ML20203L210

Prepared Statement of Ba Mehler Re Inquiry Into Leak Rate Data Falsification.Related Correspondence

ML20203L210
Person / Time
Site:	Three Mile Island
Issue date:	08/21/1986
From:	Mehler B GENERAL PUBLIC UTILITIES CORP.
To:
Shared Package
ML20203L206	List: ML20203L202 ML20203L207 ML20203L210 ML20203L215
References
LRP, NUDOCS 8608250213
Download: ML20203L210 (9)

Text

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _

r W. p y r.sryn * %

UNITED STATES OF AMERICA 00CMETED USNRC NUCLEAR REGULATORY COMMISSION BEFORE THE PRESIDING BOARD kg g) p5;g4 CFFICE OF 5:xynyy 00CMETinG 6 SEyemp' SRANCH

)

In the Matter of ) ~

)

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

FALSIFICATION )

)

i PREPARED STATEMENT OF BRIAN A. MEHLER My name is Brian Allen Mehler. I currently reside in Palmyra, Pennsylvania. I have been employed by GPU Nuclear Corporation as radwaste operations manager at TMI Unit i since January of 1984.

I started employment with Metropolitan Edison in May 1967 at the Crawford Generation Station in Middletown, Pennsylvania. I began work at Three Mile Island in October 1969 as a CRO trainee at Unit 1. I served as a licensed CRO in Unit 1 until about September 1976, when I went to Unit 2 as a shift foreman. I became a shift supervisor in April 1978, by 1

which time I was licensed as a SRO on both Units. At some time after the TMI-2 accident, I ceased being licensed on Unit 2.

Later, I was asked by the company to give up my license on Unit

1. After giving up my Unit 1 license, I worked for Ron Toole for approximately three months as an operations and maintenance coordinator. After that, I attained my current position.

8609250213 860021 PDR ADOCK 05000320 T PDR

b As a shift supervisor, my basic responsibility was to supervise the shift foremen at each Unit and, during' periods when I was responsible for both Units, to ensure that both were operated in a safe manner. I reported to Mike Ross in Unit 1 and to Jim Floyd in Unit 2. I had regular contact with Jim Seelinger in the context of the Plan of the Day (POD) meetings. I had only occasional contact with Gary Miller.

On the daylight shifts, I spent most of my time in the control' room, discussing the scheduled evolutions with the shift foreman. I spent less time in the control room on the back shifts, however, since during those shifts I was expected to make frequent tours of the plant.

It was my responsibility to schedule surveillances so that they would not interfere unduly with plant operatiot.s.

Because the performance of leak rate tests had no impact on plant operations, I had little direct involvement with leak rate testing. As a shift supervisor, I was responsible for ensuring that leak rate tests complied with technical specifications and surveillance procedure requirements. In practice, I became involved in leak rate testing on the 11-7 or 3-11 shifts only when a control room operator or shift foreman informed me of problems with that surveillance test. In general, I assumed that acceptable leak rate test results were being achieved unless I heard otherwise. On the 11-7 shift, however, I expected to be presented with leak rate test results so that I could include these results in my morning report. It was very unusual for me to run a leak rate test myself.

b POD meetings and shift supervisor meetings occurred on a regular basis. As part of my basic duties, I attended POD meetings at each unit. The primary purpose of POD meetings was to discuss the plant evolutions scheduled to be performed that day. Although there was no set agenda for shift supervisor meetings, we typically discussed personnel problems and plant problems. I do not recall any discussion at those meetings of difficulty in obtaining satisfactory leak rate test results.

During 1978 and 1979, I was aware of the limiting conditions for reactor coolant system leakage, which provided that unidentified leakage not exceed one gallon per minute (gpm) and that identified leakage not exceed ten gpm. It was my responsibility, as well as the responsibility of every other licensed individual, to ensure that leak rates were within these limiting conditions.

As I understood them, the procedures concerning unidentified leak rate test results exceeding 1 gpm required that we determine the validity of that result. One means of doing this was to run another leak rate test. In addition to running another leak rate test, I assumed that the operators assessed the validity of the unsatisfactory test result by evaluating it against all the other plant indications. If it was concluded that the unsatisfactory leak rate test was invalid, the test results that prompted the search would be discarded. If the test result exceeding 1 gpm was deemed valid, the procedures provided that we identify and reduce b

leakage to within specified limits within four hours, or failing that, begin to shut down the plant. Leakage would be identified by sending a crew member out into the plant to attempt to locate the leak and quantify it using a graduated cylinder.

I was aware, during 1978-1979, of the general requirement that the reactor coolant system and makeup system be maintained in steady state during leak rate tests. Thus, I realized that water additions, borations and deborations should not be made to the makeup tank while a leak rate test was being performed. I certainly believed that the same source should be used when recording initial and final reactor coolant system temperature, pressurizer level, makeup tank level and drain tank level.

Prior to the accident, it did not occur to me that an exception or deficiency had to be filed with an unsatisfactory leak rate test. I do not recall receiving any instructicns from anyone to file an exception or deficiency with leak rate tests.

The results of leak rate tests would be communicated to the next shift through the status board in the shift supervisor's office and through the morning report. I would include in my morning report the most recent valid leak rate test result.

