ML20087K905

Forwards 12 Addl Affidavits in Support of 840301 Petition to Defer Any Futher Licensing Decisions Until Completion of Comprehensive Reinsp,Mgt Audit & Audit of Design Verification.W/O Encl

ML20087K905
Person / Time
Site:	Diablo Canyon
Issue date:	03/23/1984
From:	Clewett J, Devine T GOVERNMENT ACCOUNTABILITY PROJECT, SAN LUIS OBISPO MOTHERS FOR PEACE
To:	Gilinsky V, Palladino N, Roberts T NRC COMMISSION (OCM)
References
CON-DSB-101 NUDOCS 8403260364
Download: ML20087K905 (9)

Text

_ - _ _ _ _ - .

V - -

^tUL GOVERNMENT ACCOUNTADILITY PROJECT e$NcP Institute for Policy Studies 1901 Que Street. N.W., Washington, D.C. 20009 (202) 2am2

'84 MAR 26 A8:28 If0C hr} g Mar h 23, 1984 Nunzio J. Palladino, Chairman Victor Gilinsky, Commissioner Thomas Roberts, Commissioner James Asselstine, Commissioner ,

' N[]", ,~ . ~ ~ ~ g g,f,, ,pk3 I

Frederick Bernthal, Commissioner , _ , , ,

U. S.. Nuclear Regulatory Commission -

1717 H. Street, N.W.

Washington, D. C. 20555 - -

Dear Commissioners:

On behalf of the Mothers for Peace, this submission formally presents 12 additional affidavits in. support of their March 1, 1984 petition to defer any further licensing decisions on the Diablo Canyon nuclear power plant, until completion of a comprehensive reinspection, management audit, and audit of design verification efforts to date.

The 12 affidavits represent the views of six current and six former employees, out of 47 interviewed in the Government Acco'ntability u l Project's ongoing investigation to date.

i In the March 1 petition six national organizations joined the Mothers for Peace. Due to time constraints, there has not yet been sufficient opportuni'ty to review this disclosure with those parties. If the f Commission or the staff demonstrates a good faith intent to seriously review the issues, counsel will organize ,them into 215 new allegations,

. as well as significant supporting documentation for' previous charges.

The 215 allegations include 51 charges of falsified documents, including 32 cases of material false statements in documents reviewed by the ,

l staff prior to its March 19, 19.84 recommendation to let the plant go critical.

Seven of the 12 witnesses have . submitted previous affidavits.

They are supplementing the record, because the staff refused their-  ;

repeated requests for followup interviews to raise additional issues, and expose false or misleading-statements in,1984 PG & E rebuttals.

to the NRC. In one case, the staff had never even bothered to make initial contact with a witness who disclosed severe hardware effects from ongoing Quality Assurance (QA) violations.

The staff's aloof posture over the last month. explains why the Commission will vote Monday on the basis of a record-that is delibe-rately incomplete and deviates severely from reality.- The staff refused to receive evidence that would have rendered untenable its March-19 recommendation. The staff's misconduct violates the Commission's own claimed policies, abdicates its public trust, and raneges-on commitments made to employees who risked -- and in.some

. cases sacrificed -- their careers.in an attempt to uphold their professional ethics and defend the public safety. Unfortunately, 8403260364 840323 PDR ADOCK 05000275 0 PDR

}

the staff wasn't interested. The analysis below summari~es the staff's recent deception to the whistleblowers and the Commission, as well as a few illustrative highlights from the evidence the staff didn't want to hear. The affidavits are enclosed as Attachments i 1 through 12.

