ML20069N334
| ML20069N334 | |
| Person / Time | |
|---|---|
| Site: | LaSalle  |
| Issue date: | 11/30/1982 |
| From: | Devine T, Garde B INSTITUTE FOR POLICY STUDIES |
| To: | Harold Denton Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8212020358 | |
| Download: ML20069N334 (20) | |
Text
,
(202)234 9382 1901 Qu2 5treet, N.W., Washington, D.C. 20009 November 30, 1982 Mr. Harold P. Denton Office of Nuclear Reactor Regulation Division of Licensing U.S. Nuclear Regulatory Commission S
Washington, D.C.
20555 RE:
C.F. Braun Independent Audit LaSalle Nuclear Power Plant
Dear Mr. Denton:
On behalf of our clients, Mr. Albert T. Howard and Ms. Sharon Marello, the Government Accountability Project (" GAP") submits the following report to the Office of Nuclear Reactor Regulation ("NRR").
This report supplements our November 19, 1982 submission. We recommend that you refrain from issuing a full-power license to the LaSalle Nuclear Power Station (Unit I) in LaSalle County, Illinois on the basis of the October 27 report submitted by C.F. Braun and Company ("Braun report").
Instead, we urge that you additionally consider other more specific ree-ommendations in the body of this report.
GAP is a non-profit, non-partisan public interest organization that assists federal and corporate employees who report illegal, wasteful, or improper activities by their agencies or organizations. GAP also monitors governmental reforms, offers its expertise about personnel issues to Executive Branch officials and agencies, responds to Congressional requests for analysis of issues related to accountable government and disseminates significant information about problems to appropriate offices within the government.
Our requests and recommendations are based on a detailed staff analysis of the C.F. Braun report, the review anc analysis of the Zack witnesses who first exposed this problem to the NRC and the public, and We have statistical analytical review by several expert consultants.
incorporated their comments into the body of this report.
Boo /
Our review of the recently issued C.F. Braun assessment of the Heating, Ventillating, and Air Conditioning System at the LaSalle Nuclear Power Station indicates that their assessment was neither statistically valid, nor judgementally sound. Further, that the conclusions reached by Braun largely do not correspond to the facts presented in the raw data, and that the raw data itself has large credibility gaps. In short, the Braun Daf GUidB2 O2IGINAL
. m / ff n.a u-b y g
~~
F212O20358 821130 PDR ADOCK 05000373 P
rcport cc it ctanda doe 3 nst Gvan cucceed ca a p;perwork cn:wir to a pap 3r-Quita cimply, it fcils to offsr a rcaponta to our clisnt's j
challenge that io cufficicnt to rscolva tha concsrns thtt prsvicusly led work problem.
In short, the Braun r4 port to postponement of full power authorization.
fails to answer the safety-related questions raised by Mr. How sing conditions, this report fails to provide the grounds for full power Marello.
authorization.
t I.
ZACK ALLEGATIONS _
Backfround In the f all of 1981 the Chicago based Zack Company, a Heating, contractor, hired Ms. Marello, Ventillating and Air Conditioning ("HVAC")
Mr. Howard, and a number of other individuals to establish a Quality Assur-ance Documentation Control office.
Zack Company had a Documentation Department that complied with 10 C.F.R. 50 oodes, and Appendix B, the American National Standards Institute (" AN i
d s Their specific assignment was to control the documentation -- purchase
("P.O.'s"), material certifications This involved the
("M.T.R."), and certificates of conformance ("C.C.").Each package represented monitoring of over 3000 purchase order " packages."
h the proof of quality for up to thousands of items used to construct t e Clinton, LaSalle or Midland nuclear power stations.
Mr. Hovard was hired as the Documentation Control Rocm Supervisor in the fall of 1981.
Ms. Marello was a clerk. They, and the three or four i
d other Documentation Control Room employees were allowed -- in fact ass gne Their task was to
-- to investigate documents contained in Zack's files.
d verify the accuracy, or identify the inaccuracies to the purchase l
them in a good position to observe the " paperwork trail" of Zack's nuc ear packages.
documents.
From approximately November 1981 to April 1982, Ms. Marello and Mr.
Howard discovered and challenged a quality assurance Their experience reflected a contractor plant sites in serious question. operating for years without regard for the l
the public health and safety.
In their employment with the Zack Company they discovered documenta-i i ertif-tion that had altered specificiations, some with " white out,"
d ith to reflect correct quality assurance approval, and adhesive stickers w t
questionable authenticity, used to modify documentation a bility vendors -- with some success -- to provide inaccurate quality and tracea s.tandards.
l certifications after-the-fact.
l l
Both Mr. How.rd and Ms. M rsllo worksd in th2 Docum:nt Control Room Although th;y rsesivsd no formalizsd Quality Angur-ance training, they, and the rest of tha QA dspirtment, did become familiar ct the Z ck herdquart:ra.
dly with the various codes, contract specifications, and regulations that allege governed their work.
Subsequent to their challenge of the QA breakdown Mr. Howard and his As Zack strove un-staff suffered a pattern of harassment and intimidation.
i successfully to meet unrealistic time demands imposed by Commonwealth Ed son Unfortunately the pressure and subcon-("Com Ed"), the pressure increased.
t tractor " rush job" has caused more time delays.
In Auguct 1981, Zack had notified the utilities of a potential non-I conforming condition under 10 C.F.R. 550.55(e), due to inadequate and inac-They also curate quality and identification records on vendor purchases.(" CAR") plan which o attached a Corrective Action Report to identify, analyze and correct all the paperwork problems at the company This CAR also outlined the proposed steps Zack would take to insure that the proper individuals responsible for this were appropriately headquarters.
disciplined.
