ML20129B192

Affidavit of H Hudson Re QA Breakdown for Work of Pullman Power Products at Facility.Qa Breakdown for Welding,Nde, Hydrostatic Tests,Vendor QA & Records Falsification Discussed

ML20129B192
Person / Time
Site:	Diablo Canyon
Issue date:	01/12/1984
From:	Hudson H PULLMAN POWER PRODUCTS CORP. (FORMERLY PULLMAN, INC.)
To:
Shared Package
ML20129A429	List: DCL-84-170, Forwards Description of Items of Concern Re Design & Const Identified on NRC 840411 Insp of Facility W/Selected Anonymous Allegers & Util Resolution.Supporting Documentation Encl ML20129A424 ML20129A475 ML20129A501 ML20129A512 ML20129A527 ML20129A569 ML20129A713 ML20129A728 ML20129A748 ML20129A763 ML20129A789 ML20129B012 ML20129B038 ML20129B053 ML20129B077 ML20129B098 ML20129B126 ML20129B192 ML20129B227 ML20129B262 ML20129B315
References
FOIA-84-745 NUDOCS 8506050144
Download: ML20129B192 (31)

Text

_

k g

AFFIDAVIT My name is Harold Hudson.

I am submitting this affidavit freely and voluntarily without any threats, inducements, or coercion, to Mr. Thomas Dev'ine, who has identified himself to me as the Legal Director of the Government Account-ability Project of the Institute for Policy Studies.

I am submitting this state-ment to evidence my concem over a comprehensive quality assurance (QA) breakdown for the work of Pullman Power Products at the Diablo Canyon.*luclear Power Plant.

There is no possible justification for allowing this nuclear power plant to go critical until the Wuclear Regulatory Comission (NRC) confinns the full scope of QA breakdown; identifies the causes; and monitors completion of a corrective action program, including a full reinspection of safety-related work at the plant.

In many instances, the reinspection may be the first legitimate quality control coverage the hardware has had.

I base this conclusion on my four and a half years experience at Diablo Canyon in Pullman's quality assurance / quality control (QC) program, including two and a half years, through 1982, during which I was the Internal Auditor. The t

basic lesson I learned is that the conclusions of a Nuclear Service Corporation audit of Pullman ars more true today than when first published in 1977--the program does not meet the requirements of 10C.F.R. 50, Appendix B; and it does not have an operative corrective action system. The latter has been demonstrated by the further deterioration in corrective action from 1979-1983. While before, the system was merely failing to identify and solve problems, now it is actively covering them up. This has been especially true with respect to welding, non-destructive examination procedures (NDE), and hydrostatic tests--all of which I learned 'were consistently uncontrolled, and that some of the procedures for the first two items were not qualified by a testing process which proves the procedures actually work as claimed.

8506050144 841211 PDR FOIA

~'

DEVINE84-745 PDR m

.., au _ -._ w== -.- = ;.:= u : ::.=-- :m _.. - -.

The system also broke down for vendor quality assurance, where Pacific Gas and Electric (PG&E) management ordered Pullman inspectors to stop reporting cracked welds found in structural steel restraints supplied by vendors such as Boston Bergen and Anerican Bridge.

As an auditor trying to work within the Pullman site and corporate QA system, I learned the cause of the QA breakdown and why it has not been corrected.

Pullman QA Management does not want to know about QA/QC violations. Management's corrective action has been to harass, threaten, and intimidate QA/QC personnel

~

who identify problems, and to dismiss those who persist. Although I exhaustively reported deficiencies, the major effect of my disclosures was to prompt orders from,the QA manager to only look where I was told, and his angry threats to "get rid of me"During one such exchange,he exclaimed Pullman's bottom line: we're not committed to building this plant to 10 C.F.R. 50, Appendix 8.

In that case I do not see any legal basis for the NRC to allow this plant to operate.

I am not opposed to' nuclear power. Rather, I believe in the technology enough to insist that it receive the proper respect.

I began working in the nuclear power industry in' 1974 at the Trojan Plant and have worked at the Humboldt Bay Plant.With the exception of two months in 1979, I worked at Diablo Canyon for Pullman from September,1978 until Friday the 13th,1984, when I was i

laid off. The layoff occurred the day after I finished a two-month series of disclosures to the NRC.

For my first three to four months on site, I was a documents reviewer. For nineteen months I worked as a weld inspector in the pipe rupture restraint program.'

In August, 1980, I was promoted to QA Internal Auditor.

1 e

O m

6

-w---

n-,----

m.

-n,

,--,--,,--,r-,,,

w,rs,,-,,

,..,-m

-~-w

,,-,,w,,,----

,--,wn

My responsibility was to evaluate and monitor the entire QA/QC program for compliance with our legal obligations. This is how I learned that Pullman does not consider 10 C.F.R. 50 a legal obligation for work at Diablo Canyon.

In January,1983, I was removed as internal auditor, but remained in the QA program to help close out Discrepancy Reports (DR) and Deficient Condition Notices (DCN), as well as to complete my pending audits. QA Manager, Harold Karner, restricted me to carrying out his specific assignments. The harass-ment was so intense that in mid-May, I resigned. Through my union, the next day I return to Diablo Canyon as a ~pipefitter. There simply had been too many headaches attempting to work within the corporate system. On my own time, at home, I finished organizing and summarizing my evidence of QA violations.

In November, I completed an initial report. On November 28. I sent it to NRC Commissioner, Victor Gilinsky. On December 6,1983, his office wrote that I would be contacted by the Office of Investigations (01). Although OI never i

called, on January 6, 9 and 12. I was interviewed extensively by a series of NRC inspectors from Regicn V.

On January 13. I was laid off.

This statement will summarize the information and list the allegations in three written reports already disclosed to the NRC. My affidavit also is to submit a written record for allegations which I have only described to the NRC in interviews and identify allegations not yet described to the NRC.

I.

QUALITY ASSURANCE 8REAKDOWN FOR WELDING With a few exceptions, from the onset of construction, the welding program for structural steel essentially has been uncontrolled--in violation of legal requirements, as well as contract and design specifications.

