ML20203L473

Prepared Statement of Wh Zewe Re Leak Rate Data Falsification.W/Certificate of Svc.Related Correspondence

ML20203L473
Person / Time
Site:	Three Mile Island
Issue date:	08/22/1986
From:	Zewe W GENERAL PUBLIC UTILITIES CORP., METROPOLITAN EDISON CO.
To:
Shared Package
ML20203L462	List: ML20203L460 ML20203L465 ML20203L469 ML20203L473
References
LRP, NUDOCS 8608260280
Download: ML20203L473 (9)

Text

-- . - - - .

.'I.

WitD COMM

1. $$G I UNITED. STATES OF AMERICA NUCLEAR REGULATORY COMMISSION .

Q g P 4 :3 2 BEFORE THE PRESIDING BOARD 0 0 Dr e

?ty' U,E i ERA $cdL*"I,'n

. iq' v

i

)

In the Matter of )

)

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

FALSIFICATION )

)

PREPARED STATEMENT OF WILLIAM H. ZEWE My name is William H. Zewe and I live in Hershey, 1

Pennsylvania. I am presently employed as the Manager of Metropolitan Edison Company's Titus Generating Station.

I began employment with Metropolitan Edison Company on February 14, 1972 as an auxiliary operator. I worked my way up i to the shift foreman's position at Unit 1 in September of 1973,

! and in April or May of 1976 I was promoted to shift supervisor. I remained a shift supervisor until August of 1982, when I became a radwaste operations manager at TMI-1. In January of 1984, I went to the Titus Station.

While at TMI-2, I was normally assigned to supervise "A" shift, which was comprised of shift foreman Frederick Scheimann and control room operators Craig Faust and Edward Frederick. I also worked with control room operators Hugh McGovern and Lynn Wright on various occasions.

8608260280 PDR 860822 ADOCK 05000320 PDR 1 "w-=' * = 'w

• ,m q . . - ,--,e-_,---yw,y- mgg .,y- -

---

r+M--7 -t'T"p -  ? -*T ?

O .

A As shift supervisor, I was otten the most senior person on site, and while I ultimately was responsible for the operation of both Units, I left the shift foremen with the responsibility of running their respective Units. The actual leak rate testing was performed by the control room operators. It was the shift foreman's responsibility to ensure that the operators performed the tests properly, and he would approve the results after receiving the tests from the operators. I cannot recall ever personally running any leak rate tests while I was a shift supervisor, but I did understand that it was partially my responsibility to ensure that the technical specifications were complied with, and it was ultimately my responsibility to ensure that the plant was operating safely.

The Unit 2 technical specifications required that we obtain one valid leak rate test of less than one gallon per minute unidentified leakage every seventy-two hours. We complied with this by running a test every day and sometimes every shift. In the event that actual unidentified leakage exceeded one gallon' per minute, we then had to either reduce the leakage rate to within limits by identifying and quantifying leakage, or enter the Action Statement.

I can recall personally inspecting for reactor coolant system leaks on many occasions. Whenever a leak was identified, I would either go myself, or have someone else go and check it for several things, including the safety implications. We would measure the amount of leakage as t ,. y i' 11 ,

,s' '

.r J

i

, c

! /

occurately as we could, then enter the leakage fiauro ipto the computer to re-categorize unidentified leakage as identifba,3 ,

leakage. Once we had measured and identified the leakage,', ewe ,' ' ' r ><

would then run another leak rate test. We would also insphct- ) /j / ,r

. , e the leak to evaluate its potential for getting worse, its- ,

i effect on other equipment, and to determine its source (thad ,

l-is, whether it was a weld leak, a boundary leak, a packing leak i +

, /

or whatever). ,

i There were often times when the operators were prevented {

from running leak rate tests because the plant was not in

steady state. Often the leak rate test results w'ould be // o i j .

determined inaccurate because the computer's test resultsidid, ,,

'f ;, ',

not agree with what the other plant instruments or pcrameters. ,

r' '

indicated RCS leakage to be.

r P(

Operators who ran leak rate tests that were inva' lid or ' '

4.

inaccurate routinely would discard those tests and initiate'a, 1 r r-new test. Before the operators discarded invalid tests, th$y .

I, were to compare the results with other plant parameters to[f '

ensure that the tests were invalid. The control room ,

' s operators, the shift foreman, and the shift supervisor, had the ability to make the determination that a test was invalid.

Although I was aware that invalid leak rate tests were ,

routi ely being discarded, it was not a requirement t, hat I be '

c.

notified before each one was thrown away. It was my understanding that all of the shifts, in both Unlts 1 and 2, followed the same policy of discarding invalid leak rate tests.

4 i

l ' I

. . - - - - --nu _ d..---e. -. , - - - - , - - , , - - - - - - - -

s Invalid leak rate tests were not the only surveillance tests that were discarded. If we were conducting a surveillance on a decay heat pump, or a nuclear river pump, and for some reason we had to turn on that pump, then we would void the invalid test and discard it. I specifically recall one time in particular when we were doing a control rod movement test in Group 3 and the reactor tripped. The control rod movement test would not have been valid then because it tripped, so we discarded the invalid test.

I remember that the computer would sometimes print negative leak rates. I attributed the negative leak rates to a variation or statistical error in the computer calculation.

Depending on their magnitude, some negative leak rates might be accepted as valid and some might be invalidated and discarded.

The decision was left to eachpshift's judgment.

