ML20072L251
| ML20072L251 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 07/01/1983 |
| From: | Doroshow J THREE MILE ISLAND ALERT |
| To: | |
| References | |
| NUDOCS 8307130228 | |
| Download: ML20072L251 (64) | |
Text
.i
/
s' u
UNITED STATES OF AMERICA
[!Ne NUCLEAR REGULATORY COMMISSION'c N
/.,.'jf
/(z, t:scu l.,i wm BEFORE THE COMMISSION l.'
JUL 1 E B3 W.
5 In the Matter of
)
\\
(""'Qs, [
METROPOLITAN EDISON COMPANY Docket No. 50- 89 (Three Mile Island Nuclear
)
Station, Unit No. 1)
)
e TMIA INTERIM COMMENTS ON B&W TRIAL RECORD c
t Joanne Doroshow Louise Bradford
' on behalf of:
Three Mile Island Alert
\\
Intervenors July 1, 1983 1
8307130220 830701 PDR ADOCK 05000289 i.
O PDR
~
i t
s TABLE OF CONTENTS I.
INTRODUCTION..............................................
1 II. SUBSTANTIVE COMMENTS.....................................
4 A.
The B&W record supports the argument that the Licensee is inherently incapable of learning from past mistakes or correcting recognized problems...........................
4,
- 1. The April 23 transient -- a precursor event.....
6 2.
PORV problems..................................
11
- 3. Condensate Folishers...........................
19
- 4. Licensee's response to the accident............
20 B.The B&W record, read in light of other new information which has come to light reflecting on Licensee's management, fundamentally undermines the credibility of the ASLB decision, and the PIDs can not be used to lawfully justify restart of TMI-l.......................................
37
- 1. Maintenance..........
7.'.......................
37
- a. Budget cuts and understaffing..................
39
- b. Repair parts -- materials management...........
42 c.
Paperwork......................................
44 d.
Overtime.......................................
46 e.
The general disorgani,zation and poor performance of the maintenance department under the supervision of current Unit 1 maintenance supervisor Shovlin.. 46 i
2.
Training.......................................
47
- 3. Management structure...........................
55 4.
Managers.......................................
59
_, III.
CONCLUSION.............................................
61
+
e
o UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE COMMISSION In the Matter of
)
)
METROPOLITAN EDISON COMPANY
)
Docket No. 50-289
)
(Three Mile Island Nuclear
)*
Station, Unit No. 1)
)
TMIA INTERIM COMMENTS ON B&W TRIAL RECORD I.
INTRODUCTION In what has become a race against time, TMIA has attempted a painstaking review of the record dev, eloped in the recently settled litigation between the Licensee and B&W.
This review is far from complete.
However, since the Commission may be set on making a restart decision in the near future, TMIA submits these comments now so that the Commissioners may have at least some indication as to the seriousness and materiality of this new information.
The B&W record, particularly when read in light of the most'recent RHR and BETA management audits and allegations by
,s TMI-2 clean-up "whistleblowers," dramatically illustrates not only that Licensee is unfit to operate TMI-1, but that the Atomic Safety and Licensing Board (ASLB) decision was wrong in finding Licensee's current competence adequate, a decision which fails to face evidence which would' plainly mandate the opposite c
conclusion.
i
_j
O.
The scope of the restart hearings has been limited to the l
question of whether this company can safely operate TMI-1.
This 1
is hardly a narrow issue.
Yet a reveiw of the B&W record firmly supports the conclusion that during the main management hearings, the ASLB either so unreasonably restricted the scope of admissible evidence, or failed to meaningfully examine the Board
/
Issues mandated for considerati'on by the' Commission's Order of March 6, 1980, CLI-80-5, 11 NRC 408, (1980), that any concrete information which impugned the competence and integrity of the Licensee management was entirely shielded from consideration by the Board.
Much of this evidence concerns problems which were deeply rooted at TMI before the accident, in fact contributed to the accident, and amazingly enough have.gontinued through the restart I
hearings to the present day.
Yet any knowledge of the true history and current state of affairs at TMI would still be t
generally unknown today had it not been for the B&W trial, revelation of the BETA and RHR managment audits produced earlier this year, and information revealed by TMI-2 "whistleblowers" Richard Parks, Larry King, Ed Gischel, and Joyce Wenger. This "new evidence" bears directly on specific management issues which were either discounted by, entirely ignored by, or not even part of the discussion before the ASLB.
Only when viewed together can it be demonstrated how deeply entrenched some problems are at this company, how these problems have seriously endangered the pubic and continue to do so, how early some problems were c
recognized'and how unresponsive management has been to them, and
f e
how the company takes corrective actions only when forced to do so while under intense scrutiny (which will soor.
- nd) and at the point where failure to take action literally threatens their license.
(This last point had been a recurrent theme throughout the restart hearings.
See, TMIA's Appeal Board brief on exceptions to the management PIDs, pp. 61-64; and, most recently, GPU President Herman Dieckamp's letters'to the Commission and Governor Thornburgh dated June 10, where the comqany promises to finally remove people clearly unfit to work in safety-related por,itions at TMI, some of whom the Commonwealth has been demanding be removed since early 1982, only in response to the Governor's strongly worded letter to the Commissioners opposing restart on these and other grounds).
Not surprisingly, the Licensee maintains that the ASLB decision thoroughly and properly resolved management issues, t
based upon the restart hearing record alone.
This record, however, as well as all prior investigations into the accident, contains a substantially different body of evidence than that contained in the B&W record on many of the same issues.
One need only consider the enormous financial stake which B&W had in discovering evidence which could prove this company's s
incompetence, and their virtually unlimited financial resources expended over the course of two years in developing this record, to realize why they were so successful at this task.
B&W's
~~ ability to' prove the company's culpability for the accident,
~
l eventually forcing settlement of the suit, illustrates not only that B&W did a thoroughly credible [ob in discovering important I
f i
4.
evidence relating to Licensee's incompetence, but that not until this litigation were the real plant conditions leading up to the accident actually known, including evidence relating directly to management culpability.
It would be unconscionable for this Commmission not to insist that the relevant evidence in this record be carefully examined on the record of the restart prbceedings, in the context of a full adjudicatory hearing, and that the credibility of the ASLB decision be fully explored,before restart.
The Commission is well aware of the deep public distrust which the people around TMI have had for this process, which they and the entire country have assiduously watched for four years.
A hasty decision will solidify that distrust forever.
At the very least, the Commission has a responsibility not to do this to a public still traumatized by the worst commercial nuclear accident I
in this nation's history -- the very people it has been entrusted i
to protect.
II. SUBSTANTIVE COMMENTS A. The B&W record supports th'e argument that the Licensee is inherently incapable of learing from past mistakes or correcting recognized problems.
N During the reopened management hearings examining the cheating incidents in November / December 1981, it was discovered that extremely lax exam testing procedures allowed significant cheating to occur on company and NRC exams from at least the time of the accident, through the 1981 NRC licensing exams.
The'ASLB
+
defined the problem as a failure to extend quality control (QC)
[
1 l procedures to exam testing which would insure exam integrity, PID 52401, ar.d to Licensee's " naivete" which would be corrected by new exam testing procedures. PID $2396.
Yet at the time the cheating was occurring, this " naive" Licensee was developing and enthusiastically presenting to the ASLB a revised training and testing program in response to severe criticism Licensee's
/
training department had received ~as a res' ult of the various investigations into the accident -- a specific subject of concern discussed in the Commission's August 9, 1979 Order, CLI-79-8, 10 NRC 141 (1979).
At the time of its alleged " naivete," Licensee
- l l
management was also already aware of the 1979 cheating incident involving VV and O, and of Station Superintendent Gary Miller's criticism of training as it related to VV's cheating.
- See, TMIA Ex. 71.
3 The cheating incidents illustrate some very fundamental I
principles about Licensee's competence and integrity which the i
B&W record clearly substantiate.
First, as the ASLB correctly points out "if the Licensee does not itself exercise the requisite quality control, quality assurance, and feed-back mechanisms to assure high-quality trainir.g and testing, it is l
.s oeyond th'e power of regulators and regulations to put an l
appropriate program in place."
PID $2327.
This principle.of r
course extends far beyond just training and testing, to all I
aspects of nuclear regulation.
Second, this Licensee is fundamentally incapable of insuring the requisite quality control, quality assurance, and feed-back mechanisms necessary to
+
insure safe operation of TMI-1.
To assume that a "new management
f s.
structure" or "new procedures" provide reasonable assurance that Licensee can now objectively police itself is to ignore its long history of failure to take independent action to correct deficiencies unless under intense scrutiny, or its refusal to recognize deficiencies even when under such scrutiny.
This failure was a central theme in B&W's case, and the trial record which we have reviewed so far is replet'e with examples.
Several' stand out in terms of their contributions to escalation of the accident, and in terms of management's direct and indirect involvement.
- 1. The April 23 transient -- a precursor event It is well known that part of the operators' failure on the morning of March 28, 1979 to diagnose the LOCA, and thus take appropriate steps to control it, w a,s the failure of either emergency procedures or training to prepare them for situations where indicators showed that pressure was decreasing while i
pressurizer level was high.
B&W 4005, p. 110 shows at least four or more transients involving loss of feedwater in the year before the accident.
Arnold at Tr. 1587.
At least thr'ee of these transients, April 23, November 7, and December 2, 1978 were the result of
\\
overcooling where the pressure level dropped so low that HPI was initiated, id. at 1598, making them reportable events to the NRC.
The April 23 event was perhaps the most significant of these events.
It not only forced a four month shut-down of the plant, 6
- i..
but during this transient, the system experienced the condition,
b
. at the time plainly not contemplated by operator emergency procedures or training, where pressurizer level increased while pressure dropped, B&W 4059.
But because Licensee presumably failed to attach significance to this aspect of the transient, (see, e.g.,
Keaton dep. at 225), it did not bother to modify emergency procedures or training to instruct operators what to do in the event the system experie'nced the' condition again.
- Thus, on March 28, 1979, when the strip charts again showed pressurizer level high while pressure was dropping, the operators were unable to diagnose the LOCA resulting in steps which led to core damage.
It is simply not credible to excuse management for this negligence.
At the time, management considered the event so significant that it had GPU Service Corporation form a special task force to investigate the incident, headed by Robert W.
Keaton, (after Met Ed submitted its own report on May 4, 1979).
4 See B&W Ex. 186; Keaton dep. at 96.
Even Arnold testified that both he and Herbein considered the event very significant.
Arnold Tr. at 1536.
Yet neither Arnold nor Herbein seemed to have learned the most fundamental lesson from the event.
In fact, Herbein stated in his trial depostion that he was unaware of anyone at Met Ed, including x
himself, who knew before the accident that saturated conditions in the reactor coolant system were in fact the key to a LOCA or that a pressurizer level increase could result from the presence of saturation in the' reactor coolant system.
Herbein dep. at 101, 106.
