noted that I felt the review team's written procedures lacked depth, but that from the June 10th presentation it was clear to me that the team had definite ideas that would be finalized (and should be documented) as their review progresses.

Please document whether or not the Expert Panel felt the review procedures had sufficient detail. If not, how did the panel make their judgment on the ade-quacy of the procedures (on subsequent verbal discussions?) and what should be added to future written procedures?

I Item 4 This discussion seems to back away from a previous Expert Panel position that f for high frequency, high acceleration, low to moderate magnitude earthquakes, ,

their guidelines were conservative. Please clarify if the Panel's position has

• l changed.

l item 7 <

{

Although this discussion addresses an issue raised at the June 10th meeting, I'm not sure it directly answers a specific question. That is, in performing a margins review, can we use the values in the first column of Table 2-1 of  !

NUREG/CR-4482 and make a statement about the ability of the plant to withstand l earthquake up.to the .3g level?  !

A-10 t

JUl. A6 gd6 Item 9 I recommended that we delay the start of a separate study on CDFM vs. PRA fragility for the following reasons:

1. The end date for the proposed study was such that it could not be used in the Maine Yankee review.

2. EQE's proposal presumably was going to use both methods and thus there would be redundancy if a separate study was run concurrently.

3. FY 86 funding limitations wouldn't permit this study along with the other proposed activities.

It's a particular sore point with me that after repeated requests for more detailed 189 work statements and review procedures, the extent of EQE's CDFM/PRA fragility comparison was not stated until I asked the question at our June 10 meeting. Now I see that the Panel feels EQE's use of the PRA fragility approach is "less desirable" and that the number of CDFM examples is presumably inadequate to make any kind of a comparision. In light of these comments. I wonder ififfope of EQE's proposed use of the CDFM and FA methods was known and was a factor during LLNL's subcontractor selection process.

It would be useful t'o know what kind of study the Panel had in mind for

" confirming the CDFM method". It seems to me that building on the current Maine Yankee review would be the most efficient way to go.

I Lh Dan Guzy Mechanical / Structural Engineering Branch Division of Engineering Technology

Enclosures:

As stated cc: h v. t. nichardson N. Anderson A-11 .

f

\

FutureRuourcu A.uociatu,Inc.

2000 Center Street Sulte 418 Berkeley. CA 947(M 415-526-5111 September 26,1986 Mr. Daniel J. Guzy Engineering Branch Division of Engineering Safety Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D. C. 20555

Dear Dan:

This letter is a reply to your memorandum of July 16,1986 in which you raise several questions about the seismic margins program, both in general and in particular regarding the trial review at Maine Yankee.

I will address the points one-by-one, going down your memorandum. Therefore, a copy of your two-page memo is attached to thir. letter for convenience.

1. Your General Comment: The documentation of the NRC/LLNL review team's procedures has been insisted on by LLNL, and my understanding is that this documentation will be an integral part of the " deliverable" package that the review contractors will provide. The Expert Panel did not comment in detail about whether the procedures had sufficient detail, since at the time of the Panel's conference call of June 27 these procedures were still being developed. However, I am certain that the Expert Panel will comment on them later, because these procedures are intended to be a model of what should be followed by a review team implementing the " final" procedures to be finalized next year.

2. Your " Item 4" Comment: The Expert Panel believes that a separate study is required of the issue of high-frequency, high-acceleration, low-to-moderate-magnitude earthquakes. The Panel believes that its current guidelines are conservative, but that a separate study is needed to reaffirm this, to clarify areas where the conservatism may be inadequate (or may be excessive), and to understand this issue better.

3. Your " Item 7" Comment: This issue should be clear already, but I can clarify it again, by restating what the Expert Panel believes. The Panel believes that the first column should be used for " review level earthquakes" up to 0.30 g; the second column if the review level earthquake is between 0.30 and 0.50 g; and the third column for above 0.50 g. Suppose that the review-level earthquake is selected as, say, 0.29 g, so that the first column should be used. If, using the first column, a plant is found to have no Booleans (loosely, " accident sequences") that are A-12

. l t

Mr. Daniel J. Ouzy -/- September 26,1986 vulnerable, then the conclusion is that the plant-level HCLPF is at least as large as the review level earthquake being reviewed against. If on the other hand, a plant-level Boolean is identified, then the plant-level HCLPF can be quantified, for example at, say, 0.26 g (or whatever).

4. Your " Item 9" Comment: The Expert Panel has continued to believe that a separate study is needed to examine the relative merits of the CDFM and the PRA-fragility ,

approaches to determining HCLPF values. The sooner this is begun, the sooner it will be completed. What is needed is a comparison of several different classes of equipment with several examples selected for each class. The first task is to establish a well-defined working definition of the CDPM method. Following this, for

, each specific equipment item being s,udied, the analysts should compute the HCLPF value using the CDFM method and compute a separate and independent HCLPF value using the fragility method. It should also be valuable, if possible, to use more than one group in the study, so as to illustrate any analyst-to-analyst differences in interpretation.

The key aspect of the' study is assuring that all assumptions made in each analysis for each equipment item are documented, so that differences between the HCLPF '

results may be traced and understood rather than merely stated.

After this part of the study is completed, the analysts should be expected to make recommendations for any modifications in either the CDFM or the fragilities

. method, as needed to help resolve any differences found.

The overall goal is to develop a method sufficiently " cut and dried" (that is, prescriptive) that it can be used routinely by the general engineering community for performing HCLPF calculations for seismic margins reviews.

i Your final comment is correct-building on the current Maine Yankee review may be the most efficient way to proceed on the CDFM comparison study at this stage. But that probably means starting as soon as we can, since getting a study in place probably will take a few months.

I hope that this letter has answered your comments if not, please let me know.

Sincerely yours, Robert J. Budnitz Chairman Expert Panel i RJB/sa Attachment CC: bhlef fanO R Arry, LLNL A-13 l r' 1 .

APPENDIX B PEER REVIEW GROUP CORRESPONDENCE B-1 -

e

.e __ y

PEER REVIEW GROUP CHARTER R. J. Budnitz, Chairman The objective of the Peer Review Group is to assure that the trial seismic margins review is executed in a fully competent and professional manner, uses methods that are at the state-of-the-art, follows the guidance established in NUREG/CR-4334 and NUREG/CR-4482, and takes cognizance of all relevant information. The sponsors of the study (Lawrence Livermore National Laboratory for the NRC and Maine Yankee as the plant owner) desire to utilize the results of the study, and require the Peer Review Group's assurance that  ;

the study is technically sound.

I To accomplish its objective, the Peer Review Group will be provided full access to all materials, information, and methodologies that are inputs to and used by the study team. Access to the study team itself will occur through scheduled meetings to follow the study's progress. The Peer Review Group will also review draft reports and participate in walkdowns of the plant. Formal reporting and interface for the Peer Review Group will be through the Group's chairman to LLNL.

4 0

B-2

FutureResources A.uociates,Inc.

2000 Center Street Suite 418 Berkeley. C.\ 04704 415-520-5111 4 June ,1986 Mr. J. Thomas Duke Power Company Design Engineering Department 422 South Church Street P.O. Box 33189 Charlotte, North Carolina 28242

Dear Mr. Thomas:

This letter is being written on behalf of Lawrence Livermore National Laboratory's ' Seismic Margins Program". As you may know, I am the chairman-of a newly-constituted " Peer Review Group" that will be assisting LLNL in carrying out a trial ' seismic margins review' of the Maine Yankee plant in the coming months. I am grateful that you have agreed to participate as a member of this Peer Review Group, and I look forward to interacting with you. I know that Bob Murray of LLNL has given you more details.

The purpose of this letter is to confirm the tentative dates of the first Peer Review Group meeting, to provide you with a tentative schedule for the rest of the Peer Review Group's activities, and to provide you with a copy of the Group's charter. (The charter is attached to this letter.)

The tentative schedule is as follows:

1) The first Peer Review Group meeting will be held at the Maine Yankee site on July 21-22-23. The first day will be used for a Group meeting, and the next two days will involve a walkdown of the plant along with the margins review team and the utility personnel.

Bob Murray of LLNL should provide you with travel details soon.

2) In Septemaer, it will be necessary to devote 1 day (perhaps 2 daysJ to a review of the progress to that date; the study _ team will have provided their tentative findings and issues to us. There will not-be a Peer Review Group meeting but we will discuss this by telephone.

3) In November, there will be a 1-day Peer Review Group meeting to go over what has been accomplished to date. This meeting could go to a second day but I hope it will not. The site of this meeting and its exact date is not yet known.

l B-3

.1

. . _. ~ .

w

~

.J. Thomas, Duke Power Company page 2 -- 4 June 1986

4) There will be a 1-2 day cannitment in February,~ 1987 to-review

.the final study report in draft form. 1This meeting will~be at a site to be determined, but it will probably be held in Washington so that the Peer. Review Group can interact with the

. relevant NRC staff members.

i-The Peer Review Group reports formally to LLNL, and specifically to

-Dr. Robert C. Murray, who is the designated LLNL contact._ However, our. report will be a public document that will also be of use to thel. '

Maine Yankee group, other utilities, as well as EPRI and the NRC. . The

'eport format will be a letter from me as Chairman, reporting on the-r

~ Group's findings. It has been agreed that there will be an opportunity i

i

_for-any Peer Review Group menber to write a separate minority report' if needed, but the expectation is that the Chairman's letter can. capture" all of the review comments.

i :I hope that this letter and the~ attached Charter can' clarify your role

as a member of the Peer Review Group. If you have any questions, I will be happy to discuss them with you. I'look-forward to meeting you at-

Maine Yankee on July 21.

With warmest regards,

! Sincerely yours, i

Robert J. Budnitz

cc

R.C.Murray e

f

?

J I

l

! 1 l

d 1

l i

4 B-4 4

0

+ 0 UNITED STATES 8 ~ ,%

NUCLEAR REGULATORY COMMISSION h WASHINGTON, D. C. 20066

\ **"* /;

JUL 7 1986 MEMORANDUM FOR: Mr. James Richardson, RES Co-Chairman Seismic K , gins Working Group FROM: Newton Anderson, NRR Co-Chairman Seismic Margins Working Group

SUBJECT:

C0peiENTS ON THE NRC SEISMIC MARGINS PROGRAM AND TRIAL PLANT REVIEW I have recently received a copy of a letter from Dr. Budnitz, Chairman of the Seismic Margins Peer Review Group to members of that group. The letter transmitted a tentative schedule for Peer Review Group meetings and a copy of the Peer Review Group Charter for the Maine Yankee Trial Review.

I have some concerns about both the charter and the schedule. I believe that the Charter does not reflect what my understanding is of the Peer Review Group function. It should be clearly articulated that we expect the Peer Review Group to not only provide "... assurance that the study is technically sound",

but to endorse the technical results. At the conclusion of the Maine Yankee review they should be expected to provide their best judgment with regard to both the review procedure and the seismic capability of Maine Yankee, based on their collective expert opinions and backed by their professional' reputations.

The other problem I have with the Charter is the statement on formal reporting.

The Peer Review Group should not report through LLNL'. They should report as an independent group, providing their results to both LLNL.and NRC without any intermediaries. LLNL should provide administrative support to the group but I feel strongly they should maintain their independence.

If my comments with regard to the Charter are accepted, then I feel that the proposed schedule for Peer Review Group meetings is inadequate. It appears to me that the limited number of meetings and allotted review time is not sufficient to allow the group to " stake their professional opinions" on their assessment '

of the study results. The Peer Review Group members should decide what level of involvement they need based on what we expect of them.

I would also like to express my concern about some problems I see with the relationship between the NRC/LLNL Seismic Margins Program and the EPRI Seismic Margins Program. My basic concern is that'there is a significant difference in the approach and in the scope of review. We need to work hard to bring these programs together. My understanding is that EPRI sees the NRC program as a study to " develop a requirement" to do a margins review and that the EPRI program is to provide the utilities with a tool for performing plant specific reviews to meet that requirement. If this is a reasonable way to proceed, then we certainly should work with EPRI to ensure the two programs are compatible.

I do not mean to imply that we are not coordinating with EPRI. I know we are; but I do not see these differences being resolved.

B-5

i

, 1 I suggest that we hold a meeting with the NRC seismics margins working group

in early August for the purpose of discussing these. issues and developing a

[ plan for resolving them.

I I ~ )

Newton R. erson. NRR Co-Chariman Seismic Margins Working Group l cc: Seismic Margins Working Group Members Dr. Robert Budnitz K Ge % Cimittps; LLNL m -

T. Spets

8. Sheron R. Bosnak I

i l

l i

t i

j i

l l

B-6

FutureResource.oA.uociates,Inc.

2000 Center Street Suite 418 Berkeley. CA 94704 415-526-5111 1 August 1986 1

TO: U.S. Nuclear Regulatory Commission D. Guzy, Office of Nuclear Regulatory Research P. Sears, Office of Nuclear Reactor Regulation Maine Yankee Atomic Power Company D. Whittier, Manager, Nuclear Engineering and Licensing FROM: Robert J. Budnitz, Chairman of Peer Review Group for the Maine Yankee Seismic Margin Review Study REF: MINUTES, FIRST MEETING OF THE PEER REVIEW GROUP These are minutes of the first meeting of the " Peer Review Group" that is reviewing the technical competence of the " Seismic Margin Review Study" that is being undertaken on the Maine Yankee reactor plant under sponsorship of the U.S. Nuclear Regulatory Commission.

The meeting was held on Monday, July 21, 1986 and continued on Tuesday and Wednesday, July 22 and 23. Attending besides the chairman were Dr. John Reed; Mr. Loring Wyllie; and Mr. James Thomas. Mr. Thomas' attendance was limited to only the morning of the first day. Absent was Dr. Michael Bohn, the fifth member of the Peer Review Group.

Also attending were representatives of the NRC; the Maine Yankee staff and their Yankee Atanic Electric associates; Lawrence Livermore National Laboratory staff; LLNL's subcontractors performing the review, Energy Incorporated and EQE Inc.; and R. Kennedy, a consultant to Maine Yankee. The signup sheet for the opening session is attached as Attachment A; however, not all of these i individuals participated in all of the meeting sessions.

Atta:hment B shows the agenda for the first day's session. On the second day, a plant walkdown was done by most of the attendees in teams of 4-6 each, preceded by 7. radiological protection briefing and followed by a de-briefing session in which technical issues raised during the walkdown were discussed. On the third day, another walkdown in the morning was followed by still another session

, ir which technical issues were discussed. The Peer Review Group then adjourned

.s session just after noon, and the study team continued its discussions and its walkdown, which would last for the remainder of the week.

