ML20207N629

Gpu Nuclear Corp Proposed Findings of Fact & Conclusions of Law (Submitted to Presiding Board in Form of Recommended Decision).* Mgt Did Not Perform,Condone or Know of Leak Rate Test Manipulation.Certificate of Svc Encl

ML20207N629
Person / Time
Site:	Three Mile Island
Issue date:	01/09/1987
From:	Blake E GENERAL PUBLIC UTILITIES CORP., SHAW, PITTMAN, POTTS & TROWBRIDGE
To:
References
CON-#187-2182 86-519-02-SP, 86-519-2-SP, LRP, NUDOCS 8701140325
Download: ML20207N629 (143)

Text

2. /f 7-00LMETED.

USNRC UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION

'87 JAN 12 P12:I3 BEFORE THE PRESIDING BOARD

~

CFF ~

L O Lv . -

)

In the Matter of )

) o -32 INQUIRY INTO THREE MILE ) Docket No. LRP ISLAND UNIT 2 LEAK RATE )

DATA FALSIFICATION ) ASLBP No. 86-519-02 SP

)

GPU NUCLEAR CORPORATION'S PROPOSED FINDINGS

, OF FACT AND CONCLUSIONS OF LAW (SUBMITTED TO THE PRESIDING BOARD IN THE FORM OF A RECOMMENDED DECISION)

Ernest L. Blake, Jr.

John N. Nassikas III SHAW,.PITTMAN, POTTS & TROWLRIDGE 2300 N Street, N.W.

Washington, D.C. 20037 (202) 663-8000 l Counsel for l GPU Nuclear Corporation l

Jcnuary 9, 1987 I

8701140325 870109 ges Acoes05000gj0 q 9

m .

l l

l l

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE PRESIDING BOARD

)

In the Matter of )

)

INQUIRY INTO THREE MILE ) Docket No. LRP ISLAND UNIT 2 LEAK RATE )

DATA FALSIFICATION ) ASLBP No. 86-519-02 SP

)

GPU NUCLEAR CORPORATION'S PROPOSED FINDINGS OF FACT AND CONCLUSIONS OF LAW (SUBMITTED TO THE PRESIDING BOARD IN THE FORM OF A RECOMMENDED DECISION)

Ernest L. Blake, Jr.

John N. Nassikas III SHAW, PITTMAN, POTTS & TROWBRIDGE

. 2300 N Street, N.W.

Washington, D.C. 20037

-(202) 663-8000 Counsel for GPU Nuclear Corporation Jenuary 9, 1987 l

l l

(

1 l-

TABLE OF CONTENTS I. INTRODUCTION AND PROCEDURAL BACKGROUND ................... 1 II. FINDINGS OF FACT .,....................................... 16 A. What was the organizational structure of TMI-2 during 1978 and 1979? ........................ 16 (i) Management .................................... 17 (ii) The Operations Department ..................... 25 (iii) Bodies Outside of the Chain of Command ........ 34 B. What Technical Specifications and procedures were relevant to leak rate testing? ................ 35 C. How were the Technical Specification 3.4.6.2 requirements for reactor coolant system unidentified leakage interpreted and implemented by control room operators (CROs), shift foremen, shift supervisors and on-site and off-site

management? Following the discovery by an NRC 4 inspector in October 1978 that Technical Speci-fication 3.4.6.2 requirements were not properly interpreted or implemented, what corrective action was taken by management personnel? Was the corrective action taken sufficient to insure l compliance with the Technical Specification 3.4.6.2 by the personnel performing and reviewing the leak rate surveillance tests? (CLI-85-18, ,

j 22 N.R.C. at 880, Issue (a)) ....................... 45 l

! Training ...................................... 46 (i) l (ii) General Practice Prior to October 1978 ........ 50 (iii) Events Related to the Issuance of the LER ..... 52 D. What difficulties, if any, were operators experiencing when conducting leak rate surveillance tests required by Technical Specification 4.4.6.2.d? Who knew about these difficulties? What corrective actions were taken? Did operators feel pressure to obtain leak rate surveillance test results which did not exceed technical specification limits? If so, what type of pressure was perceived or exerted and who wss responsible?

. (CLI-85-18, 22 N.R.C. at 880, 1ssue (b)) ........... 80

-ii-l l.

1 .- -. - - -

(i) Method Errors ................................. 82 (ii) Instrument Errors ............................. 84 (iii) Oscillations in Plant Conditions .............. 86 (iv) Inherent Conservatism of the TMI-2 Leak Rate Test ................................ 86 (v) Persons Who Knew of the Difficulties .......... 88 (vi) Corrective Actions ............................ 96 (vii) Pressure to Obtain Leak Rate Results Which Did Not Exceed Technical Specification Limits ................ 99 E. Were unacceptable leak rate surveillance test results required by Technical Specification 4.4.6.2.d discarded? If so, who knew of, condoned or directed this practice? Were unacceptable leak rate surveillance test results discarded in an attempt to hide them from the NRC7 (CLI-85-18, 22 N.R.C. at 880, Issue (c)) ........................................ 101 F. Did operators manipulate data or take other actions during leak rate surveillance testing in an attempt to improperly influence test results? Who performed, condoned, directed or was knowledgeable of data manipulation or other improper actions during leak rate surveillance testing? (CLI-85-18, 22 N.R.C. at 880, Issue (d)) ...................... 108 III. CONCLUSIONS ............................................ 123 APPENDIX A: List of Exhibits l APPENDIX B: List of Witnesses and Testimony and Documentary Material Bound into the Transcript APPENDIX C: Correlation of NRR and Stier Test Numbers for Each Leak Rate Surveillance Test Performed l

l Between 9/30/78 and 3/28/79 l

I

-lii-I

f UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE PRESIDING BOARD

)

In the Matter of )

)

INQUIRY INTO THREE MILE ) Docket No. LRP ISLAND UNIT 2 LEAK RATE )

DATA FALSIFICATION ) ASLBP No. 86-519-02 SP

)

GPU NUCLEAR CORPORATION'S PROPOSED FINDINGS OF FACT AND CONCLUSIONS OF LAW (SUBMITTED TO THE PRESIDING BOARD IN THE FORM OF A RECOMMENDED DECISION)

I. INTRODUCTION AND PROCEDURAL BACKGROUND

1. On March 24, 1980, Harold W. Hartman, Jr., a control room operator at Three Mile Island -- Unit Number 2 ("TMI-2")

prior to the accident, publicly alleged that reactor coolant sys-tsm ("RCS") leak rate surveillance tests (" leak rate. tests")l/

ware at times purposely manipulated and records of unacceptable I rssults were discarded to cover up the fact that over an extended pariod of time the results of the tests exceeded Technical Speci-fication (" Tech Spec") limits for unidentified leakage. Hartman alleged that the computer program for calculating leak rates was 1/ The leak rate tests were used to assess whether primary sys-

tcm leakage surpassed limits contained in the facility's techni-col specifications. The leak rate test is commonly known by sev-gral names or acronyms, such as

" Leak Rate," "LRT," " Reactor Coolant Inventory Balance," "RCIB," or " Mass Balance." General-ly, the term " leak rate test" will be utilized in this decision except when dictated otherwise in quoting or paraphrasing testi-mony or documentary evidenco.

. I l

unreliable, frequently yielding unrealistic results. This made it more difficult to get " good" leak rates. Hartman further al-

'lcged that operators at TMI-2 sometimes manipulated leak rate tost results by inputting wrong data into the computer, adding i hydrogen gas to the make-up tank during leak rate tests, adding w2ter to the make-up tank during.a leak rate test and not in-putting the addition into a computer, and adding water to the make-up tank while performing water transfer operations involving other tanks. Hartman specifically alleged that shift supervision was aware of such improper conduct. Inquiry Into Thres Mile Is-lend Unit 2 Leak Rate Data Falsification, CLI-85-18, 22 N.R.C.

877 (1985).

2. Shortly after Hartman made these allegations public, the Nuclear Regulatory Commission ("NRC") Office of Inspection and Enforcement ("I&E") began an. investigation. In the early stages of this NRC investigation, the United States Department of

-Justice ("DOJ") was advised of evidence uncovered by the NRC, and on April 28, 1980, DOJ assumed control of the investigation. I&E rcmained involved only to the extent of providing assistance to DOJ. Board Exh. 1-A, Stier Report, Vol. I at 2.2/

3. On April 16, 1980, Metropolitan Edison Company

(" Met-Ed"), then the operator of TMI-2, retained the law firm of l

2/ -Appendix A provides a list of exhibits offered or received l in this proceeding. See Tr. 5221; Board Order of Nov. 19, 1985.

1 l

L ._

Faegre & Benson to conduct an investigation of Hartman's allega-tions. Id. Because of the criminal nature of the investigation being conducted by DOJ, Faegre & Benson could not gain access to key witnesses. Id.; Board Exh. 2, Faegre & Benson Report, Vol. 1 ct 13. However, they did interview Hartman for two days in April 1980 and performed extensive technical analyses of leak rate testing at TMI-2. Board Exh. 2, Faegre & Benson Report, Vol. 1 at 2; Board Exh. 1-A, Stier Report, Vol. I at 2. Their report was issued on September 17, 1980. Board Exh. 2, Faegre & Benson Report.

4. While the criminal investigation of leak rate testing was pending, the NRC received a status report from its Region I personnel who had conducted the original investigation. That report-was presented on June 3, 1983, and summarized the findings of I&E up to the point where the investigation was turned over to DOJ. The NRC subsequently instructed its Office of Investiga-J tions ("OI") to investigate TMI-2 leak rate test practices. On June 27, 1983, OI began an investigation that was also limited because of the pending DOJ investigation. Like Faegre & Benson, OI was not able to interview critical witnesses. Board Exh. 1-A, 1

Stier Report, Vol. I at 2-3.

4

5. On November-11, 1983, the DOJ investigation resulted in cn eleven-count indictment returned in the U.S. District Court, Middle District of Pennsylvania (Criminal No. 83-00188), charging Met-Ed with criminal offenses arising out of leak rate practices n . -n, - - - - - - . , . . . - . . - , - - - , , , _ . . , _ . - - - . -

L ct TMI-2. On February 28 and 29, 1984, Met-Ed entered into a

-plea agreement with the Government ending the criminal prosecu-tion. Met-Ed pleaded guilty to one count of the indictment and nolo contendere to six other counts of the indictment. Id. at 3; nre also id., Vol. V(A) at Tab 3 (Statement of Metropolitan Edison Company with respect to the Plea Agreement).

6. When the Government and Met-Ed entered the plea agree-mint with the Court, both the United States Attorney and Met-Ed entered statements into the record. Id., Vol. I at 3. In urging the Court to accept the plea agreement, U.S. Attorney David Queen, inter alia, stated that the evidence developed in the Grand Jury inquiry did not indicate that any of the directors and i officers of GPU Nuclear Corporation ("GPUN") from its inception in 1982 as successor operator of TMI-2 to Met-Ed to the date of the indictment, or any of the directors of Met-Ed " participated in, directed, condoned, or was aware of the acts or admissions that are the subject of the indictment." CLI-85-18, 22 N.R.C. e.t 879.

l

7. After the Court accepted the plea agreement, Met-Ed and the NRC moved to obtain the release of the evidence presented to the Federal Grand Jury. On April 10, 1984 and June 25, 1984, the

Court denied those motions, and the evidence on which the Grand Jury relied has remained sealed. Board Exh. 1-A, Stier Report, Vol. I at 3.

._ . . = _ - . ~ . . . .

L l

8. In January of 1984, GPUN retained Edwin Stier, a former Director, New Jersey Division of Criminal Justice, to conduct an '

I independent investigation into leak rate testing at TMI. Id. at I j

4. Stier's report examined the attitudes and behavior of TMI-2 personnel toward leak rate testing during the full year of TMI-2

! operation, the 222 leak rate tests for which records presently l

l cxist, and statements made by individuals possibly involved in j I

leak rate testing. See id. at 11-16. The Stier Report, entitled "TMI-2 Reactor Coolant Inventory Balance Testing," was issued on September 5, 1985. Board Exh. 1-A, Stier Report.

9. The Commission asked OI to examine whether Michael l Ross, Manager of Operations at TMI-1, had participated in, 1

directed, or condoned leak rate falsifications at TMI-2. OI in-terviewed Ross and others under oath regarding Ross's involvement at Unit 2, reviewed pertinent records, and concluded that Ross's role at TMI-2 was minimal. In its report of April 16, 1984, OI found that during the period falsifications took place, Ross was present at TMI-2 only the minimum time necessary to maintain his TMI-2 license and was not involved'in the falsifications. See CLI-85-18, 22 N.R.C. at 879.

10. In July 1984, the NRC Staff issued NUREG-0680, Supp.

No. 5, dealing with the restart of TMI-1. Among the subjects discussed in that report was management involvement in leak rate testing at TMI-2. The NRC Staff relied on two sources of in-  !

formation: (1) the statement issued by the U.S. Attorney and l

l (2) all of the evidence that had been gathered by the NRC up to that time, including evidence developed by OI in its then pending investigation. Board Exh. 1-A, Stier Report, Vol. I at 4.

11. OI issued a report on August 15, 1984 summarizing its findings as of that date. In his cover memorandum, OI Director Ben Hayes described his report as follows: "(I]t does not set forth the facts and evidence obtained as a result of a completed investigation but sets forth the information accumulated by the NRC since May 1979." Id. (citing Memorandum, Ben B. Hayes to NRC Commissioners, Three Mile Island Nuclear Generating Station Unit 2/ Alleged Falsification of Leak Rate Surveillance Test Data (1-83-010), August 15, 1984 at 1).

12. Both the August, 1984 OI Report and the July, 1984 NUREG-0680, Supp. No. 5, indicated that the NRC Office of Nuclear Rsactor Regulation ("NRR") and OI would jointly continue an in-vastigation of leak rate test practices by some individuals who htd been licensed at TMI-2 or had held dual licenses for TMI-1 1

End TMI-2. NRR worked with OI on investigations of 10 licensed operators 3/ to determine their involvement, if any, in improper 3/ The ten operators were Raymond R. Booher, Joseph R. Congdon, Martin V. Cooper, Craig C. Fatst, Edward R. Frederick, Carl L.

Guthrie, Theodore F. Illjes, Hugh A. McGovern, Adam W. Miller, cnd Dennis I. Olson. NRC chose to investigate these individuals bscause they continued, at the time, to be licensed NRC opera-l tors. Although Booher and Olson were no longer licensed by the time NRC completed the investigation and issued the NRR/OI Rsport, the Report included findings, conclusions, and recommen-dations concerning all ten. See Board Exh. 5-A, NRR Report.

cctivities associated with leak rate testing at TMI-2 before the cccident. NRR performed a technical evaluation of 161 leak rate tasts performed at TMI-2 during the period September 30, 1978 to March 28, 1979.4/ In addition, joint NRR/OI interviews were con-i ducted with 13 former control room operators and two shift foremen. NRR prepared a report on each of the 10 individuals under investigatic.n, identifying the individual's role in leak rate testing irregularities and evaluating the individual's cur-rant performance.5/ Board Exh. 5-A, NRR Report, Vol. 1 at Enclosure 1. On April 1, 1986, the joint NRR/OI Report -- "Re-cults of NRR's Investigation and Evaluation of Ten Licensed Oper-ators Involved in TMI-2 Preaccident Leak Rate Testing Irregu-1Erities" -- was issued. Board Ext. 5-A, NRR Report.

13. In an Order issued in the TMI-1 Restart Proceeding on February 25, 1985, the Commission stated that it would institute 4/ Stier reviewed all 222 leak rate tests at TMI-2 for which rscords presently exist. Stier used different numbering of tests from NRR. Generally, the Board will refer to the number assigned to the test by NRR (e.g., NRR Test No. 1). Stier test numbers will be used for those tests which NRR did not review or when Stier's evaluation of the test is being discussed (e.g., Stier Test No. 1). NRR tests are found in Board Exh. 5-A, NRR Report, Vols. 2-4. Stier tests are found in Board Exh. 1-A, Stier Rcport, Vols. IV(C)-(K). Appendix C provides a list correlating the NRR and Stier test numbers.

5/ The Board excluded as evidence in this proceeding portions of the NRR/OI Report regarding current performance. The individ-l umis' current performance and recommendations for actions that

may be taken with respect to anyone involved in leak rate falsi-fications were beyond the scope of our fact-finding process. See CLI-85-18, 22 N.R.C. at 883-884.

a separate hearing apart from the Restart Proceeding to develop the facts surrounding the RCS leak rate data falsifications at TMI-2 prior to the March 28, 1979 accident in sufficient detail to determine the ultimate status of those possibly involved. The .

Commission's Order specifically excluded those individuals whom -

the U.S. Attorney at the sentencing hearing of Met-Ed had stated were not involved and those individuals whom OI already had re-viewed and found not to be implicated in its TMI-1 leak rate in-vestigation (i.e., Michael Ross). Metropolitan Edison Company, et al. (Three Mile Island Nuclear Station, Unit 1), CLI-85-2, 21 N.R.C. 282, 298-299 (1985).

14. In December 1985 the Commission issued a Notice and I"

- L Hearing for this proceeding. CLI-85-18, 22 N.R.C. 877. It directed this Board to address the following issues:

(a) How were the Technical Specification 3.4.6.2 requirements for reactor coolant system ,

unidentified leakage interpreted and implemented ,

by control room operators (CROs), shift foremen, shift supervisors and on-site and off-site manage-ment? Following the discovery by an NRC inspector in October 1978 that Technical Specification 3.4.6.2 requirements were not properly interpreted or implemented, what corrective action was taken by management personnel? Was the corrective ac-tion taken sufficient to insure compliance with the Technical Specification 3.4.6.2 by the person-nel performing and reviewing the leak rate sur-veillance tests?

(b) What difficulties, if any, were opera-tors experiencing when conducting leak rate sur-veillance tests required by Technical Specifica-tion 4.4.6.2.d? Who knew about these difficulties? What corrective actions were taken? ,

Did opert. tors feel pressure to obtain leak rate surveillance test results which did not exceed .

technical specification limits? If so, what type of pressure was perceived or exerted and who was responsible?

(c) Were unacceptable leak rate surveillance test results required by Technical Specification 4.4.6.2.d discarded? If so, who knew of, condoned or directed this practice? Were unacceptable leak rate surveillance test results discarded in an at-tempt to hide them from the NRC?

(d) Did operators manipulate data or take other actions during leak rate surveillance testing in an attempt to improperly influence test results? Who performed, condoned, directed or was knowledgeable of data manipulation or other im-proper actions during leak rate surveillance testing? This would include, but is not limited to the following:

(i) inputting the wrong data into the plant computer; (ii) adding hydrogen gas to the make-up tank during the test in an attempt to influ-ence make-up tank level indication; (iii) adding water to the make-up tank during the test and either not including the addition in the computer calculation or underrecording the addition in the computer; (iv) taking advantage of differences or inaccuracies in plant instrumentation (e.g.,

make-up tank level indicators) in an attempt to influence parametera critical to the leak rate surveillance test calculation; (v) taking or failing to take any ac-tion in violation of technical specification requirements?

CLI-85-18, 22 N.R.C. at 880-881.

15. The Commission's Order and Notice of Hearing of December 18, 1985 established the procedures for the hearing.

Pursuant to the Order and Notice of Hearing, the hearing was not

_g_

conducted under 10 C.F.R. Part 2, Subpart G, except as noted in the Commission's Order. The Presiding Board (" Board") did have the powers specified in 10 C.F.R. Il 2.718(a),-(e), (f), (h),

(i), (j), and (k). The hearing was conducted using a legislative hsaring format. Id. at 882.

16. The Commission directed the Chief Administrative Judge, Atomic Safety and Licensing Board Panel, to appoint a three-parson Presiding Board to rule on petitions to intervene, to con-duct any prehearing procedures and the hearing, and to render a rscommended decision setting forth the facts surrounding the fal-cifications and identifying those individuals who participated in, or knew of and condoned, or by their dereliction or culpable nsglect allowed, the leak rate falsifications at TMI-2. Id. at 881.

17. On December 20, 1985, Administrative Judges James L.

Kelley, Glenn O. Bright, and Jerry R. Kline were appointed to serve as the Board. Judge Kelley was appointed the Chairman. On August 27, 1986, Administrative Judge James H. Carpenter replaced Judge Kline, who was unable to continue because of a schedule conflict. Appointment of Presiding Board to Conduct a Legisla-tive Hearing, 50 Fed. Reg. 53,489 (1985).

18. Any person who had an interest that the hearing may have affected was allowed to petition to intervene. If the Board determined that the petitioner had an interest that could be affected and the petitioner was likely to contribute to the de-velopment of an adequate record, the petition was to be granted.

CLI-85-18, 22 N.R.C. at 881-882.

19. On December 31, 1985, as a supplemental notice to ,

potentially interested individuals, the Board sent a letter by certified mail to about 120 present and former employees of Met-Ed who were associated with TMI-2 between February 2, 1978 and March 28, 1979. The group represented those employees who might have been involved in or had knowledge of the RCS leak rate data that was the subject of this inquiry.6/ Memorandum and Order, February 14, 1986 at 1. We enclosed a copy of the Commis-1.

sion's Order and Notice of Hearing of December 18, 1985 and in-vited those interested to file a petition to intervene by January 30, 1986.

20. Following the Commission's Order and Notice of Hearing and this Board's supplemental notice, the Board received peti-tions to intervene from the following: (1) twenty-five present and former employees of Met-Zd (" Numerous Employees"),2/

6/ One of the objectives of this proceeding was to exculpate individuals whose names have arisen in connection with falsified l leak rate testing at TMI-2. Memorandum and Order, July 16, 1986 s at 15. After conducting a comprehensive evidentiary hearing, the s Board has determined that the record warrants the exoneration of any addressee of our letter whom we do not discuss in this opin-ion. Of course, as to those whom we do discuss, our decision governs.

2/ The following employees sought intervention: Charles D.

Adams, Raymond R. Booher, John A. Brummer, Kenneth P. Bryan, (Continued Next Page)

(2) John M. Kidwell, a former employee of Met-Ed, (3) John G.

Herbein, a former officer and employee of Met-Ed, (4) Gary P.

Miller, an employee of Met-Ed, (5) GPUN, (6) Marvin I. Lewis, and (7) Marjorie M. and Norman O. Aamodt. Each of the first five pe-titions alleged facts demonstrating an interest of the petitioner that this proceeding could have affected, and a likely ability to contribute to the record -- the standards for intervention estab-lished by the Commission. For those reasons, the Board granted the first five petitions listed above.g/ Id. at 4.

21. The Numerous Employees filed oppositions to the peti-tions to intervene from the Aamodts and from Lewis. We initially had questioned whether the Aamodt and Lewis petitions met the standards for intervention in this proceeding and had called for further information in our Memorandum and Order of February 14, ,

1986. We subcequently received written responses to our request from the Aamodts and Lewis. In addition, Mrs. Aamodt attended (Continued)

Joseph J. Chwastyk, Mark S. Coleman, William T. Conaway, Joseph R. Congdon, Craig C. Faust, James R. Floyd, Edward R.

Frederick, Leonard P. Germer, Carl L. Guthrie, Gregory R.

, Hitz, Sr., Kenneth R. Hoyt, Theodore F. Illjes, George A. Kunder, Walter J. Marshall, Hugh A. McGovern, Brian A. Mehler, Charles F.

Mall, Adam W. Miller, Frederick J. Scheimann, Bernard G. Smith, and William H. Zewe.

g/ On May 7, 1986, Bryan, one of the original Numerous Employees, and Kidwell withdrew as parties to this proceeding.

Memorandum and Order, May 22, 1986 at 13. Both men, however, submitted prefiled testimony and testified in this proceeding.

See Tr. 4539-4610 (Bryan); Tr. 3285-3399 (Kidwell).

h.____ . . _.._ __

the prehearing conference on March 7, 1986 and argued her enti-tlement to intervention. We subsequently allowed the Aamodt in-tervention but denied Lewis' request. Memorandum and Order, March 26, 1986 at 1.

22. The NRC Staff did not participate in this proceeding as a party. See CLI-85-18, 22 N.R.C. at 882. The Staff made avail-able to the parties and to the Board relevant documentary materi-al within its possession. In addition, the Staff provided testi-mony and assistance to the Board to help ensure that the hearing record was fully developed.

23. Under the Commission's Order and Notice of Hearing, no discovery was conducted. The Commission intended the hearing to serve as the fact-finding mechanism. Id.

