ML20126H532
Text
_._;._...__.. _
._:f 4
,..a.
.-2-.
. Ja _ _. __ _ _ _ _...
9s-0 [.
Y ($ Uhs [o UNITED STATES e
~~
g NUCLEAR REGULATORY COMMISSION g
g r;
- E REGION IV
,d[
611 RYAN PLAZA DRIVE, SUITE 1000
{
ARLINGTON, TEXAS 76011 Dockets No. 50-498; 50-499 MEMORANDUM FOR: File y/g/
g G. L. Madsen, Chief, Reactor Projects Branchb,
THRU:
W. A. Crossman, Chief, Project Section 3 FROM:
H. S. Phillips, Resident Reactor Inspector South Texas Project
SUBJECT:
RESPONSE TO CONGRESSMAN UDALL'S LETTER OF OCTOBER 30, 1981 (AITS H07003343F04)
In reading the Congressman's letter, I noted that the letter correctly classi-fied the Quadrex Report findings as " questions raised". This is certainly an accurate characterization since the report essentially refused to reach con-clusions other than to say that HL&P would need to probe deeper to make conclu-sions on the Quadrex Report findings which were characterized as " indications of potentially weak areas".
I do not agree with other parts of the letter and believe these areas may be cleared up by providing a detailed discussion of the Quadrex Report and the NRC role in this matter.
Introduction The Quadrex Corporation Report is a consultant's technical review of design.
As such, the report is written in an advisory style rather than a classical scientific and/or engineering style. That is, the report does not reach conclusions but merely gives findings.
Page 1 of the subject report states, in part, "Tne purpose of this review was to ascertain the overall adequacy of the STP design". However, paragraph 2 on page 1 recognizes.the limitations of achieving the stated purpose since it states, in part, "An exhaustive review of the design work accomplished by each -
was requested to carefully devise a sampling program to detennine the B&R engineering response _to known unique nuclear industry issues and problem areas.
Implicitly, it was expected that such a sampling program would provide sufficient insight regarding the adequacy of the technical work-perfonned by each discipline."
The words " implicitly" and " expected" indicate at the onset that the review would not be expected to result in conclusions. One must also recognize that such a sampling does not describe the sample size nor the inherent dangers of findings based on an inadequate sample. The sample is certainly not statistical.
8506100354 840620-PDR FOIA mIe OLASSPIS4-393 PDR
2 ~.:--.-..-
.... 1 -. L L
L.i
_~_c.
_L.
. - _ _ __a. x _ _
^
@w p
Memo for File 2
November 10, 1981 The last paragraph on page 1 states, in part, " Ordinarily it is difficult to 1
draw generalizations from particular responses to specific questions. On the other hand, the chosen sampling approach does provide indications of poten-i i -
tially weak areas in the engineering design of the STP plant".
In this para-graph, the report recognizes not only the difficulty of drawing a conclusion, but also the difficulty of drawing a generalization. Granted this, the report states that only " indications of potentially weak areas" are provi.ded by this approach and only further review by HL&P would allow HL&P to conclude that these questionable design areas are inadequate now or will be in the future if not corrected.
The NRC staff and interested intervenor groups must consider the limitations of the report as well as the areas which were considered questionable.
In i
other words, one cannot lift parts of the report and present them as absolute fact and reach conclusions when the report itself refuses to reach conclusions.
Although some areas may be found deficient, nothing short of a full review by
~
HL&P and Bechtel can determina the validity of Quadrex's " indications of potentially weak areas". Until such a review produces at least preliminary 4
or perhaps final conclusions, there is no basis for concluding that the design is fundamentally flawed. This is not to imply that some deficiency will not be found since no system, including the design system, is perfect. In fact, construction deficiencies reported to date by HL&P indicate some deficiencies will be identified. NRC reviews of design activities during the life of the project have not identified such generic problems. Therefore, design must be
]
considered adequate until a full indepth review concludes othemise.
I Detailed Comments on Congressman Udall's Letter The letter stated in part, "The report raises implicit questions as to how the deficiencies of the kind enumerated in the report could have escaped detection by the NRC".
4 i
i Connent: The NRC has identified isolated problems in the design engineering but no violations have been identified to date which show the STP design is i
fundamentally flawed. A trend relative to design problems began to appear in late 1980 and early 1981. The NRC Resident Reactor Inspector (RRI) identified this trend to RIV Construction Supervision in April,1981.and documented the apparent trend in a memo dated May 1981. This trend was' identified indepen-dent of and without knowledge of the licensee's efforts in this area only i
three months after the licensee initiated a consultant's review on January 16, i
1981. The Quadrex Review was completed March 31, 1981 and was documented i
in a report dated May 7, 1981; however, the NRC did not know the report was completed until August 18, 1981.. It appears that the NRC recognized a trend
-l developing shortly after the licensee recognized the same trend. Considering the NRC's limited number of inspectors, the NRC did a very good job of identi-fying the problem.
'9 y' T *
-f--
t ---
4,-
.-e#,-y ew,,,--v 4
... -,,y-
.n,..,
.w.