The minutes of the meeting will be presented as numerically ordered topics, as follows:

B-7 e

5

l l

Minutes of Peer Review Group, Meeting of 21-23 July 1986 page 2

-1) .The first session began with a briefing by Robert Murray of LLNL, who presented the background of the project. Murray discussed the purpose of the trial margin review, the development of the methodology by the NRC's " Expert Panel of Seismic Margins", the schedule of the review, its structure, and-the relationships among the parties. During the course of Murray's presenta-tion an important connent by A. Thadani of NRC was made in which Thadani pointed out that any licensing action that might result from this study would only.

occur after its completion. NRC, according to Thadani, was awaiting the results

because there are methodological issues as well as technical issues about the Maine Yankee plant that are being studied in this trial review.

2) Robert Budnitz discussed the interactions that have occurred so far between this study effort and a parallel effort being undertaken by the Electric Power Research Institute, which is studying the seismic margin at Duke Power's Catawba station. Budnitz pointed out that recent discussions have assisted both study teams to understand in what ways their respective methodologies are similar or different. Continuing interactions will be encouraged.

3) James Thomas discussed the importance of interactions between this study and related technical work being done under the NRC's A-46 program and under the industry's SQUG (Seismic Qualifications Utility Group) effort. Taking cog-nizance of these other efforts will be important for the study team.

4) William Henries of Yankee Atomic then discussed the current status of the Maine Yankee plant in terms of its seismic capacity. He provided background

, on the history of the seismic design of the plant, recent utility-sponsored studies that provided information to assist them in understanding their plant,

, and recent actions taken to modify the plant's seismic performance.

5) David Moore of Energy Incorporated, who is leading the systems-analysis team performing this review under LLNL subcontract, made a presentation that provided information on the approach being talen. Moore's viewgraphs are attached to these minutes as Attachmert C. He pointed out that there are a few key issues that must be addressed, and asked for discussion and guidance i from the group on how best to approach them. For exaqiple, because a walkdown I inside the containment will not be possible, it will be necessary to make some l assumptions about the presence of small LOCAs, and the approach being taken to l this issue was covered. Also, the handling of the possible sequences in which control rod function might be compromised was discussed, and Moore's review l approach to reactor internals and boric-acid safety injection was covered.

6) A lunch-time break was taken so that the Peer Review Group could go into

, executive session. After that executive session, the Peer Review Group

! suggested that minor modifications to the draft PRG charter should be made to reflect more accurately the actual approach being taken. The revised charter which was discussed by the PRG with those present is shown as Attachment D.

The differences between this version and the earlier draft are a more explicit l B-8

4.

i Minutes of Peer Review Group, Meeting of 21-23 July 1986

, page 3 statement of the fact that the Peer Review Group is an independent group reporting to NRC and Maine Yankee, and a clarification of the Group's role vis-a-vis assuring that the_ methodology is being followed.

7) M.K. Ravindra of EQE Inc. made a presentation in which he discussed his group's approach, as the fragilities subcontractor for this effort, in carrying s out their work. His viewgraphs are attached as Attachment E. Technical topics covered included Ravindra's approach to HCLPF determinations using the fragilities method and the C0FM (conservative deterministic fragilities method) for analyzing capacities. The handling of seismic capacities for those numerous components 1

(valves, etc.) which can only be studies in a sampling way rather than in a 100 %-analysis way was covered. Also, a discussion took place on how the study will examine those ' Group B' components that. require study to assure that there are not unusual issues involved in their capacities.

8) The structure of the Final Report was covered, so that all parties present could understand how it will be structured, who will write which sections, and its schedule. The process whereby the Peer Review Group will be able to review a draft version of this report toward the end of the project was ditcussed. It is anticipated that there will be a meeting of the participants to enable the 3

Peer Reviewers to interact.

9) The PRG's reporting was discussed. It was agreed here, as in earlier meetings, that a letter from the PRG Chairman would provide the Group's final consnents on the study. Concurrence by the other members, with the opportunity for any j individual PRG member to provide individual comments as a minority report if l desired, would be the method used.

1 10) An extensive discussion took place on how the systems analysis team would I handle 'non-seismic induced failures' in their systems analysis, when a potential I accident sequence might involve both these and seismic-induced failures. The Energy Incorporated study team will do an analysis which will incorporate these non-siesmic-induced failures where their presence will make a significant diffe-rence in the overall risk profile of the plant.

11) The Monday session ended just after 5:00 PM.

l 12) On Tuesday (July 22) a radiological briefing and a plant walkdwon in si::all

groups took most of the day. A debriefing session from 4

30 to 7:00 PM ended j the day. At that session, various technical issues uncovered during the day's l walkdown were discussed, to allow the next day's walkdown to be more effective.

l An extensive disussion of the capacity of the DC batteries took place, along with discussions of the capacities and functions of several tanks. Some of

the walkdown teams had not completed their entire first-pass tours, so this dis-cussion was partly of the character of assuring that these teams looked at and studied plant features that some of the other groups had highlighted.

B-9 i

. .,--- ,,.---., - . - -. . -_.. - - - - . . - . - - - - - . . , - - - - ..-.- -.-, - _ . i

d l . I l

. Minutes of Peer Review Group, Meeting of 21-23 July 1986 page 4

13) During this session late on Tuesday afternoon the role of'the Peer i

Review Group in interacting with the study participants was discussed. The l independence of the PRG must be assured, but it is also important that tech- 1 nical issues of concern to the PRG be provided to the study team rather than

'kept to the end'. The handling of this aspect of the project was covered.

J l 14) On Wednesday morning (July 23)the plant walkdown continued, until 11:00 when the meeting of the Peer Review Group continued for just over one hour.

In this final session, PRG members discussed again the issue of assuring that independence for their work would be a fact as well as a perception. Also, 3

the interactions between the NRC-sponsored study team.and the utility staff were discussed, to assure that there would not be improper influence by the utility over the project outcome.

15) The Peer Review Group's final discussion and comments prior to adjournment were of the character that the study seems to be 'on track' in a technical sesne so far, although of course this is a very preliminary observation, and was made by only three of the PRG members (Budnitz, Reed, Wyllie), the other two (Thomas, Bohn) being absent.

16) The PRG meeting adjourned just after noon on Wednesday,' 23 July.

17) The next meeting of the Peer Review Group will be held on September 30, at a location in the San Francisco Bay area that will be identified by the LLNL team soon. The subsequent meeting of the PRG will be held in November in con-junction with the second plant walkdown. It will be held at the Maine Yankee plant site. While dates have not been firmly set, the dates of November 17-18-19 were written down tentatively, and are being held by all participants pending further developments.

! 18) These minutes, written by the PRG Chairman (Budnitz), are being sent to i the other members for their review and comment, after which they will be made

{ final. The draft version was not circulated outside the Peer Review Group.

Robert J. Bud z, Chairman

! Attachment A: Attendance signup sheet, 7/21/86

! Attachment B: Agenda for first day's session, as prepared by R. Murray of LLNL

Attachment C

Viewgraphs for D. Moore's presentation (Energy Incprporated)

' Attachment D: Revised charter for Peer Review Group Attachment E: Viewgraphs for M. K. Ravindra's presentation (EQE Inc.)

s 1

i

_______,__--,._._-_--c .

l l

l l

Minutes of Peer Review Group, Meeting of 21-23 July 1986 )

page 5 s

ADDITIONAL ITER 4:

(This item was not in Budnitz's notes for the meeting, but Reid and the others remembered that it was discussed.....probably while Budnitz was out of the room. It is presented here because it is deemed important additional guidance from the Peer Review Group):

19) The Peer Review Group's interpretation of NRC's requirement on the input spectrum to be used in this review is that a NUREG/CR-0098 spectrum, using 50th percentile amplification factors and anchored to 0.30 g zero period acce-leration, is the proper ground motion to use. No variability in this input spectrum is to be used. Reference for NRC's guidance is the memorandum from D.M. Crutchfield to P.M. Sears, dated 7 May 1986.-

4 o

I i B-ll l

r- -- - .- - --._- , , ,_ , e.. ---,-. - . m..---- - - _

y

Gldi-Y 2(,/fl{ $ HNNE yk(W c , -

NA.M. .. . __ @ @ T \' _ _ _ _ 3 it.1A si'f Hi . b 119 ._I.._. h<a_kL #cL l uY._ h' 2l )~ YN AsNoK . . TNAb@ ..

. NRc ._

(3oQ yn- ys c.5...

-. GAR 7tt Cus1utHss - .J.Lb/L C+tQf22- +9_49

-_bb Mt4Vy ._ . ._ L.L. lM- __ ._ . _

__.__(LT._i7l:03Af

._So}...$necly._ -S$ C . - ._. .-. . . 7N 27Z- ?/O_.

-'O e _ 1%m- 4_. _ _ __ . . D ~ la. h> _c o ._. bo'O.373-W/2 Lory W .lue... y .D.egea le. lk Asoc__ (43r)392-691_2. _

Ra _ be<)_. _qBA (415) 'Mi- 3 7-P-0 4 v,2__ M oer c __ ____ _. h ag y_ % c.. (2of)__gM-00_go_ _.

Mgt.Guu.t _._ hm.ev Lc. (tou).Bn. - c.or.] __ .

M.JL. Qavind r^- E 9E Ine- __ h ! 4 ) J s 2 .? S L 9_ ._

Da.s Auly ._ .._ . K C- _ _ . 3 .e l l - 2 1 3 ~ ~? ") t h . .

                              - dew._ Yo un G - _ _. ._. E~s>%.v_.ZMc __ _ .__.. ( 2._o 1 2 4 M? . 2 e o                                           -

• B .Il. ._ b m s ..._ _. _. Was A r*~ e. _ _. (qc)_rpetoo ._

__ Niluh_ChokM . NRC _ . . . . 3ol 192..-8347 _

                              %e Tussin as_                                            __. _ _ _ L LN L. _ ._ ._ .                                                   ._                   LALS)49.-G758-_.
                    ._           EC - Af - _. _ ..- _ $0 & . . _ .Nc., _ _                                                                                                             .

fl ES) ~.] O hCl .. _.

4. N__ . ._ _ . .

                                                                                                                      .              _                   . . _ . _ ,             _               _7."~            . _ . _ _ .

f-q" d, he. .. . .. _ cvir). 41c me

                            )/JH ..I' bWIll////

Y/) AIKf2 /770MI C .

                                                                                                                                                                                      .-. _pt7 - ?;-                   .._

_. .Tl2.(C I b. G.kRT . .. . N C N2 S . _ _ _ . . 30L._85 f 2. S co$ _. ._ . .. _. To t. h e. C u m b t.t' _YAn b M o mic. __ _ G l.~l - B~l t. - 910* _ I/5 . Ddk/ _ . . .. MA / sw .

                                                                                                                                                                                ...._._. 8 -((d" M [..._                             _

_.. M w;wb mE. h. tt4 - E 2. - 9 7 9 9 B-12 i

ATTACHMENT B Mooday Meetlag Agenda Maine Yankee Atoanie Ptnser Station Staff Building July 21,1986 9:00 Project Overview /R. C. Murray 9:15 Yankee Atomic / Maine Yankee Plant Overview 9:45 Systems Status / Energy Inc. 10:30 Break 10:45 Fragility Status /EQE, Inc. 11:30 Open Discussion (collect Health Physic paperwork) 12:00 Lunch (at plant) 1:00 Peer Review Group Discussion /R. J. Bukitz 5:00 Adjourn i B-13 . l

4 ATTACHMENT C SYSTEMS ANALYSIS ENEMY INCORPORATED - TEAM: DAVID MOORE JON YOUNG MARC QUILICI e 9 0 e i g,p -

         .'s.

SYSTEMS ANALYSIS PROCEDURES STEP 2 - Initial Systems Review STEP 4 - First Plant Walkdown STEP 5 - System Mooeling STEP 6 - Second Plant Walkdown STEP 7 - System Moceling Analysis STEP 8 - Margin Evaluation of Components and Plant 9 J l I k i r,' N B-15 i ( ' ( l 5 i

  . , - . . ,           -.   -- --... ---.,--. -,.-,,. ,                ,   . - - - - -   - - - - - - - - . ~ .  . . . - - , , . . - . . . , - -   .
                                                                                                                                                     .- - . ~ . . . ,       . . - -     -. ..      , . - - .

GROUP A SYSTEMS HPSI HIGH PRESSURE SAFETY INJECTION AFW AUXILIARY FEEDWATER(includes EMERGENCY FEEDWATER) ASDHR ALTERNATE SHUTDOWN DECAY HEAT REMOVAL BAT BO'IC ACID TRANSFER PPC PRIMARY PRESSURE. CONTROL SPC SECONDARY PRESSURE CONTROL # s, ACP AC POWER DG DIESEL GENERATORS DCP DC POWER PCC PRIMARY COMPONENT COOLING WATER SCC SECONDARY COMPONENT COOLING WATER SWS SERVICEWATERSYS[EM ACTUATION (includes RPS, SIAS, and maybe RAS, CSAS, CIS) F e B-16

j. ' 43

2 i f SYSTEM: HIGH PRESSURE SAFETY INJECTION (HPSI) SAFETY FUNCTION: Injec- borated water into the reactor vesset immediately after a LOCA. Also for feed and bleed, post-accident core cooling and additional shutdown capability durin5 rapid cooldown of RCS. i SYSTEM COMPONENTS: Tanks: TK-4 Refueling Cavity Vater Storage Tank Pumps:  ?-!c A (N.C.) Charging (HPSI) Pump

                                    , P-lab (S)           Charging (HPSI) Pump P-165 (Spare)       Charging (HPSI) Pump Heat Exchangers:     E-3A                Residual Heat Exchanger E-3B                Residual Heat Exchanger

. SUPPORT SYSTEMS: AC Power: al60V Emergency Bus 3 i al60V Emergency Bus 6 DC Power: Air: HVAC: y Pump Cooling: PCC P-14 A-7, 3 Lube Oil Pumps SCC P-14B-2, 3 P-145-7, 3 Actuation: SIAS Low Pressurizer Pressure High Containment Pressure e 1 i

      .                         \                                                                     ,
          ,,                                      B-17 l
                                                                                                      \

1/I , . l

 '                                                                  N                                     M

c. c 4 o a y

               ~                                                    a                                     b 5I                                        4        15                            s        1-
    !T                                                     +                                      ?         ?