And Arnold still maintains in testimony that "[alny
+
time we were identifying a loss of coolant accident or had in
I l
~ mind a loss of coolant accident,....the pressurizer level and pressure would both trend together."
Arnold at Tr. 1459-1461.
This concept was accorded the title "the Rock of Gibraltar" by GPU attorneys at the B&W trial.
Id. at 1461.
Thus, not only did the President of the company fail to learn the most important lesson from the April 23 transient, which could have prevented escalation of the accident, but he failed to learn the very same' lesson from the accident itself. This perhaps says more about Licensee's fundamental inability to learn from the past and respond with appropriate corrective action than any other single incident in this trial record.
Further, during the April 23 event, operators reduced HPI after initiation, Keaton dep. at 141, precisely the same action they took during the 1979 accident.rgsulting in core damage.
Contrary to statements in his deposition, id. at 137, Keaton indicated in his notes during task force discussions that
" operator doesn't really have good knowledge of what is happening
- seems to react too quickly according to procedures rather than thinking," B&W Ex. 337 at 3819.
Yet Keaton can not recall what action was taken as a result of th'is finding, id. at 177, nor can he recall any concern by the task force whether operators x
reduced HPI for correct reasons, or any concern about operators' ability to identify and correctly distinguish between a steam line break or a LOCA. Id. at 186.
~rn~ addition, during the April 23 transient as during the 1979 accident, operators were extremely confused by the number of alarms which went of f at the time ofthe event.
While'GPU's i
I
. 9
' internal accident investigation, headed by Robert Keaton, concluded that during the accident, control room operators were besieged with irrelevant alarms, B&W 356, Keaton claims to have no recollection of the similar concern of operators on April 23, id. at 128, or of a May 3, 1978 memo to management written by operator Frederick who was on duty at the time of the April 23 event, in which he stated, The alarm system in the control room is so poorly designed that it contributes little analysis of a casualty.
The other operators and myself have several suggestions on how to improve our alarm system.
Perhaps we can discuss them sometime, preferably before the system as it is causes severe problems.
B&W Ex. 264. Keaton dep. at 128, 129.
It does appear that some action was at least attempted to deal with the alarm problem.
Specifically, an " alarm window correction program" was recommended','""with a priority to escalate."
B&W 79.
However, the program was explicitly disapproved by Met Ed corporate management, B&W 767, and thus never put into effect, Significantly, Superintendent of Maintenance at TMI, Daniel Shovlin, claims to have been entirely unawar.e of the operator's confusion due to the number of alarms, and of the program to correct the problem, despite the fact that s
the program, which he has "no idea whatever happened to," but could have permitted an easier diagnosis of the accident,'was his express responsibility. Shovlin dep. at 62, 63.
During the ASLB hearings, TMIA introduced as an example of an item of deferred maintenance TM1A Ex. 40, which was a four year old maintenance work request to repair a spurious alarm at o
TMI-1.
The work was assigned a " priority 3" -- the lowest
priority which, according to former Station Superintendent Gary Miller, virtually guaranteed that it would never be fixed.
B&W 360 at 22.
See, TMIA management findings !16.
The ASLB, while refusing to define the four year delay as improper, commented that "the priority on fixing it need not be high"...
- however, "a delay of almost four years seems long in view of the fact that it should be fixed evsntually.N PID 5299.
The Board's view here illustrates the narrowness of its approach in defining examples of safety-related maintenance deferral, particularly in light of the April 23 event and the accident scenario, and GPU's own accident investigation.
Despite the resources invested by the company in establishing a task force to investigate this incident, culminating in production of a detailed report, B&W 186, they apparently never even bothered to send a copy to the NRC or i
report the tech spec violations.
Keaton dep. 148.
And e
Licensee's blase attitude toward the necessity for proceeding to the next step with corrective actions can perhaps best be illustrated by Maintenance Superintendent, and current TMI-l Maintenance Supervisor Shovlin's virtually complete lack of recall about the entire April 23 event.
See, generally, Shovlin dep. at 58 et sea.
Indeed, it is readily apparent that if management had attached appropriate significance to a number of actions that day, the accident could have been avoided altogether.
+
1 i
- _. _ _ ~
I
. 2.
PORV problems i
As the Commission is well aware, the PORV was the valve l
l which stuck open during the March 1979 LOCA.
In examining the PORV's history at TMI-2, the B&W record provides in dramatic detail a story of longstanding and well-known problems with th-PORV some of which greatly disturbed the plant operators, and a f
managment which took virtually'no steps'to prevent or correct recognized problems, resulting eventually in a leaking valve and falsification of leak rate data by operators who felt driven by management to keep the plant operating no matter the risks to the public health and safety.
Hartman dep. at 85, 198, 148.
The PORV in place at the time of the Unit-2 accident had a history of problems.
Seiglitz at 5765.
In 1974, it was transferred to an operating Unit I,,where it remained until late 1975 or early 1976.
While at Unit 1, the PORV operated at 250 t
volts, 25 volts more than at Unit 2, and GPU's own post-accident investigation of the PORV determined that the PORV may have suffered some damage as a result of the different voltage in solenoids -- a fact never believed significant enough for discussion at " plan of the day" (POD) meetings where day to day operati6n and maintenance problems and plans were discussed.
Seiglitz at Tr. 5768.
In July 1975, Lee Rodgers of B&W sent a letter to Met Ed management outlining problems discovered with the PORV and suggesting a specific preventive maintenance program.
B&W 881.
There exists no evidence that any such preventive maintenance h
t-program was ever instituted.
Seiglitz.at Tr. 5786; B&W 4036.
F e,
In September, 1977, the PORV had leakage problems and was removed for in-house repairs.
B&W 4033; Sieglitz at 5769. On March 28, 1978, the PORV stuck open for four minutes due to an electrical failure at the plant, causing HPI initiation and a reactor trip.
See B&W Ex. 4005, p,
110. Fiske Tr. 1601.
- Also, according to Gary Miller, there was general pre-accident
/
knowledge of trips at TMI-2 whe*re the PORV opened and closed.
In late March or early April, 1978, an indicator light was to be installed in the Unit 2 control room to show demand indication to the PORV solenoid.
See Shovlin dep. at 56.
In the fall of 1978, an incident occurred where the PORV light was on but the valve stayed shut, an event of which GPUN President Arnold was unaware.
Arnold at Tr. 1640.
In December 1978, a request was made for a better indi' cation of PORV position than the solenoid light, and Seiglitz signed off on a priority 1 s
[
change modification to accomplish this, asking that operators be o
trained on the changes made.
Seiglitz at Tr. 5799, 5802-3, B&W 767.
This request was denied by corporate engineering two weeks before the accident, and no specific instructions were given to the training department about operators' concerns as to the lack of accurate indicators of PORV position. Seiglitz at Tr. 5806.
N Between April 10 and April 19, 1978, the highest PORV tailpipe thermocouple reading was 132.9*. B&W 4040; Seiglitz at Tr. 5846. At some point close in time to the April 23 transient, possibly in the 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> before the event, the PORV cycled 50
. times. B&W 337 at 3820.
This could have had an adverse effect on
+
PORV reliability.
See, discussion, Keaton dep. at 174.
As a r
.r
_.c.i..----
---,,,w,c
---,----m-y..y
--~m---+w, w - --
y
-w w
- result of the transient, the plant was shut down for four months, during which time no maintenance was done to determine if the PORV had sustained damage as a result of this challenge, Seigl.tz claiming to'have been entirely unaware of it.
Seiglitz at Tr.
5781-83.
When the plant returned to operation after four months PORV temperatures began reading 180* -- 50* higher than permitted by procedures.
Emergency procedures in fact required that upon obtaining such high temperatures in the discharge line, the block valve was to be closed to test for valve leakage. The company failed to comply with these procedures, for which they were cited by the NRC in the October 25, 1979 Notice of Violation.
Seiglitz, Supervisor of Unit 2 maintenance, claims to have been so ignorant of the situation that he did not know that i
180* was even an elevated temperature, claiming it was not his c
job to know.such details.
Seiglitz at Tr. 5809, 10.
This is an extraordinary admission, if believable, by the man in charge of Unit 2 maintenance -- reliance upon whom Superintendent of Maintenance and current Supervisor of Maintenance at Unit 1, Daniel Shovlin, insists justified his own ignorance about plant condi ti*cn s.
Shovlin dep at. 195.
Further, Seiglitz says he never thought to ask if 180*, which appeared on his daily maintenance log,' Seidlitz at 5815, could signify leakage, because no leak rate was being monitored. Id. at 5811.
According to Seiglitz, these high temperatures were thus never
- discussed at " plan of the day" (POD) meetings which Seiglitz attended daily.
Id. at 5817, 5844.
+
1 l
- - -, - - -,, - - - - - - - -. - - -, - - - _ _., - -. = -.
r Thus, the plant continued to operate without performance of the required diagnostic testing.
And hardly coincidentally, it was at this time that the operators began falsifying leak rate data, running leak rate tests repeatedly and throwing away " bad" tests to get " good" ones, violating its license to keep the plant operating.
See, Faegre & Benson, Results.of Faegre & Benson e
Investigation of Allegations by Harold 'd.
Hartman Jr. Concerning Three Mile Island Unit 2, Vol 1 at 20.
According to Hartman, th'e operators suspected leakage from the pressurizer relief valves.
Hartman dep, at 85, 120.
Every shift supervisor and shift foreman knew leak rates were being falsified through the time of the accident, stating, "I thought that it was just the fact that everyone knew that these leak rates were hard to get, and that we had to take devious means to get soma sometimes so we could stay operating."
Id. at 148.
High PORV readings continued into the January 17 through i
January 31, 1979 shut-down, during which time no work was done to determine if any of the three valves above the pressurizer were leaking.
Zewe at Tr. 2250-53.
When the plant reopened, PORV temperatures jumped to 190*, and the code safeties' temperatures jumped to 200* from a previous 110 to 120".
Seiglitz at Tr.
\\
5824.
Control room operators other than Hartman admit to at least knowing of the elevated temperatures, and that the PORV had been reading high since 1978.
Zewe at Tr. 2249; Frederick at B&W Ex. 5007BB.
l l
At the February 1 POD meeting, the identified leak rate was determined to be up, and Seiglit$ claims that he beli'eved
r o l
' this was caused by the code safeties, not the PORV.
However, he also has no recollection as to why suddenly after a two week shut-down, perfectly good code safeties would suddenly be leaking with high temperatures.
Seiglitz at Tr. 5823, 5824.
seiglitz recalls no discussion of what may have gone wrong during the two week shut down to have caused this to happen.
Id. at 5822.
But Seiglitz supports his theory by' coming to the
" intuitive" conclusion at the February 2 POD meeting, without performing any calculations or analyses, that the high temperature readings were the result of the accumulation of convection, conduction, and radiant heating, and that the PORV readings were primarily the result of convection type heating --
heat rising up through the air off the top of the pressurizer --
and thus not the result of leakage., Id. at 5829-31.
