ML20083D105

Intervenor Exhibit I-MOSBA-105,consisting of 940731 Ltr Re Util Response to NOV & Proposed Imposition of Civil Penalties

ML20083D105
Person / Time
Site:	Vogtle
Issue date:	04/18/1995
From:	Mccoy C GEORGIA POWER CO.
To:
References
EA-93-304, OLA-3-I-MOS-105, NUDOCS 9505240332
Download: ML20083D105 (56)

Text

.. . - . - . . ~. . .- -

r- ~*.J ;..J,;. n.;

, n - f4(DS8A

'/05 ExhibitMpage I off 4.c . .a r-  ::

C. K, McCoy COPY Nsw Georgia Power A

O- . $l5 '"" 95 m -3 P5 :09 w ay ou July 31,19hFILE OF SECREMRYCKETI% & f NUCIEAkkkhblATORY COMMISSION '

5,ov aq eu-7 Docket No. 50-424 Docket No. f*-o bo' A -1 Official Exh. No. M+ lof License No. NPF-68 in the mat:er of vos I, Lw f .1

~

Stff IDtiJT!FIED v' Mr. James Lieberman App 3 cant RECEIVED Director, Office of Enforcement intervenor V REJECTED U.S. Nuclear Regulatory Commission Canto oftr Attn: Document Control Clerk Contractor DATE C+fG~9T Washington, D.C. 20555 Omer witness & Ce v Reporter T. 2 :1.% s u GEORGIA POWER COMPANY VOGTLE ELECTRIC GENERATING PLANT REPLY TO NOTICE OF VIOLATION AND PROPOSED IMPOSITION OF CIVIL

Dear Mr. Lieberman:

Pursuant to 10 C.F.R. 6 2.201 Georgia Power Company ("GPC") submits the enclosed information which responds to the Notice of Violation ("NOV") issued to the Vogtle Electric Generating Plant ("VEGP") and forwarded by the NRC's May 9,1994 letter to Mr. H. Allen ['

Franklin, President and Chief Executive Officer of GPC. The NOV alleges five (5) separate violations of 10 C.F.R. I 50.9 " Completeness and Accuracy of Information." That regulation -

requires a licensee to assure that information provided to the NRC is " complete and accurate in all material respects."

As an initial matter, please rest assured that GPC and its employees fully appreciate and - l support the goal of this regulation and recognize their ongoing obligation of full candor and  :

accuracy in providing material information to the NRC. Moreover, GPC concurs with your- ,

statement in the NOV transmittal letter that, in the nuclear power industry, when errors are made, they will be promptly corrected, lessons will be learned, and corrections to procedures and training developed to improve future performance. The employees of GPC associated with its nuclear plant operations have learned from this experience and have spent the last four years since the event seeking to improve performance of its plants. We will continue to learn from .

our mistakes, or the mistakes of other licensees, and will implement our lessons learned in a safe, professional, and responsible manner.  :

O 1

~

9505240332 950418 PDR ADOCK 05000424'-

O PDR

GeorgiaPower d

- Mr. James Lieberman July 31,1994 Pare 2 GPC is aware that these violations are of significant regulatory concern to the NRC.

GPC is appreciative of NRC's recognition that the inaccuracies which are the subject of the violations did not have an effect on the safety of plant operation. Safety is paramount at GPC, as is regulatory compliance. Intent to comply with NRC rules, regulations or orders is not at issue here; GPC always intends to obey the law. What is at issue is whether mistakes were made, human mistakes by well-intentioned employees. As you will see in this reply,~ GPC admits centain of the alleged violations, but it is abundantly clear that at all times public health and safety was protected. GPC is committed to this overriding principle and nothing in this reply or its attachments should be construed otherwise.

GPC wants to assure that one central message is not lost in the NRC's detailed review of the enclosed responses. GPC recognizes fundamental failures in its performance in 1990.

It failed to maintain and use a single source document for diesel generator ("DG") starts and runs containing correct, consolidated, retrievable data and defined terminology. Inaccurate information resulting from personnel error was included in the April 9 presentation to Region II. Three starts of the IB DG were treated as " successful" and included in data provided to the NRC. While the problems experienced in these starts would not have prevented the engines -

from operating in an actual emergency, their inclusion was not supportive of GPC's intended message that the DG starts were reliable. GPC as a licensee also failed to identify the error in l

( the April 9 presentation, as documented in its April 9 letter to the NRC, until August 1990.

Concern had been expressed within the organization about the erroneous information on multiple occasions during this period of time, and two opportunities (April 19 and June 29) to identify the error were missed. GPC can do better, and it will.

On March 20,1990, during a refueling outage at VEGP Unit 1, GPC lost off-site power  !

and, when Unit l's A DG failed, GPC declared a site area emergency ("SAE"). (The other Unit  ;

1 DG was unavailable due to scheduled maintenance during the outage.) GPC immediately i recognized the imperative need to identify the causes of the event prior to returning Unit 1 to l operation, to coordinate recovery activities with the NRC, to obtain NRC concurrence in '

conducting major recovery actions, and to provide the NRC with all relevant and material J information. This was done, in many ways, over many days and with acknowledged success.  ;

The record clearly reflects GPC efforts to provide material, relevant information concerning this  ;

event, including problems encountered in recovery and investigating the reasons for the 1 A DG's  !

failure. During the course of the events, the NRC met with GPC representatives, interviewed GPC personnel, directly observed recovery activities, requested and received specific documents and records, and discussed ongoing recovery activity with many workers. By April 3,1990, with NRC team members providing objective oversight of GPC's technical review, high jacket water temperature sensors on the 1 A DG were identified as the probable component which failed O

700?75

GeorgiaPower A  !

p Mr. James Lieberman Q July 31,1994 Page 3 to perform their intended funcUon.8 GPC's open, candid and professional approach did not go unnoticed. By letter dated July 20,1990, the NRC Region II Administrator said:

Plant VEGP was fully responsive with regard to quarantined equipment, preservation of records or damaged equipment that  ;

may have been related to the event, availability of individuals for questioning, and conduct of separate investigation. Letter of Mr.

Stewart D. Ebneter to Mr. W. G. Hairston, III, entitled

" Completion of Confirmation of Action Letter Commitments".

Unfortunately, being fully responsive did not result in being painstakingly complete and precisely accurate in all cases. Although each of the incidents of alleged incompleteness or '

inaccuracy arguably are not " material", the collective performance of site and ' corporate '

personnel was below the standards which GPC expects its employees to observe.

However, we do not agree that our faulty performance was as pervasive or as significant as the NOV alleges. 'Ihis will be carefully demonstrated in the enclosed responses as we O

provide you with our perspective of these events and identity our differences to you. These differences reflect reasonable, contrary opinions of the responsibilities assigned to the involved individuals, and whether those responsibilities were fulfilled. In some cases GPC differs with ,

the NOV's analysis, affirming our earlier opinion that certain mistakes made -- or not prevented  ;

- were due at least in part to poor record keeping practices. In so doing, GPC has a broader  :

view of the " root cause," and explains conditions, acts, failures to act, and surrounding -

circumstances which bear on the events and the way they interacted to' produce the results at  ;

issue here.

GPC continues to believe that the absence of a single source document for DG starts and runs, containing timely and correct data, using commonly defined terminology, and reviewed ,

by qualified personnel, was pivotal in the underlying difficulty in providing accurate diesel start  !

data. The use of such a document would have permitted reverification and review of base data over time and eliminated the need for repeated attempts at data compilation and interpretation, i

8 The NRC's Incident Investigation Team (IIT) leader observed: l "So as far as all these sensors that are currently in quarantine, the ones that are high jacket water temperature are the ones that are of the most interest to this l event." (IIT Document 257, p. 58 and " Tape 30," April 3,1990.) .

O

)

i foo m S , i I

, - . , . . . - . _. --,-.-n. + . , .. . - --w.vn-- - .e -,-

l GeorgiaPower A Mr. James Lieberman July 31,1994 Pane 4 Its absence led to repetitive and compounding GPC reviews and NRC concerns. Simply put, different people used the same or different documents at different times to develop " start counts" with different terminology, covering different durations thereby producing a cavalcade of well-intentioned, but nonetheless ineffective communications. Record keeping should be recognized as a factor in these events, because the absence of an accurate single source document could lead to more problems in the future.

Iraportantly, GPC's attached responses also provide the NRC with additional, new information either not developed or not considered in prior NRC reviews. GPC requests that this new information be considered carefully, not only because of its significance but also, in some instances, its compelling nature. Three particular areas stand out: (1) the efforts of the 4

Unit Superintendent in the development of the transparency used at the April 9,1990 presentation at the NRC's Regio: II offices which is described in the response to Violation A; (2) information given to the NRC concerning " dew point measurements" of diesel control air between April 6 and April 12,1990, which is described in the response to Violation B; and (3) the efforts of Technical Support personnel on April 19,1990 to compile diesel start data for use by those site managers tasked with assuring the LER's accuracy, as described in response to Violation C. This additional information will also be useful in the NRC's analysis of the responses to Violations D and E.

O The NRC's bases for Violations D and E are predicated, in large part, upon a limited number of surreptitiously recorded conversations in June and August,1990. By their terms, these tape recordings do not reflect all of the facts and circumstances surrounding these events.

Moreover, the physical nature of hiding a tape recorder oftentimes resulted in an expressed concern being captured on tape, and not the statements which address or resolve the concern.

More extensively developed transcripts, therefore, are enclosed for your review and we urge that they be read fully.

l Only one GPC employee knew of the tape recording. Despite opportunities to assure an accurate and complete information flow within GPC and, in turn, to the NRC, he did not do so.

He clearly was not open and cooperative with his co-workers about resolving his own concerns.

As the transcripts reveal, sometimes he was non-responsive to direct questions, or vague and indirect when he did respond. In June of 1990 he secretly taped GPC employees as they searched for complete and accurate data and analysis to give to the NRC, while simultaneously withholding relevant and material information he po<w<wd. If, instead of withholding his information, he had fully shared it with his fellow employees, the result might have been an earlier resolution of these problems.

A few other comments are noteworthy to assure that GPC's responses are not misinterpreted or taken out of context. Four years have passed since these events. The NRC O

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

Georgia Power A Mr. James Lieberman July 31,1994 l Page 5 l has conducted a substantial and time-consuming review and has alleged that violations of NRC requirements occurred. It is important to recognize, however, that several of these violations resulted from GPC's attempts to provide information above and beyond the minimum. For example, GPC identified certain errors and informed the NRC of them, yet GPC has been faulted for the accuracy of its explanations of why those errors occurred. GPC identified conditions such as poor record keeping which " set up" personnel performance failures, yet GPC l is faulted for not assessing personnel performance. GPC is faulted also because it failed to uncover information which was available only to the NRC as a result of allegations and tape

)

I recordings. What is at work here is a fundamental difference in perspective: during the course i of these events, GPC concentrated its efforts on identifying and fixing problems with plant l operations in order to satisfy itself and the NRC that the VEGP could be operated safely after I the SAE. This was done and history has confirmed the wisdom of the restart decision and the continued improved performance of the plant. Now, however, the focus is on the performance i of individuals.2 We will always hold our employees accountable for their actions. But finding fault with good faith efforts by a licensee's employees to identify underlying causes, report them and fix them, has the potential to affect adversely open and effective communications between a licensee and the NRC. Some licensees may perceive that self-analysis that fails to find all "causes" brings with it a greater penalty than no self-analysis at all. Also, the failure to recognize that events have multiple primary causes may mislead future analysis.

O l i

2 One member of GPC's Event Review Team was prophetic. Long after the Team had issued its report, on June 29,1990 he saw the future, and said so, little realizing the accuracy of his words:

Team Member: I recall sitting in the war room the night of the event recommending that we keep a detailed log of everything we do, then we can reconstruct it. And it worked for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and they decided that - somebody decided its too cumbersome, too much work. And I do see many of these events, when they get big and take more than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, you don't keep a record, you are doomed for disaster. Because "who shot John" becomes more of an issue than what really happened. You just continually spin your wheels on what you did and who said what and what was the real test that was performed; what were the perturbations put on the system and under what conditions was it done, and everybody forgets. People get tired and they don't take notes. [ Tape 187, GPC transcript, page 31]

GeorgiaPower d Mr. James Lieberman  ;

July 31,1994 I Pane 6  !

Regardlets of the outcome of the NRC's review, GPC will continue its policy of providing more than the minimum information required and of continuing its policy of learning from its mistakes. It will not permit this experience to chill the mutual trust and effective communications with the NRC which GPC has encouraged and enjoyed over the last several years.

GPC has taken extensive action to reinforce its policy of open, accurate and candid communications with the NRC. First, GPC officers responsible for VEGP operations up to and including the President and Chief Executive Ofiker, have been personally involved in the review l of the NOV and GPC's response. A major lesson learned from this review is that internal  !

openness, accuracy, and candor in communications is a prerequisite for accurate and complete statements to the NRC. This " lesson learned" reinforces and validates the efforts in the Summer of 1990, to strengthen internal communications between the corporate office and the VEGP site l by, among other things, holding manager team-building meetings. The NRC's frank l observations to GPC officers in May of 1990, concerning our operations contributed to the recognition at that time of this weakness, l l

Second, after the issuance of the NOV GPC's Executive Vice President-Nuclear i Operations, sent a letter to nuclear operations employees which stressed the importance of l O effective communications and the effective resolution of concerns. A copy is attached. In addition to the required posting of the NOV, copies of 10 CFR i 50.9 were posted, and employees urged to read the documents.

Third, the Senior Vice President-Nuclear Operations, held meetings at both GPC plants and solicited comments and observations from large groups of plant employees. A copy of the outline for his prepared remarks is attached. These meetings were effective in providing a forum for open and self-evaluating communications, and were observed by NRC Resident Inspectors. By example, the meetings reinforced the "in full view" atmosphere which GPC strives to achieve in its relations with the NRC.

The Senior Vice President - GPC in addition to counseling with the VEGP General Manager, as discussed in the response to the Demand for Information regarding the VEGP General Manager, has also counseled the Unit Superintendent. 'Diis counseling focused upon his performance failures which are the subject of the NOV. This review also focused upon ways in which the Unit Superintendent could improve his attention to detail to ensure that his work is thorough and precise and that he communicates clearly with others.

• Finally, notice of availability of copies of this reply will be posted and circulated for reading by VEGP employees.

O

,,,,, l

4 GeorgiaPowerkh ,

Mr. James Lieberman O July 31,1994 Page 7 One final matter deserves comment. Because no enforcement conference was held prior to issuance of the NOV, this response is GPC's first real opportunity to give a complete explanation, from its perspective, of these events. It would be helpful for the parties to meet and confer, in person, to discuss the NOV and this reply. Mr. H. Allen Franklin, President and Chief Executive Officer of GPC, who has been involved in the preparation and review process of this reply, is available for such a meeting. This suggestion is not made lightly. GPC believes that the issues raised in the NOV and this reply can best be explained verbally so that neither the tone nor the content of this reply is misunderstood. GPC urges the NRC to hold such a meeting at a time and place mutually convenient to the parties.