24. Only the Board was allowed to call witnesses or to question them. The Board also had the power to issue subpoenas if necessary to compel the attendance of witnesses. Id. Prior to the commencement of the hearing, we made available to the par-ties a list of the individuals that we intended to call as wit-nesses. Memorandum and Order, March 26, 1986 at Attachment A.

We invited the parties to submit recommendations regarding addi-tional witnesses. See, e.g., Memorandum and Order, March 26, 1986 at 10; Memorandum and Order, May 22, 1986 at 3; Tr. 3604-05 (Kelley, J.).

25. Because the Board had exclusive authority to call wit-nesses, we considered all witnesses to be " Board witnesses." We even extended the designation of " Board witnesses" to experts (Rockwell and Stier) who had prepared reports as paid consultants to Met-Ed and GPUN and in the conventional licensing case would have been expected to appear as witnesses for GPUN. Memorandum and Order, April 3, 1986 at 3. Although these experts were paid by GPUN, they appeared as Board witnesses. Tr. 216-18 (Kelley, J.). In addition, the Board proposed and subsequently adopted a "no access" rule to promote on-the-record discussion of the issues and equal party access to the facts and to minimize burdens on technical witnesses. See Memorandum and Order, April 13, 1986 at 3; Memorandum and Order, May 22, 1986 at 13.

The rule prohibited counsel for any of the parties from commu-nicating with these Board witnesses prior to their appearance at the hearing. See Memorandum and Order, May 22, 1986. We subse-quently modified this rule to allow the parties to contact these Board witnesses to discuss their conclusions and opinions con-cerning individual culpability for leak rate falsification at Unit 2. Memorandum and Order, August 7, 1986 at 3.

26. Before each witness testified, we invited the parties 4

to submit questions in writing to the Board that they believed we should pose to the witness. The Board had the discretion to use the questions suggested by the parties. CLI-85-18, 22 N.R.C. at 882. All witnesses testified under oath.

27. The hearing commenced on September 8, 1986 in Bethesda,

, Maryland. It consumed 33 hearing days and resulted in over 5,000 transcript pages. Forty-seven witnesses appeared and testified, most of whom filed prefiled testimony.9/ Twenty-five exhibits were entered into the record. See n.2, supra.

28. The following findings of fact address the specific i' sues the Commission directed this Board to consider. See V 14, supra. The findings are divided into six major parts. The first -- - -

part of the findings covers the organization of TMI-2. The sec-ond part gives an overview of the TMI-2 Tech Specs and procedures concerning leak rate testing. These first two parts provide a j background and overview necessary to an understanding of these findings and conclusions. The third part discusses the conduct of leak rate testing at TMI-2 in 1978 and 1979, including training on the subject, and events and actions relating to an NRC inspection of TMI-2 in October 1978. The fourth part addresses the difficulties operators were experiencing when 9/ Appendix B provides a list of witnesses and transcript cita-tions to their testimony, as well as a list of documentary mate-rial bound into the transcripts.

Because Wright did not adopt his prefiled testimony, we had his prefiled testimony bound into the record, not as substantive evidence, but as a brief accurate summary of a longer statement placed in the record (Board Exh. 6, OI Report, Exh. 18, Wright Interview) and regarded as substantive evidence. Tr. 2662-63 (Kelley, J.).

John J. Blessing was subpoenaed by the Board. However, he did not comply with the subpoena and did not testify. See Tr. 4332, 4542 (Kelley, J.).

e

conducting leak rate tests, who knew of those difficulties, the actions taken to correct those difficulties, and whether opera-tors felt pressure to obtain leak rate results that did not exceed Tech Spec limits. The fifth part covers the documentation _

and retention of leak rate tests at TMI-2. The sixth part addresses whether operators manipulated data or took other ac-tions during leak rate testing to influence test results improp-erly and who performed, condoned, directed, knew of, or by dere- l liction or culpable neglect, allowed manipulation of leak rate tests or other such improper action.10/ The final portion of the opinion presents our conclusions regarding leak rate testing at TMI-2.

II. FINDINGS OF FACT A. What was the organizational structure of '

TMI-2 during 1978 and 1979?

29. The Operations Department at TMI-2 was resronsible for the administration of the leak rate test. The Supervisor of Operations headed the Department. The chain of command below this position consiste' of the shift supervisors, shift foremen, control room operators ("CROs"), and auxiliary operators. The Supervisor of Operations, together with the Unit 2 Superintendent of Technical Support, reported to the Unit 2 Superintendent. The '

10/ GPUN's proposed findings do not address the individual re-sponsibility of those below the management level of Superinten-dent of Technical Support and Supervisor of Operations.

Unit 2 Superintendent reported to the Station Superintendent, who was the highest member of management at the facility. The Sta-tion Superintendent reported to the Manager of Generation, who in turn reported to the Vice President of Generation. After the

' Station Superintendent became Station Manager in March 1979,.the

-position reported directly to the vice presidential level. The offices of both the Vice President and the Manager of Generation ware located in Reading, Pennsylvania, about 50 miles from the fccility. Board Exh. 1-A, Stier Report, Vol. I at 47, 131-32; id., Vol. VI(F), Herbein 2/8/85 Interview at 9; Herbein, ff.

i

.Tr. 5268 at'6-7.

30. In addition to the chain of command, several bodies ex-isted to assure independent review of plant activities and to

< provide management with a source of information'concerning opera-tion of the plant. These groups included the Plant Operations Rsview Committee ("PORC"), the Generation Review Committee

("GRC"), the Quality Assurance Department ("QA"), and the General Office Review Board ("GORB"). Board Exh. 1-A, Stier Report, Vol.-I at 132.

(i) Management

31. John G. Herbein was the Vice President of Generation for Met-Ed, stationed in Reading, Pennsylvania. He was responsi-ble for the overall operation, maintenance, administration, qual-ity assurance, and related technical engineering support

activities at the nuclear, fossil, and hydro generating stations owned and operated by Met-Ed. Herbein, ff. Tr. 5268 at 3-4.

32. The Manager of Generation Operations (Lawyer) reported to Herbein. In March 1979, the TMI Station Manager (G. Miller) began reporting directly to Herbein rather than to the Manager of Generation Operations. Id. at 7.

33. Herbein relied on the chain of command and the formal review committees to bring issues requiring his input to his at-tention. On a day-to-day basis, he received information primari-ly from the managers who reported to him. About every 4-6 weeks, he visited TMI to meet with station management. During those visits, he periodically would speak with employees to indicate that upper management was interested in and supportive of them and that the Reading corporate organization was concerned about activities at TMI. Id. at 5-7.

34. Lawrence L. Lawyer was the Manager of Generation Opera-tions, stationed in Reading, Pennsylvania. He was responsible for the maintenance and operation of Met-Ed's nuclear, fossil, and hydro generation stations. Lawyer reported to the Vice Pres-ident of Generation (Herbein). The Station Superintendents of the generating stations, including the TMI Station Superintendent (G. Miller) until March 1979, reported to Lawyer. In March 1979, when the TMI Station Superintendent's title changed to Station Manager, the Station Manager began reporting directly to the Vice

~ President of Generation rather than to Lawyer. Board Exh. 1-A, Stier Report, Vol. VI(G), Lawyer 11/10/83 Interview at 5-7, 9; G. Miller, ff. Tr. 5039 at 4-5.

35. Generally, Lawyer received daily station status reports during morning conference calls to each station superintendent.

The TMI Unit Superintendents participated in the call with the TMI Station Superintendent. While Lawyer did receive daily copies of the Daily Plant Status Reports, he primarily depended on the daily morning conference calls and other phone calls from the plant to identify problems that needed his attention.11/ In cddition to monthly meetings with his station superintendents, which occasionally took place at TMI, Lawyer usually traveled to TMI once a week to meet with the TMI Station Superintendent.

Board Exh. 1-A, Stier Report, Vol. VI(G), Lawyer 11/10/83 Inter-view at 7-10, 19-22, 62-63.

36. Gary P. Miller held the title of Station Superintendent at TMI until March 1979 when the title of the position changed to Station Manager. This position, the highest level of management stationed at TMI, was responsible for the supervision of TMI-l and TMI-2 and was responsible for compliance with the operating licenses, Tech Specs, and all applicable regulations. While 11/ The Daily Plant Status Report listed plant parameters, including unidentified leak rate, and provided areas for comments end special problems. G. Miller, ff. Tr. 5039 at 13. See, e.g.,

Board Exh. 1-A, Stier Report, Vol. V(B) at Tab 10.

l

Station Superintendent, Miller reported to the Manager of Genera-tion Operations (Lawyer). When the position title changed to Station Manager, Miller began reporting directly to the Vice President of Generation (Herbein). Persons in three principal operating positions reported to the Station Superinten-dent / Station Manager. They were the Unit 1 Superintendent, the Unit 2 Superintendent, and the Site Maintenance Superintendent, who was responsible for maintenance at both units. G. Miller, ff. Tr. 5039 at 2-5; Herbein, ff. Tr. 5268 at 7; Board Exh. 2, Faegre & Benson Report, Vol. 3B, Exhibit 70 at 2.

37. Insofar as operation of the units was concerned, Miller relied principally on the Unit 1 and 2 Superintendents. Typi-cally, he spent most of the time from 9 a.m. to 5 p.m., Monday through Friday, in meetings. He did not regularly attend meet-ings involving plant operations. Miller estimated that he spent from 20% to 40% of his time away from TMI. While his goal was to visit the Control Rooms about once a week, Miller in practice was not able to visit them that frequently. G. Miller, ff. Tr. 5039 at 6, 12, 14; Tr. 5041, 5066-67 (G. Miller).

38. Miller kept abreast of the daily status of the units primarily through twa mechanisms, namely the Daily Plant Status Report and a daily morning conference call typically involving Miller, the Unit 1 and 2 Superintendents, and his superior in Reading. On those days when he was at TMI, Miller tried to re-view the Daily Plant Status Report each morning. While he did l

not recall-seeing shift supervisor' turnover notes attached to the-rcports, Miller believed, based on documents shown to him during the Stier Unit 2 leak rate investigation, that he did receive them. Miller recalled that he relied on the morning call as his-principal daily source of information on the status of the units.

G. Miller, ff. Tr. 5039 at 12-14.

39. On matters of plant operations Miller expected his sub-ordinates to use normal channels. His office was outside the so-curity fence at TMI precisely so that employees would deal with their unit managements rather than coming directly to him on rou-tine operation matters. Id. at 15.

40. Because of his nut.erous responsibilities, Miller neces-sarily depended on his subordinates to bring to his attention operating matters that involved safety or unit availability. He was not in a position independently to seek out such problems.

Id. at 11; Tr. 5066 (G. Miller).

41. Until late 1978, Miller concurrently held the title of Unit 2 Superintendent at TMI along with the Station Superinten-dont title. Joseph B. Logan, who had been hired by the company in January 1378 for the position of Unit 2 Superintendent, for-mally assumed the position toward the end of 1978 after obtaining en NRC Senior Reactor Operator License and familiarizing himself with the unit and staff. G. Miller, ff. Tr. 5039 at 3; Board Exh. 1-A, Stier Report, Vol. VI(G), Logan 3/27/85 Interview at 3.

I

1

42. Miller believed that while he himself held both the Station Superintendent and Unit 2 Superintendent positions in 1978, James L. Seelinger, the Unit 2 Superintendent of Technical Support, substantially discharged the responsibilities of Unit 2 Superintendent. Miller testified that because of his station responsibilities he depended on Seelinger to supervise day-to-day operations. G. Miller, ff. Tr. 5039 at 3-4; Tr. 5050-53 (G. Miller). Seelinger took issue with Miller's view of him as acting Unit 2 Superintendent. Seelinger did allow that his and Miller's perception of his role in Unit 2 could have differed.

Seelinger clearly did not view his authority as broadly as Miller did. See Seelinger, ff. Tr. 4623 at 4-6; Tr. 4627-28 (Seelinger).

43. The Unit 2 Superintendent reported to the TMI Station Superintendent and was responsible for the administration, opera-tion, and maintenance of TMI-2 and for ensuring compliance with the Tech Specs. Administrative Procedure ("AP") 1010, " Technical Specification Surveillance Program," gave him specific responsi-bility for ensuring compliance of the Surveillance Test Schedules with the Tech Specs. Board Exh. 1-A, Stier Report, Vol. V(A) at Tab 6 (Unit Superintendent Position Description); id., Vol. V(C) at Tab 17 (AP 1010, Section 2.1.).

44. The four principal operating positions reporting to the Unit 2 Superintendent were the Unit Superintendent of Technical Suppport, the Supervisor of Operations, the Supervisor of

Maintenance, and the Supervisor of Radiation Protection and Chem-istry. G. Miller, ff. 5039 at 4; Board Exh. 1-A, Stier Report, Vol. VI(G), Logan 11/18/83 Statement at 1.

45. As Unit 2 Superintendent, Logan generally kept abreast of plant status and activities by participating in the morning conference calls between TMI and management in Reading, attending plan of the day (" POD") meetings,12/ and reviewing the Daily Plant Status Reports, shift supervisor turnover notes, control room logs, and documents generated by PORC. He also made tours of the plant and visited the control room several times each day.

Board Exh. 1-A, Stier Report, Vol VI(G), Logan 3/27/85 Interview at 20-22, 27-31, 33.

46. James L. Seelinger was the Unit 2 Superintendent of Technical Support at TMI until December 1978, when he became the Unit 1 Superintendent and George Kunder replaced him as Superin-tendent of Technical Support. Seelinger, ff. Tr. 4623 at 1; Kunder, ff. Tr. 4800 at 1.

12/ The POD meetings Logan attended were held daily in each unit early on the 7 a.m. - 3 p.m. shift. The primary purpose of those meetings was to discuss plant problems and plant evolutions scheduled to be performed that day. The typical attendees in-cluded the Unit Superintendent, the Supervisor of Operations, the Superintendent of Technical Support, one or more shift supervi-sors, and a representative from the Maintenance Department.

Tr. 4678-79 (Seelinger); Bryan, ff. Tr. 4540 at 2; Mehler, ff.

Tr. 3842 at 3; Board Exh. 1-A, Stier Report, Vol. VI(G), Logan 11/18/83 Statement at 1-2.

47. The Unit 2 Superintendent of Technical Support reported to the Unit 2 Superintendent and was responsible for supervision .

of the technical aspects of plant engineering and for ensuring TMI-2 operation complied with the Tech Specs. The lead engineers .

assigned to TMI reported to him. In addition to other responsi-bilities, the Unit 2 Superintendent of Technical Support chaired the PORC. Kunder, ff. Tr. 4800 at 1; Board Exh. 1-A, Stier Report, Vol V(A) at Tab 6 (Unit Superintendent of Technical Sup-port Position Description).

48. Seelinger perceived his responsibilities as principally being charged with the writing and approval of the procedures for operating TMI-2. Board Exh. 1-A, Stier Report, Vol. VI(J),

Seelinger 4/4/84 Interview at 7. He typically attended the POD meetings. Tr. 4678 (Seelinger).

49. Kunder saw his relationship to the Operations Depart-ment as advisory, with no direct role in the operation of the unit and no responsibility for operational decisions. He usually learned about operational problems through members of his staff who performed the engineering tasks assigned to the department.

Kunder, ff. Tr. 4800 at 1-2. Typically, assignments to the de-partment did go through him; however, plant personnel could deal directly with Kunder's engineers in accomplishing tasks.

Tr. 4814-16 (Kunder). Kunder normally attended most of the POD meetings. Board Exh. 1-A, Stier Report, Vol VI(G), Kunder 3/6/85 Statement at 26.

(ii) The Operationc Department

50. The Operations Department at TMI-2 was headed by the Supervisor of Operations, James R. Floyd, who reported to the Unit 2 Superintendent. He was responsible for the overall admin-istration of his department and for ensuring compliance with the Tcch Specs. The shift supervisors and two operations engineers rcported to him. The Supervisor of Operations was required to hold an NRC Senior Reactor Operator License. Board Exh. 1-A, Stier Report, Vol. V(A) at Tab 6 (Supervisor of Operations Posi-tion Description); Floyd, ff. Tr. 4894 at 1-2; Marshall, ff.

Tr. 4380 at 1; G. Miller, ff. Tr. 5039 at 2-3. An Administrative Procedure 1012, " Shift Relief and Log Entries," required the Supervisor of Operations to review and sign the Control Room Log cnd Shift Foreman's Log at least once per week. Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 18 (AP 1012, Section 2.2).

51. Floyd delegated the actual running of TMI-2 to the chift supervisors and expected them to discuss any problems they hrd with him. He perceived his job as primarily that of a "cri-cis fighter," living out of the control room and the shift super-i visor's office with close contact with CROs. Floyd, ff. Tr. 4894 ct 2; Tr. 4969, 4974-75 (Floyd).

52. One means by which Floyd gave direction to the Opera-tions Department was through the periodic issuance of Operations MImoranda, which were compiled in a binder in the Control Room.

Tr.;4943, 4999-5001 (Floyd);-Tr. 2637 (Coleman); Board Exh. 1-A, Stier Report, Vol. VI(I), A. Miller 3/22/85 Interview at 84.

53. Floyd believed he attended the FOD and shift supervisor m etings, but rarely PORC meetings. Tr. 5031 (Floyd). Although f

ho.was on the distribution list for the Daily Plant Status Re-

, ports, he did not scrutinize them. He viewed them as the means l

l~ by which his superiors were advised of plant status on a daily t

! basis. _Floyd, ff. Tr. 4894 at 6-7.

i

54. The TMI-2 Operations Department work force was divided initially into five groups, referred to as " crews" or " shifts,"

that would rotate among_three daily shifts of 7 a.m. - 3 p.m.,

3 p.m. --11 p.m., and 11 p.m. - 7 a.m. Board Exh. 1-A, Stier l Rsport, Vol. I at 47. Prior to the establishment of the six-chift rotation, certain individuals were specified for relief.

After addition of the extra shift, however, it became easier to provide relief personnel since_a relief shift became a part of the normal shift rotation. Tr. 2329 (Russell). A shift supervi-sor, who held an NRC Senior Reactor Operator License, headed each shift. The license authorized him to direct the licensed activi-i ties of licensed operators and to manipulate the controls of the plant. The license itself stated that "[ijn directing the i

j licensed activities of licensed operators and in manipulating the j controls . . . the licensee shall observe the operating proce-dures and other conditions specified in the facility license

{

j . . . ." Because the shift supervisor was responsible for i

4 I

l cverseeing Operations Department activities at both Units 1 and 2, he divided his time between TMI-1 and TMI-2 during his shift.

The shift foreman for each unit reported to the shift supervisor.

Scard Exh. 1-A, Stier Report, Vol. I at 47, 49; id., Vol. V(A) at Tab 6 (Shift Supervisor Position Description); Board Exh. 10-A, SRO License.

55. The shift supervisor was responsible for ensuring that plant operations did not jeopardize the health and safety of the cmployees and public. The shift supervisor was also responsible for ensuring that the surveillance testing program was conducted in compliance with the Final Safety Analysis Report ("FSAR").

Beard Exh. 1-A, Stier Report, Vol. V(A) at Tab 6 (Shift Supervi-cor Position Description).

56. The turnover notes prepared by the shift supervisor on the 11 p.m. -

7 a.m. shift were attached to Daily Plant Status R: ports distributed at TMI. Id., Vol. I at 137; id., Vol. VI(H),

M:hler 2/28/85 Interview at 76-77; see, e.g., id., Vol. V(B) at Tab 10.

57. Monthly shift supervisor meetings were held at TMI.

The minutes of the meetings indicated that personnel matters were

~

the main topic of discussion. Smith, ff. Tr. 4331 at 2; Bryan, ff. Tr. 4540 at 2; Mehler, ff. Tr. 3842 at 3; Board Exh. 1-A, Stier Report, Vol. II(B), G. Miller Summary at 5.

~

58.- A shift foreman, who held an NRC Senior Reactor Opera-ter License, directly supervised the TMI Operations Department Staff on shift at each' unit. The license authorized him to i

-direct the licensed activities of licensed operators and to ma-t nipulate the controls of the plant. The' license itself stated that "[i]n directing the licensed activities of licensed opera-tors and in manipulating the controls . . . the licensee shall observe the operating procedures and other conditions specified in the facility license . . . ." The CROs reported to the shift i foreman. Board Exh. 10-A, SRO License; Board Exh. 1-A, Stier R port, Vol. V(A) at Tab 6 (Shift Foreman Position Description);

! id., Vol. I at 48.

59. Like the shift supervisor, the shift foreman was re-i i

aponsible for ensuring that plant operations did not jeopardize the health and safety of the employees and public and that the curveillance testing program was conducted in compliance with the FSAR. Board Exh. 1-A, Stier Report, Vol. V(A) at Tab 6 (Shift Foreman Position Description).

60. The shift foreman would assign work among his subordi-notes and would personally become involved in the resolution of any problems encountered on his shift. The manner in which shift foremen performed their responsibilities varied. Some foremen i

cpent more time in the Control Room than others who devoted more i

l of their time to inspecting the plant. Id., Vol. I at 48-49.

l r

E e

l i

i d 1

16 1 . The outgoing shift communicated information to the

^cncoming shift through shift turnovers. Floyd, ff. Tr. 4894 at 7 ~; Tr. 2623 (Coleman).

62. Although Surveillance Procedure ("SP") 2301-3D1, "RCS l

Inventory," did not define specific responsibilities for the chift foreman, AP 1010, " Technical Specification Surveillance

- Program," provided that "[d]ata sheets will be signed by the per-con performing the task, and reviewed and approved by his foreman where-required by the forms and procedures." As a general rule, the shift foreman would approve leak rate test results for filing by signing the leak rate test sheet. Board Exh. 1-A, Stier Report, Vol. I at 48; id., Vol. V(A) at Tab 6 (Shift Foreman Po-l cition Description); id., Vol. V(C) at Tab 17 (AP 1010, Section 3.2.2).

d

63. AP 1012, " Shift Relief and Log Entries," required the shift foreman to maintain a Shift Foreman's Log. The shift fore-min was responsible for review and sign off of the log at the completion of each shift. The procedure did not specify that the log contain any information relating to leak rate tests. Id.,

Vol. I at 44; id., Vol. V(C) at Tab 18 (AP 1012).

64. The CRos, under the direction of a shift foreman, were responsible for operating the plant. Each CRO was required to hold an NRC Reactor Operator License. The license issued to the i operator, which authorized him to manipulate all controls of the l

- , - , , , - ,____,__-,,,,_..~,.-m._,_. - - , - - _ _ _ _ . - - _ _ _ _ . . . . - - - - - - - - - - - - - - - - , - - - . - - . - - - -

1 l

l plant, stated that "[i]n manipulating the controls . . . the '

licensee shall observe the operating procedures and other condi-tions specified in the facility license . . . ." Board l

L Exh. 10-B, RO License; Board Exh. 1-A, Stier Report, Vol. V(A) at 1

Tab 6 (Control Room Operator Position Description).

i i

j 65. Two to four CROs ccmprised each shift, possibly includ-ing one or more CRO trainees whom the NRC had not yet licensed.

l The CROs were responsible for directing the work of several aux-iliary operators. Board Exh. 1-A, Stier' Report, Vol. I at 48; McGovern, ff. Tr. 3148 at 2; Conaway, ff. Tr. 3097 at 2.

Although the division of responsibilities among the CROs on shift did vary, one CRO was assigned to the control panel, where his rzsponsibilities included maintaining the Control Room Log. The rosponsibility for " switching and tagging" equipment to control its proper use and for taking readings from plant instruments was mesumed by the second CRO on shift (or divided between two or more CROs on crews that included a total of more than two CROs).

The CRO responsible for taking readings usually would also be re-sponsible for performing the leak rate test. In practice, the CRO duties overlapped somewhat. The switching and tagging CRO might make entries in the Control Room Log, and the CRO assigned to the Control Panel might, on occasion, perform a leak rate tcst.13/ Board Exh. 1-A, Stier Report, Vol. I at 48-49.

i 13/ For each of the leak rate tests, Stier and NRR attempted to d3termine individual assignments for each of the individuals on abift. See Board Exh. 1-A, Stier Report, Vol. III(A), Tables 1-2; Board Exh. 5-A, NRR Report, Vol. 2, Table 11.

66. Like the shift foremen, the CROs used turnovers to com-municate information from shift to shift. Floyd, ff. Tr. 4894 at 7; Board Exh. 1-A, Stier Report, Vol. VI(G), Illjes 2/7/85 Inter-view at 45; id., Vol. VI(D), Frederick 3/12/85 Interview at 127-128; Board Exh. 6, OI Report, Exh. 24, Faust Interview at 6.

