ML20083E941

Proposed Findings of Fact & Conclusions of Law on Design Qa. Idvp Fails to Provide Assurance That Facility Design Conforms to Licensing Criteria & Regulatory Requirements. Certificate of Svc Encl

ML20083E941
Person / Time
Site:	Diablo Canyon
Issue date:	12/23/1983
From:	Ordin A, Strumwasser M, Van De Kamp J CALIFORNIA, STATE OF
To:
References
ISSUANCES-OL, NUDOCS 8312300052
Download: ML20083E941 (100)

Text

.

00CMETED

~* USNEC UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISS ION '83 DEC 29 N0:34 BEFORE THE ATOMIC SAFETY AND LICENSING APPEA4: BOARD:

In the Matter of )

)

PACIFIC GAS AND ELECTRIC COMPANY ) Docket Nos. 50-275 0.L.

~

) 50 -32 3 0.L.

(Diablo Canyon Nuclear Power )

Plant, Units 1 and 2) )

)

PROPOSED FINDINGS OF FACT AND CONCLUSIONS OF LAW OF GOVERNOR DEUKMEJIAN JOHN K. VAN DE KAMP, Attorney General of the S tate of Calif or nia ANDREA SHERIDAN ORDIN, Chief Assistant Attorney General MICHAEL J. STRUMWASSER, Special Counsel to the Attor ney General SUSAN L. DURB IN ,

PETER H. KAUFMAN, Deputy Attor neys General Attorneys f or Governor George Deukmejian 3580 Wilshire Boulevard Suite 800 Los Angeles, Cali f ornia 90010 (213) 736-2102

..]A U

J

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSDiG APPEAL BOARD

)

In the Matter of )

)

PACIFIC GAS AND ELECTRIC COMPANY ) Docket Nos. 50-275 O.L.

) 50 -32 3 0.L.

(Diablo Canyon Nuclear Power )

Plant, Units 1 and 2) )

)

PROPOSED FINDINGS OF FACT AND CONCLUSIONS OF LAW OF GOVERNOR DEUKMEJIAN JOHN K. VAN DE KAMP , A t torney Gen e ral of the State of Calif or nia ANDREA SHERIDAN ORDIN, Chief Assistant Attor ney General MICHAEL J. STRUMWASSER, Special Counsel to the Attor ney General SUS AN L. DURB IN , s PETER H. KAUFMA:.',

Deputy Attor neys General Attorneys f or Governor  !

George Deukmejian 3580 Wilshire Boulevard Suite 800 Los Angeles, California 90010 (213) 736-2102 .

, UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSDiG APPEAL BOARD

)

In the Matter of )

)

PACIFIC GAS AND ELECTRIC COMPANY ) Docket Nos. 50-275 0.L.

) 50 -32 3 0.L.

(Diablo Canyon Nuclear Power )

Plan t , Units 1 and 2) )

)

PROPOSED FINDINGS OF FACT AND CONCLUSIONS OF LAW OF GOVERNOR DEUKMEJIAN Governor Geor ge Deukmejian hereby respectf ully submi ts the following Proposed Findings of Fact and Conclusions of Law on Design Quality Assurance.

DATED: December 23, 1983 JOHN K. VAN DE KAMP, Attor ney General of the State of California ANDREA SHERIDAN ORDIN, Chief Assistant Attorney General MICHAEL J. STRUMWASSER, Special Counsel to the Attor ney General l SUSAN L. DURBIN, PETER H. KAUFMAN, Deputy Attor neys General By / . #h j <- -~

]

l MICHAE' J. STRUMWASSER Attorneys f or Governor George Deukmejian 3580 Wilshire Boulevard Suite 800 l Los Angeles, California 90010 l (213) 736-2102 i

. 'IOPICAL INDEX INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 FINDINGS OF FACT . ......................... 5 I.

THE IDVP HAS FAILED 'IO PROVIDE ASSURANCF THAT THE DESICN OF DIABID CANYON 'JNITS 1 AND 2 CONFORMS 'IO THE LICENSING CRITERIA AND REGULATORY REQUIREMENTS . . . . . . . . 5 A. To Meet the Requirements for a Verification Program, the Applicant Must Show Either That the Dasigr.

Fully Conforms to Licensing and Regulatory Requirments or That the Design Has Been Subjected to the Same Level of Scrutiny It Would Have Received Under an Adequate QA Program . . . . . . . . . . 5

1. The IDVP and the Staff Made Contradictory Claims as to Whether the Verification Program Demonstrated That the Design Conforms to the Licensing Requirements . . . . . . . . . . . . . . . . 6

2. 'lhe Fact That a Lawful QA Program Would Not Guarantee an Error-Free Design Does Not Render the Existence of Uncorrected Errors Irrelevant . . . . 8 B. The IDVP Has Failed to Show Either That the Design Fully Conforms to Licensing and Regulatory Requir ments or That the Design Has Been Subjected to the Same Level of Scrutiny It would Have Received Under an Adequate QA Program . . . . . . . . . . 9 l 1. It Is Virtually Certain There Remain Undetected Errors That Cause the Design to Fail to Ccnply With the Licensing Criteria .. . . . . . . . . . . . . 9

2. The IDVP Has Not Subjected the Design to the Same l

Level of Scrutiny That It Would Have Received Under an Adequate QA Program . . . . . . . . . . . . 11

a. The IDVP's verification of the seismic design consisted of a sampling of the ITP design review work ............. . . . . . . 11

b. The IDVP's verification of the non-seismic design consisted of a sample of certain parts of the original non-seismic design work . . . . . 11 l i. The IDVP failed to verify the work of

, contractors performing safety-

) related design work . . . . . . . . . . . 12

11. The IDVP suple of IGE's non-seismic design consisted of a non-randcm sm ple of that work . . . . . . . . . . . 14 1' .

.J

, C. None of the Reasons Given by PG&E for the Board to Ignore Its Failure to Meet the Licensing and Regulatory Requirements is Valid . . . . . . . . . . . . 16

1. The Claim That There Renain No Undetected Errors Having Safety Significance Is Unfounded . . . . . . 16

i. The evidence does not support any conclusion about the safety significance of the errors identified by the verification program . .,. . 16

11. There is no basis for making any -statemenet about the design errors that renain undetected ..................17

2. There Was No Evidence to Support the Suggestion That the Undetected Errors Could Be Ignored in Licensing Because They Would Be Detected in Preoperational Testing . . . . . . . . . . . . . . . 18

3. Vague Claims About Overall Engineering Quality Are No Substitute for Either License Compliance or an Adequate QA Program . . . . . . . . . . . . . 19 D. By Ignoring the Scientific Requirements for Drawing Conclusions frm Sarnple Observations, the IDVP Has Rendered Its Findings About the Sampled Design Work Inapplicable to the Unsampled Portions of the Design . . 20

1. The IDVP Violated the Comission's Requirements arri Its Own Comitznent by Failing to Avail Itself of Capetent Statistical Expertise . . . . . 20

2. Contrary to the Claim of the IDVP, It Had Never Been Relieved of the Obligation of Having the Verification Reviewed by a Cmpetent Statistician . 20

3. The Opinions of the IDVP, Staff, and PG&E Were Arrived at Without the Benefit of Cmpetent Advice on the Applicability of Statistics to the IDVP . . . . . . . . . . . . . . . . . . . . . . 22

4. Statistically Valid Rand s Sampling Is Essential to Generalization fra Sample Observations to the Unsampled Portions of the Population . . . . . . 23

5. It Is Possible to Conduct a Statistically Valid Design Verification Program . . . . . . . . . . . . 24

6. Proper Sampling Is Not Only a Legal and Scientific Requirenent, But Plainly Necessary to the Objectives of the Verification Program . . . . . . . 26 E. The IDVP Has Not Verified the Design of Unit 2 . . . . . 27 ii.

)

'II.

• THE ITP DOES NOT PROVIDE THE ASSURANCE MISSING FROM THE IDVP THAT THE DESIGN OF DIABID CANYON MEETS THE i

LICENSING CRITERIA AND APPLICABLE REGULATORY REQUIRD4ENTS . . 28 A. The ITP's Seismic Verification Program Provides Only a Limited Substitute for a Lawful QA Program . . . . . . 29 B. The ITP's Non-Seismic Verification Provides No Basis for Drawing Anf Conclusions About Those Portions of the Design Not Sampled by the IDVP . . . e . . . . . . . . 31

1. Prior to the Hearing, the ITP's Phase II Program Was Not Represented as an Independent Verification Effort . . . . . . . . . . . . . . . . . 31

2. At the Hearing PG&E Sought to Recharacterize the ITP as a Verification Program of Larger Scope Than the IDVP ..................s . 31

3. There Is No Substantial Evidence to Support PG&E's Claim That the ITP Reviewed 45 Percent to 50 Percent of the Non-Seismic Design, Nor to -

Support the Claim That the IDVP and the ITP Togeth9r Reviewed 75 to 80 Percent . . . . . . . . . 34 C. The Probabilistic Calculations Presented by PG&E Panel 6 Do Not Constitute Reliable Evidence on Any Material Issue Before This Board . . . . . . . . . . . . 36 D. To Ccrnplete Verification of Diablo Canyon, PG&E Must Establish for Each Portion of the Design Either That Itis Free of Class A or B Design Errors or That It Has Been Subjected to the same Scrutiny It would Have Received Under a Legally Adequate QA Program . . . . 38 III. 'INE SEISMIC VERIFICATION PERFORMED BY THE ITP AND THE IDVP IS DEFICIENT IN THAT IT DOES NOT DEMONSTRATE THE SEISMIC QUALIFICATION OF ALL STRUCTURES . . . . . . . . . . . . . . . 39 A. PG&E Used Incrrect Methods to Model Soil Structure Interaction for Containment . . . . . . . . . . . . . . . 39 B. PG&E Has Modeled the Soil Springs Under the Auxiliary Building Inconsistently, and Has Failed to Take into Account All Effects That the Springs May Cause . . . 40

1. The DCP Has Not Considered a Fixed Base Analysis . . 40

2. The DCP Has Not Considered Rotational Effects of Flexible Soil Springs . . . . . . . . . . . . . . . . 41 C. The DCP Has Failed to Qualify All Equipnent in Containment for the Anticipated Effects of an Uplifting of the Containnent Base Mat . . . . . . . . . . 43 iii.

_a

, D. %e DCP's Use of Translational and 'Ibrsional Inputs to the Ebel Handling Building Has Not Been Shown to be Valid . ... ....... . . . . . . . . . . . . 45 E. Pb Cbnsistent, Doctrnented kpresentation of the Properties of the Soil and Ibck Surrounding the Diesel Ebel Oil Tanks Has Been Produced by the DCP or the IDVP . . . ............. ......45

1. We data used to detemine soil properties was detemined without a @ prograrn, and tests have not been redone by the DCP or the IDVP ......45

2. ITR 68 presents contradictory infomation about the backfill around the diesel oil fuel tanks . . . 46

3. ITR 68 presents information about the rock around the diesel fuel oil tanks that contradicts data used to qualify the auxiliary buildirg . . . . 47 I F. Ib Consistent, Documented nepresentation of the Properties of the Soil Surrounding the Auxiliary Saltwater Piping and Circulating Water Intake Conduits Has Been Presented by the DCP or the IDVP. . . 48

1. Se data used to detemine soil properties was determined without a QA program, and tests have not been redone by the DCP or the IDVP. . . . . . . 48

2. Data on low strain shear nodulus of the ASW and CWIC soil are not documented adequately . . . . . . 49

3. Data on shear modulus at high strain presented in I'IR 68 do not accurately represent in situ soil properties . . . . . . . . . . . . . . . . . . 50 G. %e IDVP Did Not Perfom Independent Modeling of the Seismic Besx>nse of the Diablo Canyon l Structures, Altiotqh Such Modeling Would Give Increased Cbnfidence in the Design of h ose 1

Structures . . . . . . . . . . . . . . . . . . . . . . . 52 IV. PG&E HAS FAILED 'IO ANALYZE ALL 'IHE JEI' IMPINGEMEtE I/RDS IT IS REQUIRED 'IO CDNSIDER. . . . . . . . . . . . . . 54 A. PG&E Has Misconstrued the Requirements of Its License Application With_ Respect to Analysis of Jet Impirgement Ioads Inside Contairrnent. . . . . . . . 54 D. PG&E's Misconstruction of the Criteria Has Ieft Dree Lines UnrevieWed. . . . . . . . . . . . . . . . . 56 i iV.

i I

l l

V. 'IHERE IS 10 ASSURANCE 'IHAT 'IHE PIANT AS BUILT IS

. CONSISTDTT WIni EE DESIGN ANALYSES . . . . . . . . . . . . 56 A. Diablo Canyon Has Had a History of Cbnfiguration CDntrol Deficiencies. . . . . . . . . . . . . . . . . . 56 B. %e DCP's Cbnfiguration Cbntrol Practices Have ret SignificantJy Improved Upon PG&E's Historic

~

Deficiencies . . . . . . . . . . . . . . . . . . . . . 57 C. %e Verification Progra Ibes Ibt Supply the Missing Assurance hat the Discrepancies Have Been Reconciled . . . . . . . . . . . . . . . . . . . . 58 i

VI 'IHERE HAS BEDJ NO VERPIFICATION 'IHAT SIE WESTINGHOUSE DESIGN OF SAFEIY-REIATED EQUIPMD7T MET LICDJSING CRI'IERIA . . . . . . . . . . . . . . . . . . . . . . . . . . 59 VII 'IHE YERIFICATION PRCXBAM'S MEIHODS ARE INADEQUATE

'ID ASSURE 'IHAT ALL 'IHE IOCfr CAUSES OF EE DIABID CANYON DESIGN ERRORS HAVE BEEN ASCERTAINED AND 'IHAT ALL GENERIC CDNCERNS STEMMING FROM WOSE ERROTG HAVE BEEN IDENTIFIED. ...................62 A. %e Verification Progra Failed to Identify the Ibot Cause of Each Design Error . . . . . . . . . . 62 B. %e Verification Progra Failed to Identify the Ibot Cause of Each CA Error . . . . . . . . . . . . 64 C. %e Failure to Cbrrelate Ibot Causes and OA Deficiencies Precludes Assurance That Similar Errors Will be Discovered and Will tbt Recur. . . . . . 65 D. %e IDVP Failed to Recognize Organizational Dysfunction and Managenent Shortconings as Ibot Causes of the Design Errors . . . . . . . . . . . . . . 67 VIII 'INE APPLICANT HAS FAILED 'IO PROVE 'IHAT 'IHE RECDTT DESIGN MODIFICATIONS WERE PEREDIMED UNDER A QUALITY ASSURANCE PROGRAM WHICH MET 'IHE REQUIRDiENTS OF APPENDIX B. . . . . . . . . . . . . . . . . . . . . . . . . 68 A. %e PG&E CA Progra for November 1981 % rough August 20, 1982, Was Deficient. ............68 B. PG&E Has Failed to Prove %at the CA Progra In Place Since August 20, 1982, Meets Appendix B. . . . 71 CONCLUSICNS OF IAW. . . . . . . . . . . . . . . . . . . . . . . . . 73 v.

J

INIPODUCTION In 1968 and 1970, respectively, applicant Pacific Gas and Electric Ccrnpany (EE) was authorized to construct thits 1 and 2 of the proposed Diablo Canyon 1.rlear Power Plant in San Inis Obispo (bunty, California.

(Incket ?b. 50-275, 4 AEC 89; Ibcket !b. 50-323, 4 AEC 447.)

By 1981, following modifications in response to discovery of the Hosgri Pault less than two miles fran the plant, thit I was reaching conpletion. Ch July 17, 1981, the Atanic Safety and Licensirg Ebard issued a partial initial decision authorizing issuance of a license to load fuel and test Unit 1 at up to 5 percent power. (Pacific Gas and Electric Chnpany (Diablo Canyon Nuclear Power Plant, thits 1 and 2), IBP-81-21,14 NRC 107.) Che of the issues addressed in that decision was the adequacy of ME's quality assurance (OA) program, which the Licensirg Ibard resolved in two pages with a finding that the Diablo Canyon OA program for design, construction, and operation "have been and are in conpliance with" 10 C.F.R. Part 50, Appendix B ( Appendix B) . (14 NRC at 116.)

'Ihe Nuclear Regulatory (bnmission (Chnmission) authorized inrnediate effectiveness of that decision (Pacific Gas and Electric Chnpany (Diablo Canyon Nuclear Power Plant, thits 1 and 2), CLI-81-22,14 NRC 598 (1981)), and on September 22, 1981, the Staff issued to ME License No. DPR-76 authorizirg loadirg and low-power testire of thit _1.

Within days of the issuance of the low-power license, an error in the seismic design of thit 1 was discovered. It was subsequently learned -

that the error had been caused by the transmission to a ME design contractor of an midentified drawirg of thit 2 that was taken by the contractor to be for Unit 1. Since the two units are mirror images

, 1.

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

, of one another, the error (the mirror-image error) led to the errors in the placement of certain seismic modifications. Shortly after the discovery of the mirror-image error, several additional errors were discovered that led the Cbumission to conclude that there had been serious stortcmings in the OA program for Diablo anyon in violation of Appendix B and that there was doubt that statements made in the license application regarding the ability of structures, systems, and ompr.ents (SS&Cs) to withstand the effects of earthquakes inay not have been true. The Q2nmission therefore entered an ordar (Ormission Order) suspendirg the license on Ibvember 19, 1981.

(Pacific Gas and Electric Cbnpany (Diablo Canyon Nuclear Pcer Plant, Unit 1), CLI-81-30,14 NRC 950. ) In suspendirg the license, the 0:mnission established requirements for restoration of the license, includirg the satisfactory otrnpletion of an " independent design verification prcgram" (ID/P) . (Ctanission Order, Attachnent 1.) %e Staff issued on the same date a letter (Staff letter), supplementirg the requirenents for the IDVP.

The IDVP was established, with Teledyne Ergineerirg Services

('Iuledyne) retained as the program manager. In addition, PG&E hired Bechtel Power (brporation (Bechtel) and formed a joint management team and supprtiry technical organization, called the Diablo Canyon Project (DCP) to complete the project.- %e DCP conducted what it calls the Internal Technical Program (I'IP) to carry out PG&E's verification responsibilities.

On June 7,1982, the Joint Intervenors (JI), a consortityn of interested local parties, moved to reopen the record on design quality assurance, citirg the mirror-image error and the errors that followed. A similar notion was filed on August 16, 1982, by Gwernor H3 mund G. Brown on behalf of the State of California. On January 3,1983, George Deukmejian i

l 2.

succeeded Governor Brown, and on January 28,1983, (bvernor Deukmejian filed a !btice of Substitution, assming representation of California. %e motions to reopen were opposed by PG&E ard by the !Ax: lear Ibgulatory Cbmission (NRC) Staff. Ibwever, both the applicant and Staff subsequently witidrew their opposition, and on April 21, 1983, the Ibard issued an order reopening the record on the issue of design quality assurance.

On July 19, 1983, at the Ibard's direction, the Governor and Joint Intervenors filed " specifications of the precise matters they mntest in the reopened proceedirg." (Order of July 6, 1983.) Applicant and the Staff objected to a nmber of those issues cm the ground that they failed to meet the basis and specificity requirements of 10 C.F.R. section 2.714(b); IGE also disputed the relevance of the IWP and the right of the other parties to challenge its a3equacy.

ne Ibard indicated in its Order of Atgust 16, 1983, and confirmed in an Order of Atgust 26, 1983, their view that the evidence that led to the reopening of the record, agreed to by PG&E and the Staff and concurred in by the Board, established the basis for the reopened proceedirg and stood in the place of the " contentions" required by section 2.714(b) .

(Order of Atgust 16, 1983, pp. 2-4.)

ne Ibard went on to note the unusual nature of the proceeding:

f "Normally, an effectively functionirg design quality assurance program ensures that the design of a nuclear power plant is in mnformance with I the design criteria and ammitments set forth in an applicant's PSAR

and FSAR. In the case of Diablo Canyon, however, this confidence has been seriously eroded by the existence of significant evidence that the design quality assurance program was faulty (i.e., it failed to emply with 10 C.F.R. Part 50, Aprendix B). Hence, there is now substantial uncertainty whether any particular structure, system or omponent was designed in accordance with stated criteria and cmmitments.

1 3.

l l

2

'. "In these ciremstances, the Ctminission mandated the IDVP to provide after-de-fact assurance that the Diablo Canyon design confonned to the various design criteria and could receive an apparently operatirj license even though the applicant apparently failed to osnply adequately with the (bumission's quality assurance regulations. %e applicant established its ITP for the same purpose.

R ese verification activities, if properly conceived and carried out, are to substitute for, and suppleent, the applicant's design quality assurance program in order to denonstrate that the Diablo Chnyon plant is correctly designed . . . .

. . . The real issue in the reopened proceeding has, in effect, moved beyond the question of what deficiencies existed in the applicant's Diablo Canyon design quality assurance program to the question whether the applicant can demonstrate that the IDVP and the ITP verify the correctness of the Diablo Canyon design. As previously stated, this is necessary because the apparent weakness of the applicant's design quality assurance program has rendered the design of the plant's structures, systems and cmponents suspect. %e applicant's use of the ID/P and ITP therefore may be loosely analogized to an affirmative defense to the apparent failure of the Diablo Canyon design quality assurance program to conply with the (bmmission's regulations. As such, the applicant has toth the burden of going forward and the burden of proving the adequacy of the IDVP and ITP to establish that the plant is properly designed . . . ." (Order of August 16, 1983, at 4-6 (fn.

omitted).)

On September 7,1983, the Ibard issued an order setting forth the issues to be litigated. % e Governor and Joint Intervenors filed addi-tional contentions concerning late-developing issues, Irincipally regarding the final technical reports of the IDVP, which the Ibard approved.

In the course of discovery, the Governor and Joint Intervenors cane to be satisfied with the resolution of a number of issues, which they withdrew. The final set of issues they litigated is given in Appendix A to these Pindings of Fact and Cbnclusions of Law.

Fifteen days of hearirgs were held October 31 through Ibvember 21, 1983, at Avila Beach, near the Enwer plant site in San Luis Cbispo (bunty.

l Ebrtymne witnesses testified and are identified in Appendix B. We Board adnitted 141 exhibits into evidence, which are identified in Appendix C.

l l

4.

l

FINDINGS OF FACT I.