As pressure mounted in early spring of 1982 to have the LaSalle nuclear plant load fuel, the QA department at Zack fell under greater pressure tothat detaile close out nonconformance reports ("NCRs")Mr. Howard regused to provide a final report tation deficiencies at LaSalle.On March 1, Zack submitted 99 remaining NCRs to Com Ed.
Zack warned it was unlikely that necessary documentation to correct deficiencies to Com Ed.
This frank admission did not deter the utility and NRC In April 1982 Com Ed received permission could be obtained.
rush to begin operations at LaSalle.
to load fuel.
Mr. Howard, acting as a spokesman for the On April 13 and 15, 1982 contacted an individual in the entire Zack Quality Assurance department, This Consumers Power Company's Midland Project Quality Assurance department.
individual had represented to Mr. Howard and other members of the department k to that they should feel free to bring any allegations or problems at ZacHe also gua I
Midland's internal grievance system.
and protection from losing their jobs.
On April 18,19 and 20, an aud'.t team from Consumers and the Bechtel The QA departm at anticipated a Corporation arrived in the Chicago office.
However, on complete investigation and professional support for its effort.
April 30 the entire department was dismissed, On May 3, 1982, the first working day following the purge, Mr. Howard He provided specific allegations f
began a series of contacts with Region III.about LaSalle and to a lesse i
offer of full cooperation with an NRC investigation. months, when Mr. H k
to respond to their allegations, which had cost them their jobs, they too After 2 their information to the press and then to gap.
at a public meeting between the NRC, Com Ed, the On July 19, 1982 i
Illinois Attorney General's office, and representatives of the public includ ng e
D 9
GAP Mr. Jame3 Kepplir, Regi n III AdminktraiE8, c11cgitirn3 had not be:n pur;uid in thm cpecici invctitigntion ong thtt "...it 19 our conclu3iin that no rcr:on Gxiits to L:Sallo.
Zack cliegnticn3 --
preclude the LaSalle Unit 1 from going beyond zero power."
One week later, the NRC Commissioners overrode Mr. Keppler's rec-Following receipt of Mr. and Howard and Ms. Marello's July 26, 1982 whistleblowing disclosure which included three affidavits, numerous memo-ommendatiora.
l t randa, and attachments -- the Commission delayed licensing the LaSalle p an i
until the NRC staff had a more accurate assessment of the Zack allegat ons and their safety implications.
cover letter to NRC Chairman Nunzio Palladino, GAP In a July 26, 1982 described the efforts of the Zack witnesses, the Quality Assurance breakdown I
GAP also at Zack, and the evidence supplied by the fonmer Zack employees.
b outlined the inadequate NRC oversight that severelyucompounded the HVAC pro -
Finally, gap presented three specific requests to lems on all three sites.
the NRC Commissioners" Require the NRC to conduct a full investigation of the whistleblowers' evidence before permitting full power operations at the (1) facility.
("OI") to replace Require the Office of Investigations (2)
Region III in its ongoing investigation.
Require the Office of Inspector and Auditor ("OIA")
to f
investigate the performance of Region III's Office of Inspection and En orce-(3) ment.
Unfortunately, the recommendations we identified for immediate action The Commission conditioned license approval on a positive were not followed.
evaluation of investigations into Mr. Howard and Ms. Marello's disclosure that led to the Braun report.
In light of the C.F. Braun audit policy in monitoring the third party effort.
this may have proven to be a mistake.
l Further, although OI has not yet completed its work, we are skeptica The Region III OI investigator, Mr.
of the framework for its investigation.
IE James Foster, conducting the Zack investigation is one l
Mr. Foster still is Although the acronyms are dif ferent, LaSalle.
providing a second opinion to his own previous exoneration of Zack at complaint in May.
I l
I Finally, we are unaware of any OIA investigation into Region III IE s failure to pursue serious health and safety allegations.
GAP representatives, Mr. Howard and Ms. Marello, d C.F.
On August 11, 1982 inter alia, participated in a Region III meeting in which Com Ed propose l i dis-Braun and Company ("Braun")rective action program on design implications fr closure.
13,19821Gttir wa prszntsd our com;nts on the propwe 24, 1982, wm gubmitted In an Augurt Afttr a c5cond meeting on AuguttOur concarna of August and Oriun cudit of LaSn112.
4, 1982 1sttir.
i l w:ra wall-furth a comments in a SIptember September about the flaws in the proposed Craun racsmmtnt obv ous y justified.
On November 9, 1982 we received a four volume report of the C.F.
Salle.
Braun Company's independent analysis of the HVAC system at La GAP Interim Report summarized, and in pome cases In our November 19, 1982 A November 19, 1982 provides detailed justification, for our concerns. Commission Office of Nuclear Reac tor letter we urged the Nuclear Regulatoryto take a number of specific actions in response to Regulation ("NRR")
B Braun report as submitted:
Require CECO to recall C.F. Braun to continue with the (1)
LaSalle HVAC review.
Restrict the LaSalle license to 48% power.
(2)
Request Region III to consider enforcement action against (3)
CECO for failing to supervise subcontractor work.
vendor inspector to audit / review (4) Assign a Region I?
which disregard Braun's conclusions in light of specific C.F. Braun statements 10 C.F.R. 50, Appendix B.
In particular, we take We reinforce those r,equests in this report.
lle to go to full power serious exception to any consideration of allowing LaSa S lle. Our current prior to a complete resolution of the safety concerns at La a analysis follows.
II.
SUMMARY
OF BRAUN FINDINGS Although the Braun report found that Zack's alleged quality assurance that conclusion
("gA") breakdown did not have a significant hardware ef fect, Overall, Braun does not match the detailed findings in the body of the report.
")
CECO agreed to repair 46 items, and twenty-four Sargent and Lundy ("S&L The report also drawings had to be revised to match the as-built condition.