.vm..-._
.__. w
=.
- 5..
O. :.
u b
Memo for File 3
November 10, 1981 The whole NRC program is based on reviewing, inspecting and/or auditing only a very small sample of the licensee's design, procurement, and construction activities. This small-sample approach relies very heav'ly on the premise
~
that the license'e and contractors are responsible for controlling the activi-i ties since the licensee holds the license and has the manpower to perform such reviews. In view of the small sample and the limited NRC manpower, this approach may sometimes only identify isolated deficiencies instead of a trend.
The following consnents demonstrate that this program philosophy worked in this case:
The NRC Vendor Inspection Program had identified isolated deficiencies in the past several years but NRC construction. inspections did not identify any trends in design engineering until NRC Report 81-07, dated April 13, 1981.
Follow-up inspection on Show Cause Order Item V(A)9, Audits, showed that the HL&P audit group had identified generic problems in the B&R Engineering
-Organizations. In April 1981, the RRI reviewed the construction deficiencies reported in 1979,1980 and 1981 and determined that a trend may be developing because of the high number of construction deficiencies reported and the number which were related to design.. The RRI conducted interviews of several licensee and contractor personnel and learned that the STP Engineering Model at Cress-park B&R Office had over 1600 hold tags applied. On May 7, 1981, the RRI 4
drafted a memo which indicated that an adverse trend may be developing in B&R. engineering and the ramo requested a special team to inspect / audit the licensee and the AE. This draft was faxed to Region IV on May 7,1981 and the final typed memo, dated May 27, 1981, followed. This memo also referenced problems identified ty* two other NRC inspectors.
4 On May 13 or 14th,1981, the RRI and D. Sells, NRR Project Manager, were at the hearing in Bay City, Texas. The RRI' told the Project Manager that poten-tially adverse trends had been identified in B&R engineering and, in response to the May 7,1981 draft memo, a special RIV inspection team was to be sent to the Houston offices. The Project Manager stated that he had just learned from Jerry Goidberg, HL&P VP, Nuclear Engineering and Construction, that a consultant had been hired to review the B&R Engineering Organization's work and much of the work was completed and he promised to fonvard a copy of the report to.the NRR Project Manager.
On July 20-24, 1981, RIV Vendor Inspection Branch subsequently performed an inspection at the B&R Engineering Offices and documented the inspection in NRC Report 99900502/81-03, dated August 19, 1981.
During the week of August 10-14 1981, the RRI was requested to participate in an investigation of several allegations which related to the HL&P Houston Audit Group. One of the allegations stated: "Frazar/ Barker disagreed with HL&P QA personnel who wanted tc' issue a stop work order.to B&R Design Engineer-ing relating to the. fact that E&R Design Engineering was falling behind which might produce construction errcrs. While an NCR had been issued on this topic, 4
f c
e,.-., -
r-4,
. a.-
~... -
.. - - - - -.. ~... -.-
@v:.
O 1
Memo for File 4
November 10, 1981 e
HL&P QA thought a stop work order should be issued." During the investigation on August 18, 1981, the RRI requested all HL&P audits, consultant reports and applicable correspondence that related to B&R Design Engineering Organizations be furnished for review. At this time, the licensee produced the Quadrex Report but restricted the physical review to the HL&P Manager QA office. A brief look at the subject report indicated that serious questions had been raised. The RRI told the licensee that the report raised some serious con-cerns and a more extensive review was necessary. The licensee stated that a copy would be made available at the site.
On August 24, 1981, Region IV Management contacted the RRI after they learned of the RRI's concerns about the Quadrex Report. The RRI stated that the review of the report was incomplete, therefore, an adequate briefing on the subject would not be possible until the review was complete. The RRI completed the review and gave a briefing to RIV Management on August 27, 1981. The RRI highlighted the questions raised by Quadrex and stated that if an NRC inspec-tion confirmed such questions or concerns to be facts then B&R Design Engineer-ing would need help from another AE or, if this were not possible, may need to be replaced. The'RRI reconsnended a special team, composed of engineers capable of validating computer programs, models and design systems, to confirm or deny 4
I l
the Quadrex Report findings. The RRI requested that someone in Region IV, who currently works with Architect Engineers, review the report to determine if the NRC should perfonn such an indepth review.
On August 31, 1981, Region IV sent an individual to site to review and evaluate the subject report. He confirmed that the report raised questions that would be serious if such questions were found to be facts.
On September 8,1981, the RRI attended a RIV-Licensee Management Meeting on the subject report. An NRC Headquarters representative also attended the meeting. The licensee briefed Region Management and attendees relative to the background, significance, and actions to be taken et.7 result of the subject report. Meeting highlights were documented in Fi Memo for File, dated September 11, 1981.
On September 23,.1981, the licensee announced that South Texas Project respon-sibilities would be reallocated and Bechtel Power Corporatien would be resp 7n-sible for engineering, design and construction management services necessary to complete the project.