The techniques to circumvent quality assurance included unqualified welders;

-l

.=-

.v

=-

- =

u,

.~ r = ~.

t.

4 unqualified welding procedures; use of welding procedures so irrelevant for the assigned work that, in effect, safety-related welding was widely conducted without procedures; reliance upon unqualified inspection procedures to check the quality of the welds; informal changes of contract specifications without the required administrative review or distribution; falsification of records; and harassment and intimidation of.QA personnel who identified and attempted te obtain. corrective action against the violations. The abuses occurred both during original construction, and during the current modifications due to the Bechtel/PG8E seismic design review program.

The list below represents a more detailed summary of the allegations and evidence that form the basis for the-above conclusions. The list is I

drawn primarily from my November 28, 1983, disclosure and attachments to Commissioner Gilinsky, which are enclosed as Exhibit 1.

1.

Weld procedure Code 7/8 for piping and plates has been used improperly to weld numerous forms of structural steel on pipe supports. What happened is that Pullman substituted American Society of Mechanical Engineers (ASME) pipe welding procedures for the American Welding Society ( AWS) struc-tural steel procedures, as implemented. This practice exceeded the legally-approved limitations for use of the procedure. The limits were logical, since the two types of jobs have little in common.

Pipe welding involves working around a circumference. In structural steel welding the axis of the weld is on a. straight plane (Exhibit 1, at 2).

i 2.

Code 7/8 has been used improperly to weld tube steel on pipe suppo rts. Tube steel involves a different type of metal than the P-1 material covered by ASME procedures.

This is significant, because the NRC has identified i

1 i

e o

.r_,,

~

1 use of the same metals as a precondition to use ASME procedures for AWS work.

In fact, tube steel welding is so unique that the AWS Code has a special sec-tion for it (H., at 2-3).

-3.

Code 7/8 was improperly used to weld threaded weld studs

which bolt plates to civil steel on Class I safety-related pipe supports.

The type of welding used for these studs is not listed within Code 7/8, and

~

it bears almost no resemblance to the ~ work legally covered by Code 7/8 (M., a t 2 ).

4.

The welding for threaded studs did not even honor the require-ments of Code 7/8, which calls for the use of a backing bar. Instead, process

~ sheets operated by the construction department imposed backgrinding, which is a totally different operation (H.).

5.

Code 7/8 has been used to weld.at least eight pipe support joint configurations, including flare bevel groeve welds, and double bevel groove welds, not covered by Code 7/8. Each of these configurations repre-

~

sents a unique welding task and legally must have its own approved weld procedure specification detailing the joint configuration (M., at 3).

6.

Process sheets that guide quality control coverage did not consistently call for inspection to verify the fitup of flare bevel j

groove welds; one of the joint configurations not covered by the 7/8 pro-l cedure in the first place. That leaves the quality of the ensuing welds l

doubly unreliable. This uncontrolled work has been occurring as part of l

the current design modification construction work (H.). I have read a PG&E memorandum asserting that QC fitup inspections are not required for flare bevel welds.

That memorandum is not sufficient to overrule engineering h

+

-w,...

.-, w m, v.

,---,,-,,---.--..-....-w,--,,.--,-

-.------,-----------+,-----,--...---4.-,----I-'1,v'-

I i

6 i

specification ESD 264, which requires inspections of groove welds and full penetration welds.

7.

Code 7/8 has been improperly used on pipe rupture restraints to weld five types of metal different from the ASME approved P-1 material..

. These restraints prevent a pipe ruptured during an earthquake from whipping back and forth, which could damage the rest of the equipment (M., at 4).

.8.

Code 7/8 was improperly used to weld two structural steel shapes on pipe rupture restraints that are not covered by the procedure--W shapesandtubesteel(M.).

9.

Code 7/8 was improperly used for at least 11 joint config-urations not covered by the procedure itself. These joint configurations were not generically prequalified per the AWS Code and were without Procedure Qualification Records and/or were not detailed on the Weld Procedure Specification (M.,at4-5).

10. The result of the procedural breakdown was uncontrolled welding. To illustrate, in one example, pipe rupture restraint square groove l

.we ds were conducted without any established or documented procedure that applied to the work in question. In some instances, welds had been completely removed without any QC record of their disappearance. The records reflected QC accepted welds where none existed.

For documented repairs, there was only

{

erratic QC coverage due to unexplained procedural changes that deleted the

~

requirement for nondestructive examinations (M., Attachment 2).

~

t 4

e e

d l

,-,-----,.-..-,~,-,--~.._.-..-,,,.,w2.1:,,,-.,.nn,n

• .,-.J,'*

• .-.*L "L-."~~~

i 11. Pullman has recognized the error of applying ASME welding procedures to AWS work in an uncontrolled manner and issued Welding Technique Specification No. AWS 1-1, in an attempt to clarify the proper use of Code 7/8 on AWS work. But the scope of. corrective action was inadequate.

It only covered the. work in a weld crack repair program on pipe rupture restraints (H., at 5-6). The misuse of Code 7/8 far exceeds the use of AWS 1.1.

The crack repair program only covered about one-fourth of the pipe rupture restraints, and none of the pipe. supports.

12. AWS 1-1 failed to fully correct the improper use of Code 7/8 for welding in the weld crack repair program. The procedure uses a steel not contained in the list of acceptable AWS base metals, without evidence that it had been individually qualified to prove its reliability (M., at 6).

13. The above violation was approved on December 20,1979, by V. J. Casey, who signed off as Cognizant Welding Engineer. Sixteen days earlier, however, he had been appointed Pullman's Assistant QA/QC manager.

according to an interoffice memorandum. To my knowledge, Mr. Casey has never been listed on the Pullman organizational chart as a Cognizant Welding Engineer. The only way his approval would not represent a false statement is if he were simultaneously a construction and QA official. That would be a violation of the NRC's requirement for a QA program independent of construc-tion (M., at 6-7).

s 14.

I also have serious reservations about Mr. Casey's qualifica-tions, based on his judgment in the field. In 1978 Mr. Casey was my

~

supervisor when I began as a welding inspector. He instructed me to measure fillet welds by the throat, when the AWS Code requires the measurements from y

4-=c--e,-------,-.--,---,,.--.,._.,,,.....,.,,.--,r w

-,-e-,

--&.-m

,-.e.

