ML20203L227

Rept to Duke Engineering & Svcs,Inc,On Allegations of Willfuness Related to Us NRC 971219 Demand for Info

ML20203L227
Person / Time
Site:	Yankee Rowe
Issue date:	02/25/1998
From:	Mcgarry J WINSTON & STRAWN
To:
Shared Package
ML20203L159	List: ML20203L155 ML20203L162 ML20203L193 ML20203L227 ML20203L245 ML20203L269
References
NUDOCS 9803050407
Download: ML20203L227 (29)

Text

REPORT TO DUKE ENGINEERING & SERVICES, INC.

ON ALLEGATIONS OF WILLFULNESS RELATED TO THE U.S. NUCLEAR REGULATORY COMMISSION DECEMBER 19,1997 DEMAND FOR INFORMATION Winston & Strasm J. Michael McGarry, III Mark J. Wetterhahn February 25,1998 388 1888K 81888 6 W PDR

APPENDIX D TABLE OF CONTENTS Page I. EXECUTIVE S UMM ARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D- 1 II. INTRO DUCTI O N . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D -3 III. INVESTIGATIVE AND DELIBERATIVE PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . D-4 IV. LEGAL ANALYSIS OFWILLFULNESS .... ............................D-5 V. FI ND I NG S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D -7 VI. RESPONSE TO SECTION III OF THE DEMAND . . . . . . . . . . . . . . . . . . . . . . . . . . . D.9 A. Deliberateness . . . . . . . . .. .. ...............................D-9 B. Response to Sections III.A and B ..... ....... . ..... . . . . . . . D-9 C. Response to Section III.C . . . . . . . . . . . . .........................D-17 D. Response to Section III.D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D- 19 VII. CONCLUSION . . . . . . . . . . . . . . . . . . . ........... ... . . . . . . . . . . . . . . . . D- 2 6 wamow n. D-ii

--- - I

REPORT TO DUKE ENGINEERING & SERVICES, INC.

ON ALLEGATIONS OF WILLFULNESS RELATED TO THE U.S. REGULATORY COMMISSION J)ECEhfBER 19.1997 DEhiAND FOR INFORhfATIQN L EXECUTIVE SUhihfARY Winston & Strawn was requested to determine whether certain allegedly inadequate engineering analyses and materially incomplete and inaccurate information relating to the SBLOCA analyses used by Maine Yankee Atomic Power Company to demonstrate compliance with 10CFR50.46 was the result of willfulness, either deliberateness or careless disregard, on the part of personnel of Yankee Atomic Electric Company (" Yankee Atomic"), many of whom recently became part of Duke Engineering and Services, Inc. ("DE&S"). To respond t< (; 3 request, Winston &

Strawn reviewed the NRC's December 19,1997 Demand for Infonnation (" Demand"),

reviewed relevant documents and interviewed a number of witnesses. These actions were taken in conjunction with an expert Technical Review Team whose function was to determine the technical adequacy of the SBLOCA analyses in the areas raised by the NRC.

Willfu! ness, as used by the NRC, embraces a spectrum of actions ranging from deliberate intent to violate or falsify to and including careless disregard for NRC requirements. Deliberate misconduct is defined as "an intentional act or omission that the person knows would cause a licensee to be in violation of any regulation or other NRC requirement." On the other hand, a finding of careless disregard indicates that the person acted with reckless indifference to a requirement or with disregard or utta unconcern of the consequences of whether there was compliance. The existence of a reasonable justification for an action would defeat a charge of willfulness despite the fact that the action was ultimately found to violate NRC requirements.

We found no actions on the part of any individuals associated with me specific issues contained in the Demand that would involve deliberateness. We found all individuals to be open, honest and communicative to us. We found no specific intent to v:olate any NRC regulation. Our evaluation, therefore, focused on whether there existed careless disregard for Commission requirements.

There were four specific allegations in the Demand, the first two of which had a common factual underpinning. The NRC alleged that because not all points of the SBLOCA spectrum could be reliably calculated by the code used by Yankee Atomic for the Maine Yankee facility, the requirements of 10CFR50.46 were not met. The Technical Review Team concluded that the standard industry practice, as utilized by experts in the LOCA field, was that the code should have the capability of analyzing all points within the prescribed spectrum. Yankee Atomic had taken the position that the identification of the limiting break, combined with a sufficient understanding of the physical phenomena which were occurring over the entire small break region, provided compliance with 10CFR50.46. The Technical Review Team recognized that NRC expectations regarding 10CFR50.46 were not completely documented, but rather had been sumsaw.m D-1 l

communicated to the LOCA community through its interactions with the NRC. Utilizing the appropriate legal standard, and bearing in mind that the Yankee Atomic LOCA Group was isolated and not part of the LOCA community, it was determined that the applicable regulation could be read as it was by Yankee Atomic. Actions to implement the LOCA Group's interpretation of the regulation did not evidence reckless indifference and, thus, no careless disregard was found.

With regard to the second issue, the NRC asserted that Yankee Atomic caused a violation of 10CFR50.9 in that Yankee Atomic provided inadequate information to Maine Yankee regarding the SBLOCA which did not reveal the code inadequacies discussed above. Many of the same considerations apply to this issue as to the previous one. In addition, the Technical Review Team found that on the whole the document submitted to Maine Yankee was sufficiently complete and accurate when judged using the perspective of its intended audience, i.e., one knowledgeable in the field, such as an NRC reviewer. Under the circumstances, we found that careless disregard of the regulations was not present.

The third issue relates to an assertion by the NRC that Maine Yankee had not provided a technical basis for one element of the SBLOCA analysis, the loss coefficient for the split downcomer nodalization, and, as a result, there was overprediction of core cooling and overstatement of the conservatism of the model. The Technical Review Team determined tnat the modeling approach utilized was reasonable ar d consistent with industry experience. The Technical Review Team determined that a deficiency in the quality assurance of a confirmatory calculation existed, but there was a reasonable explanation as to why it occurred, and that the evaluation undertaken by Yankee Atomic was appropriate. No inadequate analysis existed and the issue of deliberateness or careless disregard did not arise.

With regard to the fourth and last issue, the NRC asserted that Yankee Atomic memoranda used an unacceptable Best Estimate model rather than an approved Evaluation Model in evaluating the effect of a decrease in steam generator pressure on the peak clad temperature in the small break region when it should have known that such analysis would form the basis of a 10CFR50.59 analysis. The Technical Review Team determined that methods other than the approved Evaluation Model could have been utilized for the work undertaken. However, the Technical Review Team felt that limitations on the use of such methods should be stated. Here certain of the memoranda mischaracterized the Best Estimate model as the approved 10CFR50.46 Evaluation Model for the facility.

With regard to the use of the memoranda, we determined that it was understandable to have failed to contemplate that such work products would be used in a 50.59 analysis inasmuch as they either represented scoping calculations and/or were not thought at the time to be addressing design basis issues. However, it is also understandable that a Maine Yankee employee not expert in LOCA analyses would use these memoranda in performing the assigned 50.59 analyses. In evaluating whether careless disregard had occurred, we noted that the issue of whether decreased steam generator pressure impacted the design basis did not mature until late 1992 (i.e., some months after receipt of the last memorandum), as evidenced by the Maine Yankee m:cwoms D-2

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

NRC Resident Inspector's suggestion that a 10CFR50.59 analysis should be performed. We i

found that the appreciation of the necessity for evaluating degraded conditions pursuant to 10CFR50.59 was in a state of transition at the time, with the NRC publishing Generic Letter 91-18 in late 1991. We concluded that while a misstatement was made that the Best Estimate model was the licensing basis SBLOCA analysis and while it was used without prior discussion with the NRC, the actions of Yankee Atomic personnel did not meet the test for careless disregard of the regulations in that we could not conclude that there existed a reckless disregard or careless indifference towards its responabilities or the consequences of the actions taken under the circumstances assumed. Indeed, on the very date Maine Yankee's 10CFR50.59 analysis was internally approved, YAEC furnished Maine Yankee a draft 50.59 analysis on the precise subject based on its RELAP5YA Appendix K Evaluation Model.