I i t r. e I. I. e i n

                                                                       $                          e         f
    $-                                                     $l
                                                                                                  =r--:

A I je c a m

                                                                                     ~

Sc 55 Y ' i ;E cg t I E r og m A= $ E  : 2. i= S 4= r m m A -i i- =i 4 o a v if ,a i

                                                                                           $'iu +

_ __ _. J

                                                                                               =

d n Z g$ H M y

                                  $                                                                  F E<3         *f        r                                                                   W Si         v d'

e A Lt. y T

                   .             toa                                                                 f F3 to                          <

w, o . l 7 X r o o A 7 N 2- m ( d. W c[ M M f-- m h 5

                                        =

C Z u', i '

• 4

        ;fec y     s: sM t-i             i 1

C I SI F m e, , ' 1 &c }e' I c e l

        .                                T
                                         !                    B-18

• i

                                                                                                                                                                                             ~

j .s - N ' i

              ,                                                                                             VALVE TABLE
!                                                                                                                                                                    Operating l                                                                                                                                 Power                    Normal     Position           Fai1 j                         valve                        Gescription (50V)                Position     (Actuation)    Position AFW-A-101                    Flow control to SG E-1-1                                              120VAC 1A                          0     C(SLR)               0 (1201A1)

ATW-A-201 Flow control to SG E-1-2 12cVAC 1A 0 C(SLR) 0 (120181)

;                         AFW-A-301                    Flow control to SG E-1-3                                         . 120VAC IA                          0     C(SLR)              -0 (1201C1) i 1

AFW-A-338 Flow control isolation valve (AFW-A-101) 120VAC 3A 0 C(St.R) i (120SA) AFW.A-339 Flow control isolation valve (AFW-A-201) 120VAC 3A 0 C(SLR)

!           3                                                                                                                  (12058) 1                          AFW-A-340                    Flow control isolation valve (AFW-A-301)                              120VAC 3A                          0     C(SLR)

(1205C) BA-A-32 Boric acid VCT, isolation valve C C(SIAS) C (210Z)

BA-A-80 Boric acid VCT (botton) isolation valve O C I SA-F-30 Boric acid flow control valve C C(SIAS) C (210Y)

SA-M-36 Emergency boration isolation valve MCC 8A C 0(MAN) 'AI SA-N-37 Emergency boration isolation valve MCC 7A C 0(MAN) Al Clf-A-32 HPSI pump 8 discharge to charging header BATT-1 0 C(SIAS)- (255) Cll-A-33 . IIPSI pump A discharge to charging header 8ATT-1 0 C(SIAS) _ 3 (254) 2 j CC-2/1

         ,          -                                                                                          .                                                               ________-________l

4

                                  .                           e            o          e o e mamme                             ammme.

A 41

       =

1 d$

• l l

i q&d"w A j 4 t-Y e e f'O . N - (D = Y  % Y fdl Y . U X W - m 3 a O g O N W o-M Oo a Y  % & Y Y K  % 4 - i- 1

   .      ese a

W i d X W K x  %

       )"";-

M g }fy]l  % i Y g

               } d                     /    n         %       4           %        --

O b f s) $ b O $ d ~d

                           &                                              a            y            $          d    Jr JI
                                            .r {#

b3 5< $M I $ U U $ s- ~ .~

                        ,1 i;             -  .1   ~ *-     a. e            e         ~

e , . 5 ~ ~ r F. ~ 5 r, t n - E @ J }

              -                       5                       W                o os                           -

B-20 ,

                                                                                                                        ,   ,        l

[

         /*

i e l  ! l l l

                       ! ,         i      ?

i l I l 1 I I ,! , Il  !- i  !! l _ oM i 5!  ! dh3 D

                                                                                                  \P"                           .

3 s l l xi . l 8 j <  ! i

                                                      /                               !                       v.                !

vl ,

              %i                     N           N N                                         N          N e
                      ]

8!s x , x i' t- < < x x x x h. ei i= *

    ,          A. b v
             <      t*N ifly':l' J.

e. lt'1

                     .I ylJ*l x     N     .

X P,i W W 4 J]

   ~ ~ ~

g Ij* A V ll N

                          *i s   ;' .   -t al -r e i..isi-                        , ,     o
                                                                                                                  ,          l J/,$!dJ2,5;IiNlN!h}lfl%/Sl2 4,                                                                         M        **     U    #

i , l {-ll;l'll',Ii l i '. lh I =>gJ 3 3 1 l& !!,- I

                 -                     ,~, 2                       ,

s.x  :

      .                                                                  B-21
                                                                                                      -r

Se'sst:/ - RPs EW/AFV RCS Integ. HPSI/F&B Status t037 OK OK CD OK CD Transfer-

                 'Wtry                                 -

• MT EW/AW HPSI Status Doen/Reclass OK OK o.

                                                   .          CD     ,

CD CD

                  \

Seismic Nrgin Event frees (Preliminary) B-22

 \

                .                                                                                                                                                ATTACIMENT'D' i

PEER REVIEW GROUP CHARTER (21 July 1986) The objective of the Peer Review Group is to assure that the trial seismic: margins review, following the guidance established in NUREG/CR-4334 and . l j NUREG/CR-4482, is executed in .a fully competent and professional manner, uses i methods that'are at the state-of-the-art, and takes cognizance of all relevant. information. The sponsors of the study (Lawrence Livermore National Labora - tory for the NRC and Maine Yankee as the plant owner) desire to utilize the f j' results of the study, and require the Peer Review Group's assurance that.the ! study is technically sound. 'l j To accomplish its objective, the Peer Review Group will be provided full access i to all materials, information, and methodologies that are inputs to and used ! by the study team. Access to the study team itself will occur through sche- ) duled meetings to follow the study's progress. The Peer Review Group will also review draft reports and participate in walkdowns of the plant. A formal report for the Peer Review Group will be made by the Group's chairman to NRC and Maine Yankee. It is understood that the Peer Review Group's report will be a public ) document. . I l 4 i i l 1 i l B-23 s l 1

    , - - , . - , , . - - - - - - - . ,            , - - .      - - , . , - ..,,,,-,,n.      -
                                                                                                     ,,..-,-,-,, - - ,--,,,,-,,,.-   , . --, .. ,,-,,,,-- -, _ ,., ,,, , ,,,._ e,      -

ATTACHMENT E l i NRC SEISMIC MARGINS PROGRAM TRIAL PLANT REVIEW FRAGILITY ASPECTS STATUS REPORT 4 I PRESENTED BY M.K. Ravindra , G.S. Hardy - P.S. Hashimoto l S.W. Swan 1 i I EQE incorporated Newport Beach, CA 1 PRESENTED TO PEER REVIEW GROUP i 4

 ;                                                                                          JULY 21,1986 i

i j B-24 i i

                                                                                                                                                                                                   @'{b@

.a 4

   .gw-..   --,-_.   . . . . . ,- .y.. -ewyy,,, ,, ,- _,y,__ ,._. , , , , - , . _ - , . . . , . . , - . -. -

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

OUTLINE e initial Screening e identify Target Areas for Walkdown e Walkdown Procedures d e Documentation e Outstanding issues 1 B-25 i-- - -. -- -- .- -

l l INITIAL SCREENING e Categorize Maine Yankee Components into Generic , Component Categories identified by Panel e Pre-screen Components in or Qu_t Based on Panels Recommended Guidelines e identify Areas of Concentrated Effort for Each Generic Component Class c B-26 S-IS_5, m 6-

I l

   .                                                                                                 ,,      l l

l MAINE YANKEE SEISMIC MARGINS REVIEW SYSTEM: HIGH PRESSURE SAFETY INJECTION fHPSI) COMPONENT INITIAL SCREENING lli DMI TK-4 Refueling Cavity Water Storage X Tank P-14A Charging (HPSI) Pump X1 P-14B Charging (PSI) Pump X1 P-14S Charging (HPSI) Pump X1 E-3A Residual Heat Exchanger X E-38 Residual Heat Exchanger X , Bus 5 4160V Emergency Bus X2 ' Bus 6 4160V Emergency Bus X2 P-14A-2,3 Lube Oil Pumps (Pump cooling) X1 P-148-2,3 Lube Oil Pumps (Pump cooling) X1 P-14S-2,3 Lube Oil Pumps (Pump cooling) X1 Actuations X3 5 1 Anchorage must be inspected during walkdown and verified adequate. 2 Cabinet anchorage & attached component anchorage reust be inspected during walkdown and verified adequate. 3 ctual components yet to be identified. B-27 5,@5,

MAINE YANKEE SEISMIC MARGINS-REVIEW SYSTEM: Auxiliary Feed Water (AFW) COMPONENT INITIAL SCREENING IN QUI TK-21 Demineralized Water Storage Tank X P-25A Emergency Feed Pump X1 P-25B Emergency Feed Pump X1 P-25C Emergency Feed Pump X1 T-1 Turbine for P-25B (Powered from X MainSteam) Bus 5 4160V Emergency Bus X2 Bus 6 4160V Emergency Bus X2 E-86A 011 Cooler X E-868 Oil Cooler X E-86C 011 Cooler X Actuations X3 Instrumentation X3 1 Anchorage must be inspected during walkdown and verified adequate. 2 Cabinet anchorage & attached component anchorage must be inspected during walkdown and verified adequate. l 3 ctual components yet to be identified. l B-28 l

IDENTIFY TARGET AREAS FOR WALKDOWN e Locate All Components for Walkdown on Plant Layout Drawings e identify Buildings and Areas Requiring Access for Walkdown I e Based on initial Screening-Identify Specific Component Areas for Concentrated Review, ie (Anchorage, Lateral Restraints, etc.) e Develop walkdown Data Sheets for Each Generic Component Class Defining Areas of Concentrated Walkdown Effort l l l B-29 WP GM l

WALKDOWN PROCEDURES e Perform Walkdown for identified Components e Address Areas identified in initial Screening Requiring Concentrated Review e Confirm Screening Criteria of Panelis Satisfied for Each Component e Identify Areas That May Require Additional Review l B-30

                                                                                    ':3' 3 5
                                                                                    . CGM

STRUCTURES SCREENING PROCEDURE , 1 l Component Screenine Comments Structures Out Structures housing Group A components are typically categorized as Class I in the FSAR. Class I structures were designed for the 0.1g hypothetical earthquake using the ACI 318-63 and AISC codes. Any gross structural deficiencies will tie identified by review of the design drawings and walkdown. Structure Out Class I structures are either cast impact integral with each other or are separated by three inch gaps. This will be confirmed by walkdown. Block Walls in A comprehensive screening procedure has been developed. Yard tanks in Information from drawings will be supplemented by walkdown. Soil Liquefaction Out Structures and yard tanks are founded on rock. Control and in Presence and adequacy of safety wiring Battery Room will be confirmed by walkdown Ceilings l Dams, Levees, in Nearby dike will be reviewed in l and Dikes walkdown B-31

                                                                            ,         E
                                                                            !E: :$.E       -

l BLOCK WALL' EVALUATION PROCEDURE l l REVIEW AVAILABLE INFORMATION. IDENTIFY ALL BLOCK WALLS USING MAINE YANKEE

SUMMARY

TABLE.  ; I LOCATE BLOCKWALLS ON EQUIPMENT  ! LAYOUT DRAWINGS. WALKDOWN. i i _ i WALLS SUPPORTING WALLS WHOSE WALLS SUPPORTING WALLS WHOSE GROUP A COLLAPSE LIFELINES (I.E. COLLAPSE COMPONENTS COULD RESULT CABLE TRAYS, COULD RESULT IN IMPACT PIPING,ETC.) IN IMPACT ONTO GROUP ONTO LIFELINES A COMPONENTS (I.E. CABLE - TRAYS, PIPING, 4 ETC.) lSCREENEDIN WILL IMPACT ARE THESE ARE THESE CAUSE DAMAGE LIFELINES LIFELINES TO GROUP A PART OF PART OF COMPONENTS? GROUP A GROUP A-SYSTEMS? SYSTEMS? BY SYSTEMS BY SYSTEMS ANALYST ANALYST I Y N " i i N ' ISCREENED INI iSCREENED OUTI WILL IMPACT ISCREENED OUTl CAUSE DAMAGE TO i GROUP A l LIFELINES? 1 Y N l SCREENED INI ISCREENED 0011 1 , Y g . N )

                                                                                      . lSCREENED IN) ISCREENED OUTl PERFORM FRAGILITY ANALYSIS TO CALCULATE HCLPP.

B-32 .

    .,_ _ . _     q.       ,     _ _ _ , ,                   ,      ,, , , _ . - _ _ _ , _ , _ , .              n    7 -,      ,,...,,,..%.-,.a           ,,,,,_y_..m,,.,_r.m.p-     .,

l DOCUMENTATION e Review of Design Calculation / Qualification Test Reports e Conclusions of Previous Studies e Walkdown Data Sheet e Walkdown Photographs e HCLPF Calculations B-33 s -

OUTSTANDING ISSUES , l e Lack of Qualification Data e Extent of Review for Components identified as C e CDFM Method Requires Site Specific Spectrum e Reactor Internals B-34

                                                     '.-22'2 y,= .