---

.w--,. - - -,,.

=--,eew

.g.

?

the leg of the weld. For approximately two months. I inspected welds to the wrong standard, because Mr. Casey gave me a makeshift gauge not designed to measure fillet welds. Other inspectors informed me that Mr. Casey has changed the rules on the spot for equipment anchor modifications in the containment.

They stated.his instructions were to work to a " relaxed" engineering specification ESO 243.

15. Through loopholes in its Engineering' Specification ESD 223 Pullman improperly exempted itself from AWS design, fabrication, and erection requirements for all structural steel pipe, support welding. Writing off the rules in this fashion violated the PG&E contract specifications. To my know-ledge, there is no documented authorization from PGAE to deviate from the Code requirement, which is still in the contract (Jd., at 7-9).

16. PG&E contract specifications on welder qualifications were changed.without required review and authorized approval.

The rules were 4

changed through a cryptic, unexplained note.

The changes involved the qualifications standard for all rupture restraint welders before July 10, 1979. The use of ASME qualificction standards for welders doing unrelated AWS work mirrors the breakdown in welding procedures. Again, however, the 1979 corrective action only applied to rupture restraints (JA., at 9-12).

17. The PG&E contract requirement for Charpy, or notch impact,

strength tests, was waived for Code 7/8 and other welding procedures. Charpy l

tests are necessary to be sure the welds installed under the procedure can meet relevant design and professional code requirements for strength.

1 Deleting this requirement was a serious step, which should have gone through the Con, tract Specification Change Notice process to assure proper engineering review and approval.

Instead, in January,1974, a PG8E piping superintendent l

~

S

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

. g.

removed this significant QA check with a one-word penciled response, "No",

when Pullman asked in a letter if weld procedures for rupture restraints required Charpy impact tests (Id., at 12-13).

18.

In violation of still unrevised contract specifications, specific corrective action commitments on relevant Nonconformance Reports (NCR), and relevant procedures for the weld crack repair program, none of the full penetration welds less than 9/16 in. thick among rupture restraints were ultrasonically tested.

This means that the welds in rupture restraints since July,1979, were not fully covered by quality control tests in a signiff-cant number of cases. PG&E engineers accepted the loopholes to Pullman's program in July,1979, again without the required review and approval, and without revising the relevant contract specification that was being ignored (Id., at 13-15).

19.

Another weld procedure, Code 88/89 for carbon steel piping, has been used to weld pipe support structural steel shapes and plates.during both original construction and repair work in the current design modifications.

Structural steel shapes and plates are not covered by Code 88/89 (Id..at 16).

20.

In violation of the contract specification, Code 88/89 has been used to weld carbon steel plates and structural steel shapes to rupture restraints with two welding processes, Shielded Metal Arc Welding (SMAW) and Gas TungstenArcWelding(GTAW).' GTAW is not covered by.the relevant AWS Code (,Idf) d 21.

In August,1979, PG&E issued Welding Technique Specification No. AWS 1-3 to clarify the use of Code 88/89 for AWS welding. Unfortunately, the " solution" again repeated the problem. AWS 1-3 covers a welding process, (GTAW) and a base metal (A-515) not covered by the relevant AWS code provision (Id., at 16-18).

i 9-!SP

,.t,#

7.t

- 1

?

• 9_ _ _ - - _ _

• 9 t'1 4.*.

t,.

_,,,.,y-_-.

_e.

a 22.

Pullman also substituted welding procedure Code 92/93 for pipe rupture restraints when the process sheets specified that the work would be done to Code 7/8.

The Pullman Assistant QA manager accepted the switch in an August 15, 1978, memorandum without changing the process sheets--

which left a record of work to a different procedure than was actually used.

(1d., at 18). The only records accurately reflecting the weld procedure used were the weld rod requisition forms (1d., at 21-22).

23. The informal approval of the welding procedure switch was based on a false premise--that both procedures were qualified to unlimited thickness and were technically equivalent.

In fact, they only bear a passing resemblance.

For example, Code 7/8 does not include a type of welding in Code 92/93 that is only universally approved by the AWS for welds up to 1/4 in.

thickness. Nor did Code 92/93 have its own procedure qualification test to I

verify its reliability on the welds greater than 1/4 in. thick.

In effect, that welding was uncontrolled and its quality is legally indeterminate. The two welding procedures are also different with respect to joint configurations, joint details, tacking the joints, weld processes to be used, backing bar requirements, and welding techniques, such as the allowable heat input from AMPS and maximum volts. The controls for clearly distinct special processes cannot be legally intermingled through a remorandum (1d., at 18-21).

24 Contrary to contract specifications, welders qualified to ASME-based Code 92/93 were used for structural steel welding without being properly qualified to the AWS Code. The switch was accepted on August 15, 1978, Interoffice Correspondence, rather than through an accountable procedure with re, view, authorized approval and a Contract Specification Change Notice (Id., at 20-21 ).

• 1 **,,* * " -.,......, __

• _

-e,

• *_. - _ - - -.+:-,__,a-

• • Ph 19Qt?Gm

11

25. An April 14, 1983, Discrepancy Report on 1972 welding in the Spray Ring Piping System for the Unit No.1 containment dome. DR f4713, failed to identify an organizational breakdown far more significant than the issue it disclosed (variations between the SMAW weld process used and the process reported in the process sheets). DR M713 also revealed that the process sheets and rod requisition forms referenced different weld rods 1

than had, in fact, been used. The response of the QA/QC manager was to accept the violation as is. The DR did not mention one of the most signiff-cant violations: the production department substituted an unauthorized, unapproved procedure and process for the procedure which had been properly selected and approved by the QA system and the third party authorized inspector from the State of California.

This was done in order to avoid delays when QA issued the wrong weld rod for Weld Procedure 128. Production could not wait i

to correct the weld rods, so the foreman just changed the procedure.

In other words, the production department's " solution" was to achieve compat-ibility by making the procedure as wrong as the weld rod. DR #4713 endorsed the procedure switch (!d., at 23-25).