1982. The noted that five welds out of 65 failed visual tests on September 1, report noted "some" incomplete welding qualification record deficiencies.
(Report, at 7, 18.)
They understate the extent These overall statistics are misleading. Initially, our review of Braun QC-1 and rate of identified deficiencies.
ided or inspection reports found that 28 out of the 335 inspections were voIn an cancelled completely.
h partially completed because of inaccessibility, due to location or hig l
NRR 117 incpection raports that citid dasign Further, th]
radiation 1sv313.
violations involved 193 individual discrepancies.
Sixty three of the discrepancies involved missing or defective The inspection reports generally did not provide specific quantifi-But if each example of welds.
cation for the number of welds in each violation. faulty " welds" only welds were missing or deficient.
The body of the report concludes that seven Zack welding i h they were-acceptable.
the seven procedures were not qualified for all the positions in wh c were used.
Appendix C quantifies what Braun meant by "some" missing w qualification records.1982 memorandum in Appendix C shows 23 cases out of 113 whe l
performance qualification tests are not acceptable. memorandum in Appendix C on 1 l welders ment to an October 26, 1982 demonstrates that 12 position qualifications were unaccepta qualification records for all the positions which they had previously been reviewed.
approved to perform.
a more accurate summary is that Braun found 193 As a result, Two discrepancies in 117 out of 307 relevant inspections, a 38% rate. In 35 welder procedures out of seven did not qualify unconditionally. c for relevant procedures, a 21% rate.
We have still understated the significance of what Braun found.
For example, Braun switched its reporting system in the middle and stopped reporting violations for uncoated welds and missing welder identificat for specific hardware.
were defined -- out of the conclusions, we found 16 cases where there was no welder identification available for the material covered in the 335 originally-scheduled inspections, or about a 5% rate where the welds cannotBrau be traced back to a particular craftsman.
determine how much the as-built condition varied from the original approved l
design, because so many of the earlier revisions to drawings are missing.
(Report, at 24.)
To say that these findings are not significant strains credulity.
To illustrate the significance of Braun's inability to confirm the qualifi-cations of craftsmen on 23% of the welding procedures, a similar case At an October 28, occurred at the Zimmer plant with 20% of the welders.
1982 public meeting Commissioner Gilinsky asked Region III official Dorwin Mr. Hunter replied, "It Hhnter if that failure rate was " highly unusual."At Zimmer the finding sparked NRC actio would be absolutely unusual."
m
n y
Braun, by contrast, dismisssd tha Entira requira a maneiva remedial program.
concept of unverified welder qualifications as insignificant.
What is particularly unfortunate at LaSalle is that it is not even possible to draw reliable conclusions on the work performed by suspect welders, since some 5% of the weld inspections could not be traced to iden-
- Significantly, ten of the 15 actual inspections that tifiable craftsmen.
involved unidentified welders also led to findings of welding violations.
A survey of examples in the Braun report appendices remokes any Examples doubts as to the seriousness of HVAC hardware flaws at LaSalle.
where the as-built condition failed to match the design included-- items I
(QCl-78, 245); a hanger three feet away that were of insufficient thickness from the location on the available drawing (QCl-219); missing nuts, washers, and improper bolt spacing door plugs and plates (QCl-88, 224, 238, 318, 335);
as well as numerous instancas where the work was done and projection (QCl-120);
in the wrong location or the drawings were not current.
Welding hardware problems included-- missing welds (QCl-29); partial welds and burn holes (QCl-177); cracked welds (QCl-25, 65); and welds with corrosion and porosity defects (QCl-ll7). To illustrate the conservative nature of our statistical summary to quantify r?.e number of bad welds, the QCl report disclosed that "most" welds in particular inspections had excessive corrosion (QCl-25) or porosity defects (QCl-ll7). One inspection report tnat did specify numbers identified ten cases of cracked or poor quality welds, I
with at least one and possibly two leaks (QCl-295).
The discrepancies were not limited to welding and inaccurate design To illustrate, Braun found air leaks, including one around a problems.
valve (QCl-190); a hole i-a duct (QCl-228); and eight holes in one inspec-tion that had been " repaired" with tape (QCl-38).
It is particularly important that CECO's QA program had missed all f
Yet it was ready to vouch for the quality of the of these discrepancies.
HVAC system as well as to recanmend operating LaSalle at full power.
While Braun and the utility denigrate the significance of these findings, in our opinion they evidence violations of six out of 18 quality II (personnel qualifications);
assurance criteria in 10 CFR 50, Appendix B--
III (design control); VI (Document control); XVI (corrective action); XVII (qaulity assurance records); and XVIII (audits). In short, to deny the significance of the discrepancies Braun identified is to deny the signi-ficance of 10 CFR 50, Appendix B.
Braun's empirical findings (if not its conclusions) strongly support our call for NRC enforcement action against i
CECO.
e I
1,.
_ _...._ -.m
.. -., -. ~ _,... _. _ _. -
..m-,___~
~
-U-NRR J
III. FLAWS IN BRAUN REPORT _
The findings summarized above are too serious to perinit full power There is no question, however, operation without further investigation.that the report conclusions are flawe estimates both the scope and safety significance of the hardware defects.
We have organized our critique into six areas-- 1) lack of informational
- 3) faulty methodology independence; 2) lack of institutional independence; incomplete disclosure,of results; for the scope and nature of fact-finding; 4)
- 5) suspect subjective evaluatione instead of conclusions supported by authori-failure to verify all necessary corrective actions.
tative citations; and 6)
E Our specific analysis follows.
Lack of informational independence.
A.
4, 1982 comments on Braun's proposed program, we In our September emphasized, The point of this proposal is to provide a " fresh" look at the facts -- not to publish CECO's inspection r.nd We do not contend test results under Braun's signature.
that NRC-sponsored laboratory tests should be duplicate 3.
