ML20129J501
| ML20129J501 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 01/21/1980 |
| From: | Udall M HOUSE OF REP., INTERIOR & INSULAR AFFAIRS |
| To: | Ahearne J NRC COMMISSION (OCM) |
| Shared Package | |
| ML20127A105 | List:
|
| References | |
| FOIA-84-311 NUDOCS 8507230099 | |
| Download: ML20129J501 (125) | |
Text
{{#Wiki_filter:W .:- .
%
- s . . :,4 d . 2. < =s . a s .. ..;.~,,;.,,..
s . w . , e . < . .. *. '-- * --
.- y.m.o. .. O
.s...-.s..<,...u.
.... c Mgl;Trr osJ.;I, ::. AO :. 5.'.. : ;J7 A!Rs g,,,,..,,,
- s. .
. + e i n.=3 .
- n. . ,. .h ..: A. ..ys r.g. .;%... tw u c r-c . .- u . .s r.
,.es. g;,g , g. .. .; e e s. p,cn :, re i.-~c.;;
- s. e . e .
r-e. . . .. 3..e
. s ..s,.,.e... .. .es.a.
c cu, e.c. ::::3 new n n== secc..u em.sc .- s, : ,,...,.a.s3.%s s w ..s.....c
. m. .. m. .,... m ,.w... ... . . m s
.ert. -sa. ... ..... een ussis.4 io.s .
- u s. . v. r : ..
E n. ..sa.n.e s s. .s,r. es.a s a. ca=v c. ev.rwonta
. sa.
sw.
...,..ev.--
. s .. e4w n.
January 23 , _3980 e .. ,r, e e,..c, ---cc.72 - - s.; L - . a.u,.a ,,.n m. su.a. -.. ; +).:-
..n s rse a.
. a. r. sum. ===. ,,,,,,,C
..e =.vi a - - - -
W8 A.8 88 S. G.w es tv.s & WJ. ; ...
- . .n o n. = = n. us. I
.'T* EGG.Seeft.9 % pg ., % '. 4 4 WS.4 P A. ;;.:
rTt A. esws.wv. Ps. ..,.c
-*=ensa a.m. .. ~~-
. e ,. . s
<a$<-... ~~" $
..a e<segg7., m.g, 89 wa. , .gf.
/t.4T .A. .. un.m p Tet. 43 e , tape, ...
p.9 menewsgr. 3,s.g.
*.T *.64 6 &. al.re. .= . . .
. E; The Honorable John Ahearne
- -"i Chairman, Nuclear Regulato y Commission f; Washington, D.C. 20555 . . . .
.. .+.
=u
- p. :.=
Dear Mr. Chairman:
"s On March 28, State a.nd Federal officials were tinaware {3 r --#
of information indicative of conditions at Three Mile ~# Island. As.a. result of being inadequately informed, F. ..~. c.overnment officials were severelv. imo.eded in fulfilline - f.;:c.".~. their responsibilities with regard to protection of the e.' ]
' An important cuestion, public therefore, health and safety.is whv on March 28 these officials and the .
~.
.-=-
"~
public were cenle: important information. of its I believe that the Commission to the best ability should address the following before closing its incuiry into the accident. . Station Manager Gary Miller has stated that, and "The e
- 1. r in-cores were reading anywhere from 2500 cr so, 1 1 picked 2500 (;) it could have :'een higher than that.
.Su you know, I was looking for a gross indicater ard I had it." What did Mr. Miller believe to be Did Mr.
the significance of such temperatures?
' Miller believe on March 25 that during theWhat day was cortions of the core had been uncovered?
.i r. Miller's understanding on. . March steam-cladding i ..
28 as to the .> temperatures at wnich a s:..gni:1can ~ chemical reaction might occur? . .
- 2. Is Mr. John Flint correct in his recc11ection that, .-
"These [in-core) temperatures were monitered for i
the rest of the day [ March 28] to follow .what was - happening to the core?" Jan 30
, . . . . 1/23..To EDO for Prepare Tseply for Signature of Chairman *..Date due Comm:
Cpys to: Chm,Cmrs , PE, GC, SECY,0CA to Ack. ... 80-0160 8507230099 850506~ PN [d P --311 OS
-2.% E ncrahla Jchn Ahecrne
, .?ago 2 Which of the TMI supervisers present at the ;..~~.
3. plant on March 28 believed some of the fuel rods [." 3ssi had been cooled primarily by steam for portions if= .. of the period between 6:00 a.m. and 8:00 e.m. on _ March 28? What was the understanding of these Es - supervisors on March 28 as to the claddinc temoerature at which a significant steam-cladding chemical C. g.. 7 .. reaction might occur? ,v --
."=:.... -
- 4. How does each of the TMI supervisors, who !s=:r-
* =
was present on March 28 and,who did not believe , that portions of the core had been uncovered, , 7_g interpret the hot-leg temperature data incicating g.z g superheated conditions? ,
====
- t. . ==
- 5. Is Instrument Man B correct in his recollection gjsyi
= =;:.
that the instrument technicians had told Mr. Ivan ,
?
Porter of the results of their measurements of g.;jg the in-core thermocouple voltages? ., u.: .:.
- 6. Are Messrs. Chwastyk, Ross and Frederick [;.7"
, i.m.
correct in their recollections that in the e. J afternoon of March 28 Mr. Miller was aware of the Ikdr pulse on the pressure recorder and/or actuation T3 of safeguards svstems that occurred at approximttely 1:50 p.m.? .
"I
- 7. Is Mr. Illjes correct in his recollection that on March 28 he was told about the pressure pulse, p" and that the possibility of a hydrogen explosion , ,:; ,.
~
was discussed in the evening? [ l:.. . Which cf the TMI supervisors who were present .; S.
'celieved en March 28 that the pressure recorder : _
nac ;ncicarec a reaa r se in containment pressure at 1:50 p.m. and which of the supervisors believed the recording to have been the result of spuricus electric signals? , i~
- 9. What cf the TMI supervisors present on March 28 were aware en that day of the following: .
-- actuation of containn.ent sprays;
-- negative pressure pulses on instruments that used containment pressure as a reference:
t L
F*
' 75~46 **O".orah1C Joh". J.h5arnO 4
,Iy C'O , 3
* - ~
-- a rise in containment temperature G.2 3
- =-i .,
^
. recorded by one or more containment temperature * - = =
- sensors? :.t:- .e-
-:::.~.: .:.
- 10. Did Mr. Miller and/or other Metropols. tan
*7
.g;.:.;.;
Edison officials have an obligation to report on.. .. March 28 any or all of the following to state and . .. Federal officials: 2.T..:-.f
-- in-core thermocouple data indicating . f"li
' temperatures in excess of 2000 degrees F. ;
@EEE .-=
=:.
-- computer printouts of in-core ..: gel-
~;s%
thermocouple voltages indicating. presence of superheated conditions (and partial core uncovery); .bf
..=.
-- hot-leg temperatures indicative of t;- ::.
superheated conditions prevailing for most of @{ the period between 6:00 a.m. and 8:00 p.m. ; . t;;q.g
. ;=.. a ME
-- the c.ressure pulse recordinc, and other . . . . _
manifestations of a hydrogen detenation in the : ."t;- containment at 1:50 p.m.; and, .
...34, a
-- uncertainties during the day as to -.-
whether the core was covered? _,
~' - -
In order to assist you in your consideration of this matter, I am enclosing a draft working paper prept. ed by
- Ti Co:=.ittee staf f. I would also appreciate Cc:nission .t
~2 cc==ents on this working paper. ,
Sincerely, __ Y Ju>n 4 iy
. MORRIS K. CDALL
- Ch a i~~.an . .
Enclosure L O. 4
p - a NU 3/4/80
,'[.Y.
f% hp s Draft Workinc Paper -
~
Ear ly Indications of Core Damace at TMI-2
~
Introduction ' At.approximately 6:00 a.m. on March 28, the water level
.?: . a in the TMI-2 reactor pressure vessel' fell below the top of
'the heat producing portion of the fuel rods. During the next .
l 1735i . .. _.
' .r / -
'7 hours a substantial portion- of the reactot core was cooled
- 9 . .
intermittently by steam produced in water boiling within ' he t , pressure, p;.s. . i' vesselc The steam'was a significantly less efficient coolant than water, and for part of this period radioactive decay of fission products generated more heat within the fuel than , was removed by steam flowing past the fuel cladding. Therefore, 3 cooling was inadequate; portions of the zirconium t
}
cladding reached temperatures at which it reacted with steam, producing zirconium oxide, hydrogen and an additional
~
significant quantity of heat that needed to be removed from the pressure vessel. Prior to the zirconium reaching temperatures at which rapic oxidation occurred, increased pressure from gases inside thE. fuel rods and loss of strength due to the higher temperatures caused the zirconium cladding to bulge and split. Kemeny Commission staff concluded that all or nearly all 36,000 rods balloned and ruptured, releasing the radicactive gases contained therein. (Kemeny, Core Damage, p. 4.) Kemeny Commission staff also concluded that-during the period of core uncovery fuel temperatures in the upper 40% to 50% of the core exceeded 3500 degrees 7. , and temperatures in 30% to 40% of the core exceeded 4000 degrees F. (5ormal fuel temperatu5es during cperatien are about 650 degrees F. I K regulatiens cm et r' i f e t h.s t- s .r.e *r' o
- r*** . cve'c mc -
- ct r lit A. 7 c i C: n e d c;, . tks6 .u m i.,
..J
q
\, .-.s.
- p. -
, . ,- zirconium cladding temperatures shall not exceed 2200 degrees F.
. 5 in the event of' a'ccidenEs which m'ust be taken. into account in the reactor design.) Kemeny staff estimated that ultimately
~ - -...-..... - -. ~ . - -
_ between 44% and 63% of the core's 20 tons of ' zirconium cladding were oxidized, producing thereby some 1000 - 1300
. pounds of hydrogen and ~ causing some 60% to 70% of cladding to
, . 7 .~i . . C":
' ~
gg .
.~
" i.:. eJ ,; . become sol embrittled that it lost its structural integrity.
N? . ~ f.*i? :',;& :.
'IP.4,'.. (loid. , jp,. [1) ._ -,The NRC's Special Inquiry Group estimates '
s, .+...... w.
. . less claddiEg oxidization; i.e. 31%-35% leading to production of 720-820 pounds of hydrogen. [SIG, Vol II, Part 2, p. 283.]
(NRC regulations required that reactors be designed so that no ,
, more .than 1% of the zirconium oxidizes in the event of accidents of a kind that must be taken into account in the. reactor design.)
Kemeny et al. concluded that as a result of oxidation and embrittlement of the fuel rod cladding, several feet of the upper part of the core fell into the gaps between the fuel rods, causing partial blocking of the flow of steam or water 4 that removed heat from the damaged fuel. p During the course of the accident, the TMI reactor pressures, temperatures and core geometry evolved into such an unusual l
' state that the operators were unable to achieve a relatively, l stable cooling condition until late in the day of March 28.
By 9:00 a.m. information available in the control i l room indicated that the core was severely damaged. P.igh i radiation levels in the reactor building indicated that a i large percentage of the fuel rods had been damaged to the l l y. l
. w.
7 j
.- = a
,, . . - - - - * :- . . : . r : - - 3 . :- . . -
t,
. point where they lost radioactive gases contained t.herein;
. -instruments showed that"the reactor-pressure vessel and
- . - . . . . . . ~ . . . . . . . . . . . . . . . . . . . . . . . -.
.a
. ass 6ciate piping contained steam ra.ther than water; and
,,--- inumerous temperature readings indicated that the core had been
- - . .. ~ "
em ; .. uncover'ed and the zirconium cladding of the fuel had been y- ..:. u.. C. . , . . .
. =.
y; .. . - .g.. . . .
.; . a. . ~ . oxidized with the concomitant production of hydrogen. ~-
4* .- .. The NRC's Special Incuiry Group concluded that the foregoing and other manifestations of damage would have warranted the following actions at 9:00 a.m. Monday, March 28: Actions hich Appear Warranted .. .. s., .. .,- .
- 1. Advise or order plant that core appears to have* been .
uncovered and thereby damaged. Marisus spi flov should be provided to recover the core, cpen PORV block valve for
- aximum relief flev. .
v . .
^-..
r .- 2. Advise State officials that the core has been badly da aged . and has released a subst antial amount of radicactivity. Ihe l plant is now in a condition not previously analyzed f or ccoling system performance. Presucing that full SPI flov is 4 turned on, advise the state that if the cooling systems do not function adecuately, portions of th e cor.e could begin to - melt which could lead to significant offsite releases in a fev hours. If the cooline systems are successful, evidence cf that' success should be'available in a few hcurs. Reco= = en d to State officials that they begin a precaustonary evacuation of the first few miles aroun d the plant with an alert for a larger radius (10 miles) evacuatics which may f oll ow. Ivacuees from the inner zcne of fev iles' radius should be noved to locatiens at least 20 miles dis' tant. ( SIG, vol. II, part 3, p. 268) e O
. v .. .
. . .i .
l
. . , u: .. .
. \ . . s.
.. , .. _4- -
I A. po, ar..e.n..t..ly..' a.t. ..this very hour when the SIG concluded that a partial evacuation w'oull have been warranted, Pennsylvania Lieutentant Governor Scranton called the TMI-2 control I .
- g. . .. .~~
;,i,;~ room and was briefed on the situation by Station Manager
.n. . , . .
,,n',",.. w. .. . . .. .
- . .' "- Gary Miller: While the exact nature of the information
~
~
- .w
~
con'veyed to the Lieutenant Governor is unclear, it must
.have been .
-- - f ar dif ferent f rom the picture that culd .
have been readily inferred from knowledge of events that took place since 4:00 a.m. because it led Mr. Scranton to announce at a press conference at 10:55 a.m. that:
... he had been advised that everything was under control and that there was no danger to the public health and safety."
. (SIG, Vol. II, part 3, p. 382.)
While most of the core damage had been done by 9:00 a.m.
~
- the situation remsined precarious until approximately 8:00 p.m.. The control room supervisors struggled for most of the day to reach a stable cooling configuration. Uncertain at 11:30 a.m. as to whether cooling water was reaching the entire core, a decision was mcde to depressurize the system to the point where a low presssure cooling system could 4
be used. Eut the depressurization effort was unsuccessful and at approximately 3:10 p.m. it was abandoned, apparently in fdbr of a repressurization strategy. The new objective , was to make sure the recctor pressure vessel and associated piping was filled with water so th:: the main reactor ceclan;
s
- , _ , - y;
. .~..... . , . . . . . . . . . . . . .
pumps could be used to circulate water between the reactor
~ '~'
pressure vesse1 and'the' steam generators from which heat would be removed in accordance with dstablished pro'cedures.
. .. ...............a.c.....~.........................
At approximately.1: 50 p.m. while the depressurization
.-.c
~
\
effort was still under way, a large pressure surge was
. . .- .. observed in the. reactor building. The surge resulted from
- . . +
- r. -
y77. F . .3- . u a.hydrog.en explosion or fire (subsequently referred to as a
~.-
. ,y .
o . .
"deflag ration") that occurred when hydrogen produced as a
, =v 4
. ..a.
byproduct of oxidation of the zirconium was released in the reactor building where it reacted chemically with hydrogen.*
. Meanwhile, federal and State officials remained une, ware that there had been (or continued to be) a real danger of a meltdown. On Thursday, March 28, NRC Chairman Hendrie briefed the Energy and the Environment Subcommittee, leaving
*All references to hydrogen explosions or fires in this discussion concern the hydrogen detonation outside the pressure vessel in the containment building that occurred at 1:50 p.m.
on March 28, the first day of the accident. This is not the potential explosion that directly concerned the NRC in the period March 30 - April 1, when the Commission believed there was a possibility of a detonation or fire within the pressure vessel resulting from generation of oxygen which might react with the hydrogen bubble that was believed to occupy a Eubstantial volume at the top of the pressure vessel. Subsecuent ar.alyses appear to indicate that there was no mechanism under conditions then prevailing within the pressure vessel that could lead to net production of oxygen, and therefore there need not have been serious concern about an explosion within the pressure vessel. While there may have been no danger in this period from an explosion within the pressure vessel, other danger s haC not passed and a further deterioration cf reactor systems might have lead to a major radiological release. m A
r. s
-.*c .
the impression that'while.the accident had been the most
'. severe in the history of the nation's nuclear power program, there was and had been littl.e..da.nce.r of a core meltdown.
Chairman Hendrie stated: .
"The . result /of the equipment malfunctions 7 has been - ~
apparently some overheating in the primary system . , , , that caused the venting. There apparently has been, SF . judging by the radiation levels, some core damage
. in the sense that fuel rods have' leaked. gaseous
. radioactivity, I rather doubt there is melting. .
There'does not seem to be any indication of that.*h (Transcript, p. 23) It was not until the following day, Friday, March 30 that.,the Commission came to realize the full implication's of the accident. Based on infor: ration indicatdag the occurrence of the March 28 hydrogen explosion in the reactor building and the exceedingly
'high radioactivity levels in the reactor cooling water, the
, Commission concluded that damage had been severe and there was still danger of a major radiological release, resulting gg either from another hydrogen explosion er from failures
~
of equipment which was then performing under conditions for which it had not been designed. Kemeny staf f ec,cluded that some of the uranium oxide
=
fuel may have become liquid at the temperatures which er:ceeded 4000 degrees, even though the temperatures were less than the uranium oxide melting temperatures of 5200. The reason for the fuel liquifying at temperatures less than the uranium oxide melting temperature is that it was in contact with zirconium and the melting point of the zirconium / uranium ~ oxide mixture was lower than the melting point of the uranium oxide standing in isolation. m
During the period March 30 thrcugh April 1,. serious consideration was given to the need for evacuation of persons living,within 5 p.iles of..Three Mile. Island. Bad officials known the status of the reactor on March 28, 'it is likely . (as suggested in the excerpt from the SIG quoted above) e that a precautionary evacuation would have been called on the first day of the accident. Beca'use Federal and state officials were not provided information necessary for - them to asses's the danger, they were prevented from fulfilling their responsibilities with regard to protection of the public health and safety. *
'o T date, three major TMI inquiries have been completed; none have adequately addressed the reasons for failure of TMI management to provide significant information to of ficials of the NRC and State of Pennsylvania, information that the latter officials needed to comprehend the nature a
of the situation confronting them. The first such TMI inquiry, that c'onducted by the NRC's Office of Inspection and Endorcement and reported upon in NUREG 0600, virtually ignored the question. While the TXI inquiry conducted by the President's Commission en Chree Mile Island (the Kemeny Commission) did not pursue the question in detail, Commissien member Governor Bruce Eabbitt states in a supplementary
.O
-g-3 D view incorporated in the Commission report:
"It now appears that there is evidence to indicate that MET Ed technicians understood, within a few hours of the accident,~~ that the nuclear core had been~ uncovered'and,'that th'is specific information was transmitted to supervisory personnel at the ~
plant early Wednesday. There seems to be little question that the technicians who took the temperature readings that morning understood what they found. The real question is what * happened to this information and whether it
. was transmitted to the appropriate management personnel. It certainly did not get transmitted to responsible public officials, including Lieutenant Governor Scranton during a meeting with Met Ed that af ternoon. "
, The third major TMI inquiry was that conducted by the NRC's Special Inquiry Group. The SIG concluded variously
~
that ... the evidence failed to establish that Met Ed management or other personnel willfully withheld information from the NRC," and, "The f act that NRC and B&W did no better than Met Ed/GPU in reporting critical information up the managment chain and acting upon it tends to support our conclusion that there is no evidence to show willful withholding --- of information by Met Ed from NRC." [SIG, Vol. I, p. 159-160.) In reporting the SIG results)
~
l$to the NRC, Group Director Mitchell Rogovin stated with regard to the cuestion of withholding information:
.. There was a mind set er a ' group psychology that apparently tended not te believe that the
_ cora was uncovered. They didr'; even think of :he possibility of the core being uncovered. People went along in that fashicn." . (RogcVin/!;RC Transcript, p. 57.) M
l
.sf -
9-c,...f.
. 1 i
SIG Deputy Director George Frampton stated to the Commission
^
that:
'"Our conclusions are,less.important than the. -
\
I
'evidendi itself. The evidence will all be made or has been'made public this aftern'oon -
and I think on the question of coverup, - l there is conflicting evidence. There is definitely conflicting testimony. There - l 44E are things that don't make coherent sense ! l to us. We made a judgment about the weight of the evidence. We found that, while - l there is some evidence to suggest that there l'
' is an intentional withholding of information, that the wei.ght of the evidence doesn't <
support that, others could come to a different conclusion based on the evidence that we developed." (Ibid, p. 60) - The following discussion addresses circumstances relevant to resolution of the cuestion as to whether TMI management withheld from state and federal officials information that should have been provided these officials. Emergency Command Team , In considering the kinds of data that were available, it is I useful to keep in mind the emergency management organization and procedures established by Station Manager Gary Miller. Upon his arrival at the TMI control room shortly after 7:00.a.m., he organized a group for the purpose cf analyzing informatic: and dete mining what actions should be taken. In a statement dated , l May 7, 1979 and presented to the E&E Eubconmittee on May 24, ; l Miller described this command structure:
. : . 3 :n: ..: u : :u:: v. :.: i e :. i:: : ., . .:u-: : --:e e." gs.,
- q :* ee* e o-.** .,, . =*
- g e o .* <:.q* *
.. .:: us :.. . u .u ve : u : s : : su: u , .-
3 l
. s.*
- 10 -
e D '
' ' ine first concern I has as : arrives in the Cont n1 Rece at a:pr:si ately 7:05 =es ,
- seccre fully cognitant of the sitaatica as it esistec ane cece ! fully ece-stcoe tre :ast c:esittees anc the resistien e ergesty. : irre:iately :::= ces ;e :f t*e
. ::*tr.1 R:: : and a:::leter lentor ;e ste to tire:t tre necesse y evc16ticas in tse vital a*eas to assure tsat tne :.slic was srtte: tee..tsat . .
the release -as wenitorec. 1.54% comsnicatiefts were oCrJ-ring anc t,*at tne Ciant was tr wint in sta:s t: a statie c3Scition. The etpr4md set W9 =nich I ju!! Cestr*Lec, ret frec6ently in SwgM:w* tre cay. The ge:v: presentet Unit concttiers, statss cf L-ergen:y 714e actions, snares crintens, risevasse tetenical cata, ant race re =encaticr.s. *tscassicas =ere 6 held witi Managnet, ar.c/or 44tcock & 'Jilcox. Lync'.sv-g the 5 tate, the ht:. anc a . following tmse evaluations. I sace astrecrtate cecisions anc se directed tne i=:le ar.tatica to tt.e Oct.trc,1 aoora and infermed einers noth insice ano outsice tne *
,:. slant as necessary.