This reply has been developed after substantial inquiry under my supervision and other GPC officers. The reply was reviewed by certain individuals familiar with these events and by the VEGP Plant Review Board for accuracy and completeness. While I do not have personal -

knowledge of all the facts as stated, I and others have thoroughly reviewed and evaluated the information. Based on all these efforts I have a high degree of confidence in the reply's accuracy. The information provided in this reply is true and correct to the best of my knowledge and belief. We are available to provide any clarification, expansion or verification which you should require. Mr. C. Kenneth McCoy states that he is the Vice President-Nuclear (Vogtle Project) of GPC and is authorized to execute this letter on behalf of GPC.

Yours very truly,  ;

k C. Kenneth McCoy SWORN TO and subscribed before m ' blst'da oflt114A / ,

)'. ()

4 M ETdROldff0C C/  ;

My Commission Expires:

'E res f9,g,

[ NOTARIAL SEAL)

O

_-m__,_ , , - , - - - -

h b

'GeorgiaPower A Mr. James Lieberman  ;

- ['. July' 31,1994

, Page 8 xc: Geornia Power Comnany Mr. J. Beasley, Jr. ,

Mr. M. Sheibani NORMS U.S. Nuclear Regulatary Comminion Mr. S. D. Ebneter, Regional Administrator '

Mr. D.S. Hood, Licensing Project Manager, NRR Mr. B.R. Bonser, Senior Resident Inspector, Vogtle

Enclosures:

1. May 11,1994 letter from W. G. Hairston, III to employees (example)

2. Remarks of Jack D. Woodard, May,1994

3. Executive Summary - Reply to Notice of Violation; EA93-304

4. Responses to Violations A through E

5. Answer to Notice of Violation O

5

O l 700775 I

I Iip ~ n2~cEr. . '$d 3- .

hiepu~om eca *:4715 Ma.kng Ado <nss 40 inverness Conlot Parway ~

Post on c. Son 1295 8.rmengham. Alabarn.135201 *

a. Teleonone 205 868 558I .

• v sxstwn *ectre system W. G. Hairston. lli e .cun.. v,c. P,.s.o.ni  !

, , sucwar oo.<.nons May 11,1994 7

i 4

1 TO ALL GEORGIA POWER EMPLOYEES ,

By now each ofyou have been made sware of the recent Notice of Violation and i proposed imposition of a $200,000 civil penalty against Georgia Power Company.

'Ihe Company is still evaluating this document, both'its factual conclusions and the legal options, and will prepare an appropriate response. The purpose of this letter,  :

though, is to assure all of our employees that Georgia Power Company remains l firmly committed to a full, open, complete and accurate communications policy.-

with the Nuclear Regulatory Commission, any of the Company's regulatory authorities, and with each other. Regardless of the outcome of the Notice of

)

)

Violation, all of us should consider it our personal responsibility that when called upon to communicate with the Nuclear Regulatory Commission or its staff, whether orally or in writing, we will do our best to ensure that the information < a provided is complete and accurate in all material respects. This is our obligation by law, this is our obligation by the terms of our licenses, but more importantly, it is the right thing to do.

We should all remember, and take seriously, that the policy of Georgia Power Company is to conduct its business affairs in an honest, ethical manner and to -

comply with all laws and regulations affecting the Company. Important to our success as a company is our success at compliance with our legal obligations.

Ifyou have a concern which you wish to raise, then you are encourage ^d to do so.

Georgia Power Company's policy is to encourage its employees, and employees of its contractors, to communicate their concems to their supervisors, which they are free to do at any time. If an employee concern cannot be resolved through this traditional channel, or if the employee wishes to pursue the matters through the concerns program, then use of that program is encouraged. In short, the Company wants you to feel free to raise any concem which you may have and has provided O

All Georgia Power Employees May 11,1994 multiple ways for you to do so. You will be treated with respect, you will be treated with courtesy, and a fair and reasonable response will be provided promptly and completely. Of course, you may always go directly to the Nuclear Regulatory Commission if you wish and the way to do this, as well as the relevant phone numbers, is posted on numerous bulletin boards throughout the work areas. Rest assured that you may raise your concerns without any fear of penalty or retaliation.

Let's all work together as a team, and dedicate ourselves to safe and efficient i i nuclear plant operations. We all have a community ofinterest in the success of our company, we all have a community ofinterest in full, open, complete and accurate communication with ourselves and with our regdlatory authorities. Let's pursue ,

these goals to the best of our individual abilities.

l O

LA). .

W.G. Hairston, III l

l l

O .

VOGTI F G

5/11/94 4:15 p.m.

I. DISCUSSION OF POLICY OF OPEN COMMUNICATION AND THE LETIER TO ALL EMPLOYEES 9

By now each of you have been made aware of the recent Notice of Violation and proposed imposition of a $200,000 civil penalty against Georgia Power Company. The Company is still evaluadng this document, both its factual conclusions and the legal i options, and it will prepare an appropriate response. The purpose of this meeting, though,.

is to ensure you all that Georgia Power Company remains firmly committed to a full, open, complete and accurate communication policy with the Nuclear Regulatory Conu..ission, any of the Company's regulatory authorities, and with each other.

Regardless of the outcome of the Notice of Violation, all of us should consider it our personal responsibility that when called upon to communicate with the Nuclear Regulatory Commission or its staff, whether orally or in writing, we will do our best to ensure that the information provided is complete and accurate in all material respects.

This is our obligation by law, this our obligation by the terms of our licenses, but more ,

importantly, itis the right thing to do. I encourage you to read the Notice ofViolation and read 10 CFR 50.9 which are posted on the plant bulletin board.

We should all remember and take seriously, that the policy of Georgia Power Cornpany is to conduct its business affairs in an honest, ethical manner and to comply with all laws and regulations affecting the Company. Important to our success as a company is our success at compliance with our legal obligations.

If you have a concem which you wish to raise, then you are encouraged to do so.

Georgia Power Company's policy is to encourage its employees, and employees ofits contractors, to communicate their concems to their supervisors, which they are free to do at any time. Rest assured that you may raise your concems without any fear of penalty or ,

l retaliation. If an employee concem cannot be resolved through this traditional channel, or if the employee wishes to pursue the matter through the concerns program, then use of that program is encouraged. In short, the Company wants you to feel free to raise any concern which you may have and has provided multiple ways for you to do so. You will l be treated with respect, you will be treated with courtesy, and a fair and reasonable response will be provided promptly and completely. Of course, you may always go directly to the Nuclear Regulatory Commission ifyou wish, and the way to do this, as well as the relevant phone numbers, is posted on numerous bulletin boards.

U.

SUMMARY

OF EVENTS l In March,1990 Vogtje Unit I was in a normal refbeling outage with one emergency diesel generator and one offsite supply transformer tagged out of service for routine maintenance. While in this condition, a truck backed in:o a transmission line support for the other supply transformer for offsite power to the unit. When the remaining emergency diesel generator attempted to start, it tripped due to a false trip signal resulting in a loss ofpower to plant safety systems. The diesel generator was subsequently started manually to restore power until offsite power us restored.

In the investigation of the causes of this event, the issue of the reliability of the diesel generators us one of the issues which needed to be resolved prior to retuming the unit to O

y 1

I operation Our employees, often under the observation of NRC inspectors, conducted

& extensise investigations and testing of these diesels before the unit was restarted.

U Subsequent to these investigations, a meeting was held with the NRC to discuss the event and all the conective actions taken to prevent recunence and ens'ure the unit was ready to return to service.

During this meeting, infonnation was provided reganiing the in estigation and testing of the diesel generators which included a summary of the number of successful test starts done on each of the diesels subsequent to the investigation to demonstrate reliability.

1 This infonnation was gathered by plant employees and was later found by one of our employees to have been in error. His error was reported verbally to the NRC. It was several months before all the confusion and errors were resolved.

While we continue to believe that all employees honestly and diligently attempted to provide accurate and complete information to the NRC, and the Notice of Violation did not attribute the error to willful conduct, clearly there are some lessons we should learn 1

1 from this experience, ne purpose of this discussion is not to debate the Notice of Violation- that is still under evaluation. ,

III. LESSONS LEARNED q 1

In light of this event and the NRC enforcement action, I would like to reiterate two important policies that are key to our operation:

1. We must always provide complete, accurate information regarding our operation to the NRC. This open and proactive sharing of all relevant and significant infonnation 1

^g-is essential--even ifit goes beyond the scope of an information request. It is important to be precise, accurate and complete in information provided and to identify the bases and qualifications of data provided.

2. All employees have an obligation to raise any concems they have to their supervisors,-

and to follow through to ensure the concerns are addressed. Supervisors and managers must be sensitive to concems raised, and must ensure the concem is resolved and appropriate feedback is provided to the person who raised the concern.

That includes any concems about the accuracy ofinformation. Even though we have particular employees and managers primarily responsible with developing and verifying letters, LERs and similar submittals to the NRC, each of us is responsible to call attention to any errors or inaccuracies in them. We also should suggest additional information which would assure that a complete and balanced message is being sent.

Supervisors and managers, as well as co-workers, have to be sensitive to concems raised; they must ensure that the concem is understood and resolved; and they should provide appropriate feedback to the person who raised the concem. It is sometimes not enough to resolve an issue in ymg mind-you need to be sure that the issue has been resolved in the other person's mind too. Sometimes you know the resolution as a matter of common sense or past experience, but you need to share that common sense or experience with your co workers.

IV. OUR OWN SELF-INTEREST Following our policies will obviously provide assurance that we fulfill our legal obligations under our license. Following the policies will also serve our long-term best interests on a broader scale as well. We need to be aware of those self-interests as we fe the various emotions that result from this case. Our natural, human reaction to a major proposed violation, as this one is, resulting from information which wits provided to the NRC, may be draw back, to think that ifless or the bare minircum ofinfonnation has O

1

. _ __ - . . . - - - . _ _ . . - - ~- . ---

7 been provided, or if no concem has been raised, no problem would exist today. In other wdrds, "you can't get into trouble if you don't say nothin'."

That approach is totally at odds with our two policies. The best hopes for our industry, and the continued success of this plant, are vitally dependent on the continued trust of the public in our actions. If we do not provide accurate and cornplete information to the NRC, we will lose that trust. If we fail to resolve concems once raised, we will not provide complete and accurate information to the NRC. One of the most effective means l l

of building and keeping the trust of the public in us may well be in our communications l

with the NRC.

1 O

1 l

1 1

l O

l i

(:e ., . " b Executive Summary ~

Reply to Notice of Violation; EA93-304 The purpose of this summary is to provide a concise articulation of the positions GPC has taken in its response to the Notice of Violation and Proposed Imposition of Civil Penalties dated May 9,1994. Because it is a summary it should not be used as a substitute for a thorough )

reading of the filed response. GPC admits that it failed to recognize that errors had occurred j in providing the initial diesel generator start counts to the NRC and failed to correct that l information in a timely manner. This should not have occurred, but it did, and GPC is I accountable for its actions.

Not all of its actions, however, deserve regulatory violations described in the NOV. In l some instances, GPC, either because of its view of the evidence or by offering new evidence, denies the allegations. Where it does, it provides a full explanation for its position. I Violation A:

The violation alleges that GPC made inaccurate statements in a letter dated April 9,1990 which was then sent to the NRC. The inaccuracies were that the 1 A and IB DGs had been started 18 and 19 times respectively, without problems or failures, n GPC admits this violation. The error in the April 9 letter did not support GPC's intended V inessage -- a resolution of concerns about DG reliability. The cause was the performance error of the Unit Superintendent and the misunderstanding between the VEGP General Manager and >

the Unit Superintendent of the beginning point of the number of starts. In addition, better record keeping practices might have prevented the violation. GPC does not agree that the VEGP General Manager gave inadequate instructions to the Unit Superintendent or inadequately assessed his work product.

Violation B: [

Here the allegation is that GPC's April 9,1990 letter to the NRC was incomplete because when it discussed air quality of the DG starting air system, GPC attributed " initial reports" of high dew points to faulty instrumentation. The NOV says that high dew points also occurred due to other causes, namely GPC's failure to use air dryers and repressurization of the air start system. >

GPC denies this violation on the very reasonable ground that " initial reports" referred to the discussions about DG air start quality in late March and early April,1990, not to a review of all historic dew point data. GPC provides new, documented information that it kept the NRC informed of dew point information, including actual high dew points on the 1A diesel control air. This violation seems to be a relatively straightforward misunderstanding that GPC is glad O

A j (, '. r

• M to have the opportunity to correct.

Victation C:

'Ihis violation arises out of a second failure of GPC to get an accurate count of DG starts. The NRC implicitly compares GPC documents: the April 19 LER (LER 90-006) and

. the June 29 cover letter to the revised LER. In the first document, GPC states that after the control systems of both DG's were subjected to a "com'prehensive test program" they were started "at least 18 times each . . ." The second document, ostensibly referring to the same

" comprehensive test program" says that there were no more than 10 and 12 consecutive

, successful starts respectively on both DGs. Since the numbers are obviously different, yet the starting point of the count appears to be the same, the NOV concludes that something must be inaccurate.

4 GPC believes that the statements in both documents are accurate (there had been at least

, 18 consecutive starts without failures or problems on April 19, as the LER said) but must

acknowledge that the LER's accuracy was fortuitous. It is admitted that no common definition existed for " comprehensive test program" among the various managers and because of this

. ambiguity, the NRC reasonably could conclude that LER 90-006 was inaccurate.

4 Simply stated, the reason for this ambiguity is inadequate attention to detail on the part of the managers who were aware of the potential ambiguity. Somewhere in the LER drafting process the term " comprehensive test program" should have been defined and commonly

understood.

GPC contends, however, that little, if any, materiality flows from this inaccuracy and that I comparing the two sets of numbers with different definitions of " comprehensive test program j of the control systems" merely creates a distinction without a difference. Under either scenario, the DG's were reliable and the LER goes on to identify correctly that the DG sensors used to monitor high jacket water temperature were the most probable cause of the DG failure. Further, j GPC informed the NRC of the LER's ambiguity and corrected it on its own. Accordingly, GPC j asks that this be treated as a non-cited violation. .