THE ID/P HAS FAILED 'IO PROVIDE ASSURA?CE DIAT THE DESIGN OF DIABID CANYON UNITS 1 AND 2 CONEOPMS 'IO DIE LICENSItG CRITERIA AND REGULA'IORY REQUIREMENIS A. 'Ib Meet the Beguirements for a Verification Program, the Applicant Must Show Either 'Ihat the Design Fully (bnforms to Licensing and Begulatory Requirements or That the Design has Been Subjected to the Same Invel of Scrutiny it Wuld Have Feceived Under an Adequate OA Program

1. Under the Ctanission's regulations, in order to qualify for a license an applicant must demonstrate that it has constructed the plant "in conformity with the construction permit and the application as amended, the provisions of the [Atanic Energy] Act, and the rules and regulations of the Ccranission . . . ." (10 C.F.R. S 50.57(a)(1).)

2. In evident understandirg of this requirement, the IDVP, with the approval of the Chnmission, established a system for identifying and classifyirg design errors. Wat system distinguished errors that did not have regulatory significance fran those that caused the design to deviate fran the licensing criteria. Wus, the,IDVP identified a questionable aspect of the design as an " error /open iten" (EDI), and, if it was con-firmed that a given EDI was in fact the product of an error, categorized the error as an " error class A, B, C, or D." (ID/P Final Ibp. , PG&E Exh. 90, pp. F.2 F.2-5. ) An error class A, the nost serious category i

1 of error recognized by the IDVP, caused " design criteria or operating limits cf safety related equipnent [to be] exceeded," requiring physical i modification or chages in operatirg procedures. (Id. , at F.2-4. ) A l

class B error likewise caused safety-related design criteria or operating i

l l 5.

limits to be exceeded, although the deviation might be resolvable "by means

, of note realistic calculations or retesting." (Id. , at F.2-5. ) By contrast, class C and D errors were the result of erroneous ergineerirg that produced a design that still cmplied with licensirg requirements.

(Id. , at F.2-4. ) %us, the structure adopted by the IWP was well suited to determine whether Diablo Canyon Unit 1 met the requirement of section 50.57 that it conform to all licensirg and regulatory requirements.

(1) %e IDVP and the Staff Made Cbntradictory Claims as to Whether the Verification Program Denonstrated hat the Design Cbnforms to the Licensirg Requirements

3. We IWP made contradictory claims regardirg the results of the verification program:

(a) In the IWP Final Report, it is stated that the IDVP "has effectively identified mcertainties in the ccupliance of the design with license application criteria." (IDVP Final Rep. , PG&E Dch. 90, pp. ES-7, 2.0-1.) %e report also states that ". . . the scope of the IDVP review ms sufficient, and the procedures utilized to identify concerns effective, to provide reasonable assurance that those aspects of the design work on DCNPP-1 performed by PG&E or service-related contractors which did not meet the license application criteria have now been identified." (Id. , pp. 6.2.5 6.2.5-2. )

(b) At- the same time, the IDVP Final Repcr' c disclaimed havirg " identified each and every error or questionable aspect of the design product of PG&E and its contractors or the design process they utilized." (Id., <

p. 6.2.5-2.)

(c) In its prepared testimony, the IDVP testified that the scope of its program was a3 equate "to provide reasonable assurance that those 6.

aspects of the DCNPF1 design which did rot meet the criteria of the

. license application have been identified." (IDVP Panel 1, ff.

Tr. D1458, at 1/2-31. ) But the next page of its prepared testimony contains the followirg:

"Q.27 Ibes this mean that the IDVP identified each and every deficiency in compliance with the criteria of the license application?

"A.27 lb. 'Ihe IDVP was not intended to do this, nor could any reasonable indeperdent verification program. . . ." (Id., at 1/2-32.)

4. Similarly, the staff evaluation of the IDVP is contradictory on whether or not em pliance with the license has been assured:

(a) In SSER 18, ". . . the Staff concludes (1) that the verification efforts undertaken have identified all significant design deficiencies that may have existed and (2) . . . appropriate corrective actions have been and will be taken to ensure that the design of the facility conforms to the licensirg crf teria."

(b) At the hearirg, Mr. Schier1irg, Diablo Canyon project manager for the Staff, testified that the goal of the IDVP, against which it may be measured, is the demonstration that all deviations frm the licensirg

! criteria have been corrected. (Tr. D2654-55.) But Mr. Knight, Staff Assistant Director for (bmponents and Structures Digineerirg, testified that the IDVP was only required to assure that there remain to "significant deviations fran licensire criteria." (Tr. D2656.) He l defined "significant deviations" as violations of licensirg criteria i

that would cause a loss of function of a safety-related structure, systen, or emponent. (Tr. D2657.) Mr. Knight could cite no authority in the regulations for the Staff failirg to require omnpliance with l

7.

license criteria on the ground of a claimed determination that a deviation had no safety significance. (Tr. D2659.) In deposition Mr. Schierlirg had used the term "significant design deficiencies" as equivalent to error class A or B. (Tr. D2831-35.)

(c) Both Mr. Knight and Mr. Schierlirg acknowledged at the hearity that the IDVP did not establish the absence of any remainiry deviations fran licensirg criteria, merely that there are none that would, in the Staff's view, amount to a "significant deviation." (Tr. D2655, D2705.)

5. In fact, as we find below, neither the IDVP nor the ITP (which are collectively referred to as the " verification program")

denonstrated that the design of the plant meets all licensity criteria.

Indeed, they proved the contrary: that it is virtually certain there renain numerous mdetected design errors that cause the design to violate applicable licensity criteria.

(2) The Ebet 'Ihat a Lawful OA Program Wuld not Guarantee an Error-Free Design Ibes not Ibnder the Existence of Uncorrected Errors Irrelevant

6. In an effort to respond to the evidence of remainirg errors, PG&E has cited the fact that even a plant designed with a legally suffi-cient OA prcgram in operation will likely have scrne design errors. (PG&E Exh. 89, pp.13-14; Hubbard, ff. Tr. D2084, at 11; Tr. D2130, D2369-73.)

7. Ibwever, there is no basis for ccnparity the ntanerical results i

l of the verification program with the design quality of othe plants in the country becuse there exists no numerical standard for the error rate of a design that was subject to an adequate OA progrenn (Apostolakis, ff.

Tr. D2313, at 9-10,19-20; Tr. D2715. )

i

8. Mr. Hubbard testified that the objective of a OA program is l

8.

'. not to achieve any specific maximtzn number of residual errors; instead,

. confidence in the design is a product of the @ process itself, the knowledge that if one satisfies the 18 criteria of Appendix B, one can have an adequate cx>nfidence in the design product. (Tr. D2135. )

9. One requirement of Appendix B is that suitable design control procedures be established and implemented. (Criterion III.) %e design control procedures properly will include a 100 percent review of all design work by a second designer or engineer plus approval of all work by a supervisor. (Tr. D401.) %is requirement of Criterion III of Appendix B is in addition to the required inspections (Criterion X), audits (Criterion XVIII), and corrective action measures (Criterion XVI).

10. %e Board finds that, since it is impossible to establish the adequacy of the design by ccanparirg the number of errors to the ntsnber of errors to be expected in a plant designed with an adequate @ program, an after-the-fact verification can establish the adequacy of a given portion of the design in two ways: (1) by denonstratirg that that portion of the design is error-free or (2) by subjecting that 1x)rtion of the design to the same review it would have been subjected to had there originally been an adequate @ process in effect.

B. %e IDVP has Failed to Show Either %at the Design Fully Conforms to Licensing and Regulatory BeqJirEments ur hat the Design has Been Subjected to the Same level of Scrutiny it Would Have Received Under an Adequate @ Program (1) It Is Virtually Gertain %ere Resnain Uhdetected Errors %at Chuse the Design to Pail to Ctznply With the Licensing Criteria

11. %e only quantitative results of the IDVP's verification of the non-seismic design were reported by Dr. Kaplan, the DCP's expert witness on "probabilistics" (Tr. D1333), who reviewed the results of the 9.

l l

ID/P's non-seismic verification work and found nine class A or B errors in 911 design elements sampled by the ID/P. (PG&E Panel 6, ff. Tr. Dll60, at 48.)

12. If one assumes this was the result of a randon sample of the design work, Ir. Kaplan estimates that the median error rate for the nor.-seismic design work is about 1.3 percent. (Id. , at p. 51; Tr. Dll68.)

13. ExtraInlatirg the median error rate for the IDVP sample to the total non-seismic design leads to the conclusion that there remain 40 undetected class A or B errors in the non-seismic design not reviewed by the ID/P. (Tr. D1169. )

14. Although Dr. Kaplan testified that the ID/P sample was weighted in favor of that portion of the design where errors were nore likely (Tr. Dil67), which would militate in favor of a lower estimated ntrnber of remainirg class A or B errors (Tr. D1169-71), the evidence was clear that Dr. Kaplan was in error in his understandirg of the objectives of the IDVP's selection of an initial sample. In fact the IDVP asserted no stch objective in its selection of the sample (ID/P Final kp., PG&E Exh. 90, p. 4.7.1-1; ID/P Panel 1, ff. Tr. D1458, at 1/2-21), and Dr. Cboper, the ID/P program manager, testified the IDVP did rot seek to weigh the sample in favor of that portion of the design most likely to have errors (Tr. D1549). Indeed, both Dr. Cloud and Dr. Cboper testified that one could not know a priori where errors were nest likely to occur.

(Tr. D1547, D1549. )

, 15. In addition to the certainty that there remain undetected class A or B errors in the non-seismic design work, Dr. Cloud testified that it is very likely there remain undetected class A or B errors in the 10.

corrective action program work. (Tr. D1543.)

(2) 'Ihe IDVP has not Subjected the Design to the '

, Same IRvel of Scrutiny 'Ihat it R>uld Have Received Under an Ade.juate @ Program (a) 'Ihe IDVP's verification of the seismic design consisted of a sampliry of the I'IP design review work

16. 'Ihe IDVP program management plan for Phase I originally contemplated samplirg the seismic design of mit 1 to determine the nature and extent of any design errors. (ME Er.h. 88, pp. 8-9; DCP Phase I Final Rep. , M E Exh. 91, pp. 1.5.2 1.5.2-2. )

17. By June 1982, the extent of the design errors identified in Phase I led to a change in the program, with the DCP institutire a 100 percent review of the seismic design and the IDVP verifyirg that work on a sanple basis. (ITR 8, ME Exh.100; IU/P Panel 1, ff. Tr. D1458, at 1/2 1/2-13.)

18. 'Ihe samples drawn by the IDVP for verification were not drawn at randan bat rather purposefully, accordity to what the IDVP called "engineerirg judgment." (ME Exh. 88, Appen. D, p.12. )

19. Because, as we have found, a legally sufficient @ program would have subjected 100 percent of the design work to review, the Ibard finds that the IDVP's verification of the seismic design of mit I does not constitute a substitute for having subjected that portion of the design to j an adequate @ program in the first instance.

! (b) 'Ihe IDVP's verification of the non-seismic design consisted of a sample of certain parts of the original non-seismic design work

20. miike the seismic dasign review, where NE indertook a broad review of the design product that was sampled by the IDVP, the 11.

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

4 non-seismic review consisted cf a review by the IDVP of a sample of the original design ~ work done by PG&E and certain of its contractors.

(i) 'Ihe IDVP failed to verify the work of contractors performing safety-related design work.

, 21. 'Ihe IDVP was required to review the safety-related design work of all service contractors for Unit 1. (Cbnmission Order, n E >

! Exh. 86, Attach.1; Staff Ietter, PG&E hh. 87, pp. 2-3, encls. A, C.)

22. 'Ihe IDVP developed a list of safety-related service

[ contracters who contributed to the final design of thit 1. (IDVP Final Rep., PGEE Exh. 90, 5 4.1.4; ITR 5, PG&E m h. 97; ITR 9, PG&E Exh. 101; ITR 29, PG&E Exh.121.)

23. 'Ihe IDVP selected same of those contractors for verification and onitted others. (IDVP Final Rep. , PG&E Exh. 90, p. 4.1.4-3. ) 'Ihe elimination of service-related contractors was not authorized by the Ctanission and was objected to by the Staff. (SECY-82-414, PG&E Exh.157,~

encl. 6, p. 4. ) Staff approval of the IDVP expressly restated the -

requirenent that "all" service-related contractors be included.

(SECY-82-414, PG&E Exh.157, Figs.1, 3, enc 1. 6, p. 4; W. D2755. ) A subsequent objection by the IDVP to the requirement of including all contractors (Gov. Exh. 47, p.1) was never responded to by the. Staff i

(Tr. D1468) .

i

24. Dr. Cboper sought to justify the anission of rtain contractors on the ground that the Staff Ietter permitted the use of sanpling. (Gov. Exh. 47; 'It. D1462. ) However, the letter authorizes sanpling within the work of each contractor, not sampliry among contractors (Staff Ietter, PGEE Exh. 87, pp. 2, 3), a distinction the IDVP failed to 12.

. draw in plannirg and carryiry out its program (Tr. D1469).

25. Amorg the contractors excluded fran the IDVP's verification were the followirg:

Westirghouse, which, in addition to beirg the vendor for the nuclear steam supply system (NSSS), did safety-related design work.

(Tr. D1471-77. ) IWP review of Westirghouse was limited to checkirg the design interface between PG&E and Westirghouse. (Tr. D1477. )

Western Canada Hydraulic laboratories, which did safety-related design work. (Tr. Dl478) Western Canada had been found by PG&E to have had no formal @ program applicable to its Diablo Canyon work.

(SECY-82-414, PG&E Exh.157, enc 1. 5, p. S.)

-- Stafco, which prepared the list of safety-related SS&Cs and participated in updatirg the FSAR. (Tr. Dl486.) Although the IDVP did not consider it design work, its witnesses agreed that the work was safety-related. (Tr. Dl488.) PG&E had included Stafco amory the list of contractors for the IDVP. (ITR 9, PG&E Exh.101, p. A46. ) PG&E had found Stafco to have had a deficient @ program. (SECY-82-414, ME Exh. 157, enc 1. 5, p. 5.)

26. 'Ihe IDVP also excluded the work of R. L. Cloud and Associates and Teledyne Dgineerirg Services, which p ovided safety-related design services. (IWP Final Rep. , n e Exh. 90, p. 4.1.4-3; Tr. D1483-85. ) 'Ibe Cloud organization had been found by PG&E to have no formal @ gogram.

(SECY-82-414, PG&E Exh.157, enc 1. 5, p. 5; 'It. D1483. ) The exclusion of l these contractors was based on the need to avoid a conflict of interest, i

since both were IDVP contractors (IDVP Final Ibp., PG&E Exh. 90,

p. 4.1.4-3; 'It. D1484), and on the IDVP's interpretation of the Omnission 13.

l

• Order and Staff Ietter as not requirirg inclusion of contractors performing work after June 1978 (Tr.1494-25).

27. Altinugh Mr. Knight testified that the Staff implicitly intended to prmit the IIVP to exclude contractors not significantly contributirg to the final design (Tr. D2755), the IINP did not claim that the foregoirg excluded contractors could be mitted on that ground (IDVP Final Rep. , ME hi. 90, p. 4.1.4-3) .

(ii) The IINP Sample of ME's ?bn c ismic e Design Cbnsisted of a Ibn-Randm Sample of 'Ihat Wbrk

28. 'Ihe IINP sought to verify that portion of the non-seismic design of Unit 1 for which PG&E performed the design work by takirg a sample of the PG&E-designed SS&Cs. (IDVP Final Rep. , ME hh. 90,

p. 4.7.1-1; IDVP Panel 1, ff. 'Ir. D1458, at 1/2-21. )

29. 'Ibe sample was selected frm a list of PG&E-designed SS&Cs.

(IDVP Final Rep. , PG&E hh. 90, pp. 4.1.2 4.1.2-2; Tr. D261-62. ) Sme PG&E-designed SS&Cs were mitted frcra the list, although sme of them may have been included incidentally in the IINP's verification sample.

(Tr. E306-07, D1492, D1496-98. )

30. 'Ihe IDVP did not select its sample of SS&Cs randmly but rather by purposeful selection of design work that had attributes desired l by the IDVP. (IDVP Final Rep. , PG&E hh. 90, pp. 4.7.1 4.7.1-2; IDVP Panel 1, ff. Tr. D1458, at 1/2-21. )

31. 'Ihe II7v'P sample consisted initially of three systems: the auxiliary feedwater system (AFNS), the 4160-V electrical system, and the control rom ventilation and pressurization systen (CRVPS). (Ibid.)

32. As a result of " generic concerns" raised in the non-seismic 14.

t

. verification of the three systems, the INP caused all applicable safety-related systems in Unit 1 to be reviewed for five specific errors:

lack of redundancy in shared-power equipnent, incorrect selection of design tenperatures and pressures, erroneous envircrrnental consequences of postulated pipe ruptures outside contairrnent, failure to meet circuit-separation criteria, and erroneous analysis of the effects of pipe-ruptures inside contairynent. (IWP Final kp. , PG&E Exh. 90, pp. 4.8.1 4.8.1-3; DCP Phase II Final kp. , EG&E Exh. 92, p.1-20; IDVP Panel 1, ff. Tr. D1458, at 1/2-24. ) 'Ihe IDVP verified the results of that work on a samplirg basis. (DCP Phase II Final kp. , PG&E Exh. 92,

p. 1-20.)

33. Alttough PG&E represented the IDVP non-seismic sample as amunting to 30 percent of the non-seismic design (Tr. D1419), in terms of the total number of " design elements," the IDVP reviewed only 911 out of 3,968 design elements (PG&E Panel 6, ff. Tr. D1160, at 48), or less than 23 percent.

34. 'Ihe Ibard finds that the IWP's non-seismic verification constitutes a satisfactory substitute for an adequate OA Eragram for those contractors sampled by the IDVP. 'Ihe Ibard further finds that the IDVP's non-seismic verification is a satisfactory substitute for an adequate OA program for the three systens sampled by the IDVP - the 4160-V electrical system, the AEWS, and the CRVPS - plus those elements of the non-seismic design reviewed in the IWP's resolution of the five generic concerns.

35. 'Ihe Board finds that, with respect to the balance of the non-seismic design, the INP review does not constitute a satisfactory substitute for an adequate QA program.

15.

, (C) None of the Reasons Given by n E for the Board to Ignore its Failure to Meet the Licensirg and Ngulatory Requirements is Valid

36. ME advances a number of grounds for this Ibard to ignore the evidence of remaining mdetected errors causing the design to violate

.tegal requirements and for us to excuse the failure of the verification program to scrutinize the design as fully as an adequate OA program would have. For the reasons given below, we find none of the grounds valid.

(1) The Claim %at %ere kmain no Undetected Errors Having Safety Significance is thfounded

37. While the evidence is uncontradicted that there remain undetected errors in the design of the facility, EE, the IDVP, and the Staff claim that none of the errors is safety-significant. (SSER 18, Staff Exh. 36, pp. C.5-2, C.5-6; IDVP Final Rep. , ME Exh. 90, p. 6.2.5-2. ; ME Panel 6, ff. Tr. Dll60, at 64-65; D1545, D1558, D2265-66. )

(i) Se evidence does not support any conclusion about the safety significance of the errors identified by the verification program

38. %e IDVP expressly declined to offer any opinion about the safety significance of the design errors it detected that it classified as causing a violation of the licensing criteria. (IDVP Final kp., ME Exh. 90, p. 6.4-1. ) Although one could state one's judgment, the IDVP did i not perform the analyses necessary to have a reportable conclusion about the safety significance of the identified errors. (Tr. D1555, 1559.)

39. Both nE and the Staff offered opinions about the safety l significance of the known design errors, opining that none of them constituted a safety-significant error. (DCP Phase I Final Rep., ME 1

l Exh. 91, p.1.8.6-1; DCP Phase II Final Rep. , nE Exh. 92, p. 5-1; EE 16.

l l

Panel 6, ff. Tr. D1160, at 54; Tr. D2696. ) 'Ihis conclusion was not based on any systematic study of the known errors by PG&E or the Staff. (DCP Phase I Final Rep., PG&E Exh. 91; DCP Phase I Final kp. , PG&E Exh. 92; Tr. D2689.)

40. Ibth Dr. (boper and Prof. Apostolakis testified that one could not reliably formulate an opinion on the safety significance of an error without doing a detailed analysis. Dr. (boper testified such an analysis would have to assess whether the error could cause a failure of a systen and, if a failure was possible, to identify and evaluate the chain of events followire the failure. (Tr. D1555, D1557.) Prof. Apostolakis testified that to properly assess the reduction in the margin of safety caused by a design error, one would have to perform a probabilistic risk assessment (PRA) to determine " failure paths not perceived by knowledgeable ergineers" and the potential for multiple failures of redundant omponents.

(Apostolakis, ff. Tr. D2313, at 10-11.) Even Dr. Implan testified that a PRA would be the soundest way to assess the safety significance of a design error (Tr. D1393), although toth he and Prof. Apostolakis agreed that scmething less than a " full-blown PRA" would be adequate to assess the risk fran certain errors. (Tr. D1394-96, D2374. )

41. No such PRA has been done for Diablo Chnyon. (Tr. D1194. )

42. 'Ihe Ibard therefore finds that the opinions of PG&E and the Staff to the effect that the identified design errors did not create a substantial safety hazard lack adequate basis to support a licensing decision in this case.

(ii) 'Ihere is no basis for makirg any statement about the design errors that remain undetected

43. While the IINP expressly declined to offer any opinion about 17.

~

, the safety significance of the class A or B errors it did find (Tr. D1554),

. the ID/P considered itself capable of expressirg an opinion on the safety significance of the errors that remain mdetected (IDVP Final Rep., ME Exh. 90, p. 6.2.5-2; 'Ir. D1557-59) .

44. Since the nature of the undetected errors -- their location, the extent of deviation fran criteria, and the conditions mder which the errors may manifest thenselves -- are not known, no specific analysis of them could have been done, and none has been offered in evidence.