Easically I set so this Irier;ency cortune team in the early neurs as 1 arrivet at i tee gian anc ine ractatien e*e seaty =as it. ;-:g*ess, ty esseettally femns y se91er pec:le ir.to a netween tc s.:ervile, centsci the e t-ge*: aat at: rt tc re siie tringing the ;1 ant te a safe concitice.. *. tess .as :st ir. charge e' 0:e-ataces . tc cire:t the $nift 56:e viscr. *r. Cstiei was ;61 in ctarse cf raciatten c:nce-ns, inc1 weirs raciatten surveys. onsite
- o'* site tea-s. at::. tattlity cae:'s , ssse-aly I cf sec;1e. getting in centact =tth the (me-secci *cate:1 1:atice (I"!) . etc. , *r.
Seeiinte* was in cre ell c*.arge of L'ntt t. *Se !*e-f eaty Cent si 1:stten, the tait 1
- nt-il 1:r anc tc ass.re that all f acets cf ne e e ge*cy pian =ece
- lic ec.
P.r. *.cian was enarsee to assare tnat a11 tme rec.t-ec t*ccecces sec sians we'e reviewed and Ic look thr:Wsh each tc assdre that eve y *ttS was ::ve'ec.
- Mis inclWCef the F*:cetWres f:* Deth emeegency gia* amt fer tSe uit itse1#, anc *g
'.a4 ;r v*ce De assu r ance that all actiens =eet tetrs tase* ant te te sgre the nc***i.
catiet calls we-t Pace. that P.anage9ett .as ectifie:, etc all ccm.eicatices e e
, in*tlace. *r. kee %:ge's has re:Weste: tc ;*ctice te:9r.ical assistance ;Ius Itna.wg
*=fts nts home c'*tte as %e c:61c. *r. Asecer .es in c*a*;e cf te:*sicai se::cet anc cc-rv icatiens anc M* . Incolin was 'm ca.a'se c' e a ge*:3 e.atntenance.
m
- sr *
%c ; t1tc aF t Wain $d*t SecsVse c# 'y If81Ri"I. I f'1I 8 II"'"I EI\1IIt c '
' ' * ' 'II IC*# 1C'** '
t%41 tre's =as cintmal release c# racicact'* ty 8"C *E'I gf tira 1 that8 was retwirec-
- ne
- C*' . t*e ;*es s.*e .
act N W NW W " a s ta te tr.a t I F.a f *e ve r See* s c*.::le :
. C *:"e c 10 .a s e *.C'
- c* s s 'iP:t t initieas:le. ec=ever. 15e : *t*:. 3:tt ee*a**ec tair as ca*. te test **'ec 1:
sa;a'ately. AI*. cf tre 4ettegs c* Ine ::e.a*: tas e'e rel: tr .re int *t 16 t s ts:-s 's ;*f tce te a ta?? et :s:*e e 4: 4 ::*rt e.c e: 'r:. tne *.:rt t
*: . 4 c t *e to:* s * :* ma n 'er .a t ::*e * *e :* se ' . . a t *te*.a's : :*ste:
- te
- ' a' a*c * *t tase : ;e* va- 2* *: .: -" te . s 4:a "
( r. c.: . ....
~
, 28 = .a4 yr . -) c,- J 2 e,6 . s e
- e . . .. e . e e e g
The following ciscussion accresses tne various xincs or information available to plant management which should have indicated that the reactor core had been severely damaged. Temperature Data . By observing the temperature at variocs locations
~'
y in the reactor cooling system, reactor operators should be able to determine the adhey of core cooling and khether there could be - steam srwees in the reactor
- x. , / vessel' or 'in the hot-leg pipes leading from the pressure vessel to the steam generators. Such temperature measurements enable operators to know, in the event the system contains steam, whether the steam space has
. expanded to the point where it encompasses a significant portion of the reactor core.
Temperature data can be interpreted appEoximately as follows:
-- If temperatures at the top of the core are less than the temperature at which water boils for :the prevailing
-. pressure of the reactor cooling system, the fuel rods are cooled primarily by water and cooling is adecuate. This 3.s called a "sub-cooled" condition.
-- If te peratures at the top of the core ecual the temperature at which water bcils for the prevailing pressure, the fuel rods are being cooled either by water or a steam-water mixture; cooling is probably adecuate. This .
~
is called a " saturated" condition, a..d can be converted to a "sub-cooled" condition by increasing the system pressure at which point the steam in the mixture condenses to water. d
c s ,
+
\ .
~
.a g
-- If temperatures at the top of the core are -
-greater than the temperature at which water boils ' for 3
the prevailing pressure, some portion of the fuel' rods is being cooled by steam; cooling may be inadequate tA ' and the temperature of some of the cladding is higher than the steam temperature. Steam at any temperature above the saturation temperature is called "superheated" and is indicative of inadequate cooling. .Such conditions are, at least, a warning sign that insufficient cooling. water or steam-water mixture may be making contact with the fuel fuel rods, and thatArod surface temperatures may be on the verge of rising to levels 1 where the zirconium cladding will react chemically with steam, producing circonium oxide and hydrogen.
-- Temperatures at the top of the co're in excess of 2000 degrees are indicative of inadequate core cooling and of a cladding steam chemical reaction.
A typical B&W reactor normally operates at a pressure of 2185 pounds per square inch (psi) and a het leg piping temperature of about 604 degrees F. This means that the hottest water temperature is approximately 45 degrees below the boiling point. In order for the water to boil at this temperature, the pressure would have to drop frem 2155 psi to about 1600 psi.
.. A. -
Hot-1riig -
., A principal indicator of system conditions at TMI was the temperature of fluid (water or steam) in the. hot leg pipe,*
which
- isths " pipe ~ that' c'arries *: hot' water f rom the reactor to the steam generator. Hot-leg temperature data at TMI ,
, was normally presented on a computer printout and on a s, trip Mh ., . chart recorder mounted in the reactor control room. The computer
, was progra:r.ed to record data between 520 degrees F. and 620 degrees F.; when, temperatures were outside this range, the computer printed question marks. The strip-chart was capable-of recording temperatures up to 800 degrees F.
- See Figure I-A, I-B 9
e 9 s'% e h 6 s I i
=
FIGURE I-A - 130 - Q<
* \.
~ OE 2-
-O
= 3:
O ' * .4. O
. *Aw 2=
= *
- A e E v - - ,
Su f" OO soO u- s
*= O
=,
- w I C i -
= -
(. . i, > ~
=
n i ~, :. . . -t s.'. /n5, i ., -
~3o2 -
/
i ~ l 1- G
~ 'ET M .
Es ,< === s <B- s> ... M '" G. #. ...' ei.e
-~
;U i => -
;g
-; ..w-ssex t w.,, /, . .. -,/
si - < i . i
== -4
- i. .
, o4> . j"d
=-
40 - m i W[ 5 2 .E/ =E =
* : E .
=I = o I =- , /, -
w -i -av 3
-l i s .
V: e= om n 7, -)q,e,
. .<i I ul o 2_
.g is..4 ,' , ,
"" f y "
- - /
- I .= m.- t . p -
s/ Q3 I u g: s.I
- s ' '*n D w.+
- t - ef, .
..v t-__ _--_ a u , - -
- y3= =p w. .- 3
/, 55 -
. ,.s . .- ,
f
".... .a.,
g , d., E 5E, m on p] o e .
/,
p="r.n : &r_-- =-; - - -:- n : e~ ;1
/
/
=
. s%e .
Wr r g g
,, g o ,,
y-y o OC d, c [ s A N . :- . E L v
~
D
<~ 13w' C
fu
/s
. o y w- .,
4- ! h.P. e e f s : l n t= = U 4 ( =-v.ee__=.1 = %*g
~
~
i l' D h :> 3
.A w 55 E h, . .- ci I
;4 -= ,.v - -e i g ;
., y >,
-> "i' c \c.Wa= -- -- M M ol
/ q, g-t.* m ,,.
- > x. - -
I i T. rp=- -= = . ..p3 2 y : y..-., . =
> Ca< k ,.
- k .- u- ,
, n 3 g
m o w > u c 6{+ u.s yEe. (s
\ O #
a C b* tl ' j
= w Y -
{ A .
, (
. ,Y K D-M "E-q h I I j X IA4 !- l I j a 1 i, w :
---- A Jn # '7 -
i l I go i - E I 7 I S
*:s [ q: :P l
w 15 j-4 o . .. I --
= - -
o= L i E$ E
~8 o l -s@
90 n oi-r c= e- I L. + y O Dirls 1 l i 2 e, o
< k l-ol*:Gi h:
V : l l -.3 ,. l ; j s- d -l
- v. l r---------
g g .
? I
" i e .
! i ,
L-------_-- J I l :.y ( y= =, p. 2w s **-. L 3 ", E
.) . t 1 em P I J Primary Water t n 2.0 botus eel secondaiy water l' a 1200 psig Electromatic M Steam .
bh relief valve (EhV) N Steam /tlydrogen , 4
- l r RelicI b!Ock ValVO
- ~kA S .
, s, .
*' . +1 , , ,j, !
[. ln-[-- s . -:
. :,- ?,
-> r. .
.*~< ... L :)e ,, ; .
- i. , =
'.,, 1
+ .s 1 R; s
r x ,. .e -[Gham w .
= ; :, :.s , >a 7 ywca w;. *n o .* + :: Pump:- ,, ,, g ; Pump; H
?* j
. g ..
. ofI
^< ?.v. . . . ?/. ?. . .
,. j a+
7 4
, 1 AA '
- ?N??oil*?#??! O
* * *' Ii"
,. . .,- ,f.
eWF +
#l
.' ._b!!
k!
' $N:!;[S[ **
j . . . ::ef .
* =t f H,
g
,. we y l
4 s,
- . . ::9a:
p,1.,.:j: ,,
* ~
' + ,-Q ,7
- i , , *
$ *+
q. p j
,4.,
/ 5 'r, % . 3,.s',.+ , .;+j ,
2N, e .3
. . . w , m_ , . ,, , 9
,- m . ,* m ;v p
.r.
,, s... .
9 r4 p1 y .. .
- +
, - , ,1 4 c . .. <- _ O
.~. ; b *l
-t .g -
( t
) H 3 ,. ,, &
s Loop A - Loop B y
. Core dryout and heatup continuing.
Hyrfrogen generation by zirconium / water reaction in hotter regions. ,
x 1 ;,- t ,,
, _ 1'4 _ +-
Apparently owing to the computer being incapable of
*. ~
. indicating temperatures in excess of 620' F. , Station Manager Gary
. Miller,..soon,after he, arrived at,about 7:05 a.m., directed that a meter able to_ indicate the full range of hot-leg temperature E be set up in the control room: n.
N41 When I turned to focus on plant conditions, an initial.. concern-was that the hot-leg indication was off-scale. I-asked that an extended scale readout' device be, connected
~
-to,the hot leg RTD (resistance temperr.ture Weasuring device).-
-(E&E, TMI-2, part 2, p. 297,) 3 Miller stated that:
~
... the extended hot leg temperature readout device C indicated 720'F." ( Ioid. )
- exceeding Jaur-temperature A. 660'F at the then prevailing pretsures (which ranged up to 2200 psi)-clearly indicated the precence of
~
superheated steam ; i.e. that the core had
~
- J been uncovered.
The readings from the hot leg instrument are also referred to by Ivan Porter, TMI-2 Instrument and Control 'N engineer, and by John Flint,-an engineer employed by Eabcock and Wilcox who was stationed at TMI. Porter had been l responsible for setting up the instrument, and he told NRC investigators that after intially questioning the validity of the readings he had checked some of the temperature 4
, readings taken from instruments within the core (see below) and that: \
to me it confirmed that,what i was seeing on the RTD. That we had temperatures greater than 700 degrees in the plant, since 700* degrees was full scale on the computer and I was reading greater than 700 on the hot leg RTD." (I&E, Tape 237, p. 15.) .
*This-is in reference to the in-ccre thermocouples discussed below. / The computer could indicate the in-core temperatures up to 700*
whereas it could indicate the hot-leg temperatures cnly tes620*. 4
o -
., j [.', - .
- x. .
John Flint-recalled that at..this time (sometime af ter 9 : 00 a.m. )
~^ "
... Ivan' Porter.showed me a special setup with a RTD, '
that was approximatelyJ724 degrees." (I&E, Tape 323,.
.p. 34.)
Flint also noted that he himself had monitored the strip chart recorder in the back of the reorr.which was recording the data Flint refers to in'the
- preceeding quote. .dhis chart sh:md hot-leg temperatures indicating superheated .
74 conditions in the primary system shich in turn indicated sections of,the core has been uncovered for a significant portion of the period
' between - 5 : 4 0. a.m'. "and 7 : 50 p.m. * , the ,latter being the time at which a relatively stable cooling arrangement was established. The strip chart shows that-the hot leg temperatures rose rapidly s
from'*the temperature at which water boiled for the prevailing
' pressure :(i.e. saturation temperature) to superheated temperatures following shutdown of the main reactor coolant pump at approximately-5:40 a.m.
Virginia
. Notes taken by B&W staff in'Lynchburg4cn March 28 indicate that personnel at TMI' were not using the hot-ieg temperature data as indicative of whether the core was, covered. The notes state the pressurizer temperature was being used as the basis for inferring core output te.:perature -
since the flow of water , was from the reactor pressure vessel, into the pressurizer, - and out the top of the pressurizer. The hot-leg temperature the
., sensor was measuringAtemperature at the top of the hot les where steam and/or non-condensiblo gases were blocking flow through the steam generator. Therefore, the temperature at t the-top of the hot leg was more indicative of temperatures I
d
*See Figure.II.
l \ " 8:00:3.7 10:00:37 - c 4:00:37 6:00:37 ' ' 2 3 4 5 G j , ___.'..__ ~ h RC P-1A _ . __ . . I _ II. actor contant ~ ' ~ ~ ~ ~ ~ ~ ~ ~ ~
~
{. l'innp opesalion 11C-P-2 A ~ ~ Denol.non iIC-P IO , Octetroce 3)(note 7) 11C-P-2B . l'~"~~~"..~ 7 'l-
. . }
00 I;cacIn: coolant nystem flow g m....--.
- n. , ..% A Loop A
- o. %
loop U e 30 D I O T,' Soisrce range
!;nnte v antiinformrufiale 6 . . .
range insteinnentation m8 10 4 in core
.,3 M
j.a g y,a j [/
,probabic core configuration clonge g C
m <'"t o 3 .m t01 w ', 1 0JV s .I
' 3 r
\ ,
~K. H J
- s ll Iio _-
( Intermediale range
- o. 102 m
ra 900 LoopD l f1, ar. tor e oolant syntem N ontle I 1.. inn,n atine i7, cC ; [ _ - - 9., g 750 Loop A In" Min a00 Loop A UE W ~ ~ LoopB
$ core uncovery becomes apparent
_E 450 , 1 300 probabic core configuration change D t: 600
/( . ,
e I;. .it l< >r r:onlant nynl<'m '
,{3 Loop A in;* 1 1 :inper.iline' 8L
, ,f h @ 450
%~
( - N 'P
~
- (. _ - - _ _ _ _ - _ _ _ _
/
e
~
that had existed' earlier (when the core was uncovered e.g. at 7:00 a.m.) rather than later in the morning (e.g. 11:00 a.m.) j when.it was. believed tha't.the core was probably being cooled , via high pressure injection flow through the core and'out the ' [. , pressurizer. In-core Thermocouples
. 'Other sources of temperature data were the 52 thernocouples installed inside the reactor pressure vessel above the fuel assemblies. These thermocouples sensed water (or steam) temp.eratures at particular points across the top of the core. Eac'r thermocouple provided an indication of the temperature above conditions p a particular fuel assembly while the hot-leg te:parature ser. sing device _s (discussed above) nornally indicated the averace temperature of water (or steam) leaving the reactor pressure vessel. There were, in fact, large differences between average temperatures as measured in the hot-legs and the ,
~
peak temperatures measure,d by the individual in-core thermo-in-ccre couples. It was these peak4temperatures that confirmed not only W..__ _ _ _ _ _ _ _ _ - - . _ - - _ _ _ . _ - -
that the core was uncovered, but that the zirconium cladding was reacting with steam, producing hydrogen and zirconium
~
oxide. Credibility of in-core thermocouole data. The TMI personnel showed an early and continuing'inte' rest 05 in the in-core temp 3ratures. Thev instructed the computer
~
~.-
to print these temperatures at about 8:3da.m., 8:47 a.m. - 11:10 a.m., 12:40 p.m., 4:11 p.m., . 6:30 p.m., 7:59 p.m., 8:56 p.m., and 9:56 p.m. Bedmen .6:00 a.m. and 6:00 p.m. the temperature data for at least 40% of the thermocouples (EPRI, Fig. cl-ll) kere printed as questionmarks, indicating :s
, either that the thermocouples had failed or that the temperatures were in excess of 700* F. Because some of the-thermocouples alternated between indicating question marks and temperatures .
less than 700 degrees, i t was more plausible that_the readings d were indicative of temperatures in excess of 70.0 degrees i indicative of > thanninstrument
*- malfunction.* Moreover, if manv i had been damaged while others continued to function, this in -
itself,would have been a reasonable indication that something rnajor had happened in the core resulting in disabling of the . thermocouples. : The following excerpts from interviews conducted by TMI investigators indicate that TM personnel did in fact believe the in-core thermocouples vere providing useful in'crmation .
*By approximately 12:45 a.m. en March 29 the thermoccuples were mostly indicating temperatures less than 700 degrees and of core were apparently used thereafter as a prime indicator conditions. .
g - , -. ,- .-
' '+ .
18 - s . John Flint, Babcock and Wilcox engineer stationed, at TMI told Gen,eral Public,,Util.it.ies._ investigators.in an interview held on April"20: ,
... shortly after I came in we also started calling up on the computer the incere thermocouples
%G attempting to establish what had happened in the core. Many of.them were indicating questionmarks which indicated that they were greater than their 700F range. Only one or
' two seemed to indicate that they were in fact bad. These temperatures were monitoried for the~ rest of the day to follow what was happening to the core."
Flint also told NRC I&E Inspectors that: '
~
"When we first started dumping them out (i.e. calling up the incore thermocouple data from the computer) many of them had question marks, which indicated they
, were above their normal scale of 700 degrees not printing out " bad" which would indicate that they had fciled. Over the nex: few hours these thermo-couples gradually came back on scale [;) we recovered more and more of them and towards late afternoon I believe we had most of them indicating 'on scale. "
(I&E, Tape 58-59,- p. 8.)
.. A Ivan Porter told KRC investigators:
"... I believe shortly after 7, he (station manager, Gary Miller) asked me about the readings on the in-core temperature detectors, and I punched out several of them (i.e. asked for computer printouts)..." (I&E Tape 237,
- p. 16)
Porter said he reported the results (i. e. temperatures in excess of 700* F.) to Miller who asked Pogter whether there was any way of measuring the thermoccuple voltages in order to determine how much in excess cf 700 degrees the temperatures
[ D t might be. Porter told Miller that he thought he could'get the information by connecting a digital voltmeter directly to the wires leading from the reactor core and measuring .- Iand therefore the' temperatures) thereby the voltageshbeing fed into the computer. At this
, point there are varying recollections as'to exactly who it .
was that' Porter told to make the measurements. In any event, four technicians in addition to Porter either participated in making the measurements, or observed as they were being made. One technician who the NRC investigators have , designated Instrument Man B (IKB) in deference to his wish for anon?mity, stated that:
"... Two of the thermocouples, the first few we had measured, were around.700 to 800 degrees, specific temperatures I can't cuote you, I don' t know. We had taken one off and we had measured 2600 degrees in and abous that vicinity, it was very close to that. At that time neither one of us believed than this was,a true reading because after seeing two, three that were 700, 800 degrees, . . , 2600 was hard to believe so we decided to take a few more off. ... I believe in the course of testing .
thermocouples, we had at least 10 or 12 we had disconnected th@ total. We had seen temperatures ranging anywhere from I know there was a botton of abou 690 degrees to uppers of 3700 to 4000 degrees. (I.E. Tape 315, p. 11) While cladding temperatures probably did exceed 3500*, there is disagreement as to whether temperatures as high as 4000 degrees were actually measured since it is unclear that the thermocouples would function at such temperatures. There is general agreement, however, that at least five people (Porter, Maintenance Foremen Bennet and Gilbert, Instrument Mar. Thomas Wright, and Instrument Man 3) were directly aware of the instnrent readings indicating o-
2 , , _ 20 temperatures in excess of 2000 degrees, i.e. temperatures at would'be- which there/g . significan't production of hydrogen. Porter . questioned the validity of the measurements and he tol'd the technicians to repeat them ustag another meter. The technicians (3
- did so and this' time they took measurements of 51'of the 52 -
thermocouples. Nine of the 51 were in excess of 2000 degrees. Where there was" overlap, the second set of measurements, taken
~
shortly after the first, confirred the first set." IMB told the NRC investigatcrs-
"Now the second set of readincs d_d in fact correle.te -
.the original set of readings that was taken Gith~the thermocouple reader. The general feeling at least
. amongst Jim (biright) and myself was that the readings we had were true and accurate. All five of us that were present there did in fact visually see the actual readings we had taken both off the thermocouple reader and millivot reader. All five of us did in fact verify that the millivolt reading through the conversion table was correct. So I am sure we didn't read the wrong table or the wrong .one er something of that nature.
