ML20037A294
| ML20037A294 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 06/13/1979 |
| From: | Murley T NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| TASK-TF, TASK-TMM NUDOCS 7907160399 | |
| Download: ML20037A294 (15) | |
Text
.
g
.L.
[gs,a *ecq.'*g UNITED STATES NUCLEAR REGULATORY COMMISslON
- .. ft, e
-'c j
W ASHINGTcN. O. O. 20555 N/
y r
~,,,,, : e ::/ O MEMORANDUM FOR:
The Files FROM:
Thomas E. Murley, Director Division of Reactor Safety Researca
SUBJECT:
RECORD OF ACTIONS, THREE MILE ISLAND ACCIDENT MARCH 28 - April 6 This memo recorcs my major activities and lists the individuals with wncm I had substantive contact during the Three Mile Island Accident and its i. media e after.ath, March 23 - April 6,1979.
On Wecnesday, March 28, I learned about 9:30 a.m. tha: there had been an accident at TMI.
Not much was known of the details except that there were high radiation levels in the plant and a general emergen.cy had been declared by the utility.
On Thursday morning I attended a briefing of the Comission by NRR staff where I gained the impression that the situation was generally under control, although the high radioactivity levels in the plant clearly indicated extensive fuel damage.
On Fricay, Maren 30, while on annual leave, I received a call from Saul Levine a, 4:00 p.m. asking me to come to Tony Buhl's office to get some tests started to helo resolve some problems at TMI.
On the way to the office I heard a bulletin on the car radio that NRC had announced there was a danger of core melting at the Three Mile Island Nuclear Plan:.
The following activities are listed by topic and are generaily in enronological order.
Remval of the Hvdrocen Bubble U::on arrival in Tony Buhl's office I was informed that reasurerents at the site indicated there was a noncondensible hycrogen gas buoble in
- ne reactor vessel naving a volume of 1000-1500 cu. f:. a: 1 %C osi anc 220*F.
There was some concern that the bubble was growing an -ign:
lead to uncovering the core and potentially to fuel melting.
One option being considered was to coen the relief valve on the cressuri:er and try to vent the hydrogen bubble out the valve to ?.e containment building.
It was recogni:ed this would be a tricky maneuver since it would mean that :ne one pumo operating a: the time woulc nave to be snut off and there was no assurance na: it or any of the c ner three i
puros could be restarted if :ney were suosecuently neecee.
,. A s'
b l C
4s "j
y 7,
09 7.
The Files At about 5:30 p.m., we called INEL (Larry Ybarranco, Nick Kaufman, Hank Ziele and others), described the proole= to tnam and asked if tnere were meaningful tests that could be done in the Semiscale facility to help decide whether and how to vent the hydrogen bubble.
We described tne TMI primary system layout and gave them tne imocrtant plant dimen-sions and elevations.
Later in the evening on Friday, they called back with the following information and suggestions:
A test could be run in Semiscale using nitrogen gas to simulate the hydrogen venting maneuver in TMI.
They recormended against such a maneuver in TMI, suggesting it was better to keep the plant in its then s able c;erating mode (this suggestion was relayed to the NRR sersonnel in the Incident Response Center).
After working all night to set up the Semiscale facility, the INEL staff ran a test on early Saturday morning (Enclosure 1).
The results showed that the Semiscale system could be depressuri:ed by opening the pressuri:er relief valve and turning all pumos off but about half of the nitrogen gas in the bubble remained in the primary system -- princi-pally in the steam generators.
The electrically heated rods remained cooled during this maneuver.
This information was passed on to the IRC and to B&W.
During the remainde'r of Saturday and Sunday, Wayne Lanning of the RES staff worked with 3&W engineers and INEL to establish conditions for a second test using a larger helium gas bubiale and injecting coolant from the High Pressure Injection System.
This second test was run early Monday morning, April 2 (see Enclosure 2).
The Semiscale tests provided the following general information:
They shewed that tne noncondensible gas in the bubble would not all vent out the relief valve -- in fact a: cut nalf of the gas would remain in tne primary coolant system.
They showed that.it would be dif' icult if not imoossible t:
f remove the gas frem the primary system by subsecuently restarting the primary pumps. Since the presence of gas in the high points of the steam generators could orevent natural circulation cooling, this information was a strong argument for leaving tne gas bu:ble in tne ::p of the vessel.
They Orovided useful tata for S&W in esta:lishing ne a: re:riate MP!S flow rate for their pro:osec eme gency Orc:sture in :ne ever.-
all :ri:ary pum:s ere 1:st.
JUN I O 79 The Files 0xycen production Rates On Saturday, March 31, we received a question from the staff at the IRC whether there could be sufficient oxygen gas in the primary system to form an explosive mixture and thereby constitute a threat to the reactor pressure vessel.
The answer to this question provec to be elusive.
I discussed this question with staff memoers from INEL (Sid Cohen, Concurrently, Saul Levine called Ron Ayers and Jack Liebenthal).