I. BACKGROUND The NRC staff recommendation is based on a significant shift in Commission policy toward the nuclear labor force. The staff decided to adopt the common industry tactic of smearing whist 1& blowers who voice serious public safety concerns, instead of seriously investigating their evidence and eyewitness accounts of illegality. In November 1981 the NRC announced to the Udall Committee that the agency would work more closely with nuclear workers to doublecheck the accuracy of utility records and statements, especially to assure the quality of construction. On March 19, 1984 the staff announced that in the future they would only commit to reading whistleblower affidavits and allegations. . . . _ _

The NBC is reneging on its commitment, because the workers at Diablo Canyon exposed more serious problems than the Commission was able or willing to address. The staff explained its reluctance to deal with 130 employee charges raised on March 2, because it would take two to three months to resolve the issues. As a result, the NRC will not even commit to talk with them except at whim.

The new policy represents a break with tradition, as well as the.

agency's formal position. Traditionally, NRC inspectors have used the intensity and flow of employee allegations as a rough indicator of, the plant's condition. In 1982 at Zimmer, when the allegations came in faster than the staff could investigate, the Commission wisely suspended c'onstruction to learn the full effects of a quality assurance breakdown. By contrast, when the employee reports of qual-ity assurance (QA) violations and shoddy hardware at Diablo Canyon came in faster than the NRC could investigate, the staff decided to. stop talking to employees. Instead, the staff accepted at face value the utility edenials th'at anything is wrong.

-The staff justified its refusal to stop pursuing new evidence by-

) trashing the whistleblowers. In effset, the NRC officials told the

! Commission that it wasn't worth the time and money to talk with l witnesses anymore. This was a shameful tactic that has cost the NRC its credibility with the nuclear labor force. Government Account-ability Proj ect (GAP).immediately recommended to its clients that .

they should no longer communicate with the NRC, except pursuant to subpoena. The reason is tha.t it is not in their best interest to -co-operate with a coverup. All clients have agreed enthusiastically.

Some explained that they would not have talked further with the NRC in any event after what happened.

1 1

Wee. 4* e m aww w eee ,,em

-we n.ameeem= h Meh m __ * * - = = w- *w

, _'" j g _sq gqqg y_gm,,o ,A9, g.cg wp . . - + - , -

w -- - -e -- m' *85 -9"'"

- _. . -- - ~.

-3

~

f The staff relied on false statements -- both its own to the )

, Commission, and PG&E's to the staff -- as the basis to reject employee charges. Last Monday the staff assured the . Commissioners that a program of followup interviews with whistleblowers had verified the accuracy of the inspectors' findings. In theory, these interviews

! would assure that the NRC understood the issues prior to resolution, 1 and would help catch any inaccuracies in PG&E's responses. Unfortunately, the staff did not meet with the whistleblowers again, except' in isolated i instances. Over a two week period prior to the Commission vote, the staff refused to speak with the whistleblowers. During that period GAP explained with increasing urgency that the employees had found numerous, significant false statements in the PG&E answers to the whistleblowers' original charges. The staff said they didn't.have time to talk with the employees until after the March 19 meeting.

On March 19, the staff announced that the employees were wrong and not worthy of further interviews.

A review of .just a few highlights from the evidence that the staff refused to consider leads to only one conclusion: on March 19 i the staff tarnished its own credibility, not the whistleblowers'.

II. EVIDENCE OF MATERIAL FALSE STATEMENTS

' The evidence summarized below illustrates a much larger problem at Diablo Canyon

false or misleading statements. Some'of the statements are false by omission; 1.e., they don't include facts that would change the conclusion. Other statements flatly contradict internal

~*

records at the plant. Still others are so incomplete that they are meaningless. GAP has obtained evidence of false statements on the seismic design review; construction. quality assurance for the Pullman; Atkinson and Foley contractors; and inaccurate weld symbols on drawings l throughout Diablo Canyon, that effectively left.the welders and in-spectors trying to follow instructions in the technical equivalent of

. a foreign language.

!e i

The seven examples below highlight false statements concerning the quality of work by the Pullman Corporation at Diablo Canyon back to 1972, the period when critical safety equipment was first installed.

l The false statements were contained in internal files reviewed by the NRC, or in letters received by t,he NRC during February. Whatever the context, the staff accepted the data at face value. As a result, the j

Commission has been duped again.