He further claims this was discussed at the POD meetings. Id.
f However, the rest of the record provides no support for Sieglitz's position.
(See also, Seiglitz at Tr. 5906.)
In fact, in Met Ed's official response, dated December 5, 1979, to the NRC's Notice of Violation dealing with the item of noncompliance for failure to comply with the pressurizer system failure procedure, Licensee maintains that the high PORV N
temperature readings were likely the result of conductive heating, not convection heating. Id. at 5833-4. B&W Ex. 707.
And Superintendent of TMI-2 Logan, in attendance at the POD
~' meetings, has stated
- that he could not tell which of the three valves were leaking.
Tr. 5857-60.
Further, Gary Miller who received the daily POD reports, statbd in his B&W depodition that
{
l
(
s
. - he thought the PORV was leaking.
Tr. 5853.
And operator Frederick heard that some supervisors and engineers thought it was one of the code safeties, some thought it was the PORV.
Frederick at Tr. 3539.
Still, neither Seiglitz nor any other management official attending the POD meetings recommended complying with procedures by shutting the block valve, rationaliz,ing that it was unnecessary because, according to Seiglitz, the PORV could not have been leaking, and also because the valve might stick in a closed position.
Seiglitz at Tr. 5877.
And meanwhile, as
~
leakage continued to increase from the relief valves, operators were having increasing difficulty in obtaining a " good" leak rate, and were " fudging" them continuously to keep the plant open.
Faegre and Benson at 17.
The leakage may itself have
.s significantly deteriorated the valve.
Seiglitz at 5882.
Yet as Unit 2 Supervisor Logan stated,
"[w]e made a management i
decision -- as long as it stayed within the tech specs, they would live with the leak rates until cold shut down."
Tr. 5860 (emphasis added).
Further, in a January 25, 1979 letter Dresser Industries, the PORV's vendor, informed Met Ed management of design changes
\\
on the PORV which could be incorporated on existing valves at TMI-2. B&W 4037.
Receiving no response from Met Ed, Dresser sent a telegram to TMI asking why no response had been received.
That_teleg. ram was sent on March 27, 1979.
B&W 4038. (Seiglitz stated that he did not receive the letter until the fall of 1979, at which time it was fowarded to Met Ed's legal and engineering
. departments, and he says that he never received the telegram at i
all.
See, Seiglitz, Tr. 5787-90.)
These accumulated instances of management resistance to diagnose the high temperature readings, and take diagnostic and corrective action with regard to the PORV of course raises the question of what exactly did the POD attendees know about the leak rate falsification which th,ey failed to rev'eal in trial testimony?
The obvious question is had management already decided to tolerate the leaks until Unit 1 came back on line, so as not to cost the company $500,000 per day in replacement power?
Fiske Tr. 114.
It is hard to imagine that by this point, j
management did not know, in the face of all of these symptoms indicating that leakage was abnormally high, that the operators were deliberately falsifying leak rates to keep them within the license limits, particularly since oniy management could benefit from such a scheme, i
But even without reaching this issue, Licensee's own e
consultant investigation in Hartman's allegations defines another
" fundamental" problem which in itself results in nearly as damning a conclusion against Licensee's management:
either 1) the repeated " bad" leak rate results suggested that "real"
\\
unident'ified leakage exceeded 1 gpm and therefore the plant should, as a matter of safety, be shut down; or 2) the system that had been devised to measure unidentified leakage was
_.. unreliable and inadequate for the task and should be overhauled to the the job it was intended to do.
Faegre & Benson, supra, at 35.
+
Even when viewed in a light most favorable to the Licensee, its own consultant report, issued September 17, 1980, shows that management failed to respond to the obvious problem that its system for identifying leaks in the reactor coolant boundary was not working, and at that point, management @new they were operatingunderapotentla'llyunsafecondithon. 'Yet they continued to allow this unsafe condition.,to exist so that on.
March 28, 1979, operators had bscome so condictioned to the elevated temperatures that "the abnormality was obscured or rationalized away resulting in a delayed closure of the isolation valve."
NRC's Notice of. Violation, p.- 2.
See Keaton dep. at 506.
Moreover, even after the accident and subsequent production of its consultant report, Licensee still made no effort to
.m determine who exactly was involved id or had knowledge of the leak rate falsification scheme, particularly management officials i
i (Faegre and Benson could draw no conclusion regarding " pressure" from management to obtain " good" leak rates, since no company personnel other than Hartman were available for interviews, Faegre & Benson, supra at 13, 36)<.took no action against any individual involved, and to this day denies the-fact that leak i
rates were falsifiedk See, statement of GPU Pr$sident Herman Dieckamp, NRC Commission meeting, May 24, 198?.
In addition, Licensee failed to turn the report.over to the Commission and'the
_. parties until recently, and withheld it entirely from the ASLB.
Had the report's existencp been known during the A5LB hearings, it would have at a minimum invalidated the PT,D's findingrqand p
s
/
\\g s
7 v.
r
\\r
,f ' conclusions on Boa rd Issue 10, previously favorable to the l
Licensee.
See, PID ! 506; generally, !!461 et sec.
These incideats raise questions of momentous proportions.
Until they are thoroughly and completely resolved, it is inconceivable that the Commission could determine that Licensee possesses the necessary level of integrity and respect for the public's health and safety and for the NRC regulatory process to '
be entre =ted with the operation of Unit 1.
- 3. Condensate Polishers similar patterns can be detected by evidence in the record concerning the malfunctioning condensate polishers, which started the accident.
Malfunctioning condenstate polishers had been a long-standing problem at TMI.
Evidence exists that the solenoids on the condensate polisher valves wer either wired improperly or disconnected, and therefore could not perform their function of
~
maintaining the position of the valves upon loss of air or 1
j.
electrical power. B&W 368.
In addition, the operators were frequently faced with the problem of water getting into the instrument air lines, which operate the discharge and inlet valves,' causing these-valves to shut, resulting in a shutdown of i
- fe'e'dwater' pumps and thus a loss of feedwater.
This problem could s
i have been entirely avoided by installation of an automatic bypass valve around the condensate polishers.
See B&W 368.
Recommendations for " automatic capability of the CO-Y-12" was " general knowledge "
Id. After a 1978 incident, shift supervisor Zewe'sent a memo urging such changes.
Zewe dep. Tr.
l
'120.
Hartman testified to telling s6ift supervisor Bernie Smith k
. six months before the accident that an automatic bypass valve on the condenstate polisher system should be installed.
Hartman dep. at 43, 44.
He later characterized the problem as a "nigh tma re. "
Faegre and Benson, p.
8.
Both GPUN President Bob Arnold and Maintenance a
Superintendent Daniel Shovlin claim to being entirely unaware of
?C
..the problem, Shovlin dep. at 161, or of the 1978 incident s
provoking,Zewe's memo.
Arnold at Tr. 1497, 1642.
+
In addition, at the time of the Unit 2 accident, B&W discovered that there were fully 13 work requests outstanding with respect to the condensate polisher system.
Shovlin dep. at 165.
Gary Miller cites management's failure to respond to operator complaints and suggestions for improvement of the condensate polisher system as.an example of the recurrent pattern
-a by management of general unresponsiveness to operators' complalnts, which sometimes developed into major safety problems.
B&W 360 at 17, 18, 20. Not surprisingly, nothing was done to correct the condensate polisher problem, and the result was the 1979 accident.
And perhaps most important for present purposes, Arnold still does not consider this problem as representing any threat to the safe operation of the plant.
Id. at 1498.
N 4.
Licensee's Response to the Accident.
Generally, the Commission's purpose in ordering Unit l's operating license suspension and establishing the Unit 1-restart
-, hearings was to discover whether the Licensee had learned the.
lessons of the accident and could be trusted to safely operate a nuclear reactor.
The Licensing Boar 6'has presented'three-
. ' decisions which answer this decision in the affirmative, finding that Licensee had forthrightly confronted problems and pre-accident conditions which contributed to the accident.
One aspect of Licensee's response was presented in testimony by Messrs. Keaton and Long, who headed up Licensee's own internal task force investigation into the acciden.t, and whose testimony, relied heavily upon the ASLB, " offered a postive view on the actions taken by Licensee's management in connection with the TMI-2 accident."
PID 1464.
Notably, Licensee never presented its task force report to the ASLB, and never presented it to the Commission until Commissioner Gilinsky specifically requested it from the company at the October 14, 1981 oral argument on immediate effectiveness icsues.
Moreover, not only was the Faegre and Benson report on the Hartman allegations explicitly concealed from the ASLB in connection with Licensee's testimony on Board Issue 10, it was I
never presented to the Commission in response to Commissioner Gilinsky's request, and was never even alluded to in Licensee' final accident investigation report issued three months af ter issuance of the Faegre and Benson report.
In-reviewing the final task force report, it becomes
~
apparent that its true significance lies not with the final conclusions, per se, but rather how the conclusions in the final version, produced over a year after the.first draft was presented l
~~ for d5hment to compa'ny management, changed from the original report.
After each draft was submitted to management for comment, the task force rescinded si'~nificant findings"of s
e
culpability which had been previously reached, unanimously, by members of the task force.
Unaware of course that any earlier drafts of any report even existed, neither the ASLB nor the parties ever questioned Mr.
Keaton on some'of the more notable modifications made in the r
various drafts.
What follows is a comparison of some of the most important report conclusions as they evolved through the series of draft reports.
Each is identified by letter designation for purposes of comparison:
B&W 347: Draft Interim Summary Report dated September 28, 1979:
- a. The overall conclusion of this investigation is that the TMI-2 accident was a direct result of a wide-spread complacency regarding nuclepr safety.
- b. Of the various areas investigated by the task force none was found to be totally satisfactory.
- c. Little'or no consideration was given to the critical area of man-machine interface, so that the operators were left without ready access to certain critical information while being s'ubmerged in superfluous and distracting alarms.
~
- d. Operating and surveillance procedures were incomplete, contradictory, and at least in one case in violation of the technical specifications.
At critical moments dur'ing the accident, important and significant precautions to procedures were not observed with resultant actions in conflibt with the proceddrdi
j requirements.
Watch-standing and turnover practices were insufficient to detect significant errors in safety system alignment.
- e. The approach to operator training was routine at best, and concentrated primarily on preparation for the license sxams rather than helping the operator to face I
difficult and complic'ated situations. The total system included no reliable provision for incorporating actual operating experience into the data base available to the operator.
Finally, the regulatory process itself concentrated on unreal events in a manner that encouraged blind spots,
- f. The existence of unsatisfactory features in almost every aspect of the TMI-2.. operation clearly reflects either lack of management awareness or management t
acceptance of such conditions.
The reasons for this management problem are not yet understood and will be explored further by the task force.
B&W 349: GPU Task Force Interim Summary Report, Second Draft, dated October 6, 1979:
This report contains no specifically designated l
\\
" conclusion" section.
However, it adds-some very significant additional findings, which are commented upon in a handwritten note at the end of the report.
The note reads, "This is pretty much a re-hash of what l
has already has been said.