Violation D

The underlying allegation is that GPC's June 29 correspondence (a cover letter and a revised LER) is inaccurate and incomplete. There are three examples given in the NOV that establish the basis for this violation. First, the NOV alleges that the cover letter is incomplete since it promises to clarify the April 9 letter but doesn't do so.- GPC denies this. The correspcmdence clarifies an ambiguity over the meaning of a successful start, (or one "without problems or failures") by using commonly understood terms that appear in the NRC's Reg Guide L ~

, j

l l

1 .,%_ l t . m .. . . .-

i 1.108. The time period applicable to the start count is also made clear by referencing a specific l j time period (in the revised LER) and the first successful test in accordance with VEGP l Procedure 14980 (in the cover letter). The cover letter does not, however, identify that the April 9 letter was inaccurate and, hence, a violation is admitted for incompleteness.  ;

The second example says that the cover letter is inaccurate because diesel record keeping 4

practices did not cause the different diesel start numbers reported in the April 19,1990 LER and the June 29 letter. GPC disagrees. GPC continues to feel that pertinent diesel start records were neither timely processed nor evaluated and that, if they had been, this event would  :

probably not have occurred. GPC concedes, as it must, that personnel error contributed to the i ambiguity of the LER and thus admits that the third example correctly states a violation. j The reason these violations occurred are that certain on-site managers did not resolve a l concern raised by an employee about the accuracy of the June 29 cover letter. The SAER audit, '

which did not attempt to recreate the first start count attempt on April 9, is conceded to be  !

overly narrow in scope. 'Ihe employee who raised the concern also contributed to this violation, .

since at a minimum, he should have shared with his fellow employees his detailed, written  !

evaluation of the pertinent events.

Violation E: l l

Here the NRC alleges that the following statement in GPC's August 30,1990 letter was O inaccurate: "The confusion in the April 9th letter and the original LER appear to be the result of two factors. First, there was confusion in the distinction between a successful start and a i

valid test . . . . Second, an error was made by the individual who performed the count of DG j starts for the NRC April 9th letter."

GPC denies this violation. The phrase " confusion in" the April 9 letter and the LER was meant to refer to the confusion created by those documents, and not to a " root cause" of the l miscount in the April 9 letter. GPC wishes that the letter had said, "The confusion about the l April 9th letter and the original LER . . . ." If so, then it would have been clearer that the confusion evolved over time with people, specifically NRC inspectors and GPC personnel, about

]

j terminology used in the two documents. Nonetheless, the NRC understood the intended l message, as confirmed by GPC's call to the Region II representative. l Moreover, GPC contends that the inaccuracies of the April 9 letter were not due to personnel errors of the VEGP General Manager in issuing inadequate instructions to the Unit Superintendent or assessing his performance. GPC continues to believe that the personnel error of the Unit Superintendent, although possibly avoidable by availability of a single source, timely and accurate record of diesel generator starts, caused the inaccuracies in this letter.

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

GeorgiaPoner A G

V REPLY TO NOTICE OF VIOLATION EA 93-304 NRC Viohtion A:

" Contrary to the above, information provided to the NRC Region II Office by Georgia Power Company ("GPC") in an April 9,1990 letter and in an April 9,1990 oral presentation to the NRC was inaccurate in a material respect. Specifically, the letter states that: "Since March 20, the 1 A DG has been started 18 times, and the IB DG has been stared 19 times. No failures or problems have occurred during any of these starts."

These statements are inaccurate in that they represent that 19 consecutive successful starts without problems or failures had occurred on the IB Diesel Generator (DG) for the Vogtle facility as of April 9,1990, when, in fact, of the 19 starts referred to in the letter associated with the IB DG at the Vogtle facility, three of those starts had problems. Specifically, Start 132 tripped on high temperature lube oil, Start 134 tripped on low pressure jacket water and Start 136 had a high temperature jacket water trip alarm. As of April 9,1990, the IB DG had only 12 consecutive successful starts without problems or failures rather than the 19 represented by GPC. The same inaccuracy was presented to the NRC at its Region H Office during an oral presentation by GPC on April 9,1990.

O

'd The inaccuracy was material. In considering a restart decision, the NRC was especially interested in the reliability of the DGs and specifically asked that GPC address the matter in its presentation on restart. The NRC relied, in part, upon this information presented by GPC on April 9,1990 in the oral presentation and in the GPC le'.ter in reaching the NRC decision to allow Vogtle Unit I to return to power operation."

GPC Response to Violation A:

Admission or Denial of Violation A: The Violation is admitted.

Reasons for the Violation:

On Apdl 9,1990, GPC made an oral presentation to the NRC in the Region II office (RII). The presentation was in response to a verbal request by the NRC and the March 23 NRC Confirmation of Action letter and was in support of GPC's request for VEGP, Unit I restart approval. Following the April 9 oral presentation GPC submitted a letter to the NRC which contained the same diesel generator ("DG") start information that was presented during the oral presentation.1' l' References to exhibits contained throughout this Response shall follow the exhibit numbering system used in the December 17, 1993, Office of Investigation's Report No. 2 --

Prior to the April 9,1990 presentation, the VEGP General Manager was given the responsibility of preparing information for the presentation regarding issues that were raised as a result of the March 20, Site Area Emergency ("SAE"), including the special testing that was conducted on the DGs. The materials used during the presentation were assembled on transparencies immediately prior to April 9 by the VEGP General Manager with the help of subordinates. In reference to the DGs, it is clear that the VEGP General Manager desired to present an overview of the special testing, but did not intend to present a complete accounting l of all the DG starts since the March 20 event. Rather he intended to present a number of i consecutive successful starts as support for the position that'the DGs would perform their intended function. (Vogtle Coordinating Group Report, p. 2). In other words, a number of  :

successful starts of the DGs would demonstrate that the single operability test (surveillance) was  ;

not a fluke. '

The three " problem" starts (later designated as starts 132,134 and 136) which are the focus of this Violation A, occurred prior to the IB DG being declared operable.2/ Two of the problems (132 and 134) involved trip features which would be by-passed in an actual emergency.

The other start (136) involved an alarm for an emergency feature that did not result in a trip of the DG.

GPC attributes the inaccuracy to the performance of the Unit Superintendent, who was responsible for obtaining the start count information. The inclusion of the " problems" in the  :

start count of the IB DG appears to have resulted from the Unit Superintendent's beginning his count earlier in time than understood by the VEGP General Manager. Additionally, the availability of an updated single source document of recorded DG start information, which would have been evaluated by qualified personnel (c.g., Engineering Support) would have afforded the Unit Superintendent with clearer information. At the time the Unit Superintendent performed his count, record keeping practices (e.g., data sheet routing) did not permit for a timely updated document. Had there been such a document which collated all start activities with supporting i data, then this violation might never have occurred.2' l

The VEGP General Manager requested the Unit Superintendent to compile the number  ;

020R, unless otherwise indicated.

2/ The specific start number, such as start 132,134 and 136, would not be assigned until the DG log was in the process of updating on April 25,1990 (the log's list of starts was updated on May 2,1990).

2' Such a single source document would have been an updated version of Engineering Support DG start log maintained pursuant to VEGP Procedure 55038-C. A possible difference in result would be the listing of specific start activities with the specific start numbers (e.g. E-Run Bubble Test - 137; Multiple [ simulated] Starts - 138 - 141; UV Run Test - 142, etc.).

The "14" Multiple Starts placed on the transparency would then have been recognized as an aggregation of out of sequence starts. Scc, abo, NUREG-1410, Appendix J, p. J-15.

O f

..-+,,n .,- . - - - - . ~ . - . . , . . , , ~ . , = , - ., -v - - - . - - - - - - - .

of successful starts of the DG for inclusion on a draft transparency (slide) to be used during the presentation. The Unit Superintendent confirmed the number which was placed on the transparency. (Exhibit 9 p. 7). According to the Unit Superintendent, the VEGP General O Manager's instructions were to review the log books and determine how many starts had occurred without significant problems. (hL p. 3). The Unit Superintendent understood that the VEGP General Manager wanted a count of starts without "significant problems, i.e., where the diesel had started properly and reached the required voltage and frequency." The Unit Superintendent interpreted "significant problems" to be anything which would have prevented the diesel from operating in an emergency. The VEGP General Manager had the same understanding. However, as discussed in this Response, he would not have included the three '

problem starts in his count. (GPC response to NRC Staff's First Set of Interrogatories, i Response 7a and b, August 9,1993). While the VEGP General Manager may not have articulated a definitive criteria for " successful starts" when he directed the Unit Superintendent to gather successful DG starts, both viewed the task as calling for the same information. The task assigned to the Unit Superintendent was narrow in scope and straight forward, and the Unit Superintendent was competent to perform the task based on his experience and knowledge.f' A review of the context of the Unit Superintendent's task establishes that he was provided sufficient guidance to perform the task adequately and took diligent steps to perform it. He reviewed control logs and prepared a handwritten list of DG starts. (Exhibit 10, p. 26). The draft transparency was typed prior in time to the start lists prepared by the Unit Superintendent (Exhibit 113, p. 28-46), and GPC has concluded that the Unit Superintendent received a draft of the DG "Special Testing" transparency when he assembled the requested information. He assisted with the formatting of the transparency and supplied the start count numbers by modifying the transparency. (Exhibit 10, p. 26). He knew the transparency and his data were i for the presentation to the NRC relative to the Confirmation of Action Letter. (Exhibit 9, p. 5).

He did not note any starts that had significant problems. (Exhibit 9, p.15). He understood that the presentation was going to address all of the problems of the DGs that GPC had identified in troubleshooting or corrective actions and schedules for those activities.I' (Exhibit 10, p.18).  !

From recreating his count, the Unit Superintendent concluded that several post-maintenance starts were DQ1 included in his IB diesel count (Exhibit 10, p.19 - 20). Contrary to the Office of Investigations report (p. 23, evidentiary fact 34), he did not claim that he made his own I decision on the starting point of the count. He simply did not recall the VEGP General Manager's specific instructions. (Exhibit 10, p. I1). Prior to his assignment to prepare the d' The Unit Superintendent was a member of the Event Critique Team and a degreed SRO licensed Unit Superintendent.

l' The Unit Superintendent also had knowledge concerning DG starts or runs ,which were problematic. On March 23, 1990, he discussed the high lube oil trip on the IB DG which occurred on March 22,1990 (Start 132) in the early afternoon, including the precursor alarms to the trip and the likelihood that the sensor malfunctioned (Exhibit 70, Tape 10, pp. 24-25, GPC Transcript).

3-O

t 1

transparency by the VEGP General Manager, the Unit Superintendent had reviewed the logs and highlighted DG starts (Tape 19, pp. 23-24, GPC transcript, Tab A, Item 1), and during the performance of his task was observed with extensive DG start documentation by another tasked individual. (Exhibit 51 p. 2).

While the VEGP General Manager also could not recall his specific instructions regarding the point for the Unit Superintendent to start his count, he understood that the Unit Superintendent began counting 19 starts on the IB after the overhaul period. (Exhibit 12, p.

18). It should be noted that the Unit Superintendent's recall of exclusion of three " post-maintenance starts" on the engine with problems (starts.120,121 and 122) and the VEGP General Manager's understanding that three failures to start during the " overhaul period" were  !

excluded may be related. The VEGP General Manager's understanding of the beginning of the start count excluded the three starts with problems (starts 132,134 and 136) based on the time covered by the count. The Unit Superintendent'apparently evaluated each of these " post- l maintenance" starts as not having a significant problem. (Exhibit 10, p.19,20). And, since these three start problems during the overhaul period were viewed by the Unit Superintendent as not affecting emergency functions, they were included by him in his count. (Exhibit 10, p.

16, p.19).

The NRC assumes the VEGP General Manager did not determine the specific point by date or time when the Unit Superintendent began his count. However, the VEGP General Manager did determine the point when the Unit Superintendent began his count relative to activity. Activity was the subject matter of the transparency, not date and time of specific starts.

While GPC cannot recreate the relevant conversation, it appears reasonable to conclude that the Unit Superintendent, in excluding certain " post-maintenance" starts for the IB DG, was aware )

that "in overhaul" was listed on the draft transparency, and conveyed to the VEGP General Manager -- either by word or work product - that " post-maintenance" starts had been excluded.

Indeed, by stating that "after the overhaul period. . . that's when (the Unit Superintendent]

started counting these 19 starts" (Exhibit 12, p.18), or "after sensor calibration and logic testing" (Exhibit 36, p. 21), the VEGP General Manager made statements consistent with such an understanding in the relevant time frame.F The Unit Superintendent was aware that the use of the information he was providing the VEGP General Manager was for inclusion'on the transparency which would be used in a presentation to the NRC (Exhibit 9, p. 5 - for the presentation on the confirmation letter, to determine the number of starts we had had with the diesel without significant problems; Exhibit 10, p.12, - the start count would " define" the scope of the test program). He knew his F No DG starts were associated with the extensive calibration and logic testing of the DG.

" Bubble test starts" and " Multiple Starts" were associated with starts of the DGs. 'The VEGP General Manager's understanding of the Unit Superintendent's start point on the IB was reflected in taped conversations on April 19 (of which he was unaware), his testimony to the NRC in August,1990, and his conversations with co-workers in August,1990.

4 O

numbers would be used on the transparency (Exhibit 9, p. 8; Exhibit 10, p. 26). Given this contextual knowledge, any concern about the meaning of the VEGP General Manager's instructions should have prompted him to request clarification. GPC is unaware of any such concern being voiced. In conclusion, GPC has concluded that the data which the Unit Superintendent provided to the VEGP General Manager was neither exclusively oral (it was documented by the completed transparency), nor, do we believe, presented to the VEGP General Manager in a manner likely to prompt questioning of the Unit Superiritendent's performance.

The transparency describing the Special Testing of the DGs did not reference any problems or failures. (Exhibit 7). The Unit Superintendent was not involved in the preparation of the April 9,1990 letter and after the April 9 presentation, did not review the letter. (Exhibit 10, p. 60, 61). Since he was the only person who was aware that the start count information used in the April 9,1990 letter contained " problem satrts" and he did not review the April 9,  :

1990 letter, a potential opportunity to correct the inaccuracy was missed.

One of the reasons that the VEGP General Manager tasked the Unit Superintendent to review the logs and count the number of DG starts was due to the absence of a single source i Engineering Support DG start log based on data sheets for DGs. The Unit Superintendent obtained his information from a review of the Control Room's Unit Control logs and the Shift Supervisor logs. The VEGP General Manager believed that with the Unit Superintendent's experience in Operations, he would be able to obtain the requested information. The Unit Superintendent had previously reviewed the control room log to obtain DG start information for the Event Critique Team. Additionally, plant personnel, including the VEGP General Manager were aware that the Engineering Support DG start log was not up to date. Even the Supervisor -

O SAER in his review of the various sources of information available to determine the number of DG starts concluded that there was a need for more extensive updated documentation. At the time it took approximately 24 days for processing sheets to go from the control room to the engineers. (Exhibit 57, p.6). As a result, the Engineering Support DG start log would not be ,

updated on a current basis. There was not available a current source record book that a person could review to obtain the number of DG starts. As a result, since the Engineering Support DG start log was not current, it was logical for the VEGP General Manager to assign the Unit Superintendent who was familiar with the control room logs to obtain the start count information. In hindsight, had there been one current single source record containing updated information about the DGs, no need to review the control room logs would have existed and the inclusion of the problems within the 19 " successful start" count might not have occurred.

Materiality:

The NRC has concluded that the inaccuracy was material in that the NRC relied, in part, upon the information provided by GPC in the April 9 oral presentation and letter in reaching the decision to allow Unit I to return to power generation. GPC recognizes the importance of GPC statements concerning DG performance and the NRC's restart decision, but suggeits that the inaccuracy (19 versus 12) when read in context, was not significant particularly when considered with the extensive information concurrently provided to NRC experts.