1) 90 day welders log .

In Report 83-37 the staff drew the following concl.usion: "The inspector examined.the 90 day welders los and found that no void existed between 8/72 and 12/72." The staff used this'as the basis for a finding ~that a Nuclear Services Corporation (NSC) audit-was

' wrons on that point. Somehow the NRC was hoodwinked. An April 1978 l

internal Pullman memorandum on the same subj ect concluded the opposite: 1

- . L - J :T K .- T . . r:::r:: 7 L - =r, L -

.p - -

_4_

t "The're is a vo.id in the 9'O day weld log from Augnst 1972, to .

December, 1972." (Attachment 2, p. 8 .)

2) Qualifications of ouality control personnel Pullman and PG&E contend that " Nondestructive Examination" (NDE) personnel have been qualified since at least 1974 under the requirements of the relevant professional code, the American National Standards Institute (ANSI). Specifically, PG&E claims to have met the standards in ANSI N 45.2.6. NDE personnel conduct the x-rays, magnetic particle, ultrasonic and other tests to learn the quality of the hardware.

In fact, an internal May 13, 1975 Pullman (then called W.W.~Kellogg) memorandum revealed, "It should be noted that it is virtually impossible to comply totally to N 45.2.6 because of experience requirements.

We cannot hire personnel to meet the experience requirements based for the salary scale we offer." (Id. , Exhibit 10. ) As late as July 30, 1982 Pullman QA/QC Manager Harold Karner rejected an internal audit finding that Pullman did not meet the requirements of ANSI N 45.2.6 I with the following comment: "P.G.E. has not stated.in writing that Pullman must comply with ANSI N 45.2.6....We are not in violation of PGE spec ifications or our own procedures ESD 235 and 237." (Id.., Exhibit 9 .)

In other words, even by 1982 reality had not caught up with the "offic ial " rec ord. How nuch confidence can the public have in test results from Pullman's bargain basement NDE' staff? On October 13, 1977 Pullman's QA Manager explained what N.S.C. concluded after looking at the files: "N.S.C. felt that personnel should have had more prior experience before training and qualification for N.D.E. or. field in-spection. They did not' like using ex-mechanics for N.D.E. --personal opinion."

3) Backup data for NDE oualifications The staff accepted at face value PG&E and Pullman's claim that adequate backup data exists to demonstrate that NDE personnel are qualified based on training and experience.

. In fact, a September 22, 1977 Pullman corporate review of 95

inspector and NDE files found "gener'ic records deficiencies," for. i data to prove experience and qualifications. '(Id,., Exhibit 11)

Reviewing the file for Pu]1 man's NDE supervisor -- the leader of the pro 5 ram -- reveals the magnitude of the bluff. His records say he passed the three Magnetic Particle exams with flying colors -- a ,

score of 985. But records on.the three " specific exams record" the l l

following results for the supervisor: , , [and] . There i are no grades recorded for his performance on individual tests.

(Id., Exhibit 12) l

=.+.-.. ....=...........- .  :  ;....- __

J -

l

4) Whether inspections occurred In 1976 Pullman reported that the same NDE supervisor allayed Berns about 1200 suspect weld attachments by reinspecting 314 with notic particle tests in ~fourdays. That represents '7.8.f. tests per day.

y all passed. (Id. , Exhibit 17. ) Unfortunately, the maximum possible ter of magnetic particle exams that an inspector can perform in a is around 50 and even that is for the simplest versions of the B. ( Id . , p . 6 .7)

5) Rewriting history through backdated documents On September 25, 1980 an internal Pullman audit admitted that two technicians were certified for advanced (" Level II") responsibilities,

,ito " letters in their personnel files stating they are not qualified crform Level II functions...." (Id., Exhibit *S.) Pullman's tion? Bemove the first letters and have the same NDE supervisor

'dato letters to July 24, 1980 that said the opposite -- that the j men were qualified. Rewriting' history is no way to solve quality >

lems.