Bob Rit:mann of Science Applications Inc., and we understood that Sob The infor ation I received Tecesco of NRR was contacting staff at KAPL.
from INEL was based on reported data from the Cooper plant (a SWR) and Their conclusions, which was scaled down to the power level of 25 MWt.
they stressed were extremely conservative, were that the hydrogen bubble l
contained about 2.2t oxygen and that it would take at least 4 to 5 more I was later given some data cays to reach 5t oxygen concentration.
from the Advanced Test Reactor (ATR) that was purported to support the i
l data from the Cooper plant.
I found it very difficult to piece together all of the information into The Cooper SWR data were not directly applicable a consistent story.
to TMI (a PWR) although there was some boiling in the TMI core.
Similarly, the ATR is a low pressure (150 psi) reactor and was also Late on Saturday evening-I received a not directly applicable to TMI.
call from Rob Rit:mann who reported that he was not having much luck in calculating the oxygen concentration, although he believed it was below the flameability limit.
Some time after midnight on Sunday morning,*I went to the IncidentI told Response Center where Roger Mattson asked what we were finding.
him that the picture on oxygen concentration was confused, but that a conservative estimate seemed to be that the oxygen concentration in the nyocogen bubble was increasing at the rate of it per day after reactor scram.
Late-that morning (around 9:00 a.m.) Roger Mattson met with Saul Levine, Bob Sudnit: and me at the IRC prior to leaving for tne TMI site.
Chairman Hendrie, Comr.issioner Gilinsky and Commissicner Kennedy came Mattson sumnari:ed and,went throughout t.his short meeting as I recall.
the following information as the distillation of all of the input ha had received:
Flamcability limit 2F 5t 02 in pure H2 Detonaticn limit 2 12t 02 in pure H Comoustion limi:
213t 02 in H /steim i
4 Og production rate
= it 02 per day in H2 Oubtle i
Current 02 concentratien = 50 02 in H2 :ubole l
I l
l i
i J'J.*i * ' 1373 The Files 4-After this meeting, I spent little further time on tne oxygen concentration question.
Potential Pressures from Hydrocen Exolosion On Sunday morning someone suggested that I collect information on what pressures could De generated if there were a hydrogen explosion in the pressure vessel.
I found that Vince Noonan of NRR was the focus in NRC for these analyses and I therefore was involved only peripherally.
I received information that Dr. Norman Slag of Picatinny Arsenal had rade calculations showing a sharp peak pressure of 12,500 psi for ne case of a detonation of a 1000 cu. ft. bubble containing a mix:ure of 53% H '
2 12% 02 and 5% steam.
This pressure appeared to be consistent with information received by Bob Budnit: frem Dr. Bernard Lewis of Pittsburgh that pressures could reach 5 Po for def.lagration and 13-14 Po for detonation (where Po = 1000 psi is the pressure of the bubble and the system initially).
This information was passed on to Vince Noonan and I had little further involvement after Sunday afternoon.
Hydrocen Gas behavior From the beginning of my involvement in the TMI ac:icent one of the tasks being investigated, principally by Bob BGdnit:, was how to eliminate the hydrogen bubble by mechanical or chemical means.
During Saturday and Sunday I received a number of unsolicited suggestions fr m peccle calling in to NRC.
One of :ne most novel suggestiens care from Roger Billings, President of Billings Energy Co. in Provo, Utan.
Sillings speciali:es in hydrogen researen and technology,and they had done work for DOE under a contract frem 20E-Idaho.
His suggestion was to inject into tne primary coolant a large num:er of 2-4 micron diameter evacuated glass microscheres.
The idea as that the nycrogen in 50luti:n would diffuse easily througn the glass and be tra::ec in :ne 5:neres, thereby gradually'deplenishing the hydrogen in the system.
Eiilings claimed to have experience with these micres;heres.
I felt the idea was worth a look, so I authori:ed INEL cn Sunday night to nave Billings start an experimental program on hydrogen behavior.
The Billings staff began work on Sunday nignt and ran :neir first nycrogen solubility tests on Monday morning.
Altnougn the micr:s:nere icea
- rovec to be not feasible, Billings did scre very ;;;d -orx cver tne foli:n-ing three weeks that preved very neipful in understanding nyce: gen te-navior in TMI (see Ecnlosure 3 for a ccm lete recor: Of :ne' r, ort!.
l l
JLN : 3 1373 The Files Sene of the conclusions reached fr:m :neir w rk were :ne folicwing:
Adding nitrogen gas Oc TMI coolan :: try t: fccm ammonia (NH3' and remove hycrogen would not work (reaction was :ce slow)
Acding micros:neres to trap hydrogen gas wouic nc work (takes weeks)
Adding hydrogen eroxide (H 0g) as a scurce of cxygen to ecmcine 2
with hycrogen in solution would not work (02 gas evolved wi:nin minutes) r Measured H2 solucility at TMI conditions Measured H2 evolutien rate if TMI were to de:ressuri:e Provided a means for determining the pressure at which Hg saturation was reached if TMI were to depressuri:e (pressure rebound effect)
During the week of April 2-6, Billings was out in direct touch with the NRC team at the TMI site, and they continued to provide information to NRC during the following weeks.