67. The procedures implementing the Tech Spec requirements for leak rate testing imposed several requirements on the CRO.

AP 1012, " Shift Relief and Log Entries," required CROs to main-tain and sign a Control Room Log, including a record of the start cnd completion or suspension times of all tests required by Tech Specs. Board Exh. 5-A, NRR Report, Vol. 1, Enclosure 1 at 2; Board Exh. 1-A, Stier Report, Vol. I at 44; id., Vol. V(C) at Tcb 18 (AP 1012, Sections 2.4 and 3.3.17). AP 1010, " Technical Specification Surveillance Program," provided that the operator parforming a surveillance task sign the data sheets where re-quired by the forms and procedures. AP 1012 also required the tast performer to document problems encountered during surveil-icnce testing and test results not meeting test acceptance criteria on an " Exception and Deficiency List." Board Exh. 5-A, NRR Report, Vol. 1, Enclosure 1 at 2; Board Exh. 1-A, Stier Rsport, Vol. I at 45; id., Vol. V(A) at Tab 17 (AP 1010, Sections i

3.2.2 and 3.2.4). SP 23Ol-3D1, "RCS Inventory," required the re-Eponsible CRO to enter the Action Statement under Tech Spec

3.4.6.2 if a leak rate test indicated that a limiting condition operation had been exceeded. Tr. 653 (Kirkpatrick); Board

Exh. 1-A, Stier Report, Vol. V(C) at Tab 19 (SP 2301-3D1, Sec-tions 6.4 and 7.2).

68. Several auxiliary operators were assigned to each chift. Generally, they reported to the CRO who operated the panel. They were stationed throughout the plant and assisted the CROs. Their duties included operating equipment that could not b3 operated from the Control Room. Board Exh. 1-A, Stier Report, Vol. I at 48-49. -These individuals were not required to hold any NRC license.

69. The following chart provides the shift compositions for both the five-shift rotation, from September 30, 1978 to December

-1978, and the six-shift rotation, from January, 1979 to March 28, 1979:14/

September - January -

Shift Position December 1978 March 1979 A11/ Shift Supervisor Zewe, W. Zewe, W.

Shift Foreman Scheimann, F. Scheimann, F.

CRO Frederick, E. Frederick, E.

. CRO Faust, C. Faust, C.

( CRO-in-training McGovern, H.ls/

14/ The individuals on a particular shift sometimes varied pri-morily because of vacations, illnesses, and training. The actual chift composition for a particular test was reconstructed by Stier and appears in Board Exh. 1-A, Stier Report, Vol. III(A),

Tcbles 1 and 2.

11/ 1978 Shift A--See Board Exh. 5-A, NRR Report, Vol. 1, Enclo-cure 1 of Enclosure 10 at 6; 1979 Shift A--See id., Enclosure 1 of Enclosure 6 at 1.

If/ McGovern was in training until receiving his RO license in November, 1978. See id., Enclosure 1 of Enclosure 10 at 6.

September - January -

Shift Position December 1978 March 1979 B17/ Shift Supervisor Chwastyk, J. Chwastyk, J.

Shift Foreman Conaway, W. Conaway, W.

CRO Kidwell, J. Kidwell, J.

CRO Illjes, T. Illjes, T.

CRO-in-training Mell, C. Mell, C.

CRO-in-training Hemmila, E.ls/

C19/ Shift Supervisor Mehler, B. No change Shift Foreman Adams, C. after December CRO Congdon, J. 1978 CRO Cooper, M.

CRO-in-training Phillippe, M.

D20/ Shift Supervisor Hitz, G. No change Shift Foreman Miller, A. after December CRO Olson, D. 1978 CRO Wright, L.

CRO Coleman, M.

E21/ Shift Supervisor Smith, B. Smith, B.

Shift Foreman Hoyt, K. Hoyt, K.

CRO Bocher, R. Booher, R.

CRO Hartman, H. Hartman, H.

CRO-in-training Blessing, J. Blessing, J.

CRO-in-training Germer, L.

17/ Shift B--See id., Enclosure 1 of Enclosure 9 at 1.

Ig/ Hemmila was in training until receiving his RO license in D:cember, 1978. See id., Enclosure 16 at 2; Board Exh. 1-A, Stier Report, Vol. V(A) at Tab 4 (8/15/78 Shift Assignment Sheet).

19/ Shift C--See Board Exh. 5-A, NRR Report, Vol. 1, Enclosure 1 of Enclosure 5 at 1.

l l 20/ Shift D--See id., Enclosure 1 of Enclosure 12 at 1.

21/ Shift E--See id., Enclosure 15 at 2; id., Enclosure 1 of Enclosure 3 at 1.

September - January -

Shift Position December 1978 March 1979 F22/ Shift Supervisor "F" Shift Bryan, K.

Shift Foreman did not Guthrie, C.

CRO exist Hemmila, E.

CRO McGovern, H.

CRO-in-training Germer, L.

(iii) Bodies Outside of the Chain of Command

70. The Plant Operations Review Committee was an advisory group that reported to the Unit Superintendent. It was an interdisciplinary committee consisting mostly of department heads cnd key individuals from the plant. PORC was responsible for re-view of procedure changes and plant modifications, as well as ac-tivities directly affecting the nuclear safety of the operating unit. Tech Spec 6.5.1.6 set forth the specific charges of PORC.

One such charge was investigating violations of Tech Specs, including preparing reports covering evaluations and recommenda-tions to prevent recurrence to the Station Superintendent and the GRC. PORC als'o was responsible for reviewing events requiring 24-hour notification to the NRC. The Tech Specs required PORC to provide written notification within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> to the Station Su-perintendent and the GRC Chairman of disagreement between PORC cnd the Unit Superintendent. The Unit Superintendent, however, had responsibility for resolving the disagreements. Board Exh. 1-A, Stier Report, Vol. V(B) at Tab 14 (Tech Specs 6.5.1.6 22/ Shift F--See id., Enclosure 1 of Enclosure 10 at 6.

cnd 6.5.1.7); Kunder, ff. Tr. 4800 at 3-4; Tr. 4817-20 (Kunder);

Herbein, ff. Tr. 5268 at 8.

71. The General Office Review Board was an advisory group that reported directly to the President of Met-Ed. GORB reports provided the President, who was not involved in the day-to-day cperational activities of the station, with the broad perspective of maintaining nuclear safety and appropriate radiation protec-tion. Herbein, ff. Tr. 5268 at 8.

72. The Generation Review Committee was a group organized to provide an independent review and audit of activities impor-tent to nuclear safety, which included procedural changes, plant m:difications, and violations of regulations. Id.

73. The Quality Assurance organization audited and in-cpected safety-related activities, including operations, mainte-ncnce, engineering, and licensing, to ensure compliance with pro-cadures developed by the functional groups. Id. at 9. QA had a cpecific responsibility to oversee surveillance testing. Board Exh. 1-A, Stier Report, Vol I at 138.

B. What Technical Specifications and procedures were relevant to leak rate testing?

74. The TMI-2 Operating License incorporated Tech specs that established limiting conditions for operation. Two sections of the Tech Specs addressed the requirements for pressure bound-cry leakage detection, Sections 3.4.6.1 and 3.4.6.2. Each of

these sections was divided into four parts: (1) limiting condi-tions for operation establishing minimum requirements for plant cperation; (2) an " Action" section, generally referred to as the

" Action Statement," describing the steps to be taken if a lim-iting condition for operation was exceeded; (3) surveillance rcquirements, designed to assure compliance with limiting condi-tions for operation; and (4) " bases" describing the background or purpose of the limiting conditions for operation. Id. at 36; id., Vol. V(B) at Tab 14 (Tech Specs 3.4.6.1 and 3.4.6.2).

75. Tech Spec 3.4.6.1 explicitly incorporated the provi-cions of Reg. Guide 1.45, establishing the three leakage detec-tion systems recommended in that Reg. Guide. The Tech Spec re-quired radioactive particulate and sump monitoring systems in cddition to either an air cooler condensate or a gaseous ra-dioactivity monitoring system. The Action Statement of this Tech Spec described the steps to be taken in the event these systems ware not continuously in operation. Its corresponding surveil-Icnce requirements specified how plant personnel were to demon-atrate that these systems were operable. The " basis" of this TOch Spec explained that it was intended to be consistent with R2g. Guide 1.45. Tech Spec 3.4.6.1, however, did not specify the 1sakage limits that it was intended to monitor, and it did not rcquire a leak rate test. Id., Vol. I at 36-37; id., Vol. V(B) at Tab 12 (Reg. Guide 1.45).

l l

1 1

i

i l

l 76. Tech Spec 3.4.6.2 established the following leakage limits as the limiting conditions for operation:

(1) no pressure boundary leakage; (2) 1 gpm of unidentified leakage; (3) 1 gpm of primary to secondary leakage through the steam generators; (4) 10 gpm of identified leakage; and (5) 8 gpm controlled leakage.

Board Exh. 1-A, Stier Report, Vol. V(B) at Tab 14 (Tech Spec 3.4.6.2).

77. The corresponding Action Statement to Tech Spec 3.4.6.2 rcquired that if any pressure boundary leakage was found, the plant had to be in " hot standby" within six hours and in " cold chutdown" within the next thirty hours. For any other leakage, including unidentified leakage, in excess of a limiting condition for operation, leakage was to be reduced to within Tech Spec lim-its within four hours or the plant had to be in hot standby with-in the next six hours and in cold shutdown within the new.t 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />. Tech Spec 4.4.6.2, which provided the corresponding sur-vaillance requirements, stated:

Reactor Coolant System leakages shall be dem-onstrated to be within each of the above lim-its by: [ monitoring radioactive particulate at least once per 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />; monitoring the sump at least once per 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />; monitoring controlled leakage once per 31 days andJ '

(d) performance of a Reactor Coolant System water inventory balance at least once per 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> during steady state operation.

l

l l

Id., Vol. I at.37-38; id., Vol. V(B) at Tab 14 (Tech Spec 4.4.6.2).

78. The " bases" for this Tech Spec described why er.ch of the leakage limits was reasonable. They did not indicate the ccurce from which the Tech Spec requirements were derived. Id.,

Vol. I at 38; see also Wermiel, ff. Tr. 376 at 5.

79. Similar to the definitions contained in Reg. Guide 1.45, Tech Specs 1.14 through 1.17 defined the categories of Icakage referred to in Tech Spec 3.4.6.2 as follows (Loard Exh. 1-A, Stier Report, Vol. I at 38-9; id., Vol. V(B) at Tab 14 (Tech Specs 1.14-1.17); see also Wermiel, ff. Tr. 376 at 5):

IDENTIFIED LEAKAGE 1.14 IDENTIFIED LEAKAGE shall be:

a. Leakage (except CONTROLLED LEAKAGE) into closed systems, such as pump seal or valve packing leaks that are captured and conducted to a sump or col-lecting tank.

b. Leakage into the containment atmosphere from l sources that are both specifically located and I known either not to interfere with the operation I

of leakage detection systems or not to be PRESSURE BOUNDARY LEAKAGE.

c. Reactor coolant system leakage through a steam generator to the secondary system.

UNIDENTIFIED LEAKAGE 1.15 UNIDENTIFIED LEAKAGE shall be all leakage which is not IDENTIFIED LEAKAGE or CONTROLLED LEAKAGE.

1 PRESSURE BOUNDARY LEAKAGE 1.16 PRESSURE BOUNDARY LEAKAGE shall be leakage (except steam generator tube leakage) through a non-isolable fault in a Reactor Coolant System component body, pipe wall or vessel wall.

CONTROLLED LEAKAGE I

1.17 CONTROLLED LEAKAGE shall be that seal water flow supplied from the reactor coolant pump seals.

80. The Tech Specs also established requirements for re-parting to the NRC when a limiting condition for operation was exceeded. Tech Spec 6.9.1.8 provided that in such instances a rcport had to be made to the NRC within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. This had to be followed by a more detailed Licensee Event Report within 14 days d3 scribing the event, the corrective action taken, and the steps that the licensee intended to take to prevent recurrence. Board Exh. 1-A, Stier Report, Vol. I at 39; id., Vol. V(B) at Tab 14 (Tech Spec 6.9.1.8); see also Wermiel, ff. Tr. 376 at 5, 11,

81. Finally, Tech Spec 6.10 required " records of surveil-Icnce activities . . . required by these Technical Specifica-tions" to be retained for at least five years. Board Exh. 1-A, l

Stier Report, Vol. I at 40; id., Vol. V(B) at Tab 14 (Tech Spec 6.10); see also Wermiel, ff. Tr. 376 at 5.

I l 82. Although Tech Specs required other leakage detection cystems, the leak rate test became the primary means of assuring compliance with the limiting condition for operation established in Tech Spec 3.4.6.2. The TMI-2 FSAR described the leak rate 1

l . -- - -- -

tost as the " primary means of detecting reactor coolant system

'1cakage." Board Exh. 1-A, Stier Report, Vol. I at 41; id.,

Vol. V(C) at Tab 15 (FSAR); see also Wermiel, ff. Tr. 376 at 8-9; Beard Exh 22, FSAR Sections 5.2.7.3 and 5.2.7.4. It was in fact the only one of the tests prescribed in the Tech Specs that pro-vided a quantifiable means of measuring unidentified leakage.

Sne Tr. 683 (Stier, Russell); Tr. 3865 (Mehler).

83. The following three interrelated procedures implemented Tsch Spec requirements for leak rate testing: (1) SP 2301-3Dl, "RCS Inventory," controlling the performance of leak rate tests; (2) AP 1012, " Shift Relief and Log Entries," dealing with logging rcquirements; and (3) AP 1010, " Technical Specification Surveil-Icnce Program," covering reporting the results of surveillance tssts.23/ Board Exh. 1-A, Stier Report, Vol. I at 41; id.,

g23 The NRC Staff identified one additional procedure, AP 1036, Instrument Out-of-Service Control" (Board Exhs. 12-A and 12-B),

oc a relevant procedure to this proceeding. Tr. 1167-68 (Russell). Mr. Russell stated:

l l The purpose of the procedure is to describe the method of control of readout devices which become inoperable l or are strongly suspected of being inoperable, such that they are marked, documented and controlled until repair is effected. That's under section 1.1; purpose.

Under section 2.1, responsibilities: " shift supervi-sors, shift foremen and control room operators. The shift supervisor and/or shift foreman and/or control room operators are responsible to assure out-of-service instruments are identified and logged "out-of-service."

They are also responsible to assure that work requests are submitted to effect repair and remove the out-of-service stickers from the instruments and close (Continued Next Page)

Vol. V(C) at Tabs 17 (AP 1010), 18 (AP 1012), 19 (SP 2301-3D1).

84. SP 2301-3D1 stated that its purpose was to assure com-pliance with the leakage limits set forth in Tech Spec 3.4.6.2.

i It required performance of a leak rate test at least once every 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> during " steady state" in Modes 1, 2, 3, and 4, when, ac-cording to the Tech Spec definition of Modes, RCS temperature av-eraged greater than 200*F. Board Exh. 1-A, Stier Report, Vol. I at 42; see also Board Exh. 5-A, NRR Report, Vol. 1, Enclosure 1 ct 2.

85. The procedure prescribed the plant conditions required for the performance of a leak rate test. Certain " operations should not be conducted," for example, "(a) make-up or chemical cddition to the makeup system, . . . (e) boration or d;boration." " Operations such as adding water to the Make-up tcnk or sampling the RCS [ reactor coolant system] may be ac-counted for [in the test calculation] . . ., however, these l (Continued) l out the out-of-service instrument log entries."

This is the procedure which promulgates the log that we have been discussing, and I think the procedure is quite clear and it indicates that out-of-service instruments and suspected instruments are not to be used and they are to be tagged and indicated and those l tags are to remain on them until such time as repairs l are effected and they are again reliable for the pur-pose of operating the facility.

Tr. 1803 (Russell).

I l l

l

r thould be avoided if at all possible." The reactor coolant and mnkeup systems should be maintained in a " steady state condi-tion." Valve line-ups should be maintained the same throughout the test, and reactor power and RCS temperature and pressure thould not be altered. The procedure also directed use of the same instruments to obtain data at the beginning and end of the tost. Board Exh. 1-A, Stier Report, Vol. I at 42.

86. The procedure also provided for the performance of the tsst normally by means of the plant computer. Operators were clerted to the inability of the computer to perform leak rate tsst calculations when RCS temperature was below 520*F. At such times, operators were to perform the calculation manually, and the procedure provided a data sheet for that purpose. Id. at 43; Board Exh. 2, Faegre & Benson Report, Vol. 4 at Tab 2 (SP 2301-3D1, Rev. 3, Section 4.3); see also Board Exh. 5-A, NRR Raport, Vol. 1, Enclosure 1 at 2.

87. If an operator had to change the RCS inventory during the test, he had to account for the change in the calculation and complete a " Data Sheet 4" to indicate the quantity added to the RCS and the operation that caused the change. Board Exh. 1-A,

! Stier Report, Vol. I at 43.

1

'88. The procedure contained directions for responding to tsst results that exceeded limiting conditions for operation.

l The first requirement was that an operator begin a new leak rate l

l l

t:st. Next, the operator was to determine whether unaccounted-for operator action had affected the initial test.

If such action had taken place, the operator was to invalidate the test. The procedure required that the operator record such cction in the " remarks" section of the data sheet. Id.; see also Board Exh. 5-A, NRR Report, Vol. 1, Enclosure 1 at 2.

89. Operators were then required to check for leakage. If cn operator found such leakage, he was to document the leakage on a " Data Sheet 3" along with its flow rate and a description of the method used to determine the rate. The shift supervisor was then required to evaluate the safety significance of any such leakage that the operator had identified. Board Exh. 1-A, Stier Rsport, Vol. I at 43-44.

90. If, after these steps were taken, test results remained in excess of acceptance criteria, the procedure required that the process of shutting down the plant should begin according to the Action Statement of Tech Spec 3.4.6.2. Id. at 44; see also Board Exh. 5-A, NRR Report, Vol. 1, Enclosure 1 at 1.

l l

91. AP 1012 set forth requirements for logging surveillance l tosts. A Control Room Log, maintained by a CRO, was to include a l

rccord of the start and completion or suspension times of all tssts required by Tech Specs. The Supervisor of Operations was 1

rcquired to review and sign the Control Room Log at least once par week, indicating that he was satisfied with the recording

't chniques and was familiar with any operating abnormalities.

The procedure also required maintenance of a Shift Foreman's Log, but did not specify that the Log contain any information relating to leak rate tests. Board Exh. 1-A, Stier Report, Vol. I at 44; ese also Board Exh. 5-A, NRR Report, Vol. 1, Enclosure 1.at 2.

92. AP 1010 established a system for documenting the re-sults of surveillance tests that could not be performed success-fully or failed to meet acceptance criteria. Any surveillance test that had an unsatisfactory result had to be documented on an

" Exception and Deficiency List" and filed. Filing an " Exception" I

was necessary in the event of a failure to obtain " required plant conditions" or an innbility to use "an existing procedure (i.e.,

aquipment out-of-service or a procedure which cannot be fol-lowed)." Filing a " Deficiency" was necessary when a test had baen completed but acceptance criteria had not been met. Board Exh. 1-A, Stier Report, Vol. I at 45; see also Board Exh. 5-A, NRR Report, Vol. 1, Enclosure 1 at 2-3.

l

93. A shift supervisor was required to review and initial all Exceptions and Deficiencies ("E&Ds") "as soon as possible."

In *h' case of a Deficiency, he was to determine whether a re-portable occurrence had taken place. If so, he had to bring the matter to the attention of the Unit Superintendent. The E&D List l

would then be attached to the test data package and filed with the GMS Coordinator. Board Exh. 1-A, Stier Report, Vol. I at 45.

f i

l

94. The GMS-Coordinator, who~was appointed by the Unit Su-parintendent to administer the GMS (Generation Maintenance Sys- ,

j tgm) program covering all plant maintenance and testing', would inform the PORC Chairman and the QC Supervisor when the maximum allowable time interval between surveillance tests had been ex-cseded. The GMS Coordinator would forward any E&D to the PORC Chairman and QC Supervisor as soon as practicable. He would also mnintain a follow-up action log of all E&Ds. Finally, the QC Supervisor was designated to provide general oversight for all curveillance testing. Id. at 45-46.

C. How were the Technical Specification 3.4.6.2 requirements for reactor coolant system unidentified leakage interpreted and implemented by control room operators (CROs),

! shift foremen, shift supervisors and on-site and off-site management? Following the discovery by an NRC inspector in October 1978 that Technical Specification 3.4.6.2 require-ments were not properly interpreted or implemented, what corrective action was taken by management personnel? Was the corrective action taken sufficient to insure compliance with the Technical Specification 3.4.6.2 by the personnel performing and reviewing the leak rate surveillance tests? (CLI-85-18, 22 N.R.C. at 880, Issue (a))

95. In October 1978, a series of events occurred resulting in the issuance by Met-Ed of Licensee Event Report 78-62/1T ("the LER") to the NRC on November 1, 1978. The Commission, in its Order and Notice of Hearing, requested the Board to address three specific questions pertaining to these events: (1) how were the Tech Spec 3.4.6.2 requirements for RCS unidentified leakage i

l

l l

l

~ interpreted and implemented by CROs, shift foremen,-shift super-visors, and on-site and off site management; (2) following the discovery by an NRC inspector in October 1978 that Tech Spec 3.4.6.2 requirements were not properly interpreted or imple-L m:nted, what corrective action did management personnel take; and I

(3) was the corrective action taken sufficient to insure compli-cnce with Tech Spec 3.4.6.2 by the personnel performing and re-i viewing the leak rate tests? CLI-85-18, 22 N.R.C. at 880. To cnswer these questions, the Board has found it useful and neces-ocry to examine the history of leak rate testing at.TMI-2 in 1978 i and 1979, including the chronology of the LER-related events.

i-(i) Training

96. One area of inquiry pursued by the Board in this pro-caeding was the extent of training received by the operators on the conduct of leak rate testing. With only a few exceptions,24/

24/ Hartman stated that he learned on shift how physically to parform the test and enter the data into the computer, but he re-msmbered "some discussions" in the classroom about leak rate Tech Specs and requirements, and recalled doing a hand calculation once using numbers provided by an instructor. Tr. 2248-49 (Hartman). But see Tr. 2839-40 (Cooper) (recalled receiving a copy of every plant procedure, including one on hand calculating l leak rates, but stated that he did not study or perform this cal-l culation while in the classroom). Kenneth Bryan, Shift Supervi-cor of Shift _F, stated that he never specifically instructed his chift on how to interpret leak rate tests "because this subject was part of their training on technical specifications." Bryan, ff. Tr. 4540 at 3. When asked about his understanding of where operators received their training on the Tech Specs relating to leak rate testing, Bryan responded, "Well, probably from the training, training department during their formal training (Continued Next Page)  !

i the_ operators generally testified that in the area of leak rate tcsting they were trained on-the-job, not formally in the class-room by instructors. See, e.g., Tr. 2582-83 (Coleman);

Tr. 2798-99 (Scheimann); Tr. 2838-41 (Cooper); Tr. 3198, 3207 (McGovern); Tr. 3247, 3239 (Mell); Mell, ff. Tr. 3239 at 1.

97. The operators' general statements of the on-the-job em-

.phasis of their training on leak rate testing was confirmed by the testimony of Dennis J. Boltz, a witness provided by GPUN at the Board's request, with a particular knowledge of the training program provided by Met-Ed for operators who were at TMI. See qsnerally Tr. 2216-34 (Boltz). In that period of time, Boltz was en instructor in the Training Department assigned to the licensed operator training group. Tr. 2218-19 (Boltz). His responsibil-ities included the development and presentation of classroom training and the development and implementation of written weekly quizzes, annual NRC evaluations, and on-shift oral exams of (Continued) classes and on-the-job training." Tr. 4546-47 (Bryan). Joseph Chwastyk, another shift supervisor, stated that he probably ran leak rate tests "as part of my training," but he did not recall any specific training on the subject. Chwastyk, ff. Tr. 3407 at 1. It is unclear whether Chwastyk meant he probably ran leak rate tests on the job or in the classroom. Gregory Hitz, a third chift supervisor, stated that he did not remember attending a cpecific training class on the leak rate Tech Spec, "although I mny have." Hitz, ff. Tr. 3664 at 1-2.