(I .E. Tape 315, p. 18) Bennet, Gilbert, Wright and IM3 have stated with varying degrees of certitude their conclusion that the data indicated the core was uncovered. IMB told the investigators:
"... it was the general consensus amongst the instrument people there that t'e core was definitelv uncovered, we . -
kind of found it hard to believe that this many high temperatures that we had seen that all those'inceres would have been bad and the only way that they could have went bad that radically would be an uncovering of the core, and super overheating. (Ibid., p. 17) , Bennet: We had possibly- uncovered the core was the only way we could see that you eculd have obtained temperatures of 92: magnitude." ( ~_ . E . Tape 211, p. lE)
Wright:
"I feel then -that there was a ' definite sign then that the core had definitely been uncovered to the point, where it suffered damage. But it, I still say that, -
you know, I'm there to take the data. I'm not there:to analyze it. So I gave them my personal opinion as in the, yeah, I do believe we did suffer some damace there.
- ~
(I.E. Tape 310, p. 14) . Ivan Porter was responsible for reporting the thermocouple
- measurements to Gary Miller. In his statements to the NRC investigat6rs Porter is ambiguous with regard to what-he believed the significance of the in-core measurements to be. When asked whether he believed a reading of 230b degrees to be anomalous, he stated:
"I don't know. I guess I was afraid it was real."
(I.E. Tape 237, p. 20) But Porter also suggested that the measurements were not~ believeable. He said that in reporting to Gary Miller that Miller had asked him what he (Porter) thought the thermocouple measu,rements had meant. Porter told the NRC investigators that he was not sure of what his reponse to Miller had been, but that:
" ...my personal evaluation was that they (the thermocouples) had been destroyed." (Ibid., p. 19)
IMB told the investigators, however, that he had told Porter ( n'B , the that heAfel)leved the temperatures reading's indicated that core was uncovered:
. "... I personally told him that and he'was physically there to read the' readings. He saw the actual temperatures we had....This is why the first time he didn't believe it.". (I.E. Tape 315, p. 18)
IMB also said: .
"I believe Ivan (Porter) didn't really want to believe
)g ' what was really taking place. I don't know whether it was an attitude of " hey, your measurements are wrong, you guys don't know what the heck your doing or
'. whatnot." I think the general consensus throughout x.
the whole first day was number one nobody really knew. what was actually happening, number two, some that had an inkling'of what was happening didn't really want to believe what was going on. ONce you start seeing a temperature of 3000 to 4000 degrees in a core, well ... the first thing that starts coming to mind, you've got a meltdown coming. The core is uncovered." (Ibid., p. 18-19) Whatever it was that Porter actually reported to station Miller told manager Gary Miller,/g NRC investigators: . : So you know, the bottom line here was that they (the in-cores) are hot, they were hot enough that they scared you, as far as what you're looking for. It told me the reason the computer was off scale at 700 degrees... The in-cores were reading anywhere frem 2500 or so, .and I picked 2500. It could have been higher than that. But that you know, I was looking for a gross indicator and I had it. ... I know that we were superheated and all - that sort of. thing, I don' u think we tumbled to that kind of lodge (sic, logic?) but we just know we didn't have a control, we were out of centrol. We knew the situation was one we hadn't anticipated too many times here." (I.E. Tape 159, p. 51-52.) None of the TMI investigators' interviews has indicated whether Gary Miller (who,as indicated above,had been informed of the first set of measurements) saw the data obtained from the second set of readings. This data when plotted on a map cf the core showed hot regions within the core and col'd regions on.the periphery.* Instrument Man Themas Wright who was invcived in recording the data said that the data was supposed to ha.re been turned r See Figure II!.
'~ ~
'. = FIGURE III 7i
. ...A. .
2 3 4 5 '6 ~ 7,~ '8 ~9' 10 11 12 13 14 15 .f.i 1
..gj A "*-
A _,,..
. ~ ..y
~
281 325 B... B . c,24?. w l -
.wi 957 325 C .s
~
C 469 .. a .':f4 .
. Tss::
.wy 500 217 D D 1196 1926 2580 326 E...
E 599 323 555 F . F 80 2366 2378 1806 1875 234 G G 375 2272 1774 1295 2176 1852 H H 260 2452 K K 1811 632 1760 1. L 373 1566 382 296 1774 M 325 2171 2327 348 252 W
'N N 413 578 2167 1138 308 O O 356 462 P
P 291 . 475 R R 352 7 8 9 10 11 12 13 14 15 1 2 3 4 5 6 1 Figure Cl-12. f.'.ap of Core Exit Temperature ( F) 240-330 min. t i i f L
i'
. .2 3 -
e
- E over to Ivan Porter. He stated to the NRC (I.E. Tape' 310, p. 18-19) that Porter was aware of the data.
Wright recalls also that
~
Porter had remarked that it appeared to him that it' "... looked like we came, yoka k o , that looked like tiiey'd been damaged." It is not clear from this whether Porter was saying that the core or the thermocouples had been damaged. It is also. unclear is ~' whether Porter actually received sheets 6n which the data . was written, and which went unnoticed,."it seems, from ; - . March 28 until May 7. . Mike Ross, a TMI-I superintendant, who was a member of the group making decisions on March 28 engaged in the following dialogue with NRC investigators. Hunter (NRC investigator):
"Okay'. Do you recall in the discussion, think tank discussions, that the thermocopple temperatures were, in' fact, brought up in the discussions?"
Ross:
" Thermocouple temperatures were brought up to .
Gary Miller,.and I guess the bottom line they got out of that, was that they were not conclusive. It showed the core was hot, basically. I was going to say his range varied, very scattered. l He had like ... He was saying he had various temperatures scattered throughout. So, well, ! Gary and he discussed it, and basically I think the bottom line was yeah, the ccre is hot, or it is at least hot." (I&E TMI Tape #226) 4 O e
s- . That the in-core thermocouples had indicated temperatures t . iyess of 2000 degrees. (iglying p:cbable hydrogen production) was apparently not reported to the NRC during the f$rst few days
.- . NRC $2 of the accident. Victor Stello (thengDirector of Operating Harold -
Reactors,under7Denton) said he had not knowhof such nieasurements until the week of April 1 . Roger Mattson (Director of Systems Safety under Harold Denton) stated to the E&E' Task Force on May 9 that this was the first he had heard of such temperatures. M.. , 9 4 S I
s
- 24a -
') .
Stello, however, had been concerned on March 28 about data
~
-indicating superheated' conditions in the hot legs. He requested computer er.intouts..of. the.i.n, core thermocouple data. There Emsued the following dialogue between Three Mile Island and the NRC's ,
Incident Response Center in Bethesda: gg Voice: First of all, I can't get the in-core temperatures. Okay? Voice: You cannot get them? Voice: ' They print out questionmarks. Voice: They print out questionmarks? Voice: Yes. - Voice: Okay, what's that mean? Voice: That means that either the computer point is messed up -- ekay? Voice: Yes. Voice: or that the line -- you know, the -- where you sense it, that line's broken or something's messed up with that line. Okay? They were printing earlier. Yeah,
.the computer just won't -- the computer won't spit out a good number for them. They're trying all of then to see if we can get any of them to print. Okay?
(01-033-CH 2/20-MEM-10.) Based on the foregoing discussion, NRC I&E investiga: Ors wrote: - (At approximately 4:10 p.m. on March 2E) " Reported . in-enra te ,nerrt :rce unavailat'.c. Tupervisor (at TMI) reports c NRC the-; (in-ceres) are all printing cuestiens marks which means e ther the computer point
, cr the senser is malfunction ' sic). (SUREG 060~, IA-101)
J.4b - Yet practically coincident with the conversation in which the NRC was told'that the computer was printing out question-marks, the computer was in f act displayin9 not onlyquestionmarks but also two on-scale readings, one indicating that thermocouple 2 9-H was showing a temperature of 596.9 F., and the other showing thermocouple 6-L indicating a temperature of 562.1 F. Both temperatures were indicative of superheated conditions in the core and'of the likelihood that the reason for the questionmakrs v. wLs high temperature rather than a malfunction of the instuments. A reproduction of the cor.puter printcut follows:
' 23:02:13 132. 7 135.1 591. 4 474. 7 4 5 7.h 5:!1. 4? ' .G .'.76. 3 ?:Us . S O, ' 1Gn0: 13..
_1S2.5.123.9 591.5 476.0 456.5 53'i. 421. 3. 37F. 3 235. 0 9 6:11:32 DATA 0493 IM INCORE T/C 8-H TEi1P -???.? 16:11: 41... DATA Ok 9ti.. .It4 li4 CORE T/C 9-H TE;4P , 59:i.!1. .
.0 16: 11: 49 DATA 0495 I4 INCORE T/C 9-G TO49 -???.? ~
_.16 : 11: 5.7._ DATA 04 9G. _ iM I NCORE T/C .. E : . TB4P
. ... -???.t..L_
16:12:05 DATA 0497 IM li CORE T/C 9-E TEMP -?.??.? . O 16: 12:14 group TaEl:0 - OPERATOR GROUP C ~ _ _ . . . 1032 v356 0390.. 0389 04G3 03!!3 04 72 Ok 2 0 . :1469
. n.
n ,s.,2.51
. . . . 1 2 . ').
. . . ' 21' . 9 .:'. ' .1 4 7 6. 4 4 5 rs . 0 .';"...'. 4 s'"' . S ' 7> u . . .. ".. 5 . ^
16: 13: 4.h _ D AT!\ . 04 9 3_ !!.4 . l i> CORE. .T/ C 7-P T B i? .
. . . _... .???.?.
16:13:53 DATA 0499 IM INCORE T/C 7-E TS;P -???.? O 16 :1L : 03 DATA 0500' IM !NCCRE T/C 6-G TS;P -???.?' .
,. d.-. , ~, . .;t
..., w A 0 :> v 2.., - 1.,. In....n,-.
m . . n; i e
/ r. .a- G. ,. e.r.
- s. . .e- ., ....,,.,
16:14: 22 GROUP TREi:D ~ ~ ~ ~~ (._ OPElmTdR GROUP C "-~ ~ 1032 033G 0390 0329 0':03 0393 0972 Ok t.'; :1h59 16: 15:02 1S2.0 116.7 591.0 477.9 455.2 5'.13. 4".0. 5 3 7 5. 3 2r;': .1 0 16:15:54 . DATA 0502.. IM li: CORE T/C 5-H TDi? -???.? 16:1G : 03- DATA 0503 li4 !!! CORE T/C 5-R TDIP -???.? n::20: .,_. D m. ,. .n 0.e;0 4, 1 . I n. . r,. .. n :. _i /v, . ,.e - L . i e...n
.. n. _ .. . _ _ _ _ .
- c. l.3.
. . . 3..
O 1G: 16 : 22 G'iOUP T' LEND . OPERATCO._ CROUP. C.-.. 1032 O!IG 03V0 03E9 045E 0 39.t. 0477 O W..? 0400 m,
- 10.,/ ._ 03 _..e,.
3 . .e, 1 .n. 0 s-:..;. i. ._4 7.:...> i n... . ., .r.e.. .-) . 49.0. :o.. si i e ..
. .a r.a 1G :17: 54 DATA 0505 IM !NCORE T/C 7-M T'C;P -???.?
1... e.. o.r-> . .. c ,. ,ru t . . r; c; .. n e., I ;.i i p.n., .q:. . f r., : n. .:)
. e _. u.. .. .. .
_ e. .n. o. ._
- m. ~r.
a .. . .e.~~ A ,7 s: P s s .i m. e 4.n.. e s.1
- f......- e ,
-C r...,.....
- .. s . . a ; - .' -
..r._._.... n.......... v.4_:. ..:.. .. N ,. .. .. .. .
,J
; .9 q... .. .. . .. n . . . ..
n.
; . L ;, J L . .* w 's s .. -
c v40C L .. . . 41 .' ,.;
. . . (;.
- f. s . *
- 1.* . ?. g70.C.
(=~ ,... t.' .
- c . .. *
.__.0*
r .5.- -)
..__ . . . r, ....1. .. . ! a . ...A. *.r*. i... . ,s .. r. . .
3.,.. . , . .:., . .- .
- v. w.
. 2.. -a:../w e. .i. . . 1- r. . 3
. _ v. , , . e O
24c ~
. . a .
' Although 'tihe"in-dois 'thsisociouple data was printed at least 9 times between 8:00 a.m..and 10:00 p.m. on .v. arch 28 there appears to be no record of any set of such data ,
M having been transmitted to.the NRC's Incident Response , Center in Bethesda. . 9 4 e o8 9 4 s S O r
1 a Neutron Detectors TMI-2 instrumentation incl 0ded neutron detectors mounted outside _ the ' reactor pressure vessel and 52 strings of detectofs - mounted inside. Both in-core and ex-core detectors are ..h
. , installed for the, purpose of providing operational data; .
during n'ormal operations, the in-core neutron dets: tors indicate-
~
power production at various locations within the core. During the TMI accident both in-core and ex-core detectors provided data indicating the core was uncovered. The ex-core. instruments indicated increased neutron levels. These were interpreted initially as indicating that the reactor was near the critical point where a self-sustaining chain reaction might be occurring. The operators believed that this might happen as a result of insufficient bcron concentration in the primary system coolant. In actuality the
-apparent increase in neutron activity resulted from the -
fact that neutrons -- normally produced in small quantities in a reactor core even when the reactor is not running - were, as a result -of voids in the core, less likely to be absorbed prior to their escaping the pressure vessel and were therefore reaching the ex-core instruments in greater numbers. In short, the increased number of neutrons were due ,tc the pressure vessel having lost its water and not to a restart of the chain reaction.
V The in-core ~ neutron detectors also yielded data indicating that the core was uncovered and the depth of the uncovering. Once the water. level wen't below the neutron detector and the . temperature of its surroundings rose, the detectors re'sponded to the higher temperatures and became, in effect, temperature 4^ +
. sensitive devices which provided indirect indication of water level in the core. (See Figure IV'.)
The record"is unclear as the extent to which the TMI supervisors used the neutron detector data as an indication been that the core had Auncovered. The record shows that increased counting rates were initially interpreted as an indicator of the reactor becoming critical. As a result, additional
. boron was injected into the primary cooling system. During the morning of March 28, S&W engineer John Flint looked at the neutron detector data and concl'uded that in all probability a
~.,.v. the increased counting rates were due togchange in the leakage path, i.e.,as'a result of voids in the reactor coolant fewer. neutrons were absorbed by water before reaching the neutron detector. Flint told the I&E Inspectors:
" Indications for the source and intermediate range (neutron detectors) appeared to be normal, for this period of time following a shutdown condition. I did notice,-
however, that there were several blips on the recorder for source / intermediate range and in conversatica with Ed Fredericks he informed me that they thought at the - time that they were going critical and that they had added additional boron to the system. At this time, I informed them that in all probability this was not the case, that there had been a change in leakage flux path from the reacter core to the detectors and it was not in fac: the case the reactor going critical again." ILE Tape SE, 59, p. 4.)
, - 26a -
i FIGURE IV Reactor Vessel
& N S
qq ,
> Plenum -
.y ,'
n n assembly .
+ '
A A O 4
?;.:y4 in-core
[thermocouples
- A !
'f O (above 52 fuel assemblies)
.y l ,
4 . . . _ _ _ . . _ _ _ _ _ - _ . _ - I 0 DownComer Fuel assemblies x - (177 total) g - l Source O Core - range ex core neutron
@ _ detecter (2)
Self powered
/ neutron detectors O (in 52 fuel assemblies
~
for total of 364) . t1 ER
+ - - - _ .- - - S _ _ - _ _ _ _ - - - .
Lower grid assembly y-__-----------., t
% /
% I
' ' " * . = * ' " "-8'""'
=
a#,** e
/ -
Tc carie sc eacmg re x, E25:0 teact0r COCfigu'EI'On and instrument IOOEttor.S.
1
. i e
-27_
Hich Radiation Levels in Containment At 6:35 a.m. a radiation monitor mounted at.the top of the containment,bui1 ding, indicated radiation levels of about 0.1 Roentgen (R) p'er hour. By 7:30 a.m. this . monitor was indicating levels in excess of 10,000 R per ho'ur. ,g (Approximately 50% of persons exposed to this dose rate ' would receive a lethal dose in about 4 minutes.) 'These radiatio'n levels implied that radioactive fission product gases had been released from a significant fraction of the fuel rods; the only way such a release could have occurred would have been through development of cracks or perforations in the cladding. Based on this reading alone, James Floyd, a TMI
' supervisor inferred that approximately 1/8 of the fuel rods had failed. On the day of the accident, Floyd was in Lynchburg, Virginia for training on the B&W reactor simulater. . Having been told of the accident by a colleague, .
Floyd called the centrol roon at around 7:30 a.m. and was given information on radiation levels from which he inferred that the cuantity of radioactive gas in the containment building was approximately equivalent to 1/8 of that that had been contained in the fuel rods; this led to his inference that 1/2 of the fuel had been damaged. Floyd told the Kemeny Conmission that the inference he made in Lynchburg could have alsc been made by persons in the TX: control rocr on the day cf the accident. (Kems. y
-l ..
Hearing, May 31, 1979,.p. 186-188.) Kemeny Commission member Ted Taylor asked:
"Well, to whom did ' you give this information (i.e.
Floyd's inference . ' . . that it looked like at least an eighth of the cladding had failed') duringf - the course of the day on Wednesday, outside the people you are immediately involved with a. 1 l Lynchburg?" L a, '
. .Floyd responded: ,
"Nd one."
Taylor:
"No one. Were you assuming that people at the plant knew this?"
Floyd:
"Yes sir." (Ibid., p. 188-189)
Floyd later indicated that it was not until he sat in on the Kemeny hearings in lat:.May that he realized that the supervisors and management at TMI had said that on March 28 they were unaware of a substantial amount of core damage. (Ibid., p. 194) . G 9 J
o . .
. '. I l' .
Hydrocan Combustion 9 During the morning and early afternoon on March 28, a significant portion of the hydrogen produced in the zirconium
~
steam r,eaction ,was re1, eased f, rom the reactor coo 1ing system into the containment building via the pressurizer relief - a valve. At approximately 1:50 p.m. the hydrogen ignited.I ) What is pr5bably more accurately described as a fire than an . = .
. explosion caused several effects including a.28 pound per square inch (psi) pressure pulse in the containment building,*
This pulse was recorded on the strip chart that recorded containment building pressure. It was also recorded on a series of other pressure measuring devices which used the containment pressure as a reference. The fire raised temperatures in the containment from about 125 degrees to 175 degreesF,an increase of 50 degrees dissimilar to any other temperature changes cbserved that day; these temperature data were recorded on strip charts in the control room.*+The iner, eased temperatures triggered several alarms. The containment pressure pulse also actuated vari.ous emergency systems, most notably the equipment that caused water and sodium hydroxide to be sprayed into the containment building. TMI supervisors and operators have given two explanations for not having reported the hydrogen fire to either Federal er state officials:
"F:gure V-A
*' Figure V-B
""" Figure V-C p- 's See f.n. p. 5, regarding this ne: being the pctential fire and/cr explosion feared on Marc'- 30 and subsequen: days.
O
, . ' - 29a -
FIGURE V-A e e we O e $ >*7
' m O
E i: 2 g
,) 2 0 e
~E
- 6 0 5 E m E
C O ee l-- 8 w
$ y L
'O" 5 O C
.O 3 o
~-
3 u a c e< ~ s C 8 E O i- C - l l . v
- m i l I i I ! '
I m e v e a n N N N e o e v e v l (6!sc) a;nsse;d Su;pi!ns ;o: cess - h
1
': ,' - 29b -
. FIGURE -V-B
,~ .
270 ' -5 3g 0 e o OTSG "A" , 265 -
- OTSG "B" -
0
- S 26 -
- 5 o
O
~
o o 1 0 255 - c - 10 I o s I o s h I e $ e 250 - I o - 15 p e' 5 e / 0 m o I
]
4 )O 245 - c _ 20 l 0l c 240 - 25 235 - 30 I I i I I 13:50:20 13:50:30 13:50:40 13:50:50 13:51:00 13:51:10 13:51:20 Time . N . Effect of Hz ignitior, or, measured steam pressures.
X 190 . 180 -- G Supply mr
^" "'"'"*" *"
170 - penciration 0 Ambient sump area 160 - 9 Ambient let down cooler area %
~
13 H g
; 140 -
c: 12 3 130 - - - 11 fj 8 n 7 10 a e o E 120 -
" I "- 9
_ g_ 100 -- G S 10 Ambient drain tank area 11 Ambient elevation 353-1 f)0 12 Ambient elevation 353-2 80 - 13 Ambient elevation 330-1 14 Ambient cicvation 330-2 , i 1300 1400 1500 Time
._ Reactor building temperatures during hydrogen ignition.
0
c- .
' .? ,-
- 1. The explosion manifestations were caused by spurious-electrical signals. .. . . . ..
2 Because the pressure did not remain'at a high level, , the cause of the pulse was not important, and the , pg operators were too engrossed with establishing a stable cooling configuration to pay attention to it. n Neith'e r of these explanations is plausible. The first is questionable because of plant design considerations and the simultaneous occurrance of pressure and temperature effects indigative of a rea3. pressure pulse. Given their basic training in reactor engineering it is implausible that the TMI supervisors would consider the pressure pulse to be insignificant in light of the multiple indications of core uncovery and hydrogen production.
-The record contains conflicting testimony. Excerpts
. . , follow. The first concerns whether the electrical malfunctions could have caused the pressure pulse. TMI-2 Reactor Operators engaged in the following dialocue with Ed Frederick and Craig Faustg the E&E TMI Task' Force: Myers: What could have had two of those three sensing devices sense high pressure and lead them to think the pressure was high, other than high, pressure, real high pressure, actual high pressure? Is there anything that decides the pressure is going up that could have led those sensors to think the pressure was high? Frederick: A test signal. , Myers: Would a test signal go to two of them at the same time? Faus:: No. In fact, you would have to hook it up. Irederick: It wculd have to be a lengthy manual acticn to
,, get it to do it, other than actual bui ld in? pres 5ure-
? .
4 Myers: Is there anythine that veu can think of other than excess building pressure that could have simultaneously led the meter to read 28 pounds per. square inch and to turn on the spray tanks er turn on the containment spray? Frederick: No; it had to be high level pressure.. Faust: There had to be a pressure surge in the building for it to happen. (E&E TMI Part 1, p. 147) ygg At a later date (on September 11) Frederick' told interviewer's from the NRC's Special Inquiry Group that he did not believe the pressure spike to have reflected a real increase in pressure because his training had not informed him as to the possibil1ty of such a spike. Frederick stated that none of the persons present thought it plausible that the pressure in a 2 million cubic foot building could rise and fall so rapidly. Frederi'ck stated: .