But the results of previous CECO self-inspections should have no more than background significance.
Our emphasis on the necessity for Braun to provide a " fresh look" at Zack's 24, 1982 public work was consistent with Mr. Keppler's remarks at the AugustMr. Keppler further 4 comments were addressed.
meeting to which the September Obviously indicated that all of Zack's work should be considered suspect.
that precaution applied to the Zack paperwork, whose reliability is the reason for the Braun project.
The Unfortunately, Braun trampled on this suggested premise.
" independent" review in many cases was no more than a resunnary of CECO and In fact, CECO overruled Zack's previous factfinding reports and tests.
Braun's own factfinding on the basis of reinspections conducted by the In reality, the Braun report's utility's own construction department.
A more conclusions do not even provide an independent factual record.
accurate characterization of the report is that it contains the factual Braun's own factfinding was conclusions of Braun as accepted by CECO.
preliminary at best, and subject to veto by the utility.
Braun relied entirely on the NRC and CECO for material tests 1)
When materials failed one CECO of the HVAC ductwork and supports.
the utility tried again and passed the materials without ex-(Report, at 7, 14.)
As a result, the " independent" organi-
- test, The ception.
zation did not conduct any material tests, or even the retests.
necessity of the third-party independent hardware tests is due to CECO's own previous inability to maintain adequate quality records.
e
una Crcun dafaultsd en thic braic foundstion of tha program and turn d it ovsr to the utility.
Similarly, Braun did not conduct independent verification of The third party wcs satisfied 2)
CECO's pre-operational test results.
There is no reference to with a paperwork review of CECO's records.
=
any relevant field inspections to doublecheck the results on paper.
(Report, at 22.)
Braun even relied upon tests conducted by Zack and its own For 3) subcontractors, again limiting itself to a paperwork review.
instance, Braun's investigation into leak rate and air flow balance was lbmited to restudying the tests conducted by Zack's own subcontractor.
g In other instances, Braun used Zack's measuring (1d., at 5-6, 20-21.)
equipment or CECO's personnel to actually conduct the inspections and (See QC2-88, 89. )
tests.
Braun apparently did not even collect its own records, or As doublecheck to see if Zack truly provided all that was requested.
4) the third party was content to report that six welding pro-cedure specifications and 91 welder qualification records were " purported" a result, to be all applicable documents.
(jyl., Appendix C.)
In fact, the summary of the Braun report refers to seven welding procedure specifications.
In the end, Braun did not confirm that its review even (Report, at 7.)
covered all the relevant Zack records or whether any were missing --
apparently because Braun didn't check.
CECO's control of the final factfinding conclusions is illus-The utility 5) trated by the disposition of Braun inspection reports.aen supported by attached draw-rejected Braun discrepancies that hadings, by referring to other dra This means that the (See, e_.g., QC2-64 and CECO response.)
utility not only dominated the final factual conclusions, but left its report.
empirical vetoes unverifiable.
In a number of instances CECO did not even turn to its own The utility turned the 6)
QA department to overrule Braun factfinders. job over to its own c d
were accepted as final without including the relevant reinspection r.nar s.
(See, e.g., QC2-27, 55, and 75.)
In the most crude example, CECO even relied on Zack's previous damage reports to reverse Braun's inspectors on the extent of Zack's 7)
CECO rejected one Braun observation, because Zack's records i
showed that reinforcement bars were only nicked, not severed, and ni damage.
will not affect the strength.I/
1[This, of course, dilutes the standard used this July to assess the structural effects of nicked reinforcement bars outside the HVAC system, in response to charges leveled by the Illinois Attorney G f
i for its calculations in the absence of new tests to confirm the facts.
NRR Zack's rIcorda w;ra used ts the final entwrr to mad 2 tha full cyclo.
safety questions r2 iced dus to tha qu;3tionable recurtcy of Z ck'c j
l records.
Lack of institutional independence.
F B.
CECO's organizational domination of Braun campromised the independence In our Sept' ember 4, of the review as much as its informational domination.1982 comm We also warned independent judgment with respect to S&L design decisions.daa g
assistant if CECO totally controlled the technical evaluations of Bratn s findings.
That is just what occurred, especially for the Braun quality control They inspectors who actually conducted the primary factfinding work.
d CECO were subject to an illegal gag order not to reveal what they learne.
("QC")
added further pressure on the third party by auditing Braun during the course In the 34 cases where Braun observations were overruled by SEL design changes or CECO f actual rejection, Braun accepted the vetoes witho of the review.
the final conclusions represent the judgments of CECO and question.
In short, S&L, not Braun.
Initially, the groundrules for the project violated the 1)
Braun per-independence of the personnel who conducted the review.so information revealed" during their work as " confidential" and not to disclose it to anyone "except as directed by the Project Manager."
This means that Braun employees have agreed (Report, Appendix H.)
not to reveal illegalities or significant findings that may have been To the extent that deleted from the final report -- even to the NRC.information s gation, its enforcement would be illegal.
any employee who asserted his or her legal independence under the Atomic Energy Act could lead to severe sanctions, including fines even license revocation.
(July 14,1982).
CECO violated the basic groundrules of the project by CECO's action was an undisguised 2) auditing Braun during the review.
attempt at organizational bullying, and gave the utility premature As Braun stated, access to raw Braun data before the results were in.
At the meeting on August 17, it was clearly established that the C F Braun site program would not be subject to This was audit by the CECO Site Quality Assurance Group.
agreed upon so that the C F Braun independence would not be compromised.
Nevertheless, on September 10 CECO conducted whathi (Report, at 27.)
Braun has described as an " informal" audit for conformance with tec Somehow Braun simultaneously concluded that cal program requirements.
- wus, f-tha cudit did nst cov r "tha Brtun proj ct End QA program."