There were a number of general complaints from the operators concerning the accuracy of the leak rate tests performed by the computer. A small plant oscillation could throw off the test, and make the difference between a good leak rate as determined by the computer and a bad leak rate. We expected a few bugs in the leak rate testing system because .

Unit 2 had only recently come on line. While we did not ignore the computer leak rate tests, they were only one of many things we looked at to determine leakage.

As a result of the various things we have learned during the course of the TMI-2 leak rate investigations, I now feel t

.3 I

s that.the leak rate test that was available,to us in 1978 and 1979 was not an effective tool for measuring leakage. However, in 1978 and 1979 it was the method of precisely measuring leakage available to us, and we did not have the luxury of s

analyzing the test as has been done during subsequent investigations. u It is very difficult to separate what I know;now from what I knew back in 1978 and 1979. In reviewing different documents over the past several years, I have been made aware of problems 1 ,

with one of the makeup tank level instruments, and that I made entries in the turnover notes concerning these problems. I have no independent recollection of problems'with this instrumentation, and I have no recollection of lant operators ever attempting to switch makeup tank level instrumentation back and forth in order to get accurate leak rate test results.

I do not remember if all watet' additions to the reac. tor coolant system inventory were required to be recorded in the CRO log book, but that was standard procedure on "A" shift; ,

also, if water was added during a, leak rate test, the water addition would have to be computed into the leak rate program.g If the operator running the test failed to realize that an unrecorded water addition had been made, then he would not know to invalidate the test and would record the result, if it was acceptable. If the operator running the test realized that there had been an inadvertent, unrecorded water addition made while the leak rate test was being run; the test would be

a A

considered invalid and would be discarded. I am unaware of any incident where an operator added water during a leak rate test, and knowingly failed to enter the addition into the program in a deliberate attempt to alter the test result.

At Unit 2, hydrogen was added by a control room operator through a control valve on the panel (except for a period of time when the valve was not working, and hydrogen was added out in the plant by auxiliary operators). I am not sure if hydrogen was generally considered to be a chemical addition or not; however, I did not consider it to be a chemical. I cannot recall whether or not the operators were required to record

.i additions of hydrogen in their log book. In the pre-accident period, I had no knowledge that an addition of hydrogen would affect a leak rate test, and I do not remember thinking about

- it or considering it. I cannot recall any discussion among the operators concerning how a hydrogen addition would affect the makeup tank level. The first time I heard anything about hydrogen affecting leak rate tests was when Harold Hartman made his allegations.

As shift supervisor, I was aware back in 1978 and 1979 that Administrative Procedure 1010 required an Exception or Deficiency when certain problems were encountered during i surveillance testing. I do not remember considering this procedure in connection with leak rate tests.

Prior to the start of these investigations, I had no independent recollection of the issuance of a Licensee Event

=.. . , -

t . .

k 4

Report on November 1, 1978. I do recall, however, hearing at some point that a new interpretation of the technical specification required that we begin rounding off the leak rate tests. I do not remember exactly when that practice began, but I do remember that it did not last long.

3 I also have a problem remembering a Temporary Change Notice to the leak rate procedure that was issued in 1979. I have l

a trouble differentiating between what I knew then about it, and what I now know as a result of being shown the TCN several i

times over the past few years.

I felt that with all the instrumentation and parameters that we were able to observe, we were operating a safe plant.

i In 1978, I thought that we were using all of the leakage detection methods available to us, and we were operating the i

plant the way it was supposed to be operated. The leak rate test was only one of many, many items we looked at to determine i

leakage, and it just was not something that we dwelled on or i

j focused on, which is being done now.

Although I am not employed in the nuclear industry, I would l like to clear my name of any involvement in leak rate i

~

improprieties. I never falsified a leak rate test, approved the falsification of a leak rate test, or was aware that anyone else had done so. I would be most appreciative, therefore, if the Presiding Board would officially recognize that I was not l

part of leak rate test falsification at TMI-2.

4 i

.s

\

UNITED STATES OF AMERICA Of, hIJ: 0 NUCLEAR REGULATORY COMMISSION BEFORE THE PRESIDING BOARD 16 Aui 22 P4 32 C

In the Matter of )

)

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

FALSIFICATION )

)

CERTIFICATE OF SERVICE I hereby certify that I have served copies of the Prepared Statements of Raymond R. Booher, Carl L. Guthrie, and William H. Zewe by deposit in the United States mail, first class, postage prepaid, or, as indicated by an asterisk, by hand delivery, to the following pers;ons this 22nd day of August 1986:

• 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 Jerry R. Kline Atomic Safety and Licensing Board Panel U.S. Nuclear Regulatory Commission Washington, D.C. 20555

s h

• Jack R. Goldberg, Esq.

Office of the Executive Legal Director U.S. Nuclear Regulatory Commission Washington, D.C. 20555

• Docketing and Service Branch (3).

U.S. Nuclear Regulatory Commission t Washington, D.C. 20555

• Ernest L. Blake, Jr., Esq.

Shaw, Pittman, Potts & Trowbridge 1800 M Street, N.W.

Washington, D.C. 20036 James B. Burns, Esq.

Isham, Lincoln & Beale Three First National Plaza Suite 5200 Chicago, IL 60602 Michael W. Maupin, Esq.

Hunton & Williams P.O. Box 1535 Richmond, VA 23212 Ms. Marjorie M. Aamodt 200 North Church Street Parkesburg, PA 19365 Ms. Marjorie M. Aamodt P.O. Box 652 Lake Placid, NY 12946 w - & .pha nsa 1 0

James W. Moeller l

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