Hvdrocen Decassine On Monday, Acril 2, I was asked te calculste how icng it would take o remove nydrogen from the coolant to cegassing througn the le::cwn system and througn the ressuri:er spray 'ystem.
I called Glen Jenks of CRNL for information en hydrogen removai :nrougn pressuri:ers, an ne :re-viced me with scre useful data and several references :: icok u.
My calcula:icns snowed that it would :ake from 1 :o 2 -eeks :: recuce One nycr: gen in solution from 1500 sc:/kg (sa urated ::ncentra:icn a-r 350 :si and 230*F) :: 300 sc:/kg (satura ec ::ncantra-icn a-300 :si anc la0*F).
These cai:ulations also sn:wed ina: i: -Oui: ase nearly a aeek :o reduce the. hydrogen bubble by degassing thr:ugn ne ie:ccwn sys em.
The only way the hydrogen bubbie could nave been reduced fr:e 1000 cu. ft. to :ero in 2-3 days is by as:uming nign flow rates nrougn the pressuri:er soray line (15 gpm) and a nign efficiency for hycrogen removal in the pressuri:er (90%).
I am skectical :na: :nese ' lows anc efficiencies were attained in TMI.
- have nc-hear: a convincing s ory of how ne 1000 :u. ft. nycrogen Ou: ie oas re:uce: : r.0:ning in 2-3 days, i
e e
Ths Files UN I 31379 Measurement of Water Level in Pressuri:er On April 5, I received a call from Tony Buni to help wi:n one of :ne instrument problems he was working on.
The problem concernec wna would hapoen if all the water level indicators were los: in TMI.
Buhl was considering whether the resistance tem:erature device c:uld be used as a backup level indicator oy running a hign current
- nrougn the RTD and coserving the cnange in temoerature wnen tne water level drops below the RTD in the pressuri:er.
I called Hank Ziele at INEL and put him in touch with Bob Shecard of ORNL to get the details of the RTO design.
The INEL staff then ran a scoping test in an autoclave and determined that the technicue c:uld be used in erinciale to indicate water level in the pressuri:er.
I subsecuently learned that there were oractical problems at TMI :na: made Onis appr:ach no:
feasible.
Follow-on Work Since April 6. my staff and I have been involved in extensive analyses of the TMI accident.
These activities are documented in the formal transmittals from RES to EDO.
Thomas E Murley, T'.. tor Divisien tf Reac::r Safety Researen
Enclosures:
1.
"Semiscale Pressuri:er Relief Valve Venting from Three Mile Island Tyce Concitions" 3/31/79 2.
"Seconc Semiscale Relief Valve Venting test f m TM! Conditiens" 2/2/79 2.
"TMI React:r Simulation Final Re:cet,"
Billings Energy Coco., 4/20/79 c::
- 5. Levine R. Budnit:
K. Cornell J. Cummings l
l
Fridav PM and Saturdav AM Eudnit: (recollection)
In conversation with Saul Levine and Tom Murley of RES, the following divisions of responsibility were arrived at.
All tasks were to be carried out by contact with contractors and consultants around the country.
I was to be in charge of working out at what level of oxygen in pure hydrogen the threshold of combustibility would be reached, and at what higher threshold the explosion possibility would set in.
I was then to be concerned with the properties of a fast combustion event and of an explosion, such as the dura-tion of and strength of the pressure pulse inside the TMI vessel from such an j
event.
I was also to be concerned with finding any possible mechanisms that might set off such a combustion or explcsion event inside the pri=ary TMI vessel.
Finally I was to try to find experts who could suggest and evaluate methods (using che=icals or physical means) to remove or decrease the hydrogen within the pri=ary system.
I recall that To= Murley undertook to work with i'ayne Lanning to study the feasibility of hydrogen removal out the pressurizer relief valve.
This last resulted in performance of a quick experiment early Saturday morning at Se=iscale at Idaho National Engineering Lab, and another on Monday corning, April 2.
I recall that beginning on Saturday =orning Levine and Murley undertook to contact experts on the issue of how much oxygen build-up there would be in the primary system, including the crucial factor of oxygen recombination.
I I
1
o 2
Saturday AM l
Eudnit: (meno)
I worked heavily on Saturday and Sunday on the question of understanding the i
issue of hydrogen combustibility and kinetics in a reactor vessel such as at TMI.
On referral from Richard Garvin of 13M, I called Dr. Harry Petschek of AVC0 Everett Research Laboratory on the morning of Saturday, March 31, finally reaching him at home in late morning.
He responded inmediately by indicating that he and so=e colleagues could assist in understanding the issue of hydrogen combustibility and combustion kinetics in a reactor vessel such as at TMI.
Later that day and through Sunday, April 1, I spoke, two or three times, to Dr. Petschek and one or two of his colleagues.
I was their sole NRC contact as far as I know.
They worked on the question of what concentration of oxygen in pure hydrogen would be the threshold for combus-tion, particularly at the temperatures and pressures thought to be present at TMI (about 100 psi at many hundreds of degrees F), and reported back sometime Sunday on those.