"That's why none of us considered it plausible. It'u impossible to do that." ,
SIG interviewer Ron Haynes responded:
"I wouldn't say it was impossible. I thought it actually occurred."
Frederick then stated:
" Based on our training, it was impossible. It was completely foreign. If you look back through everybody's l training and the FASR and safety analysis and the building
! construction, you will not see a paragraph that prcjects that type of transient. Nor will you se'e it in anybcdy's , I training so far as -- that is se particularly foreign and unbelievable that it has absolutely no significance. l
~
,- That's why nobody did anything abcut it fcr twc days." l (5:G, Faus et al., p. I!*-2C5.)
Instrument Engineer Ivan Porter was asked by NRC investigators whether Porter thought the pressure spike could be expikined by any form of instrument malfunction. Porter responded: -
"I would think not. It did look like a real spike.
gg to me. That was when I was specifically asked if it could be real." (I&E Tape 237, p.. 35) With regard to whether he saw the pressure pulse on March 28, Porter said: -
"But somehow I have a feeling, I didn' t look at those charts until the next day. I'm not specifically sure that I was aware of it that day. I know that I very specifically remember a discussion where we looked at the chart, was asked if it could possibly be real, and I also look at the . wide ranged pressure chart and saw that [the pulse) showed up as a decrease in pressure on the wide range reactor coolant system pressure, but I sincerely believe that. that was the next day that I looked through_the stuff." (Ibid., p. 34)
It is not clear why they would be waiting until the next . day to look at the data, or if it were the next day, what it was that led them to do it then. In addition to the spike appearing to be real to Porter, he found a similar negative spike, in the reactor coclant system pressure history. This would be expected since reactor coolant system pressure was measured using the containment building pressure as a - reference; when the containment pressure went up, the reactor coolant system pressure would appear to go down. As noted above, a similar negative pressure pulse das observed in other pressure histories.
'See Figure V- E .
d
l l l t*, , 3 . Shift Supervisor William Zewe said that he had observed the pressure spike to occur at the moment the pressurizer relief valve was cycled, but that he did not associate the spike with a hydrogen explosion. He said that he had not been aware of temperatures in the core of sufficient - magnitude for the cladding-steam chemical reaction to have 6Sk occurred. Zewe was apparently unaware o' the in-core thermocouple measurements that implied production of hydrogen. Zewe said, not knowing what the spike could have that been,Athey ascribed it to an electical malfunction associated with the operation of the valve. While Zewe may have considered have been the pressure pulse toAan electrical anomoly, others seemed concerned that it was real. TMI Supervisor Joseph Chwastyk said:
... I actually saw the recorder, the pressure recorder on the building, spiked upward. I didn't know what caused it but the fact that the spray valves started indicated to me that we actuallv had some kind of pressure spike, either on the s'ensors or:in the building itself. I was not sure. The spike of course started all the building spray pumps, decay heat pumps, etc. The pressure spiked up and it was only up ,
briefly, as a matter of fact, a couple of heart beats. I know because I missed these heart beats. It came back down again." (I&E Tape 232/233, p. 9)
.Chwastyk (who arrived in the control room between 11:00 a.m. and 12 noon, and who was apparently not aware of the direct measurements of the in-core temperatures) also referred to an explosion in describing how it was tha:
he came to realize that the reacter core Might have been significantly damaged:
"It was like I said, everybcdy was pretty busy and I didn't wan: te step anybody frem wha: they were doing so I just tried te get a fee'_ for what was .. h a t. : e n i n e. bv Icckinc. around and askine. the coerators a: the panel what they were dcine. Up until the time er scretime after the e:.::lesien anf it dawned c.-
me what it was, I didr' n c .- h e .- .u i.- c'.a ma ge we h r. f . IIbif., p. 1;' t e- e..ri , - - -
. 4' Also,,TMI shift Supervisor Brian Mehler has told investigators that he believed the instruments to have indicated occurrance of a real pr&ssure pulse and not .to have been a
~
manifestation of spurious electrical signals. Mehler referred to the pressure pulse as having originated with'a chemical k.n.. .
- reaction; he stated that he did not reca'll t.hinking that .
a hydrogen detonation might have occurred. O. 'Okay, I think in your I&E interview you s' aid that
~
you thought initially that the pressure spike in the containment: when you saw it, was probably due to somebody fooling around with the transmitter. A. Yes, I said that. But then I also said it couldn't have been possible becauss building spray pumps started.
~
Q. Which meant that there had been apressure sicnal that 1 went through and started the building spray pumps? A. Right. Q. Now, after looking at it in uhat initial dismissal, did you later realize that there had been pressure in the containment that caused that spike.on the instrument? A. Yes. Q. Do you have any idea what cculd cause that kind of a rapid pressure spike? . A. I know Joe and I talked about it later on that day, about what could have caused it and I don't think hydrogen entered into it. We though: maybe some kind of chemical reaction or something hap:ened because it was up and down so cuick.
4 t e O. That is Joe Chwastyk? A. Yes. O. So you really didn't have a good diagnosis? . A. I personally'didn't think hydrogen could form that.i quick in the building to that concentration to ca'se u it in that period of time. (SIG, Mehler, p. 13-15, 10/11/79) aM "~ Mehler also recalled in his conversstions with investigators . that on March 28 he had been told not to turn on oil pumps in the containment, apparently out of concern that electrical sparks might be produced which could cause detonation of hydrogen in the l 1 containment-structure. On October 11, Mehler told SIG '
. investigators that:
il I do know sometime after the pressure spike happened we were told not to start equipment because they assumed that.
~
it [the pressure spike) could happen again and they'probably put it that there was hydrogen in there, but that was sometime g4 after 1:50. Now how far past that, I' don't know. And I do not, I said -- well, to Gary Miller I said -- he said don't start any more oil pumps and I said we don't have to, I already tested them all, because they were concerned -- but how far into the afternoon at that time, I don'- know whether ' it was 4:00, 2:00 or what, but it was sometime after." (.~2if., p. 16.) This recollection is similar to one repofted in the New York Times whereir Mehler was said to have
89 gr responded to his supervisor's warning not to turn on the pumps by saying that he had already done so. According to the Times article someone then allegedly said: -
"Well, that means we don't have any more hydrogen
- *$ in there'." .
Mr. Mehler was interviewed again on Octover 30 and at this time he was less certain that the above noted instruction not to start the pumps had been given on March 28. The following exchange took place during the October 30 interview where the questions' are being asked by Mr. Frampton of the SIG and the answers are Mr. Mehler's: Q. Since the interview that our group did with you on October 11, as indicated by Mehler Exhibit No. 2, you have become less certain that this instruction and the conversation you had with Mr. Miller was on Wednesday, . . .a March 28th. . A. Thct is correct. O. What is it that has caused you to doubt the recollection that you had before? A. I've talked to some other people that were there on the 28th, and also thinking back upon it, you know, I cannot be certain that it did happen,en the 28th. Q. In talking with other pecple -- well, let me ask you who ycu have talkee to about it? A. :'ve tal.ed 20 Gar. :'ill e r , ::ike Ross, Joe Chwastyk,
.. Eill Zewe, and ncne of ther recc11ect that instructicn being gicen en the :E:P.
37 -
. O. .Do any of them recollect such an instruction being given on the 29th?
A. I don't think.they would say specifically that it .
. i f
happened on the 29th either, but I do believe some of - , then recollect it being given. O. Do you remember which ones? W
~
A. I think Joe does. -
.a .
O. Any of the other people? A. I dcn't know. O. In your conversations with them, what is it that they have said that's made you think that your
. recollection is probably wrong that it was the 28th?
A. Well, they would have been in the room the same time I was to h, ear the instructions, and it seems funny, i if I would be the only one that remembered it happening ; on the 25th when there were other people in the room that don't remember it. Q. So this is in part conversations with other people that you have had and in part sort of a general reconstruction of events that's made you think that you're wrong.in thinking it was the 28th; is that fair to say? A. That's. fair to say, and also, you know, cuite a bit happened-tha 28th. And I did come back the 29th. It could have very well been the 29':h, and I wouldn't even say for certain it was the 29th right now. O. It's ccnceivable it could have been the 20th? A. I wouldn't want tc get definite and sa; it was -- you :r.ow , I'- just nc: certain right new which da; it was.
~
m >,---,,,w.- -- . , , , > - m-.---- w -.-y, +-- n - - - - - - - - n
7 '. ,
, .Q. Would it be fair to say that your own recollection, faulty or not, standing alone, has been that it was the 28th,
~
but that in talking to other people, you think that your recollection is most likely to be somewhat faulty and it was more likely that it was the 29th? ' A. That's correct. (SIG, Mehler, P. 15-17, 10/30/79) Chwastyk also recalls being told, ... not to restart any equipmen.t in the reactor building. And someone at the time had just finished starting a piece of equipment." Chwastyk said he thought the equipment referred to was the DC oil pumps on the reactor coolant pumps, presumably the 9
- same pumps referred to by Mehler. Chwastyk said that he thought the instruction had not been issued on Wednesday, March 28, because he remembered receiving it in the supervisor's office, and "... I don't think on Wednesday I was in the shift supervisor's office at all." (SIG, Chwastyk, p. 16, 10/30/79) ' ~
This recollection of Chwastyk is referred to in the SIG report (Volume II, part 3, p. 0147) and is part of the basis for the SIG conclusion that Mehler was probably incorrect when*he remembered the discussion, about not turning on the oil pumps, as having occurred in the shift supervisor's office on Wednesday, March 28. On the other hand., TMI supervisor Mike Chwastvh Ross did recall4 beihg in the Shift Supervisor's office on March 25:
"At time Joe (Chwastyk) would come into the think tank, (i.e. the shift supervisor's office where O
-y.
y
. x228019. Room E-352, e, x-43224, Room H-1035 supervisory personnel convened during the day for < ~
the purpose of assessing events and de,ciding upon actions to .bo taken) where we were at, and he would enter discussions." (SIG, Ross, p. 15, 10/30/79) si.
, t,v; . . - - .-. - - - - -
-[.-[ibbi$Mf((fTherSIG_. report does not mention'that Ross's recollection
- m. .: ::= . - = - - -
78 --; z 5as' to Cliwastiyk's appearance in the Shift Supervisor's office
~
;;iq :=c-J_. - *
't. W .D =J-conflictedivith Chwastyk's recall.i Tc tha't he didn't think
. .r -- r.. . - .
. k $ ~ ~ ,1-t ?-g; , : he had been in this offico on: Wednesday, March 28.
The SIG c
, e ,.: - . : : -- - .
{@3'.f.
. - . _ :5 .
- - ' report notes-but does not comment upon a further discrepancy E=-eel- -in.the-testimony of Mehler and Miller: Mehler testified f7:+; -
' me !Xjf.
; that it was Miller who gave the instruction not to start
-Q. ..-- --- .
the pumps even though the date was uncertain; i.e.- Mehler's A
- own recollection was that the instruction had been issued
;3 : w .
; .;w-",- A on the 28tha but after talking to other people, he said f
.._[e [ '* .7.t.
.... .s - .
hat his Tecollection was most likely faulty and.it was. ^
'f*,, .~~.
'~
J_ Z. '$.F. . J .2nore likely that it was tho~29th. (Ibid.) _ Miller fon the -
~>
."[.[-~ . .-other hand,-416, not recall at any time having- giverDan?~X
; , j = ; l : 2.U.a - . -. . . .
!'}.,2EP. . 'l .jIyinstructioit not- to start pumps i or even having been in the - ~~
.q g-m: q; :;.- -
- JFI, ?..
= control-room on the 29th. at the hour which~.was Meh.ler's . . ,
. 3= .3 .
-new. recollection.of-the time at which the instruction had
- 7 b:s . -
- ;-g. g ; 7
+- - -
TI:-! c ~ ~ 3een given.-- (SIG, Nol. II, Part 3, p.t14~I-148.) 3/-i
= .~ : ~1
=
.J
.~..*
Ni e -
~ ._
.{ :.
t b r _ _ see
40 - mheodore A reactor operator,/Illjes, who arrived in the control room after the detonation, told NRC investigators that he , had been. briefed on .the.. reactor b.uilding pressure pulse: *
"I was told thev had a spike on both indications of the ,
reactor. building pressure recorder. There was some discussion as to what it was. A hydrogen explosion was (I&E, discussed. This was later in the evening." pq Tape 261, p. 6) .
~
.W hen asked again as to whether the dis'cussion of a possible hydrogen burn had taken place on the first evening (March 28)
Illjes said:.
'"As far as I know that possibility was discussed that evening." (Ibid., p. 10)
At about 2:30 p.m., some 40 minutes after the hydrogen, detonation, Station Manager Gary Miller left the site for the purpose of briefing Lieutenant Governor Scranton. (E&E, TMI Part II, p . 2 7 3,) Miller has written that while in the - control room on March 28:
"I heard a noise at approximately 1:50 p.m., however, I did not associate it with the burning of hydrogen
~ ~ ~
or the actuation of the safeguards system at that . time. I was first aware of the recorded pressure. pulse and associated actuation of the safeguards system on Friday morning, March 30, 1979." , (Ibid. , p. 298) Others who were present say that Miller was aware of the pressure pulse at about the time it occurred. When Reactor Operatcr Frederick was asked whether others in.the control rocm had reacted to the pressure spike, Frederick stated:
"I think Mr. Marshall tried tc figure i cut, and Gary :*. iller was particularly interested in it."
(E&L, Par 2, p. 145) l J
.m 41 -
Zewe state'd in a deposition for the Special Inquiry Group that Mr. Miller was in the control room when the
~
pressure-pulse occurred (SIG,- Zewe et al. , 9/11/79, p. 257) , Zewe also stated that: 1...
"I found it hard to believe that an'yone who was in
%c - .
- - the control room observing anything would have missed';
- that ~(the spike) or turning off the pumps or any of the discussions at all."
Mike Ross, who was TMI-I Operations Supervisor, but acting as se'cond in command tc Miller at TMI-2 on March 28 stated when asked whether he was present when the pressure spike occurred: ..
"Yes. I was near the~ console at that time and if we are talking about the same time was around 2:00, sometime in the area. And at that time we got an ES signal and some of the components restarted, decay heat, what have you. We got building isolation again and we took care of'that and we looked back and the control' room and operator said "Jeese the spray pumps are running." We saw'a fairly we looked back at the charts at that time.
large spike on the chart and the exact pressure at this time I don't know, ... it was around 30 pounds. Jiy thought at the time and Miller was out there with us'and he cues:1oned he said, "3eese you know I thought I heard something, too." He are moving down the road there 100 miles an-hour and we looked at it and we said "Jeese the spike was so short it must have been an instrument." That was our reasoning at the time. We reached over and we said you can shut the spray pumps off now because the pressure came right back to 0, ... almost very, very rapid return and we shut the spray pumps off. I now know that spray pumps were on about five minutes when looking back because I did look back on that particular one. I personally didn't associa:e it at the m: ment with
b '
. Q .
any kind of explosion in the building. I associated
'it with an'instru=ent' problem perhaps anc u c4aiti se did Miller at the time Decause we gus; went on to somethinc else . It wasn't until the next day that we thought about anyth,ing'like that and started looking back. (ILE Tape #226, p. 4-5) [ Underline added.) -
de
, Shift Supervisor Chawstyk who had observed the pressure .
spike but did not hear the noise referred to by Gary Miller told the KRC investigators of a suggestion made to Miller on the basis of the spike having occurred upon operation of the pressurizer relief valve:
.Chwastyk: No, I did not hear the noise. But that was the point at which I had assumed that we did have some kind of explosion in the building. And that is when I suggested to Gary Miller we no longer cycle the electro-magnetic relief valve because it had ... the explosion ...
or rapid rising pressure.in the reactor building corresponded to opening the electromagnetic relief valve. (I&E Tape, #232/233, p. 18.) Chwastyk stated in subsecuent interviews with the SIG that he recalled informing Gary Miller of his concern that an explosion had occurred. On October 11 he said that en P. arch 28 he had been concerned even prior to the explosion that water should be pumped into the primary at a higher rate and :that:
"It was right after the hydrogen explosion and I
. mentioned that I correlated the opening of the valve with the detonation period that I again went to Gary Miller and explained what I thcught had happenef as
, far as the hydrogen detenaticn and the simultaneous opening of the valve, and i: :as shcrti-; after that, Gar. Miller got haci
, c re and said gc ahead anf dra.- the bubble." (SI2, C M-t a s : ,.; ,,,p_.,, 15 , 10.11. E.)
r .
+
.i
- t 2
Chi October 30, Chwastyk was asked again whether he thought he 5 had mentioned the possibility of a hydrogen explosion to Gary Miller in so many words, or discussed what that would - had - Chwastyk replied: meanorwhatphappened.tothesystem. se ,
"My best recollection of that is th,at I did relate j to Gary that we had some sort of an explosion. Whether
- I said it was hydrogen or not, I'm not sure. But I remember-distinctly putting together the operation of the valve and the spike, and I think I relayed those
' thoughts to Gary." (SIG, Chwastyk, p. 17, 10/30/79.)
When told that Gary Miller did not recall learning of the explosion until two days later on March 30, Chwastyk stated:
"Well that could very well be true. Again, I can't absolutely -- if Gary said -- I may not have told him what I thought at the time, because I really wasn't s: certain. (Ibid., p. 19-20.) .
Chwastyk was then asked again for his best recollection and the following dialogue ensued: Q. I understand. Let me ask you this: Was there any -- strike that. Let me start it a different way: When yet saw {put) this and theng it together what you thought had happened, that must have been something that gave you some cause for Concern? A. Yes. :t 2:ared the .sil cut Of s.
, ')
, Q. Did you think.that this was something that better ought not to be generally broadcast around the control room and outside?. Was there.any reason to keep this fairly ~close among the people who were there in light of the fact that -
it was fairiy alarming? ggg A. I'll say this: I didn't go out in the control' room and broadcast it, no. It did scare me, therefore, I'm sure I didn't just make it general knowledge to everybody ~in that control rdom. I'm sure I did pick out specific individuals that, my counterpart types of people, and talked to them about i: O. You said you think that you probably discussed it with Brian Mehler, and your best reccliection it that you 6f; cussed it with Gary Miller. Do you have a pretty specific recollection of who else you may have actually discussed it with on the 28th? A. I have some recollection of talking to someone from the NRC about it. At the time, I did not have the time to
-m discuss possibilities with him, and I think I related that I think there day have been scme kind of explosion in the bcilding, but I didn't know what. (Ibid., p. 20-21.)
At the end of the October 3C interview "r. Chwastyk was again asked by the Metropolitan Edison attorney participating in the deposition, about whether he had told Gary M!.ller on March'28 that he had correlated the pressare spike with a possible explosion:
r . 9 Mr. Diaz: I don't want to put words in your mouth. You recall making the inference, but you don't recall whether you conveyed that inference to Gary Miller; is that correct or incorrect? The Witness: .My best recollection is that I did related (sic is that information to Gary. That's the.best I can remember. .
. How much of that information though, what information I gave '~
him, I definitely don't remember. I do know that I gave (sic) himtheinformationofthebanhj)thevalveopeningsimultaneous13 with the pressure spike. Now, if I related that or if I put that together and told him that I thought it was a hydrogen explosion, if I thought it was an explosion at all, I don't remember. (Ibid., p. 28-29.: In sum, on May 21, l'979, Chwastyk told I&E investigators that he believed on March 2E that an explosion had occurred, and that he had told Station Manager Gary Miller that they a should no longer cycle the electromagnetic relief valve because the pressure pulse had corresponded to opening of this valve. On October 11, Chwastyk said after the hydrogen explosion, he went to Gary Miller and explained what he,
" thought had happened as far as the hydrogen detonation and the simultaneous opening of the valve." (SIG, Chwastyk, October 11, p. 18.)
In another interview on October 30, Chwas:yk repeated that he had told Miller what he thcught, that he did te ' ' ' e r his thcughts on the ... spar:- due :c the valve opera:icn causing scme scr: cf ar exp10sicr in the building." (SIG, Chwastyk, Octc':cr IC, p. 14.1 ';'.sr
.- pr=ssed as to whether he
o . 1 0, 44 e
~
- 45a -
~
actually did tell Miller that he thought there had been an explosion, Chwastyk (for the first time, on October 30) , "" - said that he could not be sure, although le thought he did. It is not clear from the record what was the basis for Chwastyk's changing his mind except that his recollection differed from Miller who had said he (Miller) did not recall learning on March 28 that there had been an explosion. Mr. Herman Dieckamp, President of C-eneral Public Utilities, stated in a May 9, 1979, mailgram to Chairman Udall:
"There is no evidence that anycne interpreted the
' pressure spike' and the spray initiatien in terms of reactor core damage at the time of the spike nor that anyone withheld any information."
4 Ms
- l s
c , Loc Entries < Two logs containing a listing of significant events - that occurred on March 28 indicate that a pressure pulse t.g , occurred at approximately 1:50 p.m. One, log lists the pressure spike as having been 4 pounds per scuare inch and accompanied by actuation of reactor buidling sprays; the other log lists the pulse as having been approximately
. 5 pounds per square inch, and does not mentien the containment-spray.s . None of the investigation reports addressed the cuestion as to why the pressure pulse was indicated to have been 4 or 5 psi when in reality is was approximately 28 psi.
In fact, .the only reference in the SIG report to the logs indicating that the pressure spike was 4 psi is a cryptic one apparently in support of the SIG conclusion that, contrary ,, to the perception of Chwastyk and Xehler, the spike was disr.issed as being insignificant; 1.e. SIG states:
"Furthermore, an entry in the centrol roen operator log book for the afternoon of March 2E nctes that at 1:50 p.m.
an engineered safeguards initiatien signal was received, the reactor building sprays came on, and the reactor building pressure spiked up no 4 psi." (SIG, Volume II, part 3,
- p. 144.)
V h
.s ,
46a -
- a
. . . . . . ~ . - . - . -
.. ..- --- ... = _ .
