ML19343D481
| ML19343D481 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 09/30/1980 |
| From: | Arnold R GENERAL PUBLIC UTILITIES CORP., METROPOLITAN EDISON CO. |
| To: | Moseley N NRC OFFICE OF INSPECTION & ENFORCEMENT (IE) |
| References | |
| NUDOCS 8105040611 | |
| Download: ML19343D481 (10) | |
Text
.s 50-320 Metropolitan Edison Company Post Office Box 480 gj Middletown, Pennsylvania 17057 717 9444041 wnter's Direct Dial Nurreer September 30, 1980 Norman C. Moseley, Director Division of Reactor Operations Inspections Office of Inspection and Enforcement U. S. Nuclear Regulatory Commission Washington, D. C.
20555
Dear Mr. Moseley:
During my interview by NRC on September 5, 1980, I referred (tr. 41) to an investigation which had been initiated by GPU in 1979 regarding the state of knowledge of various portions of our organization in the days immediately after the accident. Enclosed is a draft of a report on that investigation, pursuant to your request (tr. 43).
The enclosed draft report, dated May 3, 1979, was never reviewed and finalized, although in its draft form it has previously been supplied in connection with other investigations to the President's Commission on August 24, 1979, the NRC Special Inquiry Group on September 5, 1979, and the Hart Committee on October 18, 1979.
Sincerely, yaw n
l l
Robert C. Arnol RCA/csk Enclosure cc: Ernest L. Blake Jr., Esq.
l l
l l
810 5040Ql f/etropouan Edson Company is a t/emter of the Genera! Potyc UR!es System
h,( f..3;
'I m! Lh,"
D 4
TW 781= ns*
i
/.',.
DEVELOPMENT OF UNDERSTANDING
/) ly.n o
- 1_ HJ The object of this task is to re-create, as best as possible, the chronology and an on-going status understanding of the TMI-2 accident
- as it developed.
It is obvious from the sequence of events and the response actions taken by the aparators that perception and understanding of the transient were changing, especially in the early hours.
Levels of complication in assessing this devaloping undarstanding are introduced by the various connunication interfaces which were established about the data source. These interfaces. provided data (as well as on-going assessment) to an increasingly' growing number of people who inturn contributed to what may be regarded as a co=non understanding k<
of the accident.
As tir.e wore on the number of connunication interfaces grew geometrically,. nd be.:ame so intertwined as to give rise to an apparent common pool of understsnding. Consequently the task of unraveling the growth of understanding becomes increasingly more difficult as more co:aunication interfaces were established (later in time following the accident).
Necessarily then, the subject assessment must be attacked starting with the TMI-2 operator's understanding and procede through that of the site management, Met-Ed management, GPU management, the GPU response l
team, and the industry advisory group. Further the peripherical L.
ay s'
2.-
(
branches of BW and NRC understanding could also be examined. Figure 1-attempts to ' illustrate the challenge and the scope of the above, hgg[I An analysis of the perception ofsignificant problems and growth of knowledge during the event can be divided into three areas:
A) Efforts to maintain control of the plant during the first four hours, B) Assessment of the radiation emergency, and C) Perception of th'e non-condensable gas b<.oble in the reactor. vessel.
A.
Efforts to Maintain Control of the Plant During the First Four Hours of the Event, the main concern of the oparators was to bring the primary and secondary systems to a stable condition. Several key factors should be discussed to. focus on the basis for operator actions. Details used to formulate each of these key factors were derived from interviews with the shift suparvisor,
{
shift foreman and two control room operators.
1.
Pressurizer Level Indication - From very early into the transient, operations personnel were very concerned with pressurizer level indication.
Within five seconds after the reactor trip the operator had started a second make-up ptr:rp in anticipation of the expacted rapid decrease in
- r. : I level never occurred, and within six minutes the pressurizer level was off scale high. The operators felt they had caught the expected level decrease with increased high pressure injection. The major concern of the operators at this point was to not take the R.C. system solid.