Dr. Petschek also referred me to Dr. Bernard Lewis in Pittsburgh, who turned out to be a highly-regarded expert in just these same issues. I finally reached Dr. Lewis on Sunday :orning, April 1 (see below).
Twice during this period (Saturday, March 31) I talked with Dick Garvin abou l
hydrogen combustion.
He gave ce insights into how i=portant the back reaction is in a proper calculation of the pressure pulse during a fast burn or detonation of hydrogen in a vessel like the TMI reactor vessel.
b t
t t
i
[
t
.. _ _ ~
3 Saturdav.e' Budnitz (memo)
All day Saturday I worked, off and on, on the idea that a snake-like device might be obtained which might be inserted into the pri=ary system to recove the "3as bubble".
Dr. Richard A. Garvin of 13M first suggested this idea to me.
I telephoned Dr. Hein: Heinemann of Lawrence Berkeley Laboratory on this subject, and on his referral to Mt. Joseph Penick of Mobil 011. Later that day, Saul Levine contacted Edward Mason of A=oco on the same subject.
Although much effort was spent on the part of many people in those co=panies, and there were cany telephone calls back and forth between NRC (me) and various people, neither contact ulti=ately resulted in anything of use to us.
I contacted Mr. Penick on Saturday morning, March 31.
He said that he thought
[
Mobil could assis: NRC with advice on the availability of snake-like devices to extract gas from a TMI-like pressure vessel.
He called back later during the weekend (I recall his return contac: as occurring on Sunday, April 1) and indicated that devices such as we sought were not readily available in the Mobil Corporation, and unlikely to be available elsewhere in the petroleum industry.
The problem was that the path into the reactor vessel from the outside to the upper dome was tco tortuous for the use of the devices that did exist, and the fabrication of a special device would be quite difficult.
I believe that I was Mobil's sole NRC contact on this subject.
Our (negative) results were communicated from ti=e to time to the IRC people, mainly to R. Mattson and D. Eisenhut in my recollection.
b I
r I
i h'
l l
t l
i
~
4 i
Saturdav AM Budnitz (memo)
As a subsidiary activity on Saturday and part of Sunday, I attempted to find experts around the country who might suggest various che:ical means to re=ove hydrogen from the primary pressure vessel.
I first called Dr. Hein: Heinemann of Lawrence Berkeley Laboratory, in the morning of Saturday.
Dr. Heine= ann I
is a chemical engineer at my former laboratory in Berkeley and is a colleague I
and friend there, who spent most of his life working for : obil Oil Corpora-tion.
Dr. Heinemann discussed with me the question of addition of catalytic j
chemical agents to decrease the hydrogen in water solution.
Dr. Heine= ann gave =e the names of several catalysis chemists who might have expertise in this =atter, and also enlisted the advice of two Berkeley colleagues.
"e talked several times over the weekend of March 31-April 1, but I learned sometime on Sunday, March 31, that this proble was being attacked by l
engineers at the GPU Service Corporation, and I turned over to Bob Cutler at GPUSC the names of experts I had turned up.
]
l l
J i
4
_...~
5 i
Saturday AM Budnitz (memo)
Dr. Laura Cherubini called me on her own from her home in 3111 erica, Massachusetts, on Saturday, March 31, with a suggestion of biological /
chemical means to reduce or eliminate hydrogen dissolved in the reactor coolant water.
I do not know how Dr. Cherubini re:eived a reference to me.
The method was to use algae that trap hydrogen from solution by presence of free electron acceptors.
Since I was not expert in this satter, I turned it aver to others at NRC for follow-up. However, by the time anything more could be done with this suggestion, the perception of the importance of a
" hydrogen bubble" had di=inished, and I think that no further follow-up occurred.
8 i
I t
6 Saturdav A'! ". arch 31 Budnit: (memo)
On referral from Dr. H. J. Kouts of Brookhaven, Saul Levine and I contacted Dr. Harold A. Schwar: of Brookhaven.
I am not sure when he was first reached by Levine, but both Levine and I talked to Dr. Schwar: at various times during the weekend.
Dr. Schwar worked
-"A
- i
^u^''--"
" April 1 on calculating the production and recombination rates of oxy 3en in the D!I primary coolant water. He did these calculations at home mostly, I think; telephone contacts with him during the weekend were at his home.
He reported on the considera-tions that were involved in his calculatio-s, and ultimately showed defini-y',
tively that oxygen generation from radiclysis would not result in much oxygen 7
in the gas phase, because of the recombination reaction with the assu=ed g) large hydrogen gas overpressure and the associated dissolved hydre en.
tie V
[;. ].
vere apprised of the preliminary results of Dr. Schwar:' wor' earls on Sunday, April 1, in my memory, but it was not fired up until.sometime y.
', e '[
_. ?,
shortly after midday on that day.
Dr. Schwar: continued with his work for several days' af ter Sunday,
- April 1, and filed a description of his calcula-tion with NRC on April 24. The April 24 note to me describes well that the calculation is tricky and involves tracking down a number of chemical parameters and calculating several reaction rates.