ENGNEEmNG. PLANNNG AND MANAGEMENT CONSULTANTS 1/ 8227-09 LLNL Seismic Hargin Review BY DATE CUENT LLNL SUBJECT Plant Walkdown Data Sheet CHKD DATE FACILITY: Maine Yankee Atomic Power Station COMPONENT: Line Number & Diameter LOCATION: Building = Elevation - Room - 1.0 CONCENTRATED AREAS OF REVIEW: 1.1 Pipe Flexibility: a Piping run - a Piping to equipment - a Building penetrations = k Note: Provide details on building penetrations if low capacity is observed. Indicate dimensions and details in Section 4 sheet 2. Photograph Roll No. Frame No.s 1.2 Pipe Condition: a Corrosion - m Brittle connections =

                                = Cast iron                      =

1.3 Support Details: a Type of anchor point Directional - Full restraint - a Spans between supports - a Support anchorage details - Photograph Roll No. Frame No.s i 1.4 Joints and Connections: a Threaded - a Socket welded - B-35 graph RoH No. Frame No.s

N ENGINEERING. PLANNING AND MANAGEMENT CONSULTANTS 2/ 8227-09 g LLNL Seismic Hargin Review BY DATE CUENT LLNL SUBJECT Plant Walkdown Data Sheet CHKD DATE 3.0 ADDITIONAL COMMENTS OR OBSERVATIONS: i 4.0 Sketch details of building penetrations if low capacity is observed. i l 1 B-36 I

1

ENGtNEERING. PLANNING AND MANAGEMENT CONSULTANTS 1/ 8227-09 LLNL Seismic Margin Review By DATE LLNL Plant Walkdown Data Sheet CLIENT SUBJEC7 CHKD DATE FACILITY: Maine Yankee Atomic Power Station COMPONENT: LOCATION: Building - Elevation - I.0 COMPONENT DATA: Plant ID Number - Manufacturer - Model - Function = Photograph (overall) Roll No. Frame No.s 2.0 AREAS REQUIRING DETAILED REVIEW: 2.1 Battery Rack Anchorage: Number and size of anchor bolts - Type of anchor bolts - Description of foundation - Photograph Roll No. Frame No.s Note: Provide a sketch of anchorage plan with dimensions and indicate any foundation deficiencies observed in space provided under Section 4 sheet 2. 2.2 Battery Rack Support: Type of support = Overall dimensions - Lateral Restraints - Type of member connections l

                                                                    =

l Photograph Roll No. Frame No.s Note: Provide a sketch of the rack indicating lateral restraints and dimensions in Section 5 sheet 3.

2.3 Batteries

Type of batteries - Description of battery spacers = i B-37 Photograph Roll No. Frame No.s 1 i I l

Mi5ml ENGWEERING, PLANNWG AND MANAGEMENT CONSULTANTS , 8227-09 2/ LLNL Seismic Margin Review _ By CUENT -

                      - SUBJECT -                                                _ DATE -

P1 ant Walkdown Data Sheet CHKD DATE -

• 3.0 ADDITIONAL COMMENTS OR OBSERVATIONS:

( 4.0observed. Sketch anchorage plan with dimensions and note any fou d n ation deficiencies 1 I I i I l l 1 l \ l 1 i B-38

l l ENGINEERING. PLANNING AND MANAGEMENT CONSULTANTS 3/ a no M27-09 a LLNL Seismic Margin Review BY DATE CLIENT LLNL SUBJECT Pl ant Walkrinwn nata thnot CHK D DATE 5.0 Sketch battery rack indicating lateral restraints and dimensions. B-39 l l l l

MiEl ENGINEEmNG. PLANMNG AND MANAGEMENT CONSULTANTS 2/ 8227-09 LLNL Seismic Margin Review JOB NO JOB By DATE LLNL Plant Walkdown Data Sheet CHKD GENT SUBJECT DATE 3.0 ADDITIONAL COMMENTS OR OBSERVATIONS: l 1 i . ( 4.0 Sketch anchorage plan with dimensions and note any foundation deficiencies observed. i B-38

ENGINEERING. PLANNING AND MANAGEMENT CONSULTANiS SHEETNO 3/ a no 8227-09 n LLNL Seismic Margin Review By DATE CUENT LLNL SUBJECT Plant Walkefnwn nata Rhoot CHK'O DATE 5.0 Sketch battery rack indicating lateral restraints and dimensions. B-39 1

ENGINEERING. PLANNING AND MANAGEMENT CONSULTANTS SHEETNO I/ a no 8227-09 job LLNL Seismic Margin Review BY DATE CLIENT LLNL SUBJECT ol:nt u ]gg.,;,g;;; 33;;; CHKD DATE FACILITY: Maine Yankee Atomic Power Station COMPONENT: LOCATION: Building - Elevation - 1.0 COMPONENT DATA: Plant ID Number - Manufacturer - Model - Function - Photograph (overall) Roll No. Frame No.s ( 2.0 AREAS REQUIRING DETAILED REVIEW:

2.1 Anchorage

Number and size of anchor bolts - Type of anchor bolts - Description of foundation - Photograph Roll No. Frame No.s Note: Provide a sketch of anchorage plan with dimensions and indicate any foundation deficiencies observed in space provided below. B-40

ENGINEERING PLANNING AND MANAGEMENT CONSULTANTS 2/ g 8227-09 g LLNL Seismic Margin Review BY DATE CUENT LLNL SUBJECT Plant Walkdown Data Sheet CHKD DATE 3.0 ADDITIONAL COMMENTS OR OBSERVATIONS: Note any system interactions. i B-41

. FutureReaurce.oA.uei-Ns,Inc. 2000 Center Street Suite 418 Berkeley. CA 94704 415-526-5111 14 October 1986 TO: U.S. Nuclear Regulatory Commission D. Guzy, Office of Nuclear Regulatory Research P. Sears, Office of Nuclear Reactor Regulation Maine Yankee Atomic Power Plant D. Whittier, Manager, Nuclear Engineering and Licensing FROM: Robert J. Budnitz, Chairman of Peer Review Group for the Maine Yankee Seismic Margin Review Study REF: MINUTES AND REPORT, SECOND PEER REVIEW GROUP MEETING This is the report of the second meeting of the " Peer Review Group" that is reviewing the technical competence of the " Seismic Margin Review Study" that is being undertaken on the Maine Yankee reactor plant under sponsorship of the U . S . Nuclear Regulatory Commission. The meeting was held at the San Francisco Airport Clarion Hotel on Tuesday. Seotember 30, 1986. Attending besides the chairman were all four of the other Peer Review Group members: Michael Bohn; John Reed; James Thomas; and Loring Wyllie. Also attending were representatives of the NRC; the Maine Yankee staff and their Yankee Atomic Electric associates; Lawrence Livermore National Laboratory. staff; LLNL's subcontractors performing the review (EI International and EQE, Inc. ) ; and R. Kennedy, a consultant to Maine Yankee. The sign-up sheet for the day's session is attached as Attachment A. All attendees attended essentially the entire meeting. Attachment B shows the agenda for the day's session. The meeting began at 9:00 AM and ended at 6:00 PM. The order of the agenda was followed quite closely, although some of the times were different. All agenda topics and presentations were given, and most of the day was used for technical discussion of aspects of the on-going margins review study. The minutes of the meeting will be presented as numerically ordered topics, as follows, with commentary included: B-42 .

i

l t

t i l

1) R. Murrav of T.TML led the introductory session, which consisted mainly of a discussion concerning the future schedule 3

for the review effort. The tentative schedule, as distributed by Murray, is shown as Attachment C. The schedule is judged to be satisfactory, in the sense that the project is currently 'on schedule' and the participants .believe that it can remain on i schedule according to the schedule shown .in Attachment C. The following specific dates have been decided on for the schedule: January 15, '1987: first full draft report due from contractors, .for limited

'                                                               distribution only to Peer Review -                                                          1 Group, Maine Yankee, LLNL, 2 or 3                                                           l NRC staff'                                                                                  f 1

January 22. 1987: next meeting of Pee' Review Group, in San Francisco, attendance by i invitation of PRG only l February 12. 1987: second draft due,.to be distributed l ' to wider distribution including broadly in NRC February 19. 1987: meeting in Washington with NRC in-house " Working Group on Seismic

Margins".

It was also decided that another Peer Review Group meeting would , be held sometime after the February time ' period, to study the j final version of the report to be prepared after the February j comments are in. I i

2) Daniel Guzv of NRC gave a brief summary of the July meeting of the NRC " Working Group on Seismic. Margins". He discussed the Working Group's current thinking, including the selection of a BWR plant as the subject . of a possible second trial margins review study. Newton Anderson of NRC, who is co-chairman of the NRC Working Group, provided additional comments. This part of the meeting was mainly for information purposes and elicited very 4 little discussion.

3) M. Ravindra of EOE. Inc. gave an overview presentation of the work and preliminary findings of his group,- who are the subcontractors doing the fragilities analysis of Maine Yankee.

He discussed a list of key items that will be examined in detail, l i and another list of . items that ; the fragilities team has already 2 B-43 _ 4- - , ., e.-_-, - - t, + - < ,,_m,. - - ,.y3.-"..,,,,i.%,r e.-3.----m,.,y,-+w,_.-.y..

      . . ~ .                .                            .      -        .            -      .   -.

analyzed. This presentation. was the subject of much technical l i discussion on several items of equipment, although a few . key items were put off until more detailed discussions scheduled for the afternoon. Ravindra discussed various data that the utility had furnished to . assist the fragilities evaluation, and also discussed his group's use of various experience data in their analyses. It was emphasized that the Peer Review' Group will need access to all data that will be relied on in these evaluations, in order to review its applicability. f Among the technical topics covered in Ravindra's presentation were service water piping; heat _ exchanger supports; valves with extended operators; HVAC fans and blower supports; cable tray motion and a possible hteraction with valve . operators; cable trays themselves, including Maine Yankee pull tests; steel frame buildings (specifically the pump house); and several others. Maine Yankee agreed to provide the summary report to the Peer Review Group which presents the results of the pull tests-conducted on concrete inserts. The main thrust of this part of the session was to familiarize the Peer Review Group with the approach being .taken by the EQE team.

4) David Moore of EI International next presented an overview of the systems analysis work that his firm is doing. He discussed the first walkdown, and a few tentative ' lessons learned' that may assist others in preparing for a first walkdown, such as doing more detailed preparatory study of HVAC and actuation systems. He emphasized that the first round of systems analysis is well under way, with seven fault trees complete and a few more under development. His group is now going back to answer specific questions that have arisen, in preparation for the second walkdown. The data sources being used were discussed, and it was again emphasized by the Peer Review Group that access to those data sources will be needed. _Also, the method of combining seismic and non-seismic failures was covered.

Moore's group will be doing some system cut sets soon, to obtain early guidance for EQE on what items seem to be more importa'nt. However, they do not plan to do sequence cut sets until after the second walkdown, as is called for in the NUREG/CR-4334 and 4482 guidance. There was discussion of certain specific items, and the Peer Review Group was provided with some detailed fault trees for their study. 3

B-44

(

 - - --       - - -    --  -         , , -      - - . ~ ,

i

5) Robert Kassawara of EPRI next -gave a presentation covering t EPRI's on-going seismic margins review project at Catawba. This- t review effort is on a schedule not very different from the schedule for NRC's review at Maine Yankee: initial report due 4 early in 1987, final report due a few months later. i While this presentation was mainly intended for information, it generated significant discussion concerning the differences between EPRI's and NRC's approaches. EPRI's approach is empha-sizing the CDFM method . for HCLPF determinations, while NRC's approach has not yet settled on one or another method, although it favors the CDFM method if: it can be studied enough. EPRI's i i

approach also uses a ' success path' method for. the systems analysis rather than a . fault-tree / event-tree method. There was  ; l auch discussion about .how to cope with small IOCAs inside  ! containment, which perhaps cannot be walked down and analyzed

           =well.- EPRI's analysis at Catawba is studying some relay chatter, l

but their analysis is limited to a very few relays on their chosen success path; NRC's approach . is not considering relay chatter for the time being, because further research is needed. j The general flavor of . this discussion was that both the EPRI l I review at Cata:rba and NRC's review at Maine Yankee will be able to learn much from each other, and continuing cooperation and i coordination will be encouraged.

6) Philio Hashimoto of EOE. Inc. gave a long and detailed presentation about three specific technical items that EQE is analyzing: the RWST, the numo house, and one block wall. He presented both CDFM and fragilities analyses of HCLPF values, and his presentation provided a vehicle for extensive discussion about the methodology used, the data relied on, approximations introduced, and whether EQE's analytical approaches were generic

or only specific to the item being analyzed.

It was pointed out in the discussion that in the analysis approach used for the RWST, the deflection compatibility between. i anchor bolt and water resistance modes is not provided. This needs to be carefully investigated before the capacities ' from l l these two modes are combined. In the course of this discussion, much detail was covered that

will not be discussed here. The principal thrust . of the Peer Review Group's comments were that there is a need - for careful study of both the CDFM and fragilities methods, since the

                'results' for HCLPF seem to depend greatly on assumptions made and data chosen.          One key aspect was the Peer Review Group's l               observation that experience data, if available, play a key part t               of the underlying approach to HCLPF analysis.                  It was requested

that realistic resistance modes be analyzed rather than using 4

B-45 l ( .

        - - =              .-        -      -     ._

design assumptions which are often overly conservative and do not represent the realistic response.

7) Greaorv Hardv of EOE. Inc. gave a detailed presentation about two other items being analyzed, the diesel day tank and one inverter located in the switchgear room. This discussion followed the same tenor as the discussion of Hashimoto's work just earlier. There was extensive interaction between the Peer Review Group and the analyst team, and mucn detailed discussion of data bases and assumptions. Again, discussion of experience data played a key role.

i 8) Grea Hardy then gave a briefer presentation, for information i purposes only, about progress in tracking down fragilities l information on three key issues: the lead-antimony battaries, the reactor internals and CRDMs, and the fire water threaded ning. In the battery case, he has been unsuccessful so far in identifying relevant test or experience data. Therefore, it is the Peer Review Group's understanding that analysis will be performed with and without these batteries. For the other two cases, the information now in hand, either due to configuration 2 or fragility aspects, should be adequate for the purposes of this margins analysis. Concerning the internals and CRDMs, the comment was made that Combustion Engineering's cooperation has been outstanding, f

9) David Moore of EI, in reply to Peer Review Group questions, discussed how his analysis group is coping with GrouD B functions, especially those items . needed to support long-term l heat removal. He provided a rationale for his approach, which will be the subject of later review by the Peer Review Group. He l! then discussed in more detail the treatment of non-seismic failures, with an emphasis on those that might compromise both

, redundant trains of some function. The question of how to combine these failures with seismic-induced failures in.a HCLPF j analysis was covered, but not resolved. What approach to take i remains an open question, and the NRC Expert Panel's guidance,on this subject is inadequate. i The Peer Review Group discussed a problem with incorpcrating non-  ! seismic failures after an earthquake, if their incorporation can ~ change the HCLPF level found. The problem arises if the way they affect post-earthquake plant response at the SSE level is identical to the . way they affect plant response near the HCLPF level or near the margin-review-earthquake level, in this case, i their actual effect on ' plant seismic margin' is minor. This issue will require more discussion in the future. j 5 ) I B-46 1 i'

2 Q ! 10) There was further discussion'of how the comparison between CDFM and fragilities methods for HCLPF. determination will.'be accomplished. It is recognized that the EQE effort in this project will- not extend to CDFM-fragilities methodological comparisons beyond those few presented to the Peer Review Group at this meeting. There is.a continuing need for a more thorough way to address this issue, which the Peer Review Group = will undoubtedly comment on further later on. It is likely that the NRC Expert . Panel may need to give this issue much more careful thought after this. Maine Yankee review has been completed.