If production can overrule the QA system so easily on such casual grounds, it means that controlled welding

" procedures occurred only when tolerated by the construction department.

i Under the circumstances there can be no basis for confidence that the quality of the welding was controlled. Most significant, in April,1983 Diablo Canyon management was still satisfied with this result.

26. DR #4713 missed another equally significant violation: QC inspectors had approved all the welds after visual examination, although the I

j GTAW and SMAW welding procedures do not.look the same. The 1972 failure raises serious questions about the reliability of QC inspections at the d

"I^

~ :

l time. The failure of DR #4713 to even note the QC inspection failure demon-strates that 11 years later, the acceptance standards have not yet become realistic. Significantly, before it was issued, this DR was reviewed three times by Bechtel and PG&E management, which must assume responsibility fer a QA report that failed to disclose, at all, the most significant QA violations (H.,at25-28).

27. The breakdown in records for the weld rod and weld process sheets render it impossible to verify the qualifications of early welders by reconstructing weld rod and process records, as asserted by Pullman in response to 1977 Nuclear Services Corporation findings that the qualifications could not be established for welders in late 1972.

I demonstrated this effect of DR #4713 by applying its findings to a case study on a welder whose qualifica-tions were challenged in the original NSC audit (,Ij., at 28-30).

28. My attempts to perform my audit duties on welding led to sustained management hostility, including restrictions on my organizational freedom, harassment and intimidation, and retaliation through personnel actions. On January 28, 1983, the harassment reached a climax.

I had already

• been removed as internal auditor on pretextual grounds (infra, at 23-4) and was doing research for pending audit reports that I had issued, in this case Unscheduled Internal Audit #35 on pipe rupture restraings. I was at my desk reviewing the records on three full penetration welds that had been tested to the wrong nondestructive examination process. Mr. Karner approached and wanted to know what I was doing. When I told him, he asked if I had been directed to identify these problems. Because I was completing a pending audit of which Mr. Karner disapproved, I accurately answered, "No."

He then shouted at me that I was no longer the internal auditor and could no longer identify

- y i

g discrepancies unless he specifically ordered.me to. At the time, I was still a quality assurance employee, helping to close out DCN's and OR's. Mr.

Karner's orders to restrict my inquiries violated the raquirement for organizational freedom in 10 C.F.R. 50, Appendix B.

29. During the January 28, 1983, confrontation, Mr. Karnir also threatened that if I repeated this type of behavior, he would "get rid of me."

From his demeanor. I was unsure whether he was referring io my presence on the job, or my presence--period. Mr. Xarner's threats eventually convinced me to resign and to take a pipefitting job. The pervasive atmosphere of

\\

intimidation was too counter-productive for an employee to successfully uphold '

required QA/QC standards within Pullman's quality' assurance program.

30. Although Pullman has gotten rid of me, the company has kept the probleia of unqualified welding procedJres. When I left in January,1984, we were still working to the same welding procedures I had auditeq. Nothing has changed except that after all the notice, it is clear that Pullman and C

PGLE's violations are deliberate. There can be no excuse of ignorance.

Corrective action has been nonexistent or ineffective. There were discussions on-site of attempting to qualify Code 7/8 after the fact, which would have been ineffective anyway since it was the sponsoring procedure for considerable work that it did not describe. As of my departure, however, even that halfway step had not occurred.

II.

QUALITY ASSURANCE BREAKDOWM IN NONDESTRUCTIVE EXAMINATIONS Nondestructive examinations to test the welds and other hardware were as unreliable as the procedures to conduct the welding in the first place.

The indeterminate quality of the testing process leaves the quality of the

---,w,

---

e-

hardware in the same status--indeterminate, at best.. In some cases, NCE results were compromised due to simple manipulation at manage-ment direction. This phenomenan allegedly occurred when Bechtel and PG&E had the NDE personnel do certain ultrasonic tests (UT) over with a different approach, after the tests had identified a large number of rejectable welds.

A good illustration of the quality assurance breakdown involves 1972 tests used to measure Seismic Class I valves on the reactor coolant pressure boundary for minimum wall thickness in response to an Atomic Energy Commission (AEC) directive.

The UT procedure was not qualified by tests to determine its reliability, which was questionable anyway, because the procedure did not measure the entire surface of the valvee. There is serious question whether all relevant valves were examined, in part due to conflicting informa-

- tion in the records. Not all the equipment used to measure the valves was

. traceable and calibrated.

The former violation invalidates usage of the equip-ment.

The latter affects the accuracy of UT results by up to 48 percent, when the AEC required 98 percent accuracy.

Informal changes of contract specifications, without the required review and approval, again facilitated,

,the QA. violations.

To my knowledge, corrective action has not occurred.

The unreliability of valve measurements was representative of a general QA breakdown for nondestructive examinations.

In Internal Audit 101, I checked 21 such precedures--seven were deficient, representing three forms of nondestructive exams To date, the most significant problem remain. The basic flaw was that records were not available to demonstrate that test pro-4 cedures were qualified. After I traced the use of one procedure back to the steam generator feedwater nozzle, the QA manager ordered me not to find out where a related test procedure was. used.-

The response to my disclosure of these problems was to sit on them for over a year.

In scme instances, there still 9

.__,.-~,.-.y.3-,...,-.w-

, -~

p-.,,.7,,vv_..,_.

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

,,,.,-~,..----,-.,w-.._...-,,.m c--

m

-15 has not been effective corrective action.

QA management reneged on solutions to which we had agreed. The situation became so frustrating, that I conducted an audit on corrective action and sent the results to Pullman corporate head-quarters. The response was to reprimand me for breaking ranks, while the QA violaticns continued to be ignored. Below is a more detailed listing of related allegations.

31.

In some instances, the unreliability of nondestructive examinations is due to manipulation of the test results in order to mask deficiencies. This allegedly occurred in 1982, with respect to tests involving around 230 Unit I full penetration welds--so,me in the containment--where UT examinations revealed large numbers of rejectable conditions. Witnesses described the defects to me as voids, slag, and lack of fusion in the roots of the welds--which raise questions about weld bonding.