Impurities within the DiI cooling water were i=portant enough that information taken from conditions of normal reactor operations could not be relied upon.
$,a
-f f
ba
/
A ff Q < ',# '. h.,*.,vJ U
W s
ejOpt 6
i.
.s
- I,.
,.,. u l
k' M -*c *t t t%/
. v %g l '.
~~.
.)
,c1l h W( l. ' ' { )y
-r1 i
No"I
( &,*'
h R
co'&
f Iest"
. -,.a t-jg
.,vp c-LL A
/
r L
7 Sunday 9:15 AM Budnitz (recollection)
Meeting with "attson, Levine, Murley, me.
Mattson was about to depart for the TMI-2 site by car with Chairman Hendrie, and he asked for an up-to-date, agreed-upon set of numbers about the " bubble." We all agreed on the following set of nuchers, which were written down by =e and are in my notes:
1.
Best estimate for fla== ability limit, pure 0, in pure H, at 1000 psi =
5: 0,.
2.
3est estimate for detonation limit, pure 0, in pure H, at 1000 psi = 12%
0,.
We also agreed that the best esti= ate for total oroduction of 0, from radiolysis l
in the TMI vessel would be about 5% oxygen (as fraction of bubble) as of the present t L= e if there were no recombination. We also discussed the fact that other people a=ong our NRC contractors were actively calculating the true oxygen which would be smaller than the radiolytis production rate because of chemical recombinat1on.
We discussed together that the 5% 0 /H rati was a 3
2 pessimistic estimate, in the sense that the actual value was' surely not larger and might be smaller.
I re= ember us discussing how probable it was that the number vculd ultimately turn out to be much smaller.
I remember that we were not yet sure that a smaller number would result from the accurate calculations underway at that time. Some difference of opinion existed as to how "sure" we were of what estimates by which experts, and Mr. Levine was more confident than the others that the ex73en would not be a problem.
I reme=ber that Cc missioner 3radford came into our brief meeting just as it was breaking up, and we acquainted him with our discussion.
r e
b
~
8 Sundav AM Budnit: (=e=o)
On Sunday =orning, on referral from Harry Petschek of AVC0 Everett, I reached Dr. Bernard Lewis of Co=bustion and Explosives Research, Inc. of Pittsburgh.
He acknowledged that he had =uch expertise on the co=bustibility of hydrogen and oxygen; indeed, he is the coauthor of a definitive textbook on this subject. He and an assistant, reached at ho=e on Sunday =orning, worked through that day and part of Monday, April 2, and gave i=portant advice on the issues that governed the physical behavior of hydrogen and oxygen burning in conditions such as were thought to exist within the primary vessel at TMI.
I was Lewis' sole NRC contact during this period.
He gave infor=ation about the =ixture t f oxygen in pure hydrogen that would be a.co=bustion threshold, talked at ler;th to =e about the physical difference between co=bustion and explosion, and what would be the impact of gaseous i= purities including stea=.
He reported back his prelL=inary conclusions so=etime af ter =idday on Sunday, April 1, and his final conclusions in =idcorning of Monday, April 2.
He calculated pressure ratios (pressure within a fast burning situation vs.
starting pressure), detonation thresholds, heat release, fla=e te=peratures, and other para =eters.
His insight was valuable in providing a perspective on which para =eters were, and which were not, i=portant in =odifying the result of what was easily calculated using approxi=ations. Thus the approxi= ate calculations were refined, and the refine =ents were better understood by =e.
Dr. Lewis' best esti= ate about =idday Sunday was that the " pressure pulse" would be about 5500 psi for a fast co=bustion event (requires above about 5%
oxygen) and about 13,000 psi for an explosion (requires above about 12%
oxygen).
On Monday, Dr. Lewis revised the 5500 esti= ate downward to about 3200 psi.
l 5
1 a
F 9
'l~&.
t Sunday R$
- Av, -
M.,,,
Budnit: (recollection)
By about midday on Sunday, there emerged a reasonably reliable consensus among our outside expert contacts about two questions:
how large the short pressure pulse might be if a fast co=bustion event were to occur inside the TMI prf.cary vessel, and at what pressure the vessel would approach i s " yield t
point" and its " ultimate" point, thereby compromising the integrity of the primary system.
I was responsible for coordinating the former concern and Tom Murley brought together the information on the latter.
Our general conclusion at that time was that, for about 1000 cubic feet of pure hydrogen with oxygen just above the threshold value of about 5% at 1000 psi, the " pressure pulse" would be about 5500 psi.
This came mainly fro =
Dr. Lewis of Pittsburgh (on Monday, Dr. Lewis revised his best estimate downward to about 3200 psi).
The " yield point" of the vessel, according to Tom Murley, seemed to be coming out at about 6000 psi.
Both of these numbers had rather large uncertainties, of which I and Merley were acutely aware.
I discussed these rough results with Saul Levine, who suggested that I tell Chairman Hendrie, who was at the TMI site.