45. Instead of providing an analysis of the safety significance of each mdetected error, the ID/P relied on an assessment that, while it had not identified all of the errors, it had identified all of the

" generic" errors. (IDVP Panel 1, ff. 'Ir. Dll60, at 1/2-21,1/2-26, 1/2-31.) But a " generic" error is merely one which the IDVP judged to have had a potential to exist in a similar manner in other, unreviewed parts of the plant. (IDVP Final Rep. , ME Exh. 90, p. 3.5-4; Hubbard, ff.

Tr. D2084, at 25. ) Plainly, the likelihood of replication of an error is not determinative of its safety significance.

46. We therefore find that there is no basis for concluding that the remainirg class A or B errors have no safety significance.

(2) There was no Evidence to Support the Suggestion 'Ihat the Undetected Design Errors (buld be Ignored in Licensity Because 'Ihey Nbuld be Detected in Preoperational 'Destirg

47. 'Ibere was sczne suggestion at the hearirg that the existence of undetected design errors causirg the design to violate the licensirg criteria might be tolerated because of the likelihood the errors would be detected in preoperational testirg or operation. (Tr. Dll82-83.)

48. Both Mr. Anderson and Mr. Knight acknowledged tlat Diablo 18.

. Canyon had nearly empleted its preoperational testirg in 1981 when its license was suspended, ard none of the subsequently discovered errors were detected in that program, nor would they have been in the remaining preoperational tests or startup. (Tr. D1271, 2718.) Mr. Anderson agreed that such testiry generally cannot detect errors that would cause a failure only trider mergency conditions. (Tr. D1210-12.)

49. %e Ebard finds that the possibility of subsequent detection of design errors neither has been proven likely nor provides a valid ground for excusirg empliance with all licensing criteria.

(3) Vague Claims About Overall Ergineerirg Quality are no Substitute for Either License Cbnpliance or an Adequate OA Program

50. PG&E offered testimony to the effect that, in the view of the witnesses, reviewers found " generally high quality engineerirg work."

(PG&E Panel 6, ff. Tr. D1160, at 51.)

51. %is claim is not material to this case. %e reported results of the verification program -- specifically the indication of whether the design meets licensirg criteria -- are tne relevant measures of design quality. 'Ib the extent that the verification program seeks to qualify the design by showirg the absence of errors, the actual results l show that the design, while it may be, in the judgment of sme, of high quality, plainly is not free of errors causing violations of the license.

52. 'Ib the extent that the verification program is offered as a review of the design cmparable to a lawful OA program, general impressions about overall quality are even less material. Se Chnmission has never indicated that a licensee would be relieved of its obligation to meet Appendix B by representations that spot checks of parts of the design I

sinwed those parts to be of gererally high quality.

19.

l l

1 l ,, ,

I

, 53. These requirenents are even nore important in this case. As l

\

Mr. Imight testified, the breakdown in @ at Diablo Canyon is tmprecedented in his experience. (Tr. D2660-62.) The Ibard finds that assertions about i overall engineerirg quality cannot substitute for a lawful @ gogram or excuse the continued presence of design errors in violation of the licensirg criteria.

D. By Ignoring the Scientific Requirements for Drawing (bnclusions Frczn Sample Observations, the IDVP has Rendered its Findings About the Sampled Design Wbrk Inapplicable to the Unsampled Portions of the Design (1) The IIVP Violated the Chnmission's Nguirements and its own Q2nmitment by Pailirg to Avail Itself of Cbnpetent Statistical Dcpertise

54. The IDVP not only failed to provide a legally sufficient substitute for the missirg @ gogram required by the regulations, it also failed to comply with the Chnmission Order.

55. In both its Phase I and Phase II program management plans, the IDVP conmitted to " arrange for an evaluation" of the entire Phase I and Phase II program "by an expert in the applicability of statistics to ergineered systems." The plan for Phase II included the ccanmitment that "the statistician retained for Phase I will be used in conductirg Phase II." (PG&E Dch. 88, app. C, p. 3; PG&E Bch. 89, app. C, p. 3; Tr. D1508-12, D1514-15.)

56. The IDVP never retained a statistician for either Phase I or Phase II to review either the design or the results of the IDVP prcgram.

(Tr. D1513. )

(2) Contrary to the Claim of the IDVP, it had Never I Been klieved of the Obligations of Havirg the l

1 Verification Peviewed by a (bnpetent Statistician

57. Dr. (boper testified that he believed his ccrnmitment to l

20.

l l

, retain a statistician for the IWP was annulled by the adoption of SECY-82-414 (PG&E Dch.157), which contained reference to the Staff's view that "[r]igorous statistical techniques are largely inappropriate for a design verification program." (Id. , encl.11, p.11-3; Tr. D1508. )

Dr. Cboper also cited the major expansion of the sample as a ground for the IIVP no longer considerirg itself bound by the ommitment. (Tr. D1516.)

58. In fact, the approval contained in SECY-82-414 expressly incorporated the IWP Phase II program management plan (PGLE Dch. 89) "as modified by [a letter not in evidence and not cited .sy the IDVP in its testimony] and the additional requirements noted in this attachment [11]

.... (SECY-82-414, PG&E Dsh.157, enc 1.11, p. 2 (emphasis supplied) .)

The program management plan, which also includes the quoted statement repeated in SECY-82-414 that the staff considers " rigorous statistical techniques . . . largely inappropriate" also contains the IDVP cm mitment to have the program reviewed by a statistician. (PG&E Dch. 89, app. C, pp. 1, 3.) 'Ihe plan also includes "' [t]he staff ['s] support (for] the use l of statistical techniques in establishirg the sample size'" (id., p.1) and qtntes the Staff's statement that

"'. . . we did not think that the program plan did use statistic sanplirg [ sic] methods to the extent that it could have, and our conclusion that whoever does this program should expand the use of samplire criteria'" as far as passible. (Id., p. 2.)

The plan also quotes the Staff's opinion that:

"We also think these statistics can be employed in devisirg the sanplirg plan in a number of areas, that it cannot substitute for judgement in picking systems throughout, and the program plan was silent, by and large, on its use of statistics in a ntuber of areas we think that could profit fran a hard look to see to which extent questions and reliability can be examined throuah the design process."

(Ibid.)

1 l

l 21.

l l

I I

t

'Ihe plan confirns that these conments came durirg that portion of the Canmission areeting "that served as the basis for the Ctanmission vote to the effect that the scope and technical approach should be enlarged to include consideration of statistics in the samplirg plan." (Id., p. 3.)

59. Dr. Cboper admitted that he never advised the Ctanission that he intended not to retain a statistician and ha never received express approval for the decision. (Tr. D1519. )

60. Alttough the IWP Final MIrrt implies that the DCP retention of Dr. Kaplan was intended to fulfill the INP's conmitment (PG&E Exh. 90,

p. 3.5-8), Dr. Cboper testified that he did not so consider the hiring of Dr. Kaplan (Tr. D1514), and Dr. Kaplan did not understand himself to be servirg that function (Tr. D1333) . Furthermore, the arrangement between Dr. Kaplan and PG&E did not conform to the requirements of independence for INP contractors. (SECY-82-89, PG&E Exh.156, p. 4; SECY-82-414, PG&E Exh.157, p. 7; Staff Panel 1 (testim. ca contention 1), ff. Tr. D2649, enc 1. 3 to attach. 1, p. 2; Tr. D2760-63.)

61. In fact, it was the Staff's mderstandirg that the cxmmitment to retain a statistician had been incorporated in the Omnmission's approval of the program management plan (Tr. D2761), that the conmitment was still l

effective (Tr. D2758-59, D2763), and that the retention of Dr. Kaplan (by i

I the I'IP) constituted a discharge of that asnmitment (Tr. _D2759, D2763).

(3) The Opinions of the IWP, Staff, and PG&E Wre Arrived at Without the Benefit of Ctsnpetent Alvice on the l Applicability of Statistics to the IWP

62. 'Ihe opinion of the IWP that rigorous statistical techniques were inapplicable to design verification was reached by the IWP program manager without consultirg with any person expert in the applicability of 22.

, statistics to engineered systems. (Tr. D1522, D1526. )

63. 'Ihe views of the Staff concernirg the inapplicability of rigorous statistical systens to design verification was arrived at by Staff menbers who were themselves not experts in statistics and without the benefit of the views of those manbers of the Staff who possess such expertise. (Tr. D2764-66. )

64. 'Ihe DCP witness on statistical issues, Dr. Kaplan, was not retained by the DCP until mid-1983 (Tr. D459-61), when he was brought in "for the express purpose" of rebuttirg the allegations of Mr. Hubbard in an affidavit on the use of statistics ('IUr. D1408-14). Dr. Kaplan does not consider himself a statistician (Tr. D1333, D1349), has no training in statistics (Tr. D1347-49), has never published in a reference journal in the field of statistics (Tr. D1348-49), and is not familiar with the literature on statistical samplirg (Tr. D1329, D1363-66).

65. 'Ihe Ibard finds that the applicant and Staff have failed to s}nw by campetent evidence that statistical techniques are inapplicable to design verification.

i (4) Statistically Valid Parxkrn Samplirg is Essential to Generalization Fran Sample Observations to the Unsampled Portions of the Population

66. Both Prof. Apostolakis and Prof. Samaniego testified that randon samplirg is essential to a valid generalization to a population.

( Apstolakis, ff. Tr. D2313, at 15,19; Samaniego, ff. Tr. D2392, at 5, 7, 8 ; Tr. D2353-54, D2394-95, D2399-D2401, D2425-26, D2430-33, D2514-22. )

Randan samplirg is not an essential of all scientific inquiry, but it is an essential of any generalization of probabilities fran samples to I populations. (Tr. D2430-33.)

1 23.

l

\

f

• 67. 'Ibe use of randcm samplirg is no less a requirement of Bayesian statistics as it is of classical statistics (Tr. D2343), and the use of a non-randam saple to update a prior using Bayes 'Ibeorem is " simply wrong," "ireppropriate," and "cmpletely invalid." (Tr. D2394-95.)

68. 'Ihe Ibard finds that randm sampliry is a requirement of any valid generalization of observations frm sample data to the msampled por-tions of a population.

(5) It is Possible to (bnduct a Statistically valid Design Verification Program

69. Prof. Apostolakis testified that one can employ statistics in a design verification program. (Apostolakis, ff. W. D2313, at 13-14.)

And Dr. Kaplan described his own testimony as demonstratirg that cne can derive error rates frm a smple of the design of a nuclear power plant usirg " statistics-like" technigms. (Tr. Dll63. )

70. Although Prof. Apostolakis confirmed that one could set up the randm samplirg problem in such a way that it would be extremely large arr3 extensive, such as by samplirg frm the population of all design decisions (Tr. D2340), he also was of the opinion that the problem could be structured in such a fashion as to make it manageable (Tr. D2341) and that the question of how one characterizes the population is distinct from the question of whether one randmly samples frcm the population (Tr. D2353).

71. tbne of the reasons given by Ir. Kaplan why he considers 6

" classical statistics" inapplicable to design verification is valid:

(a) 'Ihere is, as Dr. Kaplan testified, a need for a well-defined set of elenents. (PG&E Panel 6, ff. W. D1160, at 5.) But he demonstrated that such a set could be developed. (Tr. D1163. )

(b) Although Dr. Kaplan asserted that the elements must all be equal and 24.

well mixed, he acknowledged that one could validly take a randm sample of a hetertgenous population or one could stratify the population and take randa samples within strata. (Tr. D1272-97.)

(c) Dr. Kaplan testified that the elements of the population must be readily identifiable in terms of the attribute (PG&E Panel 6, ff.

Tr.1160, at 5) - in the mntext of design verification, one must be able readily to identify an element as either containirg an error or not (Tr. D1279) . But he acknowledged that this is not a problem of statistics, but a problem of design verification, which beset the IDVP fully as much as it would one performing a statistical study.

(Tr. D1279-890. ) And, indeed, the definitions used in Dr. Kaplan's own testimony for significant errors were, he acknowledged, "at least as fuzzy" as the definitions used by the IDVP. (Tr. D1280. ) .

(d) Dr. Kaplan stated that classical statistics focuses on frequencies, which he considers a limitation on its applicability to design verification. (PG&E Panel 6, ff. Tr. 1160, at 5.) But both Prof.

Apostolakis and Prof. Samaniego testified that the issue of randm smpliry is independent of the Bayesian versus classical statistics debate -- that both Bayesians and classical statisticians use randcm samplirg. (Tr. D2343-44, D2394-95, D2404.)

(e) Dr. Kaplan cited the " interconnected nature of design" as a reason why statistics was inapplicable. (PG&E Panel 6, ff. Tr. D1160, at 6, 42-43; Tr. D1288.) But Mr. Anderson agreed that if one were interested in capturirg this integrative quality, one could, with a detailed review of all the systems, enumerate all of the integrative aspects, and Dr. Kaplan agreed one muld then sample randcmly frm that list to 25.

study whether there were errors in the way those relationships were treated in the design. (Tr. D1289.) Prof. Samaniego testified to the same effect. (Tr. D2405-07.)

(f) At the heart of Dr. Kaplan's objections to the use of statistics was his view that statistics failed to make adequate use of non-frequency information - particularly the opinions of experts. (PG&E Panel 6, ff. W. Dll60, at 11-12,15-16, 42, 54, 68-69; W. D1286.) Dr. Kaplan and Prof. Samaniego, towever, agreed that expert opinion can be incorporated into the statistical framework by stratifyirg the pcpulation accordirg to the experts' advice. (Tr. D1304, D2404-05. )

7 2.- In the final analysis, the Ibard finds that the issue is not between classical and Bayesian statistics, but between samplirg in accordance with statistical principles -- a requirement of both Bayesian and classical statistics -- and samplirg non-randmly, a practice that the evidence stows to be invalid to adherents of both schools. %e Ibard finds that no valid reason has been given why one cannot employ statistically valid samplirg techniques in a design verification program.

(6) Proper Samplirg is not Only a Legal and Scientific Requirement, but Plainly Necessary to the Objectives of the Verification Program i

73. %at samplirg for a verification prcgram be done on a randm basis frm the wrk of each design group is an obvious requirement of any effort to provide assurance of an error-free design.

74. Dr. Kaplan testified that different design mntractors will vary in their error rates. (Tr. D1309. ) nus, samplirg frm the work of one contractor provides to reliable information about the likelihood of errors in the work of sme other, msampled, contractor.

26.

, 75. Likewise, changes in the ocanposition of a design team can be expected to alter the likelihood and distribution of errors. (Tr. D1309.)

In the 15 years during which the Diablo Canyon design was evolving, the ME design teams experienced the expectable personnel ~ changes. (Tr. D1309-10.)

76. 'Ihe Ibard therefore finds that a verification program that fails to employ statistically valid samplira techniques and that fails to sanple fran the work of all design groups cannot provide the requisite assurance that the design conplies with licensirg requirements.

E. 'Ihe IDVP has not Verified the Design of Unit 2

77. 'Ihe scope of the IDVP was limited to Unit 1. (Qmnission Order, PGE Exh. 86; Staff Ietter, PG&E Exh. 87; IDVP Final Rep., ME Exh. 90, p. 1.1-1.)

78. 'Ihe IDVP did not review the design of thit 2. (IDVP Panel 1, ff. Tr. D1458, at 1/2 1/2-36. )

79. While there are many similarities in the designs of thits 1

and 2 (PGE Panel 1, ff. Tr. D224, at 28-29), the two units are generally mirror images of one another and not identical (Tr. D2274). Among the differences are different reactor internals, differences in thermal output, resultirg in differences in accident analyses, differences in the design of the electrical generator, and differences in equipnent vendors. (ME Panel 1, ff. Tr. D224, at 29-29.) Although the piping and instrumentation schenatics are identical (ibid.), those merely show functional

, . relationships and not physical relationships (Tr. D2775).

80. ME witnesses testified that the two units were originally designed to the same design criteria (PG&E Panel 1, ff. 'It. D224, at 29),

but they did not claim that the designs themselves were identical. In 27.

fact, a substantial mount of design work unique to Unit 2 has been done and will be done, with PG&E having, at peak, arne 400 to 500 engineers working exclusively on Unit 2. (Tr. D1319. ) Both PG&E and Staff witnesses acknowledged the fact that there are opportuiities for design errors in Unit 2 that would be migue to that unit. (Tr. D1321-22, 2776. )

81. Mr. Schierliry testified that he expects the Staff to issue a separate supplement to the SER to cover thit 2. (Tr. D2778-79.)

82. Uhtil recently, the two units were designed by a single design organization mder a cmmon @ program. (PG&E Panel 1, ff.

Tr. D224, at 29, Tr. D2772. ) Sme of the design work for Unit 2 was done before November 1981 (Tr. D1320), mder the progran the IDVP and Staff found to be significantly deficient (Tr. D2772).

83. In the opinion of Prof. Apostolakis, the IINP does not provide a valid basis for assessirg the design quality of thit 2.

( Apostolakis, ff. Tr. D2313, at 20. )

84. 'Ibe hrd finds that the ID/P has neither proven the design i

of Unit 2 to be free of errors violation licensing criteria nor subjected l

Unit 2 to scrutiny cmparable to that of a lawful m program.

II.

THE ITP DOES NCTT PROVIDE 'INE ASSURANCE MISSING FROM 'IHE IDVP 'IHAT 'IHE DESIGN OF DIABIO CANYON ."EETS 'INE LICINSE CRITERIA AND APPLICABLE REGUIKIORY REQUIREMENIS

85. Although we have found that the ID/P satisfies neither requirement we have established for a substitute for a legally sufficient

@ Irogram, our inquiry is not over. It re . ins for the Ibard still to determine whether the I'IP can meet either test for a verification program:

the denonstration of a design free of errors causirg violations of license 28.

l

O

~

, criteria or the review of the design omparable to that provided by an adequate OA progrm.

86. We have already found that the verification program has denonstrated the virtual certainty of remaining class A and B errors. While that does not prevent the I'IP (or the IDVP) frm trying to prove that spe-cified prtions of the design are error-free, PG&E has mught to make no such showing. Rather, it has focused on the question of safety signifi-cance, which we have already rejected as the appropriate criterion for the present licensing decision.

87. %e remaining question, therefore, is whether the ITP has subjected the design to the same level of scrutiny it would have received fran an adequate OA program.

A. The ITP's Seismic Verification Program Provides Only a Limited Substitute for a Lawful CA Program

88. %e I'IP's review of the seismic design of' thit 1 - which is referred to as the Phase I program, as distirguished fran the Phase II program for the non-seismic design -- consisted of a nore extensive review than its review of the non-seismic design. Ibwever, it did not canprise a

100 percent review of all elements of the seismic design. Se Phase I Final Report makes it clear that many systens and conpanents were reviewed l only on a sampling basis, others only on a " generic" basis, and others only for the presence of a sirgle error or a predetennined list of errors.

89. %e PG&E Phase I Final Ibport states that the following systems and camponents received a seismic review only on a _ sampling basis, i

l not a 100 percent review basis:

I Snall bore pipirg: We DCP used unspecified samplirg techniques to i

select a sample of all snall bore pipirg for review of rtain I

29.

1

. potential design errors; twenty pipelines were addressed out of an unspecified ntrnber plantwide (ME Phase I Pinal Pep., ME Exh. 91, S 2.2.2.3.3.) his was done in adition to the " generic" review discussed, infra.

Instrument tubing supports: A " representative" (undefined) ample of 88 tubirg supports was reviewed solely to determine whether the supports were affected by revisions to respnse spectra. Wis sample was limited to the portions of the annulus affected by the changed spectra. (ME Phase I Pinal Rep. , ME Exh. 91, S 2.6.1. )

90. %e following systems and canpnents were reviewM by EE for only selected and tredetermined potential design errors, and for no other possible design errors:

Quipnent in general received an initial review solely to determine if the applicable resp 3nse spectra had charged. If spectra did not charge, no further review was done. (ME Phase I Final Ibp., EE Exh. 91, S 2.3.1.3.)

Specifically, electrical equipnent was reviewed solely for changes in applicable spectra. Chly if spectra charged was any further design review done. (Ibid., SS 2.3.2.3.2 & 2.3.2.3.4.)

IWAC equignent received a spectra charge review, with no further review undertaken unless spectra had charged. (Ibid., S 2.3.3.2.)

-- Instrument tubirg and tubirg supp)rts: %e sample of 88 tubing supports discussed, supra, as reviewed only to determine if spectra had changed, and received no further review if they had not. (Ibid., S 2.6.1.)

91. Che set of conponents was reviewed by EE on a " generic" basis, reviewed for a specific list of factors predetermined by the ITP or 30.

, the IINP as ptential causes or areas of design errors:

-- Snall bore pipirg: In addition to the samplirg review for certain types or potential design errors, a so-called generic review was done for specific problem areas identified by ITP and IINP reviews.

92. Finally, the followirg systeas and ccrnponents were systenatically excluded by ME fran review:

-- Large bore pipirg analyzed by Wentinghouse: nis piping was not reviewed by PG&E. Where data input charged, the revisions were supplied by ME to Westinghouse, but no review was done by PG&E of the analysis. (Ibid., S 2.2.1.1.)

Equipnent: Only equignent associated with safety systems designed by PGLE was reviewed by the ITP. tb other equipnent was reviewed.

(Ibid., S 2.3.)

93. % e Board finds that the seismic review consisted of a partial verification of Krne SS&Cs against scrne criteria.

B. We ITP's Non-Seismic Verification Provides no Basis for Drawirg (bnclusions About Those Portions of the Design not Sampled by the IDVP (1) Prior to the Hearirg, the ITP's Phase II Program was not mpresented as an Independent Verification Effort

94. %e Phase II Final kprt (PG&E Exh. 92) describes the function I

of its Phase II ITP in terms of furnishirg information to the IDVP, managing records and information, and " normal supprt activities." (Id. at 1-5.)

95. tbwhere in the DCP Phase II Final Nport is there any descri >-

l tion of an ITP verification program beyond the scope of the IDVP.

