ML19309G290
| ML19309G290 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 10/08/1979 |
| From: | Wilson R GENERAL PUBLIC UTILITIES CORP. |
| To: | GENERAL PUBLIC UTILITIES CORP. |
| Shared Package | |
| ML19309G289 | List: |
| References | |
| TASK-TF, TASK-TMR NUDOCS 8005050517 | |
| Download: ML19309G290 (17) | |
Text
8005050Fl?
Interview Summary U
Richard Wilson Director of Technical Functions GPU Service Corporation Mountain Lakes, New Jersey This interview was conducted by Ronald Haynes, Hartmut Schierling, Dennis Allison, and Barry llorvick of the NRC Special Inquiry Group on October 8, 1979 in Trailer 11 at the Three Mile Island plant site from about 10:00 a.m.
to 2:00 p.m.
Mr. Wilson had previously been interviewed by the NRC Office of Inspection and Enforcement, CIE, investigation team on June 1,1979 (TMI Tape 292).
A transcript of the previous interview, which de' scribed Mr.' Wilson's basic
- involvement, had been reviewed prior to this interview.
Mr. Wilson had also been previously interviewed by the NRC Special Inquiry Group on August 29, 1979. A summary of that previous interview was available to us prior to this interview. Thus, some of the matters discussed in this interview fol-lowed material from the previous interviews and do not stand alone.
Mr. Wilson was shown a copy of the standard witness notification (Attachment A).
lie was told that the interview was voluntary, notes would be kept and he would be provided a copy of our summary of the interview.
We asked Mr. Wilson to what extent he had been interviewed about the accident.
IIe indicated that, aside from the previous interviews mentioned above, he had been interviewed during the previous week by Bruce Mann of the President's Commission staff. The subject of that interview had been primarily upcoming i
2 recovery efforts rather than the immediate response to the accident. Wilson did not know whether or not a summary of the interview would be prepared by the President's Commission staff.
Mr. Wilson said that his recollection of specific times and names for March 28, 1979 and the next few days was somewhat vague.
Wilson's previous _0IE interview had indicated that he had learned that the reactor coolant pumps were off at about 10 a.m. on March 28 and this had been a point of increased concern for him.
Robert Keaton's previous OIE interview (TMI Tape 297, June 1,1979) had indicated that at about 11:00 a.m. Wilson had told him the accident was more serious than previously thought and instructed Keaten to terminate Keateh's meeting with the engineers who were going to the site and have the engineers leave for the site forthwith. We asked Mr. Wilson what he could recall about the state of knowledge and concern around this time.
Mr. Wilson said that at about that times 1.
Prior to the time in question, Wilson had been aware of the plant trip and the safety injection actuation. He had known it was not a normal plant trip but had not considered the situation alarming.
2.
At the time in question, he had the impression,from conversations he heard with the control room, that natural circulation was not working.
This was a general impression he obtained from what those in the l
control room said, but his recollection was that they did not say specifically whether or not natural circulation was working.
3 He was aware that. the operators had been unable to use the decay heat 3.
removal system for core cooling because reactor coolant system pressure could not be lowered enough.
4.
There was a concern about whether the reactor coolant pumps might have been damaged and thus no longer usable.
He was not aware of discussions with the control room concerning specific 5.
temperatures.
Wilson believed that he had become aware of radiation readings later than 6.
the time in question.
Robert Keaten's previous OIE interview had indicated that, by the afternoon of March 28, the GPUSC personnel in Mountain Lakes had enough temperature informa-tion to be concerned about whether natural circulation was working. We asked Mr. Wilson what he could recall about specific information obtained and analyses performed around this time.
With respect to this question, Mr. Wilson stated the following:
1.
He ceuld not recall any specific temperature information nor any specific analyses or inferences such as determining the amount of superheat in the hot legs or inferring a steam bubble in the core, on March 28.
His impression was that data coming from the control room were rough rather 2.
than specific. He believed most of the information came from Lee Rogers, Joe Iogan and Gary Miller. To his recollection, most conversations yielded interpretations of data by the control room personnel rather than data --
l
4 1.e., statements such as the plant couldn't be depressurized or the reac-tor coolant pumps couldn't be started rather than statements such as the We noted that Mr.
temperature is 575 F or the pressure is 1100 psig.