The Board does not consider these few statements to affect the validity of the consistent observation made by virtually all o r. the operators that leak rate training was provided on-the-job.

cperators in both their qualification and requalification pro-grams. Tr. 2219 (Boltz). Boltz specifically agreed with the B:ard's summary statement that responsibilities for instruction en the theory and practice of the leak rate tests "were entirely

~

given over to on-the-job training." Tr. 2231-32 (Boltz).

98. Boltz provided a document to the Board describing the CROs' initial training program. See Board Exh. 16, " Training De-pertment Administrative Meanorandum Number 5, Change 2";

Tr. 2219-20 (Boltz). That document explains that when an auxil-icry operator advanced to " Category IV CRO," he immediately cntered a CRO training program consisting of (1) specific study casignments, (2) oral checkouts in which the individual actually parformed or simulated performing certain evolutions, (3) written tests, (4) oral examinations, and (5) classroom sessions. Board Exh. 16, Training Memorandum at 1. It then sets forth the actual contents of these five aspects of the program. The sole refer-ence to leak rate testing in this document is to an operator's capability "of using computer for calculations of leak rate."

Id. at 12.

l

99. In response to Board questions on the source of in-struction on specific topics relevant to leak rate testing, Boltz consistently pointed to on-the-job training. Training on l

AP 1010, which required the documentation of unsatisfactory tests on an E&D List, was conducted on shift and supervised by the chift supervisor. Tr. 2222 (Boltz). Thus, instructions on how 1

to handle an Exception or Deficiency were handled, not by the Training Department, but by the members of a particular crew.

Tr. 2223-24 (Boltz). Similarly, consideration of AP 1012, which d2 scribed requirements for operating logs, occurred on-the-job rather than in the classroom. Tr. 2224-25 (Boltz). Boltz also could not answer whether the training program made clear the m2aning of the requirement of " steady state" conditions in the curveillance procedure because that procedure "would have been ttught on shift rather than in the classroom." Tr. 2227-28 (Boltz).

100. The Training Department "had to rely heavily" on on-the-job training with qualified personnel on-shift "to support the training and qualification of the new people that would be placed into the control room positions." Tr. 2225 (Boltz). The formal training provided by the Training Department was a " text-book type of training." Tr. 2226 (Boltz). Although training ma-tsrials may have covered the Tech Specs and surveillance proce-dures (Tr. 1396 (Stier); see also Tr. 2839 (Cooper); Tr. 2582 l (Coleman); Tr. 4009-11 (Olson); Tr. 4104 (Hemmila); Tr. 4229-30 l

l (Booher)), actual training on the procedures and Tech Specs per-taining to leak rates was left to on-shift practice and instruc-l tion. The Board notes Boltz's view that since leak rate testing was "an every shift operation that occurred when the unit was on line and at hot conditions, there apparently was deemed to be no training need since it was a normal evolution." Tr. 2226-27 (Boltz).

l

1 l

l l

101. The record in this proceeding belies the apparent be-

]

lief by the Training Department that on-the-job training on leak

(

rcte testing was adequate. The record indicates a remarkably consistent lack of knowledge and appreciation among the operators i

l of the technical and procedural requirements, and the critical reasons for those requirements, relating to leak rate testing.

The Board finds that the reliance by the Training Department on the Operations Department for instruction on the proper conduct of leak rate tests reflected a misplaced assumption about the cufficiency of on-shift training experience in this instance.

Adequacy of instruction alone, however, cannot excuse the indi-vidual operators for bearing personal responsibility for inappro-priate conduct "far removed" from the requirements of the sur-veillance procedures and Tech Specs and resulting in part from the operators' erroneous interpretations and applications of these requirements both before and after the events of October 18, 1978. See Tr. 1395-97 (Stier).

(ii) General Practice Prior to October 1978 l

102. From the period of time when leak rate testing began in March 1978 to October 1978, ths najority of Operations Department l

parsonnel, including CRos, shii 'oremen, and shift supervisors, ganerally worked under an inter tation that Tech Spec 3.4.6.2 l

l end its corresponding surveilla requirement, Tech Spec 4.4.6.2, required entry into the Action Statement only if they were unable to obtain a leak rate test result of 1 gpm or less

ence in a 72-hour period. See, e.g., Congdon, ff. Tr. 2709 at 2; Board Exh. 1-A, Stier Report, Vol. VI(G), Illjes 2/7/85 Interview at 14; Cooper, ff. Tr. 2835 at 5; Hitz, ff. Tr. 3644 at 3; Bryan, ff. Tr. 4540 at 2; McGovern, ff. Tr. 3148 at 3; Tr. 3151-52.

(McGovern). This assumption enabled operators generally to fol-low a practice whereby tests greater than 1 gpm.were discarded cnd tests of 1 gpm or less were filed. See, e.g., Tr. 2715-16 (Congdon); Hit , ff. Tr. 3664 at 3; Adams, ff. Tr. 3776 at 2; Tr. 3614-15 (A. Miller); see also Board Exh. 1-A, Stier Report, Vol. I at 58-59. Operators at times would. search for leakage or

" eyeball" plant parameters for indications of excessive leakage after obtaining a test greater than 1 gpm, but at other times i

they would not make any efforts to determine the validity of a test before discarding it. See, e.g., Tr-. 2714-16 (Congdon);

i Tr. 2522-24 (Faust); Tr. 3109-11 (Conaway); Tr. 4117-18 (Guthrie); Tr. 4241-47, 4252-53 (Hoyt); Smith, ff. Tr. 4331 at 4; Tr. 4359-62 (Smith); Tr. 4545-46 (Bryan). Operators in effect ware simply going through the motions of conducting leak rate tests to satisfy a procedural requirement that they interpreted to allow repeated running of tests in a 72-hour period without immediate entry into the Action Statement after conducting the first valid test with a result greater than 1 gpm.

l

(iii) Events Related to the Issuance of the LER 103. In early October 1978 Seelinger, who was the Unit 2 Su-1 parintendent of Technical Support at that time and therefore a ksy member of on-site management, became aware of the common practice among operators of running tests until they obtained one under 1 gpm and discarding all tests above 1 gpm. Tr. 4745 (Seelinger). Seelinger testified that his " initial reaction" was that he did not think such a practice was permissible. Id.

Sselinger believed that the Tech Specs required operators to enter the Action Statement immediately after obtaining a leak rete greater than 1 gpm. Tr. 4765 (Seelinger). Seelinger, how-cver, cited two incidents that led him away from implementing his initial interpretation.

104. First, when Seelinger stated to a shift supervisor that ha did not believe the prevailing operator practice was permis-cible, the shift supervisor informed him that his interpretation

of the Tech Specs would shut down TMI-1. Tr. 4745 (Seelinger).

l Snelinger stated that although he resolved in his mind at that time to raise the issue of interpretation with the TMI-1 PORC, l other tasks occupied his time and he had not approached the TMI-1 PORC prior to the inspection on October 18, 1978 by NRC inspector Donald R. Haverkamp, which brought the interpretation question to the fore. Tr. 4746-47, 4756 (Seelinger); Board Exh. 20, i McVerkamp Testimony at 2.

r I

105. Second, Seelinger stated that he met with Floyd at about the time of the shift supervisor's comment and, after I - "looking closely at the technical specification," Floyd and Sselinger agreed on the following interpretation:

4

[T]he leak rate test was to be set aside if it exceeded the criteria on the basis i of the fact that.the plant was not in i steady state operation and would allow

[ the running of another, or another, and we d.idn't specify the number of leak rates.

'fr . 4746 (Seelinger). Although setting aside tests greater than

, 1 gpm unidentified leakage may have been intended to allow opera-tors to look for trends, the practical effect of the interpreta-

-tion was that if a test was less than 1 gpm, operators presumed that steady state prevailed, but if a test was greater'than 1 cnxn, they presumed that steady state did not exist. See Tr. 4753-56 (Seelinger). Seelinger stated that at this meeting hs had argued for an interpretation of the Tech Specs that would rcquire immediate Action Statement entry on obtaining a test re-f cult greater than 1 gpm, while Floyd had argued for an interpre-tation that required one test under 1 gpm every.72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> and entry into the Action Statement only at the end of such a 72-hour period. Tr. 4764-65 (Seelinger). Seelinger,.in other words, claimed that Floyd held an interpretation of the Tech Specs that allowed operators to continue their practice of repeatedly run-ning tests without entering the Action Statement until 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> had elapsed from the last test less than 1 gpm and that Floyd's l

l l

'"once every 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />" interpretation prevailed over Seelinger's "immediate action statement entry" interpretation when Seelinger daferred to Floyd's interpretation. See Tr. 4764 (Seelinger).

106. Floyd stated that he did not recall this discussion, but he did not disagree with Seelinger's recollection.

Tr. 4928-30 (Floyd). Floyd emphasized that his understanding was that a valid test greater than 1 gpm required immediate entry into the Action Statement. Tr. 4930-34 (Floyd). He testified that he believed at the time of this meeting that "[i]f a leak rate is valid, no matter what its numbers are, if it can't be in-validated for cause it's then a valid leak rate and you enter the action statement." Tr. 4930-32 (Floyd). Floyd stated that the practice of discarding tests " grew out of Operations and was car-ried from Unit 1 to Unit 2 by me, the supervisors, and the foremen." Floyd, ff. Tr. 4894 at 2. Floyd, however, apparently thought that cperators were discarding only tests they had "in-validated for cause," and such a test was "a worthless piece of peper." See Tr. 4930 (Floyd); Floyd, ff. Tr. 4894 at 2-3.

107. Floyd, in other words, stated that he always held the interpretation of the Tech Specs that Haverkamp insisted on in October 1978. Tr. 4930-31 (Floyd). See discussion below. Floyd thus stated that paragraph 2 of a later October 20, 1978, Opera-tions Memorandum he wrote to " correct" the leak rate practices, following the Haverkamp events reflected merely a continuation of what he thought had been the practice all along. Tr. 4936-37 l

l

(Floyd). If Floyd's recollection is accurate, the interpretation ha held would be precisely the marching order of the November

-LER, and Seelinger and Floyd in their meeting of early October would be in agreement over a proper interpretation of the Tech Spec. See Tr. 4930-31 (Floyd). The Board finds such a scenario

improbable. However, we are unable to determine whether the testimony of Seelinger and Floyd on their early October meeting is inconsistent or indicative of simple miscommunication. It is possible that Seelinger did not inform Floyd and that Floyd was I otherwise unaware of the discarding of not only invalid but also l

-valid tests greater than 1 gpm and that Floyd did not inform Szelinger of his firm belief that a test greater than 1 gpm that was valid required immediate entry into the Action Statement.

i The nuances of a disagreement that occurred over eight years ago may never be definitively resolved. What *7 clear in the record, however, is that at least one member of on-site management, Stelinger, became aware of an erroneous interpretation and imple-msntation by operators of Tech Spec 3.4.6.2. and failed to re-cpond effectively. Whether Seelinger's failure to respond was partly based on an understanding or a misunderstanding of Floyd's Toch Spec interpretation is irrelevant to this finding by the Board.

108. Sometime after the Seelinger and Floyd meeting, the matter of the proper interpretation of Tech Spec 3.4.6.2 briefly CF.me up at a POD meeting attended by Seelinger but apparently i

l l

t

without significant analysic or conclusion. See Tr. 4752 (Seelinger). Because Seelinger "did not particularly care for the interpretation" and he "was not particularly crazy on having the NRC involved with that interpretation or finding out about that interpretation," he subsequently advised one or more shift cupervisors "that the tests were not to be left out and lying cround," where the NRC might find them. Tr. 4756 (Seelinger).

109. With this background, we now turn to the details of NRC inspector Haverkamp's discovery in October 1978 of inappropriate leak rate practices. The record indicates that on October 15, 1978, at 7:27 p.m., a leak rate of -0.3504 gpm was obtained.25/

Ste NRR Test No. 12. On October 16, 1978, at 7:35 p.m., a leak rate of 2.5645 gpm was obtained, substantially above the 1 gpm limit set by the Tech Specs. See NRR Test No. 12A. On October 17, 1978, at 1:27 p.m., a leak rate of 2.0738 gpm was obtained, again substantially above the Tech Spec limit. See NRR Tast No. 12B. On October 18, 1978, at 5:13 a.m., a leak rate of 1.7754 gpm was obtained, again well above the Tech Spec limit.

Sne NRR Test No. 12C.

25/ For purposes of this discussion, the Board assumes that such c relatively small negative leak rate was a " valid" leak rate i less than 1 gpm. See Tr. 4901 (Floyd). Our assumption is sup-ported by the record, which indicates that small negative leak rates were considered acceptable. See, e.g., Tr. 2797-98 (Scheimann); Tr. 4901 (Floyd); Tr. 2529-32 (Faust); Coleman, ff.

Tr. 2579 at 5: Tr. 2652-53 (Coleman).

l 110. The shift supervisor turnover note from the midnight to morning shift on October 18 stated, "Still could not get a leak rete -- 1900 today is deadline Doing hand calculations." Board Exh. 1-A, Stier Report at Tab 10 (10/18/78 Daily Plant Status Rgport, p. 104, Item 11). Brian Mehler, shift supervisor of the shift that was on duty that morning, said he was aware of the n;ed for a good leak rate and the fact that the 72-hour clock was running out. Tr. 3893 (Mehler). At 7:35 a.m., a leak rate of 1.2939 gpm was obtained. See NRR Test No. 12D. Charles Adams, the shift foreman on Mehler's shift, testified that Floyd usually came into the Control Room around 6:00 a.m., and, though he did not recall exactly what time Floyd arrived on October 18, he re-called Floyd sitting at the computer console, where leak rate tests were run. Tr. 3797 (Adams). Floyd did not recollect the morning events of October 18, but he did note that he frequently est at the computer console and sometimes started a test for an operator by typing "RCSL." Tr. 4898-99 (Floyd). Mehler did not recall Floyd's presence in the Control Room "immediately in the morning, early" but implied that Floyd was there before Haverkamp arrived. Tr. 3893-94 (Mehler). At 8:59 a.m., a leak rate of 1.3219 gpm was obtained. See NRR Test No. 12E.

111. At appres'.mately 9:00 a.m. on October 18, Haverkamp, who was on-site conducting a routine inspection of plant opera-tions, arrived in or near the Control Room, overheard a discus-cion by a CRO, shift foreman, and shift supervisor about bad

tzsts. Board Exh. 20, Haverkamp Testimony at 2-3. Haverkamp joined the discussion and saw some bad tests. Id. Haverkamp in-tarrupted his routine operations inspection schedule to question Muhler and/or Floyd about the tests. Id. at 4. He then learned of the interpretation being given to the Tech Specs and left the Control Room to discuss this with Seelinger in his trailer. Id.

HLverkamp went directly to Seelinger because Seelinger was Hsverkamp's " primary point of contact for TMI-2." Tr. 2113 (Haverkamp). Discussing the type of day in, day out relationship ha had with Met-Ed, Haverkamp testified that "when it came to plant operations, technical questions about engineer-related questions, I spoke frequently with Mr. Seelinger, probably . . . during each inspection." Id.

112. Haverkamp recalled that, shortly after he began his mseting with Seelinger, Floyd entered Seelinger's trailer office and joined the discussion. Board Exh. 20, Haverkamp Testimony at 4-5. Haverkamp clearly recalled Floyd telling him, in effect, that unidentified leakage test results must be calculated to be under 1 gpm only once every 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> to be in legal compliance with the Tech Spec surveillance requirements. Id. In Floyd's view, any number of test results could be greater than 1 gpm as long as acceptable results were obtained once every 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />.

Id. Haverkamp informed Floyd that his interpretation was clearly incorrect and stated that the Tech Spec limiting conditions for operation limits must always be met, in accordance with the

cpplicable Tech Spec facility operating modes, including the re-Eults of leak rate tests that were conducted more often than re-quired by the 72-hour Tech Spec surveillance frequency. Id.

Floyd did not recall this conversation but had no reason to be-lieve Haverkamp's recollection was incorrect. Tr. 4926-27 )

(Floyd).

113. In the discussion with Seelinger and Floyd, Haverkamp was informed that plant operators were attempting to identify any leakage sources to reduce unidentified leakage to less than 1 gpm. Board Exh. 20, Haverkamp Testimony at 4-5. Seelinger as-cured Haverkamp that the plant would be operated in accordance with the applicable Tech Spec Action Statements and informed him that the matter would be referred to PORC for its review as a

_potentially reportable occurrence. Id.; see also Tr. 2050-51 (Haverkamp); Tr. 4630-32 (Seelinger). Also during this discus-sion, the possibility of rounding off test results.was discussed after Haverkamp raised some questions about the specified limit being "1" gpm as opposed to "1.0" gpm. Board Exh. 20, Haverkamp Testimony at 8-9; Tr. 2111-13 (Haverkamp); Seelinger, ff.

~

Tr. 4623 at 7; Tr. 4770-72 (Seelinger).

114. Sometime on the morning of the 18th, Seelinger met with G. Miller and informed him of the Haverkamp-related events.

Tr. 4722-27 (Seelinger). Seelinger recalled that during this meeting, a telephone conversation took place with a Met-Ed manag-or in Reading about how to handle the situation. Id. Although Scelinger is the only member of either on-site or off-site man-egement who remembered such a telephone conversation, no one dis-

.putes that it took place. The message Seelinger recalls rsceiving during that conversation was to review the item for re-portability and act appropriately. Id.

115. At 10:16 a.m., a leak rate of 1.0246 gpm was obtained.

Sne NRR Test No. 13. At approximately noon, Mark Bezilla, the procedure coordinator and PORC secretary for TMI-2 (Board Exh. 1-A, Stier Report, Vol. VI(A), Bezilla 2/22/85 Interview at 1) at Seelinger's request brought Floyd three tests, NRR Test Nos. 12C-E, to analyze. Tr. 4913-20 (Floyd). On Test No. 12C, whicn indicated a leak rate of 1.7754 gpm at 5:13 a.m., Floyd wrote that " rounds off high but is corrected by leak rate 10/18/78 7:35:27 start time ie into action statement at 5:13:02 out of it at 7:35:27." Tr. 4913-14 (Floyd). On Test No. 12D, which indicated a leakage of 1.2939 gpm at 7:35 a.m., Floyd j wrote, "OK by roundoff, JRF, 1200, 10/18/78." Tr. 4914 (Floyd).

l On Test No. 12E, which indicated a leakage of 1.3219 gpm at 8:59 l

a.m., Floyd again wrote. "OK by roundoff." These tests indicate that by noon on October 18, a practice of rounding off test re-nults was in effect. Mehler specifically recalled that at some point in time Floyd came to him and advised him that he was going to be able to round off leak rates as a result of a discuscit Floyd had had with Haverkamp, whose opinion allowed rounding off to the nearest whole number. Board Exh. 1-A, Stier Report,

Vol. VI(I), Mehler 3/15/85 Interview at 18. Mehler then rounded off Test No. 13, which had been run that morning, and wrote on the test " Net Unidentified Leak Rate rounded off To Nears whole Number 1 GPM." See id. at 19, 37-39; see also id., Vol. VI(A),

Adams 3/13/85 Interview at 44-48.

116. At 12:21 p.m., a leak rate of 0.1081 gpm was obtained, later corrected by William Fels 2p/ to indicate a leak rate of

.283 gpm. See NRR Test No. 14; Tr. 4520-31 (Fels). Probably in the early afternoon, Haverkamp was shown this test, with an acceptable leak rate of 0.1081 gpm, and was informed that some -

amount of leakage had been identified and a computer input error --

found and corrected, both reducing unidentified leakage. Board Exh. 20, Haverkamp Testimony at 6-7; Tr. 2054-57, 2131 (Haverkamp). These statements about an identification of leakage and a correction of a computer input error were repeated in the later LER and are addressed below. Suffice it to say here, how-ever, the accuracy of these statements are nct supported by the record and the apparent reason for obtaining acceptable tests --

rounding off -- was not disclosed to Haverkamp the afternoon of October 18.

2@/ Between March 1978 and March 1979, Fels was the engineer as-signed to start up and check the TMI-2 computer systems. Board Exh. 1-A, Stier Report, Vol. II(A), Fels Summary at 1. He was .-

the individual who installed the computerized leak rate test pro-gram at TMI-2. Id., Vol. VI(C), Fels 2/25/85 Interview at 2.

117. At approximately 4:30-5:00 p.m. PORC met to review re-portable occurrences. Tr. 4780-81 (Seelinger). It is likely that at this meeting PORC determined that a reportable occurrence had occurred when the limiting condition for operation for Tech Spec 3.4.6.2 was not invoked at 7:35 p.m. on October 16 when a leak rate of 2.6 gpm was obtained. Id.; see also Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 31 (Seelinger 10/19/78 Letter to Grier ("Grier Letter")).

118. In the afternoon of October 19, Seelinger informed Haverkamp, who was still on site, that PORC had met to review operation of the facility during October 16-18 with unidentified -

leakage greater than 1 gpm and that PORC had determined the mat-ter to be a reportable occurrence. Board Exhibit 20, Haverkamp Testimony at 6-7. At that time, Seelinger showed Haverkamp a draft copy of the intended prompt notification letter to Grier at the NRC, presumably without the first paragraph, which refers to a 3:30 p.m. conversation between Seelinger and Haverkamp. Id.;

see Board Exh. 1-A, Stier Report Vol. V(C) at Tab 31 (Grier Let-ter). But see Board Exh. 1-A, Stier Report, Vol. VI(A), Bezilla 2/22/85 Interview at 2, 19 (where Bezilla, TMI-2 Procedure Coordinator and PORC Secretary, stated that Seelinger asked him to draft a letter to Grier "to confirm" his prior conversation with Haverkamp).

119. Later in the day on October 19, Scelinger sent his let-ter to Grier informing him of the reportable occurrence. Board

.Exh. 1-A, Stier Report, Vol. V(C) at Tab 31 (Grier Letter).

Bezilla wrote the body of the Grier letter; Seelinger probably made corrections to it. Id., Vol. VI(A), Bezilla 2/22/85 Inter-view at 15-22. The letter reported the discovery of a reportable situation at 10:00 a.m. on October 19 when it was determined that the Action Statement had not been entered at 7:35 p.m. on October 16 after obtaining a leak rate of 2.6 gpm. Id.,

Vol. V(C) at Tab 31 (Grier Letter). Seelinger speculated that the reference in the Grier letter to discovery of a reportable situation at 10:00 a.m. on the 19th " reflects a conversation or something I do not recall it directly with Mr. Floyd when he was brought on board, relative to our intentions to go forward with the prompt report." Tr. 4781 (Seelinger). At that time, to in-form the Operations Department of the actions and intentions of PORC concerning the events of October 16-18, "Mr. Floyd was brought on board as a final item for a final discussion relative 4

to the reportability." Tr. 4781-82 (Seelinger). Seelinger also stated that one of the reasons for the 10:00 a.m. reference "may have been the requirement to report within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> on a prompt reportable occurrence." Tr. 4782 (Seelinger); see also Board Exh. 1-A, Stier Report, Vol. V(B) at Tab 14 (Tech Spec 6.9.1.8).

120. The Board dismisses the vague recollection by Seelinger that the reference to 10:00 a.m. on October 19 reflects the time Floyd was " brought on board" about goir.g forward with the prompt report. The letter itself speaks in terms of a reportable

cituation " discovered" at 10:00 a.m. "when it was determined that" the Action Statement had not been entered at 7:35 p.m. on October 16. Such a " discovery" more likely occurred the morning of October 18 at the meeting between Seelinger, Floyd, and H verkamp. We find more plausible Seelinger's statement that concern about a failure to meet the 24-hour deadline on a prompt rcport lay behind the October 19 10:00 a.m. reference. Whatever the underlying basis, the statement in the letter is clearly in-cccurate.

i 121. The Grier letter also contains the statement that 4

"[u]nidentified leakage was reduced to [ Tech Spec) limits at 0735 (n 10-18-78" but does not state that rounding off was the method of reduction. We find that the statement is misleading.

122. On October 20, Floyd issued an Operations Memorandum to the TMI-2 shift foremen and supervisors explaining in the first two paragraphs, respectively, that (1) the leak rate computer program had be'en modified to round off and (2) "If the unidentified leakage is equal to or greater than 2 gpm, then the 4-hour time clock commences with the output time of the printout i

io., the time the final data was taken." Board Exh. 1-A, Stier Rcport, Vol. V(C) at Tab 34 (Floyd 10/28/78 Operations Memoran-dum); see also Tr. 4934-44 (Floyd). Floyd wrote the first para-graph after he learned that Fels was reprogramming the computer to round off "to tell the operators that the computer was going to be lying to them by rounding off numbers." Tr. 4935 (Floyd).