In conclusion, we determined that, while in certain instances there may have been inadequate analysis associated with the SBLOCA analysis, there was neither deliberr.teness nor careless disregard resulting from the deficienc;es discussed in the Demand. We found no willfulness on the part of the two individuals or any other Yankee Atomic personnel. We therefore believe that the Manager and Lead Engineer are capable of conducting their activities in the future in conformance with NRC requirements. We conclude that there was nothing developed as a result of our investigation on the conduct of Yankee Atomic and/or DE&S personnel that would prevent the improvements that we understand are being made from being successful and resulting in DE&S' activities being conducted in full compliance with NRC requirements and expectations.

II. INTRODUCTION On December 19,1997, the U.S. Nuclear Regulatory Commission ("NRC") sent Duke Engineering and Services, Inc. ("DE&S") and Yankee Atomic Electric Cotapany ("YAEC" or

" Yankee Atomic") a Demand for Information (" Demand") concerning the pmvision of

" inadequate engineering analyses and materially incomplete and inaccurate information to an NRC licensee," namely Maine Yankee Atomic Power Company (" Maine Yankee" or "MYAPCo"). The subject matter of the Demand was a Loss-of-Coolant Accident ("LOCA")

analysis used by MYAPCo to demonstrate compliance with 10CFR50.46 and Appendix K to 10CFR50. Specifically:

YAEC prepared the small-break LOCA analysis which was utilized by MYADCO during its operating Cycle 14 and to support its reload analyses

, for operating Cycles 14 and 15. See YAEC-1868. " Maine Yankee Small Break LOCA Analysis"(RELAPSYA SBLOCA analysis). YAEC also prepared the large-break LOCA unalysis utilized by MYAPCO for Cycles 14 and 15. See YAEC-1160. " Application of Yankee WREM-B ASED Generic PWR ECCS Evaluation Model to Maine Yankee" (WREM LBLOCA analysis).'

' Sm Letter of December 19,1997, transmitting Demand at 1.

t wamown. D-3 l

The Dem:.ad required DE&S to submit responses to the NRC within 30 days, which date was extended pursuant to DE&S' request to February 27,1998. Among the matters raised in the Demand was an NRC request for "[a]n explanation why the NRC should not consider the inadequate analyses, which apparently caused MYAPCo to be in violation of NRC requirements, to be the result of willfulness, either deliberateness or careless disregard, on the part of YAEC and/or DE&S personnel." DE&S had purchased part of Yankee Atomic in late 1997, including the LOCA Group which was the focus of the Demand.

Winston & Strawn was requested by DE&S to address the willfulness aspect of the NRC's

Demand. In response to this request, two attorneys from Winston & Strawn, J. Michael McGarry and Mark J. Wetterhahn (the " Willfulness Review Team"), both with a background in nuclear energy law, worked closely with the Technical Review Team chartered by DE&S to 3 report on the technical issues contained in the Demand. The Technical Review Team consisted of experts in the field of LO'?A analysis and the other subjects contained in the Demand.

The instant report describes the investigative process, the legal standards utilized, the determination as to willfulness and the bases for our findings for each of the four specific issues contained in Section Ill.A-D of the Demand, and the Willfulness Review Team's conclusion as to the ultimate issue contained in Section V.B of the Demand.

III. INVESTIGATIVE AND DELIBERATIVE PROCESS At the outset, the Willfulness Review Team reviewed the Demand and associated documents to better understand the NRC's perspective, to identify the documents the NRC relied upon and to define the factual statements attributed to Yankee Atomic personnel. Additional relevant documents were also revi-wed on a selective basis.

Because a determination of v.illfulness involves a mixed question of fact and law, it was necessary to rely upon deterrinations of fact made by the Technical Review Team. For example.

the premise of NRC question V.B is that there was one or more " inadequate analyses" which caused a violation of NRC requirements which, in turn, resulted from willfulness on the part of YAEC and/or DE&S personnel. If the analyses discussed in Section III.A-D were determined in whole or in part to be technically adequate as iudged by the appropriate standards, then the need for a willfulness determination as to that element is obviated. However, even should a determination be made that any of the analyses discussed in Section III of the Demand was in one or more ways inadequate, the deliberations and recommendations of the expert panel are necessary to determine the nature and extent of that deficiency from a technical standpoint and in relation to the NRC regulations.

As a result, the Willfulness Review Team closely coordinated its efforts with the Technical Review Team. In conjunction with the Technical Review Team, a number of questions were developed in order to assure that the Demand was completely analyzed and each of the issues, 2

Demand at 18.

. m m osw u cs D-4

,, ______ _ _ l

premises and facts contained therein was questioned. The Willfulness Review Team att:nded the l pre-interview meeting of the Technical Review Team. We were also present at all the interviews l conducted by the Technical Review Team, including those of the Manager and Lead Engineer of the LOCA Group. Generally, during those interviews, the Willfulness Review Team elicited background information and conducted some follow-up questioning directed to matters contained in this report; the bulk of the questioning was by the members of the Technical Review Team.

Interviews and ensuing discussions took place over the course of four days. During the meetings, the Willfulness Review Team observed and participated in the deliberations of the Technical Review Team. Utilizing the questions previously generated as a guide, we challenged the Technical Review Team's conclusions to assure that they were understandable, defensible, and in a form that could be readily understood by an informed person and thus could be utilized as input to our deliberations.3 IV. LEGAL ANALYSIS OF WILLFULNESS In analyzing the question of willfulnus, we used NRC guidance and precedents and judicial analyses in formulating a seres of questions to guide our deliberations. The term " willfulness,"

as used by the NRC, " embraces a spectrum of violations ranging from deliberate intent to violate or falsify to end including careless disregard for requirements."4 Thus, at one end of the

" willfulness" spectrum are violations involving a " deliberate intent to violate or falsify."3 At the same end of the willful spectrum is 10CFR50.5 (1997), the NRC deliberate misconduct rule, which prohibits deliberate violations of NRC regulations and deliberate submittals ofinaccurate materialinformation to the NRC. For purposes of the deliberate misconduct rule, " deliberate misconduct"is defined as "an intentional act or omission that the person knows . . . [w]ould cause a licensee to be in violation of any rule, regulation," or other NRC requirement. Thus, for a violation of an NRC regulation to be " deliberate," there must be both an intent to take a certain action and knowing intent to violate an applicable requirement.

Examples of deliberate behavior, as previously cited by the NRC, include: a secur4y superviser reteaching exam materials after several guards answered a question incorrectly, allowing the 3

We have reviewed the report of the Technical Review Team entitled " Yankee Atomic SBLOCA Technical Review Report (" Technical Review Report"). While certain portions of that report are cited herein, we have based our f'mdings on the Technical Review Team's report in its entirety. A number of factual statements in this report are not contained in the Technical Review Report. These statements were based upon interviews conducted and information received.

NRC Enforcement Policy IV (C),60 Fed. Reg. 34,381,34,385 (1995).

11 smwowms D-5 1

,, _ _ _ ..o__

guards to change their answers, then submitting the guards' revised answers for grading;6 making false statements on an application for access authorization;' and a contractor providing false information about the capability of its product. These cases are examples of actions by individuals with specific knowledge that their actions violated NRC regulations.

Dehberate behavior is, however, not the only standard applied to prove willfulness. At the other end of the spectrum of" willful" violations are those that result from careless disregard.

" Careless disregard" is not defined in NRC regulations; however, a clear pronouncement by the NRC Stuff regarding " careless disregard" was provided in the context of proposing the deliberate misconduct rule:

Careless disregard has been described as a showing of disregard for a governing statute or an indifference to its requirements . . . . A finding of careless disregard indicates that the person acted with reckless indifference to th requirement, or with disregard (or utter unconcern) of the consequences or whether there was compliance. This recklessness involves, at a minimum, an unconcern as to whether a requirement was or will be violated, or a situation in which an individual blinds himself or heiself to the realities of whether a violation has occurred or will occur.8 The NRC's description of" careless disregard" is similar to language that has been accepted by the Supreme Court. The Supreme Court has also held that if an individual makes a reasonable, good faith effort to determine what constitutes a violation of the law, then he cannot be acting with careless, or reckless, disregard. Accordingly, a willful violation cannot result if a licensee has considered NRC requirements and reached the erroneous conclusion that its actions will not violate relevant statutory or regulatory provisions. The existence of a licensee's reasonable justification for its actions should therefore, defeat a charge of willfulness, despite the fact that -

the licensee undertook an action that was ultimately found to violate NRC requirements.