6) Xeroxing signatures The signatures on weld process sheets -- which insure the work was done in an ad hoc manner -- were phoney. A blank sheet was signed then xeroxed! This is evident from a review of multiple weld process pts -- the signatures are too perfectly identical. The phenomenon 3 has been confirmed with engineers from the early years. (Id. , p. 8 .)

In short, we do not quarrel with the staff's finding that PG&E its contractors discovered their own QA violations internally. But g not enough to find the problems unless you fix them. The worst

-nso is to be aware of the violations and then cover them up. That

$ hat happened at Diablo Canyon.

7) cracked welds on component cooling water system PG&E defended Pullman's practice of welding on water-filled pipes, ah can cause cracks in the welds and in the pipes themselves,.by

<=ing that the primary cause of cracking would be the presence of rogen, which PG&E then asserted was minimized.

In fact,. PG&E intentionally overlooked the maj or cause of king when welding on water-filled lines, namely the water itself, h quenches the weld and can cause cracks in the weld and "under-

" cracking in the piping itself. (Attachment 7, at p. 3) l Q --- . - . .- ...

,3 . .

l The above represent only a few of the many false statements for t which the NRC-staff refused to receive evidence over the last month.

It hardly illustrates the scope of the staff's own coverup, however.

Some of the crudest examples involve the. issues screened out of the staff's inspections as not.sufficiently significant. For instance,

had the staff looked it would have learned that until 1978 Pullman's corporate audits did not follow.the legal quality assurance standards of 10 CFR 50, Appendix B. The early audits were crude, informal .

effortss where construction personnel " audited" themselves on. occasion.  !

They were no substitute for a minimally-acceptable, professional program. (M. , p2-5)

At least until 1982 Pullman's program for pipe. supports and. rupture restraints -- the subj ect of the current seismic - design review -- did not honor 10 CFR 50, Appendix B. It is ironic.that the staff was not

interested

10 CFR 50, Appendix B represents the heart.of the NRC's

regulatory program for quality assurance.

! III. CONDITION OF THE PLANT Through the previous two petitions whistleblowers described the effects of the quality-assurance breakdown on the condition of the plant -- an nncontrolled welding program that resulted in " pathetic" welds, as well as braces for the electrical system that were so loose

they could be moved by hand. Unfortunately, the staff would-not even accept a whistleblower's offer to point out examples of the shoddy hardware, let alone require a reinspection. .

The ~ QA violations above similary were not mere technicalities.

j The consequences could threaten public health and safety. [For e'xample, PG&E has overlooked the' fact that cracks are actually occurring in the 4

field for welding on water-filled lines. Weld,ers have said that when they weld on the water-filled pipes, their weld freezes on contact, showing that cracking is likely. Surface cracks have already been observed. (Attachment.3, at 2-3.)

l i

Similarly, on February 24, 1984 the integrity of the plant's containment was breached when one of the main containment airlock doors "

blew open. Workers were sucked from outside into the inside of the

~

contain=ent building by the force. Maintenance crews explained that the hinges for the fourteen-year-old door were worn out.before the-i plant began operation.

i The NBC and PG&E scoffed at the-incident,. explaining-that the -

door was c3csed-in three minutes. -That: misses the point. How secure is the containment if a few workers can manually override a mechanical safeguard system? 'Apparently the staff was not concerned.

i "nder the plant's safety technical specifications (tech specs),

priorL e fuel loading and periodically thereafter a surveillance pro-L cedure must;be performed to assure containment integrity. 10 CFR 50, l Appendix J, section (b)(1) calls-for the airlocks to be tested prior .