(2) At the Hearirg PG&E !bught to & characterize the IIP as a Verification Program of Iarger Scope 7han the IDVP

96. In its direct testirrony at the learirg, PG&E Panel 1, for l

i 31.

i i

l

the first time, sought to characterize the l'IP Phase II review as a "self-contained and separate review" (PGEE Pc.nel 1, ff. Tr. D224, at 13),

though not a "cmplete review" nor a " total review" (id. , at 16, 20) . 'Ihe panel offered no quantitative measure of row much of the non-seismic design work the I'IP had reviewed. Similarly, PG&E Panel 6, which presented detailed figures on the fraction of the non-seismic design verified by the ID/P, made only passirg reference to the verification by the ITP, almitting it did "not attempt to quantify it" (PG&E Panel 6, ff. Tr. Dll60, at 64) .

97. Fbr the first time on redirect, PG&E Panel 6 was asked to provide a quantitative description of the scope of the " separate" ITP Phase II review. Mr. Anderson opined that the I'IP reviewed "smewhere between 45 and 50 percent" of the non-seismic "ergineerirg work" (Tr. D1419), which he added was "over and above" his estimate of 30 parcent of the engineering work he claimed the IDVP had verified (ibid.) . 'Ihis led him to the opinion

--expressed for the first time by any PG&E witness. Ch redirect on the sixth and final day of PG&E's case-in-chief -- that between 75 and 80 percent of the non-seismic design had been verified by either the IDVP or the I'IP (Tr. D1419-20) , well over half of that by the ITP alone.

(3) 'Ihe I'IP did not Cbnduct a Verification Program Independent of the IDVP

98. (bnspicuously missing frm the documentation of the ITP is any of the information one would expect to see reported on a verification program -- information that is reported for the IDVP seismic and non-seismic verification and for the DCP Phase I program. 'Ihere is no enumeration of the systems and cmponents reviewed, nor even of the ntsnber i

of such systens and cmponents; there is no identification of the criteria

against which the design of each was verified; there is no discussion of

(

32.

l I

who did the review; there is no reprt of the nunber of errors of each class -- A, B, C, and ' D - detected.

99. The IWP Final Report (PG&E Exh. 90) does refer to "an additional design verification effort to assure the overall adequacy and design of the plant" as one purpose of the DCP (id., Et 1.4-2), but the reprt goes on to describe the DCP's Phase II program thusly:

"The DCP Phase II activity is issue-oriented, respndirg on a case-by-case basis to specific non-seismic concerns identified by the IDVP and to results of internal DCP activities.

"Although the DCP is takirg corrective action with respect to specific concerns, the DCP efforts were not intended to be a emprehensive program equivalent to that performed for Phase I." (Id. , at 1.4 1.4-5. )

100. Mr. Anderson testified that this " review" was "not nearly as systematic as the IDVP's review" (Tr. D1426), and not docmented as a review process at all (ibid.). Instead, this " review" was really just a by-product of "puttirg the files . . . and . . . the paperwork in order." (Tr. D1427.)

~

He admitted that this ITP review was "not of the same depth" as the IDVP.

(Tr. D1426, D1427, D1428, D1429. ) Unlike the IWP, ". . . where they were lookirg at a systen in its totality, all of the requirements, all of the conmitments," (Tr. D1427) lookirg at three systems in a "very in-depth review, [w]e did not spend that much time on three systems." (Tr. D1428. )

"We just didn't look at it, at the systems as tieroughly as they did."

(Tr. D1429. ) "I just don't think we looked at all of the details to the extent that they did." (Ibid.)

101. The lesser scrutiny the ITP gave the design in its review is evidenced by its results. In its review, the IDVP found nine class A or B errors in what Mr. Anderson characterizes as 30 percent of the design work.

33.

(PGEE Panel 6, ff. Tr. Dll60, at 48.) In reviewing a slice of the design Mr. Anderson characterizes as being 50 percent larger than the IDVP sample (45 percent to 50 percent versus 30 percent), the ITP found only seven "open ita s." (Id. , at 64. ) So the ITP found substantially fewer design errVrs than the ID/P in a substantially larger sample than the IDVP's sample.

Since the ID/P disclaimed havirg selected its sample for a higher-than-average error rate (Tr. D1549), there is no basis to conclude that the ITP found rewer errors than the IDVP because there were fewer to be found. %e most reasonable inference is that the I'IP found fewer errors than the IDVP because its review was less likely to find them, the natural consequence of what Mr. Anderson characterized as the lesser depth of the ITP's review.

102. Accordingly, the Ibard finds that the ITP's review of the non-seismic design is not functionally equivalent to the verification per-formed by the IDVP and not an adequate substitute for the absent CA program.

(3) There is no Substantial Evidence to Support PG&E's Claim That the I'IP Beviewed 45 Mrcent to 50 Percent of the ibn-Seismic Design, Ibr to Support the Claim That the IDVP and the IIP ibgether reviewed 75 to 80 Percent 103. In his redirect testinony at the hearing, Mr. Anderson estimated that the IDVP reviewed 30 percent of the ron-seismic ergineering work. (Tr. D1419.) The direct testimony of PG&E Panel 6 indicated the IDVP sanple consisted of 911 of the 3968 " design elernents" omprisirg the non-seismic design work - less than 23 percent. (PG&E Panel 1, ff. Tr. Dll69, at 48.) Mr. Anderson offered no basis for estimating that the 23 percent of i

the " design elements" amounted to 30 percent of the "ergineerirg work."

(Tr. D1436-38.)

104. Ibr does Mr. Anderson offer any reliable basis for his est anate that the DCP reviewed 45 percent to 50 percent of the " engineering 34.

l

work." He admitted at the hearing that at his depasition he had estimated the IINP and I'IP smpled together (Tr. D1430, D1433) at 50 percent or above

-- leavirg the I'IP saple around 20 percent accordirg to his figures.

(Tr. D1426-30.) He explained his doublirg of his estimate of the ITP's saple on the basis of havirg given the matter "further review." (Tr. D1430, D1431, D1434.) Mr. Anderson could give no grounds for the estimate other than his " judgment." (Tr. D1440. ) D . Kaplan's effort to " salvage" the quantitative claim (Tr. D1423) consisted of a reference to the shading in the diagram in the prepared testimony (PG&E Panel 6, ff. Tr. D1160, at 66) -

although he had previously aimitted that the diagram does not show the frac-tion of the design work the IINP actually sampled (TR. D1323) .

105. Mr. Anderson also testified that the 30 percent he estimated the IINP verified and the 45 percent to 50 percent he thought the ITP reviewed could be added together for a 75 percent to 80 percent of the non-seismic design reviewed by the verification program. (Tr. D1420. ) But on cross-examination, he acknowledged that there is some overlap between the 30 percent he attributes to the IINP and the 45 percent he assigns to the I'IP (Tr. D1430), so that simply addirg the two figures together as he had would not "in a strict sense" be correct (Tr. D1431) . Yet Mr. Anderson maintained that the 75 percent to 80 percent figure was still correct.

(Tr. D1431. )

l 106. Mr. Anderson admitted on cross-examination that his nunbers l

estimatirg the fraction of the non-seismic design examined by the verification program were "rather sof t." 'Ihe Ibard agrees and finds that the evidence does not support any conclusion about the actual fraction of the design reviewed by the I'IP or the conbined verification program.

l 35.

C. %e Probabilistic Calculations Presented by PGE Panel 6 does not Cbnstitute Reliable Evidence on Any Material Issue Before h is Board 107. Dr. Kaplan presented a probabilistic analysis, purporting to employ the opinions of the DCP and the results of the IDVP's non-seismic design verification to calculate, usirg Bayes' Theorem, the likelihood of there remainirg an mdetected safety-significant error in the design of Unit 1. (PGE Panel 6, ff. 'It. D1160, at 53-63.) Ebr the reasons given below, the Ibard finds the analysis mreliable and irrelevant to the licensing of Diablo Canyon:

(1) The analysis focuses on the likelihood of a safety-significant error, rather than violation of the licensirg criteria. Since the Ibard has found that to be the wrong question, the answer is irrelevant.

(2) %e analysis does little more than restate the opinions of the DCP, which dominate the probabiistic results. In applying Bayes' Theorem to the verification problem as he characterized it, Dr. Kaplan consulted with the DCP about their opinions on the likelihood of a safety-significant error, instructing then to disregard the information obtained fram the sample. (PGE Panel 6, ff. Tr. D1160, at 56-56; Tr.

D1219-24. ) % is quantity, which is called the " prior distribution" (i.e. , prior to the addition of other evidence), was .001-in other words the DCP was 99.9 percent sure there were no safety-significant errors even without reference to the IDVP results. (Id., at 57.) When

he then enployed Bayes' Theoren to ccmbine the IDVP results with the

! prior distribution, he obtained a " posterior" probability of there renainirg a safety-significant error in thit 1 of .00003. (Id., at 63.)

Plainly, the results of the analysis are dcrinated by the DCP's asstrnp-l 36.

, tions. Dr. Kaplan goes fran a prior he elsewhere describes as

" essentially certain beforehand" that there is no safety-significant error (Tr. D1388) to a probability only slightly snaller. Prof.

Apostolakis slowed that a different prior distribution, which he thought a reasonable person might have had in late 1981 (Tr. D2319-20), when subjected to the Kaplan analysis with no other nodifications would yield a Insterior probability of at least one safety-significant error of 5 percent (Tr. D1260-61, D1268-70)--more than a thousand-fold increase.

(3) We analysis relies on inherently suspect opinion evidence. Prof.

Apostolakis testified that there exists a body of evidence that people in general tend to have great difficulty assessirg probabilities, "to deny mcertainty, misjudge risks, and express unwarranted confidence in their judgments." ( Apostolakis, ff. Tr. D2313, .at 6. ) %e: literature also slows people have special difficulty dealirg with rare events.

(Id., at 7.) He also cited studies -- including one he cn-authored with Dr. Kaplan showirg that experts tend to underestimate risks. (Id. , at 6-7.) his evidence underscores the weakness of an analysis that relies on opinion evidence that is too mreliable to serve as the basis for a licensirg decision.

(4) n e results of the probabilistic calculations are based on a non-random smple and therefore invalid. PG&E and the IDVP agreed that the IDVP sample was not randanly drawn. (PG&E Panel 6, ff. Tr. Dll60, at 41-42; IINP Panel 1, ff. Tr. D1458, at 33-34.) Both Prof. Apostolakis and Prof. Samaniego testified that randan samplirg was essential to drawing valid inferences. (Apostolakis, ff. Tr. D2313, at 15; Samaniego, ff.

Tr. D2392, at 5.) Prof. Samaniego stated the proposition nost strongly:

37.

that the absence of randm samplirg rendered Dr. Kaplan's use of Bayes' Theorem "wrorg," " inappropriate," and "cmpletely invalid."

(Tr. D2394-95.) ME neither recalled Dr. Kaplan or otherwise sotxjht to rehabilitate his testinony. %e Ibard therefore finds that the quan-titative results presented by Dr. Kaplan must be disregarded.

D. 'Ib Ctznplete Verification of Diablo Canyon, ME Must Establish for Each Portion of the Design Either hat it is Free of Class A or B Design Errors or h at it has Been Subjected to the Same Scrutiny it Would Have Received for Under a Legally Adequate CA Program 108. Frcm the foregoirg follows what ME must do to meet the requirenents of a verification program and to qualify the design of Diablo Canyon Units 1 and 2 under the Ctznmission's regulations:

(1) Ebr those portions of the design we have found above to have already been verified, no further review of the verified design is required.

(2) Ebr thase portions of the design that we have not found ME to have denonstrated verification, but where ME believes it can provide documents sufficient to meet the standards for a verification program we have established here, ME may use existirg documentation to demonstrate the design has been verified.

, (3) Ebr the ITmaining partions of the design, ME retains the option for each such partion of either (a) demonstratirg in a statistically valid l

manner that there the design is free of class A or B errors or (b) subjectirg that portion of the design to a 100 percent review.

The Ibard finds that until ME has done so, the Ctznmission cannot have the necessary confidence that the design meets the licensirg criteria and regulatory standards. Se Ibard further finds that once ME has crrnpleted l

this task, it will have satisfied the requirements of a verification I

38.

1

, program, overcame the past shorteamings of its design CA program, and qualified t .^ design of the facility.

III THE SEISMIC VERIFICATION PEREDRMED BY THE ITP AND 'IEE IDVP IS DEFICIENT IN THAT IT DOES N0r DEMONSTRATE 'IEE SEISMIC CUALIFICATION OF ALL STRUCIURES A. PG&E Used Incorrect Methods to Model Soil Structure Interaction for (bntairrnent 109. As part of its nodeling of the seismic response of the contairrnent buildirg, the DCP included an analysis of the interaction of the tehavior of the underlyiry foundational material (the rock underlyirg contaniment) durire the appropriate earthquakes, and the effects of that behavior on the seismic response of the contairrnent buildirg. %is is known as a soil structure interaction analysis. In that soil structure interaction analy-sis for containment, the DCP used an axisymmetric soil mass representation, a nodel that represent the foundational material as a cylinder. (Tr.D655.)

110. In performirg the soil structure interaction analysis, the DCP applied the same earthquake notion input to both the side and bottom boun '

daries of the soil mass model, a procedure inconsistent with current soil structure interaction nodelirg. (Tr. 0655, D656, D2294-98) . 'Ihis boundary i

motion input procedure produces notion at the free surface that varies from each edge to the center of the soil mass. %is also is not aansistent with current methods of soil structure interaction analysis. (Tr. D661.)

• 111. Finally, the DCP applied a correction factor to the model to off-set the discrepancies in free surface notion fran the true design spectra caused by its method of inputtirg the same earthquake notion at toth bottan and side toundaries of the soil mass. Use of such correction factor is not 39.

I __

consistent with current practice. (Tr. D659-60.)

112. Ibwever, and largely due to the fact that the soil mass under con-tairnent is hard rock, the correction factor used by the DCP has produced results sufficiently conservative to be acceptable. (Tr. D2638-39.)

, B. ME Has Pedeled the Soil Springs thder the Auxiliary Building Inconsistently, and Has Failed to Take Into Account All Effects That the Springs May Cause (1) 'Ihe DCP Has Ibt Cbnsidered a Fixed Base Analysis 113. DE modeled the soil mass mder the auxiliary building by asstnirg a fixed base (i.e., very stiff or rigid foundational material under the structure) at elevation 85 feet, and soil springs (a way of representirg less stiff, nore flexible soil or rock) at elevation 100 feet.

Expert testimony and exhibits dealirg with the value of the soil repre-sented by the soil .=prirgs used to represent the soil under the auxiliary buildirg at elevation 100 feet varied widely. Estimates rarged fran 3500 feet per second shear wave velocity by the IDVP in one ITR (IDVP, "Best Estimate," Table 7, ITR 55, ME Exh. Ib. 145), to 2700 feet per second by ME (Tr. D701), to 2500-2700 feet per second by the Staff (Tr. D2515), to approximately 1500 feet per second by the IDVP in a second ITR (based on conversion of ompressional wave data in Figure 15 of ITR 68 to shear wave

! data; see Tr. D2226) .

114. 'Ihe DCP used a fixed base soil structure interaction analysis for the auxiliary buildirg at elevation 85 feet for all earthquakes. 'Ihe IDVP concurred with this apprcach, notiry that the Ibsgri sport provides for a fixed base analysis for foundational material of 3500 feet per second shear wave velocity (ITR 55, ME EKh.145, p.19), a value that indicates very stiff rock. 'Ihe IDVP felt that the approach was justified for Ue DE 40.

and the DDE because of the stiffness of the rock (ITR 55, page 32).

115. Ibwever, no fixed base soil structure interaction analysis was perfomed for the auxiliary building at elevation 100 feet, despite the IDVP's estimate of the foundational material at that elevation as 3500 feet per second shear wave velocity, the same shear wave velocity used to justify the use of a fixed base analysis at 85 feet for all earthquakes.

116. Ib justification was presented for use of the fixed base analysis at elevation 85 feet and lack of such fixed base analysis at elevation 100 feet (Tr. D700), despite the fact that IDVP estimated shear wave velocity of the rock at toth elevations as beire 3500 feet per second or higher.

117. An increase in the forces in the shear walls of the auxiliary buildirg could occur if the foundational material at elevation 100 feet were not asstned to be flexible, as would be the case in a fixed base ana-lysis. (Ibesset, ff. Tr. D2206.) All analyses have asstned flexibility in the soil at elevation 100 feet (ITR 55, PG&E Exh.145, Table 5). %e results of Table 5 in ITR 55 seem to indicate that the base shear force increases with increasing stiffness of the sprirgs, and then decreases again. (ITR 55, PG&E Exh.145, Table 5.) It has not been shown, lowever, that the base shear would continue to decrease with a fixed base. Ib showirg has been made by the DCP that the shear walls are qualified for any increase in forces that might ccne frcm a fixed base, other than by citirg the possibility of inelastic deformation not provided for in the licensirg criteria. (Tr. D713-15. )

(2) %e DCP Has Ibt (bnsidered Ibtational Effects of Flexible Soil Sprirgs 118. %e analysis of the auxiliary buildirg has not accounted for all effects that might occur if the foundational material represented by soil 41.

. sprirgs is as stiff as the IINP estimate of 3500 feet per second shear wave velocity, as discussed above. It is also the case that the DCP has not considered all effects that could occur if the soil sprirgs are as soft as the low estimate implied by Figure 15 of ITR 68. (PG&E Dchibit 155.)

119. Variation in the stiffness of the soil of the magnitude described above can produce variations in the seismic response of the auxi-liary buildirg. ITR 55, PG&E Dchibit 145, page 25, states that a 20%

increase in response spectra acceleration would cccur over the values of the soil sprirgs considered by the IINP in its parametric study of soil sprirg values. 'Ihis 20% increase in response spectra calculated by the IIVP was based on a lower bound for the value of the soil sprirgs of 2000 feet per second shear wave velocity. ITR 55 notes that that much of an increase is acceptable to the IINP specifically because this lower tound value is conservative (ITR 55, page 25) . Although the IINP testified that the soil sprirgs did not have much effect (IINP Panel 2, ff Tr. D1843 at 3-9), the IINP parametric studies showed a large variation in the seismic response of the base at elevation 85 feet of the Auxiliary Buildirg with variation of t?n soil sprirg values (Tr. D2470-71, D2220) .

120. A buildirg with a foundation that is both enbedded and at dif-I ferity elevations, as is the auxiliary buildirg, will experience varyirg earthquake notions at the nore deeply and at the nore shallowly embedded portions of foundation, respectively. (Tr. D2524.) This variance in motion l experienced by the different portions of the base cnuld produce a rocking l

l motion, which would vary with the stiffness of the soil. (Ibesset, ff. Tr.

D2206, at 11; Tr. D2526.)

121. 'Ihe IINP parametric studies did rot address the effects on the l

! 42.

l

, seismic response of the auxiliary buildiry of considering the soil sprires to have a value of 1500 feet per second shear wave velocity, as Figure 15 of ITR 68 implies (see Findirg 113, supra). Material of this value would be hard soil rather than rock, and would be softer than r.n lowest value of soil sprirg material considered by IDVP. (ITR 55, PG&E Exh.145, Table 5; Tr. D2226.) Testinony of the NRC indicated that any rocking notions of the auxiliary buildirg referred to in Findirg 120, supra, would be increased by such softer soil. (Tr. D2526.)

122. Despite the possibility that flexible soil springs could produce rockirg notions of the structure, that could be increased by an asstanption that the soil sprirgs are softer than the IDVP asstrned in its parametric studies, no analysis has been done by PG&E or the IDVP of the effects on the seismic response of the auxiliary buildirg of rotational effects.

123. A fixed base analysis of the soil structure interaction for ele-vation 100 feet of the auxiliary buildirg, would be consistent with the fixed base analysis done for elevation 85 feet, if the shear wave velocity of the foundational material at both elevations is 3500 feet per second.

The DCP has failed to perform such a consistent analysis, or to qualify the auxiliary buildirg for any effects predicted by such an analysis.

124. 'Ihe DCP has assumed soil sprirgs of a different shear wave velo-city at elevation 100 feet than at elevation 85 feet. Ibwever, the Ibard finds that the DCP has failed to do an analysis for all credible values of the soil sprirgs, and has failed to qualify the auxiliary buildity for all effects, includirg rocking notions, that such soil sprirgs could cause.

C. _'Ihe DCP Has Failed to Qualify All Equipnent in Contairrnent Ebr the Anticipated Effects of an Uplifting of the Containment Base Mat 125. General Design Criterion 2 of Appendix A to 10 C.F.R., Part 50, 43.

requires that all structures, systems, and cmponents important to safety be designed to withstand the effects of natural phenmena, such as earth-quakes. In addition, licensing criteria specify maxim m allowable stresses for which various systems, structures, and camponents are qualified and to which they may be subjected.

126. Dcpert testimony by Dr. Ibesset established that the contaiment base mat, the large concrete slab that is the foundation of the building, might separate frm the rock beneath containment and lift up if the con-tainment experienced accelerations over 0.4g, well under Hosgri accelera-tions. (Ibesset, ff. Tr. D2206, at 8.) All parties testified that uplift might occur at Ibsgri accelerations (0.679). (Ibesset, id.; Tr. D675; D2507-10; D1888; D1875; Dl876; Dl954-55.)

127. Such an uplifting of the contairrnent base mat would cause a charge in the center point of contact (the centroid) between the base mat and the underlying rock (Ibesset, ff. Tr. D2206 at 6-7), producing a shift in the axis of rotation (Tr. D2213-14), and increast.,3 vertical accelera-tions not accounted for in the DCP analysis of mntainment (Tr. D1881, D2215-16), as the buildire rises frm the ground very slightly at one edge.

The increased vertical accelerations frm uplift could cause increased ver-tical stresses on equignent in wntairinent. Such increases muld be on the ortler of 10-15%, but could be especially significant for equiprent whose qualification is controlled by the vertical cmponent (Tr. D2216; Tr.

D2214-15; Tr. D2510).

128. ?b analysis has been done by the DCP or the IDVP of the effect on any equipnent inside mntairunent of increased vertical acceleration due to uplift (Tr. 228) .

44.

129. %e testimony clearly established that uplift of the contairment base mat may occur at Ibsgri accelerations, and may occur at lower acce-lerations. All parties agreed that uplift could cause increased accelera-tional in the vertical direction on the equipnent in contairnent, and that no one has done any analysis to determine whether all or any equipnent in contairunent is qualified for such accelerations. Thus, the Ibard finds that the DCP has failed to qualify all safety-related equignent for possi-ble effects of earthquakes, in violation of General Design Criterion 2.