(Id.)
That cc artion 10 unv:rifiablo cinca noni of th2 CECO Eudit documints This organizational interference is in-are included in the report.
If Ceco had legitimate concerns, it should have gone to the excusable.
NRC rather than taking matters into its own hands. At a minimum, the NRC should require CECO to emplain its action, to disclose the results and Braun's response, and to produce all relevant records from the audit.
Braun chose not to question aunt nstance where S&L responded i
3)
For instance, on
, to discrepancies by changing the design requirements.
October 18 in one sweep S&L eliminated 14 Braun observations about clips
- in the wrong locations with a design change that allowed the clips to be Braun did not comment.
l located almost anywhere on the relevant item.
(Report, at 7.)
Similarly, when Braun found a generic deficiency where certain tubing was 25% too thin (3/16" instead of the required 1/4"),
Sargent and Lundy responded with a generic design change allowing all the tubing to be as thin as it was in fact. Although it had originally deemed the discrepancy as significant for LaSalle's safety, Braun again concurred without comment.
(Id., at 8.)
Braun's decision to sacrifice its independent judgment to that of Sargent and Lundy is particularly inappropriate, based on S&L's record At that plant its engineering judgments endorsing the status at Zimmer.
quo have been overruled repeatedly by the American Society of Mechanical Engineers (" ASME" ) and the NRC. The judgment flaws were so severe that last November NRC Region III called for a generic evaluation of S&L's (See IE Report 50-358/81-13.)
work by the NRC's vendor inspection team.
26, 1982 enforcement conference, S&L representatives At an October repeatedly disagreed with angry NRC inspectors who reported that S&L was designing the plant af ter the fact on the basis of " trial and error" to justify the existing construction deviations.
Braun also sacrificed its independent organizational judgment 4) to CECO, again without question. Examples where Ceco overruled Braun's judgment without challenge include observations against-- two welds with cracks and porosity (QC2-06); a bolt that is over 254 shorter than called for on the drawing (3" instead of 3-13/16"), on the basis of a (QC2-24); a hanger that CECO reported previous nonconformance report was stronger because the welder was in a difficult location than called for in S&L's original drawing, raising more questions about S&L's initial design judgments (QC2-31); missing welds (QC2-80, 85); horizontal members that are 3" long instead of 4" (QC2-91); members that are 3/16" thick instead of the required 1/4" (QC2-109); and a duct overhang that was off location by 3'5" (QC2-113).
In light of CECO's organizational domination of the judgments, as well as its control of the f actfinding conclusions, there is almost no basis In fact, Braun to conclude that Braun's report is an independent review.
was no more than a subcontractor, whose report is being publicly released.
The reality of Braun's project simply did not match its mission to provide an independent structural check on the QA breakdown for Zack's HVAC work at LaSalle.
J
ubs Fiulty rcthodology for tha scope and_
C.
n*tura of fretfinding.
Regardless of independence, a sound methodology is the necessary premise to have confidence that Braun's work could verify the safety of Zack's HVAC system.
In our August 13, 1982 public comments, we urged that Braun disclose the selection criteria for its proposed methodology, and cite to relevant authorities to support its proposed QC inspection. plan.
We.also urged that Braun consult with the whistleblowers to target items that are particularly suspect and may require hardware tests.
t Braun chose not to accept these suggestions, so our November 19 J
interim report severely criticized the methodology for lack of either a true random or an intelligently-targeted review of items whose quality is in question. Additionally, we are concerned that Braun chose not to increase the size of its inspection sample after finding a significant number of On balance, Braun's methodology was too limited initial discrepancies.
and superficial to make conclusions about quality for a system where the paperwork fails to answer serious questions about the quality of materials and workmanship.
Initially, the Braun findings are compromised by definitions 1) either nonexistent or so vague that they neutralize the significance For example, Braun limited of the report's conclusions on safety.
its work on nonconformance reports to a review of whether the disposi-Unfortunately, tion was complete, and of " technical justification."
the report does not define what constitutes " technical justification."
to qualify as a potential safety concern, a design Most significant, discrepancy had to qualify first as a confirmed "daservation" and then Unfortunately, the definition of " finding" is circular--
as a " finding."
"An observation which has been identified as a potential safety concern."
The only additional criteria for the definition are that the observation requires " extensive repair" or is a "significant deviation" from the design.
(Report, Appendix B-3.)
In other words, there is no definition The safety conclusions in the of what constitutes a safety concern.
Braun report are nearly totally subjectiva..
An analogous flaw is the failure to cite any authority for 2)
This omission under-the QC inspection procedures used in the review. It is also in contrast cuts the legitimacy of all Braun's conclusions.
to Braun's qualifications standards and audit procedures, which are referenced to 10 CFR 50, Appendix B and relevant professional standards.
(13., Appendix A. )
It should be no surprise that Braun failed to cite authority 3) the project was basically limited to for its inspection procedures:
visual inspections, which are too superficial to meet the difficult challenge of confirming quality for Zack's suspect work, or even as Braun failed to the sole technique in a normal QC inspection program.
It only conduct or require any additional chemical or pull tests.
~
't
NRR ~
u;ed Hond:structiva Tsating in rara instincOc, and did not taka s.ny radiographs.
These steps are essential to establish qualify when work has been performed by unidentified welders whose qualifications are erratic; or when material traceability can never be established due to records
=
In addition, visual inspections never generated or lost over time.
(See, e,.g.,
of ten uncover problems that require more intensive tests.
di'storted AC2-103, involving ducts which visual inspection found wereFinally, A and discolored.)
'Mr. Howard informs us that his and Ms. Marello's whistle closure covered items for which the technical specifications used ASME g
In sum, Braun's QC program diluted Zack's Class I QA requirements.
inspection standards in order to evaluate Zack's work.