O

,m -

The cause of the 1 A DG failure during the SAE was one focus of NRC and GPC

\

concerns. Immediately following the event, personnel conducted several trouble shooting starts on the 1 A DG to determine, if possible, the cause of the event. The DG started and ran without O difficulties or problems each time. The plant staff's focus was shifted to bringing the IB DG out of overhaul so that one functioning DG was available for backup power. Testing and trouble shooting on the 1 A DG was deferred while engineering personnel concentrated on the IB DG.

The problem starts which are the focus of the Violation A were known to the NRC experts. During the period March 22 through April 7, interaction between NRC and GPC personnel through meetings, interviews, document production, personal observation, telephone conferences and a review of planned testing clearly established that (1) GPC openly discussed all of the problems it experienced with the Calcon sensors, including those associated with starts 132,134 and 136 on the IB DG and (2) numerous NRC personnel, including Region II ,

personnel (Kenneth E. Brockman, Milton D. Hunt and Peter A. Taylor), were aware of the '

problems associated with the IB DG.

During the presentation to the NRC on April 9,1990, the VEGP General Manager was prepared to discuss components associated with the historic problem starts of the DGs. He presented a transparency of quarantined components which directly followed the transparency j which set out the Special Diesel Testing. This transparency identified six sensors associated with l the IB DG. With the presentation of this transparency, he was identifying specific sensors which caused problem starts coming out of overhaul on the IB DG.

Prior to the April 9,1990 presentation, the NRC was also aware of the special testing that was conducted on the IB DG. Milton D. Hunt, NRC Region 11 Inspector, witnessed special O testing for the determination of IB DG operability. (NRC Inspector Report 90-05, dated April 26, 1990). He was satisfied that the IB DG was operable; every test that was conducted on the IB DG while he was at the plant site was successful. He was not concerned with the occurrence of any failures on the IB DG prior to his witnessing of the testing. Even if GPC had shown failures on starts or runs prior to the undervoltage ("UV") test, Mr. Hunt's opinion regarding operability or the return to criticality would not have been affected. Mr. Hunt's opinion regarding restart was based on the testing he witnessed while at the plant site. (Exhibit 21).

Peter A. Taylor also believed that any failures during the trouble shooting test period would not have made a difference in the operability or reliability determination. (Exhibit 22). According to Mr. Taylor, during the presentation, had the NRC known of the problems during trouble shooting, they would have required more tests be conducted prior to restart, hul no m0Is additional icsis than ERIN actually conducted between April 9,1990 and return to power.

(Exhibit 22). Stewart D. Ebneter, Regional Administrator had final responsibility for the decision on allowing restart, and relied on the technical input from his staff, which included  ;

Milton D. Hunt. (Exhibit 18). It appears the observation of the testing, as well as the testing procedures themselves, rather than correspondence describing the number of successful starts, were influential in affecting NRC personnel judgment regarding operability and root cause identification.

I GPC strives to fulfill its obligations under 10 CFR 550.9 by assuring the completeness

- and accuracy of allinformation submitted to the NRC. However, for enforcement purposes the inaccuracy in the start count information should not be considered material in light of all the information the NRC was given regarding the SAE and the special testing of the DGs. The statements of NRC personnel support the position that they relied on the observation of and results of the special testing conducted on the DGs for a determination that the DGs were operable.2' GPC has not treated the inaccuracy lightly but, due to the interaction between GPC and the NRC during this time period, we respectfully request the NRC to reconsider whether the difference between 19 and 12 successful starts is material.

Corrective Action - Results Achieved :

GPC is committed to open, complete and accurate communication with employees and the NRC. GPC is keenly aware of its legal obligation to ensure that all information provided to the NRC is complete and accurate in all material respects. This commitment and awareness are exhibited in managements interaction with employees and NRC personnel. Since the SAE GPC has taken various actions in an attempt to inform the NRC of relevant information, to ensure that areas where GPC can improve are acted upon, as well as remind employees of GPC's commitment to open communication and the resolution of employee concerns.

On April 19,1990 the Vice President - Vogtle Project telephoned Ken Brockman of NRC Region II to assure that the IIT and Region II participants at the April 9 presentation understood the basis of the April 9th start count numbers in order to avoid any miscommunication (Exhibit  ;

36, p. 27). This evidences GPC's efforts to correct a potential ambiguity of the April 9 letter  !

when brought to management's attention. l On May 8,1990 the Vice President - Vogtle Project held a managers meeting to discuss the NRC's negative perceptions of GPC's approach to regulatory obligations. (Exhibit 74, p.

8). The Vice President - Vogtle Project informed the group that the Executive Vice President -

GPC, the Senior Vice President - Nuclear operations and he had been requested to attend a meeting in Washington with the NRC. He described in detail how the discussion addressed the NRC's perception of the personnel at Vogtle. The NRC believed that the personnel at Vogtle "have a cowboy cavalier attitude". (Exhibit 74, p.10). The tape illustrates GPC's commitment to openness. The Vice President - Vogtle Project led a very frank and critical discussion regarding personnel attitudes, performance and communication with the NRC.

In July,1990, GPC nuclear officers held two meetings in Augusta, Georgia for VEGP managers. The first meeting, on July 11,1990, at the Pinnacle Club, was a dinner, followed by discussion concerning such issues as teamwork, personnel policies, and open and effective

2' Peter A. Taylor stated that while he was not aware of any troubleshooting test failures during the IIT, any failures during such testing would not have made a difference in an I operability or reliability determination.

O E____.___ _ _____ ~ - - ~ ,

communication between groups within the organization. A follow-up meeting with a less social tone was held on July 24,1990, at the Holiday Inn. Among other items, GPC officers stated their goals of fostering better communications between the corporate office and the plant site, of greater overall candor in dealing with the issues to assure broad awareness of important safety issues, and of the corporate office's desire to support, not over manage, on-going efforts at the plant. These meetings, including the May 8th plant meeting, evidence GPC's sensitivity to NRC perceptions at that time, and efforts to enhance the information flow and cooperation between functional groups associated with the VEGP. As NRC Resident Inspectors and managers have observed, since this period improvements in communication have been commendable, and a culture of openness pervades external and internal discussions. The possibility of omitted material information or of materially inaccurate information is felt to be significantly less as well.

Weaknesses were identified in the June,1990 audit of DG record keeping practices conducted by the Safety and Engineering Analysis group. Successive revisions to procedures were thereafter initiated to provide additional assurance that data sheets which record start attempts (officially " Completion Sheet 1") were completed, and to place the completed Completion Sheet in a DG Logbook maintained and located in the control room. This Logbook provided a comprehensive collation of data sheets which was readily available. Today a bound logbook with consecutively numbered Completion Sheets is maintained in the control room.

When filled out by operators, a single impression creates two " originals," with the second (yellow) copy directly forwarded to the Engineering Support DG system engineer for completion of the diesel test evaluation section. The original remains in the control room. In contrast to past practice, the operators, rather than the system engineer, assign an identifying start number to each successive start and the " successful start" category has been deleted from the test O evaluation section of the Completion Sheet. VEGP Procedure 149801 (Diesel Generator Operability Test," Rev. 35, Sections 3.7 and 3.8 and VEGP Procedure 13145-1 " Diesel Generators," Rev. 35, Section 2.2.9).

GPC has made a concerted effort to enhance the accuracy and completeness of all communications to the NRC. GPC executive management and Region II officials, and GPC site management and NRC Resident Inspectors have periodically discussed openness and quality of communications.

On May 11,1994 the Executive Vice President - GPC forwarded a letter to employees informing them of the issuance of the Notice of Violation and reinforcing the company's policy toward openness in communications with the NRC. The letter also addresses GPC's philosophy in dealing with employee concerns (Tab A, Item 2).

On May 11,1994 the Senior Vice President - GPC visited both Hatch and Vogtle nuclear ,

plants and talked to large groups of employees regarding the Notice of Violation. He also  !

discussed the company's policy toward open communication with the NRC and stated "we must always provide complete, accurate information regarding our operation to the NRC." The NRC l

. O ,

l

Resident Inspector, attended these meetings. The Senior Vice President - GPC emphasized the point that it is essential that there be open and proactive sharing of all relevant and significant information, even ifit exceeds the scope of an information request (Tab A, Item 3). Managers and supervisors, in turn, shared these observations with their subordinates.

Notice of the NOV with a copy of 10 CFR 50.9 has been posted at both plant sites for all employees to read.

The above actions show GPC's continued commitment toward open and accurate communication with the NRC. Additionally some of.the above addressed GPC's policy regarding employee's concerns. While employees are encouraged to raise any and all concerns, it was stressed that it is not always enough to resolve the concern in one's own mind but one needs to be sure that any concern in the minds of others is also resolved. GPC is very focused on the appropriate treatment and resolution of all employee concerns. The discussions with employees was also aimed at preventing the Notice of Violation from having the effect of chilling any internal or external communications.

Corrective Action to be Taken:

GPC's Reply to the Notice of Violation will be made available for employees to read.

If an Order is issued by the NRC regarding the Notice of Violation, it will be posted for all employees to read.

The Senior Vice President - GPC will send a letter to the Vice Presidents of Hatch and Vogtle regarding the adequacy of established record keeping practices during an "off-normal" event. The letter suggests that upon the occurrence of "off-normal" events, managers consider the adequacy of existing documentation practices and whether additional action is prudent to preserve relevant information, including the need for more frequent updating of logs.

O

NRC Violatian B:

I

" Contrary to the above, information provided to the NRC Region II Office by GPC in O an April 9,1990 letter was incomplete in a material respect. Specifically, the letter states, when discussing the air quality of the DG starting air system at the Vogtle facility, that: "GPC has

! reviewed air quality of the D/G air system including dewpoint control and has concluded that l air quality is satisfactory. Initial reports of higher than expectui dewpoints were later attributed to faulty instrumentation."

This statement is incomplete in that it fails to state that actual high dew points had occurred at the Vogtle facility. It also fails to state that the causes of those high dew points included failure to use air dryers for extended periods of time and repressurization of the DG air start system receivers following maintenance.

The incompleteness was material. In considering a restart decision, the NRC was especially interested in the reliability of the DGs and specifically asked that GPC address the matter in its presentation on restart. The NRC relied, in part, upon this information presented by GPC in its letter of April 9,1990 in reaching the decision to allow Vogtle Unit I to return to power operation."

GPC Resoonse to Violation B Admission or Denial of Violation B: The violation is denied.

Reasons for the Denial:

O GPC's April 9,1990 letter addressed, accurately and completely, the on-going events at Plant Vogtle related to concerns about " dew point" data. The statement about " initial reports" referred to the resolution of a current issue which was first identified and reported to NRC representatives in the April 5-9, 1990 period (i.e., specific reports of higher than expected l dewpoints referred to reports of measurements that had been taken during the recovery from the SAE). To suggest that the letter sought to either identify or explain au " higher than expected dew points" is to take GPC's statement out of context. His would give it a meaning which is inconsistent with the actual understanding of both GPC and NRC representatives at the time.

Prior to the NRC's decision to allow Vogtle Unit 1 to return to power operation, GPC kept the NRC informed of dew point information, which included providing the NRC with written inforrr.ation on actual high dew points on the 1 A diesel control air and oral information on other engines. Documents in the possession of the NRC substantiate the context and meaning of the statement, and an understanding of the statement's meaning by NP.C representatives and of information conveyed to the NRC prior to restart. None of these documents are identified in the Notice of Violation, the February 9,1994 Vogtle Coordinating Group analysis, or the Office of Investigations report 2-90-20 dated December 17, 1993, and so, for case of reference, will be discussed here.

l l

ne Statement's Context and Meanine The April 9,1990 letter identified certain short term corrective actions. Beginning on

p. 3, the April 9 letter sets out, in summary form, these actions which had been or were being ,

implemented by GPC. The letter concludes " Rated on the above discussion, we have completed the appropriate corrective actions necessary to safely operate the unit.' ne short term corrective actions included 1) operator training, 2) modification of the under voltage start ,

feature, 3) ongoing evaluation of the possibility of design change for the jacket water high '

temperature trip function, 4) review of air quality, 5) prospective review and identification of long term corrective action, and 6) prospective laboratory evaluation of quarantine components.

The bases for GPC's conclusion of satisfactory current air quality of the DG's control system was also provided in summary form: (1) recent unacceptable dew point measurements were  ;

attributed to faulty instrumentation; (2) internal inspection of an air receiver on April 6; (3) periodic replacement of control air filters; and (4) daily air receiver blow downs.

There can be little doubt that the letter was discussing the current situation and it is unduly strained to say the statement was intended to describe all past maintenance issues. Said another way, GPC's statement regarding " dewpoint control" conveyed GPC'sjudgment on April 9,1990, that air quality in the starting air system relative to moisture or " humidity" was i satisfactory at that time. Although higher than expected dewpoints had, in fact, been recorded during the Plant's recovery from the SAE, these post event measurements acts erroneous, and faulty instrumentation Eas the reason. Representatives of the NRC's Incident Investigation Team (IIT) were informed of these high readings.

GPC Informs the NRC of Erroneous. Post-Event Dew Point R** dines The IIT transcribed many conversations between the NRC and GPC representatives regarding dewpoint control. Air quality, including the possibility of small debris or moisture

)

in the DG air system, was discussed at an IIT meeting on March 28,1990. (IIT Document 145,  ;

p. 95-97. Tab B, Item 1). In response to a question from the IIT, GPC committed to review the last historic dewpoint and, in addition, take new dewpoint readings. Both the IIT and GPC j were attempting to identify the cause of the 1 A DG spurious trip on March 20,1990.

Between March 28 and April 3, as a follow up to the IIT request, GPC tested the air quality for moisture and conducted a review of the control air filters. (IIT Document 257, p.

59-60. Tab B, Item 2). GPC stated that, based upon tests done, the quality of air was satisfactory, and air quality was not considered the root cause of the l A diesel trip on March 20,1990.

The " initial reports" of higher than expected dewpoints arose on April 6. The reports were made to GPC management and, in turn, to the IIT. (IIT Document 203, p. 4. Tab B. Item 3). The IIT team leader (Id. p. 4, line 16-18) indicated that the IIT may have been informed of the situation prior to the morning of April 6. In any event, the VEGP General Manager explained that on April 5 he had learned that the test of dewpoint on March 29 was 1 unsatisfactory for the 1 A diesell' He further stated that the preliminay indications were a bad dewpoint sensor instrument. His was the " initial reports." The " jacket water test" on the engine

, had been placed on hold, while a bleed and feed on the air storage tanks had been staned. The l basis for the General Manager's belief that the test instrumentation was suspect included recent

" bad" readings on the IB diesel. " Cooper people" (i.e., representatives of the DG vendor) had I been contacted to verify GPC's belief that any immediate problem associated with the controls of the diesel did not call into the question the operability of the engines. (IIT Document 203, l pp. 5-7. Tab B, Item 3). A new dewpoint instrument or equivalent was being sought on the morning of April 6. (Id. p. 7)2'.