I was also informed that Bechtel and PG&E management responded by manipulating the UT procedure in a manner that would lower the number of rejected indications. The welds were then " accent (ed) as is" (Jd., at 15).

In other instances, the QA violations are more deeply rooted.

The case of Engineering Specification ESD 234 for ultrasonic measurement of valves on the reactor coolant pressure boundary is a micrccosm of the break-down. On January 18, 1982 I initially reported QA violations through Internal Audit #101.

I tried again in November, with unscheduled Internal Audit #34.

l On January 2,1984, I finished a report to Commissioner Gilinsky on this still uncorrected problem, which I have since forwarded to the NRC inspectors at Diablo Canyon.

It is enclosed as Er.hibit 2.

32. There is no evidence that the ultrasonic thickness measuremnt O

e

16-procedure was qualified through tests to demonstrate the 98 percent level of I

accuracy required by the AEC. The valve measurements were conducted with an

{

uncontrolled procedure, and therefore cannot be accepted as the basis for conclusions about the quality of the valves. In my audit, I could neither find evidence of a Procedure Qualification Record (PQR), nor a Procedure Qualification Test (PQT) (Exhibit 2 at 2-3).

33. There is no evidence of " procedure verification tests,"

required by ESD 236 for the transducers,that take into account the cu. ves, ridges, and irregularities that exist on every valve and significantly affect the measurements (1d., at 3).

34 Management appears to have conducted the measurements without any qualification test, despite prior warning that the procedure was too unreliable to support its findings.

An April 17,1973, " Interoffice Corres-pondence" had disclosed:

1 i

3.

The transducers available are adequate for flat smooth surfaces. There are no adapters, shoes l

or wedges available should they become necessary.

4.

At this time, it appears the transducers supplied may not be the correct. type for thickness readings.

If this is true, we will have to order new transducers, f

5.

The effect of surface contour and roughness must

~

be tested prior to making any reportable results.

6.

There is no available equipment on the U.T. equip-ment for review.

4L It is doubtful that any meaningful results can be obtained at this time and it is definite that none can be reported until the above-mentioned problems are solved.

(id,., and related attachments)

• O

?.

35. Pullman QA manager Harold Karner improperly refused to take corrective action in January,1982, when I disclosed the lack of pro-cedure, qualification records or tests for ESD 236 and ESD 244, the UT Thickness Gauge Procedure. The problem remains uncorrected. His excuse was that these procedures were only nondestructive measurements rather than nondestructive tests, and therefore did not represent "special processes" whose quality must be controlled (Id., at 4).

That semantic distinction is irrelevant.

The reason to require reliable, controlled procedures is to assure the quality of sensitive, safety-related hardware. Indeed, in 10 C.F.R. 50, Appendix B, Criterion X, the terms " examinations, measurements or tests" are used interchangeably.

The safety-related purpose for qualified NDE procedures is magnified for ESD 236 ESD 236 was instituted in response to an AEC directive to the nuclear industry after discovery of valve problems at a series of plants.

36. Mr. Karner's manipulation of definitions is wrong.

UT measure-ments constitute a special process which must be qualified.

They are a special process because they are uniquely created to perform a specific quality-

'related function.

Further, PG8E contract specifications and 10 C.F.R. 50, Appendix 8. Criteria IX, " Control of Special Processes," identify nondestruc-tive testing as an example of special processes, not as the berndary of the, concept.

37. UIA #34 of 254 Valve Wall Thickness Data Reports demonstrated that the Data Reports are incomplete and, therefore, are not traceable, as required.

For example, none listed the size, shape, or manufacturer's designation for the transducers that performed the wall thickness.

The ESD O

4 i

1 236 Documentation Packages do not provide any information on the testing equipment beyond the seria1' numbers.

In some cases, there were not even serial numbers for the UT machines and the micrometers used as a mechanical backup measuring device (H., at 5-6).

38. The Data Reports offered unreliable, inconsistent information.

For instance,19 reports listed two different UT machines as having conducted the same valve measurement. Serial numbers for UT thickness equipment and micrometers could not be verified independently. Ten percent of the valves checked physically had serial numbers different from those listed in the Data Reports.

In many Data Reports, original information had been whited-out and altered without signature or explanation (M., at 6).

i 39.

Necessary records to demonstrate calibration of the measuring i

equipment were not consistently available. To demonstrate the potential i

effects, on three UT measurements whose accuracy was tested, the pre-and post-calibration checks showed variations of 10 percent, 48 percent, and 2.6 percent (M., UIA #34, Attachment 5).

The maximum error permitted by the AEC was 2 percent.

40.

The AEC acceptance standards were violated when valve l

measurements from equipment that failed minimum reifability standards (#39, supra) were used to accept the valves as sufficiently thick (H.).

41.

Forty-two Data Reports disclosed that the valves were below the minimum thickness, but on the paperwork they were marked as " accepted" without explanation (M.).

42.

In 11 cases, the measurements were incomplete. The records simply skip results for required areas of the valve, such as the flat pad at the bottom (M.).

{

, ~,.. -.

_-y-

j

43. In 14 valve locations, there was no documented evidence that j

l the valves had been examined at all (M.).

44.. There was no documentation to indicate that weld repairs on the valves were controlled, as required by the AEC. To illustrate the absence of verifiable controls, the Data Reports do not have a requirement to list whether valves were weld-repaired, or the weld procedure used (,I_d.. at 7).

d i

45. During my research for UIA #34. I discovered that none of the valves meet AEC and PG&E design requirements. Westinghouse, the manufacturer, had explicitly declared that they "were not designed to meet the minimum wall thickness requirements of ANSI B16.5"--one of the relevant professional codes l

listed by the AEC in 1972. By comparing Westinghouse's communication with PG&E contract specifications, I learned that the valves also do not meet the design requirements in the contract (H.).

46. To my knowledge, there still has not been any corrective action on this problem.

If there had been good faith attempts I should have been contacted as the originator of the audit. I remain available to help follow through.

47. Similar to UT thickness measurement procedures, nondestructive test procedures lacked documentation of Procedure Qualification Records or Tests.

In IA #101, I found this flaw in seven procedures out of 21 examined.