6 Pennsylvania Power & Light Co.. EA 94 212 (May 9,1995) (issuing a Severity Level III violation to the licensee for violating 10CFR50.9(a)); Darryl R. Zdanavage, IA-95-011 (May 9,1995) (issuing a Severity Level III violation to the exam proctor for violating 10CFR50.5(a)(2)).

Eg, Juan Guzman, IA 96-018 (Apr.19,1996); Gerald O. Eckard, IA 94-016 (Aug. 3, 1994).

55 Fed. Reg. 12,374,12,375 (Apr. 3,1990) (citations omitted). &nhe 52 Fed. Reg.

49,362,49,365 (Dec. 31,1987) ("The concept of ' careless disregard' goes beyond simple negligence . . . [it] connotes a reckless disregard or callous . . . indifference toward one's responsibilities or the consequences of one's actions.")(citations omitted). In sum," careless disregard" does not require a conscious decision to violate a known requirement;" careless disregard" requires that a siolator act with indifference (beyond mere negi:gence) to the applicable requirement.

srawwas D-6

V. FINDINGS The fin: lings on the individual issues outlined in Section III of the Demand and the actions of the individuals identified therein must be viewed in the context of the history of development of the SBLOCA code, RELAPSYA, by Yankee Atomic for Maine Yankee. In addition, key decisions made by individuals other than the persons named in the Demand significantly affected the actions of the named individuals, the ahernatives available to them, and their decision making process. Perhaps equally important is an understanding of the internal reporting relationships, responsibilities, and philosophy of YAEC and its relationship with Maine Yankee.

In response to a decision by YAEC management to brcaden their analytical capabilities to suppori their customers, a decision was made to initiate the RELAPSYA/ MODI development project in 1980 to analyte loss of coolant accidents in accordance with the requirements of 10CFR50.46 and Appendix K to 10CFR50. The RELAP Code was or!ginally developed by INEL. In 1983, YAEC submitted a request to the NRC for approval of RELAP5YA as the code to meet the requirements ofII.K.3.30 of NUREG-0737 related to post TMI issues?

While the NRC review was in progress, Yankee Atomic applied the RELAP5YA code to the resolution of an issue for Maine Y:nkee raised by the NRC telated to reactor coolant pump trips. During this time frame, it was recognized that there were some inherent difficulties in the RELAPSYA code which would require significant worl; to overcon'e. Certain individuals within the LOCA Group, including the Manager and Lead Engh eer, although the former did not hold a management position at the time, requested that YAEC management permit them to develop a new code without the same inherent limitations as the code then under review by the NRC.

Management decided to proceed with the use of RELAP5YA. While the full details and the wisdom of such a management decision are well beyond this review, it indicates the mindset of the I OCA Group that they were to use RELAP5YA and overcome its difficulties in meeting the technical requirements of the plants for which they provided a SBLOCA Evaluation Model.

Also as determined by the Technical Review Team, the process by which approval of the IGLaPSYA cNe was sought from the NRC in 1983, and which extended over a period of six years, was flawed in that in parallel with the approval of the Evaluation Model, plant specitic applications were not performed. This resulted in the approval of a generic code which was not significantly challenged and whose application to a r.pcedic reactor, ca, Maine Yankee, was unproven, significantly lengthening the time necessa y to obtain a final specific analysis which could be implemented for Maine Yankee.

In our view, these factors were an influence on the LOCA Group to make RELAP5YA work for Maine Yanke and increased the risk to both Maine Yankee and YAEC should that code not be demonstrated to comply vmh the requirements of 10CFR50.46. These critical decisicas were made prior to th: individt. . designated as Manager in the Demand being the incumbent in the position; he and the LOCA Group were saddled with these handicaps.

id. at Ill.B.

.mumu D7

It was also clear that the YAEC LOCA Group was working in relative isolation from the LOCA community which was dominated by the nuclear steam supply system vendors and those companies selung core reloads. The Yankee Atomic LOCA Gcoup was relatively small and, although allindividuals interviewed appeared to be knowledgeable about SDLOCA computer models and phennmena and are technically qualified, they were challenge:1 by resource limitations. Yank,:e Atomic *nanagement did not assure sufficient experience and depth in the staff of the RELAP5YA project team, in particular at the beginning of the project in the early 1980's. The Technical Review Team expressed its view that "a LOCA development program requires a very significant resource commitment which includes a sustained critical mass of expertise and continuity of key personnel.""

After the completia of the NRC's generic approval of RELAPSYA in 1989, a situation arose which led to some of the later deficiencies. During the generic RELAPSYA approval process, the LOCA Group was able to communicate directly with the NRC in furtherance of the licensing of that code; thereafter, application of the generic code to Maine Yankee became a hiaine Yankee licensing matter which lea Yankee Atomic personnel, including tLise in the LOCA Group, to believe that they could have no direct interface with the NRC."

The structure preordained by hiaine Yant:ce and Yankee Atomic management also appeared to inhibit the free flow of communications between the Yankee Atomic LOCA Group and their ultimate customers at hiaine Yankee, although hiaine Yankee personnel appear to have been briefed periodically on the difficulties that the Yankee Atomic LOCA Group were encountering in implementing RELAPSYA for hiaine Yankee. Formal communications from the LOCA Group were funneled to a hiaine Yankee project group at Yankee Atomic and from there to a single point of contact at hiaine Yankee to be disseminated to the appropriate hiaine Yankee organization. Submittals intended fer the NRC were made by a hiaine Yankee licensing group which seemingly had little contact with the Yankee Atomic LOCA Group. While there may have been legitimate msons which resulted in such lines of communications (which reasons we did not explore and did rat consider to be relevant to our inquiry), the fact is that communications were le3s than ideal and, in our view, contributed to the difficulty in resolving the issues discussed in the Demand.

l Discussions with Yankee Atomic personnel revealed to us a rather uneven understanding of the NRC requirements related to design basis, the requirements of 10CFR50.59, and of the term

" safety-related." Inasmuch as these events occurred in the late 1980s and up to 1993, we l attempted to judge this lack of appreciation z.nd the resulting actions by YAEC employees by l

contemporaneous standards. Clearly, the emphasis with regard to meeting NRC standards was to assure adequate safety as opposed to today's focus on compliance. It was also clear that the M. at Ill.D.4.

M.

M. at Ill.D.6.

I smxmm a D-8 l

i employees of the LOCA Group considered themselves a technical resource and not part of the direct support of Maine Yankee operations. The LOCA Group was largely isolated and had minimal experience in operations: their entire emphasis was on technical development of information that could be used in the implementation of 10CFR50.46 and Appendix K to 10CFR50.

With this background, we turn to the issues posited in the Demand for Information, Section Ill.A D.

VI. RFRPONSE_TO SECTION Ill OF THE DEM AND A. DeWarv.ntss it is our conclusion, as bolstered by the opinion of the Technical Review Team, that there was no deliberateness revealed in our investigation on the part of any YAEC or DE&S employee or, for that matter, any other person within the scope of the issues

! covered by the Demand. We found all individuals that we interviewed to be open, honest, and cooperative. If anything, they appeared to be overwhelmed by the situation before them, and they stilliacked an appreciation as to how matters that they perceived to be appropriate technical decisions -- or, at the most, technical differences among experts in their specialty - had turned into an issue of willfulness in which their veracity was being challenged.

We did not encounter a situation where individuals tried to shift the blame to others; all were willing to defend what they believci' ' valid technical decision making within the scope of the NRC regulations as they ..derstood and interpreted them. Thus, we have determined that to the extent that there were inadequate analyses as discussed below and in the report of the Technical Review Team, which could have resulted in Maine Yankee being in violation of NRC requirements, they were not the result of deliberateness on the part of YAEC or DE&S personnel. We have no indication of any individuals proceeding despite knowledge that his or her actions were wrong. Thus we conclude, using the standards set forth above, that no deliberateness was involved.