By April 9, GPC had performed additional post-SAE dewpoint readings. (IIT Document 206, p. 4, lines 7-10. Tab B, Item 4). On the same day, NRC representatives were informed

, of the dewpoint readings obtained by new instrumentation. One dewpoint reading at 60.9* on the diesel 2A receiver, was attributable to the air dryer being turned off on Friday (4/6). (IIT 1 Document 206, p. 5. Tab B, item 4). The IIT team leader apparently understood this fact. '

(Exhibit 28, p.128). The IIT team leader indicated to plant personnel that Unit 2 diesel-related air quality history was not of substantial interest; "we just need the information that shows us to what extent air poor quality might have had an impact on the operation of unit I A diesel."

(IIT Document 206, p. 6, lines 23-25. Tab B, Item 4). He also observed: "you also brought up another good point, which is that, you know, the way you got into this thing here recently was you thought you had bad air, but the instrument was bad." (Id. p. 8, lines 15-16). This NRC i statement is remarkably similar to GPC's statement which is the basis for Violation B. Both accurately and completely describe recent, not historic, developments associated with dewpoint measurements of the DG control air following the SAE.

_A Rule of Remmn Should Be Aeolied to The April 9 Control Air Statement In promulgating 10 C.F.R. 50.9(b), the NRC has stated that it intended to apply a " rule of reason in assessing completeness of a communication." 52 Fed. Reg. 49366, December 31,  ;

1987. A discussion of higher than expected dewpoints in the distant past attributed to " system i air dryers being out of service" and " system repressurization following maintenance" was not  !

l 1

1/

The associated Maintenance Work Order is dated March 31,1990. MWO 19001513.

2' That " initial reports" were associated with comments to the IIT in the April 6 time frame is confirmed by other contemporaneous documentation. On April 5,1990, a facsimile transmission of the draft April 9 letter was made to Plant Vogtle. At that point, the draft letter described a test of the " jacket water system temperature transient during engine starts" as "in progress." (Tab B, Item 5, Exhibit 28, p. 85), nis draft paragraph was completed in the final

April 9 letter with inclusion of the test results. (Tab B, Item 6, p. 2, paragraph f). The final April 9 letter also discussed the air quality issue which had been resolved shortly before. These modifications, occurring over a period of a few days, highlight and underscore the efforts of GPC to provide current information on DG issues.

4

=-- r- w -- . - ,, .-- - ._- -

I reasonably necessary to completely describe the short term corrective actions associated with high dew point readings after the SAE.

(l L/

Moreover, changes in preventive maintenance (PM) practices in late 1988 made more distant dewpoint measurements much less informative about air quality than recent data.' See, for example, IIT Document 05-221-90, indicating improved PMs.8 Additional Information Providad After Anril 9.1990 and Prior to Rectut GPC provided pertinent historic data on dewpoint prior to restart. On April 9, based upon a review of preventative maintenance (PM) documentation, the NRC was informed of PM ,

results which showed unacceptable dewpoints. (IIT Document 206, pp. 7-8; Tab B, Item 4).

GPC offered the actual numbers from the PM packages, including the IB-train diesel package from March,1989 when the dryer was replaced. (IIT Document 206, p. 9, lines 3-14, Tab B, 8 NRC Inspection Report No. 50-424/425, 90-19, Supplement 1, dated November 1,1991, when compared to GPC's statements to the IIT during the 4/6 - 4/11 time frame, demonstrates the completeness of information conveyed to the NRC:

In 1988 GPC's PM program for the DG were improved; prior to that time dewpoint measurements were not consistently taken; (IIT Document 233, pp. 6-7.

Tab B, Item 7).

\ -

I The daily blowdown of receivers assured freedom of moisture; (IIT Document 233, p. 7, lines 1-5. Tab B, Item 7). l GPC's internal inspection of the DG air receiver after the SAE formed a credible basis for concluding the components had not degraded due to moisture; (IIT Document 233, p.7, lines 9-12. Tab B, Item 7).

Dewpoint measurements were above specifications in instances due to air dryers out of service following the SAE; (IIT Document 206, pp. 4-5. Tab B, Item 4).

Dewpoint measurements above specifications were due to instrument problems; (IIT Document 206, p. 4. Tab B, Item 4).

Filters on the control air system which were pulled in early March,1990 looked l new, and did not appear to have been affected by " dirty" air. (GT Document 206, l

p. 9. Tab B, Item 4). j Applying a rule of reason, the information in the April 9 letter was a complete explanation of the basis for GPC's closure of dewpoint concerns which arose subsequent to the SAE.

O  !

i

Item 4). The NRC requested a table of historic measurements (IIT Document 206, p. 7, lines 12-19; Tab B, Item 4); the 1 A diesel was of greatest interest. The next day the General Manager stated that he would like to obtain information if compressors were out of service for long periods: "that's not information we have." (Tape 40, p.17-18, April 10,1990, Tab B, Item 1

Early on April 10,1990 dewpoint measurements on the 1 A diesel were reportedly faxed {

to the NRC. (IIT Document 233, p. 6, lines 7-9, Tab B, Item 7; Tape 41, p. 74, Tab B, Item 9; IIT Document 05-202-90 has a date of April 11, Tab B, Item 10). GPC thereafter orally ,

explained that good, consistent data earlier than the transmitted data had not been obtained. '

GPC's belief remained as it had been on April 9 -- that th'e current air quality of the diesel was satisfactory, although "during that period of time" in 1988, one of the air dryers was out of service for maintenance. GPC also explair.ed the inspection of the control air filters and the lack .

of observed rust or corrosion products. GPC further referenced the one air receiver which was  !

visibly inspected (on 4/6).U' In light of these communications, documented by the IIT, the {

NRC was provided historic data. The fact that the April 9 letter had not also indicated such data i is therefore immaterial.

Thus, GPC believes that its denial of this violation should be accepted by the NRC and this violation should be withdrawn. 1 O

n' Discussions with the IIT on April 11,1990, followed a GPC morning meeting. (Exhibit 66 of the OI Report, tape 41). In addition to the consensus reached at the end of that meeting concerning air quality, the General Manager expressly told the participants that he intended to inform the IIT of the fact that the preventative maintenance program in 1988 was "not as good" as the current program. (Exhibit 66, p. 48). This he did. Conversations also indicate that the knowledgeable GPC engineers considered historic dewpoint as only tangentially related to current air quality conditions.

O

l NRC Vichtion C:  !

" Contrary to the above, information provided to the NRC by GPC in a Licensee Event O Report (LER), dated April 19,1990, was inaccurate in a material respect. Specifically, the LER states: " Numerous sensor calibrations (including jacket water temperatures), special pneumatic l

leak testing, and multiple engine starts and runs were performed under various conditions. After the 3 20-90 event, the control systems of both engines have been subjected to a comprehensive test program. Subsequent to this test program, DGI A and DGlB-have been started at least 18 times each and no failures or problems have occurred during any of these starts."

These statements are inaccurate in that they represent that at least 18 consecutive successful starts without problems or failures had occurred on the DGs for Vogtle Unit 1 (I A DG and IB ,

DG) fol*owing the completion of the comprehensive test program of the control systems for these DGs, when, in fact, following completion of the comprehensive test program of the control systems, there were no more than 10 and 12 consecutive successful starts without problems or failures of 1 A DG and IB DG respectively.

The inaccuracy was material in that knowledge by the NRC of a lesser number of consecutive successful starts on I A DG and IB DG without problems or failures could have had a natural tendency or capability to cause the NRC to inquire further as to the reliability of the DGs."

GPC's Resoonse to Violation C Admission or Denial of Violation C: Violation C is denied, as stated, but GPC admits to the ambiguity of the LER.

GPC believes that the basis for this violation is a comparison of the April 19 LER and the June 29 cover letter to the revised LER. Both documents ostensibly refer to the end of the

" comprehensive test program" yet have different DG start counts. The LER uses "at least 18" and the cover letter says "10 and 12." In fact, there had been at least 18 consecutive successful starts without problems or failures on the 1A and IB DGs going back in time as of April 19, 1990 (the date of the LER). There had also been at least 18 consecutive successful starts l without problems or failures after the " comprehensive test program of the control systems" as

{

defined by the VEGP General Manager. The staterne:2 are accurate only because additional successful starts had occurred after April 9; the statements would not have been accurate as of April 9.2 The additional starts in the intervening time period fortuitously made the statements i accurate. GPC acknowledges that the statements were ambiguous and that a common understanding of the " comprehensive test program of the control systems" phrase was not assured prior to submittal of the LER to the NRC.

&GPC officials in Birmingham thought that the phrase "at least 18 consecutive starts" in the LER found its basis in the April 9 letter.

O

Reasons for the Ambiruitv in the LER.

O The conversations reviewed by GPC in preparing this response demonstrate the efforts of GPC representatives to assure accurate and clear communications with the NRC, including LERs, consistent with the GPC's policies and NRC regulations. Several revisions of the draft LER were made to clarify information or statements. The statement which is the subject of violation C is accurate, based on the definition of the " comprehensive test program of the control systems" as understood by the VEGP General Manager and as conveyed to the General Manager

- Suppon, the Technical Suppon Manager and the acting VEGP Assistant General Manager -

Plant Support. This definition specified the beginning point of the stan count on the IB DG as understood by the VEGP General Manager. (GPC Tape 58, pp. 21-26, Tab C, Item 1; Exhibit 36, pp. 21,23-25, proposed stipulated transcript with NRC staff. Docket Nos. 50-424/425-OLA-3; ASLBP No. 93-671-01-OLA-3). The LER statement was fortuitously accurate only because of additional successful stans of the DGs between the time when the count was performed through April 19th.

Several factors contributed to the failure of GPC to assure an unambiguous statement.

Some are acts of individuals, others are conditions in which the individuals performed their tasks. First, the dm[1 LER being reviewed on April 19th in the afternoon was based, in part, on the April 9 data. The VEGP General Manager represented that this data had been verified correct by the Unit Superintendent who had gone through the operators logs, and therefore could be relied upon in finalizing the LER. He informed others, including the Technical Support Manager, the General Manager - Support and the acting Assistant General Manager - Support, about the Unit Superintendent who developed this earlier data. He also confirmed that the start count was of starts beginning after the comprehensive test program of the control systems of the DGs.' (GPC Tape 58, pp. 8-9, 26-30; Exhibit 36, pp. 9, 26, 30, Tab C, Item 1). The VEGP General Manager did not differentiate between the two Unit 1 DGs, and did not inform these other managers that the beginning point of the count on the 1 A DG might be different than for the IB DG. Although he was informed that employees working for the Technical Support Manager were verifying the LER statement's accuracy, his strong verbal assurances about the beginning point of the count and accuracy of the numbers were relied on by these managers in subsequent deliberations. Recorded conversations do not indicate that he advised that verification by Technical Support employees should be used to confirm the accuracy of his recollection. However, by established practice Technical Support would sign-off on 'he LER document prior to execution by corporate officers.

Second, the presence of an updated Engineering Support DG start log, which could be readily reviewed to confirm the Unit Superintendent's prior count, would have affected the April 19 verification efforts. (Tab C, Item 2, pp. 76-78, stipulated transcript with NRC staff and provided to the staff on June 22,1994; Docket Nos. 50-424/425-OLA-3; ASLBP No. 93-671-Ol-OLA-3). That updated log would likely have assured an accurate understanding of the April 9 data. In this regard, GPC finds the following facts extremely persuasive:

4

1) Prior to the Unit Superintendent's efforts, the Technical Support Manager informed the VEGP General Manager that the data compiled by Engineering Suppon was the traditional source of start data. For example monthly INPO reports are based on the log. (Exhibit 38, p.10 -11; p. 83);

2) ne acting Assistant Plant General Manager - Support on April 18,1990 advised the General Manager - Support that accurate statements about DG stans from March 13 through March 20 were not a problem, but that making accurate statements about DG starts up to April 18 was more difficult (Tab C, Item 3, Tape 53, Side B, p. 8 -

approximately 70% through recording).  :

3) On April 19th, when the acting Assistant Plant General Manager was tasking the verification efforts of the draft LER, he was informed by the Technical Support Manager that "the real key is that it's got to come from [ logs maintained by Engineering Support ,

employee] Kenny Stokes. . . And it's got to come from Kenny Stokes because - you know, I'm just talking about the -- telling the NRC people because Kenny Stokes is the one who makes the calls of ' valid' or ' invalid.'" (GPC Tape 57, p. 76, Exhibit 34, p.

122; Tab C, Item 2). Knowing that the DG start log kept by Engineering Support was not up to date, the acting Assistant Plant General Manager tasked his subordinates to review control room logs for the IB DG from 3-23 through 4-9 inclusive; this review would be sufficient to "do the job." (Exhibit 34, p.124; see, also, p.122, lines 22-24).D' This conclusion was after another subordinate told him that "we can't get it quickly is my problem. We don't have that information available. That's got to come from. . .." (GPC Tape 57, p. 78, Tab C, Item 2).

Had the " key" updated log been available, no doubt exists that it would have been used by these persons. No reasonable doubt exists that the task of verification would have been substantially shonened. No need would have existed to contact the Unit Superintendent, as was done belatedly on April 19th. Instead, several opportunities  ;

would have been present to assure an accurate statement in the LER. Foremost would j have been additional time for Technical Support employees,1) to determine (such as by  ;

review of outage schedules) the date and time of the sensor recalibration and logic testing l on each of the DGs and,2) to confirm that such date and time was after 1731 on 3-23 l and that 19 starts without problems or failures had occurred since that time. The specific I DG start that represented the beginning point for the count, i.c. after completion of the i comprehensive test program as defined by the VEGP General Manager, would have been  ;

identified. They could have then designated the beginning point of the count by date and i time, or specific reference to the start (e.g. E-Run Bubble Test) in lieu of the phrase

" comprehensive test program."

l D'Later conversations indicate that the compilation of starts went through April 19.

4) The Unit Superintendent's start lists were not readily available by the reviewers of the draft LER. This necessitated a new, "from scratch" attempt at data compilation by the Technical Support employees. (GPC Tape 57, pp. 78-79, Exhibit 34, pp. 89-90, Tab C, Item 2).

An updated single source document, if available, would have substantially decreased the degree of reliance placed on the VEGP General Manager's assurances. His statements could have been independently confirmed without significant effort. Rather than merely orally questioning the Unit Superintendent about his historic data compilation, site managers on April 19 could have shown him a written " total start history." He could have directly referred to the document to show his inclusion of starts which were of concern to the acting Assistant General Manager (i.e., IB starts between 3-20 and 1731 on 3-23). The reviewers then would have known not to rely on the prior data. At that point the two site managers would have a clear conflict between the VEGP General Manager's understanding of when the Unit Superintendent's count began and the Unit Superintendent's recollection. With the conflict shown, any concerns about the ambiguities in the " comprehensive test program" phraseology would have become moot.