Floyd testified that he wrote the second paragraph to reemphasize o procedure that he thought was longstanding but which he real-ized, after analyzing Test Nos. 12C-E at noon on October 18, had not been followed. Tr. 4935-37 (Floyd). The second paragraph wcs a poorly phrased and therefore unclear attempt by Floyd to instruct the operators about the need to enter the Action State-m:nt immediately on obtaining a valid test equal to or greater than 2 gpm.27/ In Floyd's opinion, Operations Memoranda were more important than other material placed in the Control Room to ba read by the licensed operators. Tr. 4943-44 (Floyd).

Although the record is clear that Floyd was unsuccessful in his effort through this memorandum sufficiently to instruct the oper-ators (see 11 135-137, infra), the Board finds that Floyd did make an effort to change improper conduct by the operators on Test Nos. 12C-E by issuing one of the more important types of paper an operator is required to read -- an Operations Memoran-dum. Unfortunately, the message intended by the memorandum did not affect the practice of the operators, whose practice of re-paatedly running tests and discarding bad tests until a good test was obtained went on unimpeded. The Board finds that Floyd, as the member of on-site management responsible for the memorandum and the overall supervision of the operators, had the duty to 27/ Floyd's reference to 2 gpm is appropriate because rsprogramming the computer to round off meant that leak rates of 1.01-1.49 gpm would round off to 1, an acceptable number, whereas leak rates of 1.50-2.49 gpm would round off to 2, an unacceptable number.

l I

tcke whatever followup actions were necessary to implement the m:morandum, including specific discussions of the subject with chift supervisors. We were unable to determine that any such ac-tions were ever taken.

123. On October 19, after Seelinger told Haverkamp about PORC's determination that a reportable occurrence had occurred end showed him a draft copy of the report being sent to the NRC Rsgion 1 office, Haverkamp discussed the unidentified leakage mntters during a phone conversation with either his supervisor, Richard Keimig, or the backup project inspector for TMI-2,

' William Lazarus. Board Exh. 20, Haverkamp Testimony at 7-8. On October 20, Lazarus, in Haverkamp's absence, prepared an input to the I&E daily report on the Tech Spec violation. Id. Haverkamp bblieved that on the same day, Lazarus also reviewed Licensee's prompt report of occurrence (Grier Letter). Id. (citing Haverkamp Testimony Exh. B, Action Control Form, Report of Nonroutine Events, and Exh. C, Grier Letter).

124. During the week following the October 18 inspection Haverkamp made one or more telephone calls to a person or persons 1

in NRR to determine the acceptability of rounding off. Board Exh. 20, Haverkamp Testimony at 8-9. Haverkamp was informed that ths rounding off of test data was not an allowable practice and conveyed this conclusion by phone to Seelinger, who stated that tha rounding off practice would not be employed. Id. Seelinger tdmitted that in this conversation between Haverkamp and him,

- - - , - -.,-r - -- ---r,

which Seelinger estimated occurred 10 days after the inspection, or approximately on October 28, he "wasn't terribly straight-forward" with Haverkamp about the extent to which the rounding-off practice had been relied on to achieve acceptable results.

Tr. 4772-74 (Seelinger). The record indicates that the rounding-off practice was actually terminated beginning on October 27.

Sre Board Exh. 1-A, Scler Report, Vol. I at 143; compare NRR Test Nos. 15-24 (October 19-26), where all net unidentified leak rate results have been rounded off by the computer, with NRR Test Nor. 26 ff. (October 27 and following), where net unidentified Icak rate results indicate termination of rounding off and return a to the practice prior to the reprogramming of the computer effec-tive October 19 of printing out results to four digits to the right of the decimal point (i.e., ten-thousandths). Seelinger in this conversation with Haverkamp near the end of October had a clear opportunity to reveal the extent of the practice of rounding off.

125. Sometime after the decision by PORC that a reportable occurrence had occurred, Seelinger asked Bezilla and James Stair, the licensing engineer responsible for writing up LERs (Board Exh. 1-A, Stier Report, Vol. VI(K), Stair 3/1/85 Interview at 2),

to assist him in drafting the 14-day follow-up LER. Id.,

Vol. VI(A), Bezilla 2/22/85 Interview at 19. Stair testified that he spent a total of about 1-1/2 hours on work related to the LER, including talking with the cognizant engineer, on whom he l

.rolied for factual information, drafting the original words of

.the LER, and attending the PORC meeting where the LER was re-viewed and edited. Id., Vol. VI(K), Stair 3/1/85 Interview at 11-20; see also Tr. 4641 (Seelinger); Board Exh. 1-A, Stier R port, Vol. VI(A), Bezilla 2/22/85 Interview at 24. PORC met end provided input to the LER on October 31, when a final draft LER was prepared and submitted to typing by Stair.23/ Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 32 (Accountability Checksheet).

2g/ The precise order in which the Narrative to the LER was drafted remains somewhat unclear. Concerning the first phrase circled in the handwritten Narrative (Seelinger, ff. Tr. 4623 at attachment (Handwritten Narrative)) - "At 1000 hrs on October 19, 1978" -- Stair stated that ordinarily the time and date of the discovery of an occurrence that he would use in his rsport would come from the cognizant engineer. Board Exh. 1-A, Stier Report, Vol.'VI(K), Stair 3/1/85 Interview at 17. Stair stated that the handwriting in this circled phrase looked similar

'to his. Id. at 18. If the handwriting was his, Stair would have changed the date based on information from the cognizant engi-nser. Id. Bezilla testified, however, that he possibly wrote in felt tip the incorrect date in the first line of the Narrative.

Id., Vol. VI(A), Bezilla 2/22/85 Interview at 24; see Seelinger, ff. Tr. 4623 at attachment (Handwritten Narrative). Whether Stair or Bezilla wrote this phrase is not of consequence because the phrase repeats the inaccurate time and date provided by Snelinger in the Grier letter.

Stair also stated that PORC at its meetings typically would mnke its own editorial or technical changes, which would be in-corporated into a draft report. Board Exh. 1-A, Stier Report, Vol. VI(K), Stair 3/1/85 Interview at 8-9. Because his practice was not to incorporate a change after PORC approved a report, the report sent to off-site management had to be precisely what PORC approved. Id. at 18-19. Seelinger believed that at the October 31 PORC meeting, he wrote in the changes to the Narrative in response to what PORC wished to edit and revise. Tr. 4644-46 (Seelinger). Although Stair drafted the original words of the Nnrrative, Seelinger wrote most of the dark inserts and changes.

Tr. 4640-44, 4661-64 (Seelinger); see Seelinger, ff. Tr. 4623 at 4 attachment (Handwritten Narrative).

126. Sometime after the October 31 PORC meeting, Seelinger in his capacity as PORC Chairman sent the LER with the one-paragraph Narrative to George Troffer, Manager of Quality Assur-ence,.in Reading.29/ See id. at Tab 28 (Seelinger Memorandum to Troffer); Tr.- 4647-51 (Seelinger). On November 1, 1978, Herbein formally submitted the completed LER with a three-paragraph Nar-rative to Grier at the NRC. See Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 29 (Herbein 11/1/78 Letter to Grier). These two i varsions of the LER Narrative (i.e., the one-paragraph version

-drafted at the site and forwarded to management at Reading after i PORC approval and the three-paragraph version that ultimately cccompanied the LER forwarded to NRC by Herbein) contained sever-i al erroneous or questionable statements in common: (1) the refer-ance to 10:00 a.m. on October 19 when a determination was made that the Tech Specs had been violated; (2) the statement that

. rsduction of unidentified leakage to within allowable limits was accomplished at 7:35 a.m. on October 18; (3) two statements con-corning discovery and correction of errors in inputting data to the computer; and (4) the statement that appropriate personnel would be instructed on the relevant requirements of the Tech Specs and surveillance procedures. The three-paragraph Narrative 4

29/ ~ It is unclear whether the one-paragraph narrative was in hendwritten or typed form at that point in time (Tr. 4647-51 (Seelinger)), but a comparison of the handwritten and typed ver-oions indicates that they are virtually identical. Compare Sselinger, ff. Tr. 4623 at attachment (Handwritten Narrative)

with Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 28 (Typed Narrative).

l

olso contained a significant and totally erroneous phrase explaining how reduction of excessive leakage was accomplished at 7:35 a.m. on October 18. The one-paragraph and the three-paragraph Narratives contain a number of erroneous statements relevant to the Board's assessment of the sufficiency of the cor-rsctive actions taken by management personnel to insure opera-tional compliance with Tech Spec 3.4.6.2. These statements are discussed seriatim.

127. The first phrase of both Narratives repeats the inaccu-rate time and date that appeared in the Grier letter --

10:00 a.m. on October 19 -- when Seelinger stated the determina-tion was made that the Tech Specs had been violated. Although oither Stair or Bezilla may have written this phrase, the source of the phrase was Seelinger. See 11 118-120, supra. The Board finds that Seelinger had sufficient knowledge to know that the time and date of 10:00 a.m. on October 19 were inaccurate in both his letter to Grier and in the Narratives and was in a position to cure the inaccuracy.

128. Again repeating a misleading statement of the Grier letter, the two Narratives state that action was being taken to rcduce the unidentified leakage to within allowable limits and was accomplished at 7:35 a.m. on October 18. Again, the Board finds that Seelinger was fully aware that the action being taken to reduce unidentified leakage was simply to round off, not to identify or correct such leakage. See Tr. 4685-4701 (Seelinger).

Thus the " accomplishment" at 7:35 a.m. of reducing leakage to within allowable limits was merely a paper change with no tangi-ble consequence. We conclude that Seelinger's failure to correct this misleading statement is consistent with his admitted lack of straightforwardness concerning rounding off in his conversation with Haverkamp at the end of October. See 1 124, supra.

129. After the transmittal of the one-paragraph Narrative to Troffer in Reading, the format of the Narrative was changed to three paragraphs, minor changes were made, and, of significance, the phrase "by determining a portion of this to be identified leakage from the Reactor Coolant System and to be well within the limits of Tech. Spec. 3.4.6.2.c" was added to describe, purport-sdly, how reduction of unidentified leakage was accomplished at 7:35 a.m. on October 18. See Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 29 (Herbein 11/1/78 Letter to Grier);

Tr. 4685-88 (Seelinger). This Board has been unable to substan-tiate that this statement about how excessive unidentified leak-age was reduced has any basis in fact. The true reason for the

" reduction" appears to have been rounding off, which was not dis-closed to the NRC, nor apparently to Met-Ed management including Miller and Herbein. Although the Board has been unable to deter-mine the source and reason for this erroneous language, Seelinger testified that it is unlikely that someone.in Reading changed the Narrative without first checking with someone at TMI or without first-hand knowledge of the TMI activities associated with the LER. Tr. 4651-61 (Seelinger).

l

130.13un Narratives also contain two incorrect statements concerning the input of data to the computer: (1) "In addition,

.~ it was discovered that errors in inputting data to computer ccused-indicated unidentified leakage to be greater than actually was occurring" and (2) " Input data for the computer program which calculates unidentified leakage has also been clarified." See Board Exh 1-A, Stier Report, Vol. V(C) at Tab 28 (One-paragraph Narrative). Fels testified that input data had not been clari-

~

fied as of November 1. Tr. 4529 (Fels); see generally Tr.

4512-31 (Fels). In fact, the.only involvement Fels had with cny gsneric computer problem associated with the leak rate test oc-curred sometime between November 9 and November 22, well after the LER had been drafted, approved, and sent to the NRC.

Tr. 4514-17, 4528-29 (Fels). In that period of time, Fels dis-cussed the possibility of program errors and decided to add a note to the program heading stating that a decimal point must be entered with leakage values. Tr. 4515 (Fels); compare NRR Test No. 40 (11/9/78) with NRR Test No. 41 (11/22/78). The only in-volvement prior to November 9 Fels could recall having was limited to a specific test, NRR Test No. 14 conducted on October 18, where Fels may have told the operators involved to rado the calculation using the correct sign because the leakage sign convention looked wrong. Tr. 4520-31 (Fels); see NRR Test No. 14.

131. Seelinger testifed that he partially drafted the first computer-related sentence and that he entirely drafted the second ene. Tr. 4640-44, 4661-4704 (Seelinger). In any event, he obvi-ously approved both sentences at the time of the drafting of the one-paragraph Narrative. Seelinger stated that his practice was to rely on Fels for computer matters. Tr. 4704 (Seelinger).

S2elinger also stated that, having heard the testimony by Fels, ha believed that " Fels and I may have been on a different wave-length" when they discussed a possible program problem, with Fels stating something specific and Seelinger hearing something gener-ic. Tr. 4706-07 (Seelinger). Finally, Seelinger stated that he vaguely recalled that within the few weeks following October 18, hs went to Bezilla, the PORC Secretary, and asked him to request a write-up from Fels on the suppcsed clarification of computer input data, but Bezilla was " unsuccessful" in obtaining such a write-up. Tr. 4706-07 (Seelinger). Seelinger felt "some sense of resistance" but did not understand why and did not pursue the matter. Id. The request for a write-up was apparently withdrawn at Seelinger's request. See Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 30 (TMI-2 PORC Action Items,Section III).

132. Most significantly, the Narrative states that "[t]he appropriate personnel will be instructed on the requirements of the applicable sections of the [ Tech Specs) and the requirements to immediately invoke applicable Action Statements when the pro-1 visions of limiting conditions for operation are not met." See l

1

- .- . = . __

id. at Tab 28 (One-paragraph Harrative). Seelinger substantially contributed to the drafting of this sentence. Tr. 4709 (Seelinger). The sentence at the time it was written was not in- l L

, accurate; rather, it stated that appropriate personnel "will be instructed," a present promise of a future instruction. The

} racord, however, demonstrates.that any subsequent instruction was

. inadequate.30/ See generally Tr. 4708-15 (Seelinger). See 1 135, infra. '

133. Walter J. Marshall, one of two operations engineers as-signed to Floyd, was given the responsibility for implementing and following up on the LER; in particular, the PORC Action Item specified that Marshall was to insure that the Narrative sentence on instruction of appropriate personnel was "decumented by Ops

! rsview" of the LER. See Board Ei -A, Stier Report, Vol. V(C) e at Tab 30 (TMI-2 PORC Action Items,Section III); see also Tr. 4389-96 (Marshall). The PORC Action Item initially was given a due date of November 20 and was signed by Seelinger; however, the " Document Review" form (i.e., " signature sheet") required re-view of the LER by November 10. See Board Exh. 1-A, Stier Raport, Vol. V(C) at Tab 30 (PORC Action Item Sheet and Document

30/ Only one operator, Dennis Olson, testified that he changed

( his practice following the LER. See Tr. 4007-09 (Olson); Olson, ff. Tr. 3911 at 3-4. Olson stated that he was " pretty sure" that l out of the LER came the word not to discard bad tests and not to i

accept negative tests. Tr. 4007-09 (Olson). Even Olson, however,

-nowhere suggests that he received instructions on the require-msnts to enter the Action Statement.

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

I l

Rsview Form). Someone other than Marshall filled in the signa- l ture sheet requiring the 10-day turnaround. Tr. 4396-97 (Mar-chall).

134. Marshall stated that the phrase " documented by Ops re-view" meant "that the document went to the Control Room with the cover sheet and that ths operators had the opportunity to initial and read it." Tr. 4393 (Marshall). Marshall explained that he fulfilled his follow-up responsibility simply by making sure the LER and a signature sheet were placed in'the three-ring notebook in the Control Room with other LERs and by reviewing the binder p2riodically to see if any CROs, shift foremen, and shift super-visors had still not signed the sheet indicating their review.

Tr. 4397-4400, 4408-09, 4414-16 (Marshall); see also Tr. 4713-14, 4664-68 (Seelinger); Tr. 4943-44 (Floyd). If someone had not signed the signature sheet, Marshall "would tell the group to take a look at it and sign it." Tr. 4409 (Marshall). No one has claimed that Marshall had any further responsibilities concerning this Action Item.

135. The signature sheet itself indicates that Marshall was successful in his efforts to make sure the operators had reviewed the LER: all the CROs, shift foremen, and shift supervisors, with the exception of Bryan, whose initials are not correct, signed the signature sheet indicating their review. See Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 30 (Document Review Form); Tr. 4574, 4608-09 (Bryan). Mere placement of the LER in l

t the Control Room binder evidently did not constitute sufficient instruction of appropriate personnel. Operator after operator testified that they never received any clear instruction on the correct interpretation of the Tech Spec following the LER. See, e.g., Tr. 3699-3700 (Hitz); Tr. 3619-20 (A. Miller); Tr. 4455 (Phillippe); Tr. 4573-80 (Bryan); Tr. 2718-19 (Congdon);

Tr. 2967-69 (Zewe); Tr. 3858-59 (Mehler); Tr. 3818-19 (Adams);

Tr. 3244-46 (Mell); Tr. 3083-84 (Illjes); Tr. 3115-16 (Conaway);

Cooper, ff. Tr. 2835 at 6-7; Bryan, ff. Tr. 4540 at 5.

136. In sum, operators had at least three possible ways of rcceiving instruction on the requirements of the Tech Specs and curveillance procedures pertaining to leak rate testing. First, superiors like Seelinger and Floyd who were aware of the need for

) such instruction following Haverkamp's inspection could have di-rectly communicated with the operations Department personnel.

The record is clear, however, that effective direct communica-tions never took place. Second, Floyd's October 20 Operations Memorandum to shift foremen and supervisors was an attempt to en-cure that operators were instructed on proper leak rate practice.

The cryptic language of the critical second paragraph of the Mem-orandum, however, led to the failure of this attempt also. The record indicates no further efforts by Floyd to elucidate the meaning of the Memorandum. Seelinger even speculated that opera-tors may have interpreted Floyd's putting an "X" through his October 20 memo after learning that the rounding off practice

chould be terminated to mean a cancellation of the entire memo, nat only the first paragraph, which dealt with rounding off, and c return to the practice prior to the LER. See Tr. 4680-81 ,

(Seelinger); Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 34 (Floyd 10/20/78 Operations Memorandum). Third, placement of the LER itself in a required reading book for all operators appar-Ently resulted in such a reading, but in virtually no instruction to them or change in practices. The record does not indicate any further meeting or memorandum to fulfill the promise of in-otructing appropriate personnel. The record is clear that opera-tors did not receive adequate instruction by any means on the rcquirements of the Tech Specs and surveillance procedures per-taining to leak rate test practices.

137. The Board concludes that, despite the apparent review of the LER and the October 20 Floyd Operations Memorandum by the operators, the vast majority of the operators following the Hnverkamp inspection did not change their general practice of re-paatedly running tests and discarding bad tests based on misun-darstanding the Tech Specs to require one good test every 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />.

138. Initially scheduled for completion by November 20, 1978, the PORC Action Item had its due date extended to "no later than December 15, 1978" because " plant problems and test program" htd " overshadowed" the item. See Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 30 (TMI-2 PORC Action Item Extension). Despite

r-I the force of the language "no later than December 15, 1978," all cetion relating to this item inexplicably was not completed until March 5, 1979. See id. (PORC Action Items, signature of Ron War-ran, PORC Chairman); see also Tr. 4397, 4404-05 (Marshall).

139. Seelinger testified that he made the following personal offorts to implement the corrective actions described in the Nar-rative: (1) assuring himself that Floyd followed Haverkamp's in-otruction, (2) instructing Marshall by memorandum to inform oper-ntors of the correct interpretation, (3) discussing the matter in ,

o POD meeting and on second shift with a shift supervisor, (4) satisfying himself that appropriate computer changes had been meds, (5) advising the Unit-1 Superintendent O'Hanlon in October 1978 of the NRC's interpretation of the Action Statement rcquirement; and (6) assuring himself after he became Unit-1 Su-p0rintendent that TMI-1 operators would take comparable action.

Scelinger, ff. Tr. 4623 at 8; see Tr. 4669-79 (Seelinger); see niso Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 36 (Seelinger 8/9/84 Letter to Palladino) ("I promptly informed appropriate individuals at both TMI-2 and TMI-l of [the requirement to invoke the Action Statement immediately after performance of an unsatis-fcetory leak rate test]. On two following occasions in TMI-2 within a few days of my discussion with [Haverkamp], I again con-firmed with Operations supervisory personnel that they were aware of [that requirement]."). Without parsing each one of these 1

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

personal efforts for details and merits, the Board finds that, based on the record cited in these findings, none of these per-conal efforts were successful.

140. In sum, the Board finds that the efforts to implement i the corrective actions articulated in the LER Narrative were in-cdequate. The members of management directly responsible for these efforts were Seelinger and Floyd. There is no indication in the record that any members of management above Seelinger and Floyd, namely G. Miller and members of off-site management, had cny direct responsibility for the implementation of any correc-tive actions arising out of the LER. Furthermore, it is unclear that these members of management even understood the problem or were tipped-off by language in the LER. Such individuals had a

, certain amount of overall responsibility in the organizational structure, but not the type of direct responsibility on which we believe the Commission intended for us to base our findings in this proceeding.

l

D. What difficulties, if any, were operators experiencing when conducting leak rate surveillance tests required by Technical Specification 4.4.6.2.d? Who knew about these difficulties? What corrective actions were taken? Did operators feel pressure to obtain leak rate surveillance test results which did not exceed technical specification limits? If so, what type of pressure was perceived or exerted and who was responsible? (CLI-85-18, 22 N.R.C. at 880, Issue (b))

141. The Commission directed this Board to determine (1) whether operators were experiencing any difficulties when con-ducting lenk rate testing; (2) who knew of these difficulties; (3).what corrective actions were taken; and (4) whether operators felt pressure to obtain test results under 1 gpm. 'CLI-85-18, 22 N.R.C. at 880.

142. The surveillance requirements of Tech Spec 4.4.6.2.d required that R2S leakages be demonstrated to be within pre-scribed limits by "[p]erformance of a Reactor Coolant System water inventory balance at least once per 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> during steady state operation." Board Exh. 1-A, Stier Report, Vol. V(B) at Tab 14 (Tech Spec 4.4.6.2).

143. The Tech Specs required performance of the leak rate test only when the plant was at certain parameters for re-cctivity, percentage rated thermal power, and average coolant temperature. These parameters determined the plant mode. When the plant was in Modes 1 through 4 (i.e., power operation, otartup, hot standby, and hot shutdown) the leak rate

curveillance requirement was applicable. The surveillance requirement was not applicable during cold shutdown or refueling.

Id. (Te' 'ec'3.4.6.2 and Table 1.1).

144. TMI-2 first entered Mode 4, which required performance of leak rate tests, on March 11, 1978. The first documented leak l

rate test was performed on March 22, 1978 and may have been the first test actually performed. Id., Vol. I at 67.

145. The causes of the difficulties the operators were experiencing when conducting the leak rate tests generally fall l into one of three categories: (1) method errors, (2) instrument inaccuracies, and (3) oscillations in plant conditions. The dif-ficulties in each of these categories had some potential to cause the test to reflect a result other than actual unidentified leak-cge. In addition to the difficulties the operators were cxperiencing in conducting the leak rate test, there were certain conservatisms associated with the implementation of the leak rate tost. These conservatisms added to the difficulty in obtaining en unidentified leak rate under 1 gpm; however, they were not re-1cted to accuracy.

146. The record indicates that the degree of error caused by occh of the difficulties was not neceasarily cumulative. In some instances the errors cancelled each other out (Tr. 888 (Kirkpatrick)) or the errors were not present because the plant conditions at the time of the test were not such that the error was a factor (Tr. 891 (Russell)).

I

l l

l (i) Method Errors l V

147. The technical experts identified thirteen method errors j l

that could have produced leak rate test results at variance from the true values. The degree of variance would have been depen-dont on the conditions existing at the time each test was per-formed. Board Exh. 1-A, Stier Report, Vol. IV(A) at III.3-III.4; i

Kirkpatrick, ff. Tr. 376 at 18-20 and Attachment 4. MPR stated:

In general, . . . as long as the measured conditions (temperatures, levels, prensure,

, etc.) do not change from beginning to end of tests, there are no additions to or removals from the rector coolant system, and the iden-P tified leakage is small, then these errors all tend toward zero. Essentially these ap-proximations represent inherent limitations in the ability to account perfectly for

changes in conditions which occur during the RCIB test.

f

Board Exh. 1-A, Stier Report, Vol. IV(B) at A.9-A.10.