11) John Reed asked a question of the NRC staff whose answer was not fully covered, since the NRC staff hazard experts were not present. . The question concerned exactly what was intended by NRC staff in their choice of the margin review earthquake being used in the Maine Yankee analysis. After some discussion, it was decided that it would be assumed that the median NUREG/CR-0098 spectrum anchored to 0.30 g ZPA represented a uniform hazard

- spectrum at the 84 % confidence level. NRC staff will attempt to obtain some clarification on this issue soon.

12) James Thomas raised a general question of how systems interactions aspects should be dealt with in the analysis, such as when a valve may hit a support during strong earthquake motion. Some generic guidance is needed to provide analysts with an acceptable approach, which should be usable in most cases l witaout doing a detailed and expensive calculation. This aspect was not resolved in the meeting, but will be given further thought.

l l OVERVIEW COMMENT: It is the road consensus of the Peer Review Group that the Maine

Yankee t 1 margins review project is being accomplished so far j with acceptable technical competence. In the course of its j review, the Review Group discussed several technical issues that j have been subjects of continuing difficulty ~in the analysis, and in some cases these difficulties may not be resolved during this

trial margins review project. The Peer Review Group recognizes

that one key objective of this trial review is to uncover such l issues, especially methodological issues or issues of inadequate l data. Its overview comment is that the analysis team, including 4

3 4 ~ 6 4 B-47

I LLNL, EI International, and EQE experts, is carrying out a fully acceptable analysis so far, within the constraints of the project scope and subject to the comments in the detailed discussion above. The Peer Review Group looks forward to further interactions' as the analysis proceeds through a second walkdown and then to final analysis and documentation of the project results. That completes this report and minutes. 4

                                                     /

Robert J. B6dnitz Chairman, Peer Review Group 1 4 7 B-48 i l l t

ATTACHMENT A 9[o[{6 Peea 72evisu/ Mw . QAge Po si r,o M 7g oc a

 ? SsA G?s                                                 3oj.492-gw c vsaT2clMni>D8 pau Gu ty                   av nc/r2 c f -           3 01 - + f1' 7 6' 9 

MiiheJ Boln. 9a .duk . Led i 20:- 2+6 -Me? G% E~qtt 4% s'sa .Ozra d a l & A % e. 51 Lhvn'd 7.66/trY-ce98

 % 0 g 1Ao % d o              e.o.s,. h.                7 M - 35L- 9 E99 M . E . R a.vi w a r          Eqe      1%               7 i 4 - s s 2 - 9 2.9,9 N.C.       h vr               LLiVL-                   40 - 4 A'A- 030 8 R,9. KA m m                  EMl                      4 tf FU-3mf
 '3. s . Th o n.s              D d e Pc.w w              % v s')3 - y s> a LT8T                     h86 6 5 7354-
  &[kAcw   T/7/'     $//V6H4         f'. hod Jhel.Aki,.(Bu.H.         7N - 777 - 2/c.]

p, 5 Msc[4.&c.Mm/<[,w #47 - 423-35 2.( n eit co a u v w ,,u,,a u.a % e. e,n - a n. 3 e o o 31// llennep 'l

  ~5oh o Y (2.eed              GL%                     (41s)?6MLl L b c 9-l) & 4 .t W L.L1 6           e b oll, k W o e   4 i S'- 392. - (#r.C 2-E 6 In: ,
                                                                            ~

bt re $ Nord3 Q i +) 9~.2 - / a > > R vti;e 6,4 32 f4 L. Wif) n 6.:,'! / G '.t f l A w w ; c) q l- L sil- (ap5)4g7. qq4g, AA Aluem OQ (so O qu- h- 6 ) B-49 i

ATTACHMENT B AGENDA Seismic Margins Program Peer Review Group Meeting San Francisco Airport Clarion Hotel September 30,1986 Introductions Budnitz/Murray S:00 a.m. 9:10 a.m. Briefing on NRC Working Group Meeting Guzy/ Anderson 9:30 a.m. Walkdown Summary Ravindra 10:00 a.m. Systems Summary - Moore ' Kassawara 11:00 a.m. Briefing on EPRI Program 12:00 noon Lunch 1:00 p.m. Fragility Summary HCLPF Comparisons Hardy /Hashimoto Tank Methodology Hashimoto 3:00 p.m. Important Open Issues Batteries Hardy Internals /CRDM Hardy Threaded Fire Water Piping Hardy 5:00 p.m. Report Outline Murray Second Plant Walkdown Schedule 1 6:00 p.m. Adjourn i l B-50 1

   ., ,-,   ,     , _ -    ,   ,           ,v_,_y   , - . . _                 . - . . . ,   _,         n ,_ ,                    -_.____,_,,

           ~                                                                              _.__

ATTACHMENT'C I i SEISMIC MARGINS PROGRAM - Tentative Schedule September 23,1986 Federal Express mailing to Peer Review Group September 30,1986 Peer Review Group meeting at San Francisco Airport Clarion Hotel

   - October 1,1986         Analysis Team meeting at San Francisco Airport Clarion Hotel October 22,1986       ~ Analysis Team meeting at Energy Incorporated, Kent, -

Washington November 10,1986 Second pre-walkdown status summary Federal Express package-November 17-19, 1986 Analysis Team second plant walkdown at Maine Yankee. November 18-19,1986 Peer Review Group meeting and additional walkdown at Maine Yankee December 8,1986 Preliminary plant HCLPF available December 10-12, 1986 Possible exchange meeting with EPRI and Analysis Team meeting (most participants will be at the Symposium on Current Issues Related to Nuclear Power Plant Structures, Equipment and Piping to be held at North Carolina State, December 10-12, 1986).

    -January 1987           Briefing at the NRC on program and resukts
    -February 1987           Draft (inal report submitted to the' NRC I

B-51 ~ i , l l

FutureReaurcu A.uociatu,Inc. . l 2(MX) Center St evet Suite 1IN Berkele.v. C' \ D-8701 415-526-511I l 5 December 1986 TO: U.S. Nuclear Regulatory Commission ' D. Guzy, Office of Nuclear Regulatory Research P. Sears, Office of Nuclear Reactor Regulation Maine Yankee Atomic Power Plant D. Whittier, Manager, Nuclear Engineering and Licensing FROM: Robert J. Budnitz, Chairman of Peer Review Group for the Maine Yankee Seismic Margin Review Study REF: MINUTES AND REPORT, THIRD PEER REVIEW GROUP MEETING This is the report of the third meeting of the " Peer Review-Group" that is reviewing the technical competence of the " Seismic Margin Review Study" that is being undertaken on the Maine Yankee reactor plant under sponsorship of the U.S. Nuclear Regulatory Commission. The meeting was held at the Maine Yankee site on Tuesday and Wednesday. November 18-19. 1986. Attending besides the chairman were all four of the other Peer Review Group members: Michael Bohn; John Reed; James Thomas; and Loring Wyllie. Also attending were representatives of the NRC; the Maine Yankee staff and their Yankee Atomic Electric associates; Lawrence Livermore National Laboratory staff; LLNL's subcontractors per-forming the review (EI International and EQE, Inc. ) ; and R. Kennedy, a consultant to Maine Yankee. The sign-up sheet for the first day's session is attached as Attachment A. All attendees attended essentially the entire meeting. D. Whittier of Maine Yankee, whose name is slot on the sign-up sheet, attended the I second day's session. Attachment B shows the agenda for the meeting, which lasted all l day Tuesday and half of Wednesday. The Tuesday meeting began at 9:00 AM and ended at 5:30 PM. Wednesday's meeting began at 8:30 AM and ended about nmn. The order of the agenda was followed , ' , '

 .          closely, although some of the times were different.                               All agenda topics and presentations were given.                          Most of the meeting time B-52 i

was ~ used for technical discussion of aspects of the on-going margins review study. In addition, on Tuesday afternoon all attendees participated in a walk-down of the plant to study

                                .particular aspects of the design relevant to its seismic margin.

The minutes of the meeting will be presented as numerically ordered topics, as follows, with commentary included: ' R. Murrav of T.TML led the introductory session, which con-

1) )

! sisted mainly of a discussion concerning the future schedule-for i the review effort. The schedule is judged to be satisfactory, in the sense that the project is currently ' on - schedule ' ~ and . the  ; participants believe that it can remain on schedule. The follow-ing specific dates have been decided on for the schedule: , December 11. 1986: meeting between NRC margins team and EPRI margins study team at Ra-leigh, NC January 15. 1987: first full draft report due from , contractors, for limited distri-3 buticm only to Peer Review Group, Maise Yankee, LLNL, 2 or 3 NRC staft January 22. 1987: next meeting of Peer Review Group, in San Francisco, attendance by 4 invitation of PRG only February 12. 1987: second draft due, to be distributed to wider distribution including , Droadly in NRC l , February 19. 1987: meeting in h'ashington with NRC . in-house " Working Group on Seismic Margins". I Another Peer Review Group meeting is planned for sometime after the February meeting, to study the final version of the report to be prepared after the February comments.are in. Members of the Peer Review Group requested permission to visit the offices of the EQE and EI subcontractors toward the end of

the current phase of the project, in order to review ongoing work prior to the completion of the subcontractors' reports.- It was l

agreed that a visit to EI could occur any time after early Decem- , ber, but that visits to EQE should not be held until early Janu-ary. A meeting on the fragilities is scheduled for January 6 at l

                                       ->                                               2                                                        .l

, 'j B-53

       , '/

,^

       ,     .R +
              ~

y . .g' '- ..l_, , , , - . . , - - _ . - , , , - ~ , , - ---~~,-~~r- ex,-~

I i l EQE. The Peer Review Group wishes to emphasize the importance of obtaining access to various analyses and calculations as early as

is feasible, to allow for optimum opportunity for review.

l

2) M. Ravindra of EOE. Inc. gave an overview presentation of the l work of his group, who are the subcontractors doing the fragili- l ties analysis of Maine Yankee. He presented slides that included I preliminary tables on HCLPF values for various components and structures, which elicited much discussion from the Peer Review Group. The question of how best to document and present these results was covered in some detail.

The issue of how to treat dependencies was discussed, especially in regards to the CDFM approach to determining HCLPF values. (For the fragilities approach, the methodology for treating dependencies is more straightforward.) , Discussion about how to standardize the CDFM approach was lively. It was agreed that this issue, including comparison of the CDFM and FA approaches, would be a major topic in the Peer Review Group's final review work for the Maine Yankee project. Some of this discussion covered specific items of equipment at Maine Yankee, which the Peer Review Group agreed would be concen-trated on during their afternoon walkdown of the plant. It was emphasized that the Peer Review Group will need access to all data that will be relied on in these evaluations, in order to review its applicability. This especially includes various aspects of the earthquake experience data base. In this regard, i G. Hardy of EQE agreed to send the recent SQUG report on about 20 classes of equipment to J. Reed.

3) David Moore of EI International next presented an overview of the systems analysis work that his firm is doing. He discussed the preliminary system ' min cut sets' that have been~ developed, including a few interesting failure modes that are still being investigated in detail.

Specific technical topics covered in Moore'c presentation in-cluded load sequencing, starting the turbine-driven emergency feedwater pump after an earthquake, and the reactivity-control systems. The group discussed EI's process for pruning the trees during their development. Moore agreed that his team would publish the entire un-pruned trees, for later use, along with the pruned trees that will be used in the quantification process. 3 B-54 5 l _._ _ _ _ . _ ~ - _ _ . _ _ . . _____ _ _ _ . - . _ _ , _ _ . -

l l l An extensive discussion occurred about how to' combine non-seis-mic-induced failures with seismic-induced failures in the systems analysis. This discussion covered many aspects of this issue. The guidance given to the analysis team is as follows: l The systems analysis team should identify all  ! non-seismic failures that, in combination with a l seismic failure, produce the undesired end-point ' (core damage) with a frequency above a certain 1 cutoff. (The cutoff that EI is now using will include non-seismic failures whose contingent failure probability is above about 0.01 for. failure of a single train, and above about 0.001 for failure of two trains or of a full safety function.) The key . seismic contributors thus identified will then require determination of

;                                  HCLPF values by the fragilities team.                           However, the analysis will ngt quantify an overall HCLPF plant value that includes these non-seismic failures.             Instead, the - combinations of seismic failures and non-seismic failures thus identified will be documented and discussed, so that deci-sion-makers can be aware of their existence and the HCLPF values associated with the seismic failures identified.                   This documentation should be done in such a fashion that a complete HCLPF re-evaluation               that      includes      the non-seismic failures is possible.                                                                      ,

i l The EI group's main work in recent weeks has been development of  ! the system fault trees, which were distributed in preliminary form to the Peer Review Group. EI's main work in the coming period will be to finalize these,' based in part of the new infor-l mation,being gathered at the current walkdown. i i

4) An extensive discussion took place about how to analyze Maine Yankee's batteries. The issue came up because the fragilities team has been unable to develop any defensible technical basis for calculating fragilities for the current lead-antimony bat- l teries. The MY plant will be changing out 2 of the 4 ^ battery I racks during their upcoming shutdown in March, 1987, and may l change out the other 2 battery racks at a later shutdown in a future year. Therefore, the question arose as to which configu-ration should be analyzed in this study. The Peer Review Group's i guidance is that the study team should analyze both the configu-ration in which only 2 of the 4 racks will be changed, and the ultimate configuration in which all '4 will be changed. If the )

l 4 i l B-55 d

l technical basis for developing fragilities for the existing batteries cannot be developed, this fact should be stated in the report.

5) Daniel Guzy. of NRC gave a ' brief summary ~of the meeting- the previous week of the NRC " Working Group on Seismic Margins". He discussed the Working Group's current thinking, including the -l selection of a BWR plant as the subject of a possible second trial reargins review study. The NRC's current thinking is that i they may decide to perform the planned BWR trial margin review in full collaboration with EPRI (same plant, same study).