I spoke by phone with Hendrie and relayed this to him, expressing the uncertainties to him as well as the nu=erical best esti=ates m I
T D~ r ~l Immediately thereaf ter4-I Ua' ~raquested to give, and did give a briefing on s
to the other fo'r NRC Cec =issioners, all gathered together the same subject u
in the :TRC Incident Response Center. This 15-=inute briefing was attended also by Len 31ckwit (NRC Ceneral Counsel), and about 15 other NRC people.
This briefing can be found verbatim in the Co==ission transcripts.
/
.5
, st Y
./
_. ),
E w ***
r~s t j m,.
+._,,,.
1 A
r UI g
UNITED STATES
[ '
.c y %i NUCLEAR REGULATORY COMMISSION
"/
C WASHINGTON. D. C. 20555 5
. *:gg gg MEMORANDUM FOR:
R. J. Mattson FROM:
T. M, Novak
SUBJECT:
CHRONOLOGY OF EVENTS: TMI-2 HYOR0 GEN BUBBLE I have reviewed Enclosure 3 of your August 21 memo and offer the following additional comments.
With regard to any discussions I had concerning the potential for a H 2 explosion inside the reactor vessel, the only conversation I had occurred the evening of March 31. The discussion took place between V, Stello and myself at the NRC trailer complex at TMI, At that time Stello infomed me of a concern staff personnel in Bethesda had regarding the potential for a H2 explosion inside the reactor vessel, I recall Stello mentioning he did not believe the concern was real but was unable to convince Bethesda staff personnel to change their view, With regard to any information I may have received from B&W regarding the presence of free 02 in the vessel, I have no recollection of any information transmitted to me on this subject, I do recall having several conversations with S&W personnel during the evening hours of March 29 and for the first few hours of March 30 regarding a variety of subjects, but primarily concerned with the state of core coolability and the size of the gas bubble inside the reactor coolant system. My personal recollection is that the members of the staff were evaluating the pros and cons of maintaining the conditions that existed in the reactor coolant systens as opposed to depressurizing the system, In reviewing the notes that were available in the IRC, there was indication that B&W had discussed free 02 after 10:00 p.m on March 31, 1979, I discussed these notes with Don Davis of the staff and he does recall discussions between staff personnel with regard to the amount of 0; that might be cresent in the gas bubble, He stated that it was possible he had discussions with S&N (Nitti) regarding this concern, Don Davis also thought that Steve Hanauer was the staff member with whom he had these discussions, I talked to Steve Hanauer on this subject and he ncted that he first arrived at the incident center at 2:00 a.m. on " arch 31 and recalls discussions with Davis regarding H2 solubility but none with regard to 0.
2 n
Thomas M, 'lovak 0
m,, a,r on! W L.M ' Y
- u
!(pa "*c oq fo, UNITED STATES
)3ec3 j NUCLEAR REGULATORY COMMISSION C
WASHINGTON, D. C. 20555
- .'t o.
%.a 7 j g
j
- ..=
AUG 41973 NOTE T0:
Roger J. Mattson, Director Lessons Learned Task Force
SUBJECT:
CHRONOLOGY - HYDR 0 GEN BUBBLE CONCERNS Suggest you add the following comments to your chronology. The spanfof time for which this comment applies is uncertain in my mind, but to the best of my recollection, it was late Saturday evening through Sunday morning, six a.m.
Title the comments as Stello's recollection.
During the evening of Saturday, March 31, and morning hours of April 1, I asked Mat Taylor to look into the hydrogen problem from the point of view assuring the need to start the containment atmosphere hydrogen recombiner and the possibility of adding oxygen to the hydrogen bubble believed to be inside the ~ reactor vessel, such that a burn or explosion of a hydrogen-oxygen might result.
This issue was discussed on and off during that period and my best recollec-tion is that it was concluded that no net oxygen would evolve as a result of radiolysis since the back reaction due to the hydrogen overpressure would force all of the oxygen produced to be recombined with the hydrogen to form water. Neglecting the hydrogen overpressure, results of analysis by Mat Taylor suggested an oxygen evolution rate of about 30 standard cubic feet per day could occur. With this evolution rate, it would take many weeks to reach a flammable mixture.
Mat Taylor's advice to me was that I should not have any concerns for the potential of a hydrogen burn or explosion within the reactor vessel.
I believe I also requested others to seek information from representatives of various companies that were located at the site.
I believe John Collins returned with some information that suggested GPU (and possibly B&W) also did not consider the hydrogen in the vessel to present a hazard from a burn or explosive nature.
As a second coment, suggest adding a new title. Early p.m. Sunday, April 1 -
titled Stello's recollection.
I recall outlining for the Chairman the basis upon which I did not believe the nydrogen bubble in the reactor vessel could reach an explosive potential.
Basically, I reiterated the results of the reasoning we had gone through that morning and previous night.
I requested the Chairnan that before any further action be taken on this issue, that I be given some time to cor. tact some other experts to see if I could get additional tecnnical information.
)
C 0
1 7
O
Roger J. Mattson !.UG 4 '975 At approximately 3p.m. on the afternoon of April 1, I obtained that information.