D. %e DCP's Use of Translational and 'Ibrsional Inputs to the Fuel Handling Building Has Not Been Shown to be Valid 130. Cbntention 3(o) challenged the validity of certain aspects of the DCP's modelirg of the fuel handling building, particularly how the DCP used the motion of the auxiliary buildirg to determine the tration to input to its rrodel of the fuel handlirg buildirg. ?b findirgs are yet possible regardirg the validity of the DCP's use of translational and torsional inputs to the fuel handlirg buildire, since no written analysis confirming the DCP use of input to the base of the coltrnns has yet been subnitted to the parties or the Ibard. %e Staff has been given information orally, which it feels would be adequate if confirmed, but has not had confirmirs and verifiable written substantiation. (Tr. D2528-30.)

E. No Cbnsistent, Doctrnented Representation of the Properties of the Soil and Ibck Surrounding the Diesel Fuel Oil Tanks Has Been Produced by the DCP or the IDVP (1) %e data used to determine soil properties was determined without a OA program, and tests have not been redone by the DCP or the IDVP 131. %e testimony of PG&E, the IDVP, and the NRC regardirg soils pro-perties of the backfill surroundity the diesel fuel oil tanks ma all based on the work of Hardirg Lawson Associates. (ITR 68, PG&E Exh.155, page 1; 45.

. NRC Panel 2, ff. Tr. D2463, at 24; Tr. D763.) 'Ihe IDVP performed a quality assurance review of Hardirg Lawson Associates, and found that it imple-mented no quality assurance program for the Diablo Canyon soils work. (ITR 68, PG&E EKh.155, p. 2.) PG&E has not reperformed any field tests, labora-tory tests, or samplirg done by Hardity Lawson Associates. Also, it has not verified this work except by a general review for reasonableness of results. 'Ihat review appears to have been done by only one er two persons, and it is not clear if any of this review was documented. 'Ihe testimony seems to indicate it was not. (Tr. D767; D773; D775; D798.) 'Ihough, the ID/P reviewed the geologist's log.and the data for reasonableness of results, and cmpared data to the literature, it performed no physical tests. (ITR 68, PG&E EKh.155, pp. 4-5. )

(2) ITR 68 presents contradictory information about the backfill around the diesel oil fuel tanks 132. In Figure 14 of ITR 68 (PG&E EKh.155), contradictory information is presented as to the soil properties of the soil surrounding the diesel fuel oil tanks. 'Ihe same soil is represented by a shear nodulus versus mean effective stress curve with an exponent of 0.18, and by a curve with an exponent of 0.7. 'Ihese exponents differ by a factor of nore than three, and this difference implies a large uncertainty in representation of the soil properties. (Ibesset, ff. Tr. 2206, at 18; Tr. D2539.) Examination of each of these exponents reveals nore uncertainty as to soil properties.

Figure 14 of ITR 68 purports to describe soil that is sandy clay (Tr.

D771). Ibwever, the exponent of 0.18 is very low for a sandy clay (Ibesset, ff. 2206 at 18; Tr. D2537), and the exponent 0.7 is very high, beirg closer to a sand (Ibesset, ff. Tr. D2206, at 18; Tr. D2538; Tr.

46.

, D779). 'Ihe ITR does not make clear either the nature or the properties of this backfill soil.

(3) ITR 68 presents information about the rock around the diesel fwl oil tanks that contradicts data used to qualify the auxiliary building 133. 'Ihe diesel fuel oil tanks are near the auxiliary building, and the material surrounding them is very near or beneath that building. (ITR 68, ME Exh.155, Figure 5, p. 71.)

134. ITR 68, Pigure 15, presents otrnpressional wave data which, when converted to shear wave velocities, show a shear wave velocity for the rock around the tanks and around the auxiliary buildirg at 100 feet of about 1500 feet per second (Tr. D2227), a value for which no soil spring analysis has been performed (Findirg 122, supra). Strains associated with a lbsgri event might reduce the shear wave velocity of the rock around the auxiliary buildiry to 1000 feet per second or lecs (Tr. D2227), based on extrapola-tions fram the data in Figure 15 of ITR 68. Ib analysis has been performed for rock of this value (ITR 55, ME Exh.145.)

135. NRC and IDVP witnesses testified for the first time under cross-examination that the data in ITR 68, Figure 15, should be disregarded as "less reliable" for calculatire the shear wave values of the rock around the auxiliary buildirg than other data gathered from up-hole and cross-Inle tests performed by Harding and Lawson in 1967 and 1973. (Tr. D2548; Tr.

l D3122; Tr. D3112.) 'Ihe IDVP witness, Dr. Cloud, offered as his only reasons for discountirg tha data in Figure 15 of ITR 68, his judgment that the test method used to gather the data in Figure 15 is less reliable by same unquantified margin at this site than the up-hole and cross-Inle tests winse results were relied on in qualifyiry the auxiliary building. (Tr.

47.

l .

, D3112; D3125; D2010-ll.) No other evidence was adduced that the data in Figure 15 were not correct.

136. We cross-hole and up-hole data mgarded by the IDVP as of greater reliability than the data in Figure 15 of ITR 68 were gathered by Hardirg Iawson Associates (Tr. D3125; D3137) durirg the period of time for which they had no OA. prcgram. (ITR 68, PG&E hh.155, p. 2. )

137. %e Board finds that the data in ITR 68 and the testimony revealed pervasive inconsistencies in the properties claimed for the back-fill soil around the diesel fuel oil tanks. No conclusions as to the pio-perties of that soil can be drawn, based on the conflictirg and internally inconsistent evidence before the Board.

133. He evidence also revealed that no consistent, well-documented representation of the properties of the rock surroundirg the diesel fuel oil tanks and underlyirg the auxiliary building at elevation 100 feet.

Upon the evidence before the Ibard, it is not possible to find that the soil sprire values used in the seismic analysis of the auxiliary buildire are correct, nor is a findirg possible as to what the properties of the rock surroundirq the diesel fuel oil tanks are.

F. No (bnsistent, Doctnented Representation of the Properties of the Soil Surroundiry the Auxiliary Saltwater Pipirg and Circulatirg Water Intake Conduits Has Been Presented by the DCP or the IDVP (1) ne data used to determine soil properties was determined without a CA prcgram, and tests have not been redone by the DCP or the IDVP 139. %e testinony of PG&E, the IDVP, and the NRC regardirg soils pro-perties of the backfill surroundirg the diesel fuel oil tanks was all based on the work of Hardirg Lawson Associates. (ITR 68, IGE hh.155, p.1; NRC Panel 2, ff. 'It. D2463, at 24; Tr. D763.) %e IDVP performed a quality 48.

. assurance review of Hardirg Iawson Associates, and found that it imple-mented no quality assurance program for the Diablo Canyon soils work. (ITR 68, ME hh.155, p. 2.) ME has not reperformed any field tests, labora-tory tests, or samplirg done by Hardirg Lawson Associates. Also, it has not verified this work except by a general review for reasonableness of results. Wat review appears to have been done by only cne or two persons and it is not clear if any of this review was documented. %e testimony seems to indicate it was not. (Tr. D767; D773; D775; D798.) hough the ID/P reviewed the geologist's log and the data for reasonableness of results, geologist's log and cmpared data to the literature, it performed no physical tests. (ITR 68, M E h h. 155, pp. 4-5.)

(2) Data on low strain shear modulus of the ASW and CNIC mil are not documented adequately 140. he evidence shows that the soil data in Figure 22 of ITR 68 (ME hhibit 155), are not reliable as representations of the low strain shear modulus of soil surroundirg the auxiliary saltwater pipirg since the source of the data, both as to sample test site and as to testirg program, is not presented. We data were gathered frm testing performed near the auxiliary saltwater pipirg and frm near the diesel fuel oil tanks (ITR 68, E E h h. 155, p. 43). Ibwever, the soils frm tohse sites are of different types (ME Panel 2, ff.

D651, at pp. 85-86), and Figure 22 does not specify which data were collected frm which site (ITR 68, ME hh.155, Figure 22; Tr.

D2555). Further, ITR 68 states that tests were performed for values of 500-3000 psf (ITR 68, ME hh.155, p. 43), but Figure 22 shows data extendirg to 4500 psf. Ib explanation is given in the ITR for how, where, or by whom the tests were done that yielded these data.

49.

d

. 141. 'Ihe Ibard finds that the unreliability of the data in Figure 22 casts doubt on the curve for shear modulus at low strain that appears in that figure. Further, the unreliability of the data al.so casts doubt on the curves in the ITR that depicts shear nodulus at high strain. In soils analysis, curves depictirg variation of shear modulus with strain are

" normalized" (i.e., adjusted) at low strain values, so that inaccuracies in low strain curves, such as Figure 22 of ITR 68, may result in curves that do not accurately represent shear modulus at either low or high levels of strain because of improper normalization. (Tr. D2222.)

(3) Data on shear nodulus at high strain presented in ITR 68 do not accurately represent in situ soil properties 142. Figures 23 and 24 of ITR 68, IGE Dchibit 155, show data points derived fran laboratory tests conducted by Harding Lawson Associates. (ITR 68, PG&E Dch.155, pp. 42-43. ) Soil samples gathered in the field and tested in the laboratory are often disturbed in the process of collection, storage, and transport. (Ibesset, ff. D2206, at 20; 'It. D2558; D806.)

Therefore, laboratory values are generally lower than field values. In soil properties testirg (Ibesset, ff. Tr. D2206, at 20), laboratory test data are usually adjusted by a correction factor to account for phenomena such as sample disturbance, and to make the data more representative of in situ condition (Ibesset, ff. Tr. D2206, at 20; Tr. D2558).

143. Ibwever, appropriate correction factors are often difficult to determine, because they depend on the degree of sample disturbance. NRC expert witness, Dr. Cbstantino, said he could not estimate a particular correction factor triless he saw the soil sample ocune out of the tube. (Tr.

D2559-60.) Dr. Ibesset testified that appropriate correction factors for the laboratory data could be in the rarge of a factor of two. (Tr. D2223.)

50.

, Dr. Cbstantino for the NRC estimated a factor smewhat less than two. (Tr.

D2560.) IGE expert witness Dr. Seed testified that the appropriate correction factors for the ASW backfill would rarge frm about 1.25 at low strain levels to about 1.0 at high strain levels, based on the fact that the backfill was highly empacted and would be less disturbed during sampling than an uncmpacted soil. (Tr. D3118. )

144. tb correction factors were applied to the laboratory data pre-sented in Figures 23 and 24 of ITR 68, PG&E Exhibit 155, to make them nore representative of in situ conditions. (Tr. D806; D2557-58.) This failure to apply correction factors causes in these figures to be inaccurate repre-sentations of the properties of the ASW and CWIC soils. Dr. Ibesset testified that the curve in Figure 23 of ITR 68, representirg the high strain values of shear nodulus varying with shear strain of the backfill, would have to be raised, if the test data in that Figure were corrected to more realistically represent in situ values. (Tr. D2224.)

145. Raisirg the nodulus curve muld also affect the dampirg curves for the ASW/CWIC mil. Des. Ibesset and Costantino testified that the dampirg values for the auxiliary saltwater pipiry backfill shown by the curve in Figure 24 of ITR 68 (PG&E Exhibit 155) were already very high for the type of soil represented. (R3esset, ff. 2206, at 21; Tr. D2564.)

! Dr. (bstantino stated that he would estimate the magnitude of dampirg as lower by a factor of two than what is presented in Figure 24. (Tr. D2564. )

i 146. Ebth Dr. Ibesset and Dr. Seed testified that if the lateratory test data points in Figure 24 of ITR 68 were corrected to make them nore representative of in situ conditions, the curve showirg dampirg in that Figure would drop. (Tr. D2224; D3131.) 'Ihe seismic response of the auxi-51.

L

liary saltwater pipirg and circulating water intake conditions muld be affected by a variance in the shear nodulus and dampirg values of the surroundirg backfill. (Tr. D2224-55.) Therefore, unless the true proper-ties of the backfill are known, the seismic qualification of these systems renains unproved.

147. 'Ibe Ibard finds that neither PG&E nor the IDVP has produced evi-dence demonstrating the soil properties of the soil surrounding the auxi-liary saltwater pipiry and the circulatirg water intake mnduits in a con-sistent and well-doctanented fashion. 'Ihe seismic qualification of these systems has not been shown and remains in doubt until the soil properties have been properly demonstrated and the pipirg and conduits shown qualified for such soil properties.

G. 'Ihe IDVP Did Not Perform Independent Modeliry of the Seismic Besponse of the Diablo Canyon Structures, Althotqh Such Modelirg Would Give Increased Cbnfidence in the Design of Those Structures -

148. Brookhaven National Iaboratories performed independent nodelirg of the contairinent annulus and the buried diesel fuel tanks at the regaest of the Staff. 'Ihe independent modelirg done by Brookhaven consisted of obtainiry design information as to the structure, member mnnectivity, mass distribution, soils data and other features of the annulus and the buried tanks, developirg seismic response models independe:.cly of those of '.he applicant, and runnirg the models to obtain final results. All Brookhaven work was done without reference to the nodelirg the DCP had done, until final results were ocznpared. (SSER 18, Staff Ekh. 36, section 3.6.)

149. 'Ihe results of the Brookhaven modelirg of the mntainment annulus revealed the importance of floor flexibility and local nodal participation in the seismic response of the annulus, effects not previously accounted 52.

for by the applicant. %e Brookhaven independent nodelirg also produced different response spectra, includirg different spectral peaks for both vertical and horizontal responses. (SSER 18, Staff hh. 36, m. C.3-92-93, 3-95.) Sus, the independent nodelirg revealed important features of the annulus response not revealed by the original PG&E modelity. %ese insights had sufficient value that the results of the Brookhaven modelirg were relied on by the staff as a benchmark in understandirg the seismic response of the contairrnent annulus even after the annulus had been modified and reanalyzed by the applicant. (SSER 18, Staff Exh. 36, at C.3-5, C.3-9; Tr. D2574; Tr. D2575.)

150. In addition, the Brookhaven independent nodelirg of the buried diesel fuel tanks revealed that PG&E's nodel did not accurately represent the seismic response of the tanks and the fluid within the tanks. (SSER 18, Staff Exh. 36, section 3.6.6; Tr. Dl935. ) In response to the results of Brookhaven independent modelire of the buried diesel tanks, PGr.E com-mitted itself to further nodelirg and charges in its nodelirg of the tanks.

(SSER 18, Staff Exh. 36, p. C.3-99. ) %is denonstrates the vslue of the independent nodeliry of the tanks.

151. Independent nodelirg of the type done by Brookhaven for the con-taniment annulus and buried diesel fuel tanks is useful as an independent l

j and unbiased basis for determinirg the seismic accuracy of a structure design. (Ibesset ff. D2206 at 22-23; Tr. D2619.) Where, as in the case of l the Brookhaven nodels, there are differences, the independent model may indicate problems or provide fresh insight. Certainly, disagreement of two models indicates that further investigation is needed. (Tr. D1938-39.)

[

Where the results of such independent nodeliry are in close agreement. with l

53.

l 1

, the results of the original model, there is reason for increased confidence in the original results. (Tr. D1938; Tr. D1901.)

152. Alttough certain aspects of the seismic qualification of various structures at Diablo Canyon were independently modeled by the IDVP, the IDVP performed no independent nodeling, of the type done by Brookhaven Natior'al Laboratories for the containment annulus and the buried diesel fuel tanks, of the ctrnplete seismic response of any buildirg at Diablo Canyon. (Ibesset, ff. D2206, at 22; Tr. D2619-20; Tr. D1939-44.) If inde-pendent noderirg such as that done by Brookhaven of the complete seismic response of each buildirg at Diablo Canyon were done, increased confidence in the. design of each structure would result. (Ibesset, ff. 2206, at 24; Tr. D2291-92.) Increased confidence would also result fran independent modelirg of each buildirg usirg simplified models. (Tr. D2292; D2294-95.)

153. 'Ihe Ibard finds that independent nodeling of the seismic response of buildings, such as the nodeling done by Brookhaven National Laboratories for the containment annulus and the buried diesel fuel tanks, would give added confidence in the design of the buildirgs at Diablo Canyca, as an unbiased check on the design and seismic response of those buildings.

While more simplified models than the Brookhaven nodels could be used, such modelirg should be done by the IDVP for each buildirg at Diablo Canyon.

IV PG&E HAS FAILED 'IO ANALYZE ALL 'IHE JET IMPINGEMENI' IIRDS IT IS REQUIRED 'IO CCNSIDER A. PG&E Has Misconstrued the Requirements of Its License Application With Respect to Analysis of Jet Impingement Ioads Inside Contairunent 154. PG&E's license application requires it to perform jet impingment analyses for high energy line treaks occurrirg inside containment. (PG&E 54.

L

i

. Panel 3, ff. Tr. D487, at 25. ) However, M E testified that it believed the FSAR to be silent as to the specific temperature / pressure criteria to be utilized for conductirg high energy line break analyses inside mntain-ment. (Tr. D589.) As a result, M E used the criteria which had been developed for jet impirgement and pipe whip analyses outside containment.

l (Tr. D589.) Wese criteria were set forth in section 3.6-16-17 of the FSAR.

(Tr. D588-89. )

155. ME's interpretation of the criteria mntained in that portion of the FSAR led it to the mnclusion that an analysis was only required for lines which were subject to both high temperatures and high pressures.

(Tr. D589.) Subsequent to reachiry this conclusion, ME ascertained that tMugh there did not appear to te criteria for high energy line breaks inside containment, there was a listing of lines inside contairrnent, at l page 3.6-11 of the FSAR, which definitely had to be analyzed for such breaks. (Tr. D589-90.) %e lines listed were all subject to both high temperature and high pressure. (Tr. D589.)

156. M E's interpretation of section 3.6.4 of the PSAR is not con-sistent with the express larguage of section 3.6.4. his section provides, in pertinent part, as follows:

"%e following criteria and definitions apply to the selec-tion of piping systems outside contairrnent for evaluation of the dynamic effects associated with pstulated pipe rupture.

"1. All systems having a service temperature greater than ,

200'F or_ a design pressure gre5ter than 275 psig are considered. )

"Open crack breaks are pstulated to occur in the nost adverse locations in piping having fluid temperature or pressure greater than the above." (Dnphasis added.) (FSAR seciiion 3.6.4,

p. 3.6-16-17.)

R us, contrary to M E's interpretation, there are circumstances in which 55.

section 3.6.4 requires the consideration of jet impirgement and pipe whip effects for pipes subject to either high temperature or_ high pressure.

157. Ibwever, in usirg section 3.6.4 of the FSAR, ME ignored the criteria for pipe whip restraints inside contairsnent at 3.6-13-3.6-14 of the ESAR, which require restraints on pipes which are subject to either high ternperature or_ high pressure (FSAR section 3.6-13-14.)

158. On cross-examination EE witness Cbnnell conceded that there is no difference between the criteria to be used for postulatirg a break for pipe whip effects and jet impirgement effects. (Tr. D600.) The Ibard finds that ME stould have utilized the criteria set forth at 3.6-13-14 of the FSAR to guide its jet impirgement analysis inside contairunent.

B. E E's Misconstruction of the Criteria Has Ieft Three Lines Unreviewed 159. In failirg to utilize the correct criteria for its jet impinge-ment analysis inside contairunent, ME failed to analyze jet impirgement effects on three lines within contairrnent. (Tr. D613-14.) As a result, the Board finds that ME has failed to meet the requirements of its license application with respect to the performance of jet impirgement analyses inside containment and that such an analysis should be performed.

V THERE IS NO ASSURAtCE 'IHAT 'IHE PIANT AS BUILT IS CDNSISTENT WI'IH 'IHE DESIGN ANALYSES A. Diablo Canyon Has Had a History of Configuration Control Deficiencies 160. In 1979, the NRC issued Investigation and Enforcement Bulletin No. 79-14 requirirg licensee and permittees to verify that facilities were beirg constructed in conformity with their designs and with the analyses l

that established that the designs met their licensirg requirements. E E's 56.

response to I&E Bulletin 79-14 disclosed discrepancies between the as-built and the as-analyzed mnfiguration of thit I such that analyses had to be redone and sme plant modifications had to be made to obtain the proper analysis. (Tr. D2946.)

161. Dr. Cloud's initial IDVP analyses of Phase I disclosed as-built vis-a-vis as-analyzed discrepancies. (Tr. D2948. )

162. DNL's analysis of vertical response in the containment annulus structure showed discrepancies between the as-built plant and the design drawings. (Tr. D2946-48.)

163. Institute of Nuclear Power Operations recmmendations for changes in configuration control have highlighted deficient nE practices with respect to mnfiguration control. (Gov. Dh.11, p.10; Hubbard, ff. Tr.

D2084, at 13-14.)

B. 'Ihe DCP's Cbnfiguration Control Practices Have Not Significantly Improved Upon ME's Historic Deficiencies 164. Generally, the DCP utilizes a canbination of M E and Bechtel CA procedures. (ITR 41, n E h h. 133, pp. 5-6.) Ibwever with respect to configuration control, the DCP uses ME Rgineerirg Department procedures.

(EE hh.161; ME Panel 1, ff. Tr. D224, at 32; Tr. D1709-10. )

165. 'Ihough ME has sought to nodify sme of their procedures to rec-tify past deficiencies (Tr. D462), configuration control errors have still been identified by the IDVP (Tr. D1639). 'Ihese errors have been identified in EDI's 1120, 1121, 1123, 1124, 1133, 1135 and 1137. (Hubbard, ff. Tr.

D2084, at 15.) In addition, ntrnerous other configuration control discre-pancies were documented in ITRs 59, 60, 61, 63 and 65. (ME hhs.149, 150, 151, 152, and 153; Hubbard, ff. Tr. D2084, at 16-17.)

166. 'Ihese errors and discrepancies were so conspicuous they were

57. ~ ~

-.w - - -

i

, uncovered by the IINP although it was not doirg a program of checking drawirgs (Tr. D1641) or of mnfirmirg the configuration control product (Tr. D1635) .