B. Response to Sections III.A and B We turn now to a discussion of the four issues contained in Section 111 of the Demand to determi 2 whether there was careless disregard on the part of YAEC or DE&S personnel which could have caused Maine Yankee to be in violation of NRC requirements. In doing so, we utilize the Technical Review Team's findings as to whe'.her there existed inadequate analyses in the specific areas delineated in Section Ill.A D.

.-. D9 cs _

A. During Cycle 14 opesn: ions, and in the Cycle 14 and Cycle 15 reload analyses, Afalne l'ankee used apparently unacceptable Evaluation Afodels which could not calculate or reliably calculate ECCSperformance.

B. Af YAPCo maintained infonnation and submitted to the NRC Core Performance Analysis Reports, in support of Cycle 14 and Cycle 15 reload applications, which apparently were not complete and accurate in all material aspects.

We treat subsections A and B of Section III of t )cmand together since they are largely based upon common factualissues. Out  ; airy was bounded by the small break analyses. Ls,. tbr. small break spectrum up to and including a break size of approximately 0.6 ft.2. The WC cowds that in order for LOCA codes to be acceptable, they must not only be capable of calculating any point on the break spectrum but must be capable of producing reliable calculations. The NRC asserts that RELAP5YA, which is the code for evaluating the portion of the break spectrum up to 0.6 ft.2, was not capable of calculating break sizes of and greater than 0.?S ft. . The NRC reasons that ifit was not possible to analyze any point on the break spectrum, it was not possible to confirm that l the limiting break within the spectrum had been identified. The NRC asserts that both t

the hianager and Lead Engineer were significant contributors to the preparation of the RELAP5YA code and should have recognized it did not meet the requirements of the regulations as interpreted by the NRC.

For their part, the hianager and Lead Engineer assert that they believed their l implementation of the RELAP5YA code, as it applied to hiaine Yankee, to be in I compliance with the Commission's regulations and, in particular,10CFR50.46 and Appendix K. They reached this conclusion based upon the successful running of the RELAP5YA code at a number of break sizes smaller than 0.35 ft. and at 0.35 ft.2 until the code terrninated. They asserted that they had identified the limiting break in the small break spectrum as being at 0.15 ft.: and had done suf6cient calculations and had a sufficient understanding of the physical phenomena which were occurring in the small break region, it, up to 0.6 ft.2, such that they could tell with reasonable assurance that a limiting break would not occur at 0.35 ft. or greater. They stated that the RELAP5YA code had not " failed" in the way alleged by the NRC, but it had terminated and had not been run further either at 0.35 ft.2 or larger because it was not necessary.

The hianager and Lead Engineer were of the view that the instabilities and oscillations which had been increasing as 0.35 ft.: was approached, were conservatively accounted for in determining the Peak Ciad Temperature ("PC1)" and would r ot have effected the ability of the code to reliably determine that the limiting break sizc for the small break spectrum had already been identined at smaller break sizes. They stated that the We use the term PCT to mean the highest temperature of the fuel cladding reached for a particular break size and also to mean the highest fuel clad temperature for any assumed break size in the spectrum, usually modified by the word " limiting."

emt . D-10 l

small break size did not contribute the limiting PCT which was in the large break spectrum covered by another code."

Our evaluation starts with 50.46(a)(1)(i). In pertinent part, that section states:

ECCS cooling performance must be calculated in accordance with an acceptable evaluation model and must be calculated for a number of postulated loss-of-coolant accidents of different sizes, locations, and other proper;ies sufficient to provide assurance that the most severe postulated loss-of coolant accidents are calculated . . . . [T]he evaluation model must include sufficient supporting justi0 cation to show that the analytical technique realistically describes the behavior of the reactor system during a loss of-coolant accident.

The Technical Review Team spent considerable time in discussing the proper interpretation of this portion of $0.46. Based upon its members' experience in implementing the cited section, they viewed that the industry standard practice i3 that an Appendix K SBLOCA evaluation model be capable of analyzing any break size within the plant's SBLOCA licensing basis. That is not to say that the calculation model had to have been run for each point or that there could not be a point within the spectrum that the model code could not be capable of generating an analytical value, but, simply, that the capability should exist. An example discussed during the Technical Review Team's deliberations was that if a code could reliably calculate peak civi temperatures at 0.3 ft.2 and 0.4 ft.2, the fact that the code was somehow incapable of performing a calculation at 0.35 ft. would not be a deficiency if the limiting break had been identified and it was understood why the code was incapable of calculating a value at that intermediate point.

Based upon discussions with the individuals involved, it was the Technical Review Team's judgment that the RELAPSYA Evaluation Model has not demonstrated the capability to analyze the complete range of the historical Maine Yankee SBLOCA break spectrum."

The Technical Review Team did state that Yankee Atomic's identification of the limiting PCT within the range of break sizes evaluated and conclusion that this was the limiting PCT for the small break spectrum based on the decreasing trend of PCT for smaller and larger break sizes, was consistent with Yankee Atomic's interpretation of 10CFR50.46 and Appendix K to 10CFR50. The Technical Review Team noted that in response to its questions, the LOCA Group provided additional verbaljustincation consisting of an explanation of SBLOCA phenomena which supported its conclusion. The Technical Review Team found that additional analysis results from other codes tend to support the M.

u g, w wm.mi D-11

Yankee Atomic position, although some of the trends are not consistent with that position.

Starting with the hypothesis that the course taken by the LOCA Group deviated from the industry standard practice to calculate PCTs over the entire small break spectrum, we examined the regulation to determine whether Yankee Atomic's interpretation,it, tne focus of the regulation was the proper identification of the limiting break size and limiting PCT, evidenced reckless indifference or utter unconcern of whether there was compliance with the regulation. We believe that one reading of 10CFR50.46 could lead to the LOCA Group's interpretation. Section 50.46 requires that "ECCS cooling performance must be calculated in accordance with an acceptable evaluation model and must im calculated for a number of oostulated Ic -of coolant accidents of different liZts, locations and other properties sufficient to provide assurance that the most severe postulated loss of-coolant accidents are calculated." (Emphasis supplied.) The focus of the LOCA Group was in providing sufficient assurance that the limiting break had been identified; this is clearly the intent of the cited regulation. By its terms, the regulation requires that the PCT for "a number" of sizes be calculated. With the limiting PCT identified, the LOCA Group believed that running the remainder of the break spectrum was unnecessary since they had conf:dence they had determined that the most severe break in the small break spectrum had been identified based upon an understanding of the physical phenomena involved.

The Technical Review Team noted that based on a broader knowledge of SDLOCA phenomena and resuhs from other codes, Yrnkee Atomic was confident that SBLOCAs were bounded by LBLOCAs and based on this expectation, Yankee Atomic accepted the results from RELAP5YA as adequate for showing compliance with the regulations.

The Technical Review Team stated its understanding that this may have been a correct conclusion, but because it was based, in part, on information beyond demonstrated results of runs of the REL ' P5YA code for Maine Yankee, the situation should have been communicated to the NRC prior to implementation. The Technical Review Team concluded that a number ofissues related to RELAPSYA implementation (as well as a number of other issues related to the Demand) should have been communicated to the NRC prior to implementation. Clearly, in hindsight, that is the case.

However, we understand the situation in which the LOCA Group found itself was not entirely of its own making and have taken this into account . our review. Maine Yankee had been told that no submittal to the NRC to satisfy TMl ltem II.K.3.31 was necessary.

Communications with the NRC were understood by the LOCA Group to be discouraged. Therefore, they did not believe they could informally communicate with the NRC. A corporate culture stressing independence, poor management oversight over the project, the LOCA Group's isolation from the remainder of the entities developing id.

omway.m u D 12

LOCA codes, interactions with Maine Yankee, and actions by the NRC also contributed to difficulties in implementing the RELAPSYA Code as applied to Maine Yankee."

With regard to our analysis of careless disregard, we considered the following elements:

1. The complete small break spectrum had not been analyzed.

_ 2. The Technical Review Team had questans as to the behaFioio5hSPCT in the unanalyzed region based upon code behavior and previous results with other codes.