Braun was also unable to verify that its methodology in fact In an 4) matached the requirements of its own QC inspection procedures.
Braun's J. S. Fiedler disclosed, "The October 4,1982 audit report, improper use of the wrong revision number to inspection form QC-1 is It has been determined that we are not utilizing the (Report, Appen-very evident.
instructions in the manner in which they are intended."(See, e dix K. )
verify torque for anchor bolts.)
The flaws in the nature of Braun's methodology were matched by Initially, the criteria for selecting the sample remain a 5)
Braun failed to explain how it selected three its scope.
mystery in key instances.HVAC systems for detailed review or why it picked that n Similarly, NRC questions on that topic at the August public meetings.B d
tion for 20% of nonconformance reports and FCR design modifications deeme Unfortunately, there is no explanation cf the
" critical to safety."
standard to meet that criteria.
(Report, at 10.)
As a result, in some instances the selection criteria for the sample are as subjective and undefined as the evaluation standards.
There are grounds to doubt the adequacy of whatever criteria Braun chose to select its review sample, because it skipped significant 6) background research necessary for effective targeting of It is and Ms. Marello to help select the cample, as we previously urged.
ment.
verified by examining the scope of documents Braun used in its review.
The list fails to include the following documents, (Report, Appendix A.)
f which Mr. Howard explains are necessary both for an informed selection o the July 26, 1982 items, and a reliable evaluation of discrepancies-- a) the Howard-Marello whistleblowing disclosure and all attachments; b) c) copies of the technical specifications for Zack's HVAC contract; original material test reports originally used to verify quality; copies
- + - - - -
-m u NRR of the Zrck chop ntnconform nca rsports, as oppored to just tha fisld NCR's; applicable HVAC purchase Orders; and rolsvInt portions of Z:ck's As a result, Braun's methodology for contract for the HVAC system.
scope and evaluation was too uninformed to be reliable, particularly in the absence of a truly random sample measured against normal evaluation standards.
One of the most basic flaws in Braun's methodology w'as the 7) failure to expand the scope of its sample after identifying s,ignificant numbers of discrepancies. For example, when Braun was unable' to prove siders, it should have inspected
_ the qualifications for more than 20% of. When NRC tests led to findings all the work of the suspect craftsmen-that from seven to ten material samples tested did not conform to chemical requirements or were of questionable tolerance, it should have expanded the sample.
(Report, at 14.)
Expanding the secpe of review may have uncovered more generic deficiencies, such as the tubes in the drywell that were undersized and led to two " findings."
(Report, at 28.)
As a The full result, Braun's conclusions are only preliminary at best.
extent of problems identified in the report remains unknown.
In at least one key instance even the scope of Braun's " sample" 8)
Braun dismissed the problem of welder qualifications is not quantified.
as irrelevant with the following comment:
Braun has inspected approximately 250 welded hangers.
Since ach hanger consists of many members, Braun has in-spected thousands of welds associated with these supports.
Some by welders whose qualifications are complete and with incomplete qualifications. Based on these some inspections Braun has determined that the weld quality is consistent on all supports regardless of who performed the welding.
(Report, at 29.) Unfortunately, the passage did not cite to any particular appendix or referenced findings to support its somewhat This is one of the most sensitive conclusions incredible assessment.
in the report, but the passage does not provide any specific empirical As a result, Braun's conclusion has all the scientific relia-basis.
bility of a hunch.
Even where the size of the sample was known theoretically, 9)
Braun could not confirm that all the necessary inspections took place.
As the checklist to an October 5, 1982 audit disclosed, "There is no way of checking to see if all inspections requested by HVAC have been performed."
(Report, Appendix K.)
Braun should be required to ccafirm i
if and how the inspections have been confirmed.
l D.
Incomplete disclosure of results, The above analysis criticized CECO for rejecting Braun observations and findings without providing commensurate supporting document,s to those The problem of incomplete disclosure of necessary data and it rejected.
The rsault 10 aislerding or unvarifiable evid;nco permettso ths rcport.
Tha significtnca io to furth r illustrEts why ths Brcun conclu: ions.
report, as released publicly, cannot support full power tuthorizztion for The conclusions cannot stand on their own.
LaSalle.
The most common misleading tactic was to substitute a
1)
For instance, the report loaded adjectives for objective data.
refers to " isolated" cases of bad welds, "some" deficient welder qualification records, "some" bolts without necessary identification,
. and "most" nonconformance reports as only involving minor proplems.
(Report, at 7, 19, and 26.)
Braun should have included the relevant
-statistics in the text or listed examples, and left the reader suf-ficiently informed to challenge Braun's adjectives.
In other instances, Braun failed to adequately describe the 2)
For instance, Appendix A refers to Zack's deouments it did reference.
NCR (nonconformance raport) log as one of the documents reviewed.
Mr. Howard explains, however, that Zack had two NCR logs, one for This ambiguity NCR's and another for those written in the field.
shop explains his query whether Braun conducted the necessary review for both sets of Zack NCR's.
(Supra, at 14.)
In other cases the report does not provide the specifics 3)
Braun dismissed the signifi-for impressive, but vague references.
cance of 7 material samples that failed to conform to chemical re-quirements out of 48 tested by the NRC, because of " permissible (Report, at 14.)
But the report variations for product analysis."
fails to specify the permissible variations necessary to check the exoneration.
Suspect subjective evaluations, instead of E.
conclusions supported by authoritative citations.
Previous sections of this report have criticized the Braun report's conclusions for lack of independence and failure to fully provide referenced Whether Braun, CECO or S&L, however, in many instances the con-documents.
In other cases clusions are totally unexplained, subjective evaluations.
the judgments contradict 10 CFR 50, Appendix B or professional standards.