0 Richard Klingaman, who had been receiving information from the plant at Met-Ed's Reading office, had noted the same tendency in his interview with 30, 1979.)
(Allison and Schierling interview of Klingaman on August us.
Mr. Wilson believed that this situation was (at least in part) a result of reluctance on the part of GPUSC personnel in Mountain Lakes, New Jersey to impose on the operators' time by asking for specific information.
Mr. Wilson recalled specifically that he had not been aware of the contain-3.
ment pressure spike or the core thermocouple readings on March 28.
He had learned of the containment pressure spike the morning of a.
March 30.
He had learned on March 30 or 31 that core thermocouple readings b.
had been off-scale high and that this meant there had been considerable steam in the core with considerable cladding damage.
It also explained the difficulty in running
- reactor coolant pumps.
He did not learn what the specific readings had been until about a week after the accident.
4.
Mr. Wilson did not h tve the impression, on the afternoon of March 28, that the operators' strategy for core cooling for part of the day was to discharge steam from the pressurizer steam space while injecting water into the reactor l
i l
5 coolant system with the HPI pumps, hopefully creating a flow of cooling water through the core. To'his knowledge, this was not the plan and operators were not trained in this cooling technique. Nor had he been aware of a B&W analysis of the feasibility of this method which had been submitted in the TMI Unit 1 FSAR.
He speculated that other GPU engineers such as Gary Broughton would have been aware of any such analysis.
(We had heard of such an analysis but had not reviewed it prior to the inter-view.)
5.
Mr. Wilson believed that, with respect to core damage, there must have been information on radiation readings in Mountain Lakes on the afternoon March 28, which indicated core damage. He believed that the impression or assumption was that there had been some core damage but not an alarming amount.
6.
Mr. Wilson did not recall a firm consensus in Mountain Lakes on March 28 as to what should be done. He considered that there were insufficient data to be confident on the correct course of action.
7.
Wilson did recall that both he and Robert Keaten felt the reactor coolant pumps should be restarted. They were aware that this had already been tried a few times.
We asked Mr. Wilson to describc his knowledge and attitude on the evening of March 28 when he went home.
Mr. Wilson ir.dicated that:
1.
He believed that the accident "had happened and was under control." There was some uncertainty due to lack of data but he believed that the accident itself was over.-
... ~.
6 2.
This belief was based on conversations with the control room and on estab-lishment of a stable core cooling mode by repressurizing the reactor cool-ant system and starting a reactor coolant pump. He didn't recall that radiation readings or off-site release information had formed any part of the basis.
3.
With respect to radiation, he had known that there had been off-site re-leases, that respirators had been worn in the control room, that there were radiation problems in the plant, that there had been high vent stack readings and a "very, very high" containment dczne radiation monitor reading.
4.
With respect to the containment dome radiation monitor readings, he recalled discussions of the reading but did not recall anybody dismissing the reading as erroneous or non-meaningful. He did not make any specific inferences from it such as percentage of core damage or percentage of a design basis release. He did believe that the reading was due to fission product gases.
We asked Mr. Wilson why he had called Mr. Arnold on the evening of March 29 and told Arnold that the situation was more serious than previously thought. In amplification of his previous response (see item 3, Allison and.ichierling Sum-mary of previous interview with Wilson).
Mr. Wilson stated the following:
1.
Mr. Wilson reiterated that this judgment had not been based on hard facts such as knowledge of the containment pressure spike on March 28, or the core thermocouple readings of March 28, or the gas Fnbble in the reactor coolant system on March 28 and March 29.
7 2.
In addition to factors mentioned in the previous interview, Mr. Wilson said:
a.
It was difficult to get specific information from the plant staff.
b.
George Kunder had requested that 2-3 engineers come to the control rooti.
Wilson believed this request was made because Kunder was coming to the conclusion that the plant was not behaving and cooling down properly.
c.
Wilson was noticing a lack of real progress -- i.e., the plant was,not cooling down.
d.