Third, the Technical Support Manager, the acting Assistant General Manager and the General Manager - Support failed to clarify the " comprehensive test program" phraseology before issuance of the LER. The two onsite managers also failed to adequately review new datalf compiled by the Technical Support employees relative to the phraseology before issuance of the LER: " ...there was no data that we had that proved the new way we were going to word this was correct. We had the data that brought it into question and we -- and we went forward without any data that proved it correct." (Exhibit 6, p. 91). Technical Support employees who compiled the new data did not review the " comprehensive test program of the control systems" revision of the LER. One of these employees questioned the finalized language as erroneous the first time that he reviewed it, on or shortly after April 20. He was the employee who was tasked on April 19 to verify the PRB-approved LER draft language and did not attend the early afternoon PRB meeting when he was collating control room log start data. (GPC Tape 57, pp.

15-16, Exhibit 34, pp. 27 29); (Tab C, Item 2, Exhibit 6, p.103).

Fourth, the acting Assistant General Manager, when directly asked whether he took exception to the proposed wording in the final LER draft language presented to him for review voiced no concern. (GPC Tape 57, pp. 26-27; Exhibit 36, pp. 25-27; Tab C, Item 2). Egih of the other managers had suggested deleting the statements related to DG numbers. Instead of adopting this approach the acting Assistant General Manager subsequently told the Technical Support Manager that the matter should not be pursued further, even after discussions with the Unit Superintendent and further review of data provided by the Technical Support employees IFThe new data compiled on April 19 was apparently not provided to GPC's corporate office.

O

failed to prove the LER correct. (GPC Tape 58, pp. 37-38, Exhibit 36, pp. 27-28, Tab C, Item 1).

Materiality GPC acknowledges that the phrase " comprehensive test program" of the " control systems of both engines" is subject to different interpretations. GPC personnel, prior to the submittal of Licensee Event Report ("LER") 90-06 were aware of at least two connotations: (1) VEGP General Manager's connotation of after the recalibration of the Calcon sensors (Exhibit 26, Insert p. 4, lines 2-6; p. 5. lines 1-6, lines 39-44); and (2) acting Assistant General Manager's i connotation of after root cause testing of the DGs ending with the first under voltage test (Exhibit 26, p. 26; p. 33). NUREG-1410 and the June 29,1990 audit reflects a third potential connotation: completion of all of the special testing up through the first surveillance / operability test used to satisfy Technical Specifications. (NUREG-1410, Appendix J, p.13)ui. A fourth 1 connotation affecting only the 1 A diesel was suggested at the PRB on May 8,1990: after the third jacket water surge test per Temporary Engineering Test T-ENG-90 016. (PRB Minutes 90 66).

l The Notice of Violation concludes that the inaccuracy in the LER was material because  !

" knowledge by the NRC of a lesser number of consecutive successful starts on IA DG and l IB DG without problems or failures could have a natural tendency or capability to cause the  !

NRC to inquire further as to the reliability of the DG." GPC questions whether this finding of  !

materiality is appropriate for several reasons.

First, as of April 19,1990 and going back in time "at least 18 consecutive successful starts without problems or failures" had occurred on the DGs for Vogtle Unit 1. The significance of this statement arises from the number of consecutive starts without failures or problems, rather than the beginning point of the start count. It was merely intended to reflect that there had been some failures during troubleshooting as the NRC staff was aware. The demarcation of " subsequent to the comprehensive test program of the control system" is therefore immaterial with respect to influencing the NRC to inquire further as to the reliability of the DGs.

Second, determinations of materiality require careful, common-sense judgments of the context of which the information appears and the regulatory setting of the information. Virrinia Electric and Power Comnany (North Anna Power Station, Units 1 and 2), CLI-76-22,4 NRC 480, 491 (1976). GPC acknowledges the regulatory importance of accurate Licensee Event Reports. Indeed, the conversations on April 19,1990 reflect GPC employees conscientiously reviewing draft LER language in addition to the DG start statements. The mternal reviews for u/ December 23,1993 Interrogatory Response of Kenneth E. Brockman to Georgia Power's October 8,1993 First Set of Interrogatories to the NRC Staff, Responses 4 and 3; Vogtle Coordinating Group Analysis dated February 9,1994, page 14.

this LER included a request to verify the DG statement by the Senior Vice President. Technical l' Support personnel were tasked to fulfill this request. Managers and officers of the licensee commented on and discussed the draft LER. These internal reviews of the LER were as sure as they could possibly be that the submission was accurate. Even with the reviews, the result

'still was an ambiguous demarcation of the beginnlng point of the DG start count. GPC submits i

that the ambiguity of the LER does not affect the significant messages of the LER.  !

i

, That the likely components causing the 1A diesel failure had been identified as Calcon sensors used to monitor high jacket water temperature; and A factor supportive of this cause was that at least 18 consecutive, successful starts l

had been observed for the diesels.

l The more important message, both for the NRC and the industry, was the identification of the ~i suspect component to prevent future events at the VEGP and other facilities. In fact, on ]

. April 23,1990 the Vice President wrote a letter to the industry DG Owners Group which  !

contained the same DG information which was included in the April 19,1990 LER. Inthis l way, GPC assured prompt and extensive notification to the industry of potential component unreliability. The ambiguous statement would not have a natural tendency or capability to -

influence an NRC decision maker in light of its context. l l

nird, the regulatory setting of the statement should be considered. LER's are prepared ,

and filed pursuant to 10 C.F.R. 6 50.73(b) which sets forth the required contents. The cause '

of each component or system failure, if known, as well as the failure mode, mechanism and effect of each failed component, if known, must be included. Other required information is an O assessment of the safety consequences and implications of the event, and a description of any corrective action. Because the IB DG was not involved in the March 20, 1990 site area J

emergency, GPC's reference to this component was not required. The omission of the number of starts of either DG after the SAE would not have run afoul of LER reporting requirements.

Simply stating that " numerous sensor calibrations, special pneumatic leak testing, and multiple l engine starts and runs, were performed under various conditions, as well as an undervoltage test without an air roll for the 1 A engine, support GPC's identification of the suspect components" would have fulfilled GPC's reporting obligations. Therefore, the ambiguous phraseology of the LER does not appear material relative to the intended message or relative to the regulatory ,

setting. De NRC is requested to reconsider whether the ambiguity was material in light of these factors.

Corrective Actions T**n and the Reenits' Achieved:

GPC self-identified the " comprehensive test program" as a definitional problem and, on June 29,1990 submitted a revised LER to the NRC. The revised LER and a cover letter corrected the potential for misinterpreting the phrase " comprehensive test program of the control systems" by applying the most restrictive definition.

Significantly, prior to submitting of the revised LER, GPC orally notified the NRC that the LER was incorrect. On or about May 24,1990, the Regional Administrator was informed that the original LER contained incorrect numbers of DG starts, and that he would be provided O " correct" numbers of starts in a revision. (April 1,1991 letter of Mr. R.P. Mcdonald to Mr.

Thomas E. Murley, attachment 3 " Response to Hobby / acting Assistant General Manager i 2.206 Petition,"Section III.3, p. 6-7). When different " count" numbers were provided to the corporate office, the Senior Vice President assigned the independent site-based Safety Analysis and Engineering Review (SAER) group to obtain the correct number for the revised LER. A second oral notification of the Senior Vice President to the Regional Administrator occurred on June 14,1990, to inform him of changed start numbers. At this time the NRC was told that an independent review to verify the correct number had been commissioned by the Senior Vice President. The General Manager - Support contacted an NRC Region II manager in the same time frame (See, Tab C, Item 4, August 22,1990 " White Paper" entitled Diesel Starts and Failure Reporting provided to the NRC Special Operational Safety Inspection team). The Technical Support Manager informed an acting Resident inspector in the same time frame that the LER had to be revised. (GPC Tape 172, p. 31; Exhibit 24, Tab C, item 5).

Other corrective actions for Violation C, more extensively discussed in response to Violation A of the Notice of Violation, include:

• Posting of the NOV with 10 C.F.R. 6 50.9; The May 11, 1994 letter of the GPC Executive Vice President to employees; Discussions between a GPC Senior Vice President and employees at both of the O Company's nuclear plants which stressed accurate communications with the NRC and the need to resolve concerns when voiced by co-workers; and

• Management observation of communications with the NRC to assure that the NOV does not adversely affect the completeness of statements.

GPC's judgment is that these actions have preserved the enhanced communications initiated as a result of a May,1990 NRC/GPC management meeting.

Request for Treatment as Self-renorted and Corrected Violation:

GPC identified the LER statement's inaccuracy conceming DG starts, orally notified the NRC of the error,W and submitted a corrected LER on June 29,1990. It would be strained to judge the revised LER as not prompt, or as " ineffective" because the cover letter to the revised WThis notification was prior to any submission of allegations submitted to the NRC. More specifically, on or about May 24,1990, the NRC Regional Administrator was informed of the LER's error.

l 1

revised LER currently is viewed by the NRC as lacking. GPC requests that the accuracy of the cover letter be considered on its own merits, and the revised LER be considered effective corrective action for the original LER. " Generally, if the matter was promptly identified and corrected by the licensee prior to reliance by the NRC, or before the NRC raised a question

. about the information, no enforcement action will be taken for the initial inaccurate or incomplete information." 10 C.F.R. Part 2, Appendix C, IX. GPC believes this request is fair because the revised LER, standing alone, would have fulfilled regulatory obligations. The i voluntary submission of additional information in the cover letter should not be viewed as  !

negating the original LER's correction. I Moreover, in reviewing this response, the Office of Enforcement should consider the )

significance of the inaccurate information. Obviously,18 is numerically different from 10/12 -

but the enforcement process should also consider whether this is a . distinction without a difference. If 10/12 is satisfactory, then merely correcting the number 18 does nothing to undermine the ultimate conclusion of DG reliability. In essence, the question is whether at least "18" consecutive successful starts in the LER, after adjustment, is materially dissimilar from "10/12" to warrant enforcement action. This is particularly true, where, as here, it appears that the NRC did not rely on this number to reach it reliability conclusion. The NRC inspector who apparently had principal, technical input on VEGP diesel engine reliability was Milton D. Hunt.

Mr. Hunt was " pretty confident" in the results of the testing aflat the UV run test and was not

,- concerned with any failures on the "B" diesel that occurred prior to his witnessing of diesel testing. (Exhibit 21). Every test that was tried on the "B" diesel while he was at the site was successful and he was satisfied that it was operable. This technical judgment was based on l Mr. Hunt's personal witnessing of DG starts. The NRC Regional Administrator, we believe, would not have been concerned with troubleshooting phase failures - "if GPC had a subsequent O sequence of successful starts, he may have still given his permission to return to criticality, but only after his conversation with NRC staff." (Exhibit 18, p. 2). 'Ihe Regional Administrator apparently relied on the technical judgment of Mr. Hunt (Exhibit 21, p. 2). In a similar vein,-

NRC h;spector Peter A. Taylor apparently would not have been concerned about failures during troubleshooting tests since such failures "are not viewed as ' failures'" and "would not have made i a difference in an operability or reliability determination." (Exhibit 22, p.1). Based on these considerations of self-identification and reporting, and significance, GPC requests treatment of Violation C as a self-reported violation which mitigates the NOV. This treatment would re-enforce the appropriate message to licensees to correct LER inaccuracies. 10 CFR Part 2, Appendix C. VIL 1

i

~ ~

O l

-n ~,

= . . _

NRC Violation D

" Contrary to the above, information provided to the NRC by GPC in an LER cover letter O dated June 29,1990 was inaccurate and incomplete in material respects as evidenced by the following three examples:

The letter states that: "In accordance with 10 CFR 50.73, GPC hereby submits the enclosed revised report related to an event which occurred on March 20,1990. This revision is necessary to clarify the information related to the number of successful diesel generator starts as discussed in the GPC letter dated April 9,1990. . .."

1. The LER cover letter is incomplete because the submittal did not provide information regarding clarification of the April 9,1990 letter.

The incompleteness was material in that the NRC subsequently requested GPC to make a submittal clarifying the April 9,1990 letter.

The letter states that: "If the criteria for the completion of the test program is understood to be the first successful test in accordance with Vogtle Electric Generating Plant (VEGP) procedure 14980-1 " Diesel Generator Operability Test," then there were 10 successful starts of Diesel Generator 1 A and 12 successful starts of Diesel Generator 1B between the completion of the test program and the end of April 19, 1990, the date the LER-424/1990-06 was submitted to the NRC. The number of successful starts included in the original LER (at least 18) included some of the starts that were part of the test program. The difference is attributed to diesel start record keeping practices and the definition of the end of the test program."

2. The last sentence in the above paragraph is inaccurate because diesel record keeping practices were not a cause of the difference in number of diesel starts reported in the April 19,1990 LER and the June 29,1990 letter. The difference was caused by personnel errors unrelated to any problems with the diesel generator record keeping practices.

The inaccuracy was material in that it could have led the NRC to erroneously conclude that the correct root causes for the difference in the number of diesel starts reported in the April 19,1990 LER and the June 29,1990 letter had been identified by GPC.

3. The last sentence in the above paragraph is also incomplete because it failed to include the fact that the root causes for the difference in the number of diesel starts reported in the April 19, 1990 LER and the June 29, 1990 letter were personnel errors. First, the Vogtle Plant General Manager who directed the Unit Superintendent to perform the start count (which formed the basis for the April 19,1990 LER) failed to issue adequate instructions as to how to perform the count and did not adequately assess the data developed by the Unit O

4

Superintendent. In addition, the Unit Superintendent made an error in reporting -

his count. Second, the Vogtle Plant General Manager, the General Manager for Plant Support and the Technical Support Manager failed to clarify and verify the p

O starting point for the count of successful consecutive DG starts reported in the April 19,1990 LER.

The incompleteness was material in that, had correct root causes for the difference in the number of diesel starts reported in the April 19,1990 LER and the June 29,1990 letter been presented, this information could have led the NRC to seek further information."

GPC Resnonse to Violation D General Resoonse: Because this Violation identifies three examples, each requiring a response, GPC first provides this general response, and then GPC will respond to each example.

In Example 1, the Staff, concludes that GPC's cover letter to the revised LER and dated June 29,1990 was incomplete because it promised "to clarify the information related to the number of successful diesel generator starts," which first appeared in GPC's April 9,1990 correspondence, and yet failed to do so. In part, GPC denies this alleged violation and will show why the June 29 letter was a clarification. GPC admits, however, that the April 9 letter erroneously stated that no problems or failures had occurred in the identified DG starts and that this error was not recognized when the LER cover letter was transmitted. Accordingly, GPC admits that the cover letter was incomplete in that regard.

In the second example, the NOV alleges that the LER cover letter is inaccurate because it attributes the difference.in start counts reported in the April 19 LER and the June 29 LER cover letter to " diesel start record keeping practices and the definition of the end of the test program." The NOV states that personnel error, unrelated to record keeping, was the cause of the aforementioned differences. GPC denies that the June 29 letter was inaccurate, primarily because ofits strongly held belief that record keeping practices contributed to the numerous, and different, DG start counts. As will be explained in its response to Example 3, GPC does acknowledge that personnel error was also a reason for the start count differences in the two pieces of correspondence.