C. The Verification Program IX)es tbt Supply the Missirg Assurance That the Discrepancies Have Been Feconciled 167. 'Ihe IINP does tr)t provide assurance that all discrepancies bet-ween the plant as built and as designed have been reconciled.

168. 'Ihe IINP has not verified that Unit 2 as built conforms to the design drawirgs an:1 analyses. (Morrill, ff. Tr. D2906, at 2.)

( 169. 'Ibe IINP can only verify work which has been performed. Thus, to the extent that neither PG&E (prior to tbvember 1981) nor the DCP (af ter-wards) has reconciled as-builts to design drawirgs and analyses, the IINP can shed no light on the existence or adequacy of such a reconciliation.

170. Insofar as the DCP's design work is concerned, the as-built reconciliation process has not progre sed to the point where even the pro-cedure, let alone the configuration mntrol product, muld be checked.

Though the IINP, in ITR 41 (PG&E Exh.133), concluded that the DCP as-built procedures had been effectively implemented (PG&E Exh.133, p. 53; Tr.

D1624-25) , there is no basis in the record for such a conclusion. (Tr.

D1624-36.) In fact, the checklists utilized by Mr. Ibedy's organization expressly stated, with respect to the as-built process, that this aspect of the program could not be audited because not enough wrk had been omipleted at the time of the tbvember 1982 audit to make a reasonable critique of it.

(Gov. Exh. 48, pp. 31-33; Cbv. Exh. 49, Ip. 31-33; Tr. D1634-36.)

171. 'Ihough the IINP performed an additional audit on March 17, 1983, that audit, which the documentation indicates was conducted in a sirgle day (ITR 41, PG&E Dch.133, p. 2; Tr. D1668), did not attempt to verify the as-58.

, built process or any other formerly tnauditable aspects of the DCP CA program, but rather was limited miely to verifyirg that the 24 conditions noted in the Ibvember audit had been mrrected (Tr. D1704-05) .

172. 'Ibe Ibard therefore finds that the IDVP cannot be said to have verified the DCP as-built process, nor the mnfiguration control product of that process.

173. PG&E's ergineering procedures do not require that all drawings of safety-related structures, systems, and cmponents be brought to as-built condition until one year after licensirg. (PG&E Engineerirg Department Procedures 3.60N and 3.7, PG&E Egh.161.) Thus, there can today be no assurance that the I'IP has uncovered all discrepancies between the plant as built and the plant as analyzed.

174. Given the history of as-built discrepancies at Diablo Canyon, the PG&E cmmitment to bringirg drawirgs of safety-related SS&Cs into confor-mance with design drawirgs and analyses does not satisfy the requirements of the Ctrnmission's agulations (Appendix B, Criteria III, VI, X, XVI) for prmpt reconciliation of the plant as built with the design documents.

(Hubbard, ff. Tr. D?084, at 17-18. )

175. 'Ihe Ibard therefore finds that it is necessary, in order to establish that Diablo Canyon configuration control practices mnform to the regulations, that all the as-built drawirgs be prmptly reviewed and their conformance with design drawirgs and analyses be verified.

VI

'IEERE HAS BEEN 10 VERTIFICATION 'IHAT

'IHE WESTINGHOUSE IESIGN OF SAFETY-REIATED EQUIkMENT MET LICENSING CRITERIA 176. In answers to interrcgatories, PG&E admitted that the verifi-

39. -

~~

. cation program failed to verify that the design of safety related equipnent supplied to EE by Westinghouse met licensirg criteria. (Hubbard, ff. Tr.

D2084, at 19-20.)

177. 'Ibe only activities the IDVP has undertaken with respect to a review of Westirghouse's effort on Diablo Canyon have been:

(a) to verify that there was a process for a:>ntrolling the l interface between ME and Westirghouse and to sample a portion of the 1

transmittal of the Ibsgri seismic design spectra frm ME to

( Westirghouse (Hubbard, ff. 'It. D2084 at 20; EE Panel 4, ff. Tr.

D1088, at 2); and (b) to verify the design interface between ME and Westirgbause with respect to the ron-seismic design of the auxiliary feedwater systen (ME Panel 4, ff. 'It. D1088, at 4) .

178. Errors have been uncovered in the Westirghouse design work. BNL disclosed the fact that 30% of the W>stirghouse samples taken by the IDVP contained errors, includirg, amorg other thirgs, the incorrect use of tau filterirg. (Hubbard, ff. Tr. D2084, at 22-24; ITR 11, ME Exh.103.)

Errors were found in the seismic qualification of the main control teard.

(Hubbard, ff. Tr. D2084, at 22-23; Tr. Dll21-24.) 'Ihe IDVP discovered the fact that transmittals frm EE to Westirghouse smetimes did not identify to which unit the information was applicable. (ITR 42, EE Exh.134, pp.

53-54; Tr. D1646-50, 1670-71. )

179. While the IDVP found that ME did not identify to which unit information transmitted to Westirghouse referred, the IDVP did not consider such a failing to be significant (ITR 42, ME Exh.134, p. 53; 'It. D1648) and in fact testified that it was arguable whether the failure to identify 60.

to which mit infomation referred was consistent with good m practice (Tr. D1649) . Staff witness Haas testified that nE's failure to identify for Westirghouse the appropriate unit for which detign infomation was applicable constituted an interface problem cmparable to ear. .c ME interface problems and that such interface failirgs can have an impact on quality. (Tr. D2955-56. ) The Ibard finds that the ME-Westirghouse inter-face errors reflect an absence of sound design control practice and a violation of Criterion III of Appendix B.

180. We @ prcgram applicable to the non-seismic design work per-formed by Westinghouse for Diablo Canyon was not audited by the NRC until 1973-74. (Tr. D1090, D1092. ) %e majority of the ron-seismic design work performed for Unit I had already been ompleted by that time. (Tr. D1089-92.) Thus, ro NRC audit has assessed the implementation of the Westirghouse m program utilized for the majority of the Unit 1 non-seismic design work.

181. Westirghouse had sme experience in seismic design reanalysis and had developed what it believed to be standard seismic design spectra which would envelop all potential seismic spectra. (Tr. Dll35.) However, it discovered in workirg with the Ibsgri spectra frm Diablo Canyon that sme Hosgri spectra exceeded its standard spectra in sme instances. (Tr.

D1135.) Furthermore, Westirghouse had never before had to utilize tau filtered spectra in any other of its seismic design analyses. (Tr. D1108.)

182. Durirg the period when Westirghouse was perfoming its Ibsgri seismic design reanalysis for Diablo Canyon, the NRC was auditirg the overall Westirghouse @ program. Ibwever, it did rot specifically audit the implementation of this Westirghouse prcgram with respect to the Ibsgri reanalysis. (Tr. D1108. ) In the absence of a Diablo Canyon - specific NRC 61.

e audit of the Westirghouse non-seismic design and Hosgri reanalysis or an ID/P verification of the Westirglouse design product, no meaningful assurance has been provided that the Westirghouse design of safety-related equignent at Diablo Canyon meets all applicable licensirg criteria.

(Hubbard, ff. Tr. D2084, at 24.)

183. Accordirgly, the Ibard finds that the applicant has failed to prove that tne design of the raclear ste.nn supply system (NSSS) for either unit of Diablo Canyon was subject to a legally sufficient GA program.

184. A properly functionirs OA program of a licensee performs assessments of the adequacy of work performed by contractors and suppliers of safety-related systems and equignent. ( Appendix B, Criteria IV, VII.)

In order for the verification program to provide the level of assurance equivalent to that provided by a satisfactory OA program, the Ebard finds that a verification of the Westinghouse design work is required. 'Ihat verification should be of the actual design work applicable to Diablo Canyon and should be of a scope and depth otrnparable to the review a licen-see would be expected to make of its NSSS vendor in order to satisfy its obligations under Appendix B.

VII THE VERIFICATION PROGRAM'S MEWOM ARE INADEQUATE

'IO ASSURE 'IHAT ALL 'IME IOOT CAUSES OF 'INE DIABID CANYON DESIGN ERRORS HAVE BEEN ASCERTAINED AND 'IHAT ALL GENERIC O'NCERNS SrfEMING FROM 'IHOSE ERBORS HAVE BEEN IDENTIFIED A. 'Ihe Verification Program Failed to Identify the Ibot Cause of Each Design Error 185. Ibot cause is the underlyirg basis that precedes and usually includes an effect or result. It is the same as the basic cause described 62.

by the IDVP in section 6.3 of its Final kport (EE hh. 90). Generic

! concern refers to the potential of each identified error to exist in a similar manner in other unreviewed parts of the plant. (Hubbard, ff. Tr.

D2084, at 24-26.)

186. In the case of significant mnditions adverse to quality, Criterion XVI of Appendix B requires the identification of the condition, its cause and corrective action to prevent its recurrence. Interpreting Criterion XVI in the context of the Diablo Canyon application, the Camission order (PG&E Exh. 86, attachment 1, at part 1(a)(5)(b)) and the Staff Intter (PG&E Exh. 87, at parts 1(b), 2(b) and 3(b)) require an assessment of the basic cause of all design errors found.

187. Both the IDVP and the ITP made general statements as to the basic cause of the design errors uncovered. (IDVP Final kport, sections 6.0, 6.3, 6.4 and 6.5, PG&E Exh. 90; PG&E Phase I Final kport section 1.8, PG&E Exh. 91; EE Phase II Final Report section 3.0; PG&E Exh. 92.) However, the basic causes identified related primarily to seismic design errors.

(Hubbard, ff. Tr. D2084, at 27; Tr. D3003.) Pbreover no attempt was made to wrrelate these basic causes with each of the design errors which had been identified. In fact, the IDVP affirmatively states that an iden-tification of the basic cause of each and every design error was unne--

cessary. (IDVP Panel 1, ff. Tr. D1459, at 7-4) . 'Ihus, there is no cbjec-tive evidence that the basic causes identified by the IDVP and the ITP are in fact responsible for each of the design errors mcovered. (Hubbard, ff.

Tr. D2084, at 26-27.)

188. Absent a wrrelation between the basic causes listed and the design errors uncovered it is not possible to meet the wrrective action 63.

1 l

. requirements of Criterion 16 of Appendix B or to cunply with the Ccr:rnission's tbvember 19, 1981 order. (Hubbard, ff. Tr. D2084, at 26-27.)

B. 'Ihe Verification Program Failed to Identify the Ibot Cause of Each @ Error 189. Neither the IDVP nor the ITP sought to identify and isolate the specific @ failure which allowed the subsequently discovered design errors to have gone mdetected. (Hubbard, ff Tr. D2084, at 28-38.) In so doirg they failed to assess the potential generic concern which can result from a failure to establish or implement @ procedures designed to detect such errors. (Hubbard, ff. Tr. D2084, at 28. ) 'Ihe IDVP's rationale for the failure to assess the generic significance of the @ failures to detect pre-November 1981 design errors was that it was asstrned, in the search for other design errors, that the m prcgrams at issue were deficient.

Therefore the search for other design errors would not have been expanded had each @ failiry for each design error been identified. (IDVP Panel 1, ff. Tr. D1459, at 7-3. )

190. While this may mnstitute an acceptable explanation for the IDVP's failure to delineate a @ cause for each undetected error for pre-November 1981 design work, it does not explain why no such deter-mination was made for design errors made by the DCP, inasmuch as the DCP's Ca program was asstrned to meet the requirements of Appendix B (PG&E Panel 5, ff. Tr. D847, at 9, 10) and to have been effectively implemented (IDVP Panel 1, ff. Tr. D1459, at 8-3) .

191. PG&E, the DCP, and the IDVP all evinced the belief that @ audi-tors are not expected per se to find failures to meet design criteria.

Rather their primary function is to see if ergineerirg procedures have been

[

followed. (PG&E Panel 5, ff. Tr. 247, at 7, 8; Tr. D972-975; Tr. D1644.)

64.

l l

l 1

l

, Thus, the verification program did not look at the generic significance of  ;

. 1

@ failures because they did not believe this would enable them to detect I

~

ergineering failures to meet design criteria. Ibwever, the Staff's psi-tion has been that witinut a thorough examination of the extent and imple-mentation of @ program controls, an assessment of generic findirgs in the area of design control consistency will be inhibited. (SECY-82-414, E E Exh.157.) %erefore, this Ibard cannot conclude that aiequate assurance has been provided that all generic design errors have been identified.

C. We Failure to Cbrrelate Ibot Causes and @ Deficiencies Precludes Assurance hat Similar Errors Will be Discovered and Will Not Recur 192. %e consequences of an approach which fails to correlate root causes and m program deficiencies with specific design errors are demonstrated by the aftermath of PG&E's look Back Ieview. Che of the fin-dings of this review states that consultants were "scrnehow" not included as suppliers of nuclear safety related services. tb attempt was made to correlate any of the various causes with each of the contractors in such a way as to pinpint or verify a specific cause for any particular contract.

Furthermore the review refrained frczn even considerirg whether the cause of these errors was based on a lack of management conmitment to m or organi-zational dysfunction. (Tr. D1034-35.)

193. We review listed a number of progranrnatic and @ implementation deficiencies which had been identified and purportedly resolved:

l l a. Lack of formal training requirement in @ for personnel -

resolved in 1977 through issuance of a revised Ergineering Department Manual;

b. Deficiencies in interface control between PG&E and its ser-vice contractors -- resolved in 1977 through issuance of a revised 65.

I

, Engineerirg Department Manual;

c. Failure to assure that up-to-date license aanmitments were beirg utilized by designers - resolved through the project manager informity all design groups to verify ommitments as current prior to use. (Gov. Exh. 34, m. 5-7.)

194. Ch cross-examination Mr. DeUriarte testified that the cause of these deficiencies had been identified and no further problems had been identified with respect to those deficiencies. (Tr. D859-84.) He also testified that as of ?bvember,1981, EE's @ program fully met all its license cammitments. (Tr. D895.)

195. Almost immediately after testifyirg that all interface deficien-cies with service contractors had been resolved by 1977 (Tr. D857-62), Mr.

DeDriarte was forced to admit that, contrary to his testimony and the Icok Back kview conclusions, a deficiency with respect to the ME interface with URS Bltme was disclosed in 1978. (Tr. 0893; (bv. Exhs. 43 and 44.)

196. Itreover, in the Project Assistance Cbrporation (PAC) report (Gov. Exh. 35) on the status of the EE m program, which ME admitted contained no untrue statements (Tr. D3182), and in the EDS Nuclear Inc.

Report (Gov. Exh. 36) on the status of ME @ and Engineerirg Department Manuals, it was demonstrated that, as late as 1982, ME ht:J still failed to resolve deficiencies with respect to trainirg, interface mntrol, and I

the updatirg of license cxanitments and had not in fact yet met all its m license conmitments. (Cbv. Exh. 35, pp. 3-4 ; (bv. Exh. 36, pp. 3, 5-7. )

197. 'Ihus, the aftermath of the ME Icok Back kview demonstrates how a gogram which does not correlate design errors with root causes and @

program deficiencies can fail to prevent the recurrence of these errors in 66.

the future. Inasmuch as the verification prcgram has admittedly adopted such an approach in its review of the root causes and the generic import of the design errors uncovered, it cannot provide assurance that no further design errors will be uncovered in Unit 1 or occur in Unit 2.

D. 'Ihe IDVP Failed to Recognize Organizational Dysfunction and Management Shortccznings as Ibot Causes of the Design Errors 198. In addition to their other findirgs, the PAC mport (Gov. Exh.

35) and the EDS tuclear Report (Gov. hh. 36) denonstrated that RE's Icok Back kview had failed to discover the root causes for the @ program's deficiencies. These root causes, which M E's Icok Back kview failed to consirkr as possibilities, were:

a. Organizational dysfunction caused by multiple almost wrolly independent quality assurance programs within the various nE depart-ments which the corporate @ program is unable to control (Gov. Exh.

35, p. 5; Cbv. Exh. 36, p.1); and

b. 'Ihe low organizational level at which the EE @ Department functioned thereby reflectirg a lack of cxxumitment to @ at the highest levels of management (PAC Mport, (bv. Exh. 35, p. 6; EDS Nuclear k port, Cbv. Exh. 36, p.1) .

i 199. 'Ihese conclusions about the root causes of ME's @ deficiencies l

were echoed in the NRC Staff's assessmant of the root causes of the Diablo Canyon mirror image design errors in its final Case Study C (Staff Exh.

l 54). In that report the Staff identified the fact that ME management was ccmposed of individuals who believed that engineerirg could satisfactorily design a nuclear power plant without an overlayiry @ organization (Staff Exh. 54, p. 6) and wto failed to appreciate the potential merit of a formal institutionalized @ program (Staff Exh. 54, p. 7).

67.

i I

I

\ -. __ .. .

200. Even before the final Case Study C Ieport, the Staff had

. recognized that the management of ME bore a measure of responsibility for

. the breakdown in design quality assurance. (SSER 18, Staff Exh. 36,

p. C.5-6; Tr. D2297.) In reviewing the DCP's description of the contri-buting causes of the design errors - all of which were extrinsic to PG&E and its management (Staff Panel 5, ff. Tr. D30ll, p. 2) - Mr. Schierling adnitted that the DCP assessment did not reflect recognition by PGGE mana-genent of its own past shortcanings. (Tr. D3012.)

201. %ree separate organizations, workirg independently, have con-cluded that a fundamental problem with the PG&E @ program was a lack of high-level management omnmitment to @. %e IINP has failed to identify lack of management ccanitment to @ as a root cause of the @ deficiencies in PG&E's design effort. (IDVP Final Report, PG&E Exh. 90, section 6.0, et seq.)

202. %e Ibard finds that the failure by the IDVP to recognize as a root cause the PG&E lack of managerial commite nt to quality assurance demonstrates that the IDVP's approach to ascertainirg root causes is not adequate. We Ibard further finds that the IINP has not in fact ascer-tained all the root causes of PG&E's design @ failures and the full extent of their generic implications, and has therefore not canplied with the Ocnnission order, or with Criterion XVI of Appendix B.

VIII THE APPLICANT HAS FAILED 'IO PROVE 'IHAT THE RECENI' IESIGN MODIFICATIONS WERE PERFORMED UNDER A CUALITY ASSURANCE PROGRAM WHICH MET 'IHE REQUIREMENTS OF APPENDIX B A. %e PG&E @ Program for November 1981 nrough August 20, 1982, Was Deficient 203. Between tbvember 1981 and Atgust 20, 1982, Diablo Canyon design 68.

work was performed under the NE m Program. (ME Fnnel 5, ff. Tr. W47, at p. 10.) There is no evidence the IDVP verified that the m program being implemented by the DCP at the time was implemented effectively. (Tr.D1639.)

204. During this period DCP engineering worked under and followed the procedures set forth in the ME Ehgineering Department's Dxjineering Manual. (Tr. D3161.) At least 20 to 25 percent of the total design work performed by the DCP was performed at this time. (Tr. D3157, D1024.) In addition approximately 10% of the DCP design work resulting in nodifica-tions was performed during this period. (Tr. D3160. )

205. The Staff testified that until the spring of 1982 the EE m program was still being implemented deficiently and that corrections of trose deficiencies were still ongoing as late as mid-1982. (Tr.D3024-26.)

206. The PAC kport (Gov. Dch. 35, pp. 3-6) found that as of May 1982 the NE @ program still had not been able to fulfill all its m license catrnitments and had severe organizational difficulties which were compounded by the m department's lack of influence within the ME corporate struc-ture. Mr. Haas testified that if the PAC mport conclusions were correct, then PG&E's m program failed to meet the requirements of Appendix B with respect to the items identified. (Tr. D2969-70.)

207. Neither Mr. DeUriarte nor Mr. Skidnore became aware of the con-tents of the PAC mport until this hearing. (Tr. D3155-56, D3198-99. ) hhen confronted with the PAC report conclusions, Mr. DeUriarte testified that they were rot important enough to warrant pronpt attention. (Tr. D3205.)

Mr. DeUriarte's opinion in this regard is consistent with the fact that actions to correct the deficiencies noted in the PAC report are still inccrnplete. (Tr. D3156.) Though PG&E testified that sane of the deficien-69.

1 l

l

, cies in the PG&E @ Manual noted in the PAC mport were cmpensated for in other PG&E Manuals (Tr. D3152), neither the specific portions of these manuals where these deficiencies were overcme nor the manner in which they had been overccme were ever identified. Ebrthermore, on cross-examination, Mr. DeUriarte conceded that m deficiencies caused by missions frm the @

Manual were not the only deficiencies noted by PAC. (Tr. D3227. ) 'Ihus ,

even if all PAC noted missions iere cavered elsewhere, that fact alone would not resolve all the @ deficiencies identified by PAC.

208. Similarly, the EDS Nuclear Inc. NEnrt (Gov. Exh. 36, FP. 5-7) found that the PG&E Ehgineerirg Manual, as of May 1982, was deficient in n merous respects. In rebuttal to the EDS Nuclear Report, EG&E presented the testinony of Mr. akes, a former sployee of EDS Nuclear (Professional Qualifications of William J. Stokes, ff. Tr. D3145, at 2.) 'Ihough Mr.

Stokes had not reviewed any of the manuals for design (Tr. D3189-90) and had not participated in the draftirg of the report (Tr. D3189, 3225), he offered the opinion that the EDS NLx: lear Nport conclusions could not properly be read too broadly. However, on cross-examination, he conceded that the report's conclusions, if limited to the engineerirg rranuals reviewed, were not unwarranted. (Tr. D3188-89. ) He also admitted that the report was written in concert with a qualified ergineer who had reviewed the ergi-neering design manuals at issue. (Tr. D3223-24. ) 'Ihe actual authors of the report were not produced at t% hearirg, 209. As a result, approximately 25% of the total design work performed by the DCP and roughly 10% of its design work resulting in nodifications was performed mder a deficiently conceived and implemented @ program and an engineerity manual likewise inadequate.

70.

, B. PG&E Has Failed to Prove %at the @ Program In Place Since

. Atrjust 20, 1982, Meets Appendix B

. 210. In Atr3ust 1982, the NRC approved the DCP's @ program. Ibwever, the NRC's approval of that program was based on little nore than a review of the Bechtel 'Ibpical Report of matters included within the program and did not include a review of the actual program documents. (Tr. D2977.)