6) The topic of how the margin review earthquake was selected by NRC was then discussed. John Reed agreed to draft a paragraph for these minutes on this subject, which paragraph is as follows:

Dan Guzy discussed the question of what confidence and probability level the margin earthquake represented (i.e., NUREG-0098 rock median spectrum anchored to 0.3g peak ground acceleration). It was explained by Guzy that the NRC did not use probability concepts to establish the. margin earthquake requirements, but rather adopted the NUREG-0098 spectrum shape to be consistent with a previous licensing decision made for Maine Yankee. John Reed requested as a minimum that the NRC agree that the margin earthquake input represents a uniform hazard spectrum over the ~ entire frequency range. This understanding is consistent with the fragility method being used by the analysis team in developing HCLPF values for structures and equipment. In addition, it allows the NRC to make confidence and probability statementa in the future after the margin analysis is completed, if they so desire. t

7) The Peer Review Group went on an extensive walkdown of the plant on the afternoon of Tuesday, November 18. Items walked i

down were those that were of special interest to the reviewers, based on either their importance to the Maine Yankee margins analysis or their intrinsic methodological interest. After this l walkdown, the Peer Reviewers and the analysis team reconvened to discuss what w a's observed. Interaction with the Maine Yankee , staff and their consultant (R. Kennedy) was important during this l session. i 5 B-56

i l I l OVERVIEW COMMENTS:

1) The main overview comment here is identical to that made after the last Peer Review Group meeting. Specifically, it is the broad consanmus of the Peer Review Groun that the Maine Yankee trial marains review nroiact is beina - accenlimhed so far with accentable technical connetence. Numerous technical and methodological issues have come up, and all have been coped with well by the analysis-team. The cooperation of Maine Yankee and Yankee Atomic personnel has been excellent so far.

2) Although PRA-type information is being developed during the trial margins review, the Peer Review Group wishes to emphasize that it is_probably not usable for full-scopo PRA analysis ner se without significant additional effort.

3) The Peer Review Group also wishes to emphasize that it cannot be expected to endorse the results of the trial Maine Yankee review. If all goes well, the Peer Review Group should fir:d 1 itself able to endorse the methodology.used, and should also be  !

able to. comment on some specific technical aspects of the study. l However, an endorsement of the study results would require a j level of review effort much greater than can be expected from our , ! PRG. l I To summarize, the Peer Review Group looks forward to further l interactions as the project proceeds to final analysis and docu- , mentation of the project results. 1 That completes this report and minutes. Robert J. Budn tz Chairman, Peer Review Grcup Attachments: A = sign-up sheet B = meeting agenda e 6 B-57 l l l

NTf4 M , g fy -Q y _JN W

                                                                               // /        N NT        hat 4L            914ML

! NjML jM10)T/thl 0 U///ddi W Lkt FM, L< - &ir)sL 6J/ll~ b G Art 711 0 m a >-rcs L AlL '(41 ) 4 2. 2 - 4 l Midlael P JTolou 9 4 & l'f la l (Sas'h 246 //92Y l Bli L;sa 21.x4 CVir) 423 -47.s~8 M. V . R avah d m EQE DIN) e s a.-9a %

       .D o w 6> 0 T                            ,N ft C-                   30/ fi3 - 7t 9 9 A 1-       L e         "

lac soi- m-,,so

                                 /

G,DBVWs h&x YAses- . M7- & za -3:r2 -I - L. M 4, m b e r .h_e +. e c n su.- s m _. Roberf YenoreA JL 'd,A,1, (w*d. . 7N - 777-2 M)___ _ _

       'JN w(J                           Unck R. bumW A u. (417) cKM ut l o<-w A % II,'e.                   Ll 5 D ,r n $I4_ Es c-              4IS-3 9 2-bel._

sA7)lL Ur-,,n 61er 04 e 4/ 7- p72-riao be,) / Moore

          ~                                  s e. e %                             zos/ ers-mes Jf                               i T: ,'T%~*s                          Dukn. P,ww co.                        ?o y - 311- y6 ) a 6oh ha.rw/                            L.LNL                              415              '1AA-0308 h ri i e u S e I G2 s                  uR c. /c 2 R.                        n,. 4 e 2 - goo c
                                                           /

NJ/sr// c c#0ks#/ N#c. /NM- 3 o l - 447 P447 l Gre4 NarkJ f/QF 7:4 - FD -9aw I J l h\C.ME,L d. bEM)d EO L 'llL4 PA?_- t % ___ LLW _ __ ___ . ._ fJg . _.. 14 9st -%% D%Mhc.s tAReL.L.\ _.. 3%wEs_dere c.. .. ____Giuhtob_ __ B-58 H- - _._ __ __- _ ~ . . . . . - _ . _ . . . __

f :. I AGENDA Peer Review Group Meeting November 18-19, 1986 Maine Yankee Tuesday, November 18,1986 9:00 a.m. Introduction 9:15 a.rn. Status of Fragility Evaluation 10:C3 a.m. Status of Systems Analysis 10:45 a.m. Open Items 11:00 a.m. Peer Review Group Disciission i 12:00 noon Lunch 1:00 p.m. Plant Tours and Discussions with Plant Personnel 5:00 p.m. Adjourn Wednesday, November 19,1986 8:00 a.m. Plant Tours and Discussicas 10:00 a.m. Peer Review Group Wrap-up Meeting 12:00 noon Adjourn l

                                                                                            )

l 1 l t  ! i i B-59  ; 1 1

FutureReaurcuA.uociates,Inc. 2000 Center Street Suite 418 Berkelty. CA 94704 415-520-5111 27 February 1987 TO: U.S. Nuclear Regulatory Commission D. Guzy, Oifice of Nuclear Regulatory Research P. Sears, Office of Nuclear Reactor Regulation Maine Yankee Atomic Power Plant D. Whittier, Manager, Nuclear Engineering and Licensing - FROht Robert J. Budnitz, Chairman of the Peer Review Group REF: FINAL REPORT: PEER REVIEW GROUP for the MAINE YANKEE TRIAL REVIEW OF THE NRC SEISMIC MARGINS METHODOLOGY This is the final reoort of the " Peer Review Group" that is reviewing the technical competence of the "Scismic Margin Review Study" that is being undertaken on the Maine Yankee reactor plant under sponsorship of the U.S. Nuclear Regulatory Commission. The Peer Review Group has five members: Robert J. Budnitz (chairman) Michael P. Bohn John W. Reed James Thomas Loring Wyllie, Jr. The PRG is pleased to report that the five members are all in agreement with this report, and all en&ro it fully. i l l 1 I B-60

I. Peer Review Groun Charter At its first meeting on July 21, 1986 the Peer Review Group adopted the

 ~ following charter, which was also agreed to by all project participants, including NRC and the utility participants:

The objective of the Peer ' Review Group is to assure 'that the trial seismic margins review, following the guidance established in NUREG/CR-4334 and NUREG/CR-4482, is executed in a fully competent and profes-sional manner, uses methods that are at the state-of-the-art, and takes-cognizance of all relevant information. The sponsors of the study (Lawrence Livermore National Laboratory for the NRC and Maine Yankee - as the plant owner) desire to utilize the results of this study, and require - the Peer Review Group's assurance that the study is technically sound. To accor.plish its objective, the Peer Review Group will be provided full access to all materials, information, and methodologies that are inputs to and used by the study team. Access to the study. team itself will occur through scheduled meetings to follow the study's progress. The Peer Review Group will also review draft reports and participate in walkdowns of the plant. A formal report for the Peer Review Group will be made by the Group's chairman to NRC and Maine Yankee. It is understood that the Peer Review Group's report will be a public document. II. How the Review Was Accomolished The Peer Review Group met on five occasions during the course of this project. Attendance was perfect except that at the first meeting one member could not attend and another's attendance was cut short. The five meetings were:

1. July 21-22-23,1986, at the Maine Yankee site. During this meeting the PRG had the opportunity to walk down the plant, and to meet with all project participants. During this meeting the PRG's charter was discussed and clarified.

2. September 30, 1986, near San Francisco Airport. At this meeting the PRG received presentations and documentation from the project participants, and asked questions to clarify technical issues.

3. November 18-19, 1986, at the Maine Yankee site. During this meeting the PRG had a second opportunity to walk down the plant, to meet with all project participants, to review documentation, and to ask questions.

4. January 22, 1987, at Loring Wyllie's office in San Francisco. During this meeting the PRG met alone with no other attendees, discussed its conclusions about the study, discussed an early draft version of the 2

B-61

                                                                                           .i 4

project report, formulated verbal comments on this early draft .that .! were then relayed to the project team, and laid out the work needed a for the PRG to arrive at its final conclusions.

5. February 19,19g7, at NRC's Bethesda offices.1 The PRG met alone briefly in the morning, spent the main part of this day in a- large meeting with the entire team of project participants as well as NRC and utility representatives, and then met again at.the end of the day to finalize its comments as embodied in this letter.

Besides these meetings, each of the PRG members had access to considerable information during the course of the- project. Technical material was distributed at each of the first three meetings by the project participants

       -(including material from Lawrence Livermore National Laboratory, from their subcontractors at Energy Incorporated and EQE Incorporated, and from the             j utility). This material enabled the PRG to remain fully informed and up-to-
       ,date on project progress, technical problems encountered, and proposed resolutions of various issues.

In addition, PRG members met individually on a few occasions with project participants to allow interaction on technical issues, and of course the participation of PRG members in both walkdowns of Maine Yankee provided hands-on interactions with the project. Copies of the first version of the project's draft final report were distributed to the PRG only six days prior to its fourth meeting. About a month later, copies of the second version of the final report were distributed to the PRG only 3 to 6 days prior to its fifth and final meeting. In both cases, there was not enough time for the PRG to undertake as thorough a review as would have been possib!c if more time had been available. However, the PRG believes that there was enough time to allow the PRG to reach conclusions on each subject in its charter. This is primarily due to the opportunity for PRG interaction with the project team extensively throughout the project, includ-ing in the last several weeks by telephone, personal interaction, and study of preliminary material. Ill. PRG Findines The PRG charter requires findings on several topics. These findings form the substance of our review, and are presented here through discussion of the ' following five topics:

1) The PRG's access to technical information and to the croiect team been fully adeaunte to allow it to reach useful conclusions on each tonic in the charter. The PRG members believe that their interaction with the project I team has been significantly more extensive than is typical of reviews of this {

kind. Access to technical information has been complete and timely, within I the obvious constraints of a very tight schedule toward the end of the , j project. l 3 {

                                                                                              'l B-62 1

I I

                                                                                            .i 1

                                                                                 . _ .       .                    .       _                        _                _,m. _
                                                                                       ~

b i

2) . The nroiect team followed the nuidance established in NUREG/CR-4334

, and NUREG/CR-4482 as fully as could be exnected. In several areas, the 2 guidance provided in the two NUREG reports was found to be incomplete or

                                         - inadequate, and in each case the PRG believes that the project team success-

! fully overcame the specific issues involved. We are grateful that the PRG was consulted on the major issues, a few of which are discussed below. 4 3) The croiect team executed the study in a fully comnetent and nrofessional' manner. The PRG believes that the professional competence of the project team is outstanding, and that the execution of the project has also been outstanding. This is all the more remarkable when one considers the various

constraints on_ the project (such as the fact that this has been a research

!                                           project which was unfortunately carried out as an NRC -licensing action, which led to the problem that various uncertain researchable questions took on a licensing and therefore a financial aspect).
~

There were numerous issues within the project about which professional i disagreements occurred between various participants, or between participants ' and the PRG. All were resolved in a fully professional manner, with open - discussion and honest effort to find the best approach.

4) The croiect team used state-of-the-art methods. In fact, during the project a few advances were made in the state-of-the-art.- Among them were the approach taken to combining scismic' and non-seismic failures in the

systems analysis; the issue of how to reconcile the CDFM and fragilities j approaches to calculating the HCLPF values of components; and the way the

! systems-analysis team was able to use pruned fault trees to develop PRA type systems cut sets with considerable savings of effort.

5) The croiect team seems to have taken connizance of all relevant informa-liga in fact, much information not previously published, or available only in draft or incomplete form, was used by both the~ systems analysts and the fragilities analysts. Information from the Maine Yankee and Yankee Atomic Electric contacts was well used.

! IV. Other PRG Comments i i j A number of technical issues arose in the course of the project that deserve special comment by the Peer Review Group. These topics are the following, { discussed in the next paragraphs: $ i i i 4 .l B-63 i ) 4

  --% --- g ger -y--e ---'t.- - - -
                                    ----w-n.       g---%e --+ m.-nw--.--, e-n3q,c,,      , ,   -p  .-gn  -
                                                                                                           -,p-y. w qy-.= - ,-e -e--ysw-w.-a-y,. g- - - w,      yy.   . -y-9

c l A. selection of the review level earthquake -

  ' B. combining seismic and non-seismic failures                                                      l C.' CDFM method vs. fragilities method for HCLPF analysis D. level of expertisc required by a margins review team E. earthquake experience and test data base F. treatment of relay chatter G. correlations among carthquake-induced failures H. comments on the screening methods used.

A. Review level earthaumke- Throughout the Maine Yankee trial review, there has been confusion as to the precise interpretation of the review level earthquake. The PRG believes that the confusion has now been cleared up (in large part through PRG interaction with the project team and with NRC), and that the analysis performed for Maine Yankee has correctly utilized a consistent review-level earthquake. However, the PRG' believes it essential that more explicit guidance be provided as to what technical features must be established for the review level earthquake, and how it is to be used. (Presumably, the provision of this explicit guidance should be a task for the NRC Expert Paucl, after interaction with the NRC staff and after. seeking advice from industry representatives.) ~ A second comment on this subject, which has been given verbally to the study team, is that the write-up in the draft report version we have studied in mid February is still inadequate in its description of both how the review level earthquake for Maine Yankee is characterized and how it was used. B. Combinine seismic anel non seismic failures. The methodology used in the Maine Yankee review fe combining scismic and non-seismic failures is not a rigorous methodology, in the sense that the final plant-level HCLPF value for Maine Yankee depends on factors that are not completely specified in the guidance. For the Maine Yankee review project, the final HCLPF values for both LOCA and non-LOCA cases have essentially no dependence on this inconsistency, but in a general case there can be differences. The PRG urges that the methodology for this aspect of seismic margins reviews be developed fully (presumably, again, by the NRC. Expert Panel) and documented for general use. It is important that the way' NRC will use the results of a margins review be explicitly considered in developing guidance on this issue. It is also important in this regard to note that there needs to be clarification of how a " seismic margin' is to be interpreted in terms of being greater than the SSE level, especially if non-seismic failures affect the HCLPF value but,- of course, do not affect the SSE level to which' the HCLPF value could in some minds be compared. C. CDFM vs. franilities method for HCLPF ana!vsis. The PRG endorses the recommendation in the Maine Yankee report that further research is needed to develop the CDFM (conservative deterministic failure method) approach to -- analyzing HCLPF values, so that it can be applied routinely. We have learned. much from the comparisons in the current project between HCLPF values calculated for a few components by both the CDFM and fragilities approaches. In particular, we have learned that the CDFM method as currently set down has too much latitude for the analyst and therefore does not provide robust HCLPF values. We urge that this situation be resolved by further research. L 5 B-64

                        . -                . ~ . - . .       .                  . - . .                   -             .    . - .                                ..           --        -                . .