Wilson believed that Gary Miller.had communicated a feeling that the plant was not behaving properly in their meetings when discussing interviews of plant operators.
e.
Wilson had looked at some curves on plant data that had been gathered by Richard Lenz. He believc.l these may have indi-cated steam in the core on March 28.
We asked Mr. Wilson if he had been aware of any plans to vent the containment and/or convert the spent fuel pool into a huge charcoal filter during the first days after the accident. He said no.
We asked Mr. Wilson why GPU engineers had been placed on watch in the control room (on Friday, March 29) and what they did. He indicated that:
~
---4 8
He needed It was dif.ficult to get much data from the control room.
1.
such data for the tasks his staff (the Technical Support Group) was working on.
GPU engineers were beginning to be involved in procedure approval and 2.
they needed to obtain an accurate picture of plant status from the control room.
The GPU engineers in the control room sometimes spent considerable 3.
For example, considerable effort was time determining plant status.
expended in determining tha existing letdown system valve lineup.
GPU engineers in the control room assisted a great deal in operational 4.
planning and made a significant contribution to data reduction.
The GPU engineers in the control room were representing the Technical 5.
Support Group (Wilson in charge) and would be expected to notify that Later (the second week),
group when new situations or problems arose.
they were familiar with the long-range operational plans and were able to provide input to the plant operators on these plans.
(The planning was related to the plant in its actual condition, with a damaged core.
This condition was outside the scope of previous training or procedures.)
The mechanics Their purpose was not to watch over the plant operators.
of operating the plant were left to the operators (and the Plant Opera-tions Group with Jack Herbein in charge). The engineers supervising this effort were such people as Jack Devine, Ed Wallace, Gary Broughton, Mike Morrell, Ted Fischer and Tom crimmens. They were specifically l
r---
o 9
familiar with previous accident (safety) analyses and had reactor oper-ating experience (in different reactors, such as Naval reactors if not in large connercial reactors). In Wilson's perception, these engineers were able to relate well to the plant operators and the arrangement worked well.
Wilson felt that early in the recovery there had been two primary competing goals in plant operation. In the first place, it was very important to degas the reactor coolant system as rapidly as possible - i.e., by all means available. Although a reactor coolant pump was running, it was not known how long a' pump could be kept running. Also, the reactor coolant system pressure control might be lost - i.e., the plant' pressure might drop.
Thus, in the long term, natural circulation cooling would be necessary and degassing would be necessary for natural circulation to work. On the other hand, one means of degassing, maximizing letdown flow, increased the rate of radioactivity release. Of course, one could not expect each control room operator to resolve this conflict for himself on his shift. This type of conflict had to be resolved in the support groups.
Wilson said that working relationships with NRC personnel had been good. He didn't discern any differentiation between people from various NRC organiza-tions (NRR, OIE, RES, etc.). Generally, the NRC people appeared to be " working the problem" -- i.e., to understand the options and determine the best course of action - rather than only criticizing plans proposed by others.
10 These good working relationships extended, to his knowledge, up through Richard Vollmer and Victor Stello. From above that level, it appeared that he sometimes got statements for which he could not find a reason.
For example, with respect to the possibility of a hydrogen bubble explosion inside the reactor coolant system, he asked some people to inquire as to f
the reasons for the NRC position but these people never found any answers from NRC.
(Wilson did not know whether this was due to lack of time to follow through or what.) This problem was then turned over to the Indus-try Advisory Group which resolved the conflict.
We asked Wilson about his participation and knowledge of the TMI-2 recovery support provided by organizations and individuals outside the GPU/ Met-Ed organization. He provided the information below on the following subjects:
A.
Support Provided by Babcock & Wilcox On Wednesday, March 28, Wilson had no contact with Babcock & Wilcox in Lynchburg, Va.
He recalled that R. Arnold, GPUSC, had some telecons with Lynchburg but was not aware of any details. From his viewpoint, he would not characterize the Lynchburg activities on Wednesday as an l
extensive technical effort. Wilson did communicate with the TMI Site Representative of B&W, L. Rogers, who was in the Unit 2 Control Room.