The leak rate test was not always a reliable indicator of leakage. I merely regarded the leak rate test as the best

> tool available to quantify plant leakage.

.o k

Often, one leak rate test was run on the 3-11 shift and one was run on the 11-7 shift. As a general rule, we did not run a leak rate test during the daylight shift. I realize, however, that there were instances in which two or more leak rate tests were performed on a shift.

I did not instruct the operators under my supervision to bring leak rate tests exceeding 1 gpm to my attention as a matter of course. I assumed that those under my supervision followed proper procedures upon obtaining a leak rate test in excess of 1 gpm (that is, run another leak rate test, identify plant leakage, and so forth). In general, I do not recall any situations at Unit 2 where I believed that a threat to plant safety existed as a consequence of reactor coolant system leakage.

I was aware during 1978-1979 of the practice of discarding leak rate tests that showed unidentified leakage in excess of 1 gpm but were deemed invalid based on an observation of plant parameters or on the results of a subsequent leak rate test, or both. I assumed that those under my supervision discarded unsatisfactory leak rate tests only after observing other indicators of plant leakage, and concluding that the test was invalid. If an instrument fluctuation occurred during a leak rate test, this could also serve as a basis for concluding that the test was invalid. I did not discard any leak rate tests nor do I recall directing anyone to discard leak rate tests ahowing ex'cessive unidentified leakage. I did witness, W

8 L l l

A however, the discarding of leak rate tests showing excessive unidentified leakage and those showing negative leakage. We assumed that a small negative leak rate was an acceptable result because the tests were somewhat inaccurate due to instrument errors. Thus many such tests were filed. The only time during which I recall many leak rate tests being discarded was just before the accident, when it was difficult to get a satisfactory leak rate because of a pressurizer relief valve.

However, my knowledge of matters just before the accident is limited because during most of the period from February 16 through the accident, I was assigned to Unit 1 for part of the time as an outage coordinator for its refueling and was assigned to different shifts while I was in Unit 2 .

I recall that Jim Floyd reported that Don Haverkamp had authorized us to round off the results of leak rate tests to the nearest whole number. This policy was rescinded, however, soon thereafter.

Apart from what I have learned through post-accident interviews, I have no recollection of the incident in which an NRC inspector discovered unsatisfactory leak rate tests lying around in the control room. I have no independent recollection of any discussions at POD meetings or shift supervisor meetings regarding the Licensee Event Report that resulted from that incident. Nor do I recall discussing the propriety of our interpretation of the 72-hour frequency requirement with Don Haverkamp in the aftermath of that incident.

Y 6

a Leak rate tests at Unit 2 were logged according to their completion time only, as had been the practice at Unit 1.

There were two reasons why hydrogen had to be added to the makeup tank on a periodic basis. First, the addition of hydrogen maintained the net positive suction head for the makeup pumps. In addition, the addition of hydrogen fulfilled the necessary function of scavenging oxygen from the reactor coolant system. My observations at Unit I had shown me that the addition of hydrogen to the makeup tank produced only a slight increase or slight decrease in the level indication of the makeup tank strip chart. In any event, the level indication returned to its original position very shortly after the hydrogen addition. I am not sure which other TMI-2 personnel (if any) were aware of this phenomenon, because I never discussed it with anyone while Unit 2 was operational.

In light of the negligibility of this effect- I believed during 1978-1979 that the addition of hydrogen to the makeup tank during a leak rate test had no significant effect on the test result. I have no knowledge of any operator deliberately adding hydrogen to the makeup tank during a leak rate test in order to affect the test result.

I was aware of the requirement that any water addition to the makeup tank had to be logged in the CRO log. I was not aware, during 1978-1979, of the alleged phenomenon by which the makeup tank strip chart would register a water addition greater in volume than that actually added and logged -- thereby

6 enhancing the leak rate test result. Nor was I aware, during that time period, that anybody was manipulating leak rate test results, as has been alleged, by gradually adding small quantities of water to the reactor coolant system. I also

.-believe that it would be impractical to jog water, because the operator would have to start and stop at least one pump a number of times with split-second precision to avoid  ;

detection. I do not recall any deliberate attempt on the part l

of an operator to conduct a feed-and-bleed operation during a leak rate test so as to manipulate the test result. I do not recall any operator falsifying leak rate test results by adding water during the course of the test.

I was not aware of anyone switching makeup tank level transmitters during the course of a leak rate test so as to enhance the test result. I believe that using a faulty level transmitter to feed information to the computer could just as easily lead to an unacceptable leak rate test result as an acceptable one.

I can recall no time when I felt pressure from my supervisors to operate the plant despite the fact that I felt it was unsafe to do so. Nor did I ever feel pressure from my supervisors to tolerate improper conduct on the part of my subordinates.

I was unaware, during 1978-1979, of any manipulation of leak rate tests by TMI-2 operators. I can assure the Presiding Board that had I known of such irregularities, my 1

l

6

, w concern for plant safety would have prompted me to take immediate corrective action. I have been employed in the nuclear industry for nearly two decades and I wish to remain so employed. I believe that the Board should find that there is no evidence of my involvement in leak rate test falsification or improprieties at TMI-2, so that I may finally be cleared of I

alleged involvement in these matters.

1 I

I

_ - - _ _ - _ _ _ _ .