I i D. Level of emnertime needed. Based 'on our interactions with the Maine

Yankee trial review, the PRG believes that a review team must have certain attributes in order to carry out a successful scismic margins review. A utility I whose choice of a review team does not take. these attributes into account I will be less than fully successful in applying 'the methodology. - The main -

l attribute for the fragilities analysis team is that it must have experience in. plant walkdowns, in how to focus on the critical components, and in realis-tically analyzing seismic capacities; otherwise the margins analysis ~will be

burdened by unnecessarily conservative results beyond those already embodied.

in the HCLPF approach. The systems analysis team must be familiar with PRA methods, and will be substantially strengthened by prior experience in PRA analysis of external initiators. Of course, it is not necessary for the . analysis team'to possess expertise at the level that was needed to develop the

margins methodology in the first place, f

E. Earthaumke exoerience and test data base. In the current Maine Yankee -

a review, considerable reliance has been .placed on the earthquake experience and test : data base for various components. The . draft report we have ji reviewed contains citations of much of this data base. Even -though some ! members of the review team are among those who have developed the data base, its interpretation for the purposes of this study has been problematical l in a few situations because some of the data base is not yet publicly avall-i able. For other analysis teams in the future, this issue will be also exist. I Therefore, it is essential in future studies that the data base relied on be thoroughly documented, and that its interpretation by the analysis team be set down in detail in future written reports. Guidance to future. analysts along these lines must be strengthened (presumably,' again, by the Expert

Panel) to assure that future reviews are carried out acceptably.

;                                           In addition to the documentation aspect, it is important that the guidance be 1                                            more explicit regarding the type and configuration details for components j                                            covered by the screening guidance, because some analysts in future routinc

} applications may not be as familiar with the underlying raw experience data j as were the analysts performing this Maine Yankee trial review. i F. Treatment of relav chatter. Following the guidance of NUREG/CR-4334, the issue of relay chatter and its recovery by operating crews has not been . covered in the Maine Yankee review. The PRG wishes to point out that considerable progress in understanding relay chatter has recently occurred, and that the guidance for future analysts may be able to be strengthened in j this aspect. (This is presumably' another task for the NRC Expert Panel, with . input from electrical circuitry and systems analysis experts). G. correlations amona carthauske induced failures. The handling of I correlations among earthquake induced failures is a methodological problem. 4 These correlations can exist in several areas, including the fragilities, ~ the i responses, and the systems aspects. Although there are methodological weaknesses in the approach that was used, they do not affect the plant level i HCLPF results at Maine Yankee, and the review team's treatment is therefore i' acceptable. However, there remains a general methodological question as to l how these correlations are to be treated, which the current guidance in 4' NUREG/CR-4334 and 4482 does not cover well. The problem is that, while 6 I B-65 i . {- ,

    -,,--..~e..-mo.,e,,     w.-,-v,,,--,,v---.ww.,            ,,-.-.--y-.ya----         ,-,-vv..,www-,,,.-pr,.-w-,                 . m-r  #w. ~..-----r,.-.m    . e  ..,,-.-.--,-,3-, -m   -e.w.-.,- u e - e m  .r.

1 there are exact methods for handling these correlations in' a full scope and' realistic seismic PRA study, their treatment in the approximate HCLPF analysis requires 'some explicit guidance not yet available. This area is in need of methodological' development before the HCLPF methodology can be considered fully acceptable, in . addition, the guidance needs to be more explicit concerning when correlations should be assumed to be 100 % --- and hence treated implicitly in the basic event definitions in the fault trees -- and when they should be treated explicitly in the numerical evaluations of the cut sets. H. Comments on the screeninn methods used. The PRG has carefully

                             . studied the methods used by the Maine Yankee analysis team in its several 4                              screening tasks during the study. (This has been one of the most important reasons why PRG interaction durina this . trial review has been so valuable.)

The approach taken, which employed an iterative process between systems-i based screening and fragility based screening, seems : to have been very , successful in reducing the amount of analysis work required while retaining ! the key items for final study. The PRG wishes to affirm its endorsement of i this iterative and interactive approach. It also wishes to endorse the original guidance (in NUREG/CR-4334 and 4482) on.how important a walkdown type screening step is even for those components that are thought be screenable out based on generic guidance in the NUREGs. Only by such walkdown work (which is not necessarily a 100 % walkdown of all components) can there be high assurance that specific components do in fact fit within the ' generic categories for which the generic guidance is applicable. j 1. Additional comments on the screeninn muidance: There are a few other areas where the Peer Review Group believes that the screening guidance - should be improved. Among these are the approach to screening or analyzing batteries, especially aged batteries; the approach to walking down and analyzing fragilities for small LOCAs, in recognition of the fact that a thorough walkdown of all small lines in the primary pressure boundary may be prohibitively difficult, so that a screening approach may be needed; the . ! treatment and documentation of system successes in the event trees; and , how to treat component interfaces and ' super-components' such as multiple items on a single skid. i V. Level of Peer Review Grouo Effort i The PRG has been asked to comment on whether the level of review effort I expended by the Group has been adequate; and more generally on how much effort should be devoted to peer review of scismic margins studies in the l future. , 1) On the first issue, the PRG is unanimous in believing that the peer

review effort expended on this trial review has been a necessary part of accomplishing the project's objectives effectively. This project was exploring a new methodology, determining what are its strengths and weaknesses, and performing the research work in the context of a licensing action. This combination of ingredients would have been much more difficult to accomplish.

well without the continuing interactions that the project team had with our 7 f B-66 i t.

     , . - - -  -.n.- .,-...        ,      , ~ - . , . , . . . - . . - - - ,-- .- -     -

n, , . , , .- . . . , , .- - - . _ , . - , - ,

l l I Peer Review Group. These continuing interactions allowed for the timely resolution of several technical issues that arose during the project. If there-had been no PRG, or if the peer review activity had only been a review of the final report once completed, we believe that the success of this project in accomplishing all of its objectives would probably not have been as outstanding. We also believe that all of the project participants agree with this conclusion.

2) On the second issue (concerning peer review of seismic margin studies in the future), our PRG members are divided. Here we are considering the situation in which the margin review methodology has become relatively

               " standardized" and a review is being undertaken by a utility, on either its own initiative or NRC's initiative.

(i) Some PRG members believe that carrying out a successful seismic margin study reauires the interaction of a group of outside reviewers, experts in the disciplines involved, to provide technical review as a way of assuring the validity of the study results. The rationale for this position is that it will be difficult for a utility to perform such a review competently, especially for routine applications of the methodology in which review teams are composed of engineers who may not be truly expert already in the technical aspects of these subjects. (ii) Other members of the PRG believe that, once the methodology has been developed into a more-or less standard methodology, the necessity for

,                     peer review will be no different than it is for any other engineering study performed by utilities on their plants. In this view, the internal review procedures that utilities already use to assure the technical validity of their
                    -internally supported analyses should be just as adequate in this arena as they are in the numerous other engineering arenas aircady part of reactor safety analysis.

(iii) There is yet a third view among the PRG members that can be summarized as follows: outside peer review is not an essential attribute of e.ny routine utility supported scismic margins study, but its presence will surely enhance the study's credibility with NRC, and hence the NRC's need for its own separate review may be less. Therefore, outside peer review should be considered carefully by any sponsoring utility, especially if the margins review team itself is not fully familiar with all of the technical aspects of performing these studies. This division of views among the PRG members on the need for routine ' outside peer review does ag.t conflict with the unanimity of views on the

usefulness of the peer r. view exercise for the current project, which was a i

trial (research) project. In particular, the value of ongoing interactions throughout he course of the trial project cannot be overemphasized, l l l I 8 B-67 w yy-erg a -.ee = r_ g- -

          *m..   .-.9    ,-       ,,__r~     _     ,y_,     _q.#       y         9

l VI. Final Remarks In conclusion, it is our pleasure to acknowledge the outstanding cooperation and assistance of all parties who participated in this trial margins review. The PRG found excellent cooperation from all five parties to the review effort: the NRC staff; the utility team from Maine Yankee and Yankee Atomic Electric; the Lawrence Livermore staff; and LLNL's subcontractor teams from EQE Inc. and Energy Incorporated. The PRG believes not only that a project like this should be technically competent (it was) and timely within schedule and budget (it was), but also enjoyable for the participants, including both project team and reviewers. For making this last aspect come true, we thank one and all. i Robert J. Budnitz Chairman, Peer Review Group 1 1 r I l 9 B-68 i i 1

e - 2 t APPENDIX C .; UTILITY COMMENTS 1 i C-1 ,

   ,r     .. ~.. _           ,     , , _ . , _ ._ m ,

l 1 l

                                                                                                                                                                      )

l MAIRE BRRHEE 'AmmicNWERCOMPARU e ,6,, ,OneeSrenROAo FR AMINGHAM, MASS ACHUSETTS 01701 ENGINEERING OFFICE v- 617 872 8100 0 MAINE YANKEE SEISMIC DESIGN MARGIN PROGRAM UTILITY PERSPECTIVE i Prepared By: W. E. Henries Lead Mechanical Engineer J. T. McCumber i Project Engineer 1 December 1986 l l l l l C-2

I. PURPOSE This' report provides a description of the Utility role in a Seismic Design Margins Program. This description will assist potential generic implementation by documenting our experience with the methods utilized, costs incurred, and impact'on normal Utility operations. 3

                                          -1 5210R C-3

i II.- BACKGROUND d i Maine Yankee tuut been involved with the issue of demonstrating seismic margins for several years, with particular. focus generated by the 1982 i New Brunswick earthquake. In addition to identifying the inherent conservatism in Maine Yankee's seismic design, a series of voluntary reviews and walkdowns were initiated in 1982 to find any potential weak areas. These plant reviews were performed by known experts in the fields of structural and seismic design. Several cost 4 beneficial upgrades, primarily associated with equipment anchorage, were. identified which if made would enhance seismic adequacy.- The equipment anchorage and supports upgraded by Maine Yankee are listed in Table 1. 4 j Maine Yankee and Yankee Atomic Electric Company have been very active in the efforts of the Seismic Qualification Utility Group (SQUG). The j equipment at Maine Yankee has been favorably compared to the SQUG experience data base as it was developed. ] When assessing the applicability of the Maine Yankee experience to the i generic Seismic Design Margins Program, it should be recognized that Maine i Yankee may not represent the typical state of seismic adequacy for a plant of j its age. This is due to upgrades associated with the seismic review and enhancement program described above. J I 1 a i

                                                                                                -1 l'           5210R j                                                                                                 C-4 4

I^

l III. INFORMATION RETRIEVAL / ANALYSIS SUPPORT I l The primary role of the Utility in the Seismic Design Margins Program f was to provide a complete and accurate description of the operation and -I construction of the Maine Yankee plant. It was first necessary to provide a description of the safety-related systems, their functions, and interdependencies. The Analysis Team (AT) could then properly categorize them as " Group A" or "Not Group A" Systems (Reference 1). In addition to the systems information needed for the Group A screening, it was also necessary to provide seismic design details to the AT to aid in their seismic capacity determinations. A summary of the information provided to the AT is contained in Table 2. Once the AT systems modeling and fragilities analysis began, the Utility role evolved into a question response mode. Specific system operation or equipment anchorage / design questions were forwarded to the Utility and the requested answers or design details were returned to the AT. Many of the questions were readily answered but, occasionally, specific details were not available and required more extensive review. Examples of the latter included reactor internals design details and anchorage details for original plant equipment which was not wi. thin the scope of the previous seismic review programs. In order to obtain the necessary reactor internals and control rod drive information, a meeting with the NSSS vendor was arranged and attended by representatives of both the Utility and the AT. The meeting was a success, and the information required to perform the fragility analyses was obtained. Similarly, when original installation information was not readily available, , searches through microfilm archives, or actual field measurements were required to answer the AT questions. The availability of data is a major factor in determining the Utility workload and associated costs. The final role played by the Utility was to review the systems models and assumptions developed by the AT. This effort involves close scrutiny during the entire development process to insure the accuracy of the model and 3 5210R C-5

       - its assumptions. It is important to prevent misconceptions or easily upgraded
        " weak links" from entering into the s'odels.

Following this approach, Maine Yankee agreed to implement post-earthquake procedural guidance and to make minor anchorage upgrades to increase the confidence level of the system models and equipment fragilities. The total Utility effort was accomplished without major impact to normal operations. As discussed in Section V, slightly more than 1,000-man-hours were dedicated to the program over a ten-month period. Given advance planning and an absence of a refueling outage, permitted the engineering hours shown on Table 4 to be provided by the existing engineering personnel. 4 l 6 Y f I l l r l

                                                 -h-
        $210R C-6

IV. PLANT WALKDOWNS s General The Seismic Design Margins Program included two walkdowns of the Maine ] Yankee plant. The first walkdown, lasting a week, was performed early in the program to obtain:

1. A firsthand understanding of Maine Yankee's structures, systems, and equipment, 1

2. Equipment anchorage details, and

3. Assurance that plant-specific problem areas or seismic interactions were not a concern.

1 The final plant walkdown, lasting three days, was performed near the end of the program following the completion of pre 1Luinary fault tree and fragility analyses. The objectives of the final plant walkdown were tot 1

1. Cather additional information on systems and equipment needed to complete the analyses,

2. Obtain equipment operating and maintenance information,

3. Insure the validity of system interaction assun:ptions, and

4. Resolve open issues.

The walkdowns were each scheduled in conjunction with Peer Review Group meetings, so in addition to the AT, the walkdowns included the Peer Review Group (PRG), NRC, and Lawrence Livermore (LLNL) personnel. The specific interests of each group had to be carefully considered in the walkdown planning. 5210R i C-7

  +--                     -                              -e

l Walkdown Planning In addition to the actual walkdown activities, PRG meetings, plant access training, body counting, getting dosimetry, and miscellaneous information gathering meetings had to be scheduled within the walkdown time period. Considering all of the above, a productive walkdown required careful preparation and planning. Coordination with the AT well before the actual walkdowns was necessary-to determine the areas of the plant and specific equipment of interest. -This advanced notice allowed the Utility Walkdown Coordinator to familiarize himself and other escorts with the exact areas of the plant to be visited, as well as the system's function and anchorage details of the equipment to be inspected. Pre-walkdowns, marked-up drawings and equipment lists were all helpful tools in assuring efficient use of the limited time available. Other preparations which facilitated plant walkdown activities included:

1. Sending out all of the required health physics forms with instructions to allow walkdown members to arrive at the site with the necessary paperwork completed.