Adding it seems to be
+
s-j building levels of excuse for the operators":
t.
G
=
- g. Operatcrs may have become " desensitized" to abnormal conditions due to previous experience, some plant design features and plant conditions which existed just prior to the event.
Leaking pressurizer relief valves produced high discharge pipe temperatures before the event.
Reactor building sump pump operation had become routine doe to leakage.
High pressure injection after' trip had occurred during several previous transients, none of which was a loss of coolant accident.
Some radiation alarms were expected after a trip....
Such conditions make it more difficult to recognize valid deviation from expected performance.
For example, alarms which are " expected" may not receive the proper level of attention.
The..ajarmisnolongeravalid indication of abnormal conditions.... An operating and maintenance philosophy toward minimizing the number and duration of abnormal operating conditions can reduce the tendency to produce " desensitization."
B&W 350: Interim Summary Report dated October 17, 1979:
a.
The overall conclusion of this investigation is that the TMI-2 accident was the result of a complex i.
combination of factors,
- b. Omitted.
c.
Limited application of human engineering in the design of the man / machine interface left the control room operatcrs without ready access to some useful c
i s
a
information while being besieged with irrelevant alarms.
j d.
Investigation of the emergency feedwater valve closure revealed surveillance procedures were in violation of the technical specifications even though they had passed through the full chain of approvals.
Watch standing procedures and practices were admittedly, inadequate to detect incorrect alignment of even safety system components.
e.
The approach to operator training placed emphasis on attaining and maintaining an NRC operating license rather than a thorough understanding of the plant performance under all foreseeable situations. Simulator casualty training stressed single failures and did not deal with plant response uTder multiple failures.
Emergency procedures provided inadequate and in some cases centradictory guidance.
Errors in operator judgment delayed isolation of the leak and also resulted in prolonged operation at low primary pressure with insufficient recognition of primary inventory.
Precautions and limitation in emergency procedures were i
overlooked and not recognized as applicable.
Relevant operating experience from industry-wide activities was not effectively transmitted, received, and evaluated for use in, operator training ~ programs.
Finally, the regulatory process itself-placed undue emphasis on
+
v-m
--g e--
l
. specific unlikely events in a fashion which encouraged blind spots.
- f. The reasons for widespread existence of problems in the various aspects of the TMI-2 operation has not yet been well defined by the task force and will be further
- pursued, However, the general operational condition
. appears to indicate a. combination of a lack of
/
management awareness of problems, an insufficiently stringent standard by which to evaluate operations, and a management philosophy which accepted this situation, at least in the short run.
- g. Omitted. The following was added, however: The NSSS design which led to occurrence of HPI during non-LOCA transients initially masked recognition of a LOCA.
c.
Other problems on the secondary systems distracted at least some operators.
i In addition, the following findings were made:
- h. Review of the emergency plan performance showed that the organizational and communication concepts were inadequate.
While information flow to the cognizant state agendy regarding releases of radioactivity was generally adequate, there was at least one case of contradictory information being released.
Communications were inadequate to properly inform the NRC and utility management of plant status and
- problems, f
.w
a 1. Operation and maintenance problems in the condensate system coupled with the limited capability of the system design to accommodate transients resulted in the unit trip.
B&W 351: Interim Summary Report dated October 29, 1979.
The task force cover letter states that "[t]he conclusions and recommendations represent the unanimous current opinion of the task force."
a.
Unchanged'from last report,
- b. omitted.
i c.
Unchanged from last report.
d.
Unchanged from last report.
e.
Unchanged from last report, except the second sentence was changed slightly to: Simulator casualty c
training stressed planned response to pre 6efined single failures, and did not deel with unanticipated event or multiple failures.
- f. Unchanged from last report, except the second sentence was changed slightly to:
However, the general operational condition appears to indicate a lack of t
management awareness of problems, an insufficiently i
f stringent standard by which to evaluate operations,
\\
and/or a management philosophy which accepted this situation, at least in the short run.
9 Unchanged from last report, except the word " masked" was changed to " hindered."
+
. F
- h. Possibly unchanged from last report, however the actual copy is unclear,
- i. Unchanged from last report.
B&W 352: Interim Summary Report, dated November 28, 1979.
This report, as the previous one, was signed by all five task force members,
- a. Unchanged from last report:
- b. Omitted.
- c. Unchanged from last report.
d.
Investigation of the emergency feedwater valve closure revealed that watch standing procedures and practices were inadequate to detect incorrect alignment of safety system components.
Operator training in geperal had placed emphasis on e.
attaining and maintaining an NRC operating license; in retrospect this approach did not guarantee a thorough i
understanding of the plant performance under all foreseeable situations.
Casualty training, including simulator training, stressed planned response to predefined single failur'es, and did not deal with simultaneous multiple failures which present unique
\\
sets of symptoms to the operators, such as actually occurred at TMI.
No training had been given in response to events which were not predefined.
~'
Emergency'and operating procedures were-found.to provide unclear and in some cases contradictory guidance.
Errors in oper[ tor. judgment delaydd'
i isolation of the leak and resulted in prolonged operation at low reactor coolant system pressure.
Precautions and limitation in some emergency procedures 4
were overlooked or not recognized as applicable.
In addition, the following operator training conclusions were added:
/
-- A basic human problem was the lack of appreciation of the unique symptoms which would result from this event. The data base used for operator training and preparation of emergency procedures was limited to a LOCA which reduces both system pressure and pressurizer level. The limited attention which had been given to the consequences of a leak from the pressurizer did not focus on the difficulty in;-recognizing the leak, due to the rising pressurizer level.
The operators also did not have access to information on-the occurrence of a similar event at another plant.
Furthermore, the simulator used for operator training did not have the capability for simulating a LOCA from the pressurizer vapor space.
The operators thus expected loss of both pressu,re and pressurizer level if a LOCA occurred, and s
were not equipped to recognize a LOCA from the pressurizer vapor space in which pressurizer level did not reflect the system inventory.
-- For both unlikely accidents and expected plant transients, inadequate attention was given to
+
predicting the most probable plant response.
Training
programs and operating procedures thus did not have a sufficienct data base to provide the operator with all the needed information.
- f. Unchanged from last report.
- g. Unchanged from last report.
- h. Review of the responses to the accident showed that the organizations involved had'not perceived the magnitude or duration of requirements in the post-accident trip period.
The protracted series of events which actually occurred pointed out inadequacies in the organizational support, in the communication system, and in the predefined information flow network required for effective accident management.
The TMI experience showed that a large of f-site support i
organization is needed to assist the in-plant organization.
While information flow to the cognizant i
state agency regarding releases of radioactivity was generally adequate, there was at least one case of contradictory information being released.
Communication systems proved inadequate to properly inform utility management and the NRC of plant status s
and problems.
- i. Unchanged from last report.
B&W 354: Final Summary Report, dated March 24, 1980
~'
~~
a '. Unchangdd from last report.
- b. Omi tted.
- c. Unchanged from last repdrt.
. d. Unchanged from last report.
e.
The first sentence of the second paragraph discussed above has been changed, characterizing the basic problem as one of " software" rather than " human."
Other than that, it is unchanged from last report.
- f. The task force did not perform a thorough review of the role played by TMI management relative to the identified problems, primarily because the management structure was significantly changed from that which existed at the time of the accident.
The task force did, however, develop some recommendations for future management actions, as discussed in the next section.
- g. Unchanged from last report.
- h. Unchanged from last report. In addition, the 4
following finding was added:
Plant staff and company management concentrated on plant cooldown, emergency response and communications on March 28 rather than assessment of total core damages.
Realization of the extent of core damage resulted from the discovery of the hydrogen bubble.
- i. Unchanged from last report.
l B&W 355: Final ~ Summary Report, dated May 12, 1979
- a. Unchanged from last report.
- b. Omitted.
j l
- c. Unchanged from last report, l
- d. Unchanged from last report.
- e. Unchanged from last rep"rt.
o
- f. Unchanged from last report.
- g. Unchanged from last report.
- h. Unchanged from last report.
- i. Unchanged from last report.
B&W 356: Final Summary Report, dated December 15, 1980, with the notation: THIS REPORT IS CONFIDENTIAL FOR LIMITED
/
DISTRIBUTION.
THIS REPORT IS NOT Tb BE REPRODUCED.
The relevant conclusions are identical to the May 12, 1980 version, above.
Comparisons:
' Conclusions (a) and (b):
After the first opportunity for management comment on these conclusions in B&W 347, both were removed, never to appear again.
Conclusion (a) was changed entirely, (b) eliminated.
-s Conclusion (c): This conclusion remained through all drafts i
and the final report.
It highlights the critical problem of confusing alarms, discussed supra, which the ASLB discounted as a safety issue with regard to a TMIA maintenance work request exhibit.
Conclusion (d): The most damning aspects of this conclusion, i.e.,
that surveillance procedures were in violation of technical y
specifications even though they had passed through the full chain of approvals, made it through all drafts until November 28, 1979, B&W 352, when it was eliminated.
Up until that point, the drafts reflected a very strong task force finding that tech specs were
'indeed violated, and significantly, the NRC cited Licensee for
+
violating these tech specs in its October 25, 1979 Notice of
. 9
' Violation.
However, in Licensee's December 5, 1979 response to 1
the NRC's Notice of Violation, management did a complete reversal of task force's original position, insisting there that tech specs were not violated.
Conclusion (e): The substance of this conclusion seems to become more detailed as the drafts evolved.
By November 28, Licensee was clearly recognizing the failure of training and emergency procedures to prepare operators for conditions where pressurizer level was rising as pressure dropped.
The most important conclusion in terms of the restart case, since it has reappeared in the April 28, 1982 Special Master's report on the Reopened Proceedings, and most recently as a strong finding in the RHR managment consultant report (See, discussion on training, infra), was that "the apprcach to operator training pla'ced emphasis on attaining and maintaining an NRC license t
rather than a thorough understanding of the plant..." The i
reappearance of this finding over and over illustrates clearly Licensee's failure to correct one of the most critical training problems revealed by the accident and acknowleged by the Licensee.
Moreover, in Robert C.
Arnold's cover memo accompanying Licensee's December 5, 1979 response to NRC's Notice i
of Violation, Arnold downplays the seriousness of the trai.ning problem to the NRC, stating,
"(d]uring the period from 1975 to 1978, operators at Three Mile Island had a failure rate on their NRC written and oral
~
ERams ~ half the industry ~ aver age.
NRC performance evaluations ranked the Three Mile Island f acility above the average for comparab'le plant.s.
Metropolitan Edison does not feel that there was any sigtifigant decline in the,, Company's performance.
1-.y- - +
m 4
v
Importantly, Arnold's representation connotes a fairly positive view of training which is quite contradictory to findings by its own 1978 management audit, (which findings generally appear also in a 1975 management audit, Arnold dep.
l 350), that "the quality of operations personnel is on a continuous downhill trend due to lack of training, B&W 843 at 45229.