I to fuel-Icading. The surveillance and' test' process is1 designed to pre-l' -

\

! l x ' _. --.._.._ _ . _

.. - _ j._,,_.___ _ , _ , , , , _ _ _ ,

vent precisely the type of situation that occurred in February --

safety-related structures that are worn out before the plant even operates. A thorough search of the public documents room in San Luis Obispo, California failed to identify any description of relevant surveillance or tests performed. In short, either the surveillance and tests did not occur, or they were inadequate.

How many other structures and components are worn out already and represent dormant safety hazards missed by the surveillance and test program? Again, the staff was not interested.

IV. GHOST ALLEGATIONS The staff has not interviewe'd any witnesses concerning the 131 allegations the witnesses presented on March 1, 1984 through the-Mothers for Peace. On Monday, Earch 19, 1984 the staff announced that it was not necessary, becau.e the allegations were "similar" to others raised previously. The excuse would be irrelevant, even if it were accurate. "Similar" QA violations could represent inde-pendent causes for an accident, a possibility the staff casually dismissed. Furthermore, in many cases the "similar" allegations greatly increased the significance, and analytical and evidentiary development, of earlier issues.

In fact, the staff's rational 12ation has the same degree of accuracy as some of PG&E's denials -- next to none. The list below illustrates some of the issues introduced on March 1:

1) circumvention of QC inspection criteria, QA reporting re-quirements and formal design criteria through the " Quick Fix" program

-- an uncontrolled after-the-fact " trial and error" approach to design, in which field engineers could veto the approved design at whim, without calculations.

2) a December 28, 1983 procedural modification to ESD-223, which instructs QC inspectors to accept welds that were not required

~

by the design, and to accept the absence of welds that were required by the design -- in other words, to accept anything that does (or does not) exist -- as long as the deviation is identified on an as-built drawing.

3) abuses in the. engineering review:of the as-built drawings thap supposedly catch the errors funne11ed into Quick Fix and out of QA/QC, such as approval of suspect design deviations without any basis in support-ing documents of engineering calculations.

4) continuing discrepancies between the~ drawings relied on by operators, and the as-built condition of the plant, for such critical items as the location.of valves.

5) another Department of Labor decision finding that an employee was dismissed for writing an internal quality-control report challenging f

the lack of QC coverage in the weld qualification test booth.

6 e

• • * * * * = - a=-emmeo --4-e m, 4.,. t, _

g. ..

. _8_

6) Bechtel pressure for employees who resign to sign a "see no evil" form saying that they are unaware of any code, design or QA violations. This illegally places the employee in a Catch-22 situation, since the statement is obviously false. If later ques-tioned by the NRC, the employee must choose between denying the obvious or repudiating a signed statement to Bechtel. The former option risks criminal prosecution; the latter risks industry blacklisting.

PG&E's response to the last statement is almost humorous --

the company said that the gag form was an attempt to learn what the problems are at Diablo Canyon, by having employees sign a form stating that there aren't any problems. As one whistle-blower still on site explained.to counsel in disgust, "If they wanted to know about problems, why didn't they just give us blank paper and ask us to start writing?"

The staff's response was less_ amusing. They. ignored all these issues. These issues effectively vanished from the NRC's regulatory program, although 'if correct certain of the violations would estab-lish landmarks in quality-assurance abuse.

In short', the staff obliterated the quality of its own factual record. If the NBC were a utility, it would have violated the First Commandment of its own quality-assurance rules. 10 CFR 50, Appendix B, Criterion I prohibits utilities from sacrificing quality for cost and schedule concerns. Unfortunat'ely, 10 CFR 50 Appendix B no. longer appears relevant for Diablo Canyon, if it ever.was.

This discl6sure 1s offered.

~

. to th,e Commission in the hope that information, evidence and public safety are all still relevant for licensing decisions under the Atomic Energy Act.

Respect fully " submitted, u -

' Thomas De' vine-John Clewett NM Government Accountability Project 1901 Que Street, Northwest Washington, D.C. 20009 Attorneys for the Mothers for Peace l

l

• Jed in %2es, '

2 Sema &w.  ;

i' i