148. The technical witnesses that appeared before this Board l identified the following " method errors":

Changes in pressure from the beginning to the end of the test were not accounted for in determining the l changes in the density in the RCS and pressurizar. Id.,

J Vol. IV(A) at III.3.

The reactor coolant density was determined using the average of the hot leg and cold leg temperatures. This 4 average temperature was not necessarily the one i

i Y,_,_..~......,...-..

I I

corresponding to the true mean RCS density. Id. at III.4.

Operator-caused additions or removals from the makeup tank,' the reactor coolant drain tank ("RCDT"), or any other part of the system were not converted to volumet-ric units at RCS conditions. Id.

Leakage into the reactor coo). ant drain system was not converted to volumetric units at RCS conditions. Id.

Identified leakage other than that collected in the RCDT was not converted to volumetric units at RCS con-ditions. Id. .

Changes in the RCS volume with pressure were not ac-counted for. Id.

Certain constants used in the calculation of mass changes in the RCS, pressurizer, makeup tank, and RCDT were known with only limited certainty. Id.

The changes in mass in the steam space of the pressur-izer were not accounted for. Id.

The interpolation table used in the computer program to determine RCS density did not correctly calculate den-sity above 582*F. For temperatures above 582*F, the density at 582*F was used. Id.

The level versus volume relationship used to establish the volume collected in the RCDT was slightly incor-rect. .pd.

The volume used for the RCS was in error. Kirkpatrick, ff. Tr. 376 at Attachment 4.

The makeup tank mass change rate was in error. Id.

The pump-downs of the RCDT during a leak rate test would be understaced by amount'of inflow to RCDT during i

pump-downs because both inflow and outflow used the same level indicator. Board Exh. 2, Faegre & Benson Report, Vol. 1 at 23; id., Vol. 2 at 43.

i (ii) Instrument Errors a

149. The " instrument inaccuracies" that affected the leak rete test results fall into two groups - " normal instrument er-rors" and "special sources of instrument error at TMI-2." MPR stated:

The results of the reactor coolant inventory balance surveillance test are affected by the accuracy of the measurements of temperature, j pressure'and level used in the procedure. As with any procedure involving measurements,

. there is a limited certainty to which the "true" values of parameters are known. The uncertainty in the measurements is propagated to an uncertainty in the final result, in

this case the leakage rate. The influence of an error in a single measurement on the final result depends on the magnitude of the error and the " sensitivity coefficient" of the re-sult to that measurement.

B:ard Exh. 1-A, Stier Report, Vol. IV(B) at A.19.

150. MPR set forth in detail eight sources of normal instru-ment error. These sources were nonlinearity, hysteresis, sensi-tivity to effects other than the measured parameters, time-dependent " Aging" of transducer, limited sensitivity of instrument, calibrations uncertainty, time response errors, and instrument application. Based on an evaluation of available data, MPR concluded that the random error in unidentified leak rate in a test as a result of normal instrument error was

+ 0.71 gpm. Id. at A.19-A.22, A.26. NRR witnesses stated their cnalysis would support this estimated error. Tr. 933 (Kirkpatrick).

151. In addition to the normal instrument inaccuracies, two instrument application errors were of particular importance at TMI-2, one affecting the makeup tank level and the other af-fecting the RCDT level. The makeup tank level instrument, under certain conditions, would mistake the actual level change in the makeup tank. The potential source of the problem was hypothe-sized as a collec'-ion of water in the tubing leg of the level instrument that was intended to be dry (a loop seal). The RCDT level measurement error existed because the measured level was not temperature compensated even though it was treated as a true level in the leak rate calculation. Board Exh. 1-A, Stier Report, Vol. IV(B) at A.27-A.33. 9 (iii) Oscillations in Plant Conditions 152. The effects of oscillations in plant conditions were cbserved at TMI-2. These oscillations occurred in the reactor coolant temperature, pressure level, and makeup tank level. The.

oscillations were not in phase but appear to have had a constant phase relationship indicating they were related to a common cource. If the leak rate test procedures were exactly correct, the oscillations in plant conditions would not have affected the leak rate test results. Id. at A.34; id., Vol. IV(A) at III.14.

However, because the procedure contained the previously mentioned method errors, MPR concluded:

At worst, the effect of the oscillation would cause an additional error of the same magni-tude as the instrument errors. On a 'typi-cal' basis.the additional error would be less. Since it is independent of the instru-ment errors, it would combine with them on a random basis, producing only a minor increase on the overall expected error.

Id., Vol. IV(B) at A.36.

(iv) Inherent Conservatism of the TMI-2 Leak Rate Test 153. In addition to the difficulties induced by methodology cnd instrument errors, certain other aspects made it more diffi-cult for operators to obtain a satisfactory unidentified leak rete result. The difficulties in this class reflect conserva-tisms inherent in the TMI-2 leak rate test. They are summarized no follows:

l

The TMI-2 Tech. Specs include the leak rate test without limiting it to a measurement of any particular form of leakage, although Reg. Guide 1.45 makes reference to using the leak rate test for intersystem leakage. The unidentified leakage at TMI-2 included intersystem leakage and while it is believed some intersystem was always present, intersystem leakage was always entered "O.0" on field tests. Id., Vol. I at 40, 52-53.

TMI-2 Tech Specs and surveillance procedures did not incorporate Reg. Guide 1.45 latitude to compare values of unidentified leakage obtained by different detection methods, e.g., sump pump. Reg. Guide 1.45 states moni-toring sump data is a reliable and sensitive way to measure unidentified leakage (Board Exh. 2, Faegre &

Benson Report, Vol. 1 at 29-30), yet the TMI-2 Tech Specs provided no guidance on how to reconcile dif-ferences between detection methods (Board Exh. 1-A, Stier Report, Vol. I at 40). Tech Spec bases would seem to indicate a need for comparison to determine

" effectiveness" of leak rate tests. Board Exh. 2, Faegre & Benson Report, Vol. 2 at 18-21.

Unidentified leakage was based upon a volume rate of one gallon per minute determined at RCS average temper-ature of 581*F rather than reactor building temperature (room temperaturo). At room temperature, unidentified

leakage would be more dense such that 1 gpm of unidentified leakage at TMI-2 was approximately 0.72 gpm as. compared to 1 gpm if room temperature had been used. Board Exh. 1-A, Stier Report, Vol. IV(A) at' III.2; id., Vol. IV(B) at B.5-B.12.

TMI-2 did not have~off-set of evaporative loss in the leak rate test. Board Exh. 2,'Faegre & Benson Report, Vol. 1 at 19; Board Exh. 1-A, Stier. Report, Vol. I at 21.

(v) Persons Who Knew of the Difficulties.

154. The difficulties with the leak rate test that the oper-ators were' experiencing manifested themselves over a long period of time in a complete inability to run two tests back to back, with no power changes or operator-caused changes, and obtain the atme results. The test results were described as "unpredic-tchle," and "quite erratic." Adams, ff. Tr. 3776 at 3; Cooper, ff. Tr. 2835 at 6; Guthrie, ff. Tr. 4113 at 2; Board Exh. 1-A, Stier Report, Vol. VI(F), Hitz 4/24/84 Interview at 31; Hitz, ff.

Tr. 3664 at 4. But see Frederick, ff. Tr. 2447 at 4.

155. Herbein was not aware of the difficulties operators wore experiencing. To a varying extent, Logan, Seelinger, Kunder, and G. Miller had some awareness that difficulties ex-inted with the leak rate test. While Floyd maintained that he htd no recollection of any of these problems, the record contains Eubstantial evidence indicating he would have been aware despite his inability to recall.

l 156. James Floyd, Supervisor of Operations, characterized himself as a " crisis fighter." He allowed his shift supervisors to run the plant on a day to day basis and as long as something had not come to the crisis level, it rarely came to his atten-tion. Tr. 4969 (Floyd). With regard to leak rate testing in particular, Floyd stated that "[als soon as this plant would have bien shut down for a leak rate test I'm sure I would have been ccutely aware of the problem." Tr. 4969 (Floyd). It was his tOstimony, however, that he has "no recollection of knowing about any of these problems" and that he was " basically . . . ignorant of what was going on here [TMI-2]." Tr. 4976 (Floyd).

157. Despite Floyd's recollection, there are a number of in-dicators that Floyd would have been aware of operator problems with leak rate tests. Floyd had a discussion in early October 1978 with Seelinger concerning the interpretation of the leak rate test procedure and was aware of and involved in many of the events of October 18, 1978. See 11 105-140, supra. He ex-pacted difficulties with erratic leak rates during initial opera-tion of the unit. Floyd, ff. Tr. 4894 at 5-6. Fels believed he discussed leak rate problems with Floyd; Chwastyk recalled a Floyd briefing on leak rates after the events on October 18; Haverkamp had a very strong recollection of Floyd being present when he spoke to Seelinger on October 18th; Cooper was aware l

1

I l

f l

Floyd was getting feedback from shift supervisors concerning leak rate problems; and Guthrie was "sure" Floyd knew the operators had problems with the leak rate. Tr. 4509-10 (Fels); Tr. 3502 (Chwastyk); Tr. 2050 (Haverkamp); Tr. 2903 (Cooper); Board Exh. 6, OI Report, Exh. 30, Guthrie Interview at 17.

158. James Seelinger, Unit 2 Superintendent of Technical Support, had an awareness of the difficulties. Seelinger became aware of difficulties in early October 1978 when he learned of the practice of running leak rate tests until one showing less than 1 gpm unidentified leakage was obtained and discarding the tsat results over this 1 gpm criterion. Seelinger's knowledge of 2

the difficulties was substantially increased as a result of the 9 i

LER 78-62 event, in particular, the need to "round off" to obtain '

a result less than 1 gpm unidentified leakage. Both these events and Seelinger's involvement are discussed fully at 11 103-140, supra.

159. Following the elimination of " rounding off" as an aid to achieving a leak rate test result within the limiting condi-tions for operation, Seelinger received further notice of opera- f I

tor difficulties. Seelinger was approached by one or two shift l aupervisors requesting help or relief relative to conducting leak l

rate tests. Seelinger provided none, testifying:

]

A. . . . I did not feel I had any relief relative to tech l spec interpretability. That hac been pretty well finalized by Mr. Haverkamp's visit of October 18th. And I felt like, at the time, I believe I felt like the procedure was a de-rivative, or largely relied on what we had gotten from Unit 1, and I didn't know what to do or how to give proce-dural relief. I didn't go any further with that at that point in time.

Q. You did not spend some hours looking into this at all?

l A. No, sir, I didn't.

l Q. Did you assign anybody to look into it?

l A. No, sir, I didn't.

1

! Tr. 4736-37 (Seelinger).

160. George Kunder followed Seelinger as the Unit 2 Superin-l tcndent of Technical Support in early December, 1978. Although Scelinger was uncertain as to whether or not he informed his suc-consor of the difficulties with the leak rate test (Tr. 7783-84 (Seelinger)), Kunder had some limited knowledge of the difficul-ties with the leak rate tests. He also was aware that some ef-fort was undertaken to correct the difficulties.

161. Sometime in 1979, Kunder became aware that there was some question whether leak rate test results accurately reflected plant conditions. Kunder believed the question was whether the l calculation used to determine leakage was in error so that erro-nsously high readings were being obtained. Kunder was also aware 1

I that his department was requested to look at the leak rate test procedure or calculation to determine if a problem actually ex-isted,.and if so, to resolve it. Kunder, ff. Tr. 4800 at 2-3.

Kunder had only limited recall of how he became aware of the dif-ficulties and no recollection of what was accomplished to correct ,

l l

l j

the difficulties. Tr. 4811-12, 4834-35, 4840-41 (Kunder). Dur-ing this time, Kunder worked considerable overtime and had a very

! h avy work load, " attempting to not only take over the reins of l

. . . managing the engineering organization but also take care of the collateral duties of PORC chairmanship, beginning some prepa-1 rations for the refueling outages which [he] was responsible to coordinate the planning for, and . . . in particular, deal with a myriad of issues and problems that existed at the time to attempt to get (himself] up to speed on the details of the Unit 2 sys-

. tcms; that is, to prepare for senior reactor operators license i

. . . ." Tr. 4803 (Kunder). We believe Kunder's inability to rscall detail with respect to his awareness of difficulties re-flects his actual limited knowledge of these difficulties during the period of operation of TMI-2.

162. Joseph Logan, Unit 2 Superintendent, also had some knowledge of the difficulties operators were experiencing. Logan thought the problem was with the computer because the leak rate l

test results were inconsistent. Although he does not know when ha became aware of the difficulties, Logan does recall having had discussions with shift supervisors, Kunder, and Floyd. Tr. 5117, 5123-24 (Logan). Logan considered negative leak rate test re-sults as an indication that there were computer program errors associated with this test. Tr. 5143, 5145 (Logan). However, Logan felt that the problem was recognized and the work being done led him to believe that a solution to correct the problems

_ _ _ - - _ .~ __ _ ____ _ ~ . _ _ _ _ _ . _ . _ - _ _ . . , _ _ _ _ _ _ _ . . _ _ _ _ . _ . _ .

would be found.' Tr. 5119, 5133 (Logan). Logan does recall that

~

a change was made, but has no recollection if it was effective.

Tr. 5134 (Logan).

163. Gary Miller, Station Manager, was aware to a limited extent of the difficulties the operators were experiencing.

Miller was routinely expo ed to several sources of information containing data that, if analyzed (such as comparing leak rate

. test results day after day) might have raised a question in his mind to inquire further. See generally Board Exh. 1-A, Stier 1 Report, Vol II(B), G. Miller Summary at 4-11. Miller, however, did not carry out such a comparison: "In short, I simply did not realize that such repetitive results were being recorded."

, G. Miller, ff. Tr. 5039 at 19. The primary source Miller depended on to raise operational problems was the morning confer-cnce call. Miller did not believe that the subject of continuing leak rate surveillance difficulties was ever discussed during these calls nor does he believe that anyone ever advised him

! cbout such difficulties on any other occasion. Id. The record supports this statement.

164. The extent of Miller's knowledge probably comes from Sselinger's conversation with Miller on October 18, 1978, con-l corning the violation of the Tech Spec discussed at 1 114, supra.

While Miller has no current recollection of their conversation, ho believes it well could have taken place.31/

31/ One other possibility of notice to Miller was considered.

Stier analyzed an alleged overheard phone call by Joseph Chwastyk (Continued Next Page)

165. By Seelinger's account of the conversation, Miller probably became aware that the operators had experienced some difficulty with the leak rate test at that time. Tr. 4723-27

'(Seelinger). Miller also was aware that an LER was issued (Tr.' 5056 (G. Miller)), but for the reasons previously discussed on the accuracy and completeness of the LER (see 11 126-132, rupra), it is doubtful that the LER would have expanded Miller's cwareness. More likely the LER would have suggested to Miller that there was a one-time problem (not a consistent pattern of procedural abuse) and given Miller confidence that the matter was being resolved. Tr. 5085-86 (G. Miller).

166. The degree of Miller's awareness was best summed up by Miller himself:

I do not believe, however, that I was aware during the time period in question that the control room operators in Unit 2 were having significant difficulty achieving the accept-able leak rate surveillance test results re-quired by Unit 2 Technical Specification 4.4.6.2.d. I knew from my experience in the development of the Unit 2 Technical Specifi-cations that surveillance generally would be (Continued) botween Seelinger and another person (s) (possibly G. Miller end/or Herbein) that could have provided notice of operator dif-ficulties with the leak rate test; Stier was unable to confirm l that call. Board Exh. 1-A, Stier Report, Vol. I at 134-35.

! Based on the information developed in this prrceeding (which was not available to Stier), in particular the tertimony of Chwastyk (Tr. 3509-13 ), we believe the evidentiary value of such a phone conversation is too tenuous to make any findings with respect to G. Miller or anyone else.

]

.more difficult for Unit 2 than for Unit 1, in terms of the number and frequency of surveil- l lance requirements and the nature of those  :

requirements. As part of this, I knew that l acceptable leak rate test.results would be more difficult to achieve under the Unit 2 l Technical Specifications than they had been

for Unit 1; the Unit 1~ Technical Specifica-tion permitted us to take credit in designing l our leak rate test procedure for a .5 gpm l evaporative loss, but the Unit 2 Technical Specification did not. I thought, however,

that acceptable leak rates could be consis-

! tently obtained under the Unit 2 Technical Specification, and I was never advised prior to March 28, 1979 that that was not the case.

I G. Miller, ff. Tr. 5039 at 17-18.

167. John Herbein,-Vice President of Generation, had no r

knowledge of the difficulties. The strongest potential source of information that might have alerted Herbein of the difficulties was the Licensee Event Report 78-62/IT.32/

168. On November 1, 1978, Herbein signed the transmittal letter sending the LER to the NRC. Herbein, however, has no in-1 i dependent recollections of the LER or surrounding events.

I Herbein, ff. Tr. 5268 at 10-11. The circumstances surrounding i the preparation of the LER and Herbein's signing of the transmit-tal letter are covered at 11 125-132, supra.

i 169. The LER failed to identify that the sources of the i

problem were repetitive test results exceeding acceptance i

32/ See n.31, supra.

-e5-l i

criteria, and the Operations Department's lack of confidence in leak rate test results. The LER also failed to point out that l

the solution to the problem had been the decision to round off test results. Board Exh. 1-A, Stier Report, Vol. I at 143.

Given these inaccuracies and inadequacies in the LER, the Board j cannot impute knowledge of the difficulties operators were cxperiencing to Herbein. In fact, to the extent Herbein would have had any question of the existence and subsequent resolution of operators' difficulties, his questions might have disappeared the following January when the NRC notified him that the LER had been selected for onsite follow-up. The notification stated that "the inspector verified that the reporting requirements . . . had b:en met, that appropriate corrective action has been taken, that the event was reviewed by the licensee as required by Tech Specs, cnd that continued operation of the facility was conducted in conformance with Tech Spec limits." Board Exh. 20, Haverkamp Tcatimony, Enclosure to Exhibit E at 10-11.

(vi) Corrective Actions

! 170. Despite the fact that there were many difficulties as-

cociated with the leak rate test, only limited corrective actions i

wore taken to cure the fundamental deficiencies that caused the i

difficulties. Two such corrections are discussed below. That

! more corrections were not made or attention paid to the host of d3ficiencies with the leak rate test was explored by the Board with a number of witnesses. One explanation is that operators l l i

l l

l felt it was being cured. Chwastyk, ff. Tr. 3407 at 3; Faust, ff.

Tr. 2511 at 3; Frederick, ff. Tr. 2447 at 4. Another was that the plant was still new. Adams, ff. Tr. 3776 at 2-3. Virtually no one seemed overly concerned with the problems.

171. A major attempt to correct leak rate test problems was prompted when in February 1979, the collection rate of the RCDT b gan to increase. This high rate of collection had not been ex-parienced before at TMI-2 fer any extended period of time. Board Exh. I-A, Stier Report, Vol. I at p. 91. The leak rate test cal-culation failed to convert RCDT collection from room temperature to RCS temperature before it was subtracted from gross leakage.

A volume of a given mass of water is 1.4 times greater at reactor ccolant temperature than it is at room temperature. Id.

172. Because of this error, the calculated unidentified leak rate was overstated by an amount equal to 40 percent of the RCDT collection rate. Thus, when RCDT collection reached a rate of 2.5 gpm at room temperature, the unidentified leak rate calculat-cd at RCS temperature would be greater than the actual unidentified leakage by 1.0 gpm, even if the actual unidentified Icak rate had not changed. This rate of drain tank collection (2.5 gpm) was reached around February 25. Id. at 92.

173. Having recognized that RCDT collection was driving up the unidentified leak rate measurements, on March 16 a procedural change was made to correct the calculation error that was causing

RCDT collection to be understated. Thereafter, each test calcu-lation was supposed to be corrected for this error by hand. This change, however, did not produce the desired result. Operators continued to experience difficulty in obtaining satisfactory leak rate test results. The apparent reason is that an inaccuracy in the makeup tank level transmitter continued to cause a 20 to 50 percent exaggeration in the drop in makeup tank level. There-fore, if RCDT collection were 4.0 gpm, the error in makeup tank icvel measurement would overstate the calculated unidentified leak rate by 1.1 to 2.8 gpm. Id. at 100.

174. The principal effect of the problem with the makeup tcnk level instrument was that level changes in the makeup tank wculd be significantly overstated. The hypothesized source of the problem was related to the collection of water in the tubing icg of the level instrument which was intended to be dry (a loop scal). Id., Vol. IV(B) at A.27-A.28. This Board need not deter-mine that a loop seal existed. Kirkpatrick testified that the NRC at times believed it did exist. See Tr. 994 (Kirkpatrick).

R:gardless of the cause, there were times at TMI-2, when the makeup tank pressure changed when gas was added to or removed from the makeup tank or when the tank level changed (thereby changing the volume available for gas), that the makeup tank l

icvel instrumentation overstated the actual level change. Tr.

i 963-965, 997 (Stier); Board Exh. 1-A, Stier Report, Vol. IV(B) at A.29.

i i

175. The second, but largely insignificant, attempted cure to leak rate test difficulties occurred in the Fall of 1978. On October 31, 1978, the Instrument and Control Department reported o problem. The instruments for both the wet reference leg and the dry variable leg of the makeup tank were connected to a com-mon set of sensing lines. This arrangement made it impossible to perform maintenance on one instrument without affecting the cther. By November 9, 1978, installation was completed of sepa-( rate sensing lines with a common penetration on the makeup tank.

B:ard Exh. 2, Faegre & Benson Report, Vol. 2 at 65; Board Exh.

1-A, Stier Report, Vol. V(D) at Tab 53 (Field Questionnaire).

There was no apparent overall benefit from this change with rcspect to the subsequent accuracy of the makeup tank level men-curement.

(vii) Pressure to Obtain Leak Rate Test Results Which Did Not Exceed Technical Specification Limits 176. This Board examined the issue of whether operators felt pressure to obtain leak rate test results that did not exceed the Tcch Spec limit.

177. CROs felt pressured by shift supervisors and shift foremen to obtain test results under 1.0 gpm. Shift foremen felt aimilar pressure from their shift supervisors. The evidence does not indicate that shift supervisors experienced similar pressure from their superiors. Board Exh. 1-A, Stier Report, Vol. I at 124-25.

1 l

178. The pressure felt by the CRos was depicted as a general i conse to keep the plant on line (Id., Vol. VI(G), Illjes 2/7/85 Interview at 57-58); being asked questions about the status of  ;

the leak rate test (Board Exh. 6, OI Report, Exh. 10, Cooper In-torview at 24; Chwastyk, ff. Tr. 3407 at 6); and being told to get a good leak rate (Coleman, ff. Tr. 2579 at 2-3; Tr. 2586 (Coleman); Booher, ff. Tr. 4175 at 3; Board Exh. 1-A, Stier R: port, Vol. VI(B), Congdon 4/10/80 Interview at 2). Despite the fact that some CRos felt pressure, that pressure did not trans-late into a sense that adverse action would be taken against them if they failed to obtain a good leak rate test result (Booher, ff. Tr. 4175 at 3; Board Exh. 6, OI Report, Exh. 18, Wright In- i torview at 109-10), even if such failure meant shutting down the plant (Board Exh. 1-A, Stier Report, Vol. (VI)(G), Illjes 2/7/85 Interview at 57-58).

179. Some shift foremen felt pressure to keep the plant on line as much as possible. Board Exh. 1-A, Stier Report, Vol.

VI(B), Conaway 2/21/85 Interview at 37-38; Board Exh. 6, OI l R: port, Exh. 30, Guthrie Interview at 45-46. Again, we perceived no sense that such pressure translated into any reluctance to chut down if that was seen as necessary for any reason by the chift foreman, f

-100-1

E. Were unacceptable leak rate surveillance test results required by Technical Specification 4.4.6.2.d discarded? If so, who knew of, condoned or directed this practice? Were unacceptable leak rate surveillance test results discarded in an attempt to hide them from the NRC?

(CLI-85-18, 22 N.R.C. at 880, Issues (c))

180. The third issue specified by the Commission in this proceeding concerns whether unacceptable leak rate surveillance test results required by Tech Spec 4.4.6.2.d were discarded. If the Board found that records were discarded, it was supposed to determine who condoned or directed this practice and whether there was an attempt to hide unacceptable leak rate surveillance tost results from the NRC. CLI-85-18, 22 N.R.C. at 880. In con-n:ction with this issue of documentation, we also explored the cperators' compliance with related paperwork requirements, including logging and filing of E&Ds.