In still other examples, the evaluations in the Braun report are internally The errors involve such fundamental issues --
inconsistent or contradictory.
i.e., evaluation of chemical material tests, or the relevance of welder qualification records -- that they invalidate the report's blanket reas-In our August 13 comments we urged that Braun reference its Its failure to do so fatally surances.
evaluations to relevant professional codes.
damaged the credibility of the conclusions.
As seen above, on the most casual of empirical studies Braun 1) rejected the relevance of a widespread inability to verify welder The QA basis for the conclusion is that both Zack qualifications.
l 6
cnd S&L told Brtun at a September 21, 1982 m uting that th2ra w m "no particular coda or d: sign cpecificttion" rcquired for th3 HVAC w21d2.
(Report, Appendix C.)
This cascrtion litsrclly dafinid-out 10 CFR 50, It casts doubt Appendix B, Criterien IX for the LaSalle HVAC system.
upon Braun's judgment and independence that the third party apparently accepted the statement without challenge.
To Further, Zack and S&L appear to have misinformed Braun.
illustrate, the notes to an August 18, 1980 CECO Surveillance Report (Attachment 1) cite repeated violations by Zack for failure to suf-ficiently review welder qualifications. 'Mr. Howard points out that his document review team at Zack examined welder qualification records, and notes that if the issue were irrelevant Zack would not have issued In short, an August 2,1982 potential 10 CFR 21 Report on weld records.
the Braun report of fers the first evaluation that HVAC welding quality in general, and welder qualifications in particular, are exempt from normal QA standards.
- 2) Despite unsatisf actory chemical test results on 11 samples out of 48, Braun gave its seal of approval to all HVAC materials based on permissible variations. That conclusion was premature, however, under the ASTM standards to which Zack had been held by technical His enclosed analysis is specifications prior to the Braun report.
based on research of the relevant ASTM standards.
(Attachment 2.)
It demonstrates that due to gaps in the scope of the reported test (i.e., missing field, mechanical and elongation data), only 27%
data of the material samples can be confirmed as acceptable.
- 3) Braun also based its material evaluations on misapplied standards. The report states that--
... materials specified for ductwork and hangers are the same as those used in typical commercial and industrial The maximum design stress level is conservatively use.
18 KSi. The lowest grade of galvanized sheet metal and structural shapes available exceeds this value without exception.
(Report, at 14.)
Based on the standards used at Zack, however, Mr. Howard reports that Braun's evaluation is only accurate for ductwork, not hanger materials. The hangers were required to meet ASTM A36 standards, which specify minimums of 58 KSi for tensile strengths and 36,000 KSi for yield strength. Again, Braun either applied the wrong evaluation criteria or significantly diluted even Zack's standards.
4)
In at least one instance Braun's evaluation is suspect because of unexplained shifts in judgment. To illustrate, a September 14, 1982 Braun memorandum stated that welding procedure WFS P-5 CS is not qualified for short arc welding. An October 6 Braun memorandum found the procedure acceptable, however, without explaining the change.
(Report, Appendix C.)
5)
In another enco Brcun'c conclu3 ion 10 not credible, becausa tha cppendicas contrrd_ict tha cxoneration in tha tsxt. Braun concludsd in the text that "[ 3 /n each case it was concluded that these procedures are acceptable."
(Report, at 7.)
The appendices, however, refer to at least two procedures that either were not qualified or of only limited acceptability. The flaws for one procedure, WPS P-6 CS, Rev. 7, illus-trate the inaccuracies of Braun's conclusion:
Eleven /[f 14 tensile tests [ failed at less than the 50,000 psi minimum tensile strength required for the t
RCuSi-A filler metal (AWS 5.7).
The two lowest ten-sile strengths reported were 30,200 and 35,800 psi.
The base materials are not identified but the three specimens that broke in the parent metal failed at tensile strengths exceeding 50,000 psi.
(Report, Attachment C.)
- 6) Braun's Site Review Team ("SRT") had the responsibility to review all discrepancies uncovered by QC inspectors. The trend of suspect, unreferenced subjactive evaluations was most severe at this level. The significance is that the SRT defined-out the safety signi-ficance of the inspectors' findings. A particularly common occurrence was to overrule without explanation the discrepancies found by inspectors.
The evaluations need some explanation, since they overruled discrepancies such as welds that were 2.5" instead of the specified 5.5" (QC2-96) or accepted concave washers and oversized holes that admittedly represented
" poor workmanship."
(QC2-104.)
In other cases the unexplained evaluations were inconsistent.
(Compare QC2-70, 71 with QC2-73 for evaluation of missing bolte.)
(See also QCl-03 and 18 for examples where the inspection supervisors overruled the inspector without explanation.)
7)
In other cases the SRT's explanations are insufficient to justify the conclusions. For instance, in one example the inspector found that most of the welds were defective due to porosity. The SRT rejected the discre-pancy with the explanation that there are more welds than nec.essary. (OC 2-54 But it failed to explain how many extra welds there.are, or what the accepta-ble error rate is.
8)
In numerous instances the SRT offered explanationr~ ^ shious cre-dibility, in the absence of further information or authorit).
In one case the SRT decided that eight holes could be repaired with tape, because of where the leakage would flow.
((N 2-60)
This explanation casts doubt on the necessity for any repairs in the first place.
(See also QC 2-104, 107.)
As with unexplained rejections, Braun's supervisors also offered weak expla-nations that confirm our earlier criticisms. For example, the Braun saper-visor resolved one discrepancy with references to Zack documents that in-volved the "buyoft" of missing and faulty welds.
e I
- 9) CECO'o judgment, cf cource, wm d:ciciva. It wad Gv;n more i
ckstchy than Brcun's SRP. For example, the utility raj ctcd an ob %rv1-tien en a duct with a 3'5" execcciva ovirhing, without explinttion.