3. The decision that all points on the complete small break spectrum did not have to be run came after the LOCA Group was unsuccessfulin completing the run at 0.15 ft.2 after significant effort over a number of years of working with RELAPSYA.

4. RELAPSYA was not run for a break size of 0.5 ft.2, the limiting small break size for i the Appendix K analysis of record.

5. It was a priority of the LOCA Group to complete the project.

6. If the LOCA Group had inquired of the NRC or determined its expectations from others, it could have avoided the situation.

7. On its face,10CFR50.46 would not prohibit the combination of analysis and technicaljustification based upon an understanding of physical phenomena.

8. NRC expectations regarding 10CFR50.46 were not completely documented but rather were communicated to the LOCA community through its interactions with the NRC.

9. The LOCA Group believed that the successful running of the RELAP5YA code at a number of break sizes smaller than 0.35 ft.: and at 0.35 ft.2 until the code terminated, combined with a sufficient understanding of iae physical phenomena which were occurring in the small break region.11. ep to 0.6 ft.2, and an understanding of the differences between RELAP5YA and the code of record was sufficient to comply with 10CFR50.46.

10. Members of the LOCA Group were able to provide the Technical Review Team justification which supported their code. The Technical Review Team found that such results tend to support the Yankee Atomic position although some trends do not.

. Id. at Ill.D.I.

.mww . D-13 l

11. Although it was unable to draw a definitive conclusion regarding the PCTs for the unanalyzed portion of the Maine Yankee SBLOCA spectrum, the Technical Review Team concluded that the SBLOCA PCTs for all the analyses reviewed met the 10CFR50.46 2200*F criterion and that SBLOCAs remain bounded by LBLOCAs.

12. A combination oflongstanding institutionalissues, isolation from the LOCA community, poor management, inadequate support within the Company, interactions with Maine Yankee, and actions by the NRC placed roadblocks in the way of the successful completion of this project.

13. The technical capabilities of the LOCA Group were acceptable for performing LOCA analysis, and the project included extensive assessment of RELAP5YA te scaled test facility SBLOCA data.

Weighing all these factors, we End that there was a basis for the LOCA Group's interpretation of 10CFR50.46, albeit one in contrast to NRC and industry expectations; we also find there was a basis for the LOCA Group to conclude they had identified the limiting break in the small break spectrum based upon their knowledge of the phenomena associated with small breaks. We therefore conclude that a reasonable justification existed for the LOCA Group's actions, and thus we cannot conclude that there was reckless indifference to the requirements of 10CFR50.46 or utter unconcern of the consequences in that the LOCA Group thought it was complying with a permissible interpretation of 10CFR50.46. In these circumstances, we cannot say that an individualin the LOCA Group blinded himself or herself to the realities of whether a violation would occur. We conclude that careless disregard of the Commission's requirements was not involved with respect to the issues in Section Ill.A of the Demand.

Turning now to subsection B, the NRC asserted that Maine Yankee maintained information in its files (YAEC-1868) and submitted to the NRC Core Performance Analysis Reports ("CPAR") in support of operating Cycle 14 and Cycle 15 reload analyses which were not complete and accurate in all material respects in apparent violation of 10CFR50.9(a). The NRC further asserted that the Manager never brought the deficiencies to the attention of the cognizant Manager in ch, 'e of the Engineering Section of the Licensing and Engineering Group of Maine Yanku who was the $

manager kept directly apprised by Yankee Atomic and by the Manager on the development of the specific RELAP5YA Evaluation Models.

For their part, the involved YAEC personnel maintain that they believed that they had complied with the requirements of 10CFR50.46(a)(1) and that the documents were complete and accurate in all material respects and they adequately communicated with Maine Ycnkee personnel."

Our investigation did not reach the question whether individuals at Maine Yankee may have been kr.owledgeable in the requirements of 10CFR50.46 and RELAPSYA. Ssr n.59.

sms-m D-14

i it is clear that the Yankee Atomic LOCA Group expected that YAEC was subject to audit by the NRC with regard to the application of the Evaluation Model to Maine Yankee and its compliance with Commission regulations. We believe that this belief was reasonable, particularly in light of correspondence between Maine Yankee ari the NRC i that such was contemplated."

The Technical Review Team concluded that YAEC-1868 documents the results of the I SBLOCA analysis performed by Yankee Atomic using the NRC approvea  ;

YAEC-1300P SBLOCA Evaluation Model. It further concluded that the document was l suf0ciently complete and accurate as a summary of the SBLOCA calculations that were performed. It also found that it was understandable to its intended audience and was suitable for a licensing submittal in support of Maine Yankee. The Technical Review l l Report concluded that the amount of technicalinformat:on included was appropriate for any knowledgeable person to understand the results of the analysis. The Technical

Review Team found there was no intended concealment ofinformation which would have identified any non compliance with NRC regulations. The Technical Review Team stated that YAEC-1868 could be understood by a knowledgeable engineer not trained in l the LOCA licensing process to be complete and in compliance with the regulations. The 3

Technical Review Team concluded that the compliance statement and the supporting analyses in YAEC-_1868 would be understood by an NRC reviewer and would have led i

to interactions with Yankee Atomic?

. Our review of YAEC-1868 confirmed that there were instances which erroneously stated that the complete break spectrum was analyzed. However, we believe that the document should be read as a whole. The arguments advanced by the Manager and Lead Engineer as to RELAP5YA's compliance with 10CFR50.46 are discussed in the body of the document itself. Inasmush as this document was maintained at Maine Yankee and

, subject to audit by the NRC, we believe, as does the Technical Review Team, that any i reviewer from the NRC 2'sent to audit compliance with 10CFR50.46 would have infra. Sct ahn Technical Review Report at IV.B.

J May 8,1989 letter to C.D. Frizzle, President, Maine Yankee Atomic Power Company from Patrick M. Sears, Project Manager, NRC.

Id.

2' We believe that the standard advanced by the NRC to the effect that " language would not signify to an individual without expertise in LOCA code that RELAP5YA had failed and was not capable of calculating ECCS performance"(Demand at 11)is not the appropriate one for determining whether there had been careless disregard of the regulation. We believe the issue of concealment should be judged from the point of view of a cognizant NRC reviewer knowledgeable in computer codes used to satisfy 50.46. See Virginia Electric Power comoany (North Anna Power Station, Units 1 and 2), CLI-76 22,4 NRC

.m m . D 15

understood the compliance statements and supporting analyses and would have inquired further. Finally, we observe that the Technical Review Team felt that the document was suitable for licensing2: which we understand to mean complete and accurate, subject to anticipated NRC questions and revision, as appropriate.

In addition to the considerations previously outlined in response to Section III.A of the Demand and in light of the above, we weighed the following considerations in determining whether careless disregard of NRC requirements existed related to Section III.B:

1. YAEC 1868 could be understood by a knowledgeable engineer not trained in the LOCA licensing process to be complete and in compliance with the regulations.

2. The Abstract of YAEC-1868 states that "[e] valuations were performed over a complete range of break sizes . . . ." Other sections also centain similar language.

3. YAEC 1868 was sufGeiently complete and accurate as a summary of the SBLOCA calculations that were performed. The amount of technical information included was appropriate for any knowledgeable person to understand the results.

l

4. The scope of the analysis as contained within the report is characterized correctly and there is no concealment regmiing the " failure ' of the run at 0.35 ft.: and that no runs oflarger break sizes within the small break spectrum were run.

5. But for the May 8, lW9 NRC letter, YAEC-1868 would have been submitted to the NRC.

6. Yankee Atomic had a reasonable belief that the NRC would perform an audit on YAEC-1868.

7. The statements in YAEC-1868 are consistent with the LOCA Group's understanding of the regulations.

8. The compliance statements and supporting analyses in YAEC-1868 would be understood by an NRC reviewer and would have led to interactions with Yankee Atomic.

480,4 (1976). Such a reviewer or auditor would be able to appreciate the signiGeance of the statements contained in YAEC 1868. The experts within the Technical Review Team were able to do so.