2. Scheduling body counting during the PRG meetings (approximately 15 4

minutes for each person - took turns leaving the meeting).

3. Notifying key plant personnel of the walkdown schedule and assuring their availability, as needed,

4. Reserving meeting rooms and work rooms, and 4

5. Obtaining necessary approvals for camera passes.

Walkdown Logistfes The plant walkdowns were made up of a large group of people (around 20), of varied interests, and conducted in a limited amount of time in a l 2 limited access environment. Therefore, complex logistics were to be expected. 1

                                                                                        .\

C-8 l 5210R l

_, .... -- ~ - - . . . . - . . _ . . 5 i Plant security regulations limited the number of unbadged people which ' could be. escorted throughout the plant by one escort to five. This meant that I four or more prepared escorts were necessary at any. time. Also, it was. . {~ -important to match the escort's expertise to the group; i.e., systems engineer- ,

                    .for a group seeking plant operations information and a structural / mechanical l                      engineer for a group looking at anchorages.

The objectives of the walkdown members are generalised below:.

Analysis Team - Gather data on equipment throughout the plant.

                                                                                                                - Obtain information from plant operations, engineering, and maintenance personnel.

Peer Review Group - Obtain a good overview of the plant.

                                                                                                                - Inspect equipment of specific interest to them.

l 1

                                                                                                                - Knowledge of AT activities.

NRC and LLNL - Obtain general. overview of plant. j - General knowledge of AT activities 4 (programmatic). A preplanned tour of the plant was arranged at the beginning of the l initial walkdown for all of the walkdown members. The list of plant ~ areas and j equipment which the AT wanted to see was used to develop the tour. This j assured that everyone involved received a good overview of the plant and had a chance to inspect the equipment of interest. ! Subsequent plant tours were arranged to facilitate the specific information which had to be gathered. During the initial walkdown, the PRG could only stay three days, so an effort was made to meet their plant walkdown > needs during that time. Otherwise, the priority was to assure that the AT

  .                    obtained information needed.

J

Meetings were scheduled each evening with all walkdown participants to

discuss progress and potential problem areas. Also discussed, was the 4

schedule of the next days activities and information which may be needed. I * ' 5210R

C-9 1

2 ___,.n --_ . . . . ~ . . . . - _ _ _ , _ . _ , ,_ ..-. _. _ ,.._._ _ ._. ..-.- - . _. .. _ . -._ . _ _ - . _ . _ . ~ _ - , . - . - . .

l This allowed time for preparation that night, if required. These meetings were extremely productive, and while they made for a long day, are highly recommended. Plant Impact As a result of the preplanning associated with the walkdowns, the impact on normal' plant operations was minimized. Additional steps which were taken to lessen the impact included:

1. Staggering the time which various walkdown groups pass through HP to minimize their processing time and to avoid congestion in the change area.

2. Keeping the number of people in the Control Room to a minimum, preferably doing Control Room work during the evening shift.

3. Using simulator and simulator staff to answer operations questions I where possible.

i

4. Letting operations or security personnel know in advance of locked areas which must be accessed.

5. Letting HP personnel know in advance when access to a high radiation area is required so they may arrange proper support.

6. Having radiation work permits made out in advance.

5210R C-10

                                                                      .e i

1 V. COST . I The Utility costs to support the Seismic Design. Margins Program are summarized in Tables 3 and 4.- Several hbservations.regarding these costs are worth noting. y First, the majority of man-houra' vere spent by Mechanica1' Engineering; since that discipline has historically been responsible for seismic review , . programs at Maine Yankee; thus, thel$ulk of data retrieval and program i management fell to them. Systems Engineering'also required considerable 3 man-hours. Theirinput_wasvitalinddentifyingtherequiredsystemsand s' components and assuring plant unique operational features were accurately modeled. The Licensing man-hours and Consultant costs are attributable to the fact that this pilot research program also had licensing implications for the-Utility. .

;                                            Secon,11y, the relatively high travel costs and miscellaneous expenses are attributable to the fact that the AT consultants are all'from the West Coast, while the subject plant is on the East. Additionally, the tight program schedule requirements required frequent overnight' mailings and phone calls.                                                              ,

a-

Finally, the total program eff ort may I dary for other utilities for two l reasons. Reduction in the data retrieval time for the Seismic Design Margin l Program was greatly aided by the walkdown documentation and equipment anchorage upgrades previously performed at Maine Yankee. Since 1982, three separate plant walkdowns have been conducted at Maine Yankee as part of its J

voluntary seismic review program. As a result of these walkdowns,-many equipment anchorages and masonry walls have been strengthened beyond FSAR requirements. These upgrades, and the relatively recent design at' lyses, I allowed easy information retrieval for the Utility and fragility. screening for j the AT. On the other hand, the time involved with retrieving original design i data for those components not wituin the scope of'the previous walkdowns was j significant. The inability to readily retrieve old (16 to 18 years) design records or to obtain as-built details due to radiation or operational ' 5210R ,- , C-11

   --,--,r.,--   - ., __ y.. -    ,,e,,,-w_-     --,w   ,          .,,,,y .
                                                                            ,,-._..,.n   . ,      .,..~,.. , ,.,,-     ,,. , , _ , . .     ,r.- _ , ,. n.mo,m.      . . . , . .     ,.e~,._e

I constraints prevented the Utility from being.as responsive to AT requests as we would have liked. Other utilities should thus estimate their man-hours accordingly. i

                                                                                                                     't 1I
                                                                                                                  }.

4 I l 5210R 'l l C-12 l l l l .,

j. .

l'^_. . _ , . _ _ ..._ __ .. . . , _ , _

i VI. IMPRESSIONS / REC 0!MENDATIONS

              -The concept of using plant walkdowns and experience screening to evaluate seismic margins is excellent and long overdue. The value obtained by

.- eliminating those components shown by experience not to be seismically fragile is enormous. Just as important, the walkdown methodology provides a thorough i systems interaction review. This program is a major step in developing a cost effective method of identifying plant-specific seismic margins and, perhaps more important, easily fixed seismic " weak links." While we praise the overall thrust of the program, several recommendations are provided below which may improve future reviews.

1. Separate the seismic review from the licensing area. This change will allow closer Utility /AT cooperation without the glare of.the "public meeting" requirements.

2. Define the review earthquake with greater precision. The ongoing concerns regarding the confidence level of the ground response spectral shape and ZPA resulted in too much wasted effort.

I

3. Precisely define the need for long-term safe shutdown (Group B functions) capability. Considerable effort was expended on these components which we believed were excluded from the review scope per Reference 1.

4. Provide better field guidance on the use of the SQUG experience data base (i.e., MOV/A0V interaction, cable trays). i I

l

5. Develop improved methods of determining / defining "high confidence" levels for components such as yard tanks. As demonstrated by our I experience, the significance of yard tank capacity is high and the analytical methods for determining actual capacity seem unnecessarily conservative.

6. Decide if nonseismic failures should be included in the HCLPF determination and, if so, how.

5210R C-13

VII. REFERENCES

1. NUREG/CR-4482, "Recomunendations to the NRC on Trial Guidelines for Seismic Margin Reviews of Nuclear Power Plants," by P. G. Prassinos, M. K. Ravindra, and J. B. Savy, March 1986.

i i a I d I w J a 12-5210R C-14 l

      ,,              --+     -                    - w      gr----ee--=-e-e   + m- " '

l TABLE 1 Maine Yankee Equipment Anchorage and Support Uparades Emergency Buses 5, 6, 7, and 8 Battery Chargers 1, 2, 3, and 4 Battery Inverters 1, 2, 3, 4, and 5 Floor Mounted Diesel Generator Control Panels Main Control Board Control Roor Auxiliary Cabinets Electrical Control Board Radiation Monitoring System Heat Jracing Cabinet Air Conditioning Control Panel EHC Panel Reactor Regulating System Feedwater Regulating System Vibration and Loose Parts Monitoring Cabinet Reactor Protective System Core Loading Panel Radiation Monitoring System Meteorological Survey Cabinet Safety-Related Instrument Racks Masonry Wall Reinforcement Service Water Piping Support 5210R ., C-15 l l

                                                                                       )

TABLE 2 Summary of Information Provided for Maine Yankee Seismic Desian Martins Review o Maine Yankee FSAR o Ground and Floor Response Spectra for 0.18g NUREG/CR-0098 an 0.2g Regulatory Guide 1.60 Earthquakes o Approximately 15 Specifications o Approximately 40 Calculati6ns (or portions thereof) o Approximately 550 Drawings o Emergency Operating Procedures o Pump NPSH Curves o " Tagging" and Modification Control Procedures o Maine Yankee Technical Specifications o Maine Yankee Training Manuals o Equipment Testing Frequencies and Procedures o Abnormal Operating Procedure for Flooding of the Circulating Water Pump House o Abnormal Operating Procedure for Earthquakes o Reports on: Class I Structures Dynamic.Modeling Components Required for Accident Mitigation and Safe Shutdown Reactor Protective System Appendix R Alternate Shutdown System Maine Yankee Masonry Walls 1983 Seismic Walkdown and Modifications Seismic Review Program Summary Maine Yankee Seismic Hazard Analysis (UHS) Conservative Seismic Capacities of Maine Yankee's Reactor Containment Evaluation of LOCA-Related Loadings on the Reactor Coolant System Expansion Anchor Information and IEB 79-02 Anchor Bolt Testing 5210R C-16

TABLE 3 Maine Yankee Seismic Martins Program Costs

• Man-Hours Dollara Mechanical Engineering 585 Systems Engineering 195 I&C Engineering 15 Electrical Engineering 45 Nuclear Engineering ** 75 Licensing 100 Plant Engineering / Operations 45 2

Consultant (Includes Travel) $21,000 Travel 11,400 Miscellaneous 1,525 (Drawings, Xerox, Mail, Phone) l TOTAL *** L.QfiQ 133.925 4 .l

• Does not include cost of anchorage upgrades. l

   ** Includes PRA review.
  *** Does not include costs to review final report.

l 5210R ' C-17

                                                                     - TABLE 4 Monthly Man-Hour Breakdown

i. Month Man-Hours May 67 June 50 July -170 August 154 .

,                                    September                                                                     88 October                                                                     .159 November                                                                     171 December                                                                      66                                      .

January 100 J February 35 TOTAL 1.d60  ! t i k

                                                                                                                                                              +

4 r k l 5210R , C-18

                      .r
                                                              ~                       '

g- - 7 m y - 7i- yy y--m - -y --,p .m .r,wwwW- _~w igyyy 4 =.g-.-e--. g w v g. g-44d ^Fwg- ww- w vi ~ w

i MRRE BRIMEE 'AT0l8CNWERCOMPARU e suoy,7,, ugmg pax

       %                                                                                 (8o7) 42 2 8
                      .                        February 26, 1987 WW-87-22                                (IM-87-43 1

01 rector of Nuclear Reactor Regulation United States Nuclear Regulatory Comnission Washington, D. C. 20555 Attention: Mr. Ashok C. Thadant. Director PHR Project 01 rectorate #8 Olvision of L.icensing References; (a) i.tcense No. OPR-36 (Docket No. 50-309) (b) HRC to H.Y. letter dated (c) H.Y. to NRC letter dated

Subject:

Maine Yankee Seismic Margins Program Gentlemen: For the past alght months. Halne Yankee has been working with the MRC staff and its consultants on a seismic margins prograr. This program, was designed to desenstrate the feasibility of the margin assessment methodology, and to resolve NRC staff concerns about the adequacy of Natne Yankee's selsste design. The program has been completed and a draft report issued. The program has demonstrated that the Maine Yankee plant possesses seismic ' ruggedness well beyond the plant's original design. The program also identifled several areas where the seismic ruggedness of the plant could be effectively enhanced. The changes that we have decided to make along with the scheduled completion dates are ilsted below. Description of Schedule Completion l Change Date j

• Diesel fuel day tank anchorage upgrade done i

        ' Control Room cooler anchorage upgrade                            done                           i

• Helding cart / gas bottle tiedown done j
• Security lighting tiedown done

        ' Maine control board 41 ara tiedown                               done                           ;
        ' Strengthen blockwall VE 21-1                              1987 Outage
         Upgrade anchors for fans 44A & B 1967 Outage

• Install internal anchors for transformers 507 & 608 1987 Outage
• Replace safaty class batterles 1 & 3 1987 Outage '

l

• Replace safety class batteries 2 & 4 1988 Outage
• Refueling water storage tank anchorage upgrade 1988 Outage 8492UGDW C-19

uAswt VAwwte ATOMIC pCwsa coupANy United States Nuclear Aegulatory Camelssion Page Two Attention: Mr. Ashok C. Thadant, Director m.87 22 i l The draft seismic margins report indicates that the Maine Yankee plant could withstand setsele events well in excess of the plant's original design bests without endangering pubile health and safety. The report also indicates that the limiting coeponent is the refueling water storage tank (RMST) with a high confidence of low probattlity of fatture (helpf) of 0.21g. As indicated in the foregoing table, we plan to upgrade the seismic ruggedness of this tank, by leproving its anchorage, no later then the 1988 refueling outage. He are making an effort to complete all or part of these upgrades during the 1987 refueling outage. Although the anchorage upgrades have not been finally designed, we expect to upgrade the tank's seismic ruggedness to a helpf value in the range of 0.279 . He would appreclate your concurrence that completion of these committed modifications resolves this long standing Ilcensing concern so that we may apply our attention and resources to other safety enhancing efforts. Yery truly yours, lu!NE YANKEE ATG4fC PCHER CONPANY bhAV G. O. Whittler, Manager Nuclear Engineering and Licensing GDN/bjp Enclosure cc: Dr. Thomas E. Nurley i Nr. Pat Sears } Mr. Cornellus F. Holden l l l i M92L/ CON C-20

                                                        -          . _. _          .               . ..}}