211. %e NRC only reviewed the DCP's cemitment to Appendix B, not its actual compliance with that regulation's requirements. (Tr. D2977-79.)

212. In November 1982, the IDVP condacted an audit of the implemen-tation of the DCP @ Program based in part on checklists. (ITR 41, PG&E Exh.

133.) In that Ibvember audit, the IDVP found, anong auditable aspects of the program, there were 24 failures to meet the requirements of the program.

(IVDP Panel 1, ff. Tr. D1459, at 8-3.) These were attributed to the fact that the @ prcxjram, as of Ibvember 1982 was still just getting underway.

(ITR. 41, PG&E Dch.133, p.11. )

213. At the same time, the IDVP also determined that nore than 50 per-cent of the aspects of the DCP @ Irogram in two of its checklists could not even be audited for compliance with procedures because the DCP had not yet begun to implement these aspects of its program. (Gov. Dchs. 48 and 49.)

214. turing the period when this audit was taking place, the DCP was less than a nonth away fran what has ome to be known as the " big push" in construction of the nodifications at Diablo Canyon. (Tr. D1662-1666.)

Thus, the DCP @ program was still not fully functioning after the design work on a major Inrtion of the seismic design nodifications had taken place.

215. We NRC Staff testified that Criterion II of Appendix B requires that the @ program be in place end functioning before design work is done.

(Tr. D2980.) Under cross-examination, Mr. Ibedy conceded that a strict 71.

Q

. construction of the requirements of Appendix B would indicate that the DCP

@ program had not been implemented in a timely fashion. (Tr. D1699.)

216. In March 1983, the IINP performed a follow-up audit to confirm that the 24 failures to meet program requirements had been met. Ibwever, this audit did not attempt to verify that the program requirements which had previously not been auditable because they had rot yet been implemented had now been implemented satisfactorily. (Tr. D1704-05. ) Indeed, in this March 1983 audit nothing nore than a resolution of the 24 previously identified failures was attempted. (Tr. D1705-06. )

217. 'Ibere is no evidence in the record that any of the attributes of the DCP @ program listed in Governor's Exhibits 48 and 49 described as not auditable or not applicable have ever been satisfactorily implemented.

218. In December 1982, the San Francisco Power Division of Bechtel Corporation performed an audit of the implementation of the DCP's @

program. (Gov. Exh. 33. ) DCP witnesses conceded that this audit had unco-vered failures to cmply with procedures which were serious enough to require corrective action. (Tr. D939-47.)

219. In February and March 1983 the PG&E Ergineerirg Quality Cbntrol Department performed an audit of the DCP's engineerirg manuals and deter-mined that 50 percent of them contained at least 16 incorrect items. (Gtn.

Exh. 42. ) Mr. DeUriarte testified that such a failirg merits concern. (Tr.

D1058.) PG&E's D1gineerirg Quality Cbntrol Manager stated that such a result is not empatible with a " controlled quality program." (Gov.Exh. 40, l

p. 1.) Yet the DCP @ Manager testified that he was not concerned that his own m personnel had not uncovered these deficiencies. (Tr. D961-962.)

220. thtil March 11, 1983, the DCP had not distributed an ergineerirg 72.

l

instruction depictirg the interface matrix for cantractors involved in the design of Unit 1 and by that date the DCP had not yet distributed an engi-neering instruction depicting the interface matrix for contractors involved in the design of Unit ?. (PEI-5; Gav. Exhs. 37, 38, 39; 'Ir. D925-34. )

Inadequacies in deseign interface had been known at least since discovery of the mirror-image error. 'Ihe Baard finds the failure of PG&E to implement formal procedures for design interface is a failure to take timely correc-tive action in empliance with Criterion XVI of Appendix B and demonstrates a continuing failure to PG&E to appreciate its @ obligations.

221. Criterion XVIII requires that cmprehensive audits be undertaken to verify empliance with all aspects of a quality assurance prcgram. Che aspect of a quality assurance prcgram involves verifying or checking the adequacy of design. In the DCP, the individuals responsible for verifyirs or checkirg the adequacy of design doctznents do not perform audits. (Tr.

D976-78.) Instead, audits are performed by @ Enpart:nent personnel.

However, @ Department personnel are not required or expected to verify or check, by virtue of their audits, whether design documents meet the design criteria for them. (Tr. D972-75.) 'Ihe Baard finds that the DCP conducts to atr3its to verify whether design documents meet all design criteria and thus fails to meet the requirements of Criterion XVIII.

CONCLUSIONS OF IAW

1. 'Ibe applicant has failed to satisfy the requirements of section 50.57(a)1, in that:

a. It has failed to establish the quality of the design of Diablo Canyon Unit 1 and Unit 2 by a verification program that either establishes the design to fully meet the licensing criteria and regula-73.

tory requirements or that subjects the design to a level of scrutiny equivalent to that of a lawfully adequate CA program;

b. It has not met its FSAR comitment to perform a jet impingement analysis on all piping inside contairunent which is subject to pressures over 275 psig or temperatures over 200 degrees fahrenheit;

c. It has not demonstrated that its configuration control prac-tices conform to the requirements of Criteria III, VI, X, and XVI of Appendix B to 10 C.F.R. Part 50;

d. It has not demonstrated that in procuring the NSSS for Diablo Canyon it met the requirements of Criteria IV, VII, and XVIII of Appendix B to 10 C.F.R. Part 50;

e. It has not demonstrated that in attemptirg to ascertain the root cause and generic implications of the design errors mcovered it has conformed its conduct to the provisions of Criterion XVI of Appendix B to 10 C.F.R. Part 50;

f. It has not denonstrated that the DCP CA program under which the recent design nodifications to Diablo Canyon were conceived and imple-mented was in empliance with the requirements of Appendix B to 10 C.F.R. Part 50.

2. 'Ihe applicant has failed to emply with General Design Criterion II and with applicable licensirg criteria in that the seismic design of all structures at Diablo Canyon has not been shown to be qualified for all effects of the applicable earthquakes, and the seismic response of all structures has not been shown to be within allowable limitations specified in the licensing criteria.

3. 'Ibe IINP has failed to cmply with the cmmission order in the 74.

followirg respects:

a. It failed to verify a sample of the design work of each service-related contractor;

b. It failed to obtain and use the expertise of a qualified statistician;

c. It failed to identify all deviations in the design of thit 1 frun the licensiry criteria and regulatory standards and, instead, took it upon itself to recast the objective of the IDVP by limiting its objectives to the identification of all violations of the licensing criteria and regulatory standards the IDVP deemed to be safety-significant;

d. It failed to identify all the root causes of the design errors uncovered at Diablo Canyon.

I i

75.

Appendix A ISSUES LITIGATED IN THE REOPENED DESIGN QUALITY ASSURANCE HEARINGS Upon the motions of the Governor and the joint intervenors and in acordance with the concessions of the applicant and the staf f, the record in the Diablo Canyon operating license proce1 ding was reopened on the issue of design quality assurance.- The primary focus of the reopened proceeding, howeve r , has now moved beyond the question of whether the applicant can demonstrate that the IDVP and the ITP verify the correctness of the Diablo Canyon design. The licensing criteria contained in the PSAR, FSAR, Hosgri Report, SER, SER Supplenments and other licensing documents are outside the scope of the issues in this reopened proceeding and are not subject to challenge, except to the extent that such criteria may have been modified in the verification program. The Governor and the joint intervenors assert that the verification programs have failed in the following respects:

1. The scope of the IDVP review of both the seismic and non-seismic aspects of the designs of safety-related systems, structures and components (SS&C's) was too narrow in the following aspects:

(a) The IDVP did not verify samples from each design activity (seismic and non-seismic).

i (b) In the design activities the IDVP did review, it l did not verify sanples from each of the design groups in the L design chain performing the design activity.

(c) The IDVP did not have statistically valid samples from which to draw conclusions.

l (d) The IDVP failed to verify independently the analyses but merely checked data of inputs to models used by PG&E.

(e) The IDVP failed to verify the design of Unit 2.

2. The scope of the ITP review of both the seismic and non-seismic aspects of the designs of the safety-related systems, structures and components (SS&C's) was too narrow in the following aspects:

(a) The ITP did not verify samples from each design activity (seismic and non-seismic).

(b) In the design activities the ITP did review, it did not verify samples from each of the design groups in the design chain performing the design activity.

l A-1

^

l (c) The ITP did not have statistically valid samples from which to draw conclusions.

(d) The ITP has failed systematically to verify the adequacy of the design of Unit 2.

3. In various situations listed below, the ITP used improper engineering standards to determine whether design activities met license criteria. In some of these situations, t

the IDVP either used or approved the use of such improper i standards or did not verify them at all. '

(f) The ITP's modeling of the soil properties for the containment and auxiliary buildings was improper in that:

(i) in the soil structure interaction analysis of containment for the DE and the DDE, use of boundary motion inputs to the model were improperly used; (ii) the soil structure interaction analysis for containment for the DE and the DDE uses a 7 percent damping value for rock, which is unconservative, -

especially for the DE; (iii) the dynamic analyses of the containment for all earthquakes omit any analysis of uplifting of the foundation mat;

,(iv) the modeling of the soil springs for the auxiliary building does not specify soil properties; (v) in the modeling of the soil springs for the auxiliary building , the motion inputs to the lower ends of the springs does nat account for all soil structure interaction phenomena that could be expected.

(o) The ITP has not demonstrated, and the IDVP has not l

l verified, that the DCP modeling of the seismic response of the fuel handling building is proper, in that the DCP has not adequately justified the use of the translational and torsional response of the auxiliary building as input to the fuel handling building nor has it demonstrated the validity of the dynamic degrees of freedom selected.

(p) The ITP has not demonstrated, and the IDVP has not l verified, that the DCP seismic model of the slabs in the l auxiliary building is proper, in relation to the use of vertical and rotational springs to model the columns, and the motions used as input at the ends of the springs not connected to the slabs. In addition, in the study of the diaphragms, the ITP has not adequately accounted for the A-2 l

l l

l t

l l

l l

inplane flexibility of these slabs, and has not adequately demonstrated that stresses are within allowable limits at all elevations.

(q) The ITP has not demonstrated and the IDVP has not verified, that the soils analysis for the buried diesel fuel oil tanks is proper in that the values of the exponent shown in figure 14 of ITR 68 have not been demonstrated to be appropriate and the variation of shear velocity with depth is not properly justified.

(r) The ITP has not demonstrated and the IDVP has not verified that the soils analysis for the auxiliary saltwater piping and circulating water intake conduits is proper in that.the selection of the modulus versus strain-curve utilized is not justified.

(s) The ITP has not demonstrated and the IDVP has not verified that the seismic analysis of the turbine building is proper in that bolt bearing capacities were taken from an inappropriate source.

(t) The ITP has not demonstrated and the IDVP has not '

verified that the seismic analysis of the turbine building is proper in that the use of four different models for the vertical analysis has not been justified.

4. The IDVP accepted deviations from the licensing criteria without providing adequate engineering justification in the following respects:

a. Contrary to the requirements of FSAR Section 17.1 regarding compliance of the as-built installation with the design documents, the IDVP review of the AFWS disclosed that the as-built installation failed to meet the design drawings in that (i) a steam trap on the turbine-driven AFW pump steam

supply line is not provided and (ii) there are discrepancies in the arrangement of the long-term cooling water supply line.

b. Contrary to FSAR Section 8.3.3, the electrical design does not fully comply with the commitments regarding separation and color coding.

h. Contrary to PG&E's September 14 and December 28, 1978, licensing commitments, CRVPS equipment identified in the FSAR as necessary to maintain control room habitability during safe shutdown has not been evaluated regarding the effects of a moderate energy pipe break.

A-3 l

L - . - . __ - - _ _ _ . , _ . - _ _ _ _ _ . , - _ _ _ _ _ _ _ . _ . . .. _ _ __. _ _ _ - - . - - . . - -

1. The fire protection for the motor driven AFW pump room is not consistent with the PG&E licensing commi tment for fire zone separation as stated in its November 13, 1978, Supplemental Information for

' Fire Protection Review ("SIFPR") in that:

1) there is a large grated ventilation opening in the ceiling of the room;

2) a fire damper has gaps when it is closed.

j. The fire protection for the AFW pump room is not consistent with the PG&E licensing commitment for cable separation as stated in its SIFPR of November 13, 1978, in that

1) the pumps for the motor driven AFW pumps and the control circuitry for a flow control valve necessary for operation of the turbine driven i AFW pump are. located in a single fire zone; i

l

2) cables for some AFW circuits are not routed in accord with descriptions in the SIFPR and four AFW circuits PG&E committed to identify and

+

review in the SIFPR were not included in that document.

, k. Contrary to the licensing commitment set forth in its SIFPR of November 13, 1978, each of the three 4160 volt cable spreading rooms has a ventilation opening leading up to the 4160 volt switchgear

rooms.

1. Contrary to FSAR Section 3.6, possible jet impingement loads have not been considered in the design and qualification of safety-related piping and equipment inside containment.

q. Contrary to PG&E's December 28, 1979, licensing commitment letter to the NRC, modifications to l protect two Auxiliary Feedwater valves from the effects of moderate energy line breaks were not implemented.

r. Contrary to the licensing commitment to maintain i minimum system redundancy as stated in FSAR Section 3.6A (NSC evaluation of pipe break outside containment), four components were identified for which high energy line cracks could cause temperatures in excess of the specification temperatures of the components.

A-4

V

s. Contrary to the licensing commitment to maintain minimum system redundancy as stated in FSAR Section 3.6A (NSC evaluation of pipe break outside containment), a conduit was identified whose f ailure due to a high energy line crack could

. eliminate redundant Auxiliary Feedwater system flow.

t. Contrary to the FSAR Section 8.3 commitment to provide switchgear buses with adequate short circuit interrupting capability, the calculated duties for circuit breakers on 4160 V buses F, G, and H were above the nameplate ratings for those buses.

u. Contrary to single failure criteria stated in FSAR

' Section 3.1.1, reviews of the Auxiliary Feedwater and Control Room Ventilation and Pressurization

.s systems identified circuit separation and single I

failure deficiencies. Similar deficiencies were identified in additional verification reviews, s which included other safety-related systems.

5. The verification program has not verified that Diablo Canyon Units 1 and 2 "as ! Jilt" conform to the design drawings and analyses.

6. The verification program failed to verify that the design of safety-related equipment supplied to PG&E by Westinghouse met licensing criteria.

7. The verification program failed to identify the root

, causes for the failures in the PG&E design quality assorance l program and failed to determine if such failures raise generic Concerns.

8. The ITP failed to develop and implement in a timely manner a design quality assurance program in accordance with 10 C.F.R. Part 50, Appendix B to assure the quality of the recent design modifications to the Diablo Canyon facility and the IDVP failed to ensure that the corrective and preventative action programs implemented by the ITP are suf ficient to assure that the Diablo Canyon facilities will meet licensing criteria.

9. Contrary to General Design Criteria 44 (GDC-44) of Appendix A to 10 C.F.R. Part 50, PG&E has failed to provide adequate assurance of component cooling water system (CCWS) heat removal safety function capacity in that the maximum ocean water temperative of 64 degrees F. *

. not conservative because it has already been exceeded in l' ? ? Furthermore, a technical specification limitati in <:d - permits plant operation at reduced A-5 I

power levels in lieu of enlarging the capacity of the CCWS does

, not provide an equivalent level of safety as compliance with the requirements of GDC-44 (SSER 16 (Aug. 1983) and September 1983 ocean water temperature readings) .

h l

l l

A-6 L

Appendix B WI'ITES!ES 'IESTIFYING IN 'IEE REOPENED DESI(N QUALITY ASSURANG HEARING PANEL DIIECT 'IESTIMNY NLPBER WTINESSES BEGINS FMLOWING ISSLES APPLICANT'S WI'ITES!ES PG&E 1 G. Cranston, G. Moore, D224 1, 2, 5 R. Arderson, W. White, L. Shipley PG&E 2 H. Seed , R. Arderson, D551 3 L. Shipley, W. White PG&E 3 E. Ccnnell, G. Moore, D487 4, 9 R. Arderscn, W. Yahlstrcm PG&E 4 R. Wiesemann, J. Hoch, D1088 6 PG&E 5 C. Dick, M. Jacobscn, D847 7, 8 S. Skicinore, T. DeGriarte, G. Moore PG&E 6 S. Kaplan, R. Arderscr; Dll60 1(c) & 2(c)

! ID/P 1 W. Cooper, R. Cloud, D1458 1, 2, 5-8 l J. Krechting, R. Reedy IDIP 2 R. Wray, R. Cloud, D1843 3 J. Biggs, M. Holley ID/P 3 J. Kre:hting, W. Cooper D2)39 4 (D/ERNCR'S WI'INESSES ,

R. Hubbard D2084 5-8 J. Roesset D2206 3 G. Apostolakis D2313 1,2,7 JOINT INTERVENTS' WI'INESSES F. Samaniego D2392 1 B-1 l

~

PANEL DIRECT TESTI!ONY

. NIMER WINS!ES BEGNS POLIIMNG ISSIES SIRFF WTINESSES Staff 1 J. Knight, H. Schierling, D2649 1, 2 J. Wermiel Staff 2 J. Knight, P. Kuo, H. Polk D2463 3 C. Miller, C. Costantino, A. Philipproleulos, P. Wang Staff 3 J. Wermiel, J. liv. t D2864 4 Staff 4 J. Knight D2864 4 Staff 5 P. Nrrill, W. Haass, D2864 4 J. Knight, H. Schierling W. Altan Staff 6 J. Wermiel D2864 9 APPLICANT'S REBUITAL WIMSSES Rebuttal 1 R. Cloud, H. Seed , W. White 3 Rebuttal 2 G. Moore, S. Skidrrore, 8 T. DeUriarte, M. Jacobson, L. Gouveia, W. Stokes B-2

Al pendix C EXHIBITS AIMITIED IN 'INE REXPEtED

, MSIm GPLITY ASSURANCE HEARINGS NIMER IESCRIPTICN OETERED AINITIED GOVERNGl'S EXHIBITS Gw. 11 2/12/82 L: Wilkinson (INPO) Mielke D-2085 D-2087 Gov. 12 11/2/82 BNL Review of ITR-ll, Rev. O D-2085 D-2087 Gov. 13 10/12/81 L: Crane-Engelken transmitting LER D- 2085 D-2087 cn annulus error.

Gov. 14 11/6/81 NRC Regicn V Inspecticn D-2085 D-2087 Repor t 50-275/81-29 & 50-323/81-18 Gov. 15 10/11/82 Potential Program Resoluticn Repcrt D-2085 D-2087 EOI 7002, Rev. O (Jet Impingement)

Gov. 16 7/26/83 Open Item Repcrt - EOI 7002, Rev. 5 D-2085 D-2087 (Jet Impingement)

Gov. 17 9/13/82 Open Item Repcrt - EOI 8010 D-2085 D-2087 Rev. 0 (valve added, could result in excess pressure)

Gov. 18 3/4/83 Error Repcrt - EOI 8010, Rev. 8 D-2085 D-2087 Gov. 19 6/1/83 Program Resolution Report D-2005 D-2087 EOI 8010, Rev. 11 Gov. 20 10/4/82 Open Item Repcrt - D-2085 D-2087 EOI 8017, Rev. O (Electrical

' Cmtrol Transfer switch -

separaticn criteria)

Gov. 21 10/12/82 Open Item Repa t - D-2085 D-2087 EOI 8022, Rev. O (circuit breakers Gov. 22 3/10/83 Error Repcrt - EOI 8022, D-2085 D-2087 Rev. 5 (circuit breakers)

Gov. 23 10/12/82 Open Item Repcrt - D-2085 D-2087 Rev. O (Loca Corditions -

480 Volt Systens)

Gov. 24 3/11/83 Program Resoluticn Repcrt D-2085 D-2087 EOI 8023, Rev. 5 (Loca Corditions-480 Volt Systens)

C-1

NLPEER IESCRIPTION GTERED ADE'ITED G w . 25 10/29/82 Open Item Report - D-2085 D-2W7 EOI 8060, Rev. O (Pump Pressure Control System)

Gov. 26 3/15/83 Error Repcrt - EOI 8060, Rev. 5 D-2085 D-2087 Gov. 32 2/82 Classificaticn of SS&Cs for Diablo D-1447 D-1447 Gov. 33 1/12/83 M:Kellermann - Moore: D-936 D-936 Audit 317 Package Gov. 34 11/2/82 M:DeUriarte - Raymcrx3 D-857 D-857 trarsmittirg. Look Back Review Sunmary of PG&E Design Activities G w . 35 5/82 Review of the PG&E QA Manual D-902 D-902 for Ntz: lear Pcwer Plants: Sunmary Report Gov. 36 6/7/82 L:Polivka - Dodd transmits D-996 D-996 Project Sunmary Report fa PG&E by EDS Nuclear, Irc.