Technical Review Report at IV.B.

.m- na D 16

~ .. _ _ _ ,

_ _ _ _ _ _ _ _ - _ _ _ . _ _. __.m.____ ___

1 On this basis, we conclude that to the extent any deficiency existed, it did not represent reckless indifference to a requirement or utter unconcern whether there was compliance with the requirement. We cannot say that any individual blinded himself or herself to the realitks of whether a violation occurred. Thus, no careless disregard was involved.

C. Resnonse to Section IILC C. During Cycle 14 operations and in the Cycle 14 and Cycle 15 Cl%R, MYAPCo used an apparently unacceptable SBLOCA Evaluation Model uh!ch over predkled core cooling."

In this subsection, the NRC asserts that the LOCA Group misapplied the Alb Chambre penetration correlation, which is an empirical calculation of the penetration factor of the injected ECCS wster penetrating the downcomer annulus into the lower plenum. The NRC cmcludes that because the analysis did not provide a basis to assume full peneuation of the emergency core cooling system injection, it provides no basis to derive the loss coefficient of 600 used for the split downcomer nodalization. The NRC further asserts that these deficiencies resulted in overprediction of core cooling and overstatement of the conservatism of the model. The NRC states that the Manager and Lead Engineer should have realized during the work associated with the RELAP5YA analysis, that the Alb-Chambre correlation had been incorrectly applied, it states that adequate QA review would have revealed the errors and the unac:eptability of the RELAP5YA SBLOCA analysis described in YAEC-1868. The NRC ultimately states that based upon these errors, YAEC caused Maine Yankee to rely on an unacceptable SDLOCA Evaluation Model in violation of 10CFR50A6(a)(1).

The LOCA Group asserts that the NRC description does not reflect their reasoning concerning core penetration and the actions that they took. While admittedly there was an arithmetic error in the implementation of the Alb-Chambre correlation, the results as presented in the calculation would not cause a reviewer providing QA oversight to h:ve recognized that the Alb-Chambre correlation had been mistakenly calculated.

Furthermore, the LOCA Group states that even if the negative value of Alb-Chambre

, correlation had been discovered, a negative range,11, less than zero, would not be indicative of a meaningless result and merely would have caused them to further examine the values used in the correlation. The LOCA Group denies that such calculations also indicate other possible errors in application of the Alb-Chambre correlation.

The Technical Review Team concluded that the use of the Alb-Chambre correlation as a confirmation of th: modeling approach which includes the junction loss coefficient of 600 in the reactor downcomer is reasonable given the available data and the deficiency in u it may be argued that Section Ill.C does not allege willfulness, but merely asserts that there were specific deficiencies in the LOCA analysis and related quality assurance errors.

However, we have analyzed this issue in accordance with Section IV.B of the Demand.

l s=wmo= r, D-17 l '

i

{

l the code. Early applications of the RELAPSYA SBLOCA Evaluation Model to Maine Yankee identified excessive ECCS bypass relative to what was expected based on scaled test facility data and the results of other codes. Eventually an artificially large Icss coefficient was introduced in the junction connecting the two volumes representing a split reactor vessellower downcomer. The value of this loss coefficient was varied to obtain a balance between the expected ECCS penetration and the effect on steam venting via the break. A value of 600 was selected as an appropriate value. The amount of ECCS penetration obtained with this modeling approach was justified, in part, by the Alb-Chambre correlation. This correlation was applied to confirm that the amount of ECCS penetration predicted by IELAPSYA was conservative, The Review Team considered this approach to be an acceptable compensation for a RELAP5YA deficiency, and the use of a value of 600 obtained an amount of ECCS penetration that was consistent with industry experience. The Technical Review Team concluded that this modeling approach is not expected to result in an overprediction of core tooling, but since the calculations were not completed for all break sizes, it cannot be definitively confirmed."

An arithmetic error was made in the application of the Alb-Chambre correlation, but not identified during the quality assurance process. The Technical Review Team found this to be a failure of the quality assurance process, but was able to conclude there was a reasonaue explanation as to why it occurred. The arithmetic error did not skew the result of the calculation, which was in the range of the expected result that complete penetration was predicted. The correlation can produce results in excess of the value of 1 (in this case, a value of 8) which have the meaning of complete ECCS penetration. The Technical Review Team found that a person performing a quality assurance review is influenced by the result based on experience and expectations and the errors encountered here are the most difficult errors to identify. A correct application of the Alb-Chambre correlation without the arithmetic error would have produced negative values indicating complete ECCS bypass. It was concluded that this result would have been immediately recognized by Yankee Atomic as non-physical for the SBLOCA conditions ofinterest. The cause af the non-physical result would have been traced to excessively conservative input values. More reasonable values would then be input to the correlation, and reasonable and valid results indicating significant ECCS penetration would have resulted. Therefore, the Technical Review Team concluded that although the application of the Alb-Chambre correlation contained an error, the modeling which incorporated the loss coefficient with a value of 600 remains valid and, thus, the results of the error in rolying the Alb-Chambre correlation did not result in iuvalid input to the SBLOCA anaij > s."

M. at IV.C.

x M.

emww.oa D 18

The Technical Review Team concluded that the use of the Alb-Chambre correlation as a confirmation of the modeling approach which includes the junction loss coef0cient of 600 in the reactor vessel downcomer was reasonable given the available data and a need to address a deficiency in the code. The Technical Review Team believed that had there been a submittal of the hiaine Yankee SBLOCA application to the NRC, this modeling

• approach would have been discussed and reviewed. The Technical Review Team believed that this model could have been approved by the NRC in this form or with some revision.26 It is therefore our conclusion that thete is no inadequate analyses involved in subsection Ill.C and, therefore, the issue of deliberateness or careless disregard does nc,t arise. It is our view that the actions of the cited individuals were reasonable and there was no evidence of careless disregard.2' D. Resnonse to Section III.D D. Ml'APCo used an apparently unacceptable Best Estimate RELAPS1'A SBLOCA Evaluation Modelto calculate ECCSperfonnance, The NRC assens that hiaine Yankee performed a safety evaluation in order to determine if a decrease in steam generator pressure involved an unreviewed safety question pursuant to the requirements of 50.59 and, in doing so, used an unacceptable Best Estimate ("BE") RELAP5YA Evaluation Model in evaluating small break ECCS LOCAs in apparent violation of 50.46(a)(1). The NRC states that hiaine Yankee's March,1993 50.59 analysis relied on an analysis of the effect of reduction in steam generator pressure on ECCS performance prepared by Yankee Atomic which used, among other analyses, the Best Estimate RELAPSYA SBLOCA analysis. This document was entitled LOCA 91-04 which was approved by the Manager of the Yankee Atomic nuclear engineering department for the Manager of the LOCA Group. The 26 M.

27 While not raised by the NRC in its Demand, the Technical Review Team discussed whether the loss coefficient changes required NRC approval inasmuch as it concluded that due to the non-physical value used and the significance of this input parameter, it is an imponant model change. While the Technical Review Team recognized that the loss coefficient was literally an input to the model, it concluded that in its opinion it was an important model change (Technical keview Report at Ill.D.1, IV.C). We note that the Technical Review Team recognized that had the Maine Yankee SBLOCA application been directly submitted to the NRC, this model "could have been approved by the NRC in this form or with some revision"(M. at IV.C). Because of these factors and the confusion that -

existed due to the NRC's May 8,1989 letter, we conclude that Yankee Atomic had a basis for assuming that the NRC would have reviewed the matter in due course and, therefore, would have acted consistent with the Technical Review Team's expectations.

smtewuc, D-19

hianager received a copy of a YAEC memorandum, Impact of Lower Steam Generator Pressure on the safety analysis, NED 91-18, which relied on LOCA 9104 to evaluate the impact of reduced steam generator pressure on the small and LBLOCA analyses. A Yankee Atomic memorandum dated hiay 29,1992 entitled Steam Generator Pressure and Heat Transfer Coefficient Performance (TAG hiY-92-035) concluded that "the lower initial steam generator pressure did not affect the results of the licensing analysis,"

which conclusion was based in part on NED 91-18. The hiay 29,1992 memorandum states that it is " safety related." The hianager approved the hiay 29,1992 memorandum.