.? / . ." * ,
~
'! /- *
". .; 4 . ' .#.: .% ' .v . . J . . r: . _J ". . .*. _
O , j.;* *. .: f e/.
.-..'4.
.;. J . g_ .. ,
.. .. . . + . .../...-
i. s . . n
.- . ,e.. ,,,r- ./-.~;,,,,....--.
i.'....,:-_.... ...-. ..- w . .. , J .-r
. .__s.. .~ at.- .:'.,
I, ;,. , . .; . . , -
. =. , . . , . . ... . ,
. - .~. . ..c._......
_. -- .- n. ..--: . , - . 9 ,s. - - <
-.- . .,.-n
- c. . - . . . . .
.a r.g c r . A-d* .~.r 's. ' .r.a j
1 1 ;. i _ _ / 1.,..: .- V h ',' ..J. c a'i
# ./:.. !;; : ,
e.* v- *
\*
.u.-......--..
T ~ i L G. .. .5 L=t R._ ~ - ~^ ^. \.w..- S..
-- m.',-. W. 4h,, .- .
- i. s
. .s,c
. .' .v.. .a. _ - - . _ . . . . . . . . . _ . . . . .
a . . . . - - - . . - - _.... . . . . - y- = . . .= - - - . c./ 1 ,2 15 w .' a. !2
.f o - - - - - - - . - - - . . . . . . . . .
. . . . .% c .. ,\, ,, _
f .. f.4.. . ' . . ; .**9"._ .,.. o -
.t. .
*~
_/.,, i, L.'
?,,,g,,, < tr N 1*a.
m e i ., ,, JO., i' ,
.c c-y. c a-
-E..- .y%. c,.vs :,. ~y= !..s. uc s'.sL. , -di.% .'< !&Rf . p.. :.- W .
~
i 7 .. . . . i. r,sr l .._..,.1. u u(... w. W.U. .$. .. Z . . ..A. .:,':.~ ~,s. ./. &. -( C %. ,f.c. . .c,2. .4. % . .. . - . . . ,
! ' /
CL:mi& L..g . C.Ces -t.. .(1h t s .-.- . . . --.... -. - .. . . . _ -.. - ..... d
o 47 - s Presence of Hydrogen in the Primary Cooling System Hydrogen produced as a byproduct of the oxidation of zirconium cladding collected at high points in the reactor cooling system; e.g. at the tcp of the reactor pressure vessel, pressurizer, and upper (candy cane portion) of the-2a ~
' hot legs ( See Figure I-A, I-B; Pages 13a,13b.) Wnile.the hydrogen in'the pressurizer could be released'from the system via the pressurizer relief valve, there was no such valve in the other locations; hydrogen could be eliminated from these areas only by flushing it out or forcing it to dissolve in the water, and then'either allowing it to escape from the water in the pressurizer wherefrom it could be let into the containment or extracting it from water removed via the letdown system. A mixture of hydr' ogen ~and superheated steam in the candy cane had for most of the day obstructed filling the primary system with water, thereby preventing an ef fective use of the steam cenerators for removal of heat croduced in the fuel bv radioactive decav.
-Qata available in the control room had, in fact, indicated (1).
that there was a gas other than steam in the system. The possibh (or likely) existence of a non-cendensible gas (most likely hydrogen) in the system appears not to have been reported either to the Nr.C c officials of the State cf pennsylvania. . p.. ese = : r...i
-.. .. c... , ..c..w. . ..=.,.=. .,, =. . cc.
I
o .
, *4
- Perceived adecuacy. of procedures to assure adecuate core cooling.
Statements made during interviews conducted by accident
~
investigators indicate that during the day of March' 28 there were periods when .TMI personel n appeared uncertain as to core conditions and whether the procedures being followed and actions being taken were adequate to lead . ,.n ' "~
' to stabilized conditions.
TMI Station Manager Gary Miller told General Public Utilities invest'igators on April 12:
... We, being me and Lee Rogers, called Lynchburg pretty early, and we sat in the roca and every hour tried to figure out how to keep pumping water into it. But all we were doine._was pumpinc.
that BWST / Borated Water Storage Tank / through the electromatic to the floor. Nothing was changing, so you know we pumped 12' or 13' out of SWST and =y fear was pumping 50 feet out and the core still hot and no water in the BWST. So our goal was to somehow get some circulation going, either natural circulations from steam generators or reactor coolant pump, using HP injection the whole time. My memory is that we pumped against the electromatic at fairly high' pressures like 1800 or 2200 psi. We could have pumped a, gainst the codes, but we. assessed - that what we'd do is get the same flow through. the codes without being able to see pressure. . We pumped through there until around 11 in the . morning, at which time we decided to take a shot at getting on core flood. And the reason we took a shot at core flood -- now remember Lynchburg was on the phone with a let of good advice, but it was clear that it was my decision -- we assessed that if we could get dcwn and activate core flood tanks and we saw them dump on the core we could get assurance that the core had some water on. We couldn't tell that; we were scared that wasn't happening. Radiation was , all over the place, everything was cff scale. You sc: nothine in the cere tha: tells you about water level; you get ne pressurizer level, since it's solid; no way of drawing the hubble; I did.'t have an,. heater.s; I didn'- have ar; leticwn; and we had radiation in everi recr e went to. Didn't ever have til purps fer s: e Of the RC pumps; cculfn't get iT s 0' ~ O f the TC O~.s ; the rehd:ngs wire h:ricTfeus. L. _
r . s# - (' M e Mike Ross gave the following rationale for reducing syster. pressure in order that cooling might be achieved from what was thougMito have been a better method than 9 the one they were then using which depended upon keeping , the system pressure high and using the high pressure injection pumps:
... One, we were running out of water in BNST and we hadn't gained an inch. I mean we hadn't gained any headway in where we were trying to go our goal being one to establish some mode of cooling a reliable mode of cooling to the core. Two, we, at least I wasn't sure
, that we purposely or we in fact had the core covered and all high pressure injection was going to the core. I was not totally convinced.
I didn't have anything to tell me. Hey, high pressure injection is in fact going through the core. So based on that, we discussed going down with the goal being one, to let the core flood tanks come in and verify that yeah, the reactor was in fact covered, two, give it a drink of water. That is a little coarse I know, but give it a drink of water if it wasn't getting it, and maybe go on decay heat removal which is a forced mode of cooling. That is what we were trying to do at that time." e (IEE TMI Tape #226, p. 26) Perceptions concerninc ccre uncovery The record contains few specifics with regard to perceptions of these present at TMI as to whether the core was or had been uncovered. It appears that TMI personnel concluded that, after 9:00 a.m., the core was prebably cov'ered. The apparent basis for this conclusler was that the pressurizer water temperature was at less than saturatien level, and since the pressurizer temperature was an in ica:cr cf ccre cutput hu .i.. _ . . _ _
t 50 - temperature, pressurizer temperature was assumed to indicate that the core was probably covered. There is little in the
~
record to indicate the interpretation of the high in-ccre thermocouple temperature measurements, which in fact showed local boiling and a badly damaged core. Wi . While it was generally assumed the dore was covered af ter 9 : 00 a.m. , there is little discussion in the record to suggest what the thinking might haue been with regard to core uncovery between 6:00 a.m. and 9:00 a.m. As the foregoing discussion suggests, there was censiderable temperature data to indicate that the core had in fact been uncovered during-this period. The most extensive statement on the record is an interview conducted by Kemeny staff with B&W engineer John Flint. Flint was apparently unaware of the extended period during which water had leaked from the system through the stuck open pressurizer relief valve, and he was also
.a apparently unaware of the in-ccre thermocouple measurements .
indicating temperatures in excess of 2000 degrees. On the basis of temperature and neutrcn data, Flint did concludg however, at 10:30 a.m. or thereabouts that the core had been uncovered earlier, even though it appeared to him that by 10:30 a.n. it was again covered. Flin engaged in the f ollowing dialogue with Kemeny interviewe,rs. The answers are Flinds:
~ .
. . *r
, . t
-Q. When you_ reached the conclusion that the core had been uncovered approximately an hour to an hour and a half after you arrived, which would be something in the order of 10:00 to 10:30, did you tell anyone?
A. Yes, I'did. I believe I mentioned it to Lee Rogers at the time. D . . Q. What was his reaction? . A. 'I believe he went to discuss it with Gary Miller O and George Kunder. Q. Were you present during that discussion? A. No, I was not. Q. Did he report back to you? A. I don't remember him addressing that specific cuestion, no. Q. Did you ever find out what discussion he had with Kunder and Miller with respect to core uncovery? A. Not that I ca.4 remember, no. Q. Did you tell anyone else that you had reached the i conclusion that the core had uncovered? . A. Eill Zewe, Ed Fredericks. Q. What was their reaction? A. I would say surprised. O ,. It was news to them? A. Yes. i O. It was news to Rogers too when you told him?
, A. That is correct, so far as I know.-
[Kemey, F'in , p.22, line 2 to p.24, line 5.)
52 - Percept' ions held on Mar'ch'28'as t'o extent of core damage. The foregoing indicates the existence on March 28 of . manifold indications of core damage. Based on partial "5 information available to him (i . e. he did not know of the 20i00 degree core thermocouple measurements or the manifestations of an explosion),,. Victor Stello, then Director of the NRC's Division of Operating Reactors, told the E&E Subcommittee TMI Task Force on May 9 that he suspected oxidation of the cladding (and presumably hydrogen production) on March 28. Mr. Stello engaged in the following dialogue: Mr. Stello: My initial reaction was that considerable f ailure of the fuel had occurred, with large numbers of failed fuel rods. And that the potential for oxidation, metal-water reaction, was clearly there. The periods of time involved were, even if the heat-up were not significant, at lower tenperatures, with the top of the core uncovered for extended periods cf time, metal-water reaction was clearly possible. The rctes -- how high the temperature got I would not even indicate -- I could have cuessed at. Eut oxidation of'the cladding was something I clearly expected; failures of the cladding were clearly evident. Mr. Terrell: When did you suspect this? On Wednesday you are talking about that you expected this fuel damage? Mr. Stello: Yes, I think just from knowing what was happening te the core, you had to expect that there was quite a bit of damage tc the fu'e1. (E&E, TM: part 1, p. 4-5) 5
n
., i
^
7'
~> '
Another indicatios of there being suspicion on March 28 of hydrogen production can be inferred from a statement made by TMI supervisor James Floyd. As noted above, Floyd was at the'B&W facility in Lynchburg, Virginia on March 28 , and had inferred from limited information that about gg one eighth of the cladding had failed. On May 31, Flovd
)
told the Kemeny Commission of his suspicions.about hydrogen production when he engaged in the following dialogue with Commission member Ted Taylor: Taylor: Had you sometime Wednesday made the connection in your mind between the high temperatures in the
. core, cladding failure and hydrogen?
Floyd: Yes,' sir. . The March 29 New York Tines described a press briefing presented by Jack Herbein en March 28 which gave a different picture of the situation: Jack Herbein, a vice president of the utility that operates the plant, told a news conference held on a bluff over-
., looking the plant that "a valve failed in a shut position" in a feed pump that scuirted water around the reacter. -
Mr. Herbein, spent almost an hour describing the accident and emphasized that the rain safety systems in the almest brand-new plant had worked to preven; a very serious accident, or indeed a catastrophe. Using bland terms, he described the series of events in the plant as "not the normal evclutien" in stating
. that there was "some miner fuel failure."
This translates to the fact that sens of the pellets of enriched uranium fuel became se heated because of loss of coclant tha: the pellets nelted through the zirconium clad tubes that hold the pellets. Uni: Sc. 2 has 177 fuel assemblies, each centaining 20E reds, each cf which centain X uraniur pel'ets. _ l
" O n * ;, a f c'.: cf the-
_ e':ed thrcugh,' 'c r . Herbine said.
.+ ,
e . *. - 54 -
- s. .
On March 28, NRC issued a Preliminary Notification of Event or Unusual Occurence which gave little indication of the actual severity of damage: , Facility: Three Mile Island Unit 2 Middletown, Pennsylvania , n_ a (Docket No. 50-320)
Subject:
REACTOR SCRAM FOLLONED BY A-SAFETY INJECTION AT THREE MILE ISLAND - UNIT 2
- in The licensee notified Region I at approximately 7:45 a.m. of an incident at Three Mile Island Unit 2 (TMI-2) which occurred at approximately 4:00 a.m. at 93% power when the secondary feed pumps tripped due to a feedwater polishing system problem This resulted in a turbine trip and subsecuent reactor trip
-on Eigh Reactor Coolant Pressure. A combination of FeeC Pump Operation and Pressurizer Relief - Steam Generator relief valve operation caused a Reactor Coclant System (RCS) cooldown. At 1600 psig, Emergency Safeguards Actuation occurred. All ECCS components started and operated properly. Water level increased in the Pressurizer and Safety Injection was secured manually apprcximately 5 minutes after'actuanien. It was subsecuently resumed. The Reactor Ccclant Pumps were secured when low net positive suction head limits were approached.
At 10:45 a.m. the Reactor Coolant System Pressure was being held at 1950 psig with temperature at 220'F in the cold leg. By
' 10: 4 5 a.m. , radiation levels of 3 mr/hr had been detected a 500 yards offsite. (PNO-79-67) 9 9 I 6 I
r . r ..
' L.,
.. The March 28 Washington-Star quotes Lieutenant Governor l
Scranton:
"Everything is under control. There is and was no danger to the public health and safety."
Officials of Metropolitan Edison / General Public Utilities have said subsequently that on March 28 they did-not M '
. reach conclusions such as those of Stello and Floyd referred to .
exwe about the possibility of cladding oxidation.
. .1 In answer to a written question as to what at 11:00 p.m.
on March 28 was his estimate as to the amount of cladding that had been oxidized during the period when the core was uncovered, Herman Diechamp, President of General Public Utilities, responded: I interpreted danage to mean cladding failure. I did not think in terms of core uncovering, high temperatures, or Zr-Hp3 reactions. (I&I, TMI Part 11,
- p. 2 8 3,~ )
In response to the foregcing questien, Metropolitan Edison Vice President John Herbein stated: +. I didn't consider a steam clad interacticn producine , hydrogen until it was discussed on March 30th. (Ibic., 6. 290) . In response to the question as to what was his estimate on March 28 as to cladding oxidatien, Static,n Manager Gary Miller stated:
... I did not conclude there was an on-going steam-claddinc interaction and therefore, hac ne estimat'e of the amount of cladding failure by 11:00 p.m. en P. arch 25." (Ibid. , p. 295.)
f, * -
]
p .. w e' .
,7 * = *
- Whatever the indications of hydrogen production available to the NRC on March 28 and early March 29, little idea of these was convey'ed to the E&E Subcommittee on March 29 when, .as noted above, Chairman Hendrie stated: .
SI "The magnitude (of release of radioactive gases) suggests that perhaps one percent of the core might
. have been involved in the cladding cracks." (Transcript, - p. 60.)
P + 4 5 S O i I e ! l O { H l l J
r '
.v
~
~~~
F07A iIV3I.I- *; % = l. . . - .
. 4. -
m% '.ax. TE DanscTost _ .ia. _y,y 3, y
- g=a"- CoNMITTEEON INTERIOR AND INSU1.AR AFFAIRS .sr .se swam ?
* * "A8 * *"G-
" _ 7,;,"* Q,"gy ""*. . ,U.S. HOUSE OF REPRESENTATIVES
"* WASHINGTON, D.C. 8^ N 20513 (d'**' ,*"N"" ,,
7 8* "-- (, . '
;;7 ___~: -*- ~..~.T c.. =ll .
= ..-l:.:: _. --": ~r . January 21, 1980 *^",L*,;",.fM*" s
,, [._
~:= :;'- -':c / o
- ,~., ;;.,, -~
f( ,AIl - . . . .
=,=.,::;,':l=::=l:r . _ . ,
=~L ;,.L . ~
'::::"::. lll2R . -)0 3 g g ,,,,,g ...
The' Honorable John Ahearne .. Chairman, Nuclear Regulatory Cnmmission . . . _ . . Washington, D.C. 20555 g.
Dear Mr. Chairman:
.==.
On March 28, State and Federal officials were unaware 7.6 of information indicative of conditions at Three Mile T:? Island. As a result of being inadequately informed, J.:2.:l government officials were severely inpeded in fulfilling their responsibilities with regard to protection of the ..._.
~~~
public health and safety. An important question, therefore, is.why on March. 28 these officials and the f5
. public were denied important information. 7,[
~
. E:.
I believe that the Commission to the best of its "5 ability should address the following before closing its - inquiry into the accident. - lll. l;
- E~
l
- 1. Station Manager Gary Miller has stated that, "The [E" in-cores were reading anywhere fron 2500 or so, and :;s.;.
I picked 2500 [;] it could have been higher than that. ;;.;:~ ~
. But you know, I was looking for a gross indicator 5.;
and I had it." What did Mr. Miller believe to be == the significance of such temperatures? Did Mr. i.~fL Miller believe on March 28 that during the day ~~i&, portions of the core had been uncovered? What was 5s7 Mr. Miller's understanding on March 28 as to the E.. temperatures at which a significant steam-cladding =~ ?. . E chemical reacti'on might occur? E.
- 2. Is Mr. John Flint correct in his recollection that, E >..
"These [in-core] temperatures were monitored for XE
=~
the rest of the day [ March 28] to follow what was happening to the core?" ===
.e
- _-~;
\ at=
".'.*.~;.*:.
L :--::.
' ~=~Y . . . . . . _ . .
P . , e .
. The Honorablo J hn Ahearna Page 2 ,
5 Frr y:
?.;':: .
- 3. Which of the TMI supervisors present at the ifi plant on March 28 believed some of the fuel rods :=-
had been cooled primarily by steam for portions E5: of the period between 6:00 a.m. and 8:00 p.m. on == March 287 What was the understanding of these -" supervisors on March 28 as to the cladding temperature :_.... at which a significant steam-cladding chemin 1 === reaction might occur? Ei
. EE
- 4. How does each of the TMI supervisors, who E' was present on March 28 and who did not believe EE.
that portions of the core had been uncovered, Eii-interpret the hot-leg temperature data indicating ::= superheated conditions? E "-
==:
- 5. Is Instrument Man B correct in his recollection E:;:
that the instrument technicians had told Mr. Ivan - . . - Porter of the results of their measurements of E the in-core thermocouple voltages? EE g=_ =
- 6. Are, Messrs. Chwastyk, Ross and Frederick Fr
. correct in their recollections that in the !E afternoon of March 28 Mr. Miller was aware of the F" pulse on the pressure.recor'de'r and/or actuation i of safeguards systems that occurred at approximately [:f 1:50 p.m.? -
- e .
~
- 7. Is Mr. Illjes correct in his recollection'that ' 2'
.on March 28 he was' told about the pressure pulse, and that the possibility of a hydrogen explosion f
Ef was discussed in the evening? E== F:
- 8. Which of the TMI supervisors who were present isf
. believed on March 28 that the pressure recorder ;~l....
had indicated a real rise in containment pressure p~ at 1:50 p.m. and which of the supervisors believed E.i the recording to have been the result of spurious E.) electric signals? ==
- 9. What of the TMI supervisors present on March 28 were aware on that day of the following __
- actuation of containment sprays; [i[.?
.E=.:'1 negative pressure pulses on 5:....
instruments that used containment pressure .g;; as a reference; MET
??... .
e :' C... 3.b A
,a eee em 66@Di 6 g
j . . . J
. ,a s -
=:-
4 The Honorabla John Ahearne f.1
~
Page 3 - i_~ ns..
- a rise in containment temperature ,
recorded by one or more containment temperature == m sensors? =;:.
. 1E0--
- 10. Did Mr. Miller and/or other Metropolitan .
- ===
Edison officials have an obligation to report on _...
' March 28 any or all of the following to State and ===
Federal officials: -==
- in-core thermocouple data indicating hM
, temperatures in excess of 2000 degrees F. ; !EEh.
EE~
. - computer printouts of in-core :==
thermocouple voltages indicating presence of EE~ superheated conditions (and partial core uncovery); E"" EEE
- bot-leg te=peratures indicative of "-
'"-~~
superheated conditions prevailing for most of the period betveen 6:00 a.m. and 8:00 p.m.; g.3g.
- =
- the presst;e pulse recording and other gg manifestations of a hydrogen detonation in the Et containwant at 1:50 p.m.; and, !L.,7 EE:-
-- uncertainties during the day as to EE:
whether the core was covered? - E=. Er'
. n In order to assist you in your consideration of this j=g.:
matter, I am enclosing a draft working paper' prepared by h.. ' _ Committee staff. I would also appreciate Commission lid._. comments on this working paper. E~". .
$ Sincerely, E_
. i, IEE
/ EE.t.
l lEN?.? MORRIS K. UDALL ==-. Chairnan ==
=:- .
Enclosure Eihr
- E=i
- =2
'2 25 b
. .gg
~=*2=
ww - ..
p *
' , 1/21/80
. Draft Working P per !
~
l
. Early Indic tions a of Core' Damage' at' TMI-2 ,
At approvinately 6:00 a.m. on March 28, the water level
.in the TMI-2 reactor pressure vessel fell below the top of
~
the active portion of the fuel rods. During the next several
, hours a* substantial portion of the reactor core was cooled ;
4 ; primarily by steam produced in water boiling within the pressure 4.c6% vessel. The steam was a significantly less efficient coolant th.an water, -and for parti of this period radioactive decay of fission products generated more heat within the fuel than was removed by stean flowing past the rods. Therefore, the cooling mechanism was inadequate; zirconium cladding of the fuel reached temperatures at which it reacted with steam producing zirconium oxide and hydrogen, and generating 1
. an additional significant quantity of heat that needed to be removed from the pressure vessel. ,
~
,. : .- _: ~
. Prior to the z'irconium reaching temperatures at which oxidatiot occurred, increased pressure from gases in the rods and increased ductility resulting from the higher temperatures caused the zirconium cladding to bulge and split. Kemeny Commission staff concluded that all or nearly I all' 36,000 rods balloned and ruptured, releasing the radioactive
_ gases contained therein. (Kemeny, Core Damage, p. 4.) , Kemeny Commission staff also concluded that during the
. period of core uncovery fuel temperatures in the upper 40%
to 50% of the core exceeded 3500 degrees F., and temperatures in 30% to 40% of the core exceeded 4000 degrees F. (Normal - operating tenperatures are about 560 degrees F. NRC regulations specify that emergency systems must be designed so that peak r 1 1 - - - -
. _ _ _ -_ _ _ _m .
zirconium c12M4'g. temperatures shall not exceed 2200 degrees F. in the event of accidents which must be taken into account
.in the reacter design.) Remany staff estimated that ultimately between 44% and 63% of the core's 20 tons of zirconium *
' cladding were oxidized, producing thereby some 1000 - 1300 pounds of hydrogen and causing some 60% to 70% of cladding to l
1 . become so embrittled that it lost its structural integrity.