Conclusion (f):
This conclusion undergoes perhaps the most significant change of all.
Management's culpability is represented in each draft through the March, 24, 1980 version, at which time its pejorative nature is completely eliminated and no subsequent finding is made against management.
Notably just before the issuance of this draft, the Commission issued its March 6, 1980 order directing the ASLB to resolve 13 specific management issues including management's role in the accident, raising very real questions regarding the relationship between i
these two events.
l Conclusion (g):
The concept of operator "desensitivity" recognized by the NRC in its Notice of Violation, (see p.
24, supra), was carefully eliminated o'r at the very most, extremely watered down from what was suggested by the task force in B&W 349
-- the October 6, 1979 draft.
As was discussed, supra, continued operation with high discharge line temperature readings was a deliberate act which seriously endangered the
~~public-heal'th and safety, if for no other reason than it clearly 7
f desensitized operators to that condition during the accident contributing to failure to diagnose 5he stuck-open PORV. Indeed,
[
I l
l,it resulted in the greatest percentage of the total civil penalty imposed by the NRC -- S630,000.
Further, the Licensee affirmatively disputes this notion in its official response to the NRC, concealing any reference to its own accident investigation task force finding of several weeks earlier.
Specifically, on p.
34 of Licensee's December 5, 1983 official response to the NRC's October 25, 1979 Notice of Violation, in response to the item of noncompliance dealing with allowing operation with valve discharge line temperatures to exceed the normal 130*F for months before the accident, Licensee states, "there is no indication that this procedure or the history of PORV discharge line temperatures delayed recognition that the PORV had stuck open during the course of the accident."
i Conclusion (h): The first few drafts cite communication failures.
This is later changed to'c'ommunication " system" failures, showing how the company, through mere word engineering seemed to absolve the people involved in the communication failures.
Further, the idea for the additional paragragh cited in the March 24 draft seems to have had a specific origin.
In a Decembe.r 3, 1979 MEMO TO FILE, B&W 353, Re: TMI-2 Accident Review Task Force Open Items List, the following action item was listed:
Under Section II.B.2 "Information Flow" develop a story that the plant management and Met-Ed management were immediately drawn into communication problems with state agencies rather than being free to concentrate on the plant conditions. Develop a recommendation that casualty managers have a communication system which leaves them free to concentrate on the plant.
(' emphasis ddded).
~'
i This raises the obvious question of whether a " story" was 0
i-deliberately fabricated to excuse company management for iheir
o,
' failures in understanding how to control the accident, and for presenting misleading accounts to state officials as to the accident's severity.
The ASLB made the "information flow" issue its primary focus in resolving Board Issue 10.
Thus, the questions raised in development of Licensee's "information flow" conclusions have very direct relevance to the ASLB's resolution of this issue.
Conclusion (i):
While this conclusion is significant in indicating management's recognition of the condensate polisher's frequent malfunctioning and the maintenance department's responsibility in that area, it is also significant in comparing the report's analysis of the condensate polisher problem in light of B&W 718.
This exhibit is a handwritten cover memo from task 4
force member Ken Lucien who apparently had just completed his 4
analysis of the condensate polisher problem.
The memo reads, in
- part, This is the draft of my report on the polishing / condensate / air systems for the Investigative Task Force.
Per our understanding with R.
- Keaton, please launder this to bring it into line with your presentation of the forthcoming master task force report.
(emphasis added).
v This memo raise _s the very serious issue of dishonesty by the task force itself in its conduct of the investigation.
In light of this and other highly suspicious modifications which were made i
to the task force's original conclusions as the final report evolved, it is incumbent upon the Commission to insist that the Licensee's internal accident investigation, particularly with regard-to the role of company management, be thoroughly examined, 1
. and that the ASLB's conclusions on Board Issue 10 be reevaluated and substantially modified in light of this new evidence.
B.The B&W record, read in light of other new information which has come to light reflecting on Licensee's management, fundamentally undermines the credibility of the ASLB decision, and the PIDs can not be used to lawfully justify restart of TMI-1.
TMIA has consistently maintained throughout these
/
proceedings that had the ASLB exercised its responsibilty to conduct a fair and impartial hearing, thoroughly examining the many management issues which the Commission directed it to consider, a result favoring restart could not have been reached.
TMIA's review of the B&W record so far, particularly when read in light of the evidence already developed in the restart record, the BETA and RHR management audits, and recent allegations by TMI-l clean-up "whistleblowers," provides renewed support for this position.
TMIA will attempt to present a necessarily t
abbreviated discussion of some additional evidence in the B&W record with direct bearing on particular restart issues ostensibly " resolved" by the ASLB in support of. restart, which demonstrate just how remarkably shallow the ASLB findings and conclusions really are.
- l. Maintenance.
s Deficient maintenance practices at TMI-l was the sub' ject of TMIA Contention 5.
See, PID 1 277.
In brief, TMIA alleged that Licensee was incapable of safely operating TMI-l because of a past course of conduct which included deferral of safety-related maintenance and repair beyond a point established
+
l by its own procedures, and a disregard of important l.
l
6 1 ' safety-related maintenance, illustrated by its proposed 1979 maintenance budget cut, failure to keep accurate and complete i
maintenance records, inadequate and inderstaffed QA/QC maintenance programs, and excessive overtime.
The difficulty TMIA had in presenting its case, and the shoddy treatment given these issues by the ASLB has been briefed
-for the Appeal Board.
See TMIA.'s brief in support of exceptions to the management PIDs.
Briefly, the ASLB conclusions rest primarily on assurances by the Licensee that " maintenance work scheduling is consistently maintained, completed timely, and consistently with safety."
Shovlin, et al. ff. Tr. 13533, at 52; and NRC staff assurances that maintenance practices are satifactory.
- See, e.g., PID 1314.
See, also, TMIA's management findings 1579-81; PID 15289, 296.
G To understand how empty the ASLB's conclusions are, the Commission need only consider that the very maintenance department which allowed such dangerously deficient pre-accident maintenance problems to develop at TMI-2 with regard to such components as the PORV and condensate polishers, (see discussion, supra), was run by the same individual who now heads TMI-l maintenance -- Daniel Shovlin.
This was the same department N
which occasioned' company auditors to find i~n late 1979, after the accident, that,
--maintenance lacking in advance planning; is done on ad hoc basis.
- "No preventive maintenance done by Met-Ed - not enough people to do it; therefore all maintenance that is done is
" correcting" work.
-Met Ed labor responds to crisis.
-- Consistently get low productivity in maintenance - have had opprations analysis but never follow-through. "
-- Maintenance is the weakest area on sight.
-- Maintenance planning "out one day" is a disaster.
-n-
,w-:r
-4
B&W 32, p.
9.("The Glickman audit" commissioned by GPU to assist the company in determining how to restructure the management organization after the accident.
The audit reflects the opinion of more than 30 key GPUSC and Met Ed managers, executives, and technical and task force leaders.
Id., cover memo).
And this is the same department which prompted the following e
observation from the BETA auditors this year:
Plant maintenance at TMI-l has yet to reach the point where required equipment reliability can be reasonably assumed.
It appears from the above that the maintenance department is still woefully deficient, despite testimony to the contrary by Licensee and Staff witnesses.
Moreover, throughout this record, one sees repeated evidence of specific and often acknowleged
<5
{
maintenance problems dating back to years before the accident, some of those very same problems presented for consideration but I
disregarded by the ASLB, only to reappear in this year's BETA report.
Although relating primarily to pre-accident conditions, the B&W record does show how deeply ingrained maintenance defici*ncies are.
It is not surprising that problems still exist.
.But more importantly, the record demonstrates that the i
ASLB's glowing account of TMI-l's current maintenance department is simply incorrent.
The following issues are examples directly relevant to the restart case.
a.
Budget cuts and understaffing.
--~
Perhaps the most revealing description of Mr. Shovlin's pre-accident maintenance department is contained in B&W 360,
0 U - for'mer Station Superintendent Gary Miller's post-accident interview with Keaton.
Miller describes the difficulty he had with understaffing and budget restrictions before the accident, resulting in among other things 800-1000 open maintenance items at TMI.
Yet when questioned about this figure, Shovlin seems to characterize such a large number of open work items as.the
" normal" amount which actually, exists to, day.
Shovlin dep. at 171.
But to Miller, that situation meant simpl'y that "you are going to develop more work requests than you are going to do."
B&W 360 at 13.
The ASLB never knew of Miller's concerns because he never appeared before them, and they certainly never examined Shovlin on this, and it is quite doubtful whether they could support such a condition.
Miller's biggest problems bef, ore the accident centered
-G around the budget cuts which he supervised, particularly those dealing with the maintenance department.
He discusses his 8
i difficulties quite frankly with GPU's accident investigation team in B&W 360:
-- Miller states in answer to a question about what kinds of problems he had with budgets and personnel:
"Trying to cut the budge.t...Trying to convince people I couldn't.
Trying to convice people I needed more maintenance staffing.
Trying to eliminate the contractor and I didn't think he could be eliminated, that k'ind of thing." p.
6
-- Miller states in answer to a question ~about why he had 800 maintenance items:
" Yeah, I thought of it.
I didn't know how to proceed on it because the answer kept coming up we need manpower.
If you really think about it we had shift maintenance but we only really had two units with a million dollars and we had the ability to do one or two jobs on the back shift.
We just can't do it with the size of this plant.
You are just kidding yourself." p.+12-13.
i.
- 9 "If you look at the routine way we do business I don't believe we had the manpower to do more than the priority ones.
Don't even talk about money.
The only way you had of doing other than priority one jobs was to contract and we were beginning to limit that to zero." p.
22.
"If you take surveillance and we weren't doing PM's when Unit 2 came along we went from the 50 or 60% PM performance in Unit 1 to almost zero." Id.
" Maintenance was where the staffing problems were, and we hadn't really increased'our staffing in maintenance very much-beyond Unit I's levels of maintenance.
You are talking a maintenance staff of roughly 40 mechanical people, probably 35 mechanics and 35 instruments, 24 electrical, serving two units and I just couldn't believe you could serve two units with that number of people and work and do a reasonable job without a contractor.
The thing was really squeezing us before the accident.
Was squeezing me in the elimination of the contractors." pp. 23-24.
A number of these observations are confirmed by the Glickman audit's post-accident findings, discussed supra.
(e.g.,
"No preventive maintenance done by Met EId - not enough people to do it; therefore all maintenance that is done is " correcting" work.
Met Ed labor responds to crisis." B&W 32 at p.
9).
TMIA attempted to prove that the proposed 1979 maintenance budget cut, which would have taken effect but for the accident, was not approached with due regard for safety.
TMIA management findings 1145-47.
This assertion was rejected by the ASLB.
s PID 11322, 324.
It now appears that such cuts were sought by Met-Ed management back in 1977 or 1978.
Shovlin dep. at l'4.
Miller's observations, as well as findings made in the Glickman audit, and evidence of deferred maintenance on items such as the PORV and condensate polisher, are in direct conflict
[
withspecificfindingsoftheLicensjngBoardwhichrel..ied exclusively on the self-laudatory testimony of Licensee and the i
v.