181. The administrative procedures applicable to the docu-m:ntation of leak rate testing are discussed at 11 81, 91-94, I rupra. The plant records at TMI-2 show that the only leak rate tcsts ever " filed" were those showing unidentified leakage under 1 gpm. Board Exh. 1-A, Stier Report, Vol. I at 60; id.,

Vol. III(A) at Table 1. Stier estimated that at least 50 percent of all the tests performed were discarded because the results chowed unidentified leakage exceeding 1 gpm. Id., Vol. I at 60.

MPR Associates performed two statistical analyses of recorded tosts. Id. First, they determined that the variability of the frequency with which tests were filed was high, indicating that

-101-

tests.were filed at irregular intervals and raising the possibil-ity that significant numbers were discarded. Id. Second, MPR plotted test results as histograms, and from the apparent truncation of the resulting distributions, estimated that about half or more of the expected results would have exceeded 1 gpm l

but were not filed. Id.; id., Vol. IV(A) at V.2-V.8. Testimony cbtained from TMI-2 personnel indicated that as many as two or three leak rate tests may have been discarded for every one that was filed. Id., Vol. I at 61 (citing the Interviews of Smith, 2/8/85 at 70; McGovern, 2/6/85 at 15; Illjes, 2/7/85 at 10-11, 114; Faust, 2/19/85 at 138; Adams 3/8/85 at 81).

182. The practice of discarding leak rate test results greater than 1 gpm began at TMI-1 and carried over to TMI-2. Tr.

4325 (Smith); Floyd, ff. Tr. 4894 at 2; Board Exh. 1-A, Stier Report, Vol. I at 58. Every CRO, shift foreman, and shift super-vicor who appeared before the Board, except Frederick,33/

testified that he was either aware of the practice of discr.rding tests or personally discarded tests. Tr. 2250 (Hartman); Tr.

2544 (Faust); Coleman, ff. Tr. 2579 at 2; Tr. 2673 (Wright);

Congdon, ff. Tr. 2709 at 4; Tr. 2795 (Scheimann); Cooper, ff. Tr.

2835 at 4; Tr. 2958 (Zewe); Illjes, ff. Tr. 3010 at 2; Tr. 3110 (Conaway); Tr. 3204 (McGovern); Tr. 3243 (Mell); Kidwell, ff. Tr.

33/ Frederick testified that he gave the leak rate tests to his chift foreman and was unaware that the tests were being dis-carded. Frederick, ff. Tr. 2447 at 6.

-102-

f 3285 at 3; Tr. 3487 (Chwastyk); A. Miller, ff. Tr. 3608 at 3; Hitz, ff. Tr. at 3; Adams, ff. Tr. 3776 at 2; Mehler, ff. Tr.

3842 at 5; Tr. 4007-08 (Olson); Hemmila, ff. Tr. 4039 at 3; Tr.

4115 (Guthrie); Booher, ff. Tr. 4175 at 2; Hoyt, ff. Tr. 4233 at 3; Tr. 4325 (Smith); Phillippe, ff. Tr. 4432 at 2-3; Bryan, ff.

Tr. 4540 at 2.

183. Descriptions of the practice of discarding tests dif-fered somewhat among the witnesses. Some witnesses testified that CRos or auxiliary operators were dispatched to search for leaks before a test was invalidated and discarded. Tr. 2250 t

(Hartman); Tr. 3109-10 (Conaway); Booher, ff. Tr. 4175 at 2; Sr.ith, ff. Tr. 4331 at 4. Other witnesses testified that they compared the leak rate test to plant parameters. If they found that the leak rate test results were inconsistent with the plant parameters, they discarded the leak rate test results greater than 1 gpm. Tr. 2817 (Scheimann); Zewe, ff. Tr. 2946 at 3. Some cperar. ors testified that results over 1 gpm were given to the chift foreman to discard. Faust, ff. Tr. 2511 at 3; Congdon, ff.

Tr. 2709 at 4. Most operators, shift foremen, and shift supervi-cars testified that leak rate test results above 1 gpm were re-toined until a leak rate test result below 1 gpm was obtained; cnd then the tests above 1 gpm were discarded. Coleman, ff. Tr.

2579 at 2;34/ Congdon, ff. Tr. 2709 at 4; Cooper, ff. Tr. 2835 at 34/ Coleman testified that on one occasion he obtained a leak rote test in excess of 1 gpm, signed the test, put it on his (Continued Next Page)

-103-I

4; Illjes, ff. Tr. 3010 at 2; Tr. 3243 (Mell); Kidwell, ff. Tr.

3285 at 3; Tr. 3487 (Chwastyk); A. Miller, ff. Tr. 3608 at 3; Adams, ff. Tr. 3776 at 2.

184. James Floyd, the Unit 2 Supervisor of Operations, was cware that invalid leak rate tests were being discarded. Floyd, ff. Tr. 4894 at 2. Floyd maintained that he never directed that leak rate tests be discarded, but the practice of discarding tests carried over from TMI-1. Id. Floyd was also of the "opin-ion that blatantly bad leak rates (for example, excessively nega-tive ones) had no connection with reality; conseqpently, it was permissible to discard them." Id. at 2-3.

185. In early October of 1978, James Seelinger, the Unit 2 Superintendent of Technical Support, became aware of the practice of discarding valid tests above 1 gpm and running another test.

Tr. 4745 (Seelinger). The extent of Seelinger's knowledge of the practice of discarding tests, and the actions he took when he discovered the practice are discussed at 11 103-105, 108-113, supra. In sum, Seelinger testified that after a meeting with Floyd in early October, he (Seelinger) agreed that the Tech Specs required one leak rate test under 1 gpm every 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> and entry l

(Continued) shift foreman's desk, and a short time later, three people came cut of the shift supervisor's office and one told him that they did not want to see leak rates that exceeded the Tech Specs.

After that incident, Coleman began throwing away leak rate test results over 1 gpm. Coleman, ff. Tr. 2519 at 3.

1

-104-

into the Action Statement only at the end of such a 72-hour period. Tr. 4764-65 (Seelinger). Because Seelinger "did not particularly care for the interpretation" and he "was not partic-ularly crazy on having the NRC involved with that interpretation J cr finding out about that interpretation," he subsequently cdvised one or more shift supervisors "that the tests were not to be left out and lying around," where the NRC might find them.

Tr. 4756 (Seelinger). After Haverkamp's visit, Seelinger stated that he interpreted the Tech Specs to require immediate Action Statement entry on obtaining a test result greater than 1 gpm.

Tr. 4758-59 (Seelinger). Such an interpretation should have ended the practice of discarding tests. Id. Seelinger acknowl-cdged, however, that he never effectively put the word out to the chift supervisors to adept the new interpretation of the Tech Specs. Id. Therefore, the practice of discarding leak rate tests apparently continued up to the accident at TMI-2.

186. Operators did not log the starting time of the leak rate test. See, e.g., Tr. 2496 (Frederick); Congdon, ff. Tr.

2709 at 3; McGovern, ff. Tr. 3148 at 3; Hitz, ff. Tr. 3664 at 5; Tr. 4116 (Guthrie); Bryan, ff. Tr. 4540 at 3. Some operators did log the completion time of leak rate test results below 1 gpm.

See, e.g., Chwastyk, ff. Tr. 3407 at 3; Smith, ff. Tr. 4331 at 3.

187. No E&Ds were ever filed with any of the leak rate test results at TMI-2. Tr. 2268 (Hartman); Board Exh. 1-A, Stier Report, Vol. I at 60. Most witnesses testified that they just

-105-

did not think about using E&Ds with the leak rate test or had no i cxplanation for why E&Ds were not used with the leak rate test.

Tr. 2452 (Frederick); Tr. 2671 (Wright); Congdon, ff. Tr. 2709 at 3; Tr. 2794 (Scheimann); Tr. 2988-89 (Zewe); Tr. 3668-69 (Hitz);

t

( Mehler, ff. Tr. 3842 at 4; Tr. 4078 (Hemmila); Tr. 4742 (Seelinger). Some witnesses testified they thought that the E&D procedure may have been inapplicable to the leak rate test be-l cause the test was conducted on the computer. Tr. 2911 (Cooper);

Tr. 3018 (Illjes); McGovern, ff. Tr. 3148 at 4; Tr. 3524 (Chwastyk); Tr. 4346 (Smith); Bryan, ff. Tr. 4540 at 5. Finally, come witnesses claimed that the E&D procedure did not apply to the leak rate test because the test was a routine surveillance.

Tr. 2545 (Faust); Tr. 3018 (Illjes); Floyd, ff. Tr. 4894 at 3. .

188. Based on the evidence in the record, it is clear that TMI-2 personnel uniformly failed to document leak rate testing properly. Valid leak rate tests were routinely discarded by CRos cnd shift foremen. Shift supervisors were aware of this prac-tice. Floyd was aware that operators were discarding invalid tests. Further, James Seelinger, at one point, told TMI-2 per-connel not to leave leak rate tests lying around. This instruc-tion may have led at least some TMI-2 personnel to discard leak rate tests so that the NRC would not discover them.

I 189. In addition, TMI-2 personnel failed to log the start times of leak rate tests and failed to file E&Ds. Board Exh. 1-A, Stier Report, Vol. I at 60. Floyd failed to enforce

-106-i

the application of the E&D procedure to the leak rate test.

Floyd, ff. Tr. 4894 at 3. He claimed that E&Ds were not filed because the leak rate test was run more frequently than required by the Tech Specs. Id.

190. No members of management above Floyd and Seelinger ap-parently knew that tests were being discarded where unidentified l leak rate results exceeded 1 gpm. Floyd was the most senior per-con who admitted E&Ds were not filed when leak rate tests failed to meet acceptance criteria. See Board Exh. 1-A, Stier Report, Vol. I at 133; Tr. 4745 (Seelinger); Floyd, ff. Tr. 4894 at 3.

By their very nature, the practices of discarding tests and failing to follow procedures for documenting test results tended to conceal such conduct. Board Exh. 1-A, Stier Report, Vol. I at 135. Although the CROs and shift foremen openly followed these practices, the evidence does not show that any management offi-cials participated in their activities, and, other than Floyd and Seelinger, observed their conduct.

191. Both Kunder and Logan testified that they did not know that operators were discarding leak rate tests. Tr. 4839 (Kunder); Tr. 5138 (Logan). G. Miller also had no knowledge op-crators were discarding leak rate tests. As Station Superinten-dent, G. Miller had no direct involvement in leak rate testing.

Miller testified that he depended primarily on the morning con-ference call as a means of identifying operational problems.

G. Miller, ff. Tr. 5039 at 19. Miller did not believe that the

-107-

cubject of continuing leak rate surveillance difficulties was cver discussed during those calls. Id. It was his belief that curveillance records were retained and collected by the Surveil-lance Coordinator. Board Exh. 1-A, Stier Report, Vol. II(B),

G. Miller Summary at 1. Herbein was informed of the daily status of the plant through subordinates. Herbein, ff. Tr. 5268 at 7.

Herbein's subordinates never informed him that CRos or others l

were discarding or failing to document unsatisfactory leak rate j tests. Id. at 13. Further, since appropriate procedures gov-orning testing, logging, and E&Ds were in place (see 11 74-94, rupra), there exists, absent some exhibited knowledge on manage-ment's part of the actual practices, no cause to fault management cbove Floyd and Seelinger for dereliction or culpable neglect.

F. Did operators manipulate data or take other actions during leak rate surveillance testing in an attempt to improperly influence test results? Who performed, condoned, directed or was knowledgeable of data manipulation or other improper actions during leak rate surveillance testing? (CLI-85-18, 22 N.R.C. at 880, Issue (d))

192. The final specific issue that the Commission directed the Board to address involves the manipulation of data and other improper actions during leak rate surveillance testing:

Did operators manipulate data or take other actions during leak rate surveillance testing in an attempt to improperly influence test results? Who performed, condoned, directed or was knowledgeable of data manipulation or other improper actions during leak rate sur-veillance testing? This would include, but is not limited to the following:

-108-

r (i) inputting the wrong data into the plant computer; (ii) adding hydrogen gas to the make-up

tank during the test in an attempt to influ-ence make-up tank level indication; (iii) adding water to the make-up tank during the test and either not including the addition in the computer calculation or i underrecording the addition in the computer; 1

(iv) taking advantage of differences or l- inaccuracies in plant instrumentation (e.g.,

i make-up tank level indicators) in an attempt

} to influence parameters critical to the leak j rate surveillance test calculation; (v) taking or failing to take any ac-i tion in violation of technical specification

requirements.

j CLI-85-18, 22 N.R.C. at 880-881. The Commission gave the Board

{ the further responsibility of issuing a recommended decision j 'cetting forth its findings "on who participated in, had knowledge i of and condoned, or by their dereliction or culpable neglect al-1 i lowed the leak rate falsifications, and the facts surrounding any j cuch involvement in sufficient detail to determine the in-

! volvement of any individual who may now work, or in the future 1

j work, at a nuclear facility." Id. at 883.

j 193. We understand these two statements by the Commission to i

j b3 requesting findings on the extent to which any individual per-formed, condoned, directed, knew of, or by his dereliction or culpable neglect allowed falsification, manipulation, or other cuch improper actions during leak rate surveillance testing. In i

the area of management (see n.lO, supra), the Board has

]

-109-I

l cpecifically examined the involvement in leak rate testing of Floyd, Seelinger, Kunder, Logan, G. Miller, Lawyer, and Herbein.

! W3 heard direct testimony from all of these members of management cxcept Lawyer, whose prior statement in the record we determined to be sufficient for purposes of this proceeding. See Tr. 5341 (Kelley, J.).

194. In sum, the Board finds that the record in this pro-creding demonstrates that no member of management from, and including, the level of Supervisor of Operations and Superinten-d:nt of Technical Support and above, performed, condoned, directed, or knew of falsification, manipulation, or other im-proper actions during leak rate testing. The Board does find, h wever, that Floyd and Seelinger, "by their dereliction or cul-pable neglect," allowed such improper actions.

195. In attempting to determine the " dereliction or culpable noglect" by members of management in allowing improper leak rate cctions, the Board has not applied a standard whereby a manageri-ol position alone carries with it a basis for finding fault in this proceeding. We do not interpret the Commission order as rcquiring a standard of vicarious or imputed responsibility, but rcther one of negligence that considers the particular tasks and duties of a position and what the individual knew or reasonably chould have known. Applying such a standard in our scrutiny of both on-site and off-site management, we have found that only two 1

individuals, Floyd and Seelinger, bear responsibility. Before

-110-l

evaluating the conduct of these two individuals, we provide the bases for our conclusion that higher members of management, name-ly Kunder, Logan, Miller, Lawyer, and Herbein neither knew of nor cllowed, by dereliction or culpable neglect, any of the improper cctions identified by the Commission.

196. George Kunder replaced Seelinger as Unit 2 Superinten-dent of Technical Support in December 1978. See 1 46, supra. In that capacity, Kunder assumed a variety of responsibilities de-ceribed in 11 47, 49, supra. Despite assuming Seelinger's posi-tion, however, the record is clear that Kunder, unlike 5eelinger, had no knowledge of improper leak rate test practices. Kunder was not licensed on TMI-2 and had no direct role in the operation of the Unit and no responsibility for operational decisions.

Kunder, ff. Tr. 4800 at 1-2. He viewed his relationship to the Operations Department solely as advisory. Id. at 2. His contact

! with Operations personnel was limited to occasional meetings and intermittent conversations when specific engineering problems were brought to his attention. Id. He usually learned about op-crational problems through members of his staff who performed the engineering tasks assigned to his Department. Id.

197. During the time that Kunder was assigned to TMI-2, he never ran, signed, or reviewed a leak rate test. Id. He had no recollection of ever being present in the Control Room while an cperator was running a test. Id. Kunder testified that he never knew about the operator practice of discarding leak rates greater

-111-

than 1 gpm. Tr. 4839-40 (Kunder); see 1 191, supra. Kunder stated that he had "no role in or responsibility for performing or recording leak rates" - "[t] hat was the job of operations personnel." Kunder, ff. Tr. 4800 at 2. Sometime in 1979 Kunder did become aware that "there was some question whether leak rate test results accurately reflected plant conditions," but the Board found in 1 161, supra, that Kunder's inability to recall detail concerning his awareness of difficulties reflected his ac-tual limited knowledge of these difficulties.

198. Based on this testimony by Kunder, which the record supports, the Board finds that Kunder neither knew of nor al-lowed, by dereliction or culpable neglect, any of the improper actions identified by the Commission.

199. The Board has found that Joseph Logan, Unit 2 Superin-tendent, had some knowledge of difficulties operators were experiencing with leak rate tests, but he felt the problem was recognized and.a solution would be found. See 1 162, supra.

Logan viewed the leak rate test as an administrative mechanical paperwork requirement and relied on other cenfirmatory indica-tions to reveal excessive leak rates. Tr. 5136-37 (Logan).

Logan nevertheless stated he felt at the time that "we needed to pay attention" to leak rate tests greater than 1 gpm and, in the event of a consistent excessive leak rate, "we would have done something about it." Tr. 5137-38 (Logan). Logan testified he was not aware that operators were discarding tests greater than 1

-112-

gpm. Tr. 5138-39 (Logan); see 1 191, supra. The record supports Logan's testimony. We find that Logan's limited knowledge of leak rate difficulties, absence of knowledge of the practice of discarding tests, and general view toward the tests are consis-I tent with his position as Unit 2 Superintendent, which involved oversight from a distance, not direct supervision, of the testing -

program. See 11 43-45, supra. The Board thus finds that Logan neither knew of, nor allowed, by dereliction or culpable neglect, -

any of the improper actions identified by the Commission.

200. The Board has found that Gary Miller was aware to a limited extent of difficulties operators were experiencing with leak rate tests but that overall he thought operators, despite the difficulties, could consistently obtain acceptable leak rates and was never advised otherwise. See 11 163-166, supra. Miller did not know that operators were discarding leak rate tests greater than 1 gpm. G. Miller, ff. Tr. 5039 at 20; see 1 191, supra. Miller stated, "I did not know of any manipulation of data or other improper efforts to influence leak rate test re-suits at Unit 2. I can say categorically that I would not have condoned such conduct." G. Miller, ff. Tr. 5039 at 20. Miller, in his position as the highest level of management at the facility, clearly did not have direct responsibility for, and any significant knowledge of, the leak rate practices and problems.

See 11 36-42, supra. The record supports Miller's testimony.

The Board therefore finds that G. Miller neither knew of, nor

-113-

allowed, by dereliction or culpable neglect, any of the improper operator actions identified by the Commission.

201. In evaluating the conduct and knowledge of Lawrence Lawyer, the Board relied on his only interview in the record, an extensive one conducted on November 10, 1983, by OI.

See Board Exh. 1-A, Stier Report, Vol. VI(G), Lawyer 11/10/83 In-terview. The interview does not reflect any knowledge by Lawyer of leak rate practices or problems. See Tr. 5341 (Kelley, J.).

None of the parties to the proceeding called for Lawyer's appear-ance as a witness. See Tr. 5329-33 (Kelley, J.). G. Miller, who reported to Lawyer, stated that he never had any discussions with Lawyer pertaining to leak rate tests and felt that Lawyer could not add to the Board's knowledge of the subject matter. Tr. 5076 (G. Miller). We thus find that Lawyer neither knew of, nor al-lowed, by dereliction or culpable neglect, any of the improper operator actions identified by the Commission.

202. John Herbein was even more removed than G. Miller from direct responsibility for the operation of the leak rate testing program. As Vice President of Generation for Met-Ed, stationed in Reading, Herbein was responsible for the overall management of TMI-1 and TMI-2, two multi-unit coal-fired stations, one hydro-electric station, and fourteen combustion turbines, with a com-bined capacity of 2,576 megawatts. Herbein, ff. Tr. 5628 at 4; see 1 31, supra. The nature of Herbein's position required him to rely on the chain of command and formal review committees to

-114-

bring issues needing his review and resolution to his attention.

Herbein, ff. Tr. 5268 at 5-6; see 11 31-33, supra.

203. Herbein testified that he was not aware of, did not participate in, and did not consciously tolerate improprieties associated with leak rate testing at TMI-2. Herbein, ff.

Tr. 5268 at 13. Although Herbein signed the transmittal letter sending the LER to the NRC and therefore was aware of a failure to respond to a leak rate test outside the acceptable limits (Herbein, ff. Tr. 5268 at 13), we have found that, given th(

inaccuracies and inadequacies in the LER, the Board cannot impute to Herbein knowledge of the difficulties operators were experiencing. See 11 167-169, supra. Herbein stated that he was never made aware of the following ongoing problems: (1) operators having difficulty obtaining leak rate test results within accept-able limits; (2) operators documenting test results within acceptable limits without regard to the validity of the results; (3) operators or others discarding or failing to document tests greater than 1 gpm; (4) operators manipulating test results; and (5) Operations Department personnel not taking appropriate action after receiving test results greater than 1 gpm. Id. Herbein stated that he had no reason to believe these problems were oc-curring because his managers, the formal review committees, and the NRC never brought them to his attention. Id. at 13-14.

204. The record supports Herbein's testimony. We therefore find that Herbein neither knew of, nor allowed, by dereliction or

-115-

culpable neglect, any of the improper actions identified by the Commission.

205. James Floyd had learned in his experience at TMI-l prior to 1978, that the leak rate procedure produced inconsistent results, and, therefore, he expected similar difficulties at TMI-2. Tr. 4984, 4978 (Floyd); see also Floyd, ff. Tr. 4894 at 5-6. Because of his confidence in the shift supervisors and their experience at TMI-1, Floyd thought they would be able to handle the problem of inconsistent results; and if the shift supervisors were unable to solve a problem, he expected them to come to him for assistance. Tr. 4979 (Floyd).

206. Floyd talked to the CROs "on a daily basis" and was in the Control Room "many hours every day." Tr. 4974 (Floyd). In fact, the only time Floyd used his office was for disciplinary reasons: "The rest of the time I lived out of the control room and the shift supervisors' office." Tr. 4975 (Floyd). Despite Floyd's considerable presence in the Control Room area, he has no recollection of knowing about any of the common problems opera-tors were experiencing with the leak rate test. Tr. 4976 (Floyd). Again, Floyd expected the CRos to come to him with any questions or problems they may have had. Tr. 4974 (Floyd). Be-cause the leak rate test did not require much of the operator's time, Floyd testified, when he was in the Control Room, he did not pay attention to what the operator was doing and

"(blasically. . .was ignorant of what was going on." Tr. 4976 (Floyd).

-116-

207. The Board does not fault the general concept of a man-ager's relying on subordinates to raise problems with him. The Board, however, cannot accept the notion that by seemingly exclu-sive reliance on this management technique a member of management can wholly insulate himself from responsibility for his subordi-nates' problems, particularly where he neither discharges the requirements of his position (e.g., to review logs) nor requires those under his direction to discharge the requirements of their positions (e.g., to file E&Ds), either of which would could have led to disclosure of those problems and presumably their correc-tion. Moreover, he cannot get credit on the one hand for in-volvement with his personnel by physically claiming presence in their work area (i.e., the Control Room) while disclaiming on the other hand any knowledge of problems they were having because they did not specifically call those problems to his attention.

208. We find Floyd negligent in his conduct on two counts.

First, Floyd admitted that he should have been aware of the fail-ure by operators to record the start and stop times of leak rate tests in the CRO Log Book, as required by AP 1012, "because I was required to review the log book once a week." Floyd, ff. Tr.

4894 at 6. Floyd also admitted that he did not enforce the ap-plication of E&Ds to leak rate tests, as required by AP 1010.

Id. at 3; see also Tr. 4991-93 (Floyd). Operators thus were "failing to take [ actions] in violation of technical specifica-tion requirements." CLI-85-18, 22 N.R.C. at 881. By failing to

-117-

review the CRO Log and enforce the application of E&Ds to leak rate tests, Floyd through "derelection or culpable neglect" was allowing such improper actions to occur. These very actions, if corrected, might well have highlighted the greater underlying problems with leak rate testing practices and led to their cor-rection.

209. In 11 105-106, 184, supra, we found that Floyd knew that operators were discarding invalid leak rate tests; according to Floyd, the practice had developed in the Operations Depart-ment, and the shift supervisors, shift foremen, and Floyd himself had carried the practice from Unit 1 to Unit 2. In 1 107, supra, we were unable to determine whether Floyd, like Seelinger, also knew that operators were discarding valid tests under an errone-ous interpretation of the Tech Specs. We reiterate here our finding that Floyd knew of the improper practice of discarding invalid tests. We also find that Floyd, as Supervisor of Opera-tions, although he may not have known, by his dereliction or cul-pable neglect allowed, the improper action of discarding valid tests.