The NRC asserts that a 50.59 analysis cannot confirm that the ECCS performance will be adequate unless the 50.59 analysis uses LOCA Evaluation hiodels acceptable to demonstrate compliance with 50.46 and then reasons that the Best Estimate RELnP5YA model could not be utilized because in January,1989, the NRC approval was for use of RELAPSYA as the Appendix K Evaluation hiodel for hiaine Yankee.

Additionally, the NRC states that there were some parts of the Best Estimate model which did not comply with Part 50, Appendix K. The NRC states that it is reasonable to conclude that the hianager of the Yankee Atomic LOCA Group was aware that the Best Estimate approach deviated from the approach approved by the NRC and that a Best Estimate RELAP5YA model would not be acceptable for use in licensing matters, without NRC approval. The NRC believes it is reasonable to conclude that the hianager knew that the analysis which Yankee Atomic had performed would be used by hiaine Yankee in a 50.59 analysis or other safety analyses. The NRC concludes that in view of the intended use of the YAEC analysis, the hianager should have provided hiaine Yankee with an analysis which met NRC requirements and that by providing to hiaine Yankee an unacceptable analysis of the ef'ects of reduction steam generator pressure, Yankee Atomic caused hiaine Yankee to apparently violate 10CFR50.46(a) by relying on an unacceptable SBLOCA Evaluation hiodel to calculate ECCS cooling performance in preparing a 50.59 analysis.

%ere are two separate elements of our evaluation of this issue. The first relates to the s :ne Yankee service requests which were issued in 1990. These asked Yankee Atomic to perform what the Technical Review Team found to be a " scoping analysis" of the effect of reduced steam generator pressure on the licensing basis transients and accidents. Steam generator pressure was decreasing steadily due to fouling of the steam generator tubes. The Yankee Atomic LOCA Group was responsible for addressing the LOCA aspects of this situation."

We examined the use of the Best Estimate evaluation model to respond to those requests and the appropriateness of the response. In early 1990, due to difficulties in applyir,g the generic Appendix K RELAP5YA SBLOCA code to the hiaine Yankee plant within the prescribed time, the bianager had recommended a parallel alternative Id.

. m m m. D 20

Best Estimate Evaluation hiodel 8 lince the previously submitted topical report did not include the Dest Estimate aproach, separate NRC approval would be necessary prior to implementatiori. The hianager acommended submittal of the Best Estimate model to the NRC, presumably for NRC review and approval. The Best Estimate approach was memorialized in a memorandum dated August 1,1990. No NRC approval for the Best Estimate approach was sought by hiaine Yankee. For some time, the hianager was under the impression that the Best Estimate approach methodology report had been submitted to the NRC by hiaine Yankee for review."

LOCA-91 04, dated January 25,1991, and NED 91-18, dated January 28,1991, which incorporates LOCA 91-04, are the memoranda responses to the 1990 service requests.

In order to be responsive to the customer's request, Yankee Atomic used the only available LOCA analysis tool at that time, the Best Estimate model, to assess the effect of reduced steam generator pressure on the analysis of record. The author's characterization in LOCA-91-04 that th- Best Estimate model was the licensing basis for SBLOCA analysis at the facility was incorrect and never corrected until the issuance of TAG-MY-93 012. The analysis of record at this time was the 1977 Combustion Engineering analysis. As noted by the NRC, the hianager did not sign the 1991 memoranda."

The Technical Review Team disagreed with the NRC's premise that only approved Appendix K cvaluation models can be used in performing some scoping safety

evaluations, including input to 50.59 evaluations. The Technical Review Team found i that models such as Best Estimate models can be appropriately used provided that the l

application does not replace the analysis of record, and provided that the use of the analysis method is clearly stated and justified. It noted that if there is any doubt regarding the appropriateness of such an application, then the NRC should be consulted prior to implementing the results of the analysis. It found that Yankee Atomic's failure in this situation was in incorrectly characterizing the Best Estimate LOCA analysis as the licensing basis analysis, and furthermore not stating that the analysis used non-NRC approved methods and had restrictions on its use.32 The Technical Review Team noted that the effect of reduced steam generator pressure on the SBLOCA results for the magnitude of steam generator tube fouling and plugging that was being evaluated would not be expected to be significant. This is particularly

" January 2,1990 hiemorandom to P.L. Anderson.

" Technical Reviev Report at IV.D.

The hianager informed us that he likely would have recognized that the characterization of the Best Estimate modelin the memorandum was incorrect.

32 Id.

.mmov.m D-21

true given that the analyses of record showed the SBLOCA PCTs to be lower than the LBLOCA PCTs. The Technical Review Team found that an evaluation could have been justified without any SBLOCA analysis. An evaluation could also have been justified using the Best Estimate analysis methodology provided that suf0cient qualineation was included, and provided that the analysis of record was not replaced. Given this finding, it would be possible to conclude that no technicalinadequacy existed,it, that a non-approved LOCA evaluation model could be appropriately utilized to support a 10CFR50.59 analysis. However, we complete our analysis based on the assumption that use of the Best Estimate model was not appropriate.

We understand that the author of LOCA-91-04 may have believed he was preparing a l scoping document to answer a hypothetical question from hiaine Yankee for which the Best Estimate approach was appropriate. We reviewed the Service Requests from hiaine Yankee which resulted in LOCA 9104 and NED 91-18. We believe that based upon the amount of time permitted to complete the evaluation,it was questionable whether anyone contemplated that other than a scoping analysis was requested. The Technical Review Team agreed that a scoping analysis was requested. It is recognized that the Service Requests speak to bounding of the safety analysis; however, at the time, there was not the present sensitivity :o the term " safety analysis" or to - need to analyze design basis deviations. Because the Best Estimate model was yie; significantly higher peak c 'u temperatures than the Evaluation hiodel of recora (while still below the acceptance criterion of 2200*F set forth in 10CFR50.46(b)(1)), it may have been felt that it was more conservative to use this approach to determine whether the change in steam generator pressure caused the PCT to increase. In the circumstances, we conclude that there was a reasonable basis for the use of a Best Estimate model at the time despite the error in its representation as ths licensing basis analysis.

The second element involves a response to the April 9,1992 Service Request No.

hi-92-42, which asked for a determination of the minimum steam generator pressure that can be supported by analysis. A target value of no greater than 743 psig actual pressure was indicated as desirable. A request for the provisions of the " uncertainties that should be used with the computer and hiain Control Board indications of steam generator pressure" was also made. In response to the request to determine the minimum steam generator pressure that can be supported by the analysis, on hiay 4, 1992 Yankee Atomic stated that this portion of the request "will be completed after the mosamu D 22 l

revised SBLOCA model is completed."" The response indicated that the second half of the request had already been completed, referencing a memorandum.

TAG MY 92 030."

The May 29,1992 memorandum, TAG MY-92-035, referenced NED 91-18. It is noted i that the SBLOCA discussion in this memorandum is a very small part of the technical content and there is no mention of the referenced analysis as using the Best Estimate model. TAG MY-92 035 was referenced by Maine Yankee in a 50.59 evaluation dated April 12,1993 as part of the justification for operation with reduced steam generator pressure. Tht: analysis based on the Appendix K SBLOCA results,it, RELAPSYA, I' were forwarded to Maine Yankee also on April 12,1993, along with a draft 50.59 evaluation for Maine Yankee's use. These results were not referenced by Maine Yankee until January 13,1994, when the original Maine Yankee 50.59 evaluation was revised.

We do not believe that it was appropriate to utilize NED 91-18 which relied on a Best Estimate approach without identifying it as being a non. approved LOCA model inasmuch as at that time,ir, by April 1992, the LOCA Group knew that the Best Estimate model had not been submitted to the NRC. However, the Tec'anical Review Team did conclude that its use would have been permissible provided that the application does not replace the analysis of record and provided that the use of the i

analysis method is clearly stated and justified. This was not done. Furthermore, in 1992, 1 Yankee Atomic should have recognized that the Maine Yankee request was for more i than a scoping analysis. We believe that Yankee Atomic should have recognized that its memorandum, TAG MY-92-035, could possibly be utilized by Maine Yankee for safety related analyses.