,, , ; (Ibid., p. 1). (ERC regulations require that reactors be ,
designed so ' hat no more than it of the zirconium oxidizes in the event of accidents of a kind that must be taken into account in the reactor design.) Kemeny et al. concluded that as a result of exidation and embrittlement of the fuel rod cladding, several feet of the upper part of the core fell into the gaps between the fuel rods, causing partial blocking of the flow of steam or water that renoved heat from the damaged fuel., During the course of the accident, the TMI reactor system
+.98.48 8 dif parameters evolved into such an unusual configuration that the operators were unable to achieve a relatively stable cooling arra:1gement until late in the day on March 28. What was the likelihood, in these circunstances, that further syste.ms
' deterioration, leading to a major radiological release, would occur prior tn stabilization? Kemeny et al concluded, subject to the caveat that the assumptions used in their calculations were valid, that there was a high probability that the contai r=nt would have , held even had there been a meltdown.
Nevertheless, the answer to this "what if" question will probably bn the schjent of long and indeterminate debate. What is clear, however, is that m management, in spite of many manif estations he
. . - - , , . - m- --- - - - - -
..w n
.___ 3 _
~
of severe damage and an unknown but presumably significant on March 28 or : potential for a major radiological release, did notfreport - fully on the situation'to either the ? officials of 'the -NRC of the State of Pennsylvania. While some NRC officials were.
- = = =
aware that the core had been uncovered, it is unclear what. l. g,. . . they inferred from this with regard to the extent of damage. . 1 Nor is it clear how much of what the NRC staff kner wa.s passed on to Chairman Hendrie. At a briefing presented to the Subcommittee on Energy and the Environment on March 29, Chairman Hendrie stated that: The result (of the equipment malfunctions ] has bean apparently some overheating in the primary system that caused the venting. There apparently.has been, judging by the radiation levels, some core damage in I the sense that fuel ods have leaked gaseous radioactivity, _ I rather doubt there is melting. There does not - seem to be any indication of that.* (Transcript, p. 23) x-
- Kemeny staff concluded that some of the uranium oxide fuel may have become liquid at the temperatures which exceeded 4000 degrees, even though the temperatures were less than.the uranium oxide melting temperatures of 5200. The reason for the fuel liquifying at temperatures less than the uranium oxide melting temperature is that,it was in contact with zirconium and the melting point of the zirconium / uranium oxide mixture was lower than the melting point of the uranium oxide standing in isolation.
.. :.. . .. . .: .. . '. v_
. ~ : . . . -- .=. -.-
. . u l .1 - --J--
L. . . . ::x:: ~..1 i. ; A. lC: *
- . . ~; .-
.- .;.- ,: . d.
T 2'.
.a t. : y
-. 3 ..: L. w. ;. . .
----.-. v . _ . :. , . - :~ .-. .. ..n.L:7
. .. . . - v u. -w : .-
-. n ; . . : ?-;.t :. : :s=
- .. .. pr . ;-
.- :y.;.- . .-_;.. ;a
.~ . .. .:...... . p. .- -- -
.,- - -..~- .;- .- --._ -
.> -.- +-n-,. L = -- ...~ -~ ~
- c. .
~
_4 ,l It appears that Governor Thornburgh was even less well informed. Had Governor Thornburgh on March 28 known what might have been inferred from the instruments in the TMI-2 control room, . he would have likely given serious consideration to ordering,
.a precautionary, evacuation. Yet this was a step not seriously 4 considered until two days later on March 30 when NRC officials wesaw 4
were notified that a pressure pulse had been recorded at 1:50 p.m. . n;;;s . on March 28, indicating a hydrogen explosion
- in the containment building. This and other information led the NRC to conclude that the reactor had been much more seriously damaged than previously
- the Commission /had thought, and that a major radiological release was possible either from an explosion or from further 4
deterioration of equipcent which was already performing under conditions for which'it had not been designed. To date neither of the two major completed TMI inquiries -- that conducted by NRC's Office of Inspection and Enforcement and that conducted by the Kemeny Commission -- have substantively w
*All references to hydrogen explosions or fires in this discussion concern the hydrogen detonation outside the pressure-
. vessel in the containment building that occurred at 1:50 p.m.
on March 28, the first day of the accident. This is not the potential explosion that directly concerned the NRC in the period March 30 - April 1, when the Commission believed there was a poss,ibility of a detonation or fire within the pressure vessel resulting from generation cf oxygen which might react with the hydrogen bubble that was believed to occupy a ; substantial volume at the top of the pressure vessel. Subsequent ; analyses appear to indicate that there was no mechanism under conditions then prevailing within the pressure vessel that could lead to production of oxygen, and therefore there need not have been serious concern about an explosion within the pressure
' vessel. While there may have been no danger in this period ,
from an explosion within the pressure vessel, the danger had l not passed and a further deterioration of reactor systems ' night have lead to a major radiological release.
.l _
1
_4_ . l z It appears that Governor Thornburgh was even less well informed. Had Governor Thornburgh on March 28 known what might have been inferred from the instruments in the TMI-2 control room, he would have likely, given serious consideration to ordering , a precautionary evacuation. Yet this was a step not seriously considered until two days later on March 30 when NRC officials fehnew were notified that a pressure pulse had been recorded at 1:50 p.m. on March 28, indicating a, hydrogen explosion
- in the containment sw. .
building. This and other information led the NRC to conclude that the reactor had been much more seriously damaged than previously the Com:sission/had thought, and that a major radiological release was possible either from an explosion or from further deterioration of equipment which was already performing under
, conditions for which'it had not been designed.
To date neither of the two major completed '5MI inquiries --
~
that conducted by NRC's Office of Inspection and Enforcement and that conducted by the Kemeny commission -- have substantively w *
*All references to hydrogen explosions or fires in this discussion concern the hydrogen detonation outside the pressure-vessel in the containment building that occurred at 1:50 p.m.
on March 28,' the first day of the accident. This is not.the potential explosion that directly concerned the NRC in the period _ March 30 - April 1, when the Commission believed there was a poss,ibility of a detonation or fire within the pressure vessel resulting from generation of oxygen which might react with the hydrogen bubble that was believed to occupy a substantial volume.at the top of the pressure vessel.- Subsequent analyses appear to indicate that there was no mechanism under conditions then prevailing within the pressure vessel that could lead to production of oxygen, and therefore there need not have been serious concern about an explosion within the pressure vessel. While there may have been no danger in this period from an explosion within the pressure vessel, the danger had not. passed and a further deterioration of reactor systems .
, , , , , , might have lead to a major radiological release.
~
~
.~
, ,. . 4 s. -
- 5. -
addressed the reason for the failure on March 28 of TMI management to report fully on the severity of the situation.* An examination of data available to the managers, of &ctions - taken by the managers, and of post-accident statements by persons who were at the site shows numerous indications that on March 28 the managers were aware of information
-,. indicating that portions of the reactor core had been uncovered
*In a supplemental view in.w.@ rated in the Mmeny %i. en 'INI, Remeny M"ian nw-+mr Bruce Babbitt stated:
It now appears that there is evidence to indicate that Met Ed ,
- - technicians understood, within a few hours of the accident, that the ~;
cuclear core had been uncovered and that this specific information was, transmitted to supervisory personnel at the plant early Wednesday. There_ pee =s -to be little question that the technicians who took the < tecperature readings that niorning understood what they found. The real cuestion is what happened.to this information and whether it was j transmitted to the appropriate management personnel. It certainly did not get transmitted to respnnsible public officials, including
-, Lieutenant Governor Scranton during a meeting with Met Ed that afternoon.
4 e* ,bO ' SG l M e w G $ O em.
, , ,--- _ .,,,.,.._--.,r.. . , . - , , - , - - , - - ~ , - - , , , ,.-.--,-,-,,.,,_---r,-.- ,
, . =.
6. for several hours, and than during this time a substantial quantity of zirconium had reacted with the steam to produce zirconium oxide and hydrogen. . Among the major indicators of the core uncovery and/or 1 oxidation of the zirconium were data availab'le on March 28
.p .
showing that fuel rods were being cooled by superheated steam r4ther than by water, fuel cladding temperaturec in excess ; r . _ of levels at which zirconium midizes, neutron detector data showing voids in the pressure vessel, data indicating that hydrogen had burned or exploded in the containment building, l and data indicating the presence of hydrogen in the primary
~
cooling system. .
. Emergency CoI=nand Team j.
The following discussion addresses the various. kinds of
~
infort:ation known to plant management indicating that the l reacter core'had been severely damaged. In considering the kinds of data that were available, it is useful to keep in ! r:ind the emergency management organization and procedures established by Station Manager Gary Miller. Upon his arrival at the TMI control room shortly after 7:00 a.m., he organized a group for the purpose of analyzing information and determining what actions should be taken. In a statement dated May 7,
. 1979 and presented to the E&E Subcommittee on May 24, Miller _
2 described this command structure: .I e 0
- - . . - , _ _., . . . _.. _ , ..,_m._._..,,..,,9-D** -O---- --
v-m-e-- * , ----' mv e ---t-
l .. . . U.
,- . ': _. [
7,s : e say artmary gual mes to sew:act :?w pielte and our actions mere an attsuet to f-thietae re!aeses. West ay, rec:suoric e ecuattm if morreeted. =ith sazimum aevance . me:1rica. inn to ta-sinate the incide=st and to stabillae ce Unit. The firs
- cuncara had as I arffved is tne Castrol h.
- at apprestaately F:05. mas ta became fally capit:an.of the sit:aattes as it existed and once I fully understand i
ce :1mc c:r.dt: toms and me raffatten emergency.1 innedtately took charge of :p Can*:'o1 anse and a: pointed Sonice people ta direct the necessary evoluttans in me vital armas to asstre that the psblic ses pro
- acted, that the release was monitored, tnat commaar*ca:!srs we*e occ=rritg and tha: the, plant ws brought in steps to a stable conettfoe. The e==-=' set so. =htch I Just cescribed. set frequently t.*.rougho::t l
l t:ne :ay. Tte ;-su: grese:ted Ut*t concitions, status of Emergency plan actions, l . l shared ccie*:ns, c'sc=sset to:hefcal data, and made recommendations. Discussions wert
. held witn hnageme- amf.'or Bak: met & il11cox. Lynchburg, t.6e State. the .9RC and f=11: wing t ese evsluattoes. I made apr op-f ate decisions and so directed the t
tacionar:::att.n :s ::e w:fol Roma and informed others both inside and outside tae i
- are as r.e:assary.
l . i lasi:ali *f set u: t::1s !merpev/ catmand test in the early hours as I arrived at t9e : Tar arc One nef atim sue gency =as in progress, by esser.itally for=ing my semier Te:p's feto a par:w:-t to u;ervine, c:r.:act the rergency and report to me while l t-in;try De piast is a ssfe ton:ttfon. Mr. R:ss was put in charge of C;erations t: cd ect : e Smif. 5 oe-viscr. *r. "ktiei mes put la charge of ractation c:scer-s. incl.ctes ra:f a:1cr s:.; veys. :ns':a U*fsfie teacs, acc untability check, asse=mly i
~e c' pe:::e. ge :"n; fs :se:act wi:t :--e Eme ;er.:y Centrol Statics (1:3]Petc.. Mr.
i seel'ste =es te o.or:11 = arse :" L.-it 1. 7'.e !:ergesey Control Station. the unit 1 j' , l C=nt :1 tcxar anc t: assu t tut n!1 facats tf ne eme ;ercy plan we e foliomed. l l l Pr. *.ogan =as cmar;ed .o ass.re : at all t e required pr:cedures and plans mere l regimet ar.: t= lost :arn.sh eac: t: assure that every lies was covered this j factu:e: ce ; mee: urns f:r =ct' ecie ;e-cy plan and fer ce unit itself, and to prov'te w essa n.n:a : Sat alt ac:for:s mere teitg takes ard to be sure the nottft-catim calls we-o sede. at naragement was notiffed. and all ccumunications mere l ! to.g:aca. Pr. *se tegers was re:ueste to prerfde to:hnical assistance plus link-up with 31s home C fite as te C:ulC. P . Under ses in charge of technical su;Dort and c:zzt.ni:ati:ns an: V. !!cv113 has fr. c.arge cf emergency maintenance.
- 5ecause cf my trattisq. I felt a s:-ora; c:ligattan to t e puolic and to making sure tha: : ers mes et=taal release of -adt:ac.1vf ty and that there was evacuation f a r!anty 1
of size tr : net =as scutt,t. The om. the pressure. the fact :nat the planc was in i i a state ::.at med sever tem s=cc!ed is, coetned to make conditions almost 1 .sie-n.:~e. h:mener. . a C:r.tr.1 *a sos rematt.e4 cals as can te testiffed to sea-sta*y. 21: f ::e see:tr.;s f . e c:ss:ard taas were held in the shif: 5 oe-etsers s 0*" ice ta a cais ::cs: ere. at a poir.t reieved fr:n the *ontrol
;cas, a c ::e ce:ist:s. makts; .as . m :m:isely. at intervals dictated by tM
;!a*:. a-e *e es case lorge- : .st 2C -:: 4.0 minutes asart. .
(E&E TMI-2', Part II, pp. 253-256.) _ . . . . .. , _ e {
. i
. e e
Temneratura Data l. By observing temperature ~ levels at various locations in the primary cooling system, reactor operators are able to determine whether there are steam spaces in the reactor , vessel or in the hot-leg pipes leading from the pressu e vessel to the steam generators. Such temperature measurements also enable operators to know, in the event the system contains steam, whether the steam space has expanded to the point where it encompasses the active part of the reactor core. Temperature data can be interpreted as follows:
-- If the temperature is less than the terperature at which water boils for the prevailing pressure, the fuel rods are cooled primarily by water; cooling is adequate. {
This is called a "sub-saturated" condition.
-- If the te=perature equals the temperature at which water boils for the prevailing pressure', the fuel rods are being cooled either by water or a steam-water mixture; cooling is adequate. This is called a " saturated" condition, and can be converted to a "sub-saturated" condition by increasing the system pressure to the point at which the stean in the bubble condenses to water.
-- If the temperature is greater than the temperature at which water boils for the prevailing pressure, some or all of the fuel rods are b'ing e .
cooled by steam. --Steam at such temperatures is
< iia __.... . .. .
o -
2 g-Superheatad crnditions are indic tiva of inadequate cooling; such conditions are, at a min hum, a warning sign that insufficient cooling
'7 water is making contact with the fuel rods, and that temperatures. may be on the verge of rising to heights where the zirconium cladding interacts chemically with steam, producing zirconium oxide and hydroge.n.
Temperatures in excess of 1600 degrees are indicative of a cladding steam chemical reacti.on. A typical E&W reactor normally operates at a pressure of 2185 pounds per square inch (psi) and temperatures of about 604 degrees ?. This means. that the water temperature is approximately 45 degrees below the boiling point. In order for water to boil at this temperatur.e, the pressure would
~
have to drop frca 2185 psi to about 160'O psi. Bot-leg A principal indicator of system conditions at TMI was the temperature of fluid (water or steam) in the hot leg pipe,* which is the pipe that carries hot water from the reactor i l to tthe steam generator. Hot-leg temperature data at TMI was normally presented on a computer printout and on a strip chart recorder r.ounted in the reactor control room. The computer was progracmed to record data between 520 degrees F. and 620 i degrees F.; when temperatur5s were outside this range, the computer printed question marks. The strip chart was capable of-recording te=peratures up to 800 degrees F.
- See Figure I-A, I-B W _L______ **
- ~
* .ms FIGURE I-A - 9 a -- i,
*
- N
=< <Z .
e.=s
- *=4 .
2
= "O o 2" )
O ** "O M . 9
, 5" I< :
g a ._,.- 14 E30 E - o= >0 @ O
= u .
g
- a w
( - ga~ = g
. . . . -t ~ : .
/S
!E kf $ .' M* -
- : 3< ==3
-=I= l <3 = < . .
c: => .. x'** w .,. ,
. // I 3l =$ h ! E> * '
- 51 = < ==
.* O T
-s - h X ',, S',.*:. '* (( , ~
" 3 ,. a 3 g x . .
q== z$ f
>=
w I #,, > gwww
<O 3"
m3 j' u
< ~
ff *= = c-ll
~1 E _,1g4 .
% ,/
a' , pis M o m. o
$ < 3r I$s m '
.<g3 E
- 23 l U$
m n <a - - $ c eg 7' ) 9' 3.e
- m
'l ;; *u e- ow gg E
, y ~ e .
j f ug .
.g g
' t g ~: , . .2 .
9 r v g,d
- I am.-ene mmen.I h k e -
~
}/ , N ,
,0
-- ss< ; < y 5y
, i . j
. .&~'-.........,3...,.' m O<
Y - O I;ov ye , ,
'a
- r,___ -Q w n e.
. . . . _. I .
c- O E a o u Oa
) , m.s -_ _ --
~h at <E' W ')
" W
>- ./N .
o fr
't f-lo jJ G '-
)
o < Q {" = -
$$ M en le >. :: e 2 --e -_ . .
&s
- O3 3 5'E '
. 1-o -
.- H' ~
3 Es : E ge- --
..$j_,I.o =8 rr
(.. -5 i !' .>
- >o .El D jL(f m = %~ ' '.
8 -
.,T[
% 15 W'
. - - . t N
z E e .(2M. O!,- ,
$g-a3 $.9 u-5k.j- y
, _ i 1-f+ '
E Z z r >- ' s. l Q m s. n GB u.= N- -~ W . . . l
% I l
f ; ; c-- --- j ..-r,2_ _c n- - -- . :. = _ _q i i x ,, ., i In $ ____ . 1".g'f i '. r T . r
- I, I -
l t] 1 1 1 1 - io i 2 l < . g na .g g_ z- *
- E BI T I EE
>O l
Q m lE CE "O L i l "O I i
> D
- = 1 3-a
__ i 4
- e. "ss!>s"
=~i -
=- .
i i < I '++" l k! l :E Ig ci 1 Er c . E5=1 I ws 3 3 I
,8 p. I I
- a g ,
I 1 e ( L___________a g-g,
*" 4 0 2
. MEh m _ _ . _ _ - e ^
. - 9b -
FIGURE I-9 i 7 4 A C3 o e- :Q , 3E Ee $l if + . yg . R* yy ] ,%, pumwrwunyg,. , _-r. ,
-g W i
.- k j
? CD ;
g5 = v
- c. F Q.
- - > O
. a=
+ s';
... . 7m W o
' -=uitT J
,5+* * **a& ??h$
l e
* #' " $ #, $ % % + BWQQ~_mM 4
- Y$.&S$NWhR$5'fj$$$y>, .
e th w - . C 4 * # ,h .e.*.,..,e..,-
- e# #I#=%# p a,
g t- No u e
? # =
*a o hi C I,'
O ,- J. ar Y
"
- g aJ k ++_v .-
d- '
) '-l < *+g* a an, * , *% ; $ '
.C C l e
- %+ + a
@. EI4 3$ l O C o l N'
,*k %". +p% $ Y$
.G o C m
- l s c C) n % ,, E l *Y:*m. ~e A 4
O o y
, =, -4 4**+ Ogu ,
a r
- >* 1.
" Ou i gbc 3' _,_-
,r
#v *x *' -
,, o. o j Xt N c . .B m g m l C
3o gE e c , m
.C E
.=::.oo _o -- o - 5 u _s
.wew .; .._
; ci *_ eC WO '
s 'm' - * ~O E-y '
'I
- w :n .y . r.,; O j
- g. .-g, L u _J L g.;;
*erl 9,, 7 .
1
% eC T: u o j
? < 8% -
M . y O O M wi' Q C b'i P%r4Eh y
+1 o &
-ox w >
< ~**, W.' ? '<s?::: G s:,?sw e m m s s. g-
' C
.o C l 4 .
O Y ajO O" ~ r s c' '
- asa
~
O yg _o **# 4 4*,. s 8.e 3 # air *' . y 1
$= 0 3*, e W2C
- Onw J 3 :v.a I -
99,R
~.e m
x; die t ; l -- F+[? s *, 5, s ,t . - _', s s,s.,,. o _ :,,. -fe; ; .u *~. i .. Mf$.
--n
>s+s,*ss
> 6 .:m' .~ w. 2 __
- * ; . z o r ,.:.r %:. .? e ..
q9 2 p h
I ,
- 10 -
Apparently owing to the computer being incapable-of indicating temperatures in excess of 620* F. , Station Manager Gary Miller, soon after he arrived at about 7:05 a.m. , directed that a meter able to indicate the full range of hot-leg temperature be set up in the control room: When I turned to focus on plant conditions, an initial concern was that the hot-leg indication was off-scale. I asked that an extended scale readout device be connected to the hot leg RTD (resistance thermal device) . 297.) (E&E, TMI-2, part 2, p. .
~
I Miller stated that:
... the extended hot leg temperature readout device indicated 720 *F. " ( Daid. )
exceeding Any terperature A 660*F at the then prevailing pressures (which ranged up to 2200 psi) clearly indicated the presence of superheated steam and/or hydrogen; i.e. that the core had been uncovered.
- The readings from the hot leg instrument are also referred to by Ivan Porter, TMI-2 Instrument and Control engineer, and by John Flint, an engineer employed by Babcock and b'ilcox who was stationed at TMI. Forter had been responsible for setting up the instrument, and he told NRC investigators that after intially questioning the validity of the readings he had checked some of the temperature readings taken from instruments within the core (see below) and that:
to me it confirmed that,what I was seeing on the RTD. That we had temperatures greater than 700 degrees in
/
the plant, since 700*' degrees was full scale on the computer and I was reading greater than 700 on the hot leg RTD." (I&E, Tape 237, p. 15.)