Based on high level indication and concern of taking the system solid, the oparator bypassed Safety
- Injection, stopped MJPIC, and throttled
(
hteh4q Eved
%d 3'uskoMen;Eh,M h
i V
lhhukilbx C B/'ah:cs A> bos4 be P
SWecadI.cd A
A Aqemes 2
(
11 Shhig
&4SC
& Mt Fawymad ff NRf A\\
a 3
y 7
D IAG deh-Ed
/ ream
\\
Faseqenec y
f s
6FQ damemec Fqore 1. OoMWatcNM MNN5 whtdt onablad Oe devel.0pmetd 0- Ottc6'S;Budr l
RJ.
2
- 3. -
(
the MU-V16's.
At this point RC pressure was still decreasing, and
-judgments were made primarily on pressurizer level based on either past experience during feedwater transients or reactor trips, training, or procedural guidance.
2.
Failure of the Pressurizer Electromatic Relief Valve to Close -
During the initial reactor coolant system pressure increase due to the turbine trip, the electroastic relief valve on the pressurizer opaned, as designed, at 2255 psi. After the reactor trip, the valve failed to close as pressure decreased through 2205 psi, although the oparator did verify that the valve indication did not signal an opan valve.
This valve remained opaa for the first two hours and twenty-two minutes of the event. Thus, the control room operators had failed to recognize a constant loss of coolant through the opan relief valve for
(
that pariod. On at least three different occasions, the operators checked the computer output for the thermocouple bands on the relief valve discharge piping to determine whether the valve had properly seated as indicated; howaver, the computer data was misinterpreted and the block valve was not shut for two hours and twenty-two minutes.
0 Readings from the thermocouple were in the range of 230 -280 The oparator judged these va1Yes to be quite low compared to pressurizer
~
3 temperature (approximately 603 ) and therefore concluded that the electromatic must be closed.. The oparator did not realize that.the temparature indication was from a themocouple strapped to the outside of the discharge pipa and. based on heat losses, readings in the range of 250 were an indication of an open relief valve.
- D -
s
.4 During the tim that this valve was open, taany other indications
- of a loss of coolant accident ware present in the Control Room, such as rapidly falling RC pressure, rapidly increasing.RC drain tank pressure and temparatu're, increasing RB sump level while both sump pumps running, increasing RB temperature and pressure. Althots h'
- ,,... ira rs 1.
these indications of a,LOCA existed, the op3rators did not associate them with the stuck open relief valva. The operators. continued to focus on the initiating event (loss of feedwater) and attempted to deal with the consequences of that event on the condensate systes and the turbine heat sink.
3.
Reactor Coolaat Pump Operation - At one hour and thirtean minutes into the event, two Reactor Coolant pumps were tripped due to observed " flow fluctuations" and allowable NPSH requirements of opar-(
ating four RCP's.
Approximately thirty minutes later the remining two Reactor Coolant pumps were t' ripped based on similar concerns.
Approximately two minutes after the Reactor Coolant pumps ware
~
tripp2d, the operator began to raise steam generator 1cvel from thirty percent on the start-up range to fifty parcent on the operating range to further induce natural circulation. Within the next thirty minutes RC hot leg temperatures ware increasing to off scale (greater than 620 ).
It was realiced that natural circulation
~
was not occurring probably due to a steam bubble formation in the -
A loap (the B loop was isolated). Attempts ware made to again start a RC pump to force circulation through the core; however, attempts were aborted due to pump motor low running current. (The 2B RC pump was run for a period of about 19 minutes some 75 minutes after the
- 1ast pu=p was tripped.). The 1A Reactor Coolant prap was started
q
/
,a
-5~
t approximately fourteen hours after all pumps had been stopped.
It appears that the actions of securing the Reactor Coalant pumps were taken based on minimizing equipment damage (i.e.'to-
~
preveat pump cavitation and protection of pump seals) and not based on an understanding of what was occurring in the primary system.
B.
Assessment of the Radiation Emergency - The growth of knowledge in this area developed rapidly approximately two hours.
and forty-five minutes into the event, as the first radiation monitoring alarms were received throughout the plant. This know-ledge was accumulated quickly, and used effectively to determine action 1cvels according to the radiation emergency plan. Although the extent of fuel failure was not realized this early in the event, the perception of the significance of the radiation monitoring h
system readings.was accurate.
C.