D equally shallow testimony of the NRC staff. Miller's statements 2
concerning the impact of understaffing and budget cuts, the difficulty of handling what work had to be done, the impossibility of doing other than priority one items, and the inability of doing preventive maintenance at all, undermines the credibility of such ASLB findings as:
-- the Board finds no support in the record for TMIA's specific allegation that Licensee has under the past or oresent system departed from a company standard in failing t.' perform maintenance in a timely fashion. PID 1 289.
-- With respect to the issue of improper' deferral of maintenance work, we find that Licensee's responsive written testimony satisfies us that there was none of significance, and that there is nothing inconsistent with the written responsive testimony in the examination of Licensee's witnesses at the hearing.
PID 1 296.
-- We can and do find that Licensee's records under the old system were auditable, albeit at times with difficulty, on the basis of the uncontradicte.d and unquestioned testimony of the NRC Staff inspectors on the sample year-1978.
PID 1314.
-- GPU management, even in times of financial stress, has recognized the unique demands of its nuclear obligations and j.
has shifted available resources to meet those obligations.
See also, PID 1322, (GPU approached budget cuts with due t
regard for safety).
In light of observations by Miller, whose unique position made him particularly sensitive to the consequences of budget y
cuts, and who never testified before the Board, the ASLB canclusions on this issue seems particularly incredible.
Clearly, these findings need to be reevaluated and modified, and should not be used as a basis to justify the ad'equacy of TMI-l's current maintenance department.
+
v e
r 6
- b. Repair parts -- materials manacement.
r.
In 1978, Met Ed commissioned a final report of a Management Audit performed during the period of January 9-20, 1978, which identified problems at TMI and recommended corrective action.
B&W 843 ("1978 audit").
The auditors found, among other things, that there was " difficulty in locating repair parts known to be in the Warehouse," and that "the repair. parts documentation for /
Unit 1 is not well organized and requires an inordinate amount of research to identify the Met Ed stock numbers required for requisitioning material."
Id. at 45228.
At an April 24, 1978 management meeting held to discuss the audit, the problem of difficulty in locating repair parts known to be in the Warehouse, random storage, large quantities of unidentified parts
.n the warehouse, and multiple storage of single type items, was 4
discussed.
B&W 884.
As concrete evidence of this problem, E&W discovered that a I
spare PORV was ordered in 1975, and delivered to the island until 1978.
Yet in early 1979, the Supervisor of Maintenance was still unaware that it had been delivered.
Seiglitz Tr. 5766-67.
In fact,*an inference can be drawn from plant discussions that in early 1979, Met Ed was unable to determine from their records s
whether they had a s' pare PORV in stock and had to request help from B&W to locate the PORV.
See, id. at 5790-95.
In late 1979, the "Glickman audit" concluded at p. 14 as
-- follows: --
-- cataloguing of documentation as to Supplier, part number, etc. a disaster on Unit 1 - Should be overhauled.
+
-- There are serious problems in TMI documentation, cataloguing, storage, retrievability.
,w-g r--
TMIA raised this issue during the restart hearings with regard to a particular maintenance work request which had been given a lA priority.
TMIA managment findings $25.
The issue was dismissed by the ASLB.
PID !297.
Significantly, practically identical findings reappeared at p.
28 of the BETA management consultant report issued in 1983, demonstrating seriously poor management,.of this department and i
illustrating how unresponsive management has been to this well-documented problem which pre-dated the accident:
-- warehousing inventory records were inaccurate to the point as to be considered unreliable by job planners.
-- the principal deficiency with inventory records is nomenclature.
-- the amount of stock is excessive.
-- there is no scheme for purging stock frem inventory when technical or administrative requirements prohibit use of material present in stock.
-- the period of time from preparation of a requisition to delivery of purchased gaterial is small.
Again, with regard to this issue, the ASLB findings need to be reevaluated and modified, particularly since it was never the f
subject of specific discussion, other than with reference to one 1
TMIA exhibit, supra.
- c. Paperwork In Gary Miller's post-accident interview with Keaton, he explainbd that maintenance foremen spent only 19% of their time supervising their people, the other 70% or 80% spent with-paperwork. B&W 360 at 13.
Shovlin confirmed this in his B&W deposition.
See, Shovlin dep. at 184-186.
And Robert Arnold admitted in his deposition-that he was aware before the 1978 audit that maintenance was spending a lot of time on paperwork,
+
i.
I 4
4. and' when asked about Miller's 705 figure, Arnold suggesteds that the problem was well known to them and that they instituted efforts to mitigate the problem after the 1975 audit.
Arnold dep. at 380-383. (emphasis added).
Moreover, he admits that he knew in 1978 that paperwork problems persisted. Id. at 384.
The 1978 audit found that much of the efforts of maintenance managers and supervisor are taken up in, chasing' material and necessary documentation required for work performance, and that "many supervisors expressed frustration due to the fact that they are often " tied to their desk" by administrative tasks and excess paperwork." B&W 843 at 45219.
The Glickman auditors in late 1979 found that " paperwork is incradible."
B&W 32 at 11.
.A large part of TMIA's case before the ASLB concerned the poor administration of paperwork to the extent that safety-related records were unrevied$ble in violation of 10 CFR Part 50, App.
B, and that the situtation has not improved but has i
i in fact gotten worse.
See e.g.
TMIA management findings 1120, 76.
See also, Smith Tr. 3730.
The ASLB concluded that paperwork was no longer a problem.
See, PID 1301, et seg.
Yet the BETA report finds that in 1983, " paper generation is
\\
not only a constant source of complaint, but our observation indicates that it is real."
BETA at 110; and the RHR auditors noted, " complaints of ' working in a paperwork jungle'" again
_. illustrating the faultiness of.the ASLB conclusion.
e i
x
- d. Overtime
~
Overtime, including "the extended hccrs," was tha factor responsible for the " greatest negative impact on morale s
according to the 1978 audit, and was becoming an increasingly j
serious problem.
B&W 843'at 45217. See, also, B&W 884
(" Excess overtime -- effective course of action not developed).
Overtime was also a problem discussed by' Gary Miller in his
/
January 1979presentationtoGh,UmanagemenionTMI.
~
B&W 783.
The Glickman auditors in late'1979 discovered the feeling that
" working hours on-site are ridiculous."
B&W 32 at 11.
The overtime issue was arbitrarily dismissed by the ASLB, which failed even to',avaluate the credibility of the itnesses who' testified. See, PID 5 341, ("We could identify n"o abuses"); TMIA brief in support of exceptions to tlie PID, p.
14 et seg.
~<
The ASLB's peremptory avoidance of the, issue seems even more arbitrary in light of the strong -findings irIthe 1978 audit.
i See, PID 1343.
The issue deserves far better examination than i
s c_
that given by the ASLB.
The general di'sorganization a d poor perIormance of e.
the maintenance department under~the supervision o_f _
current Unit 1 maintenance supervisor. Daniel Shov.lin.'-
c~ s
~
t The B&W regord reveals'that the pre-accidedt maintenance
[
g rm.
department at TMI was extraordinarily disorganized, manifbstqd'
- q not merely by paperwork and warehouse confusion, but by. failures within the department to efficiently, rem dy safety-related
~
maintenance problems. ~Nuch of(the problem can be attributed to
~ 1B the incompetence and/or ignoranch of Daniel Shovlin.
Shovlin's
C n-L
+
depositions reveal that as head of' maintenance before the
'sy m-t n
,. accident, he was either so ignorant as to not understand the significance of prominent, long-standing safety-related maintenance problems; or, he had utterly lost control over his department; or, he was simply a non-credible witness.
To iliastrate Shovlin's ignorance, incompetence, and/or dishonesty, several examples from his depositions stand out:
-- when Shovlin was hired as supervisor of~ maintenance at
~
TMI, he had never before w6.rked for a commercial utility yet he felt there was nothing he needed to do to familiarize nimself with TMI-1. Shovlin dep. at 30.
-- he has never looked at a startup log, nor seen a maintenance log. Id. at 16, 17.
-- In 1975, Miller held Shovlin, among others, responsible for a " low level of interest and seriousness with which the
[ Unit 2 procedure writing effort] was generally thought of and pursued.
B&W Ex. 763
-- he has no recall of the July 30, 1975 letter sent from Lee Rogders of B&W to John Herbein recommending preventive maintenance for the PORV, whichewas never implemented.
_Id.
at 50.
-- he has no idea if any preventive maintenance was being i
performed on the PORV before the accident or if any regular i
inspection of the PORV took place.
Id. at 182, 183.
-- he did not know that the PORV failed to open during Unit 2 start up. Id. at 51.
-- he says it never came to his attention that in the fall of 1977, there was any particular problem with respect to the PORV, Id. at 52, yet in September 1977, the PORV had leakage problems and was removed for in house repairs, and in an October 1977 report describing the leakage problems,
\\
B&W 4033, Shovlbn was assigned resolution of the problems, and he signed off'on the work procedure. B&W 4034.
-- he has no recolfection of knowing at the time about the March 29, 1978 transient at Unit 2 where the PORV remained open for four minutes because of an electrical failure causin.g a.rgactQr trip. Id. at 55.
-- he is entirely unaware of the April 23 transient or of doing anything in connection with it, or of participating in any dicussion concerning the event. Id. at 59.
i L
2r
_ _ ~ '
-- he was unaware in late March or early April 1978 that an indicator light was to be installed in the Unit 2 control room to show demand indicator to the PORV solenoid. Id. at 56.
r
-- he is entirely unaware of the April 1978 " alarm window correction program" although it was a priority 1 item under his responsibility, and rejected by corporate management Id. at 60-63.
-- he did.not know that in the fall of 1978 the PORV was tested and remained shut when the indicator light showed it was open. Id. at 63.
-- he says he was not aware before the accident of leakage 2
past the relief valves at the top of the pressurizer in Unit
- 2. Id. at 125.
-- he has no recollection of any recommended repairs to the PORV. Id. at 129.
-- he claims he has no recollection of the early 1979 t
increased identified leakage or the elevated temperatures in the discharge piping, or.that temperature and pressure in the reactor coolant drain tank was increased over normal levels during the January 1 - March, 1979 timeframe, yet he recalls he was asked to look at a change in the reactor coolant drain tank capacity "seneral years" before. Id. at 149, 150.
~
-- he recalls no attempt to determine which, if any, of the relief valves at the top of the pressurizer was leaking during the January to March 1979 timeframe, despite the fact i
that he was attending most POD meetings, and got written reports of those meetings he could not attend. Id. at 152-153.
-- he has no idea when a spare PORV was ordered from Dresser ir.1978, or what occasioned their concern for obtaining a spare. Id. at 153-154.
-- he never knew the ccndensate polishing system
's malfunctioned before the accident. Id. at 161.