Example 3 is the corollary to Example 2. Its logic is that if the letter was inaccurate by not blaming the differences in DG start counts on personnel error, then it was also incomplete in the same manner. GPC admits this violation. Although GPC believes strongly that the LER cover letter was (and is) accurate, in retrospect it concurs that the letter was incomplete.

Personnel error did contribute to causing the differences in DG start counts described in the April 19 LER and June 29 letter. Unquestionably, the communication in those documents that, "no failures or problems have occurred" during the referenced DG starts needed additional explanation - a need which GPC recognized only after the documents had been sent. Said slightly differently, a violation of 10 C.F.R. I 50.9 is admitted on the basis that the LER cover O

i

,. - - - - . - . =, . - . . - - - -- -.- _, .

Y1 e

i letter was incomplem by not acknowledging that personnel error (i.e. resolution of ambiguity

- in phraseology) contributed to GPC's failure to identify and resolve the underlying errors in its 4 A April 9,1990 letter and the April 19,1990 LER. j M-Framole 1 1 Admission or Denial of Framole 1. Viointion D: -

l Example 1 is denied in part and admitted in part.

Ra==ans for Viointion. Framole 1:

The starting point for evaluating whether the June 29 LER cover letter clarified "the information related to the number of successful diesel generator starts' in GPC's April 9 letter is, naturally, the initial correspondence. In pertinent part, it said: )

"Since March 20 the 1 A DG has been started 18 times, and the IB DG has been started 19 times. No failures or problems have occurred during any of these starts."

With the revised LER GPC thought that it provided clarification of the original April 9 I letter. Vice President C.K. McCoy's own handwritten words, made in an August, .1990 l explanation say: l l

In a Licensee Event Report (LER) dated April 19,1990 (LER 50-O 424/1990-006, ELV-01545) and Revision 1 to this LER dated June 29,1990 (ELV-01729), GPC attemnted to clarify this [ April 9th information] by usine ren. muide terminolony (i.e. valid vs successful starts) and clearly definine the time neriod. (Emphasis added) (Tab D, Item 1, p.1).

A review of taped conversations occurring on April 19th confirm also that Mr. McCoy clarified a definition for the time period associated with the April 9th DG data. (Tab C, Item 1, p. 9).

The June 29th letter, using Reg. Guide terminology, also clearly defined the time period of count data (by dates of March 21 through June 7) in the revised LER. From Mr. McCoy's viewpoint, each of the documents which followed the April 9 letter attempted to clarify prior statements. The time period of the count was more accurately defined in the June 29 cover letter as beginning with the first successful test was "in accordance with Vogtle Electric Generating Plant ("VEGP") procedure, 14980-1 Diesel Generator Operability Test. . ." Also, instead of the reference in the April 9 letter to starts "without twoblems or failures" (equivalent to the

" successful starts" in the oral presentation) the LER cover letter points out the revised LER's use of commonly known terminology customarily found in LERs, or other reports, e.g. Reg Guide 1.108 terms. Instead of a subjective standard for what constitutes a " successful start" now there was a commonly understood phrase, " valid tests" that would operate as the objective O

yardstick. At a minimum, then, this point was taken from fuzzy to precise.

From the perspective of the responsible GPC officers, the cover letter was helpful in additional ways. They thought that the start count numbers referenced in the April 9 letter and April 19 LER were the same and could see plainly that the June 29 cover letter had different numbers. Thus, to them the cover letter provided additional clarification by more fully explaining its basis. It said:

1) Some of the starts in the earlier LER were part of the test program,

2) There was a definitional problem with the end of the " comprehensive test program" and

3) Poor record keeping, LL, the absence of a single source document, explained the differences.

Furthermore, the June 29 correspondence provided a start count number beginning on March 21, the same beginning time literally stated in the April 9 letter, and used " valid test" terminology which was consistent with phraseology traditionally used in NRC special reports on DG failures.E' This clearly was helpful in understanding the differences between the two earlier documents and the June 29 correspondence.

Thus, GPC contends the June 29 LER cover letter met its intended goal of providing explanatory information by correcting and clarifying the April 9 letter. It did this by its terms and, from the perspective of the GPC officers, it went beyond what was required to provide a full and complete explanation of the different start count numbers.

Knowing what it knows now, however, GPC admits that the LER cover letter should have corrected the "no problems or failures" language in the April 9 letter. There were I concerns raised: Even a casual reading of the transcript of Tape 187 demonstrates that at least j one person, Mr. Mosbaugh, was attentive to the absence of this kind of correcting language, i Other managers had the opportunity to confront the concern and resolve it, but this was not done. Thus, GPC admits that, to this extent, the June 29, LER cover letter was incomplete.

The reasons for this violation are:

1) The concern was raised but not fully addressed. Discussions between GPC site and corporate representatives afforded an opportunity to identify the error in the April 9 letter.

GPC's review of these conversations lead to the conclusion that the communication between site and corporate representatives and the failure to explore a " comment" of the former acting Assistant Plant General Manager were underlying factors in this violation. (GPC Tape 157, pp.

2-5, 9-13, 24-25, Tab D, Item 2; and Exhibit 57, p. 61). The corporate licensing engineer involved explained his rationale that the April 9 letter was not in error but, instead, was n'The June 7,1990 date for the end of the start count contained in the revised LER reflects the date through which the Technical Support personnel had developed a " valid test" start count.

O

r supported by data (the April 9 overlay). In response, the former acting Assistant Plant General Manager cryptically stated his opinion that he thought both the April 9 letter and the overlay were wrong. In this particular conversation, he neither provided supporting facts (which he O

5 possessed) nor suggested that the VEGP General Manager had instructed the Technical Support Manager to use the LER cover letter to correct the April 9 document, which he knew. (Exhibit 5, p.195,234). He remained silent. Even so, GPC admits that it had enough information to trigger additional questions to resolve the concern.

t

2) The SAER audit was probably overly narrow in scope. GPC concludes that, by June 1990, enough questions had been raised about the DG start counts, indeed GPC senior management was raising questions too, that the SAER audit probably should have attempted to i recreate the development of prior start counts. Of course, this would likely have led to the discovery of the three problem starts that were included in the April 9 numbers. Had the SAER  ;

audit examined the development of the April 19 LER " start count" or obtained the Unit '

Superintendent's start listing for verification, the inclusion of three " problem starts" in the April 9 letter would have surfaced.

3) The former acting Assistant Plant General Manager (now Technical Assistant j reporting directly to the VEGP General Manager) did not divulge that he had prepared a written l detailed evaluation which was material and relevant to the issues under discussion. As earlier I noted, GPC's review of Tape 187 has led it to conclude that the failure to explore a " comment" of the former acting Assistant Plant General Manager contributed to causing this violation. But no discussion of this violation would be complete without focusing on his opportunities to speak accurately and completely.

For example, when commenting on the June 29 cover letter, the Te:hnical Assistant said:

Mosbaugh: In addition, this particular cover letter assigns a -- attributes a reason to the errors, and whereas that statement may be correct, it is certainly not complete as to the cause of our making these mistakes and providing inaccurate information.

Greene: Mm-hmm.

Mosbaugh: We can send a half-truth out, but, you know, it seems to me at this point we  ;

ought to be coming clean. j 1

Greene: How uvuld you change the letter? (pause}

Mosbaugh: Well, it uvuld seem to me that somebody ought to explain the truth relative to the mistakes. (Emphasis added).

I What is missing, of course, is the Technical Assistant's revelation that he had already prepared detailed evaluations of DG starts and was personally involved in the LERs development and l verification.

O

)

1 After he made his concern known, he was given another opportunity for a full, articulate  ;

explanation of what to do; his response was simply not helpful. l

' Greene: Allen I don't know any - you know, I wasn't involved in the original LER and l

I don't know all the sources. I do know that we have revised the LER several times. Most PRB members are getting tired of looking at this (inaudible) the LER. We need to go ahead and just decide what we're going to submit. And that the original LER, everybody agrees, has some problems with it. This is as  ;

reasonable a way of explaining how the differences are that I can think of. You have to admit that.

I Mosbaugh: It's incomplete. {

Greene: Tell me how you would change the letter then.

Mosbaugh: We said this was going to explain the April kh letter. This doesn't explain the l April kh letter at all. (Emphasis added).

Greene: All right.

Mosbaugh: This only explains references to the comprehensive test program. The April 9th 1

letter doesn't use any words like comprehensive test program. So how did we make that mistake?

Webb: That April 9th letter also referred to 18 starts, without problems or failures.

Mosbaugh: Yeah, from the time of the event. How was that false? Why was that false?

Webb: Because there wasn't 18 starts with no problems or failures. There were starts with problems on that basis.

Mosbaugh: Well, okay, is that because we counted starts that were included in the test 7 program? No. It's a different reason.

Greene: What do you think the cause was?

Mosbaugh: The information was all available, okay.

Odom: You say the information was available, then how does it fit that the (inaudible) !

was it easily .. 7 i I

1 Mosbaugh: I don't know. You know, you're trying to ask me to state why somebody else made mistakes, okay, and I don't know how to do that. I took the same set of >

information and got right numbers. i l

Example 2 I

Admission or Denial of Examole 2. Violation D: Example 2 is denied.

Reasons for the Denial j Example 2 of Violation D addresses the accuracy and completeness of the sentence "the difference is attributed to diesel start record keeping practices and the dermition of the end of the test program." The " difference" specifically addressed by this sentence is, as recognized by the NRC, the difference in the number of DG starts reported in the April 19,1990 LER and the June 29,1990 LER cover letter. GPC has carefully considered the NRC's analysis of this sentence. It has also conducted interviews of employees with first-hand knowledge of the development of the April 19,1990 LER who were not interviewed previously by the agency.

The NRC's reviews miss the significance of. the fact that the Unit Superintendent developed a written list from control room logs prior to April 9,1990 and that a different list was developed ]

on April 19,1990 in support of the development of the LER. GPC has concluded that the NRC

is in error in concluding that personnel errors " unrelated to any problems with the diesel generator record keeping practices" was a cause of the difference in the number of starts reported O

l reasons. in the April 9,'1990 LER as compared to the June 29,1990 letter for the foll The need for an accurate record of diesel-related starts license and testing was evident immediately after the SAE. George Frederick acknowledges the importance of this need when he said, ". . .You don't keep a record, you are doomed for disaster. . .." (Exhibit 57, p. 31).

Paul Kochery, using the control room logs, fashioned a preliminary list in response to NRC inquiries (Tab D, Item 3, IIT Document 05-180-90). Kochery stated that he probably started listing these control room entries because the Engineering Support DG start log had not been updated past March 12 or 13. Similarly, tape 19 dated March 29,1990, reflects an effort by the Unit Superintendent to retrieve control room logs, so he could highlight the starts and stops and confirm the performance of the 1A diesel. 'Ihe NRC's need for up to-date data was also evident. IIT Document 212, p. 4-5 reflects NRC difficulty in developing a finaliwd, complete picture on DG starts and stops. (Tab D, Item 4). In response, the Technical Support Manager recognized that a resolution would be to make sure that the " diesel log is totally up-to-date.

That is how we keep track of our starts and stops. And we will fax you the latest copy of the diesel log." Clearly the absence of a single source document, i.e., the updated, verified, and retrievable Engineering Support DG start log, was an impediment experienced by the agency and the licensee in having a common understanding and appreciation of DG start history.

h

. Prior to the April 9,1990 presentation to the NRC Region, VEGP General Manager requested the Technical Support Manager to perform a DG start count; the Technical Support Manager informed him that his Technical Support group obtained the start count from O Engineering Support. (Exhibit 38, p.10-11; Technical Support obtained DG start log information for NRC "Special Reports" and the monthly INPO repons. Exhibit 38, p. 83).

The Unit Superintendent's compilation of starts, based on control room logs, was not precise. (Exhibit 11). A review of the list shows that IB DG starts 128 through 133 were not recorded in the control room logs.u' In addition, the Unit Superintendent appears to have made a mistake in listing an April 1,1990 " phantom" stan.. Without doubt, if the Unit Superintendent had an updated Engineering Support DG start log on April 6,1990, the omission of the unlogged starts would not have occurred.H' ,

Additionally, the lack of the single source DG log for stans after March 13,1990 also necessitated the manual development of start numbers for use in reporting, pursuant to 10 ,

C.F.R. 50.73, the valid 1 A DG failure which occurred on March 20, 1990. As reflected in April 18,1990 conversations (Tape 53, side B, approximately 70% through the recording, Tab ,

C, Item 3), the corporate office had been apprised that records in the DG log after March 13, 1990 were absent. Mosbaugh advised that making statements about DG tests after March 13-20 '

was not a problem, but "I think it becomes more difficult to make statements, you know, up to today [ April 18]." Consequently, prior to the finalization of the LER on April 19,1990, the difficulty, and possible error, inherent in making start count statements in the absence of an updated log was recognized by the acting Assistant Plant General Manager - Support.

Because of the absence of an updated DG log, on April 19 Technical Support personnel i

O were tasked to develop a list of start attempts and any associated problems from control room logs (see, generally, Response to Violation C). The list is referred to in April 19 conversation, and given the known methodology and source documents, the list would not have included starts 128 through 131 or the April 1,1990 " phantom" start on the Unit Superintendent's list. The list would have included start attempts through the afternoon of April 19. (See, Exhibit 36, p.

37, lines 12-19: "Odom counted up to the present."). For the IB diesel, five satisfactory surveillance tests under procedure 14980 were performed in this additional time frame, and one inadvertent emergency start. On June 29,1B data included starts 128 through 131, although they were not involved in the count because of the definition assigned to the " comprehensive test program."

n'The acting Assistant Plant General Manager - Support in his written allegations to the NRC also noted the absence of these starts in the Operations Department control room logs, as well as the failure of data sheet development for certain start attempts. (Exhibit 5, pp.184-186.)

HThe Unit Superintendent can not correlate the specific starts which he included in the "18" and "19" to his start count list. However, given the timing and character of starts 128 through 131, a reasonable conclusion is that the missing data affected his count.  ;

30 -

O w--m,-. , - -,y --ow- rm-w-._m--.-.. - - - - . . . . - --,,-,,,---,--.-m

Finally, GPC requests that the NRC reexamine the actual wording used in the LER cover letter in considering this response to the Notice of Violation. The sentence in question does not represent that a definitive root cause analysis had been performed on the underlying events. In O addition, the opinion-based nature of the sentence is reflected in the wording choice (i.e., "is attributed to" by GPC based on the SAER audit). A comparison is being made between two documents, not merely an identification of the cause for the first document's LER error. In the first instance DG start records were incomplete, as acknowledged by the acting Assistant Plant Manager - Support on April 18,1990 and in his July,1990 interview with the NRC. The lack of a precise " definition of the end of the test program," on April 19,1990 contributed to the original LER's error as well. Had a single source DG start document been available on April 19, 1990 and had a precise definition of the comprehensive test program been applied, the original LER would not have been in error. In the second instance, on June 29, the Engineering Support DG start log had been completed. Several additional starts omitted from prior lists had been identified. A clear difference in documentary basis had been used for the two different counts.E' Thus, GPC contends that the June 29 LER cover letter accurately states that a cause of the LER's error was record keeping practices.