Perception of the Non-Condensable Gas Bubble in the Reactor Vessel - Early in the evening of March 29th, a group of engineers met to discuss present plant status. Two of these engineers reparted to the control room to back up the oparating staff. By approxi-mately 2100, it was apparent to this group that a non condensable bubble existed in the reactor coolant system. Prior to 2300, a formula was derived to calculate the size of the gas space in the system. Cas bubble volumes were routinely calculated throughout the 30th of March and calculated volumes began to decrease late in the evening of the 30th. Based on analysis of. the 3/28 reactor building pressure spike and containment air sarple analysis which began at approximately 0400 on 3/31, it was deter:alned that the gas bubble in the reactor coolant
y -
e
.,g
-64
'(
system was prissrily hydrogen. Volume of the gas space' decreased steadily through April 2.
This was confitn2d as the increase in hydrogen-concentration of containment atmosphere leveled'off as the bubble in the reactor coolant system diffused.
Ink &es-- na 'first management co munications concerning the TMI-2 accident ware drafted in a telephone conversation between Mr. Herbein and Mr. Fabian approximately 7:15 a.m. Wednesday, March 28.
At this time, they mutually drafted a statement - for response to press inquires that related that the
~ TMI-2 reactor was ' shutdown due to a malfunction in a feedwater system.
The entire unit systematically shutdown and was expected to be.out of service for about a week while equipment is' checked and. repairs were made.
At approximately 9:30 'Jednesday morning, Gary Miller called Mr. Troffe-to relate his conversations with Lt. Governor Scranton concerning the.
uait status. During these conversations, Gary indicated that there was some fuel pin leakage, however, he noted that he didn't have any indication of fuel malt. ~The prepared statenant to the press was updated by noon, March 28.
This statement revealed radiation levels were being monitored in and around the plant and that there had been no recordings of any significant levels of radiation and none were expected outside the plant. No avacuation of the local population was indicated at that tina and that the reactor was being cooled according to design by the reactor coolant system and should be cooled down by the end of the day, March 28.
It added there was no danger of a melt down.
During the Yet-Ed press conference in Hershey, on March 29, Mr. Herbein said it was too early to tell the extent of the fuel damage at TMI-2.
However, he noted that fuel failure had been experienced during the t
accident. -He related this fuel failure to the point of turning off the reactor coolant pr:ps during the transient. He updated the plant-I
- ?u
- ' ~ *f^
status to say that a reactor coolant pump was running and co31down was proceeding and that he expected to be on the decay heat system in approximately 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />.
In response to questions from the press, Mr. Herbein related perhaps one half to one percent of the rods may have experienced some melting and that the fuel had primarily failed due to the reactor coolant system depressurization and.the need to shutdown the reactor coolant pumps. He noted that it was passible for sone steaming in the uppar core' region at that time that lead to the fuel failure.
Early in the evening of Thursday, March 29, Mr. William'Lowe, Mr. J. P. Moore _had gone to the Unit 2 Control Room to assist the operating staff. Based on observed indications this group assessed that there was a non-condensable gas bubble above the core. Later that evening, calculations began to determine the volume of the gas bubble.
(
During the press conference given on F.cch 30, Mr. Herbein revealed the evidence of the gas bubble a'bove the core. Hovaver, he notti that it appeared that the fuel assemblies were covered at that time and that decay heat renoval was progressing. He suggested at this time-that the fuel failure was caused by a momentary uncovering of the fuel during the transient.
During the press conference on March 31, Mr. Herbein revealed that efforts were underway to reduce the size of the bubble over the top of the fuel.
Initial indications indicated that the venting process was successful and that the bubble had reduced in size.
He did mention at this time, however, a concern that the venting process has lead to a build up of hydrogen in the reactor building. During the evening a m
l
es-a t. c.
s-sample of the reactor b'uilding atmosphere has been taken and that
~
at this tim there was no. danger of an. explosive mixture in the -
reactor building.
The first results of the reactor coalant analysis were received on March-30th.- Based on these results, Mr. Herbein noted in the Match.31 press conference that the core was indeed severely danged and that there was a possibility that a,very large parcentage of fuel-assemblies were in the dam ged condition. This March 31 press conference was the last held by Met-Ed.
After that time, conunications concerning the plant status were handled by the NRC.