In Wilson's opinion, Rogers did not perform the functions of a Site Representative on March 28 but was an active member of the TMI-2 re-sponse team and assisted G. Miller, Station Superintendent. Wils'on recalled that Rogers did not recommend the restart of the RCS pumps in the morning of March 28.
m m
,o, t
11 On Friday, March 30, an open telephone line had been established between the Control Room and B&W in Lynchburg, which was manned 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> a day The technical contact in Lynchburg was with D. Roy and L.
on both ends.
On about Sunday, April 1, J. MacMillan, Vice President of B&W Kosiba.
Nuclear Power Generation Division, came to the site (Trailer City) and was the management contact between GPU and B&W.
The technical support provided by B&W Lynchburg became evident on Friday, March 30, and was substantial from then on.
Typical examples were the following:
I identifying and providing miscellaneous facts
- and figures about the NSS system, specifications
/
and qualifications of components
(
evaluation and planning of short term and long range activities such as diffe;:ent methods for degassification of the RCS, natural circulation I
approaches, supportive analyses preparation of step-by-step procedures for po-tential emergencies and special operating procedures continuous availability of B&W staff familiar with details of the TMI-2 plant for technical consultation by the operating staff in the Control Room
r-s.
12 In many cases, the B&W staff in Lynchburg initiated evaluations of the plant conditions on their own and postulated problems without request from GPU and then advised the Technical Support Group (TWG) under Wilson.
B.
Support Provided by Other Organizations Most of the support to the 'IWG was provided by the NSS vendor and archi-tect-engineer (AE) organizations.
In many cases, these organizations had a representative (or team) at the site which then provided contact with the appropriate elements of their entire organization.
In addition, operators and operations staff personnel from utility companies provided support to the Met-Ed operations staff in the TMI-2 conhol room, for ex-ample, in the preparation and review of procedures. Wilson recalled that Arkansas Power & Light Co. (Arkansas 1 and 2), Florida Power Corp.
(Crystal River 3), Toledo Edison Co. (Davis-Besse 1), and Duke Power Co.
(Oconee 1, 2, 3) provided operations staff support.
While all organizations provided valuable and expert assistance, Wilson stated that individuals with experience or familiarity of the B&W design j
were most effective. Differences in NSSS and A-E terminology for some systems required that some individuals went through a learning period before becoming as productive as others. Procedures prepared by indivi-duals without B&W familiarity had to be reviewed for applicability to the TMI-2 design. Wilson stated that N. Pope, Station Superintendent for Oconee 1, 2, and 3, of Duke Power Co., provided superb assistance from the time of his arrival. He was placed in charge of one of the shifts of the Tech Support for Shift operations under Wilson.
^
y l
13 i
C.
Support Provided by Burns & Roe Wilson stated that the Plant Modifications Group (PMG) under W. Cobean, Vice-President of Burns & Roe, became effective on about Saturday or Sunday. The assignment of Cobean to this management position was a logical choice. This not only made any required manpower from the Burns
& Roe organization available via a direct management channel but also assured immediate effectiveness because Burns & Roe was the A-E for TMI-2.
Under cobean's direction they took charge of the design and engineer-ing of plant modifications and the procurement. of equipment.
At times there were approximately 200 people from Burns & Roe involved in the TMI-2 support activities.
D.
Industry Advisory Group l
Wilson was not involved in the development of the concept or in the organ-ization of the Industry Advisory Group (IAG). The first members of the group assembled on Sunday, April 1.
Their function was to evaluate broad issues and specific technical concerns, and make recommendations to GPU.
Typical examples were:
1 development of methods to achieve long term core cooling evaluation of the size and explosion potential of the
" hydrogen bubble" application of noise analysis identification of core status independent review of methods of venting the waste gas decay tank back of the containment
14 In these efforts, the IAG provided a valuable assistance. This group of experts was able to look at the issues in a rational and non-panic mode (as compared to the operations organization), and was good for " bouncing of questions and concerns and for a second guess of proposed methods."
Towards the end of April, the IAG became more involved in the development and review of procedures which frequently were qdte plant specific.