In comparison, the Quadrex Review involved 12 engineering consultant personnel who spent more than six weeks in reviewing B&R Design Engineering documents and interviewing various B&R discipline engineers. This represented more than one man year of design consultant review effort, yet, three NRC inspectors independently and without the knowledge of consultant review identified similar trends in a fraction of the time and with a fractioi of the manpower.
As discussed above, Region IV inspection not only identifie.1 the adverse trend but also escalated the potentially adverse trend to Region IV Manage-ment who took appropriate action.
4 f
g Q
t Memo for File 5
NoYember 10, 1981 The letter further stated in part, "The Quadrex Report suggests additional reason to doubt the efficacy of the Office of Inspection and Enforcement in monitoring activities at this site".
Comment: I have been the resident inspector on the STP site for almost two and one half years.
I think that I have done an outstanding job of monitor-ing activities at the site. The author of this statement shnuld be respect-fully infomed that the design work is not perfomed on site and, therefore, the design work could not possibly be monitored on site and is not the responsibility of the resident or regional inspectors who inspect construc-
~ tion activities on site. The author of this statement should be respectfully informed that NRC manpower would have to be drastically increased to perform indepth design overview of the Architect Engineers total engineering effort; however, the NRC could certainly perfom such reviews if such resources were available.
In regard to the tardiness of identifying or confronting allegations of harassment of Quality control Inspectors at South Texas Project, Region IV inspections and investigations addressed each problem that arose at South Texas Project; however, a trend was not apparent until a comprehensive investigation, which lasted several months, was conducted to identify and substantiate such allegations and determine other adverse QA trends.
Typically, Regions do not have adequate resources to conduct such extensive inspections and investigations. However, two Region IV personnel made major contributions to this investigation which resulted in violations and a civil penalty.
In view of the coment previously given. I think that the NRC identification of potentially adverse trends was timely and the NRC system worked effectively.
"It is important, therefore, that the Comission come forth with a statement presenting (A) the reasons that defects of these kinds remained so long undetected and (B) and explanation of what is being done to uncover any such situations at other plants."
Coment: The letter correctly characterizes the Quadrex Report and keeps the questions of that report in perspective, however, paragraph 2 and 3 appear to conclude that STP design has fundamental flaws or defects. The licensee has reported several construction deficiencies which were the result of the Quadrex Review / Report; however, the nature of these deficiencies do not suggest l
that design is fundamentally flawed. The STP design is also compared to Diablo Canyon where the design is known to be fundamentally flawed, however, this comparison is not accurate and no such conclusion can be reached until either the licensee /Bechtel's indepth review gives a basis for such a conclu-sion or an NRC review / overview would allow reaching such a conclusion.
9
, p g
g q
p
'v r =e'g
=
- g w
e Memo for File 6
November 10, 1981 The present NRC inspection system evaluates a very small sample of the Architect Engineer's (AE) total engineering effort. Accordingly, there will always be a high risk that adverse trends in design engineering may not be identified. Such inspections are far more likely to identify isolated deficiencies or violations rather than generic adverse trends.
Current inspection of design does not extend to the site, in that, NRC Design Engineers do not inspect the design effort on site as compared to the review of. design documents and other objective evidence at the AE facility. The NRC interface between inspections performed at the design engineer and review of design on site is a weakness in the inspection program.
NRC manpower to inspect the AE design effort would need to be greatly increased to achieve increased inspection of the AE effort and "as built design".
It would necessitate placing NRC Resident Engineering Inspectors at AE and NSSS facilities as well as the construction site. This action would give the NRC a 'better opportunity to identify adverse trends earlier and would increase the probability of identifying such trends; however, the present program does not provide such engineering inspections.
Position on Quadrex Report What is the staff position on the Quadrex Report?
Does the report conclude that the design of the South Texas Project is funda-mentally flawed?
Comment: No, the report does not conclude that STP design is fundamentally fiaued.
Does the staff, on the basis of the report, believe that the design is flawed?
Comment: Objectively, the staff cannot conclude that the STP design is flawed. However, the staff does think that the Quadrex Report has identified some questionable areas that must be addressed.
What staff action is currently necessary?
Comment: No staff action is necessary until the HL&P/Bechtel review of B&R design is completed except to monitor the in-progress conclusions reached as a result of this review. HL&P has been doing very little new safety related work during the last year. When they received the Quadrex Report, they unilaterally stopped work in many areas until questions are resolved. The recent contract dispute and resultant B&R withdrawal as project constructor will cause all safety related activities to'stop by December 1, 1981, except for caretaking activities outlined in HL&P letter (ST-HL-AE-750), dated November 9, 1981. This development will allow HL&P/Bechtel additional time to assess the questions raised in the Quadrex Report. As the letter indicated, the transition will take at least ::ix months before safety related work activity would resume.
4 k
,+n
-n
([
Q:
Memo for File 7
November 10, 1981 4
When the HL&P/Bechtel review is completed, the staff would review the results of the review. If their review was considered adequate and corrective. action, if-necessary, was adequate, the work should proceed.
+
S i
r H. S. Phillips Resident Reactor Inspector South Texas Project e
f l
l
_