9 9
O F
o l
1 1**rw
-w-
+-
.--,e
p.
sabCOCh&WI!COX wr cenettien cioua P.O. Box 1260. Lynchburg. Va. 2 *105 Telephone:(804)284 5111
~
April 15,1979 3g
,N
)
Mr. G. P. Miller Station Superintendent Metropolitan Edison Conpany Post Office. Box h80 Middletown, PA 17057 a:=
Subject:
Review of. Unit II's March 28, 1979 Transient x.
Dear Gary:
As agreed in the Saturday, April 14, 1979, get-together in the Superintendents Conference Roo=, I submit a few of the thoughts that have passed through I
my nind during and since those~ review dischssions.
I believe we all got a lot of =erory recall benefit out of that session, plus a feeling of being together on so many other thcnghts.
Personally, I believe that we all really have to pull together = ore than ve 'ever may have before in order to accesplish an enlighten =ent of Investigative Groups and the Public in general. Met-Ed and 'DII, including in no s=all bit Btni, have re' ally taken severe shots'by the sedia and the KRC in the public forum. We must also exercise our rights in the same public for:.:n to correct and educate the rest of the vorld.
I know what we did,and I also, knov that our collective actions and Met-Ed's real (not imagined) innge is of a very high technical cud moral standard. Our biggest task, as I-see it, is to bring out the facts without confusion and embellishment ir.
such a manner (not very technical) that cost people vill' understand what
~
ve are saying and thereby change all of the negative inpressions.
IT.itur al.ly,
the anti-nu%es von't listen because their minds are not allowed to ~oe spen but. there is n very large segment of the general public that, vill listen because they really do vant to knov the truth. Any one of us involved
. rust keep it in our sinds that the real end of the tunnel is to have both Unit I and II back on the line, a little safer and ve operations people a lot snsrter because of Ibrch 28, 1979; but really in that node cf operation with a lot of the general public really backing us up.
l.
E00R BRIGIN
=
The Babcock !. Waccx Cemcony / Estabtished IS67
.:u. - _-
i i
s G. P. Hille- -
b/M/79
' - - ~.., _ _
\\
. ~. ~,
~
If you have any furthr questions, please do not hesitate to contact me.
.p..
Very truly yours,
~
,w L. C. Bogers Site Operations Manager
~
s g
1,
. j,
. t
+
t.
?
e e
p e
y
- t g
e e
Q es. e*
l l
t
., e4,
h e
9 1,
t t
j
.sg 6
l
~
~
P00R ORIGINALi e.
Y*'
-~-
3
...,.,-_-..._-.._...r.--
1 OTHER REFLICTIO!!S A!!D REC 0ZIE! DATIO *!S 1.
(a)
During significant occurrences as Station Energency and General Emergency at the DEI Station, Met-Ed should have a designated
[
s * ~ w. individual on the hergency Bill, that is " qualified" and" recognized vith Ne_t-Li, G7U, 2&W, PA Stats, ITRC,' EPA, 3.stf-site officials, etc., throti hout the enerr ency time. He should provide status, data, and expected evolutions to all of these outside parties until such time as they are able.to provide their own on-site linkups.
(b)
As I observed events in the communications systen'during the long day and several days subsequently, it was apparent that several of the outside parties were given necessary infor.ation but apparently..
-in.different sections or their organizatio'ns; and they were not talking to-each other, thereby, creating additional qrestions coming from several sources.vithin the same organizations. EXA'!PLE:
The NRC people on site were on. an open ' telephone line to a " situation room," I believe at the Region El~ office (not sure), and infor:ation was flowing in a generous fashion.
Yet the NRC headquarters was in turn generating questions to the site independent of their own on-site inspectors.
In fact, they were not even tal'<.ing to thes (the NRC people) but in turn asking for plant personnel: to provide
' the ansvers and also directing questions and de= ands to B&W Lynchburg.
(The Lynchburg source for any ansvers during any developing. crisis is not an acceptable flow path for outside organizations since B&W is by riature and geography not able to be on top of rapidly changing conditions.) Another example vas' that the cc=sissioner of the NRC, as quoted in the media, was
~
in the darE and thoroughly confused. 'I submit that he and his deputies have technical advisors closely at hand durin6 significant events to interpret the information already i', the IIRC at other
, areas which need a central tie-in mechanism to allov the top
' decision makers the chance to make good ultimate proclamations.