(QC 2-113)
In numercu3 in:ttnc03 ths utility dismissed Brtun obIsrvation3 solely based upon the following versatile, if nonauthoritative explantion --
"as built analysis" -- another undefined evaluation standard.
- 10) Sargent and Lundy's evaluations underly many of the CECO responses, which were generally too sketchy to specify the S&L role. S&L's analysis was exposed in two of Braun's potentially significant safety'" findings,"
however. Braun found a generic deficiency of tubes that are, undersized
- by 25% (3/16" versus 1/4").
S&L's review rejected the finding, explaining that the maximum stress would be only 14,267 psi, or less than S&L's design
- stress of 18,000 psi.
(QC 2-88, 89)
S&L failed to delineate the effect of undersized tubes on the acceptable pressure that could be sustained by the design, even if the design only called for commercial standards. (Suora at 16. ) This unexplained assumption casts doubt on S&L's dismissal of two key safcty findings.
F.
Failure to verify all necessary corrective actions.
In our September 4 comments we recommended that Braun verify the corrective action on its findings. This is always the necessary last step for an effective QA program. It is particularly necessary when the program represents the final hurdle for full power operation of a nuclear plant.
Unfortunately, as with other crucial stages of the project, Braun either skipped discrepancies or ceded the task of following through to CDCo. As a result, the report itself cannot stand as sufficient basis for final judg-ment of.L.asalle's HVAC system. Even if Braun's analysis were comprehensive and sound, CECO still must bear the burden of proving it has honored all its commitments to the third party.
In many cases that may be impossible once the plant is operating at full power.
- 1) Perhaps the easiest way not to verify corrective action is to lose track of the original discrepancy. Our review of QC #1 and QC #2 reports revealed numerous cases where discrepancies originally uncovered by the QC inspector disappeared without any specific written rejection.
See, e.g. QC 2-69 (SRT failure to discuss three welds reported as warped and/or rustings CECO failure to discuss two of three other welds the SRT observed were defective); QC 2-72 (anchor bolts cutoff and large amounts of debris); QC 2-76 (a missing bolt and a missing nut); and QC 2-78 (SRT and CECO failure to discuss reported discrepancy that hanger horizontal and vertical members are butt-welded, contrary to the draw-ings).
- 2) Braun explicitly defaulted on verification of CECO QA commitments, instead entrusting that duty to CECO itself.
(Report, Appendix L.)
For-tunately, in some cases Braun's QC 2 reports verify that CECO repairs have been completed. Unfortunately, Braun also signed off on the QC 2 reports when corrective action remained unverified.
(See, e.g., QC 2-46, m
m-
74 cnd 79.) One unrcsolved iccue with bread'applic tSon Envogv5 fret /e5Pu
~
counitment to develop a program "to corrcet tha omiccicn of painting, gal-vanizing, fircproofing cnd inculcticn as requested by tha Sits Review Tram.
QC 2-53) Thic prograr zu3t be implementsd cnd vsrifisd befora full powxr operations are authorised, or else it may never be impletmented at all.
4
- 3) There can not be any debate that the Braun report demonstrates S&L's drawings are neither current nor complete for the HVAC system.
Even for the small sample covered, however, the report doet not verify that the inaccuracies have been corrected. In some cases when CECO chose not to L2plement corrective action on a Braun observation, it at least promised to upgrade the drawings. In othe unexplained instances, CECO did not mention correcting the outdated drawings when it rejected the sub-
- stance of Braun's disclosure of design deviations.
(See, e.g., QC 2-65, 85 The NRC should inquire as to whether those drawings will remain and 96.)
outdated, and if so, why.
CONCLUSION In our opinion, the flaws in the Braun report disqualify it from Our criticisms playing a decisive role in this NRC licensing decision.
represent more than analytical challenges. Braun ruled out four NRC QA criteria under 10 C.F.R. 50, Appendix B, from relevance for its sample selection -- Criterion VII (vendor quality assurance); Criterion VII (vendor quality assurance); Criterion VIII (material tracebility);
and Criterion IX (special process controls); and Criterion XV (disposition of non-conforming conditions). Further, in our opinion the conduct of the Braun review itself is inconsistent with five other criteria of 10 C.F.R. 50, Appendix B -- Criterion I (independence for QA personnel);
Criterion V (appropriate acceptance criteria for instructions and pro-cedures); Criterion X(proper procedures and performance of tests and inspections necessary to assure quality); Criteria XII (control of testing equipment); and Criteria XVI (verifying corrective action, including the cause of significant conditions).
The weaknesses in the report support our recommendations that the Vendor Inspection Branch review Braun's findings, and that after that review Braun should return to the LaSalle site to complete its I
mission properly. Unfortunately the Braun report has demonstrated that Mr. Howard's and Ms. Marello's fears of last spring were well-grounded in fact. Meaningful conclusions about the quality of Zack's work and the HVAC system at LaSalle can still not be drawn. In fact, the only conclusion that can be drawn is that Braun's preliminary findings demonstrate the necessity for nondestructive examinations, hardware testing to resolve nontraceability problems or unreliable Zack paper-work, and a significantly expanded sample size..
i 1
e
Th3 Brcun r; port confirm 3 thLt th3 Zick Qurlity Aszuranc3 BrscAdown was not c paperwork problem. Furth:r it confirms that tha solution does not lie in more visual tests cnd prperwork riviews. A finn 1 resolution can come only after Braun perfortas a truly independent assessment and comprehensive audit of the HVAC system. The NRC
" hands off" policy with regards to this problem has failed / it must now assume a much more involved role in protecting the public from an increasingly anxious utility.
Sincerely, t
J J+M/
Thomas Devine Legal Director hA N
Mi Billie Pirner Garde Citizens Clinic Director Attachments tr.y: td O
e
- -., -.,. _ _