After presenti.ig the information to the Chairman, I believe he was persuaded that we need not have a concern for either a burn or explosion of the hydrogen in the reactor vessel.
Following our discussion, he placed some telephone calls to Washington to discuss this information with fellow Comissioner(s).
Victor Stello, Jr.
Director Office of Inspection and Enforcement l
1 i
r p* '8 %,
UNITED STATES NUCLEAR REGULATORY COMMisslON f 'h e.[
WASHIPJGTON D. C. 20555 3
E.,; *....f August 24, 1979 MEMORANDUM FOR: Roger J. Mattscn, Director Lessons Learned Task Force FROM:
William H. Beach, Technical Assistant to the Director Division of Reactor Safety Research
SUBJECT:
CHRONOLOGY OF EVENTS: Tf1I-2 HYDROGEN BUBBLE Tom Murley was in Idaho on official business when your August 21, 1979 memo arrived.
I called him to let him know in general what it con-tained.
I told him that you 13,.1979 memo to the files on his reco,had quoted from parts of his June llections concerning TMI events.
He said that was fine and asked that I inform you of his OK.
WilTiam H. Beach, Technical Assistant to the Director Division of Reactor Safety Research 9} a n.( J.L y "
- r ~ t y ~
An /14 I ml 9
t A
a p* * *%e 0
UNITED STATES
[ *..., [,}
NUCLEAR REGULATORY COMMISSION 3
^/
f WASHINGTON, D. C. 20555
%.../ 5 lf August 27, 1979 MEMORANDUM FOR: Roger J. Mattson, Director Lessons Learned Task Force THRU:
Frank H. Rowsome Acting Director Probabilistic Analysis Staff, RES FROM:
M. A. Taylor, Sr. Reactor Safety Engineer Probabilistic Analysis Staff, RES
SUBJECT:
CHRONOLOGY OF EVENTS: TMI-2 HYDROGEN ' reflects a chronology dn the TMI-2 H2 matters that I recall' having any personal knowledge of or direct involvement with.
I hope this contributes to rounding out that chronology of events circulated by your August 21, 1979, memo.
I trust that Enclosure 1 will be discussed with those I have mentioned so that we will be advised as to the completeness and accuracy of my recollections.
Note that the specific times of day cited herein should be regarded as being very approximate.
I found the chronology in your August 21 memo to be quite enlightening about Bethesda efforts concerning the H2 explosion potential.
Because I was largely outside of these Bethesda events and at the Ti1I-2 front lines, I have no constructive comments or corrections to make.
~ l// f II -
.t. A. Taylor Sr. Reactor Safety Engineer l
Probabilistic Analysis Staff Office of Nuclear Regulatory Research
Enclosure:
As Stated cc:
S. Levine l
R. Sudnit:
l J. Murphy I
C AMg..
h D /_[3l
[ p 3,,, y v Y "
F TMI-H9 Concerns Chronoloey of Events (recalled by M. Taylor from his involvement ir. M2 matte.rs)
Thursday, March 29, 1979 s2030 M. Taylor received a call at home from T. Novak and W. Minners who were at the Bethesda HQ-IRC.
Question asked was what would be the radiolytit generation rate of H if TMI-2 reactor was to be reduced 2
to low pressure levels to permit operation of the low pressure decay heat removal system. Taylor advised that most of hic info for making such H2 estimates was at Bethesda office, but that he'd try to recall work of N10 years ago and get back to HQ-IRC shortly with an estimate.
s2200 TaylorcalledBethespaHQ-IRC(Minners/Novaket.al.)withaback-of,-envelope estimate for radiolytic H2 generation considering near-atmospheric pressure levels and non-boiling bulk coolant temperatures.
The estimate was roughly 1.4 SCF H /hr for these conditions, but HQ-IRC 2
was advised to use a 1-2 SCF H2/hr range because of considerable uncertainty about actual TMI-2 fuel conditions and fission product releases to coolant, etc.
Taylor also advised that if HQ-IRC wanted better estimates and more detail on radiolytic H2 behavior, they
~
should make contact with Dr. Bud Zittel of ORNL, or possibly prof. Reed Johnson, U.Va. (currently technical member on ASL3).
Dr. Zittel, in particular, had considerable experimental experience with post-accident radiolysis at our (i.e., AEC) request s10 years ago.
Friday, March 30, 1979 s23C0 At TMI Site M. Taylor took phone call from Chairman J. Hendrie, NRC, who asked that the following items be relayed (note on this call passed to D. Ross):
1.
Convey thanks/ congratulations to H. Derton en the good job beir.g done and on his press briefing, 2.
Convey to V. Stello that before things were over, the TMI duties could be of benefit to his waistline, I These estimates made at home relied on past excerience and analysis, recall of CRNL excerimental work and results, CRSCRT notes of 1933, MIT handbook on Reactor Safety (T. Thcmpson et.al.) and extrapolations to and assumotions about TMI-2 condi tions.
f h
3.