*This is in reference to the in-core thermocouples discussed belov The computer could indicate the in-core temperatures up to 700*
whereas it could indicate the hot-leg temperatures only to 620*.
------,s
_ 11 _ John Flint recalled that at this time (sometime after 9 :00 a.m. )
... Ivan Porter showed me a special setup with a RTD, that was approximately 724 degrees." (I&E, Tape'323,
- p. 34.) ,
Flint also Icted that be himself had manitored the strip chart recorder in the back of the reorr Wich was rmding the data Flint refers to in the prM%g qtote. ' Ibis chart shcus hat-leg tenperatures indicating superheated l a=ditions in the pMuary systen which in turn indicated sections of the core i had accumulated l had been uncavered and/or non-condensible gases (probably hydrogenhin the upper sections of the hot leg for a substantial portion of the period between 5:40 a.m. and 7:50 p.m.t the latter being the time at which a relatively stable cooling arrangement was 1 f' established. The strip chart shows that the hot leg temperatureb rose rapidly from the temperature at which water boiled for , the prevailing pressure (i . e . saturation temperature) to superheated temperatures following shutdown of the main reactor coolant pump at approximately 5:40 a.m. In-core Thermoconoles I i Other sources of temperature data were the 52 thermocouples installed inside the reactor pressure vessel above the fuel assemblies. These thermocouples sensed water (or steam) temperatures at particular points within the core. Each thermocouple provided an indication of the temperature above l conditions A a particular fuel assembly while the hot-leg i temperature sensing devices' (discussed above) in'dicated the { l I
- Scc Figure II. .
j%. . l .
~__ -- n. , 4:00:37 .G:00:37 8:00:37 10:00:37
- 1.0 0 1 2 3 4 5 6 -
--d
~~ ~~ * - -
Reactor coolant RC P-1 A ' Ptimp operation
~~ ~~~ ~ ' - ' ~ " -~~ -
l
' RC P 2A Donotes on RC-P 18
~ ~ " ~ " ~'" " ' '
~ ~ ~ ~ ' ' ~ ~ ~ ~ ~ ' ~ ~ ~ ~ ~~
(referenco 3)(nolo 7) RC P 20 - - - . - - . . . . . . . _ .
,"j ..~
00 ... Reactor ci.nlant nyntom flow
,E 60 _, ,,_,2%
C Loop A L ao j-- Loop D 0 Loop D d l/ = 5 Sourcri nnr1 informorfinto E , I[ - Gourco rnn00
- inngo inntrinnontation ,
mh ta 10 4 in core h{ g [ ,probablecoreconfigurationchange,g 3 g _ 10 ,,)
' (Sh _- _
_g M S 102 Informodlato ran00 -
$ 000 Reactor coolant system E Loop B - g outlot temperattsro L8- $ 0 750 \ r y
' --~
Loop A - 5 e 8 m 600 \ . W - ( Loop B - Loop A ) 3 j 450 , core uncovery beu res apparent . l 300 : .,
$ 600 Probable core configuration change 7 Ronctor coolant system y c A -'
I I inlet temperature {[ 450
,g9 aeg Loop E 3 e e m 300 .
oe
' S Loop B R Si. .
- - - _ - ? - _'* - - _--- - _ - _ _ _ _ - - l M
e 12 ,, average te.mperature of water .(or steam) leaving the reactor pressure vessel. There were, in fact, large differences between average temperatures as measured in the hot-legs and the , peak temperatures neasure,d by the individual in-core thermo-in-core couples. It was these pea g temperatures that confirmed not only O@ O e G 9 m e e O O 4 9 e * , - . -. -+--- , , ,, ,-- * - , -. ,
.- l that the core was. . uncovered, but that the zirconium cladding was reacting with steam, producing hydrogen and zirconium oxide.
Credibility of in-core thermocouole data. - The DiI personnel showed an early and continuing interest
~
in the in-core temperatures. They instructed the computer to print these temperatures at about 8 : 34 a.m. , 8:47 a.m. 11:10 a.m., 12:40 p.m., 4:11 p.m., 6:30 p.m., 7:59 p.m., 8:56 p.m., and 9: 56 p.m. Between .6:00 a.m. and 6:00 p.m. the temperature data for at least 40% of the thermocouples (EPRI, Fig. cl-ll) were printed as questionmarks, indicating l l either that the thercocouples had failad or that the temperatures 1 were in excess of 700* F. Because some of the thermocouples alternated between indicating question marks and temperatures less than -700 degrees, it was more plausible that the readings were indicative of temperatures in excess of 700 degrees indicativo of thangi nstr' - nt malfunction.* Moreover, if many had been danaged while others continued to function, this in itself would have been a reasonable indication that something major had happered in the core resulting in disabling of the thermocouples. The following excerpts from interviews conducted by DiI investigators indicate that TMI personnel did in fact believe the in-core thermocouples were providing useful informatici
*By approxinately 12:45. a.m. on March '29 the thermocouples were rostly indicating temperatures less than 700 degrees and were apparently used thereafter as a prime indicator of core conditions.
l 9
. ~ - --. . - . . . - - . - _ _ . - _ . , . . . , - . . _ _ _ - _ . . . _ . . , _ _ _ . _ , . . , _ _ , __,
/ .
. le -
John Flint, Babcock and Wilcox engineer stationed at TMI told General Public Utilities investigators in an interview held on April 20:
... shortly after I came in we also started cat Hng up on the computer the incore thermocouples attempting to establish what had happened in the core. Many of them were indicating questionmarks which indicated that they were
. greater than their 700F range. Only one or two seened to indicate that they were in fact had. These temperatures were monitoried for the rest of the day to follow what was happening to the core."
Ivan Porter told IGC invest.1gators:. I believe shortly after 7, he (station manager, Gary Miller) asked me about the readings on the in-core tem;:erature detectors, and I punched out several of them (i.e. asked for computer printouts)..." (I&E Tape 237,
- p. 16) .
j Porter said he reported the results (i . e . temperatures in excess of 700* F.) to Miller who asked Porter whether there was any way of reasuring the thermocouple voltages in order l 1 to deterrine how much in excess of 700 degrees the temperatures l l e l 1 e
I V
- might be. Porter told Miller that he thought he could get the information by connecting a digital voltmeter directly l to the wires leading from the reactor core and measuring ,
(and therefore the' temperatures) . thereby the voltageshbeing fed into the computer. At this
. point there are varying recollections as to exactly who it was that Porter told to make the measurements. In any event, four technicians in addition'to Porter either participated in making the measurements, or observed as they were being made. One technician who the NRC investigators have designated Instrumen' Man B (IM3) in deference to his wish for anonymity, stated that:
... 'Ivo of the thermocouples, the first few we had measured, were around 700 to 800 degrees, specific temperatures I can't quote you, I don't know. We
- had thken one off and we had measured 2600 degrees in and abq that vicinity, it was very close to that. At that time neither one of us believed that this was a true reading because after seeing two, three that were 700, 800 degrees, g 2600 was h'ard to believe so we decided to take a few more off. ... I believe in the course of testing thermocouples, we had at least 10 or 12 we had disconnected f l total. We had seen temperatures ranging anywhere from ;
I know there was a botton of about 690 degrees to uppers l of 3700 to 4000 degrees. (I.E. Tape 315, p. 11)' I l While cladding temperatures probably did exceed 3500', there l is disagree: erit as to whether tenperatures as high as 4000 degrees were actually measured since it is unclear that the thermocouples would function at such temperatures. There is general agreement, however, that at least five people (Porter, Maintenance Foremen Bennet and Gilbert, Instrument Man Thomas Wright, and Instrument Man B) were directly aware of the instrutent. readings indicating i i
. , - , . . . , . . _ , . . . . . - .., L--. ,__:.-..-.. '
- temperatures in excess of 2000 degrees; i.e. temperatures at wxild be which thereA , significant production of hydrogen. Porter ,
questioned the validity of the measurements and he told the technicians to repeat them using another meter. The technicians
. did so and this tire they ti:>ok measurements of 51 of the 52 thermocouples. Nine of the 51 were in excess of 2000 degrees.
Where there was overlap, the second set of measurements, taken shortly after the first; confired the first set. DB told the 1 RC investigate)
"Now the second set of readings dd.d in fact correlate the original set of readings that was tak'dri Qith th5 the raccouple reader. The general feeling at least
. amongst Jin Qiright) and myself was that the readings we had were true and accurate. All five of us that were present there did in fact visually see the actual readings we had taken both off the thermocouple reader
+ and nillivot reader. All five of us did in fact verify that the millivolt reading through the conversion table
~
was correct. So I an sure we didn't read the wrong table or the vror.g .one or something of that nature. (I.E. Tape 315, p. 18) Bennet, Gilbert, Wright and DiB have stated with varying degrees of certitude their conclusion that the data indicated the core'was uncovered. IMB told the investigators:
... it was the general consensus amongst the instrument people there that the core was definitely uncovered, we . -
kind of found it hard to believe that this many high terperatures that we had seen that all those incores would have been had and the only way that they could have went- bad that radically would be an uncovering of the core, and super overheating. (Ibid. , p. 17) Bennet: We had possibly uncovdred the core was the only way we could see that you could have obtained temperatures of that magnitude." (I.E. Tape 311, p. 18)
b_, / , fr L= o r= e- ... [- k 6 Wright: 3" "I feel then that there was a definite sign then that [~ the core had definitely been uncovered to the point, . a where it suffered.danage. But it, I still say that,
- you know, I'm there to take the data. I'm not there to E-- analyze it. S ; I gave them my personal opinion as in the, 52 yeah, I do believe we did suffer some damage there.
g; (I.E. Tape 310, p. 14) & ~ E Ivan Porter was responsible for reporting the thermocouple ?+ measurements to Gary Miller. In his statements to the
- =
"M kt NRC investigatbrs Porter is ambiguous with regard to what & he believed the significance of the in-core measurements 37_ l5- )E to be. When asked whether he believed a reading of 2300 l [ _ degrees to be anomalous, he stated: E [F "I- don ' t know. I gues's I was afraid it was real." ?? . (I.I. Tape 237, p. 20) I$, e . $5 - But Porter also suggested that the measurements were not 5(p 9D believeable. He said that in reporting to Gary Miller that W z es Miller had asked him what he (Porter) thought the thermocouple C k$ measurements had meant. Porter told the NRC investigators j that he was not sure of what his reponse to Miller had been, Dd but that: Et
" . . .my personal evaluation was that they (the thermocouples) h;
,r- had been destroyed." (Ibid. , p. 19) ra rr 5f IM3 told the investigators, however, that he had told Porter G (IMB E_ that heAb el)ieved the temperatures readings indicated that the em
- Ky Core Was uncovered:
h h.- 4 E - [r: m: _h
=
. .. I personally told his that and he was physically there to read the readings. He saw the actual temperatures we had. This is why the first time he didn't believe it." (I.E. Tape 315, p. 18)
IMB also said:
"I believe Ivan (Porter) didn't really want to believe
_ _ what was really taking place. I don't know whether it was an attitude of " bey, your measurements are wrong, you guys don't knew what the heck your doing or whatnot. " I think the general consensus throughout the whole first day was neher one nobody really knew what was actually happening, number two, some that had an r inkling of what was happening didn't really want to believe what was going on. ONce you start seeing a temperature of 30C0 to 4000 degrees in a core, well ... the first thing that starts coming to mind, you've got g.5 a meltdown coming. The core is uncovered." (Ibid., p. 18-1g
+
1 Whatever it was that Perter actually reported to station ; Miller told
$.- manager Gary Miller,/g NFC investigators: .. . _ :j h' So you know, the hottom line here was that they (the in-cores) are hot, they were hot enough that they scared g# you, as far as what you're looking for. It told me the i
h' M
~
reason the coIrputer was of f scale. at 700 degrees. . . The in-ccres were reading anywhere from 2500 or so, and ii?? I picked 2500. It could have been higher than that. M is&- But that ycu knew, I was looking for a gross indicator and I had it. ... I know that we were superheated and all g -tv - that scrt cf thing, I don't think we tumbled to that kind of locge (sic, logic?) but we just know we didn't have a centrol, we were out of control. We knew the situatie was one we hadn' t anticipated too many times here. " (I.I. Tape 159, p. 51-52) None of the WI ir.vestigators' interviews has indi'cated
; whether Gary Miller (who as indicated above had been informed of the first set of measure =ents) saw the data obtained from the second set of readings. Olis data when plotted on a map of the coq showed hot regicns witHn the core and cold regions on the
_ periphery.* Instr-nt Man Thc=as Wright who was involved in
- recording the data said that the data was supposed to have been turn i
- See- Figure III.
zu
- f
~
FIGURE III
'vi
.Td
~
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 .!),i a W,', A-
' A 5:
'QfM-
=-..'- 4:@
B 281 325 B :-
- - {,bhi; C . 469 957 325 -C -
. l@- ."..s D 1196 500 217 'D
,.~;._'
E 599 1926 2580 326 E' - 2366 2378 323 555 F F 80 G 375 2272 1774 1806 1875 234 G H 260 2452 1295 2176 1852 H 632 1760 K K 1811 373 1566 332 296 1774 L L 325 2171 2327 348 252 M'
!.i 413 578 2167 N N ,
O 356 462 1138 308 O, I 291 P P 1 352 475 R R 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Figure Cl-12. Map of Core Exit Temperature (*F) 240-330 min. l e
_ 19 - j _ over to Ivan Porter. ..Ee stated to the NRC (I.E. Tapef310,p.18-19) ({ that Porter was aware of the data. Wright recalls also that Porter had ramarked that-it appeared to him that it "... looked 4-like we cane, you know, that looked like they'd been damaged." (-- It is not clear from this whether Porter was saying that the JE core or the thermocouples had been damaged. It is also unclear L-
,N .
whether Porter actually received sheets on which the data was written; and Miich want unnoticed,/it seems, from . G_ March 28 until May 7. . l
? - $ ,r Mike Ross, a TMI-I superintendant, who was a member f
L of the group making decisions on March 28 engaged in the y following dialogue with NRC investigators.
@ Hunter (NRC investigator) :
C [4 g
"Okay. Do you recall in the discussion, think tank discussions, that the thermocouple temperatures g were, in fact, brought up in the discussions?"
gf.
% Ross:
usy
% " Thermocouple temperatures were brought up to !$ Gary Miller, and I guess the bottom line they psi got out of that, was that they were not conclusive.
It showed the core was hot, basically. I was g- going to say his range varied, very scattered. 4t He had like ... He was saying he had various 5 terneratures scattered throughout. So, well, g - Gary and he discussed it, and basically I think I the botton line was yeah, the core is hot, or it l R_~ is at least hot." C - 4 (I&E TMI Tape #226) 6 A A " Y i 1 i . D P . _ __t_______________._
_ _ - - _ _ _ _ _.m. l That the in-core thermocouples had indicated temperatures in excess of 2000 degrees (implying hydrogen production) was apparently not reported to the NRC during the first few days i NRC l of the accident. Victor Stello (thengDirector of Operating
. Harold Reactors under4D enton) said he had not knowhof such measurements until the week of April 1 . Roger Mattson (Director of Systems Safety under Harold Denton) stated to the E&E Task Force on May 9 that this was the first he had heard of such j _ temperatures.
R b - C . m i g m M \ W i : l l . ?,5$ ' I _ __ _ _ _ - - -
e-
%n ,. ? ...
E e t d-E^ Neutron Detectors Y h T}iI-2 instrumentation included neutron detectors mounted outside 2 L the reactor pressure vessel and 52 strings of detectors if mounted inside. Both in-core and ex-core detectors are h di . installed for the purpose of providing operational data; l- [ during normal operations, the in-core neutron dhb.,rs indicated f U-power;noduction at various locations within the core. t$ - During the TMI accident both in-core and ex-core detectors ut MU provided data indicating the core was uncovered. g
/Bi v=
The ex-core instruments indicated increased neutron levels. 9 Ti These were interpreted initially as indicating that the reactor s NI was near the critical point where a self-sustaining chain H-as Gy reaction night be occu ring. The operators believed 4F - l T that this might happen as a result of insufficient boron in ff concentration in the primary system coolant. In actuality the apparent increase in neutron activity resulted from the mz fy; fact' that neutrons -- nornally produced in a reactor core Im
$N even when the reactor is not running -- were, as a result h$ of voids in the core, less likely to be obsorbed prior to 4 ~
3[ their escaping the pressure vessel and were therefore more Q A' likely to be detected by the ex-core instruments. In short, 23 f; the increase in the number of neutrons was due to loss of i-water in the pressure vessel and not to a restar_t of the 5 . . y chain reaction. E'
} --
w g . I v N b- I
O
'lhe in-core detectors also yie.1 ed data indicating that the core was uncovered and the depth of the uncovering. Once the water level went below the neutron detector and the temperature of its surroundings rose, the neutron detectors responded to the high temperatures and became, in effect, temperature measuring devices which indicated water level in the core. (See Figure IV.)
The record is unclear as to how the TMI supervisors I interpreted either the ex-core or in-core neutron detector data, other than inferring during the first few hours i of the accident that they might be indicating that the reactor was going critical.
~
High Radiation Levels in Containment At 6:35 a.m. a radiation monitor mounted at the top , of the containment building indicated radiation levels of about 0.1 Roentgen (R) per hour. By 7:30 a.m. this monitor was indicating levels in excess of 10,000 R per hour.
$ (Approximately 50% of persons exposed' to this dose rate would receive a lethal dose in about 4 minutes.) These radiation levels ir lied that radioactive fission product gases had been released from a significant fraction of the fuel rods; the only way such a release could have occurred would have been through development of cracks or perforations i
) _ 9 ! a Y ____ __
F l - 22a , FIGURE IV Reactor Vessel .
. A A ll ; ; ;; ; ; ;;
Plenum ' jg j( assembly '
~
.. 4 A A P >
in-core &
,f [thermocouples * .
_g ;
,_._--__._________..). O (above 52 fuel assemblies -
l Dov.r. comer Il
- Fuel assemb! es
\ -
(177 total) g Source (l Core - range [-' I Y-, ex-core neutron _ ~ g _ detector (2)
- .g'~
Self powered
/ neutron detectors e '
O (in 52 fuel assemblies - for total of 364) }. g H Y _ , _ . . _ . . . _ _ - . - . - - - - - - - Lower grid assembly
's,~
~~ __-
l
-u
; To cable spreading room t
Basic reactor configuration and instrument locations.
;2OE l _ _ _ _ _ _ _ _
~
.i
~
i 1 in the cladding.
't Based on this reading alone, James Floyd, a TMI supervisor inferred that approximately 1/8 of the fuel rods l
, had failed. On the day of the accident, Floyd was in Lynchburg, Virginia for training on the B&W reactor simulator.
l - Having been told of the accident by a colleague; Floyd called the control roon at around l l 7:30 a.m. and was given information on radiation levels from which he inferred that the quantity of radioactive gas in l. n m the containment building was approximately equivalent to 1/8 of that that had been contained in the fuel rods; l i . I this led to his inference that 1/8 of the fuel had been damaged. b- - Floyd~ told the Kemeny Commission that the inference he L i 9 made in Lynchburg could have also been made by persons in e.1 the TMI control room on the day of the accident. (Kemeny Hear.ing, May 31, 1979, p. 186-188.) Kemeny Commission member 1 1 Ted Taylor asked: ll "Well, to whon did you give this information (i.e. Floyd's inference '... that it looked like at least an. eighth of the cladding had failed') during the course of the day on Wednesday, outside the gGi ~ people you are irmediately involved with at Lynchburg?" Floyd responded: pg- "No one." . Taylor:
"No one. Were you assuming that people at the plant a%.- e knew this?"
yE m
7 _ %:
- 24 .
Floyd: l "Yes sir." (Ibid., p. 188-189) Floyd later indicated .that it was not until he sat in on the Kenemy hearings in lat:.May that he realized that the supervisors and management at TMI had said that on March 23
~
they were unaware of a substantial amount of core damage. (Ibid., p. 194)
,. e
\
4 i O
- J Hydrogen Conbustion .
During the morning and early afternoon on March 28, a significant portion of the hydrogen produced in the zirconium f} cladding steam reaction w;is released from the primary cooling system into the containment building via the pressurizer relief ' valve. At approrimtely 1: 50 p.m. the hydrogen ignited. (1) What is probably more accurately described as a fire than an explosion caused seseral effects including a 28 pound per square inch'(psi) pressure pulse in the containment building,*
- s. This pulse was recorded on the strip chart that recorded 4r containment building pressure. It was also recorded on a
)['7 series of other pressure measuring devices which used the
? containment pressure as a reference. The fire raised r temperatures in the containment from about 125 degrees to
{c 175 degreesF,an increase of 50 degrees dissimilar to any other l
~ -
ten erature cranges observed that day; these tenperature data were recorded
- j. on strip charts in the control room *.**The increased temperatures
- ( w
<C I
triggered several alarms. The containment pressure pulse i g also actuated various emergency systems, most notably the I[ equipn.yat that caused water and sodism hydroxide to be sprayed l into the containment building. t TMI supervisors and operators have given tuo explanations for not having reported the hydrogen fire to either Federcl or state officials:
.,
- Figure V-A
' ~
** Figure V-B
*** Figure V-C t
(1I See f.n. p. 4, regarding this not being the potential fire and/or explosien feared on March 30 and subsequent days. l ,m \u
[g ,. . 1 . 32 ' ' j 20 - - l l-lydronon burn
.3 24 -
i '
$ 20 -
I e et in -- 4 ca
. 12 -
, d a 3 , ' U 8 - N n
< tn
, E " j 4 - $* l )' [ \ O g
-4 ! ! ! l l l 0 2 4 6 0 10 12 14:
! Time AfterTurbine Trip (hours) ( Reactor Building Prossuro Versus Time 4 I
25b ,
. . . -s .
FIGUREN-B 270 -5 0
- OTSG "A" 265 -
- OTSG "B" -
o
)
i 260 - 5 0 o O 255 - ! o 10 O
$ 0 O
a 250 - I c - 15 h ; b ! /
~
O o
" s' I
- 4 $
_ O l O 245 - 0 l - 20 1 0oB 240 - 25 2.35 - 30 y . l l l l l
' 13:50:20 13:50:30 13:50:40 13:50:50 13:51:00 1351:10 13:51:20
, , Eme N. , Effect of H2ignition on measured steam pressures. i O
,_.-.-.,...--..e , -
--- FIGURE N-C.
i- l!