Gov. 37 8/10/82 DCP Instructial No. 5, Rev. O D-934 D-934

- Design Interf aces Gov. 38 10/29/82 DCP Instructial No. 5, D-934 D-934 Rev.1 - Design Interf aces Gov. 39 3/11/83 DCP Instrtx:tial No. 5, D-934 D-934 Rev. 2, Design Interf aces Gov. 40 4/21/83 M:Ralston-Brand transmits D-954 D-954 Audit EQ-8303 on Engineerirg Dept. Manual Ccntrol Gov. 41 Graph - Variaticn of Shear McIlulus D-78 2 D-782 with Shear Strain Gov. 42 4/8/83 Table - Engineering Manual D 957 D-957 Audit Gov. 43 4/26/79 Naiccnfctmance Report D-894 D-895 G w. 44 5/8/79 Nonccnformarce Rep 3rt - D-894 D-895 Design Interfa:e Gov. 45 7/14/80 Article by Kaplan & D-1336 D-1446 Garrick "On the Quantitative Definiticn of Risk" C-2

. NueER IESGIPTION GTERED AD4ITTED

. Gw. 46 6/11/81 Article by Kaplan D-1346 D-1446

" Safety Gels ard Related Questicns" Gov. 47 10/15/82 L: Cooper - Eisenhut in D-1463 D-1464 Respcnse to Regicn V Letter ("All")

G w. 48 11/11-22/82 - R.F. Reedy Checklist D-1629 D-1630 Audit No. 014-003-A-03-6 Gov. 49 11/11-22 /82 - R.F. Reedy Checklist D-1629 D-1630 Audit No. 014-003-A-03-5 Gov. 50 12/1/82 - Audit Notes (By Norris), D-1654 D-1657 Audit No. 014-003-A03-7 Gov. 51 3/15/83 - L:Krechting - Cooper D-1564 D-1586 tensmits draft of ITR 14

(" Bullshit Ccanents")

Gov. 52 Draf t of ITR 14 (Re: Representative samples) D-1606 D-1607 Gw. 53 5/2/83 - Page 3.5-8 of ID/P D-1553 D-1554 Fimi Report. Rev. O D-1554 Gw.54 6/27/83 - Page 3.5-8 of ID/P D-1553 D-1554 Final Report. Rev. I D-1554 Gov. 55 Safety & Evaluation Report: D-2578 D-2634 Review of Diablo Canytn Turbine building Gov. 60 9/1/83 M:Rubinstein - Ranney - D-2602 D-2770 Recticns cn PG&E Evaluaticn Spot Welds C-3

NCDBER MSCRIPTION GFERED ADE'ITED PG&E EXHIBI'IS PG&E 86 11/19/81 Ccamissicn Order (CLI-81-30 D-247 D-251 PG&E 87 11/19/8 1 NRC Staff Letter I>-247 D-251 PG&E 88 3/29/82 Design Verificaticn D-247 D-251 Program Management Plan, Phase I PG&E 89 6/18/82 Design Verificaticn D-247 D-251 Program Management Plan, Phase II PG&E 90 10/10/83 IDVP Fimi Report D-247 D-251 PG&E 91 10/14/83 PG&E Phase I Final Report D-247 D-251 PG&E 92 10/11/83 PG&E Phase II Fimi Report D-247 D-251 PG&E 93 ITR 1 (Rev. 1) 10/22/82: Additional D-247 D-251 Verificaticn and Additicnal Sampling Effective 10/15/82 PG&E 94 ITR 2 (Rev. 0) 6/23/82: Evaluaticn of the D-247 D-251 Quality Assumnce Program and Implenentaticn Reviews PG&E 95 ITR 3 (Rev. 0) 7/16/82: Tanks D-247 D-251 PG&E - 9 6, ITR 4 (Rev. 0) 7/23/82: Shake Table Testing D-247 D-251 PG&E 97 ITR 5 (Rev. 0) 8/19/82: Design Chain D-247 D-251 ,

PG&E 98 ITR 6 (Rev. 0) 9/10/82: Auxiliary Building D-247 D-251 PG&E 99 ITR 7 (Rev. 0) 9/17/82: Ele:tr ical D-247 D-251 Raceway Stgarts PG&E 100 ITR 8 (Rev. 0) 10/ 5/82: Verificaticn Program D-247 D-251 for PG&E Corrective Acticri PG&E 101 ITR 9 (Rev. 0) 10/15/82: Developnent of the D-247 D-251 Safety-Related Contractor List for Ncn-Geismic Design Wcrk l

Performed Prior to 6/1/78 PG&E 102 ITR 10 (Rev. 0) 10/29/82: Verificaticn of D-247 D-251 Design Analysis Hosgri Spectra C-4

toeER DESCRIPTION GTERED ADiI'ITED PG&E 103 ITR 11 (Rev. 0) 11/ 2/82: PG&E - D-247 D-251 Westinghouse Seismic Interface Review PG&E 104 ITR 12 (Rev. 0) 11/ 5/82: Piping D-247 D-251 PG&E 105 ITR 13 (Rev. 0) 11/ 5/82: Soils-Intake D-247 D-251 S tructure PG&E 106 ITR 14 (Rev. 2) 7/25/83: Verificaticn of D-247 D-251 the Pressure, Teperature, Humidity, aM Submergence Environments Used for Safety-Related Equipnent Specificaticn Outside Ccntaiment for Auxiliary Feedwater System and Ccntrol Room Vmtilaticn ard Pressurization Systs PG&E 107 ITR 15 (Rev. 0) 12/10/82: IWAC Dtet and D-247 D-251 Duct Stpports PG&E 108 ITR 16 (Rev. 0) 12/ 8/82: Soils-Outdoor D-247 D-251 Water System Storage Tanks PG&E 109 ITR 17 (Rev. 0) 12/14/82: Piping: D-247 D-251 Additional Sample PG&E 110 ITR 18 (Rev.1) 5/24/83: Verificaticn of D-247 D-251 the Fire Protection.Provided for Auxiliary Feedwater System, Ccntrol Room Ventilation ard Pressurization System, Safety-Related Pcrticn of the 4160 V Electrical System PG&E 111 ITR 19 (Rev. 0) 12/16/82: Verificaticn of D-247 D-251 the Post-IXXA Porticn of the Radiaticn Environments Used fcr Safety-Related Equipnent Specification Outside Cantalment for Auxiliary Feedwater System ard Control Rom Ventilaticn and Pressurizaticn System PG&E 112 ITR 20 (Rev. 2) 7/25/83: Verificaticn of D-247 D-251 the Mechanical / Nuclear Design of the Ccntrol Room Vmtilaticn ard Pressurizaticn Syste PG&E 113 ITR 21 (Rev.1) 5/ 3/83: Verificaticn of D-247 D-251 the Effects of High Energy Line Cracks and Moderata Energy Lim Breaks for Auxiliary Feedwater System ard Control Room Ventilaticn ad Pressurizaticn System C-5 I

l

, NGEER DESCRIPTION GTERED AINI'ITED PG&E 114 ITR 22 (Rev. 2) 7/25/83: Verification of D-247 D-251 the 14echanical/ Nuclear Pcrticn of the Auxiliary Feedeter System PG&E 115 ITR 23 (Rev.1) 5/27/83: Verificaticn of D-247 D-251 High Energy Line Break ard Internally Generated Missile Review Outside Contairinent for Auxiliary Feedwater System and Ccntrol Room Ventilaticn and Pressurization System PG&E 116 ITR 24 (Rev.1) 5/ 4/83: Verificaticn of D-247 D-251 the 4160 V Safety-Related Electrical Distributicn System PG&E 117 ITR 25 (Rev.1) 4/29/83: Verificaticn of D-247 D-251 the Auxiliary Feedwater Systen Electrical Design PG&E 118 ITR 26 (Rev.1) 5/ 2/83: Verificaticn of D-247 D-251 the Control Roczn Ventilaticn ard Pressurizaticn System Electrical Design PG&E 119 ITR 27 (Rev. 2) 7/25/83: Verificaticn of D-247 D-251 the Instrtrnent ard Control Design of the Auxiliary Feedwater System PG&E 120 ITR 28 (Rev. 2) 7/25/83: Verificaticn of D-247 D-251 the Instrument an3 Ccntrol Design of the Control Rocyn Ventilaticn an3 Pressurizaticn System PG&E 121 ITR 29 (Rev. 0) 1/20/83: Design Chain - D-247 D-251 Initial Sample PG&E 122 ITR 30 (Rev. 0) 1/12/83: Small Bore D-247 D-251 Piping Rep 3rt PG&E 123 ITR 31 (Rev.1) 8/ 4/83: HVAC Ccrnpcnents D-247 D-251 PG&E 124 ITR 32 (Rev.1) 4/1/83: Ptunps D-247 D-251 PG&E 125 ITR 33 (Rev.1) 4/28/83: Electrical D-247 D-251 Equipnent Analysis PG&E 126 ITR 34 (Rev.1) 3/24/83: Verificaticn of D-247 D-251 Diablo Canyon Project Efforts by SEC PG&E 127 ITR 35 (Rev. 0) 4/1/83: IDVP Verificaticn D-247 D-251 Plan for Diablo Canyon Project Activities C-6

. . j

l NGEER EESGIPTION CFFERED AD4I'ITED PG&E 128 ITR 36 (Rev.1) 6/20/83: Final Reprt on D-247 D-251 Cmstructicn OA Evaluaticn of Guy F. Atkinson Ccznpany PG&E 129 ITR 37 (Rev. 0) 2/23/83: Valves D-247 D-251 PG&E 130 ITR 38 (Rev. 2) 6/20/83: Final Reprt m D-247 D-251 Ccnstructicn Quality Assurance

! Evaluaticn of Wismer ard Becker PG&E 131 ITR 39 (Rev. 0) 2/25/83: Soils-Intake D-247 D-251 Structure Bearing Capacity ard Lateral Earth Pressure FG&E 132 ITR 40 (Rev. 0) 3/ 9/83: Soils-Intake D-247 D-251 Structure Sliding Resistance PG&E 133 ITR 41 (Rev. 0) 4/19/83: Corrective Acticn D-247 D-251 Program ard Design Office Verificaticn PG&E 134 ITR 42 (Rev. 0) 4/13/83: IFR IINP Phase II D-247 D-251 Review ard Audit of PG&E Co. ad Design Ccnsultants for Diablo Canycn Unit 1 PG&E 135 ITR 43 (Rev. 0) 4/14/83: Heat Exchangers D-247 D-251 PG&E 136 ITR 44 (Rev. 0) 4/15/83: Shake Table Test D-247 D-251 Mounting of Class IE Electrical Equipnent PG&E 137 ITR 45 (Rev. 0) 5/17/83: Additicnal D-247 D-251 Verificaticn of Redurdancy of Equignent and Poer Supplies in Shared Safety-Related Systems PG&E 138 ITR 46 (Rev. 0) 6/27/83: Additicnal D-247 D-251 Verificaticn of Selecticn of System Design Pressure aM Differential Pressure Across Power-Operated Valves PG&E 139 ITR 47 (Rev. 0) 6/27/83: Additicnal D-247 D-251 Verificaticn of Envircrinental Ccnsequences of Postulated Pipe Rtptures Outside of Contairinent PG&E 140 ITR 48 (Rev. 0) 7/27/83: Additicnal D-247 D-251 Verificaticn of Jet Impingement Effects of Postulated Pipe Rtptures Inside Containment C-7

_ - - - - - - - - - - - - - I

NLDEER MSCRIPTION GFERED AINITI'ED PG&E 141 ITR 49 (Rev. 0) 6/23/83: Additional D-247 D-251 Verificatial of Ciralit Separaticn ard Single Failure Review of Safety-Related Electrical Equipnent PG&E 142 ITR 50 (Rev. 0) 7/22/83: Cmtairnent D-247 D-251 Annulus Structure Vertical Seismic Verification PG&E 143 ITR 51 (Rev.1) 9/21/83: Cmtainnent D-247 D-251 Annulus Structure - Verificaticn of Corrective Acticn PG&E 144 ITR 54 (Rev.1) 10/ 3 '83: Cmtainnent D-247 D-251 Buildirg PG&E 145 ITR 55 (Rev.1) 10/1/83: Auxiliary Building D-247 D-251 PG&E 146 ITR 56 (Rev.1) 9/24/83: Turbine Building D-247 D-251 PG&E 147 ITR 57 (Rev.1) 9/ 8/83: Fuel Handling D-247 D-251 Building PG&E 148 ITR 58 (Rev. 1) 10/ 1/83: Intake Structure D-247 D-251 PG&E 149 ITR 59 (Rev.1) 9/24/83: Large Bore Piping D-247 D-251 PG&E 150 ITR 60 (Rev.1) 10/ 3/83: Large and Small D-247 D-251 Bore Pipe Stpprts PG&E 151 ITR 61 (Rev.1) 10/ 2/83: Small Bore Piping D-247 D-251 PG&E 152 ITR 63 (Rev.1) 10/ 2/83: HVAC Ducts, D-247 D-251 Electrical Raceways, Instrument Ttbing ard Associates Stpports PG&E 153 ITR 65 (Rev.1) 10/11/83: Rtpture Restraints D-247 D-251 PG&E 154 ITR 67 (Rev.1) 9/ 9/83: Equiptent D-247 D-251 PG&E 155 ITR 6B (Rev. 1) 10/ 4/83: Verificaticn of D-247 D-251 HEA Soils Work Verificatim of HEA Soils Work PG&E 156 3/1/82 M:Dircks - Ccanissiaters Secy-82-89, D-251 D-254 Verification Program Phase II Recamen3ations PG&E 157 10/13/82 M:Dircks - Ccmnissicners D-251 D-254 Secy-82-414, Verification Progran Phase II Reconnerdaticns C-8

NEDBER MSCRIPTION GTERED AIMPITED PG&E 158 3/8/82 M:Chilk - Dircks, Commission Approval D-251 D-254 of Phase I PG&E 159 12/9/82 M:Chilk - Dircks, Cmmissian Approval D-251 D-254 of Phase II PG&E 160 8/5/83 DCP Project Engineer's Instructici D-251 D-254 No.13, Rev.1 - Unit 2 Review Procecitre PG&E 161 8/15/83 PG&E Engireering Dept. Proce3ures D-251 D-254 3.6 CN an3 3.7 "As-Builts" PG&E 162 2/18/76 Hearings before Joint Cmmittee cn Atorric Energy PG&E 163 5/17/83 IINP Program Procedure Interfa:e Bet.

ID/P Par ticipants, DCP & Designated Other Par ties C-9

NGBER MSCRIPTION GTERED ADE'ITED

, JOINI' INIERVEN@S' EXHIBITS JI 128 7/83 !EC "@ Case Study Working Paper, D-3031 D-3031 Case C," Draft Working Paper JI 129 Reedy "@ Review and Audit Report, Phase I; D-1000 D-1000 Safety-Related Activities Performed by PG&E Prior to 6M8 JI 130 7/82 BNL "In3eperx3ent Seismic Evaluaticn D-2612 D-2612 of the Diablo Contairunent Annulus Structure JI 131 7/1/82 L:Denton-Cooper Re: BNL Firr31ngs D-2613 D-2613 cn discreparcies in seismic design JI 132 5/17/83 L:Reicti-Kto, BNL Firdings Re: D-2614 D-2614 Analysis of Horizontal Response JI 133 12/82 ITR 34, Draft D-1793 D-179 5 JI 134 5/16/83 Design Criteria Menorandtzn M-65, Rev. O JI 135 2/74 3.6 Criteria fcr Protectim Against Dynamic Effects Associated with Postulated Rtpture of Piping JI 136 12/18 /72 L:Giambusso-Searls Re: Effects D-651 D-651 of a Piping System Break Outside Containment JI 137 Suninary and Evaluaticn Report ITR 9 D-2616 JI 138 11/23/82 L" Cloud-Hmh D-17 59 D-1760 JI 139 9/21/83 L:Maneatis-Eisenhut Transmitting D- D-Cmments on Draft of Case Study C JI 140 Changes from Draft to Final Case Study C - 10

NGEER MSCRIPTION GFERED AINITTED STAFF EXHIBITS Staff 36 ER Stpplanent 18 D-2465 D-2467

' Staff 37 ER SLpplement 19 D-2465 D-2467 Staff C 11/15/82 L: Bishop-Crane transmits NRC D-2465 D-2467 Impection Reprt 50-275/82-36 Staff 41 11/22/82 L: Bishop-Crane transmits NRC D-2465 D-2467 Inspection Report 50-275/82-36 Staff 42 1/10/83 L: Bishop-Crane transmits NRC D-2465 4 2467 Inspection Report 50-275/82-41 &

50-275/82-19 Staff 43 1/19/83 L: Bishop-Crane transmits NRC D-2465 D-2467 Impecticn Report 50-275/82-42 Staff 44 5/11/83 L: Bishop-Crane transmits NRC D-2465 D-2467 Inspectico Report 50-275/83-14 &

50-323/82-19 Staff 45 8/18/83 L: Bishop-Schuyler transmits NRC D-2465 D-2467 Inspection Reprt 50-275/83-27 &

50-323/83-19 Staff 46 8/3/83 L: Bishop-Schuyler transmits NRC D-2465 D-2467 Inspection Report 50-275/83-26 Staff 47 2/3/82 M: Herring-Miraglia NRC Trip Report D-2465 D-2467 Staff 48 3/3/82 M: Herring-Miraglia NRC Trip Reprt D-2465 D-2467 Staff 49 3/31/82 L:Stemberg-Crane Regicn V Inspecticn D-2465 D-2467 Reprt 50-275/82-02 & 50-323/82-02 Staff 50 5/27/82 L: Bishop-Crane Regicn V Inspecticn D-2465 D-2467 Report 50-275/82-17 Staff 51 6/29/82 L:Stemberg-Crane Regica V Inspecticn D-2465 D-2467 Reprt 50-275/82-20 & 50-323/82-10 Staff 52 10/14/82 L: Bishop-Crane Regicn V Inspecticn D-2465 D-2467 Reprt 50-275/82-30 & 50-323/82-14 Staff 53 2/11/83 L: Bishop-Crane Regicn V Inspecticn D-2465 D-2467 Reprt 50-275/82-30 & 50-323/82-14 Staff 54 9/19/83 M:Eisenhat-Ccumissicners. Board D-2910 D-2910 Notification 83-135A: Case Study C Staff 55 ER Supplement 16 D-2465 D-2467 C - 11

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING APPEAL BOARD

)

In the Matter of )

)

PACIFIC GAS AND ELECTRIC COMPANY ) Docket Nos. 50-275 0.L.

(Diablo Canyon Nuclear Power

) 50 -32 3 0.L.

)

Plant, Units 1 and 2) )

)

CERTIFICATE OF SERVICE I hereby certify that on this date I caused copies of the foregoing Proposed Findings of Fact and Conclusions of Law of Governor Deukmejian served on the following by U.S. Mail, first class (except for those persons marked with an asterisk

("*"),

to whom the envelope was posted Express Mail), postage prepaid.

Hon. Nunzio Palladino, Chairman U.S. Nuclear Regulatory Commission 1717 H Stree t, N.W.

Washing bon, D.C. 20555 Hon. Victor Gilinsky, Commissioner U.S. Nuclear Regulatory Commission 1717 H Street, N.W.

Washington, D.C. 20555 Hon. Thomas Roberts, Commissionec U.S. Nuclear Regulatory Commission  ;

1717 H S tre e t , N.W.

Washing bon, D.C. 20555 1.

. Hon. James Asselstine, Commissioner U.S. Nuclear Regulatory Commission 1717 H S tree t , N.W.

. Washing ton, D.C. 20555 Hon. Frederick M. Bernthal, Commissioner U.S. Nuclear Regulatory Commission 1717 H S tree t, N.W.

Washing ton, D.C. 20555 Atomic Safety and Licensing Appeal Board U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Hon. Thomas S. Moore, Chairman

• Atomic Safety and Licensing Appeal Board U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Hon. W. Reed Johnson

• Atomic Safety and Licensing Appeal Board U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Hon. John H. Buck

• Atomic Safety and Licensing Appeal Board U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Judge John F. Wolf, Chairman Atomic Safety and Licensing Board U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Judge Glenn O. Bright Atomic Safety and Licensing Board U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Judge Jerry R. Kline l Atomic Safety and Licensing Board

U.S. Nuclear Regulatory Commission i

Washington, D.C. 20555 Harold Denton I

Director of Nuclear Reactor Regulation i U.S. Nuclear Regulatory Commission Washington, D.C. 20555 l Leonard Bickwit, Esq.

1 Office of the General Counsel U.S. Nuclear Regulatory Commission Washington, D.C. 20555 2.

Lawrence Chandler , Esq.

• Henry J. McGurren, Esq.

Office of Executive Legal Director BETH 042 U.S. Nuclear Regulatory Commission Washing ton, D.C. 20555 S ecre tary U.S. Nuclear Regulator y Commission Washing ton, D.C. 20555 Atten tion: Docketing and Service Section Mrs. Elizabeth Apfelberg 1415 Cozadero San Luis Obispo, CA 93401 Janice E. Kerr, Esq.

Public Utilities Commission 5246 S tate Building 350 McAllister Street San Francisco, CA 94102 Mrs. Raye Fleming 1920 Mattie Road Shell Beach , CA 93449 Mr. Frederick Eissler Scenic Shoreline Preservation Conf e rence, I nc.

4623 More Mesa Drive l S anta B ar ba ra , CA 93105 Gordon Silver Sandra A. Silver 1760 Alisal Street San Luis Obispo, CA 93401 l

Joel R. Reynolds, Esq.

Eric Havian, Esq.

John Phillips, Esq.

Center for Law in the Public Interest 10951 West Pico Boulevard, Third Floor Los Angeles, CA ,90064 B ruce Nor ton, Eso. '*

Nor ton, Burke, Berry & French 2002 East Osborn Phran ix, AZ 85064 Philip A. C rane, Jr. , Esq.

• Richard F. Locke, Esq.

Pacific Gas and Electric Company 77 Beale Street, Room 3135 S an F ra nc isco, CA 94106 3.

O David S. Fleischaker, Esq.

P. O. Box 1178 Oklahoma City, OK 73101 Arthur C. Gehr, Esq.

Snell & Wilmer 3100 Valley Bank Center Phoen ix, AZ 850 73 Mr. Richard B. Hubbard MHB Technical Associates 1723 Hamilton Avenue, Suite K San Jose, CA 95125 Mr. Carl Neiberger Telegram Tribune P. O. Box 112 San Luis Obispo, CA 93402

, Virginia and Gordon Bruno P ech o Ranch P.O. Box 6289

'. Los Osos, CA 93402 Nancy Culver 192 Luneta San Luis Obispo, CA 93401 Maurice Axelrad, Esq.

• Lowen stein, Newman, Reis , & Axelrad 10 25 Connecticut Avenue, N.W.

Washington, D.C. 20036 i

l i

l l

4.

Charyle Johnson Five Cities Times Press Recorder P.O. Box 460 Arroyo Grande, CA 93420 DATE D: December 24, 1983 JOHN K. VAN DE KAMP, Attor ney General of the State of Calif ornia ANDREA SHERIDAN ORDIN, Chief Assistant Attor ney General MICHAEL J. STRUMWASSER, Special Counsel to the Attorney General SUS AN L. DURB IN ,

PETER H. KAUFMAN, Deputy Attor neys General f

I/

l By h .

6 STRUMWASSER MIpAELJ.

Attorneys f or Gover nor George Deukmejian 3580 Wilshire Boulevard Suite 800 Los Angeles, Calif ornia 90010 (213) 736-2102 i

i i

1 5.

l

. . . - -- . -.