-- he'has no recollection of an incident in the fall of 1977 involving a buildup of resin in the condensate polishers followed by the trip of main feedwater pumps or a May 1978 incident resulting in the loss of main feedwater, yet he sat in on a discuss (on at which the installation of the automatic bypass system was discussed. Id. at 162, 163.
-- he has no idea if any preventive maintenance was being performed on the condensate pol (sher system before.the e
accident. Id. at 182.
a r.
.e
he had no idea that at the time of the TMI accident 13 work raquests were outstanding with regard to the condensate polisher system, and from January 1 through March 1979, 13 more were carried out. Id. at 165.
In addition, the following exchange between B&W's attorney and Shovlin typifies shovlin's testimony during his B&W deposition.
Shovlin dep. a.t 195-198:
Mr Wise:
Now, Mr. Shovlin, I don't,.mean this to be a i
facetious question, I am just trying to understand what your position was within the organization and what your responsibilities were.
Did anybody bring any problems, maintenance problems, to your attention before the accident at TMI-2 in March 1979 on Unit 2?
l Shovlin: The significant items were handled by the
~
supervisor of maintenance on the unit.
Wise: I have asked you about the condensate polishing system malfunctioning, and you said you didn't remember that. I have asked you about the radiation monitoring systems, and I have asked you about the leakage past the pressurizer relief valves and the discharge from them, and you don't remember anything about that coming to'y.qur attention.
I might say that after the accident there has been quite a good deal of investigation about various things and those three problems have all been identified as things which occurred before-the i
accident and which were ongoing problems.
None of those i
things came to your attention before the accident.
The question is: is it true that none of the specific items that I just mentioned came to your attention before the accident?
Shovlin:
On the time frames that you are asking me to recall, no.
They could have -- I could have been invovled in several of the problems, but I don't recall or remember specifics about them.
Yes, we had leakage of the pressurizer valves.
When, in the time frames and the particular and specific problem, five, six, seven years ago, no, I don't remember.
Wibb:kreyouabletorecalltodayatanytimebeforeMarch 18,. 1979 being aware of leakage past the pressurizer relief valves at TMI-2?
Shovlin: No.
e Wise: And that is without regard to any specific time period or anything else?
y w.
.y..
..._s.
.y
. Shcylin: No. Right.
~
Wise: That just never came to your attention?
Shovlin: I am telling you right now, no.
Even assuming Shovlin is as forgetful as he claims, serious questions remain.
How can the head of maintenance competently supervise his department when his subordinates do not inform him of important and potentially dangerous plant conditions -- an arrangement which he appears to find perfectly satisfactory?
Moreover, how can it possibly be established that Shovlin has learned the lessons of the accident, the severity of which was caused in part by the incompetence of his own department, when he has no recollection of the circumstances surrounding any precursor event or malfunctioning component which caused or led to the accident?
It is utterly irresponsible to permit him to 3
continue as head of TMI-l maintenance.
2.
Training.
Licensee's training department has undergone intense scrutiny in the restart hearings because of the role improper training played in causing the accident.
Company management's role il creating deficiencies in the training department, including significant and wide-spread instences of cheating, and management's respons'e to correcting those deficiences has been the subject of much controversy.
What the B&W record adds to this issue is demonstration that training problems were long-standing and well-recognized, and that the company did absolutely nothing about them until forced to do so after the accident.
f A particular issue of controversy during the restart hearings was the adequacy of the requalification program and the related problem of the Licensee's policy on non-attendance and take-home make-up training packages and exams.
Lax procedures created an environment which allowed Supervisor of operations at Unit 2, VV, to cheat on his exam in July of 1979. PID 12272 et seg.
D&W 462 is a March 1, 1977 Memo from Mr.
Tsaggaris to a number of people, including Gary Miller, Jack Herbeln, and L.
L.
Lawyer of the training department, concerning the Unit 2 on the job training program.
Tsaggaris states in a handwritten comment, We are in trouble on this program! Progress for the last two weeks has almost been nonexistent.
All groups have fallen way off the required curves... I don't know what the problem is but we had better find out now or we will never make it by 7-1.
This matter will be discussed at G.P.
Miller department head meeting on March 3, 1977.
On June 2, 1977, Miller sent to Lawyer a memo on the training program, in which he stated,
...As is typical with every startup, we are attempting to complete a year's worth of effort in about 6 months.
The Unit 2 information at the critical detail level is just now becoming available in usable form.
B&W 774.
Milier testified in his deposition that this memo meant that he did not feel the classroom training was directly applicable to the operation of the units.
Miller dep. at 466.
On June 17, 1977, Unit 1 shift foreman T.L.
Book sent the following handwritten letter to Unit 1 superintendent James P.
O'llanlon, who later reported it to Miller:
Since taking the requal exam this past February, I have not been in a single training lecture or received any guidance as to what course of study to pursue to best fulfill the NRC requirements meaningfully.
i Also, I do not believe that sending out a casual memo I
I or documenting on green sheets that an E.P. was read on back shift constitutes good training practice.
Like all else the S/F & S/S's have become the Godhead of 60 hrs. required training per year.
Its time to put i
training back in the training dept. where it belongs and in a responsible fashion.
This means more training space, l
people and expertise.
This also means 6 shifts for CRO's, S/F and S/S's.
While I fully realize that there is no pat answer for our complex training problems, I like many other operations j
people have made suggestions to various training personnel.,
However it seems as though those fall on deaf ears or end up in the circular file.
We have been told " write up your suggestions and concerns or call us."
We did!
Nothing happened.
Besides being just plain frustrated over all of this, i
it is my opinion that it is somewhat erroneous to say we fulfill the NRC requirements when they are based on documentation of subject matter supposedly covered on shift.
Many times more hours are documented than were actually used for training.
I am willing to listen to or discuss anything on the topic with anybody.
I am willing to help solve the problem if I can help in a meaningful way.
Something must be done !!!
B&W 564; Miller dep. 477.
Several m@nths later, the 1978 management auditors made the following finding about the training department:
"The quality of operations personnel is on a continuous downhill trend."
B&W 843 at 45229.
Miller voiced i
similar concerns in his post-accident investigation interview, B&W 36C..
There he stated, i
j everytime I went to a shift foreman or shift supervisor i
meeting one of the single most emotional complaints was l
N training.
Lackaof.
Lack of real training.
l f
B&W 360 at 2.
Among the major problems with training, including one of
~ ~ Floyd'sImosh significant training shortcomings even after the accident, was the training non-attendance record. See PID 12274.
B&W 304 is a September 1, 19[8 memo by Beers, oE the t
b e
y
....w..s
%-.%,,,,-..,4
,,_.rp,.
,.gy, c..pm y_,,,-4 y_.,
.-._.,,..,,_y
,,.,,,3y,
~+e
1 '
training department, which states, but overall approximately 1/2 of the licensed people are not attending requalification training."
In a November 2, 1978 memo, Beers writes to Miller,
" decrease in attendance from last report."
B&W 776.
This caused training instructors to spend substantial amounts of time making up take-home training packages.
Arnold Tr. 1703-1704.
Even by
=
late 1979, after the accident, the Glick' man auditors found that '
i.
cases existed where one could pass the licensing tests without taking any training session.
In addition, one of the more significant revelations of the B&W record was that Mr. Zechman, the acting supervisor of training, not only did not have his operator's license, but at a time of major training deficiencies within the department, a decision was made to have Zechman spynd full time studying for his license, spending no time running the department.
Arnold at Tr. 1706.
Moreover, some time between the fall of 1978 and the accident, Zechman took the examination and failed to pass it.
Id.
Miller believed that the department suffered because of Zechman.
B&W 360 at 29.
In Licensee' response to the'NPC's Notice of Violation, dated December 5, 1979, the company downplays the seriousness of s.
the training departm'ent problem.
See p.33, supra.
- Licensee, however, assures the Commission that "[a] shift technical advisor has been added to the normal shift complement and substantial additional attention
- will be directed to the operating experience of similar reactors and the nuclear industry as a whole" and e
i-9
-u-n w
.n
~
..-p.
m-.
_ = _ _
b
. "[a] major revision and expansion in the training procrams for the operating organizations has been made..."
Licensee' performance in meeting these cbjectives, as well as in correcting the indisputable problems unmasked by the accident, such as training's failure to prepare operators for actual emergency situations, can be quite accurately evaluated by again turning to the BETA, and RHR consultant reports produced this year.
In particular, the reports make the following findings:
-- There are a number of problems associated with the STA program...Our observation is that [ proper STA's training) is not being done...There is a serious lack of understanding on the part of the shift Supervisor... on the role of the STA...
There is also a lack of understanding on the part of the STA's as to just what role they are to play, particularly during the vast majority of time that the plant is not in an abnormal mode.
BETA at 70.
-- There exists a lack of supervision of instructors in the TMI Training Department....In Eome cases, it was because cupervisors who were present did not react to situations where instructors were not performing their assigned a
tasks....In other cases, it was noted that there just was i
not any supervision present.... It would seem that this finding should be unnecessary considering the seniority and experience level of the training staff....However, based on the observations made, there should be concern over classroom performance.
BETA at 58.
(*This finding is particularly significant since the ASLB made such supervision a condition for restart. PID 1 2421]
-- too much emphasis is being placed on proving to the werld i
t' hat the training program is good and not enough on doing what should be done to produce a competent operator.
BETA at 57.
-- only 60% of those who responded agreed that the content of the last exams was job relevant and only 1/3 agreed that t_he oral portio.n of the exam tested how one would act in an emergency.
RHR.
-- most considered that the training department is not oriented to the needs of the op,erators. RHR.
p
-- there is strong agreement that there is not enough training on plant conditions. RHR.
-- operators complained of a lack of convergence between training, testing, and the ability to operate the plant.
3 out of 4 denied that training prepared individuals to pass exams and is successful at this but it doesn't prepare them sufficiently to operate.
To compound this, what is taught in training is different from what they experience in the plant. RHR.
The validity of these findings was substantially confirmed I
e by Administrative Judge Milhollin in his April 28, 1982 report on the reopened proceedings.
Thus it appears that one of the most significant causes of the accident, one of particular concern to the Commission in its August 9, 1979 order, has not been rectified.
Despite what arbitrary conclusions the ASLB chooses 4
to draw, the Commission must recognize that with such problems sti11 rooted in the training department, Licensee is simply unfit
'L at this point to operate TMI.
- 3. Management structure.
The Commission has had difficulty throughout this hearing process attaching significance to the many organizational and operational pre-accident problems, because of Licensee's 4
insistence that despite what may have gone on in the past, the organization has so significantly changed that those problems no s
longer have relevance.
- See, e.g.,
B&W 356, (GPU's final accident investigation report, " management" conclusion).
However, what the B&W record illustrates and what the ASLB fails to acknowledge is that the most fundamental organizational 4
problems at TMI have never related to the organizational structure of the company, but rather to the manner in which the organization functions.
The most pervasive functional problem 1
+.y
--+?F--
177--
-%+
=
_-.wrf-mr em7 m
~y
'1M-w S-ev w
wy -
=. _.-.