~
g !
]
I mvvr? 08888 a n a c> i l 18888 e . Es?%%% BBB 8 z Effff ?= 88888 - 33333 j C 55555 e:ne 1
.c
?
l
- 8, I u
r E y I
.3 9 8 ( t i E i 2 E.
O hO ? 5 58
= ag 8 t 3-5 se S jo
/e S 333_883.
a 8
==g222 m/ "
w m =
~ tig2ms -
~s g l
easit<i u) u) < u< o 1 mes e c) l l
! 1 I I I I I I I l I l
8 8 R ---- 8 8 ? 8 I 8 e8
- - 8 8 R- ,
(3.) eJnle;edwal R, [ .
- i. :
,,, - 2'7-
- 1. The explosion manifestations were caused by spurious electrical signals.
- 2. Because the pressure did not remain at a high level, the cause of the pulse was not important, and the
) operators were too engrossed with establishing a stable cooling configuration to pay attention to it.
, Neither of these explanations is plausible. The first is questionable because of plant design considerations and the simultaneous occurrance of pressure and temperature effects indicative of a real pressure pulse. Given their basic training in reactor engineering it is implausible that the TMI supervisors would consider the pressure pulse to be insignificant in light l
i__ of the multiple indications of core uncovery and hydrogen production. . e The record contains conflicting testimony. Excerpts
.a follow. The~ first concerns whether the electrical malfunctions could have caused the pressure pulse. TMI-2 Reactor Operators gg; I engaged in the following dialogue with EdFrederickandCraigFaustg the E&E TMI Task Force:
Myers: What could have had two of those three sensing devices sense high pressure and lead them to think the pressure was high, other than high pressure, l real high pressure, actual high pressure? Is there anything that decides the pressure is going up that could have led those sensors to think the pressure was high? 1 Frederick: A test signal. l l Myers: Would a test signal go to two of them at the ( same time? - Faust: No. In fact, you would have to hook it up. Frederick: It'would have to be a lengthy manual action to ' get it to do it, other than actual bui ld ing pressure,
? .
I
- ~
t -
. 1
~
. - . -,, - 28 :- t Myers: Is there anything.that you can think of other than excess building pressure that could have simultaneously led the meter to read 28 pounds per square inch and to turn on the spray tanks or turn on the containment spray?
Frederick: No; it had to be high level pressure. Faust: There had to be a pressure surge in the building for it to happen. (E&E TMI Part 1, p. 147) Instrument Engineer Ivan Porter was asked by NRC investigators whether Porter thought the pressure spike could be explained by any form of instrument malfunction. Porter responded:
"I would think not. It did look like a real spike to me. That was when I was specifically asked if it could be real. " (I&E Tape 237, p. 35)
With regard to whether he saw the pressure pulsa on March 28, Porter said: ,
~ "But somehow I have a feeling, I didn't look at those charts until the next day. I'm not specifically sure that I was aware of it that day. I know that I very specifically remember a discussion where we looked at the chart, was asked if it could possibly be real, and I also look at the wide ranged pressure chart and l saw that ['he t pulse] showed up as a decrease in pressure on the wide range reactor coolant system pressure, but
, I sincerely believe that~that was the next day that l I looked through the stuff." (Ibid., p. ~34) ! ~~ l It is not clear why they would be waiting until the next day to look at the data, or if it were the next day, what it was that led them to do it then. In addition to the spike appearing to be real to Porter, he found a similar negative j spike, in the reactor coolant system pressure history. This would be expected since reactor coolant system pressure
~
-1 , d was measured using the containment building pressure as a reference; when the containment pressure went up, the reactor coolant system pressure would appear to go down. As noted above, a similar negative pressure pulse was observed in other pressure hi~ stories.
- Shift Supervisor William Zewe said that he had observed the pressure spike to occur at the moment the pressurizer relief valve was cycled, but that he did not associate the spike with a hydrogen expl'osion. He said that he had not been aware of temperatures in the core of sufficient magnitude'for the cladding-steam chemical reaction to have occurred. Zewe was apparently unaware of the in-core theritocouple Ibeasurements that implied production of hydrogen. Zewe said, not knowing what the spike could have that been,Athey ascribed it to an electical malfunction associated
- with the operation of the valve. While Zewe may have considered z- have been the pressure pulse togan electrical anomoly, others seemed
*See Figure IV-B.
4 9 4 0
- - . -- . ,- , - - - - , ,- , - , , , , - - - - , - - - e ---,-e- e--, -- e,,-,. -,.e--,,.-n---- ------r-_---.,rne,.~.- , - , - -
was measured us-ing the containment building pressure as a reference; when the conhinment pressure went up, the reactor coolant system pressure would appear to go down. As noted above, a similar negative pressure pulse was obseryc1 in other pressure histories.
- Shift Supervisor William Zewe said that he had observed the pressure spike to occur at the moment the pressurizer relief val.re was cycled, but that he did not associate the spike with a hydrogen explosion. He said that he had not been aware of tecperatures in the core of sufficient i
i ,. magnitude fer the cladding-steam chemical reaction to have occurred. Zewe was apparently unaware of the in-core t.h erroco:ple I$easurements that implied production of hydrcgen. Zewe said, not knowing what the spike could have that been,Ath'ey ascribed it to an electical malfunction associated with the operation of the valve. While Zewe may have considered have been the pressure p,ulse togan electrical anoroly, others seemed
*See Figure IV-B.
i 9 4 - ..s- ,
- ,. -- -- ,. ,- ,,.,,n.-,, . - - . . - . , , . . , , .
- . - ...--,,--,-..,-.-nn- , , - , - - - - , . , . , . , - . - - - - - _ , , , . . . - , - , . , .
i g.. ..~ Y u concerned that it was real. TMI Supervisor Joseph Chwastyk said:
" I
... I actually saw the recorder, the pressure recorder I on the building, spiked upward. I didn't know what caused it-but the fact that the spray valves started indicated to me that we actually had some kind of pressure spike, either on the sensors or in the building itself. I was not sure. The spike of course
, started all the building spray pumps, decay heat pumps, l etc. The pressure spiked up and it was only up briefly, as a matter of fact, a couple of heart beats.
I know because I missed those heart beats. It came back down again." (I&E Tape 232/233, p. 9) 1
*Chwastyk (who arrived in the control room between
{ J 11:00 a.m. and 12 noon, and who was apparently not aware of the direct reasurements of the in-core temperatures) also referred to an explosion in describing how it was that he came to realize that the reactor core might have been significantly dacaged:
"It was like I said, everybody wa's pretty busy and-I didn't vant to stop anybody from what they were doing so ..I just tried to get a feel for what was happening by looking around and asking the operators j ! at the panel what they were doing. Up until the t
time or sometime after the explosion and it dawned on I ne what it was, I didn't know how much damage we had." (Ibid. , p. 18) (underline added) e 9 9 - 1__ _ __ _ _ _ . _ _ _ m
i,
~~~ -
. Brian Mehler, a TMI shift supervisor, has said (NYT,-
10/21/79, p. 1) that during the afternoon of March 28, he was told by his supervisors not to turn on oil pumps in the containment. (This instruction to Mehler was presumably based er on a concern that operation of the pumps might cause sparks b - that would ignite hydrogen in the containment.) Mehler F-g responded that he had already done so and someone then allegedly said: (& W %s "Well, that means we don't have any more hydrogen in there. " M Theodore l k, ,_ A reactor cperator,/Illjes, who arrived in the control rcom after the detonation, told NRC investigators that he h has been briefed on the reactor building pressure pulse: D "I was told they had a spike on both indications of the N reactor building pressure recorder. There was some 1 _ discu~ssion as to what it was. A hydrogen explosion was '
= discussed. This was later in the' evening." (I&E,
,3 Tape 261, p. 6) When asked again as to whether the discussion of a possible
]
hydrogc2. burn had taken place on the first evening (March 28) Illjes said:
+
l "As far as I know that possibility was discussed that evening. " (Ibid. , p. 10) [ At about 2:30 p.m., some 40 minutes after the hydrogen detonation, Station Manager Gary Miller left the site for l ( the purpose of briefing Lieutenant Governor Scranton. (E&E, l l TMI Part II, p. 237.) Miller has written.that while in the control roon on March 28:
"I heard a noise at approximately 1:50 p.m., however, I did not associate it with the burning of hydrogen
' ]
i . O i or that actuation of the safeguards system at that time. I was first aware of the recorded pressure pulse and associated actuation of the safeguards system on Friday morning, March 30, 1979." . - (Ibid., p. 298) Others who were present -say that Miller was aware of the pressure pulse at about the time it occurred. When Reactor Operator Frederick. was asked whether others in the control room had reacted to the pressure spike, Frederick stated:
"I think Mr. Marshall tried to figure it out, and Gary Miller was particularly interested in it."
(E&E, Part I, p. 145) Mike Ross,'who was TMI-I Operations Supervisor, but acting as second in command to Miller'at TMI-2 on March 28 stated when asked whether he was present when the pressure spike occurred: -
-. 13, ,
"Yes. I was near the console at that time and if we are talking about the sane time was around 2:00, sometime in the area. And at that time we got an ES signal and some of the components restarted, decay heat, what have you. We got building isolation again and we took l
care of that and we looked back and the control room l operator said "Jeese the spray pumps are running." and l we looked back at the charts at that time. We saw a fairly ' large spike on the chart and the exact pressure at this time I, don't know, ... it was around 30 pounds. My thought at the time and Miller was out there with us and he questioned he said, "jeese you know I thought I heard something, too." We are moving down the road there 100 miles an hour and we looked at it and we said "Jeese the spike was so short it must have been an instrument." That was our reasoning at the time. We reached over and we said you can shut the spray pumps off now because the pressure came right back to 0, ... almost very, very rapid return and we shut the spray pumps off. I now know that spray pumps were on about five minut.es when looking back because I did look back on that particular one. I personally didn't associate it at the moment with j
a i 1 any-kind of explosion in.the building. .I-associated it Vitt W 'iristrument problem perhaps ana u. tu.um w did Mil)ir at the time because we Just went on to something else. It wasn't until the next day that we
.. thought about anything like that and started looking
. back. (ISE Tape 4226, p.' 4-5) [ Underline added.]
Shift Supervisor Chawstyk who had observed the pressure spike but did not hear the noise referred to by Gary Miller told the NRC investigators of a suggestion made to Miller on the basis of the spike having occurred upon operation of the pressurizer relief valve: Chwastyk: No, I did not hear the noise. But that was the point at which I had assumed that we did have some kind of explosion in the building. And that is when I suggested to Gary Miller we no longer cycle the electro-magnetic relief valve because it had . . . the explosion . . . or rapid rising pressure in the reactor building corresponded to opening the electromagnetic relief valve'. (I&E Tape, #232/233, p. 18.) - Mr. Herran Dieckamp, President of General Public C.'.ilities, stated in a May 9, 1979, mailgram to Chairman Udall:
"There is no evidence that anyone interpreted the
' pressure spike' and the spray initiation in terms of reactor core damage at the time of the spike nor that anyone withheld any information."
l
. 1 . .
. l Presence of Eydsicien 'in the Prim [y Cooling System -
Eydrogen produced as a byproduct of the oxidation of zirconium cladding collected at high points in the primry cooling system; e.g. at the tops of the reactor pressure vessel, pressurizer, E-d upper (candy cane portion) of the _ hat legs (SeeFigure I-A, I-B; .Pages 10a,10b.) Wd.le the hydrogen in the . pressurizer could be released from the system via the pressurizer relief valve, there was no such valve in the other i components; hydrogen could be elininated from these only either by forcing it to dissolve in the water, and thenfallowing it to escape from the water in the pressurizer wherefrom it could be let into the' containment or extracting it from water removed via the letdown system. Eydrogen and/or superheated stean in the cam
/ cane partion of the hot-leg had for nest of the day obstruct d E- W 7 system with water, thereby preventing use of ~
the . generators for removal of heat produced in the \ i radioactive fuel by radioactive decay. Data available in the control room had, in fact, indicated (1) that there was a gas other than steam in the system. The possibly (or likely) existence of a non-condensible gas (most like".cv hydrogen) in the system appee.rs not to have been reported either to the NRC or officials of the State of Pennsylvania. I (1) See EPR /NSAC-1, Appendix TH, P. 74, P. 88. ) -
-cu 3 .
Psrceived adequacy of procedures to' assure adequate core cooling. State =ents made during interviews conducted by accident investigators indicate that during the day of March 28 there were periods when TMI personnel appeared uncertain .
. as to core conditions and whether the procedures being followed and actions being taken were adequate to lead to stabilized conditions.
TMI Station Manager Gary Miller told General Public Utilities investigators on April 12:
... We, being me and Lee Rogers, called Lynchburg pretty early, and we sat in the room and every hour tried to figure out how to keep pumping water into it. But all we were doing_was pumping that BWST / Borated Water Storage Tank / through the electromatic to the floor. NothIng was changing, so you know we pumped 12' or 13'xout of BWST and my fear was pumping 50 feet out and the core still hot and no , water in the BWST. '
So our goal was to somehow get some circulation going,"either natural circulations from steam - generators or reactor coolant pump', using HP injection the whole time. My memory is that we pumped against the electromatic at fairly. high pressures like 1800 or 2200 psi. We could have pumped against the codes, but we assessed ( that what we'd do is get the same flow through the codes without being able to see pressure. We pumped through there until around 11 in the morning, at which time we decided to take a shot at getting on core flood. And the reason we tock a shot at core flood -- now remember Lynchburg ,was on the phone with a lot of good . advice, but it was clear that it was my decision -- we assessed that if we could get down and activate core flood tanks and we saw them dump on the core we could get assurance that the core had some water on. We couldn't tell that; we.were scared that wasn't happening. Radiation was all over the place, everything was off scale. You got nothing in the ' core that tells you about water level; you got no pressurizer level, since it's solid; no way of drawing the bubble: I didn't have any heaters; I didn't have any letdown;
. and we had radiation in every room we went to. i
~ Didn'.t even have oil pumps for some of the RC -
~
pumps; couldn't get in some of the rooms; the
- ceadings were horrendous.
. t . Mike Ross gave the following rationale for reducing system pressure in order that cooling might be achieved from what was thought to have been a better method than the one they were then' using which depended upon keeping the system pressure high and using the high pressure ! injection pumph:
. . . One, we were running out of water in
~
BWST and we hadn't gained an inch. I mean ) we hadn't gained any headway in where we were trying to go our goal being one to establish some mode of cooling a reliable mode of cooling to the core. Two, we, at least I wasn't sure that we purposely or we in fact had the core covered and all high pressure injection was going to the core. I was not totally convinced. I didn't have anything to tell me. Hey, high pressure injection is in fact going through the core. So based on that, we discussed going down with the goal being one, to let the core g ficod tanks come in and verify that yeah, the reactor was in fact covered, two, give it a drink of water. That is a little coarse I know, but give it a drink of water if it wasn' t getting it, . and maybe go on decay heat removal which is a forced mode of cooling. That is what we were trying to do at that time. " (I&E TMI Tape (226, p. 26) l 0 0
6 . l
... PerceptIions held on March 28 as to extent of core damage.
The foregoing indicates the existence on March 28 of manifold indications of core damage. Based on partial information available to him (i.e. he did not know of the
^^
2000 degree core thermocouple measurements or the manifestations of an explosion), Victor Stello, then Director of the NRC's Division of Operating Reactors, told. the E&E Subcommittee i TMI Task Force on May 9 that he suspected oxidation of the ' cladding (and presumably hydrogen production) on March 28. y l Mr. Stello engaged in the following dialogue: Mr. Stello: My initial reaction was that considerable failure of the fuel had occurred, with large numbers t ' of failed fuel rods. And that the potential for i oxidation, metal-water reaction, was clearly there. i The p.eriods of time involved were, even if the heat-up were not significant, at lower temperatures, with'the g top of the core uncovered for extended periods of sjg r'
- time, metal-water reaction was clearly possible.
The rates -- how high the temperature got I would not even indicate -- I could have guessed at. But 2 oxidation of the cladding was something I clearly expected; failures of the cladding were clearly evident. Mr. Terrell: When did you suspect this? On Wednesday you are talking about that you expected this fuel damage?. Mr. Stello: Yes, I think just from knowing what was happening to the core, you had to expect that there was quite a bit of damage to the fuel. (E&E, TMI _ Part, 1, p . 4 -5 ) e n
ima -
- b. .
k ,,,
- 38'-
llF S Another indication of there being suspicion on March 28
$l h of hydrogen production can be. inferred from a statement
- made by TMI supervisor, James Floyd. As noted above, Floyd
>T ' was at the B&W facility in Lynchburg, Virginia on March 28 !
L d- and had inferred from limited information that about DC ~ one eighth of the cladding had failed. On May 31, Floyd t- told the Kemeny Cormission of his suspicions about hydrogen C
# production when he engaged in the following dialogue with W
g Con ission member Ted Taylor: N Taylor: Had you sometime Wednesday made the connection k 2 in your Ic.ind between the high temperatures in the core, cladding failure and hydrogen? g Floyd: Yes, sir. NI' k The March 29 New York Times described a press briefing LO - Q. ~ presented by Jack Eerbein on March 28 which gave a different ji ch picture of the situation: W. M = Jack Herbein, a vice president of the utility that operates Wid - the plant, told a news conference held on a bluff over-looking the plant that "a valve . failed in a shut position" g-- in a feed pump that squirted water around the reactor. 4 Mr. Herbein spent almost an hour describing the accident f-h gt- -- and emphasized that the main safety systems in the almost brar.d-new plant had worked to prevent a very serious 3-C- . accident, or indeed a catastrophe.
$ Using bland terms, he described the series of events is in the plant as "not the normal evolution" in stating
{ that there was "some minor fuel failure." Ef This translates to the fact that som.e of the pellets F; of enriched uranium fuel became so heated because of 6 ; loss of coolant that the pellets melted through the h& zirconium clad tubes that hold the pellets. . E ; r- Unit No. 2 has 177 fuel assemblies, each containing 4 208 rods, each of which contain 200 uranium pellets. 5,U
; "Only a few of them melted through," Mr. Eerbine said.
Ei
- A-
5.- E
- E E-4:
w E_- On March 28, NRC issued a Preliminary Notification'of Event ED g or Unusual Occurence which gave little indication o'f the actual E-g severity of damage: y Facility: Three Mile Island Unit 2 b Middletown, Pennsylvania
& (Docket No. 50-320) r ~
Subject:
REAC2OR SCRAM FOLLOWED BY A SAFETY INJECTION AT THREE -{ P MILE ISIAND - UNIT 2 w_ - 2-m The licensee notified Region I at approximately 7:45 a.m. of an
$: incident at Three Mile Island Unit 2,- (TMI-2) which occurred at l4 approximately 4:00 a.m. at 98% power when the secondary feed y pumps tripped due to a feedwater polishing system problem W This resulted in a turbine trip and subsequent reactor trip kl- on 'igh Reactor Coolant Pressure. A combination of Feed Pump Gk stion and Pressurizer Relief - Steam Generator relief valve 3 at' ion caused a Reactor Coolant System (RCS) cooldown. At f 1s,0 psig, Emercency Safeguards Actuation occurred. All ECCS d components started and operated properly. Water level increased M in the Pressurizer and Safety Injection was secured manually Wi approximately 5 minutes after actuation. It was subsequently R resuced. The Reactor Coolant Pumps were secured when low net pcsitive suction head limits were approached. '
ME . . . . . . . . . .- .
$ At 10 : 4 5 a .c. the Reactor Coolant System Pressure was being held gj at 1950 psig with temperature at 220*F in the cold leg. By SE 10:45 a.n. , radiation levels of 3 mr/hr had been detected j 500 yards offsite. (PNO-79-67) $e '
f! 4 - g The failure to indicate that hot-leg temperatures were in E j excess of 750 degrees F at 10:45 a.m., that in-core temperatures y in excess of 2000 degrees r had been recorded, or that the LIE l g,, containme$1t building radiation monitor was indicating readings l i( of several thousanc roentegen (R) per hour conveyed an f i ; inpression that the. situation was in much' better shape \ ( {=- than was the case.
=
E fi 2
- b -.
'~' .
b .40 - q The March 28 Washington Star quotes Lieutenant Governor Scranton: T- "Everything is under control. There is and was no
- _ danger to the public health and safety."
9 Officials of Metro'politan Edison / General Public 2
- - Utilities have said subsequently that on March 28 they did not l-h ,
reach conclusions such as those of Stello and Floyd referred to F above about the@ssibility of cladding oxidation. e > In answer to a written question as to what at 11:00 p.m.
~
3_ j on March 28 was his estimate as to the amount of cladding
',=h Y that had been oxidized during the period when the core was y
f[
;7 uncovered, Eerman Dieckamp, President of General Public f Utilities, responded: ?lE 2 . I interpreted damage to mean cladding failure. I 1[5 did not think in terms of core uncovering, high R temperatures , or Zr-Hg) reactions. (E&E, TMI Part 11, le$ ~
o.
~
2 8 5,') ~ T*g g In response to the foregoing question, Metropolitan pg Edison Vice President John Herbein stated: I didn't consider a steam clad interaction producing
' hydrogen until it was discussed on March 30th. (Ibid.,
II EF p. 290) l_E c R-- In response to the question as to what was his estimate on E y March 28 as to' cladding oxidation, Station Manager Gary' Miller p . Q stated: i t' 4 l
# " . . . I did not conclude there was an on-going steam-I h cladding interaction and therefore, had no estimate I f of the amount of cladding f ailure by 11:00 p.m. on
($ March 28." (Ibid., p. 298.) e
- l
~
4,, I
. ... :.. r
_ 41 _ Whatever the indications of hydrogen production available to the NRC on March 28 and early March 29, little idea of these was conveyed to the E&E Subcommittee on March '29 when, as noted above, Chairman Hendrie stated:
"The magnitude (of release of radioactive gases) suggests that perhaps one percent of the core might
~
have been i_nvolved in the cladding cracks." (Transcript,
- p. 60.) ,
P 1 l O e 1 O m}}