Beyond the UT thickness procedures, there were five cases where no evidence existed that NDE procedures had been qualified. As a result, the quality of work examined under those procedures remains indeterminate. These included:

1) ESD 234, for UT Inspection of Groove Welds on pipe rupture restraints prior to 1979; ESD 241, for UT examination of Safety Yoke Rods on Safety

.wm..-,.--n,

-._n

,.e,,

i Valves; ESD 246, for Magnetic Particle testing, with unknown use; ESD 247, for i

Magnetic Particle examination of welds in the crack repair program on Unit #1 Steam Generator Feedwater Nozzles; and ESD 270, for Liquid Penetrant examinations, with unknown.use. On January 12, 1984 I completed and delivered to NRC inspectors, a draft report to Commissioner Gilinsky on IA 101.

It is i

enclosed as Exhibit 3.

48.

The corrective action for procedure ESD 234, consisted of unreliable.

"a fte'r-the-fact" Procedure Qualification Tests, whose use was not controlled and accomplished using qualified procedures.

Ironically, this is the same flaw the late PQT were supposed to correct.

Further, there is no evidence that management reviewed and approved the procedures for the PQT (M.. at 2-3).

4 9.. QA Manager Harold Karner improperly prevented any corrective action for the lack of procedure qualification records on ESD 270.

Instead, he directed that the Procedure Qualification Records for a similar procedure, ESD 210, should be used for ESD 270. That is unacceptable.

If the two pro-cedures have separate numbers, there are at least some dissimilarities. Those unique features of ESD 270 inherently will not have a proven demonstration of their ability to identify defects. This QA violation remains ignored.

50. No investigation' was perfonned to detennine where ESD 270 was used. Instead, the QA manager told me to just write up what I hadlearned already as an audit finding.

51. ESD 241 for UT of the safety valve yoke rods involves the most significant violations.

In addition to the lack of a PQR, the hardware was tested from December 17-20, 1973, before the UT. procedure it' elf'was even s

issued on December 26, 1973, and prior to approval of the UT procedure O

e 9

e

.p

i by PG&E cn February 12, 1974.

The testing was totally uncontrolled for the yoke rods on these valves, which I believe control the release of radiation from the containment (H., 8 at 4)

~52.

ESD 241 was deficient because it violated instructions from Dresser, the vendor for bolts and studs. The Dresser instructions required the rods to be examined prior to threading. At Diablo Canyon, the UT's wert i

i conducted after the threading. Further, ESD 241 did not use the Dresser instructions to determine the reference point for sensitivity and the criteria to report questionable items (M., at 4-5).

53. The existing documentation for the tests fails to meet the standards both of ESD 241 and the Dresser Instructions. Required information on the testing surface and instrument calibration was not included (J.,d,.

at 5).

d 54 Both ESD 241 and the UT insoection records failed to reflect compliance with a PG&E-imposed requirement for backup inspection "with the liquid dye penetrant technique to check the yoke rod ends for indications of cracking that might extend into the threaded area of the yoke ends" (H., at 5-6).

55. No DR was issued to PG8E on ESD 241, although this corrective action had been agreed to both by Mr. Karner and the NDE supervisor, is. J.arner improperly reneged on the basis of a memorandum from John Guyler, my successor as internal auditor. Mr. Guyler dismissed the detailed, documented DR which I had proposed with the following assertion: "PPP has accomplished thislper

. instruction from PG&E. It is evident that a nonconformance does not exist and a DR is not necessary" (J,d,,, at 3-4).

Mr. Guyler's response was inadequate.

First, the procedure violated PG&E instructions (see #54, supra). Second.

muee e, O

~

22 even PGSE does not have the authority to validly instruct Pullman to violate 10 C.F.R. 50, Appendix B, Criterion IX- "Special Processes." Third, Mr.

d Guyler did not document his asserted conclusion.

56. Overall, Pullman violated NRC reporting requirements and PG&E contract specifications by only reporting the deficiencies for two out of the seven ncndestructive procedures to PG&E on Discrepancy Reports (Id,.. at 6).

57.. PG&E dispositioned the DR for ESD 246 " accept as is"..although there is no information indicating where the nondestructive test was conducted.

Since the identity of the affected hardware could also impact on the evaluation criteria, PG&E's acceptance was premature (ld., at 7).

58. The reason the locatien of work tested under ESD 24G could not be identified is that Mr. Karner improperly prevented me from looking. After I learned that ESD 247 was used for welds in the crack repair program on feedwater

- nozzles in the Unit I Steam Generator, he ordered me not to check where ESD 246 -

had been used (Id., at 6).

59. PG&E improperly dispositioned the DR on ESD 247 " accept as 'is",

although the Magnetic Tests in the procedure were referenced to ANSI standards, rather.than the relevant ASi1E Code Section I; and although the qualifications of the MT personnel conducting the test cannot.be verified from the records

~

available (Id.).

60. The corrective action for ESD 246 and 247 involved procedure qualifications after-the-fact (ld., at 7). After-the-fact procedure qualifica-tions should not excuse PG&E from accountability under NRC rules. At best, it means that the damage has been minimized. But it also inherently means that t

9 m

e e

-v y

r

--w.,.,

n,..

---

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

.n._,.

.n

,.,-,,.n.

e n,,.-m

~

23-10 C.F.R. 50, Appendix B, was violated, because special processes were con-ducted under uncontrolled conditions.

61. Even if it is acceptable to conduct procedure qualification tests after the fact, the tardy test must be performed under controlled cir-cuestances. In this case PQT's were conducted with different equipment than had been used originally Qd.). No documentation was supplied to support the asserted Corrective Action Response that the new equipment made the results more conservative.

62.

QA Manager Karner was responsible for the deliberate failure to provide reasonably prompt corrective action for IA 101. On January 18, 1982, I initially disclosed IA 101; on March 23, 1982, it was finalized after I provided Mr. Karner with additional information which he had requested. On April 6,1982, corrective action for the first finding in the audit' on lack of procedure qualification tests was approved.

Be, fore implementation, how-ever, he changed his mind. Although the official time limit for corrective action is ten days, the audit was not closed out for over another year, despite my repeated memoranda and attempts to formally notify Mr. Karner of his obligation to address the issue of unqualified NDE procedures (Jd., at 8-11).