Our next task was to determine whether the inadequacies described above were the result of careless disregard. In the 1992 time frame, utilities were only beginning to recognize that changes in plant conditions could constitute a change, test or experiment which would bring into play the requirements of 10CFR50.59. Whereas the steam ,

generator pressure was an input to various station analyses, it was not a controlled variable nor do we understand that it could be found in the Technical Specifications for 1 the facility. At the time,it was known that the LOCA small break analysis was not sensitive to the value of the steam generator pressure. This understanding was confirmed by later analysis. Generic Letter 91-18,"Information to Licensees Regarding Two NRC

" This response was contained in the YNSD Response se: tion of Service Request No. M 92 42. We understand that it was the Manager's intent from April 1992 forward that the Appendix K SBLOCA analysis be used in conjunctiun with the analysis of the lowered steam generator pressure issue. However, this intent does not appear to have been communicated to the individual at Maine Yankee performing the 50.59 review.

" We understand that as to the SBLOCA, the technical bases of the two memoranda, TAG MY-92-030 and TAG MY-92-035 are similar; bnth reference NED 91-18.

s mwm.,nca D-23

..__w _

Inspection hianual Sections on Resolution of Degraded and Nonconforming Conditions and on Operability," which was issued on November 7,1991, increased nuclear utilities' sensitivity of the need to prepare a 50.59 analyses for "as found" degraded conditions, but such appreciation does not appear to have worked its way to the LOCA Group until much later. We do not believe that the period of time between the issuance of the generic letter and the issuance of TAG hiY 92-035 was such that it was likely that the isolated Yankee Atomic LOCA Group would have appreciated the effect of Generic Letter 91-18 upc,n the Group's activities.

The most compelling piece of evidence that we obtained on this issue was an indicction that the impetus for hiaine Yankee preparing a 50.59 evaluation related to the lowered steam generator pressure issue,it, based upon the hiay 29,1992 memorandum, came from its NRC Resident inspector some months after hiaine Yankee received the hiay 29.1992 memorandum. This acts to confirm that Yankee Atomic may not have been unreasonable in not anticipating that its memo would be used in a 50.59 evaluation. If the requirement to prepare a 10CFR50.59 evaluation was not within the contemplation of the utility which was in a better position to understand its design basis, the expectations of the NRC and its own procedures to implement 10CFR50.59, this is a strong indication to us that even if Yankee Atomic should have anticipated the design basis implications of its hiay 29,1992 memorandum, its failure did not amount to careless disregard.

Also instmetive are Yankee Atomic's actions and statements in an April 12,1993, memorandum, TAG-hiY-93 012, on the same subject matter. This memorandum incorporated the lower steam generator pressure of 735 psig in the safety analysis, utilizing RELAP5YA in the Appendix K mode as promised in the Service Request. With regard to compliance with 10CFR50.59, that memorandum stated in a section entitled

" Safety Evaluation":

This memo is safety related. It provides hiaine Yankee with a single method for monitoring the full power SG pressure and overall heat transfer coefficient (UA) to verify that these parameters are within the safety analysis envelope. Since this memo theoretically expands the operatir.g space of the plant from what is actually reported in Appendix D of the FS AR, it is our opinion that incorporation of the information into Cycle 13 operation may require a 10CFR50.59 cvaluation. Attachment C contains 50.59 evaluation information (per hiaine Yankee Procedure Number 0-06-4) which can be used by hiaine Yankee to complete their formal review. It is not intended to b- a comnlete or formal 50.59 evaluation. (emphasis in original)"

It is ironic that this memorandum vias dated the same day as hiaine Yankee's approval date for its 50.59 analysis related to the hiay 29,1992 memorandum. We understand that Yankee Atomic was not aware of hiaine Yankee's 50.59 analysis at the time it issued the smtum.um D 24

l We believe that this language reDects Yankee Atomic's new appreciation of the changing attitude in the nuclear industry as to the circumstances for which a 50.59 analysis need be generated."

Our evaluation of reckless disregard considered the following factors:

i

1. The Best Estimate model was the only small break evaluation model available at the time.

2. The Best Estimate approach was yielding PCTs higher than the Evaluation hiodel of record.

3. The Best Estimate model was incorrectly indicated as being the licensing basis evaluation model.

4. In 1991, there was considerable confusion for reasons beyond the control of the LOCA Group as to the status of the Best Estimate model.

. LOCA 91-04 could have been considered a " scoping" evaluation.

6. The hianager should have known that afDrmative NRC approval of the Best Estimate model was needed and in 1992 that it had not yet been received.

7. There is no evidence that the hianager was familiar with the contents of NED 91-18 and LOCA 91-04.

8. The hianager apparently approved the hiay 29,1992 memorandum without reviewing the references, including NED 91 18 and LOCA 91-04.

9. Failing to review all references was an error, but understandable.

10. Yankee Atomic personnel should have recognized that the hiay 29,1992 memorandum had the potential to be utilized in a manner that affected the safety analysis,

11. From April 1992 forward, the hianager intended that the Appendix K SBLOCA analysis would be used 'o reanalyze a lower steam generator pressure limit than had memorandum.

Some time before this memorandum was prepared, hiaine Yankee requested that Yankee Atomic prepare draft 50.59 input for any taalyses it perforned which could affect the facility's safety evaluation. We understand that the April 12,1993 memorandum, TAG.hiY-93-012, was among the first to incorporate such an evaluation.

emmowns D-25 l

been previously assumed, but such intent was not communicated to the individual performing the 50.59 review.

12. hiaine Yankee personnel also did not contemplate the need for a 50.59 analysis relating to reduced steam generator pressure until suggested by the NRC Resident Inspector some months subsequently.

I

13. The regulatory position that degraded conditions would have to be reflected in the safety analysis via a 50.59 evaluation in the same way as affirmative changes in the design had only recently been addressed by the NRC. Apparently this guidance had not had an opportunity to have affected the thinking and actions of LOCA Group personnel with regard to steam generator pressure degradation.

14. Yankee Atomic's April 12,1993 memorandum, TAG-hiY-93 012, does contain a draft 10CFR50.59 analysis.

15. The Technical Review Team concluded that analytical techniques other than the approved Evaluation hiodel can be used to prepare 10CFR50.59 evaluations if sufficient dis.bsure is made.

We cannot conclude that there existed a reckless disregard or callous indifference toward their responsibilities or the consequence of their actions on the part of Yankee Atomic personnel. In the 1991-92 time frame, Yankee Atomic did not appreciate that steam generator pressure degradation affected the design basis; however, as this issue matured, so did Yankee Atomic's thinking as evidenced in its April 12,1993 memorandum, TAG hiY 93-012. Yankee Atomic did not reasonably have knowledge that its behavior was likely to result in a violation of NRC requirements as it reasonably could have been expected to understand them at the time. Therefore, we conclude that there has t'ot been careless disregard of the regulations.

Yll. CONCLUSION Section V.B of the Demand states as follows:

An explanation why the NRC should not consider the inadequate analyses, which apparently caused hiYAPCo to be in violation of NRC requirements, to be the result of willfulness, either deliberateness or careless disregard, on the part of YAEC and/or DE&S personnel.

The pre ei."is sections state the conclusions of the Willfulness Review Team as to its findings as eo the specific matters within the scope of the Demand. The Demand focused on the actions of two individuals in relation to whether willfulness, either deliberateness or careless disregard, existed. We found no willfulness on the part of the two individuals or any other Yankee Atomic mosow n D-26

_._J

9 personnel. We therefore believe that the Manager and Lead Engineer are capable of conducting their activities in the future in conformance with NRC requirements.

We conclude that there was nothing developed as a result of our investigation on the conduct of Yankee Atomic and/or DE&S personnel that would prevent the improvements that we understand are being made from being successful and resulting in DE&S' activities being conducted in full compliance with NRC requirements and expectations.

s mi,mm aa D 27

1 l

APPENDIX E ROOT CAUSE AND CORRECTIVE ACTION ASSESSMENT

%NEDl0%=RRM l

. _