Examole 3 Admission or denial of Examole 3. Violation D: Example 3 of Violation D is admitted.

1 Renmns for Violation D. Examole 3 As GPC has described in its response to Violation A and the other two examples of this Violation, it admits that the Unit Superintendent's personnel error was a cause of the inaccurate I language in the April 9 letter. GPC does not view the Plant General Manager's assignment of the task to the Unit Superintendent, or subsequent assessment of the data as involving performance failures. As set out in response to Violation C, GPC believes that the Unit Superintendent's error was supplanted by the independent verification efforts by Technical Support personnel who were tasked by and reported to the Technical Support Manager and the-

l 1

NThe acting Assistant Plant Manager - Support's efforts between April 20 and May 8,1990, demonstrate the ability to develop an accurate count for inclusion on an LER if (1) all relevant records are located and compiled in one location, and (2) an express definition is assigned to the end of the comprehensive test program. By the time he did this, though, he knew of the degree of difficulty of the task, nose who performed start counts on April 19 and before April 9 were not as fortunate. For them the value of a single source document would have been immeasurable.

1 O 1

I acting Assistant General Manager.E' But admitting to making " personnel errors" does little to correct the conditions which O permitted them to occur; and that's the only way to improve future performance. GPC believes that, regardless of the specific individuals involved, DG start record keeping practices and i

greater precision in defining ambiguous wording are causes which encompass or " bound" the various personnel performance failures which have been identified by either the NRC or GPC.

In GPC's opinion, the combination of accurate DG start records and a precise definition of the end of the comprehensive test program on April 19,1990 are the two essential elements which were missing that day. In addition, had they been present, the Unit Superintendent's earlier efforts would likely not have been used or, at the least, his error would have been caught and Corrected.

The failure of GPC employees to resolve a concern also contributed to the violation. The reason for this failure was an atypical situation in which the concern holder withheld detailed knowledge which he possessed and fellow employees did not fully explore his basis for the concerns. Had the information been forthcoming, a different weighing of knowledge and opinions might have resulted.

Corrective Action Taken - Results Achieved:

Please refer to pertinent corrective action sections of GPC's responses to Violations A and C. Rose actions are believed to encompass sufficient corrective action to preclude violations similar to Violation D. In addition, on January 2,1991, the new VEGP General Manager sent correspondence to each VEGP employee which addressed the essential nature of O frank and open communications, including the voicing of concerns. "The identification of issues which may adversely affect safety or health is a fundamental responsibility of each employee."

(Tab D, item 5).

l E' Violation D.3 refers mistakenly to the "Vogtle Plant General Manager" as failing to clarify and verify the starting point for the count for successful consecutive DG starts. He acting Assistant Plant General Manager, rather than the Plant General Manager, was aware of the ambiguity in the starting point for the count. (See, May 9,1994 Notice of Violation transmittal letter, p. 5).

O

NRC Violation E:

, " Contrary to the above, information provided to the NRC Region II Office by GPC in a letter dated August 30,1990 was inaccurate and incomplete in material respects as evidenced by the following two examples:

The letter states that: "The confusion in the April 9th letterand the original LER appear to be the result of two factors. First, there was confusion in the distinction between a successful start and a valid test... Second, an error was made by the individual who performed the count of DG starts for the NRC April 9th letter."

1. These statements are inaccurate in that confusion between a successful start and a valid test was not a cause of the error regarding DG stan counts which GPC made in its April 9,1990 letter to the NRC.

The inaccuracy was material in that it could have led the NRC to erroneously conclude that the correct root causes for the error in the April 9,1990 letter had been identified by GPC.

2. The statements are also incomplete. While an error was made by the Unit Superintendent who performed the count of diesel starts for the April 9,1990 letter, the root causes of the error in that letter were not completely identified by GPC.

Specifically, the VEGP General Manager who directed the Unit Superintendent to perform the start count failed to issue adequate instructions as to how to perform the O count and did not adequately assess the data developed by the Unit Superintendent. In addition, the Unit Superintendent did not adequately report his count to the Vogtle Plant i General Manager.

The incompleteness was material in that, had the correct root causes for the error in the  !

April 9,1990 letter regarding DG start counts been reported, this information could have led the NRC to seek further information."

GPC Resnonse to Violation E l 1

Admission or denial of Violation E: The violation is denied. ,

l Rammns for the Denial:

Violation E is based on two instances in which GPC's August 30,1990 letter is viewed as inaccurate or incomplete. In the first example, the NRC misquotes and unreasonably reads GPC's August 30 letter. 'Ihe statements in the August 30 letter are accurate, when taken in context. (Tab E, Item 1). In the second example, the NRC incorrectly concludes that the letter was incomplete when, in fact, the letter was complete relative to the letter's intended purpose. i

Examole 1 of Violation E:

The NRC letter of May 9,1994 transmitting the NOV at page 8 concludes that GPC's letter O stated "that errors in the April 9 letter and presentation and the April 19, 1990 LER were caused, in part, by confusion in the distinction between a successful start and a valid test" (emphasis supplied). First, the-NRC misquotes the GPC letter. ' Die letter actually states "the confusion in the April 9 letter and the original LER appear to be the result of two factors" (emphasis supplied). An earlier draft of the letter, referred to in conversations, used the word

" errors" but was revised. This revision avoided any suggestion that the distinction between the two terms by the Unit Superintendent was one of the reasons for the error in the April 9 letter.

Second, GPC's statement cannot reasonably be construed as stating that confusion between a successful start and a valid test was a cause of the error in GPC's April 9 letter, i.e.,

either confusion by the Unit Superintendent in performing his count, or confusion after April 9.

The letter by its express wording describes two factors which caused confusion about the April 9 letter: (1) confusion about the distinction between a successful start and a valid test and (2) an error made by the Unit Superintendent who performed the count of DG starts. After describing the historic submittals to the NRC in initial paragraphs, the letter accurately describes  ;

l two sources of confusion which developed from the time the April 9 letter was submitted, through the August,1990 special Operational Safety InspectionD'.

Third, the August 30 letter expressly states "our use of the term ' successful' was asy.sr l intended to imply a ' valid successful test' in the context of Regulatory Guide 1.108" (emphasis supplied). This sentence explains that no prior correspondence's use of " successful starts" meant ,

that the counts, or counter (s), applied formal Regulatory Guide terminology. Similarly, GPC l

O told the NRC special Operational Safety Inspection team a week or so before that "The 19 starts discussed on April 9 were based on operator assessments of the starts as successful using VEGP l

I l

procedures." Therefore, the allegedly inaccurate statement can not be read, in context, as stating that a root cause of the error in the April 9 letter was confusion between a successful start and l valid test. The statement was simply a recognition of past developments. Real confusion by some NRC employees about the terminology used in the April 9 presentation and letter had developed over time. (Exhibit 23, p. -1; questions posed by the special Operational Safety inspection team in August,1990 (e.g., Exhibit 1, p. 2 ". . .Those inspections. . . did develop enough information to indicate that there may, in fact, have been a ' counting problem' with respect to enumerating the number of starts and definine what actually constituted a valid start for countine ourooses"); (Exhibit 9, pp. 8-9; Exhibit 12, pp.11-12; and Tape 184, Exhibit 60, Insert, p.15; p. 20).

l E' Another way of stating this can be found in an earlier draft of the letter: "In hindsight this use of multiple terms [in the historic correspondence] to discuss diesel engine starts was r

confusing and in combination with the operations superintendent error in counting the number of starts resulted in the confusion of the April 9 letter." (Tab E, Item 2).

34 O

l l

l

1 l

l l

l Some GPC personnel also became confused about DG start terminology over time.

I Between June 29 and August 30, 1990 different connotations of the term " successful start" ultimately lead GPC to identify three IB DG starts designated starts 132,134 and 136 as O unsuccessful starts. Each had previously been treated as " successful starts." Based on new, specific definitions set out in the August 30 letter, starts 132 and 136 were treated as  !

unsuccessful, even though the diesel ran for one and a half and one-half hours, respectively.

(Tape 184, pp.12-14; Exhibit 29, p. 80 and Exhibit 40, p.73). For these reasons, GPC denies that the August 30 letter was inaccurate in its discussion of confusion which developed over time I about with the NRC and GPC reviews of the DG start counts.W l hmmole 2 of Violation E:

GPC's August 30,1990 letter was complete. The letter identified personnel error in the development of the April 9 letter's inaccuracy: "Second, an error was made by the individual l who performed the count of DG starts for the NRC April 9th letter." But the purpose of the letter was not to provide the NRC with a root cause analysis of the April 9 letter error. 'Ihe i purpose of the letter is expressly articulated in it: "However, during a recent inspection it was pointed out hat the revised LER did not adequately clarify the number of mrts in the April 9th letter." This purpose is as reflected in other contempo,raneous documents ". . .an appropriate report to clarify the number of simla reported April 9,1990. . .." (Tab E, Item 3 August 22, j 1990 OSI White paper, p. 2).

The August 30th letter fulfilled its intended purpose - it laid out all the starts in Tables and, using a new definition of " successful start" defined in the letter, identified starts which were O inappropriately included in the April 9 presentation and letter.

GPC acknowledges that it failed to timely recognize and correct the April 9th letter.

(See, generally, response to violation D). Notwithstanding the corporate office's desire'to understand why an LER revision was warranted (PRB 90-81 Minutes, p. 5 of 5, Tab E, Item

4) and attempts by GPC corporate office managers to obtain an explanation as to why the LER i was in error (Tape 157, Tab D, Item 5, pp. 3-10,12), we failed to identify and take timely l corrective action on the error in the April 9,1990 letter. Because no error in the April 9,1990 -l letter was identified, no effective action was taken to determine the underlying error until the  !

NRC August,1990 special Operational Safety Inspection. Based on knowledge gained in that inspection, GPC came to its conclusion that the Unit Superintendent had made an error by inclusion of start 134 in his count, as reflected in the transparency presented to the NRC on April 9. )

i Violation E is also premised on the NRC's conclusion that personnel errors by the i 1

l W The former acting Assistant Plant General Manager - Support also observed that confusion arose relative to the terminology used in counting starts. (Exhibit 92, Insert p. 3, lines 18-24).  :

I

)

. - - _ _ . - _ - - . .-_ D

l General Manager were a contributing cause of the errors in the April 9,1990 letter. As explained in response to Violation A, the narrow tasking of the Unit Superintendent by the General Manager, the Unit Superintendent's knowledge of the prospective use of his data and O the actual revision of the transparency by the Unit Superintendent for GPC's use on April 9, 1990, convince us that the General Manager did not fail to adequately task or provide sufficient oversight of the performance of the task. Although he was the " architect" of the DG transparency, the General Manager's actions were not a significant factor in, or " root cause" of, the April 9 letter's error.M' We also disagree that the special inspection in August,1990, should have prompted an assessment of the actions of the General Manager and the Unit Superintendent. De NRC's observations in the NOV transmittal letter come from conversations in an August 17, 1990 meeting, aficI a meeting with the NRC team leader but hafgIn the team's exit. The.

Vice-President was advised that the underlying issue of " intentional error" had been " basically l resolved..." The issue that the team had not resolved was why the LER and the cover letter )'

that sent it did not address 12 sequential starts. (Exhibit 68, p. 33, lines 8-14; see, also, p. 35, lines 6-7; p. 38, lines 6-16). Contemporaneous documentation also reflects the fact that the Vice President was told that the " intentional error" allegation had been resolved by the NRC. First, the NRC's exit notes reflect.this fact. Second, a letter from the General Manager to plant employees restated the NRC's conclusion. (Tab E, Composite Item 6). Third, the NRC did not request, nor do we suspect expected, an explanation of personnel error associated with the April 9 letter. De NRC had interviewed the Unit Superintendent and the General Manager. It knew that the start count was a " successful sequential" start count. (Exhibit 68, pp. 37-38). It also knew that "one successful start" was, in fact, a " failure" or " unsuccessful," and broke the sequential or consecutive "19." (Exhibit 29, p. 80; Exhibit 40, p. 73; Exhibit 68, p. 33, lines 19-23).u' Thus, GPC should not have held a heightened concern after August 17,1990, or '

i believed that the NRC desired anything other than a technical clarification of the start numbers.

From the perspective of GPC, the allegation had been resolved, and only a technical closure on start numbers and reporting of " invalid failures" remained open.

In hindsight GPC observes that the September, 1990 5 2.206 petition raised additional i questions about the April 9 letter and the LER, and prompted an NRC investigation , i.e. an

" assessment," of the performance of the General Manager and the Unit Superintendent, among others. But the purpose of the August 30,1990 letter was not to address these question _s, and was not an attempt of such an undertaking. The NRC's analysis unfairly and inaccurately n'" Successful starts" were provided to the NRC in the April 9th presentation. De April 9 letter describes starts "without problems or failures." The General Manager reasonably considemd these phrases as synonymous.

. M'On December 19, 1990, by an informal memorandum from Mark Ajiuni, the Vice-President was informed of documentary basis of the specific error (which correlates with start 134 of the IB DG).

p d

I i

< l

Exhibit N,page 9 of 9 implies that it was.

With respect to the Unit Superintendent's reporting of his count to the General Manager, O please refer to.GPC's response to Violation A.

modification of the draft transparency used on April 9.

His count was reported, in writing, by Materiality We request the NRC to reexamine its materiality fm' dings in light of the express purpose of the letter, as understood by both the NRC and GPC. On September 20,1990, the Vice President - Vogtle, called the NRC's Mr. Ken Brockman, who had been the Region H representative involved with the VEGP's recovery from the SAE and then the customary contact in the Region for the Vice President. Mr. Brockman and the Vice President discussed the August 30 letter and, as reflected in contemporaneous notes, Mr. Brockman " indicated they had all [the] info [rmation) and understood what occurred." (Exhibit 29, p.106, p. 72). 'Iherefore, at the time the NRC, which had conducted a detailed " assessment" did not view the letter as incomplete. These facts indicate that the materiality finding in Violation E is based on hindsight review of September,1990 allegations addressed in the OI report, rather than an examination of whether the information was material to the NRC at the time. The purpose of the August 30 letter was not to address those allegations.

Moreover, the NRC's determination of materiality is in the abstract, without a meaningful examination of whether the allegedly omitted information would have been considered by reasonable staff experts. We suggest the omitted information was neither requested nor had a potential bearing on the issues under discussion in late August,1990. The omitted information O could not have led to further inquiry, because the relevant issues had been resolved. ,

P J

4

. . - - - - - - - . - . . - - - _ - - . -- .