)

l

63. Pullman corporate QA Director A'. Eck was notified of the failure to take corrective action and improperly refused to help. Instead, he j

reprimanded me for bringing the matter to his attention. On June 14, 1982, I notified Mr. Eck, through an Interoffice Correspondence, of the overdue corrective action. He did not respond. On July 6,1982, I performed and submitted Unscheduled Internal Audit #31 to Mr. Eck on the lack of corrective j-action' required by ESD 263 within 10 days.

This time I received a response.

~ Both Mr. Eck and Mr. Karner reprimanded me for submitting the audit to Mr.

Eck directly, rather than letting it proceed through the chain of command.

-e,--

,,_.,y.

.. ~, -,.

A

^ --

m

~

This violated ESD 263, they explained. My audit was voided. Both individuals neglected to mention the violation of ESD that I had raised -

the QA violations were not getting fixed (M., at 9-10).

64 In January 1983, I was further punished for Mr. Karner's improprieties. I was renoved as internal auditor because only 5 instead of the required R audits had been closad out. Part of. the problem was due t o circumstances -

beyond my control. Mr. Karner or supervisors-were sitting on some of my audits beyond the required deadline. Mr. Karner also was loading me down with ancillary assignments and unscheduled audits were not counted.

65. On January 28, 1983, during the meeting in which Mr. Karner threatened to get rid of me for looking at quality -related issues without being assigned (Suora, Nos. 27-28), I informed Mr. Karner that he had violated 10 C.F.R. 50, Appendix B.

He responded twice that we are not committed to 10 C.F.R. 50, Appendix B, and that it was "0.K." for him to violate the Code of Federal Regulations and related contract specifi-cations.

III.

BREAKDOWN IN 00ALITY ASSURANCE FOR HYDROSTATIC TESTS.

Hydrostatic testing at Diablo Canyon from 1975 to 1978 does not have the necessary QA. documentation to prove the reliability of the tests.

In hydrostatic tests, water is run through the plant at higher pressures that. normal to see if the piping is reliable.

In February.1981, I conducted Interna 1' Audit 86, in which I learned that nearly all hydnistatic piping tests for a year, during 1980 and 1981 were conducted wtthout required QC documentation. -In April 1982 NRC inspection identified that documentation problems identified

,n,-

m

,e.,

+

~

25-

'n Internal Audit 86 were _ not properly--corrected. I became convinced i

that,

serious problems may exist with the hydrostatic tests.

In March 1983 I completed Internal Audit 106, which examined the records for 79 original hydrostatic tests and 118 retests conducted from 1975 onward. I learned that the test documentation did not have evidence of required QC oversight.

QA records, consistent procedures, or controlled test conditions. In short, i

there-has been a generic breakdown in the QA requirements for hydrostatic tests. They must be redone.

Internal Audit 106 is enclosed as Exhibit 4 My specific allegations follow.

66. The procedures for hydrostatic tests conducted before January 27, 1975 are fundamentally inadequate, due to their failure to

-include documentation requirements, and due to lost pages, the inability to even entirely reconstruct the procedure requirement.

67. Almost all hydrostatic tests and retests from 1975 onward lack required QA documentation. The most significant omission involves QC coverage documented on a piping system closecut - F98 Department Release. This activity is necessary to assure that departments performing the test comply.with procedure checklists. Unfortunately,

~

departments only complied sporadically with the requirement to complete and maintain the form whichchmonstrates compliance with the test pro-cedure.

In other cases, there is not necessary backup documentation to verify the conclusions in _ the release.

(Exhibit 4. AAR #1).

68. From December 1977 - April 1978, in 28 cases Pullman test requirement forms did not have information necessary under the

.~.--.-m v

26-

)

procedure ESD 229. Fundamental data, such as the type of fluid, pressure and temperature, simply is missing (g., AAR #2).

69. 'In 28 cases,. Pullman's NT procedure data form does not match PG&E requirements. This form is the guide used to conduct the test, so the distinctions translated into different test conditions that disqualify the results from Pullman's hydrostatic test. To illustrate, in one test Pullman's procedure only had a pressure of 2485 PSIG, when PG&E's acceptable minimum was 2812 PSIG.

70. The absence of backup documentation continued after 1978.

From March 1978 to April 1980, there were 14 hydrostatic retests without a signed QC field pipe release, dispite the conclusion by Quality Engineering' in the test records that QC had verified the results (H. AAR #3).

71. The problems with hydrostatic tests offer another example of management harassment of QA personnel.

During the May 1982 NRC inspection, I spoke extensively with. NRC representatives. After the interview Mr. Karner expressed anger at' the length of the meeting. At a later meeting, during this general time frame, he threaten to get rid of me.

IV.

BREAXDOWN IN VENDOR QUAT.ITY ASSURANCE.

Although I as not as actively involved with vendor QA as with special process and hydrostatic test procedures, I observed the symptoms of a generic QA breakdown after becoming familiar with two exampes of QA violations involving vendors. One case involved a vendor that calibrates micrometers, a precision measuring device for Pullman tools and the impact of weld repairs, among other functions. Although the vendor had a clean bill of health and was on the Approved Vendors

-e+---

, m, eye v

-~i--*.yIw,m

+=

m v

,=

,-,*.-w-- - - - - - + - -

t.

27 List (AVL) until my October 1981 audit, there was virtually no quality assurance program. Unfortunately, corrective action was solely prospective - to remove the firm from the AVL. The damage that already has been done will remain.

The second case involves 1980 and 1982 orders by PG&E for Pullman inspectors to stop reporting the large number of cracked shops welds found in Boston Bergen and American Bridge. work.These hardware defects should have been reported on.DR's, but instead were ordered to be ignored because they came from a vendor. Specific allegations follow.

72. The reliability of Pullman's Approved Vendors List is indeterminate, due to the inclusion of Microsurface Engineering. This firm only had a token quality assurance program, yet had been approved' and passed previous vendor audits. My audit demonstrated that Microsurface did not conduct audits, did not have a written procedure for calibration, conducted uncontrolled inspections, lacked traceability for use on Pullman tools, failed to disclose laboratory standards for calibration, and did not have required documentation for training of laboratory personnel.