~
i.which can be detected both in the B&W record and in the BETA and RHR reports, and which most importantly was not an issue examined i
by the ASLB, is the lack of " accountability," or the failure of the organization to take responsibility for its failures or a
r mistakes, or to instill such a sense of responsibility in its employees.
It is a problem which is characteristic of individual people within the organization as well as the co'rporate institution as a whole, as post-accident events have clearly shown.
The accountability problem manifests itself in various ways.
Gary Miller described certain aspects of the problem in Miller's post-accident interview, B&W 360.
He was particularly affected by it because at the time of the accident, he was working 70-80 hours a week because everything had to "go through" him. Id. at c
- 7. Miller explained,
[
-- You know that call in the morning (of March 28, 1979] at 4:01 told you something.
It told you I was called very i
i early and very quickly.
Not by a superintendent, by a shift supervisor.
That's the was we run this place.
The station i
superintendent was that involved. Id. at 6.
-- I don't think we as a company were willing to transfer the accountability down. Id. at 9.
See, also, Id.
at'54.
Mi,11er also describes the problem in terms of the particular
,s difficulty he had in firing people working under him.
To Miller, i
this demonstrated not only that " people have never been held accountable" but that.he simply never had the authority he needed "to make people accountable."
Id. at 55, 56.
Miller continues, I am saying accountability is....taking a chance and l
+
v the guys going to make a mistake."
Id. at 57.
Finally, at the L
l F
---t-v vwwwww*wi=-(
w y
wy7yvgy yw-ep-
--w' g-'
-r--wwrw w--v'7y-w-
w rs------*m,wr9-=wr py
. end of Miller's interview with Keaton and Long, the following exchange takes place, Long: Yeah, I know.
I think that's the key right now it is.
That's'the feeling that people have its unacceptable to make a mistake and so that goes with it.
I am not going to make the decision, I'm going to make Gary make the decision and then if there is a mistake, at least he gotta go with me.
Miller: That's two of us.
We do tend to committee j
everything.
Okay and I think that makes it hard to do anything.
But that still comes to accountability.
If a guy don't do his job do something about it.
Keaton: You're scared of getting accountability.
We had a hundred and one thousand examples of that on March 28.
(emphasis added).
Id. at 58.
Moreover, the problem was nothing new.
In the 1978 management report, the auditors concluded that "[m]ost supervisors feel we do not hold people accountable, which in turn creates a " buck-passing" atmosphere and allows weak supervision 4
to continue to perform in such a fashion.
Not only do we need to give TMI the authority to make more decisions but we need to push many of these decisions down a level or two within the TMI organization."
B&W 843 at 45208.
Also, "[mlost personnel feel they have the responsibility but not the accompanying authority... (u]pper management is involved with too much detail..." Id. at 45216.
The report also characterized the s
problem in terms of
- bypassing the chain of command" which it concluded was "somewhat widespread."
Id. at 45214.
The problem was also specifically recognized in the 1975 management auditT-See, B&W 883, p.
- 3. And Glickman's 1979 audit found that the " tendencies of upper management to meddle rather than 6
manage has,to be curbed."
B&W 32 at 21.
The problem still persists throughout Licensee's organization.
For example,
-- BETA repeatedly heard the complaint that too many decisions are made at too high a level... It was felt by those interviewed that this phenomenon originated at the level of the Office of tne President.
BETA at 112.
there appears to be a reluctance within the GPUN system to take action either to improve the performance of poor performers or to terminate their employment. BETA at 114.
-- BETA was informed by a number of people that contributing to the inability to move quickly on personnel matters is the existence of a number of high-level corporate committees, that seem to get involved in too many issues that should be handled routinely.
If this is true, it is another example where decision-making within GPUN has been elevated to the point that it takes inordinate time to get a decision, and just as important, people at the lower levels automatically push the decision upward to avoid future reversal. BETA at 94.
-- another result of this situation is the feeling at the lower levels that by sharing the responsibility for decision-making among many, somehow no one person has to take the full brunt of the blame if things go wrong.
BETA at 113.
See, also, BETA at 21.
Further, RHR reports that "two out of three deny that management has coamitted to an accountable organization which resolves problems at the correct level."
And in testimony before the Subcommittee on Energy and Environment, House Interior Committee on April 26, 1983, TMI-2 "whistleblower" Larry King spoke candidly of the confusion about organizational N
authority at Unit 2, the " organizational breakdown," the inability to decide who's in charge, decisions by those responsible overruled by managers so that departments cannot
' carry out their duti*es, and unclear lines of authority.
Perhaps most importantly, this very problem can be seen in Licensee's response to specific incidents examined in the restart
',I hearing process.
In particular, there has been a consistent
/.
pattern by the corporate entity of refusal to admit wrongdoing and take appropriate responsive action.
For example, Licensee still refuses to admit to or appropriately respond to its failure to provide accurate information to State and Federal authorities during the accident.
Licensee refuses to admit that leak rates were falsified and take appropriate responsive action as as result.
Licensee responded to the cheating incidents in a wholly inadequate manner on many levels.
See, TMIA's br~ef on exceptions to the PIDs.
The overwhelming evidence confirms that Licensee must be fundamentally incapable of taking responsibiltiy for its own wrongdoing, and that clearly it has sent a signal to its employees that this is an acceptable way to operate.
- 4. Manaaers 3
Finally, the B&W record reveals a tremendous amount of i
information about the competence and integrity of individual managers, whereas the ASLB seemed to make a conscious decision f
not to probe into the record's of particular individuals.
As an example of how the B&W litigation produced some very damning i
evidence on certain people still Within the organization, we have compiled a list of some of the more revealing statements and x
criticisms of Bob Arnold, President of GPU Nuclear:
-- when asked about the 1978 audit findings concerning " lack of trust in supervisors by upper management" and the " buck passing" at TMI, Arnold claims he does not recall those findings and would not agree with them anyway, Arnold dep.
l
-at-3 8 7-3 8 8, yet both show up in a similar form in Gary l
Miller's post-accident interview with Keaton, B&W 360, and in the the BETA and RHR reports. (See discussion on l
accountability, supra.)
s.
-- whe'n asked about Gary Miller and Robert Keaton's statements in Keaton's post-accident interview with Miller, t
B&W 360, that it is almost impossible to fire anyone at TMI, Arnold does not agree, and does not remember Keaton saying this, id. at 387-389, yet the same allegation shows up in the BETA report. (See discussion on accountability, supra.)
-- when questioned about complaints by the operations staff that the work load at TMI-2 during 1977-1978 was too great, Arnold claims he does not agree with this " assessment," but when questioned on Gary Miller's statement in B&W Ex. 360 about routinely working 70-80 hours per week, Arnold rationalizes that "the need for very heavy commitments on the part of senior members of the staff is inherent to some extent at least, in the nature of finishing up this type of project."
Arnold dep. at 369, 370. See, also, TMIA management findings 11 39, 44.
-- Arnold admits to being the " focal point" for resolution of problems demonstrated by the 1978 audit, which he also admits were the same problems found as a result of a 1975 audit which he initiated, id. at 347, 350, 353, which were the same problems found in the BETA and RHR audits.
-- when questioned about Gary Miller's assessment during his January 1979 briefing to GPU management, that there exists a
" dollar crisis" at TMI, Arnold retorts that Miller really meant that the issue isn't total dollars but " compensation packages."
Arnold dep. at 392.
-- Arnold does not recall knowing at the time that Mr.
Zechman, the head of training at TMI-2, failed to pass his operating license exam between the. fall of 1978 and the TMI-2 accident, nor did he think there was any problem with Mr. Zechman not holding an operating license, nor any problem with Mr. Zechman, while supposedly head of the training department, studying full time for the operating license he couldn't seem to get.
Arnold at Tr. 1705 1707.
-- Arnold admits that Met Ed failed to instill in the operators a sense of respect for the " upgraded" post-accident training department even in light of the TMI accident.
Id. at 1790-1792.
-- Judge Owen made the following observation about Arnold at Tr. 1732:
"[T]his witness has put his credibiltiy on the line as to being all knowledgeable about everything that went on in this company and that all was right in the world.
That was the impression that one sort of got from his direct examination, which was that he was on top of everything, they had these boards, they had this board, he personally was taking part in this, that and the other and everything was fine.
And the Floyd incident, as I hear it here; arguably attacks that status l
because you wonder whether somebody who is that j
knowledgeable about everything that's going on could have 1
h
l i been aware that somebody who was in a supervisory capacity j
was having all of this trouble without him knowing about it.
I i
-- Arnold still does not recognize that Mr. VV cheated in j
1979.
Arnold dep. at 461.
-- when questioned on how to identify a LOCA, Arnold still insists that pressurizer level and pressure trend together, despite the opposite experience of both the April 23, 1978 transient and the accident itself. Arnold at Tr. 1459-1461.
-- Arnold seems still to have a poor understanding of a i
(
~
fundamentally important concept concerning how a steam bubble will react to an increase in pressu're within a thermally insulated container, implying that condensation will take place with an increase in pressure.
Arnold at Tr.
1605.
One of the most important lessons of the September 1977 hot functional test transient, where it took days to remove the steam which had formed in the reactor coolant system, Tr. 1473, seemed to be that without some method to remove heat from the steam space, a steam bubble can not be compressed by a mere increase in pressure.
-- Arnold did not consider the recurrent problem with the condensate polishers of water getting into the instrument air lines as representing any threat to the safe operation j
of the plant, Id. at 1498, despite the fact that this very malfunction started the TMI-2 accident.
Thus, a substantially different portrait is painted of s
Arnold in this record than was permitted in the ASLB's main i
management hearings.
Clearly, Arnold's continued participation in the management structure of Licensee must end before his influence contributes to the escalation of another accident at TMI.
i III. CONCLUSION In 1976, Gary Miller warned in an ominous memo that "the current philosophy is to ignore Unit 2 with the thought that it
~will 5E acc'omplished' eventually when really needed.
I consider
~
t this an item we will pay our price for in terms of hours, feelings, and dollars."
B&W 764.
Lfcenseemayhavepa'id'a
--,2
-. ~, -.. _ - -..
...v w
.-n--
,.-,,..,,,,v,,,
,w-
=
. price, but they seemed to have learned nothing from their misfortune.
As B&W's attorney stated in his opening, "if there is anything shocking in this case, it is the pervasive pattern of appalling indifference shown by GPU management... to everything from maintenance and training to following the requirements of their own procedures which woul.d have prevented the accident."
The same patterns are there today -- the same patterns persist at TMI-1.
This record illustrates that Licensee is simply incapable of learning from its mistakes or even of responding honestly to them, and that the ASLB simply did not come to the correct decision.
Licensee cannot legally be entrusted with the enormous responsibilty of operating TMI-1.
c.
Respectfully submitted, i
e i
By:
M a
,C Joanne Doroshow Louise Bradford July 1* 1983 TMI Alert g
h 4-g
~. ~
. -, -