Wilson feels that in this regard the IAG was not as effective because of non-familiarity with the plant. He cited as an example the criteria for establishing natural circulation. While a procedure had been established very early (about Sunday) and was available in the control room to go to natural circulation (in the case of RCS pump failure), much time was re-quired to establish the criteria to be used to verify natural circulation.
E.
Management Support Wilson stated that the inclusion of executives from other utilities in the TMI-2 recovery organization did not present any problems but provided a balanced opinion. It also made available through those management channels their entire organization in whatever capacity was required.
With respect to NRC takeover, Wilson was asked if an NRC takeover during an ac-cident was a good idea. Wilson felt that NRC people could not come into a plant and learn about specific aspects of that plant in a timely enough manner. Even a resident inspector would not be knowledgeable enough to operate a plant. Wilson had not noticed any great differences in the competence of the better industry personnel assisting after the accident and the competence of the better NRC person-nel. Even if the NRC should maintain a large staff of people to be qualified on
m 15 various plants in the event takeover is needed (a truly large undertaking) it was difficult to see how they could be as familiar with the plant, its organ-j l
Furthermore,
{
ization and its resources as those who work at the plant regularly.
t Wilson believed that the critical time of an accident is the first few hours or perhaps the whole first day; so, by the time the NRC got into place behind the controls, most of the key events would have already occurred.
With respect to the need for immediate communication of hard data, Wilson went on to say that during an accident it is important to get hard data out to the right people for analysis. Emphasis on the immediate communication of important This commun-data would be more beneficial than a plan calling for NRC takeover.
ication system must be independent of the control room operators who are too busy dealing with the plant.
With respect to the availability of information, Wilson stated that every plant should have several packets of basic data to assist those providing technical support in response to an accident. Necessary, fundamental calculations can often be done without the aid of a computer if summarized data are available on Charts should be available that such things as plant and equipment constants.
show what critical instrument readings mean; for example, it is helpful to know After what various percentages of full scale steam generator level readings mean.
l the accident, it was not uncommon to spend an entire shift looking for basic data l
such as this in order to perform a calculation. Piping and instrumentation, draw-ing books should be updated and available as well as instrumentation and control These books are used frequently.
block diagrams and electrical one-line drawings.
Simple sketches of systems showing dimensions and elevations should also be availa-One ble in the packets. Also, a 10-20 person tech support staff is inadequate.
needs more people to provide coverage and assimilate the large influx (200+) of technical people during accident response.
16 d
(Note: This is similar to a recommendation made this summer by Myer Bender of the ACRS. Such basic information was envisioned by Bender as being useful to the control room operators and persons on the scene during normal operations as well as immediately following an accident.)
With respect to the possibility of hydrogen explosion, they had felt that hydro-gen was present in the reactor vessel, but did not believe it was explosive (inside the reactor coolant system); the explosion possibility was raised by the NRC.
GPU asked severa; sources to examine the problem but they were always told there was no chance of an explosion because there was no oxygen. GPU ad-vised the NRC of their evaluation of the hydrogen issue on Saturday; for some days the issue was hashed over many times by Denton, Vollmer, and Stello.
The GPU people felt real concern over the waste gas storage tanks since both H and O might be present in this tank.
2 2
By Sunday the H bubble issue was passed from GPU to IAG who agreed and resolved 2
the in-vessel explosion potential issue as a non-problem.
We asked Wilson if there were anything he felt we should know or wanted to bring to our attention regarding the accident aside from the matters we had already discussed.
I l
Wilson said that immediately following the accident the NRC and others had been well aware that a clock was running and expeditous action was necessary for a
safety.
l
17 flowever, at the -,e of this interview, expeditious action to complete the recovery seemed ;-ossible.
lie noted that much of this was out of the hands of the NRC direct., but came in other arenas such as the federal courts.
Regardless, these ;elays could constitute a safety problem and it did not appear that the s. ious investigations were addressing it.
lie noted that the urrent plant condition was not a desirable or safe " final solution." The c :.s still a clock running, although not as fast as after the accident.
Sc:. thing may go wrong in the future. Therefore, the NRC and others should be re interested in an expeditious cleanup, removal of the core and placerer-of the radioactive debris in safe long term storage con-dition.
9 l
..