The violations were so ingrained and pervasive that it is not credible to conclude they only sprang up since the ve'ndor passed an audit the previous year.

73. Corrective action for the Miscrosurface QA violation improperly was restricted to the prospective step of removing the firm from the AVL This was inadequate, because the accuracy of measurements

~

made with Microsurface tools is indeterminate. The effects of previous vio-lations will remain undisturbed.

..I

- = -. : _--., -..:m e _=

a. : --.:::: : u r : --

74.

In July 1979 Pu11ran inspectors began finding signifi-cant quantities of cracks in welds received from two vendors, Boston Bergen and American Bridge. Until 1980 Pullman inspectors wrote 19 Discrepancy Reports on the welds, which displayed a consistent pattern of

~

2.5 linear indication. The DR'.s are enclosed as Exhibits 5-)(.

On April 3, 1980, however, Mr. Marvin Leppke of PG&E issued a memorandum directing Pullman to stop issuing Discrepar.cy Repcrts on these " shop" welds. The memorandum is enclosed as Exhibit 75.

In 1982 PG&E repeated the improper restrictions on QA enforcement against the same shop' welds. This time PG&E instructed Pullman to delete shop welds from the formal walkdown program that represents a final visual check on quality. Relevant supporting documenta-tion is enclosed as Exhibit )&

V.

RECORDS FALSIFICATION Beyond instances of contradictory and impossible information in the records, in some cases I am sufficiently familiar with the cir-cumstances of false records to state that they were intentionally fal sified. Examples involve the qualifications tests for QC inspectors.

As a prospective welding inspector I failed one of sty initial test and was then given a copy of the test to study to assure passing on the second attempt.

\\

Another inspector was certified after taking-e test which upon review months later he:was found to have failed. He was ratested at that time and passed with the assistance of coaching. The test was backdated to the original test date to cover work performed during the intemin period. The latter example occured in 1980.

O e

I

~~.

I.

-._...[_.....,__._.,[.

$ _.l _.,

$.b.[

.,,. ((._ i.

r-29-

• ~

VI. CAUSES OF THE QUALITY ASSURANCE BREAKDOWN.

77. The most significant cause for the QA breakdown is the environment of repression and the predictable retaliation against QA personnel who diligently try to identify and correct QA violations. The problem goes well beyond the loss of organizational freedom. Upholding the Atomic Energy Act at Diablo Canyon can represent professional suicide.

Most significant, the sacrifice is for nothing. The violations remain, uncorrected. My own experience is a case study. Mr. Rarner threatened to "get rid of" me on three occassions when I persisted in attempts to obtain corrective action. 'Mr Karner restricted my freedom as an inspector untti I could only look at specific problems assigned by him.

I was reprimanded, verbally and in writing, for communicating with corporate QA management about such a fundamental violation as the failure to take corrective action against unqualified NDE procedures on safety related work. To add insult to injury, in January 1983 I was demoted for not finishing enough a ssignments. The demotion was due in part to Mr. Karner's refusal to act on my audits, which made it impossible in some cases for me to finish my assignments.

78. The final act of reprisal against me occurred on January 13, 1984 I was laid off from ley.iob as a ~pipefitter, the day after making my third disclosure to the Nuclear Regulatory Comission. NRC inspectors already j

had told me that site management had a copy of my first report on welding procedures, and that Bechtel 'was studying it. On Friday, 50 pipefitters were l

1 aid o.ff, supposedly due to a lack of parking space. The usual practice i

for these layoffs is to let workers from the local union stay until last.

In this instance 4 6 out of the 50 employees laid off were " travel cards"

__-___m._-----%.,--

--a-

-m..a., - - - - +

--v

-i-- e ai

- - + -< *

-~ -

1 30 from out -of-town unions. Although more travelers were available, four employees from the local were swept out with the travelers. One of the four was having conflicts with his supervisor and one had an absenteeism problem.

The other two were my partner and myself. My foreman protested to the super-visor not to lay off my partner and me, and asked for permission to pick someone else. The supervisor referred him to the resident construction manager, who refused the request and told the job steward that we had to be the ones laid off. My foreman and the job steward recounted these events to me on the day of the layoff. That day the job steward also infonned me of the perception of site that my layoff was due to " politics" and was decided " higher up".

On January 25, 1984, the day after retaliation was widely discussed at Congressional hearings, management called me back to work but not my partner.

The pattern represented by my case illustrates why a significant number QA violations have gone unreported, and why the quality of Diablo Canyon is indeterminate.

Those who persist in reporting the violations are dismissed, or harassed relentlessly until they resign, or give up and stop trying.

79. Another cause for the QA breakdown is subordination of PG&E's and Pullman's QA department to construction. Until recently, PG&E site QC did not review Pullman Discrepancy Reports. PG&E's Resident Mechanical Engineer, a con-struction offical, reviewed and approved corrective action to discrepancies. As of May 1983, Pullman Internal Audits were not submitted to PG&E site QC for review but instead submitted to the Resident Mechanical Engineer.

80. Another cause for the QA violations was lack of resources. To illustrate, from August 1980 to September 1982 Mr. Karner was the only pennanent employee in the QA/QC site management. He did not have an assistant QA Manager, and the QC Supervisor was a temporary enployee.

81. The QA breakdown was not due to PG&E ignorance. On

• 6 4

1 repeated occasions, I identified many of' th'e issues in this affidavit to a variety of officials within the PG&E supervisory and sanagement staff..

Although some officials listened and expressed agreement and/or sympathy, none of the violations were corrected. I believe that PG&E and Pullman have been gambling that the NRC will not enforce the QA laws, even if they are caught.

For the sake of the public's health and safety, I hope that the NRC calls their bluff.

I have read the above 31 page affidavit, and it is true, accurate and complete to the best of my knowledge and belief.

Sh Id Hudson

\\

SUBSCRIBED AND SWORN t is '

' day of January,1984, in California.

3 NOTARLPUBLIC

' mission Expires:

e

.e-o -