210. As for the second type of negligent conduct, Floyd admitted that the analysis of plant status he provided on October 18, 1978, was invalid because he attempted to determine the " legality" of continuing to operate the plant without asking for all of the relevant information. Tr. 4919-20 (Floyd); see 1 115, supra. In response to the Board's inquiry concerning how

-118-

Floyd could correctly answer Seelinger, who had sent Bezilla with three tests for analysis, NRR Test Nos. 12C-E, Floyd responded, "All I was asked for was to look at these three pieces of paper."

Tr. 4919 (Floyd). The Board believes that the Supervisor of Operations must initiate and probe as well as receive and ob-serve. By failing to demand the further information necesnary for a valid analysis, Floyd lost a critical opportunity to dis-cover that operators were again failing to take an action --

entry into the Action Statement immediately upon obtaining a s

valid leak rate over 1 gpm -- in violation of Tech Spec 3.4.6.2.

We therefore find that Floyd by his " dereliction or culpable ne-glect" allowed one of the improper actions enumerated by the Com-mission in its Order and Notice of Hearing.

211. Our analysis of the record relating to Seelinger dis-cussed at length in 51 103-140, supra, has led us to conclude that by his own dereliction or culpable neglect, Seelinger al-lowed certain improper actions during leak rate surveillance testing. For purposes of this discussion, the Board focuses only on those salient acts of commission or omission that relate to our finding of dereliction and culpable neglect on the part of Seelinger. In 1 107, supra, we found that Seelinger's two artic-ulated reasons for changing his " initial reaction" to his discov-ery that operators were repeatedly running tests and discarding all tests greater than 1 gpm -- namely, the comment by the shift supervisor about shutting down TMI-l and the common ground

-119-

1 e

reached in the meeting with Floyd -- are not sufficient to excuse _

him from the responsibility of his position as Superintendent of -

Technical Support. ]

212. In early October 1978, the record is clear that James Seelinger learned of the practice of discarding tests. The fact f that a shift supervisor subsequently told him that TMI-1 would shut down if Seelinger's interpretation of the Tech Specs were adopted and that Seelinger then intended to raise this issue with i the TMI-l PORC does not excuse Seelinger from the affirmative duty to act and follow through after gaining actual knowledge of a potentially improper practice. Furthermore, the meeting with Floyd at which Seelinger claims they reached a common ground of -

interpretation does not justify following an interpretation when  ;

Seelinger had actual knowledge that adherence to such an inter- l pretation in practice meant the repeated running of tests and the A 2

discarding of all tests greater than 1 gpm.  ;

213. The Board believes that Seelinger's awareness in early 3

October 1978 of the practice of repeated running and discarding of tests in and of itself is sufficient evidence for the Board to -

find that Seelinger by dereliction or culpable neglect allowed improper operator actions. Seelinger, however, allowed a bad situation to get worse. He has admitted that he was the source of an instruction to shift supervisors not to leave tests lying around to prevent the NRC from seeing them and becoming aware of -

an intepretation of the Tech Specs with which he was i

-120-i

uncomfortable. Tr. 4756 (Seelinger). Seelinger's concern about the NRC's possibly finding bad tests lying around was realized on October 18 with Haverkamp's arrival in the Control Room; at the same time, the interpretation with which Seelinger was uncomfort-able and which led him to advise the shift supervisors to keep bad tests out of sight was replaced by the immediate Action Statement entry interpretation insisted on by Haverkamp. See Tr. 4757-60 (Seelinger). Despite the relatively short duration of Seelingec's concern, the Board does not view his decision to 3

keep information froht the NRC as either reasonable or tolerable.

214. In his letter to Grier informing him of the reportable occurrence, Seelinger stated that "[u]nidentified leakage was

\

reduced to (Tech Spec] limits at 0735 on 10-18-78" but failed to state that rounding off was the method of reduction. Board Exh.

1-A, Stier Report, Vol. V(C) at Tab 31 (Grier Letter). Seelinger also testified that one of the reasons behind his erroneous and misleading reference to discovery of a reportable situation "at 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> on October 19, 1978" may have been the requirement to report within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> on a prompt reportable occurence."

Tr. 4782 (Seelinger).

215. In a phone call with Haverkamp which Seelinger believes occurred approximately ten days after the October 18 inspection, Seelinger was not " straightforward" when he informed Haverkamp about the extent to which the rounding off practice had been relied on to achieve acceptable results. Tr. 4772-74 (Seelinger).

-121-

216. Seelinger took an active role in the drafting and editing of the Narrative to the LER, which contained a number of misleading or inaccurate statements in both the version sent to management in Reading and placed in the Control Room binder and the one sent to the NRC. See Tr. 4640-44, 4661-64 (Seelinger).

217. Seelinger has described, at length, his personal ef-forts to implement the corrective actions described in the LER narrative, but the record is clear that these efforts were unsuccessful. See Seelinger, ff. Tr. 4623 at 8; Tr. 4569-79 (Seelinger); Board Exh. 1-A, Stier Report, Vol. V(C) at Tab 36 (Seelinger 8/9/84 Letter to Palladin)). One of Seelinger's per-

/ }

sonal follow-up efforts that the Board does wish to focus on in-volved his attempts to inform operators of the correct interpre-tation of the Tech Specs by discussing the matter in a POD meeting and again at a later date on second shift with a shift supervisor. See Seelinger, ff. Tr. 4623 at 8. Seelinger testified that when he brought this matter up on second shift, "I was disappointed to find out that I was addressing the same shift supervisor that I had previously addressed. . . . I had hoped I was going to get another shift supervisor to talk to." Tr. 4677 (Seelinger). It is unknown why Seelinger, instead of feeling disappointed, simply did not seek out other shift supervisors.

The Board finds Seelinger's follow-up effort in this instance to be lacking in the reasonable level of initiative and drive to be

< expected in a Superintendent or Technical Support.

-122-

218. Finally, contrary to one of the Narrative statements to which Seelinger made a substantial contribution, the appropriate personnel were not adequately " instructed on the requirements of the applicable sections of the [ Tech Specs) and the requirements to immediately invoke applicable action statements when the pro-visions of limiting conditions for operation are not met." The general practice among operators of running tests repeatedly and discarding all bad tests did not change, and their 72-hour inter-

~' ---

pretation persisted as if the series of events related to the LER had never occurred. In 11 126-140, supra, the Board examined at length the inaccurate and/or misleading statements in the LER Narrative and stated the extent to viiich Seelinger had acted im-properly or insufficiently in several different instances. For 1 purposes of our finding of dereliction and culpable neglect, we simply reassert the remaining conclusions arrived at in that sec-tion.

III. CONCLUSIONS V 220. On the basis of the detailed findings above, the Board reaches the following conclusions on the operator issues identi-fied by the Commission in CLI-85-18:

(a) Operators at TMI-2 received inadequate instruction on Technical Specification requirements related to leak rate testing and incorrectly interpreted these requirements. l l

Corrective actions following the discovery in October 1978 e

-123-

of the incorrect interpretation were ineffective and the practices continued.

4 (b) Operators experienced considerable difficulties when conducting leak rate tests. These difficulties were recog-o nized by most operators as well by some members of on-site management, but no effective corrective actions were taken.

Operators felt an inherent pressure to obtain good leak rates from shift supervision so that the plant could oper-ate, but the pressure was not excessive and does not appear to have existed at all above the level of shift supervisor.

In any event, the pressure did not translate into a percep-tion among operators that the plant should not be shut down if necessary.

(c) Leak rate test results were customarily discarded, a practice recognized and accepted by the Operations Depart-ment management and for at least some period by the Superin-tendent of Technical Support (who held the position prior to December 1978), whose discomfort led him to instruct one or more Operatons Department personnel not to leave out tests in order to minimize their disclosure to the NRC.

(d) No member of management appears to have performed, con-doned, directed, or known of any manipulation of leak rate tests, although both the Supervisor of Operations and Super-intendent of Technical Support (who held the position prior

-124-

.k

to December 1978), by their dereliction or culpable neglect, allowed such practices to occur.

Dated: January 9, 1987 Respectfully submitted, SHAW, PITTMAN, POTTS

& TROWBRIDGE 2300 N Street, N.W.

Washington, D.C. 20037 (202) 663-8000 By:  !- Me [*?

Ernest L. Blake, Jr.

John N. Nassikas III Counsel for .)

GPU Nuclear Corporation

-125-

APPENDIX A List of Exhibits TMI-2 Leak Rate Proceeding Exhibit Description Identified Received No. at Tr.Page at Tr.Page 1-A TMI-2 Reactor Coolant Inventory Balance 388,391(as 388(as Testing, Prepared for GPU Nuclear Corp. modified modified by Edwin H. Stier, September 5, 1985, at 569) at 570),

all volumes. 5104,5221 ,

L 1-B Review of NRR and OI Reports on THI-2 388,391(as 388(as -

Reactor Coolant Inventory Balance modified modified -

Testing, Prepared for GPU Nuclear Corp. at 569) at 570),

by Edwin H. Stier, August 28, 1986 5104,5221  ;

2 1-C Letter from Edwin H. Stier to Philip R. 388,391(as 388(as  :

Clark, dated September 2, 1986 modified modified at 569) at 570),

5104,5221  ;

I 2 Results of Faegre & Benson Investigation 389-89 389, of Allegations of Harold W. Hartman, Jr. 5104,5221 Concerning Three Mile Island Unit 2, September 17, 1980, all volumes, but excluding Vol. 1, Sec. IV and Vol. 2, Ch. 9. i 3 Portion of Page 12 of Report of GPU 389 389, Assessment Panel for Individuals 5104,5221 Involved in THI-2 Leak Rate Testing (in 1978-1979), January 6, 1986.

(re Herbein) .

4 TMI-2 Computer Log (October 27, 1978). 389 389, ,

5104,5221 '

5-A Results of NRR's Investigation and 389(as 389(as Evaluation of Ten Licensed Operators modified modified Ei,sp.Jl involved in TMI-2 Preaccident Leak at 569-71 at 571), ~. C v #

Rate Testing Irregularities, including 5104,5221 #$":f .)

t attachments and supporting documents. #pe?.g.,

(per the Board's 6/24/86 Order, p. 7, j@A%d undifferentiated references to .?r' f :

" management" are to be disregarded.) "((j$[

QW, lgy m fJ;f m

a$f:h EnK

l Exhibit Description Identified Received No. at Tr.Page at Tr.Page S-B Memorandum from William Russell to 389(as 389(as Harold Denton, dated September 20, modified modified 1985, with enclosed report. at 569-71 at 571), ,

5104,5221 6 Office of Investigations Report 389-90 390, entitled: Three Mile Island-2: 5104,5221 Investigation of Individual Operator Actions Concerning the Falsification of Laak Rate Test Data, including all p attachments. I 7 U.S. Nuclear Regulatory Commission, 390 390, Office of Nuclear Reactor Regulation, 5104,5221 TMI-1 Restart: An Evaluation of the Licensee's Management Integrity as It Affects Restart of Three Mile Island Nuclear Station Unit 1, Docket 50-289 (July 1984) (NUREG-0680, f' Supp. No. 5, Sec. 4.0.)

8 Memorandum of Oct. 27, 1978 390 Withdrawn 5103,5221 9-A Photographs of TMI-2 Control Room 377-381(as 381(as thru renumbered) renumbered 9-F at 391), ,

5104,5221 9-G Photograph of the TMI-2 makeup 1182 1183, storage tank level indicator, 5104,5221 pressure indicator and temperature i ,

indicator.

1 10-A Sample SRO license 1025-26 1027,5104 '

5221 Sample R0 license 1026 1027,5104 5221 11-A J. Moore Charts, Effect of 1025 5104,5221 and Oscillations 11-B 12-A Administrative Procedure (AP) 1036, 1354-55 5104,5221 Instrument Out-of-Service Control, Rev. O, March 1978 A-2

Exhibit Description Identified Received No. at Tr.Page at Tr.Page 12-B Administrative Procedure (AP) 1036, 1354-55 5104,5221 Instrument Out-of-Service Control, <

Rev. 1 August 1978 13 Drawing of peak-to-peak v. slope 1845(as 5104,5221 1 offset calculation (included in modified at Exhibit 24.) 5103-04) a 14 Stier cover letter to Board 1862(as 1862 with two strip charts showing modified at hydrogen additions (included 5103-04) in Exhibit 24.)

15-A Curriculum vitae of Dr. Harrison 2008 2008, and and Mr. Cole 5104,5221 15-B 16 Training Department Administrative 2219-20 2234,5104 Memorandum Number 5, Change 2, dated 5221 October 8, 1976 (Subject-Category IV CR0 Training Program.)

17 Excerpts from Instrument Out of 2400 2400,5104 Service Log 5221 18 RCIB Test of 2/15/79, portion of CR0 3470 3470,5104 log and MUT Strip Chart 5221 19 TMI-2 Control Room Layout diagram 3500-01 3501,5104 5221 20 D. Haverkamp prefiled testimony and 4336 4337,5104 attachments 5221 21 Bettenhausen affidavit 4336 4337,5104 5221 22 TMI-2 FSAR Pages 9.3-26 thru 9.3-30 4431 4431,5104 and 5.2-16 thru 5.2-27 5221 23 TMI-2 FSAR Section 5.2.3.4, and 4431 4431,5104 Table 5.2-12 5521 24 Stier letter to Judge Kelley, dated 5103-04 5221 September 30, 1986, with attachments A-3 ,.f

~

1

h Exhibit Description Identified Received No. at Tr.Page at Tr.Page 25 MPR letter to Edwin Stier, dated 5104 5221 g September 30, 1986 26 Letter dated Nov. 3, 1986 from 5266-67 Rejected R. Gallo to Rochester Gas and 5267 Electric Corp., enclosing NRC Region I Inspection Report No. 50-244/86-16, and Numerous Employees Follow-up Questions for Donald R. Haverkamp

~

27 Letter dated Oct. 21, 1986 from Bd. Order Bd. Order J.P. Moore to W.A. Rockwell, 1/19/86 1/ 19/86

Subject:

MUT Level Transmitter I

A-4

APPENDIX B LIST OF WITNESSES AND TESTIMONY Written Oral Witnesses Testimony Testimony Date Adams, Charles D. ff. Tr. 3776 3775 to 3841 10/15 Boltz, Dennis J. 2218 to 2234 9/25 t

Bocher, Raymond R. ff. Tr. 4175 4175 to 4232 10/28 (

Brummer, John A.* ff. Tr. 5236 Bryan, Kenneth P. ff. Tr. 4540 4539 to 4610 10/31 Capra, Robert A. 374 to 1904, 9/8,9/9,9/10,9/11, 2207 to 2209 9/12,9/16,9/17, 2313 to 2439 9/18,9/25,9/30 Christopher, Keith 2314 to 2438 9/30 Chwastyk, Joseph J. ff. Tr. 3407 3405 to 3600 10/10 Cole, Norman M., Jr. 374 to 2034 9/8,9/9,9/10,9/11, 9/12,9/16,9/17, 9/18,9/19 Coleman, Mark S. ff. Tr. 2579 2578 to 2657 10/1 Conaway, William T., II ff. Tr. 3097 3096 to 3141 10/7 Congdon, Joseph R. ff. Tr. 2709 2708 to 2782 10/2,10/3 Cooper, Martin V. ff. Tr. 2835 2835 to 2945 10/6 Faust, Craig C. ff. Tr. 2511 2511 to 2577 9/30,10/1 Fels, William J. 4489 to 4535 10/30 Floyd, James R. ff. Tr. 4894 4892 to 5036 11/4,11/5 Frederick, Edward R. ff. Tr. 2447 2446 to 2510 9/10 Germer, Leonard P.* ff. Tr. 5236 Guthrie, Carl L. ff. Tr. 4113 4112 to 4159 10/17 B-1

Written Oral Witnesses Testimony Testimony Date Harrison, Dwight H. 374 to 2034 9/8,9/9,9/10,9/11, 9/12,9/16,9/17, 9/18,9/19 Hartman, Harold W., Jr. 2239 to 2309 9/25 Haverkamp, Donald R. Exhibit 20 2042 to 2210 9/24 5237 to 5267 11/12 Hemmila, Earl D. ff. Tr. 4039 4038 to 4111 10/17 Herbein, John G. ff. Tr. 5268 5267 to 5320 11/12 Hitz, Gregory R. ff. Tr. 3664 3663 to 3732 10/14 Hoyt, Kenneth R. ff. Tr. 4233 4233 to 4299 10/28 Illjes, Theodore F. ff. Tr. 3010 3010 to 3096 10/7 Kidwell, John M. ff. Tr. 3285 3285 to 3399 10/9 Kirkpatrick, Donald C.** ff. Tr. 376 374 to 1243 9/8,9/9,9/10, 9/11,9/12 Kunder, George A. ff. Tr. 4800 4799 to 4889 11/4 1

Logan, Joseph B. 5105 to 5220 11/6 -

Marshall, Walter J. f f. Tr. 438') 4379 to 4424 10/29 McGovern, Hugh A., Jr. ff. Tr. 31L8 3147 to 3235 10/8 Mehler, Brian A. ff. Tr. 3842 3841 to 3907 10/15 Mell, Charles F. ff. Tr. 3239 3239 to 3282 10/9 Miller, Adam W. ff. Tr. 3608 3607 to 3662 10/14 Miller, Gary P. ff. Tr. 5039 5038 to 5096 11/5 Moore, James P. 374 to 1165 9/8,9/9,9/10,9/11 Olson, Dennis I, ff. Tr. 3911 3911 to 4034 10/16 .

Phillippe, Mark D. ff. Tr. 4432 4431 to 4489 10/30 Rockwell, Winthrop A. 374 to 1165 9/8,9/9,9/10,9/11 B-2 m---m--m u s- um m

Written Oral Witnesses Testimony Testimony Date Russell, William T. 374 to 1904, 9/8,9/9,9/10,9/11 2313 to 2438 9/12,9/16,9/17, 9/18,9/30 Seelinger, James L. ff. Tr. 4623 4614 to 4784 11/3 Scheimann, Frederick J., Jr. ff. Tr. 2831 2783 to 2831 10/3 Stier, Edwin H. 374 to 2034 9/8,9/9,9/10,9/11, 9/12,9/16,9/17, 9/18,9/19 Smith, Bernard G. ff. Tr. 4331 4330 to 4379 10/29 Bermiel, Jared S.** ff. Tr. 376 374 to 801 9/8,9/9,9/10 Uright, Lynn 0.*** ff. Tr. 2663 2661 to 2707 10/2 Zewe, William H. ff. Tr. 2946 2945 to 3006 10/6

• The Board did not call Brummer and Germer as witnesses. Their prefiled testimony, however, was bound into the record and reflects what the witnesses would have sworn to had they appeared. Their statements could be referred to as if the witnesses had appeared to testify. Tr. 5236 (Kelley, J.).

• • Joint prefiled - Kirkpatrick and Wermiel

• • • Because Wright did not adopt his prefiled testimony, it was bound into the record, not as substantive evidence, but as a brief accurate summary of a longer statement placed in the record (Board Exh. 6, OI Report, Exh. 18, Wright Interview) and regarded as substantive evidence. Tr. 2662-63 (Kelley, J.).

B-3

Documentary Material Bound into the Transcript Description Page Date Professional Qualifications of Robert A. Capra ff Tr. 651 9/9 -

Surveillance Data Flow Chart ff Tr. 731 9/9 -

Questions 5 thru 8 Submitted by Numerous Employees for D.R. Haverkamp ff Tr. 2169 9/24 Personal Resume of Dennis J. Boltz ff Tr. 2217 9/25 Paragraph from Page 7 of Prefiled Testimony of B. G. Smith Tr. 2262 9/25 List of Exhibits ff Tr. 5221 11/6 I

B-4 4

APPENDIX C*

CORRELATION OF NRR AND STIER TEST NUMBERS FOR EACH TMI-2 LEAK RATE SURVEILLANCE TEST PERFORMED BETWEEN 9/30/78 AND 3/28/79 NRR TEST STIER TEST No. No.

SEPTEMBER 1978 1 161 OCTOBER 1978 2 160 3 159 4 158 5 157 6 156 7 155 8 154 9 153 10 152 11 151 12 150 12A 149 12B 148 12C 147 12D 146 12E 145 13 144 14 143 15 142 16 141 17 140 18 139 19 138 20 137 21 136 22 135 23 134 24 133 25 ---

26 132 27 131 28 130 29 129 30 128 31 127 2

o See Board Exh. 5-B, NRR Report, Attachment 1.

NRR TEST STIER TEST No. No.

NOVEMBER 1978 32 126 33 125 34 124 35 123 36 122 37 121 38 120 39 119 40 118 4 41 117 DECEMBER 1978 42 116 43 115 1 44 114 45 113 46 112 47 111 48 110 49 109 50 108 51 107 ,

52 106 53 105 -

54 104 55 103 56 102 57 101 58 100 59 99 60 98 61 97 62 96 63 95 64 94 '

I 65 93 r 66 92 67 91 68 90 69 89 70 88 71 87 72 86 73 85 74 84

NRR TEST STIER TEST No. No.

75 83 76 82 77 81 78 80 79 79 80 78 ,

81 77 ,

82 76 JANUARY 1979 83 75 84 74 85 73 86 72 87 71 88 70 89 69 90 68 91 67 92 66 93 65 94 64 95 63 96 62 ,

4 FEBRUARY 1979 97 61 98 60 99 59 100 58 101 57 102 56 103 55 104 54 105 53 106 52 '

107 51 108 50 109 49 110 48 111 47 112 46 113 45 114 44 115 43 116 42

h NRR TEST STIER TEST No. No.

117 41 118 40 119 39 120 38 121 37 122 36 123 35 124 34 125 33 126 32 127 31 128 30 129 29 130 28 131 27 132 26 133 25 MARCH 1979 134 24 135 23

-136 22 137 21 138 20 139 19 140 18 141 17 142 16 143 15 144 14 145 13 146 12 147 11 148 10 149 9 150 8 151 7 152 6 153 5 154 4 155 3

--- No Test 2 156 1

DOLKE ri p UNITED STATES OF AMERICA ~ "WC NUCLEAR REGULATORY COMMISSION BEFORE THE PRESIDING BOARD '87 JAN 12 R2:14 YGC.l:

T'

)

In the Matter of )

) Docket No. LRP INQUIRY INTO THREE MILE )

ISLAND UNIT 2 LEAK RATE ) ASLBP No. 86-519-02 SP DATA FALSIFICATION )

)

CERTIFICATE OF SERVICE I hereby certify that on January 9, 1987 I served the foregoing "GPU Nuclear Corporation's Proposec Findings of Fact and Conclusions of Law" b,y hand-delivering a copy tTereof to the following persons marked with an asterisk and by maillng, first class, postage prepaid, a copy thereof to the remaining individuals and office on the follow-ing list:

• Administrative Judge James L. Kelley, Chairman Atomic Safety and Licensing Board Panel U.S. Nuclear Regulatory Commission Washington, D.C. 20555

• Administrative Judge Glenn O. Bright Atomic Safety and Licensing Board Panel U.S. Nuclear Regulatory Commission Washington, D.C. 20555

• Administrative Judge James H. Carpenter Atomic Safety and Licensing Board Panel U.S. Nuclear Regulatory Commission Washington, D.C. 20555

Mary E. Wagner, Esq.

Office of General Counsel U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Docketing and Service Branch (3)

U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Harry H. Voigt, Esq.

James W. Moeller, Esq.

LeBoeuf, Lamb, Leiby & MacRae 1333 New Hampshire Avenue, N.W.

Suite 1100 Washington, D.C. 20036 Smith B. Gephart, Esq.

Jane G. Penny, Esq.

Killian & Gephart 216-218 Pine Street Box 886 Harrisburg, Pennsylvania 17108 James B. Burns, Esq.

Isham, Lincoln & Beale Three First National Plaza Suite 5200 Chicago, Illinois 60602 Michael W. Maupin, Esq.

Hunton & Williams P.O. Box 1535 Richmond, Virginia 23212 Mrs. Marjorie M. Aamodt Box 652 Lake Placid, New York 12946 and Mrs. Marjorie M. Aamodt 200 N. Church Street Parkesburg, Pennsylvania 19356 e b sfohn N. Nassikas III

__ _