E In other taped interviews, the Shift Supervisor ~ identified two B&W people, that were assisting Unit I in their startup program, as arriving in' the
, Unit II Control Room.
Subsequent questioning. chovs that these people vere not B&W people..In fact were Scott Wilkerson (Met-Ed Nuclear,
Engineer) and another Met-Ed employee.
There vere no E&W personnel on
. site until I arrived approximately.0710 hours0.00822 days <br />0.197 hours <br />0.00117 weeks <br />2.70155e-4 months <br /> on fS):! arch 1979 (point of clarification).
3.
Respecting the normal human concern and also training towards that concern to not aggravate plant operatin6 conditions or cause'damsge to plant equipment, any action similar to securing "all" reactor coolant flov during transients must be drilled into the operators and. supervisors as g
an action that should not be automatic but tailored as a case basis, i
This type of sucgestion is going to be difficult to imple=ent since in one transient as example, securins pumps vould be absolutely correct and in another it vould be an action that would tend to aggravate the problems.
pggg gR\\Gl%
{
- W W='-
..a.,su.n
&.-- ~ ~--.----
~ ~ ~ ~
~
e=
r
(
N
/kELICOIONS & RIC0!OIENDATIONS
/
4.
A need is identified nov to evaluate a21 possible system co=munication, '
~
conr.ections bet /een the reactor building.and the outside environnen&,
such as the auxiliary building, fiel buildir;;, direct outside, etc.
y These evaluations need to look at nor=al pumping systems, E/P driven systems (press in reactor buildir.g and not other places'), D/P. following pumping actions and merely stopping of the pu= ping (siphoning actions) air-borne paths, and all of this could be a major undertaking. Needs high management type emphasis because of the auxiliary building contam-ination problem on this transient, although the installed normal systems were not supposedly lineiup to allow such flov between buildings. This needs attention and corrective action follow-up.
5 The Site Emergency Plan / General Plan needs review. My personal experience was that when I arrived at the North Bridge Gate, your plan.vas in effect -
entry was being denied to traffic.
I vns recognized by yoEF ' guards as being*needed for the problems. They gave no my 002 " red" badge and opened the gate allowing me entry.
I arrived at the area of the' Unit II turbine building access.
Steve Drabick was on duty.
I vent into the Catalytic Building and left my briefcase, picked up my hard hat and valked across the street. Steve Drabick said, " Lee, you cannot go in."
I asked him to call the Unit II' Control Room on his radio.
He did'and gave me my " green 025 security entry badge." I vent in.
However, Steve Drabick was very busy at that time turning back all of the normel day shift craft workers that had entered through the South.Eridge entry.
In fact, I think I remember a bus being there.
I talked to Jim Elanton (Catalytic supervisor) and told him that the plant had tripped and had experienced "co=plications."
The point,is,that a lot of personnel vere coming onto the site from the South Bridge and the "3rass Gate Entry",at a time when the North 3 ridge was tight and controlled. Needs a serious look at effective closing both bridges and the brass gate.with the e=ergency; condition announcement and follow up accountability at so=e emergency designated area..
6.
Need to provide continuous recordable (retrievalbe) instrumentation of the vital natura; incore T/C's is an exa=ple of ve'ry valuable data not retrievable nos and would have solved many of the advertised concerns.
A survey of. vital, needed instru=ents is certainly in order.
These do not necessarily need to be displayed in an area of the CR0 but should be,able to be recovered post transient.
t 7.- A thought about corrective maintenance needs on the CR0 available instru-mentation.' Met-Ed needs to establish a pr.ocedure for identifying priority of corrective niintenance on critical instru=entation, i.e., pressurizer i
level instrument, one channel was unreliable due to known panel switch l
problems. This problem and others like it should have a required repair j
date and techanism for repair if plans operations are to continue.
t 8.
I have been informed that EMI Lynchburg has incorporated the T:1I-2. transient i.
into the Simulator Programs. Other Utilitics are cycling ~their operators thrcuch the simulatien en a crash basis during the nicht shifts.
I submit i
that Met-Ed seriously consider a similar procran for their licensed operators acquainting everyone with the indications end actions that can be taken j
to lessen a reoccurrence. This -is in preparation for the Unit I return to cperations that we all are looking forward t a
n A
8 l
frane.
g g
. ~
= - _ a -., =.z -_.
-