Express concern about 02 evolution in IMI reactor vessel and possibility of H -0 explosion; he asked for these concerns te 2
be relayed to Bethe,sda staff so they would get moving on an assessment.
Saturday, March 31, 1979 s0100-
- TMI Site At 0200 M. Taylor called Bethesda HQ and relayed to Dr. S. Hanauer the concern about 02 evolution and the possibility of explosion in the R.V.
Taylor asked Dr. Hanauer to assure staff gave prompt attention to this matter.
s2000 M. Taylor received verbal request from V. Stello to give thought and analysis to following matters and to advise him on these before shift end:
1 Possible "what if" system scenarios and possible consequence outcomes plus best action courses that might be followed, balance calculations to estimate Zr-H O reaction 2.
Try to make H2 2
magnitudes anc where H2 sources might be, 3.
Give best judgement on whether or not H2 explosion in vessel should be of worrying concern.
Here it should be noted that Taylor recalls discussing the prior 1.4 SCF H /hr estimates with V. Stello.
(The specific time or March 31,21979, when this information was discussed with V. Stello is not recalled.
It is believed to have been during the latter part of the prior shift, sometime in the am, e.g., 0400-0800.)
s2200-M. Taylor in TMI-2 control room participated with V. Senaroya 2400 (DSS) et.al. on procedures review and progress toward hook-up of the containment H2 recombiners.
Sunday, Acril 1. 1979 s0200 At TMI Site M. Taylor contacted R. Tedesco (Bethesda) to find out what were latest assessments from Sethesda regarding reactor vessel H, 0, concentrations 2
and explosion potentials. Tedesco advised that most recent Bethesda calculations indicated the following concentrations in the R.V.
(from notes):
l 3
(radiolysis) s 26.5 ft, 09 Hj(radiolysis) siO3 ftj,Hj(Ir-H0 reaction) s61 5 ft,
3 3
s755 ft (approximata bubble size estimate) l l
L
l.
Tedesco also relayed some KAPL info on flammable / detonable concentrations, and he pointed out present estimates indicated that there would be s10 days before explosive concentratier.s would be reached in the R.V.
Taylor recalls Tedesco mentioned some bounding kinds of calculations on the effect of an explosion in the R.V. where explosive pressures on the order of 14,000 psi might be experienced. Taylor recalls suggesting again that HQ-IRC make contact with Dr. Zittel of ORNL for added insights on radiolysis behavior.
l N0800 M. Taylor advised V. Stello as to the following:
1.
H balance estimates were very rough, but these suggested a 2
Zr-H O reaction magnitude in the range of 25-32%.
[ Note these 2
estimates depended on assumptions about R.V. bubble size and on the nature (local vs. free-field) of the H2 deflagration observed at roughly 9-10 hours into the accident.
Review of pressure recorder spikes suggested local as opposed to a free-field l
deflagration in containment.]
2.
Personal judgment was that he ('!. Stello) should not worry about the explosion in the reactor vessel since this was considered to be an exceedingly remote possibility.
(Note that due to the press of shift change activities, the reasoning underlying this personal judgment by M. Taylor was not fully explained to V. Stello.
Reasoning relied importantly on recall of the ORNL radiolytic work done for AEC N10 years ago. This work covered various post-accident coolant chemistries, temperature conditions and system configurations--flowing and static. Importantly, those experiments where basic ph chemistries were involved led to highest H2 yields approaching 7-8 atmospheres (recall) before recombinction terminated the net yield of H /0 7
by radiolysis (i.e., back and forward reactions were equaT).2This suggested no net evolution of 0., should be anticipated at the actual TMI-2 R.V. conditions of 900 to 1000 psi.
Further, the ORNL work with these basic coolant chemistries which were assumed to approximate the TMI conditions where NaOH had actually been introduced early in the accident.
This ORNL work of N10 years ago also indicated an interesting 02 scavenging behavior, i.e., 09 evoluftfonwasnotstoicicmetricwithH.
This behavior further 2
suggested that any free 02 (if it possiell) existed in the bubble space inasmuch as.his might be attributable to the use of non-degassed water from the SWST) could be subject to the same scavenging behavior.
These above fact:rs when taken with the reali:ation that the H2 reducing environment at TMI procably exceeded, by far, the routine reducing environment usually usec in PWRs, plus the need for an ignition source, led to a judgment that an He explosion in the R.V. was of an extremely remote possibility.]
3.
Tedesco's estimates were described.
(
4.
Some possible "what if" scenarics and thoughts were available for discussion when Stello had further time and need to consider these.
s0900 Made phone contact with wife to arrange for more clothing, etc., for extended TMI stay. First learned that the media accounts in Washington had caused wide-spread pubife alarm / concern about H I
2 'XP CSICD-Advised wife not to worry; all was stable and going OK at TMI and that media accounts were probably irresponsibly sensationalized.
mes 4
(-
ENCLOSURE 3 MEMORANDA COMMENTING ON FINAL DRAFT OF CHRONOLOGY l
r jaa neg,
'o, UNITED STATES 4
+[
,' i NUCLEAR REGULATORY COMMISSION
"/
C WASHINGTON, D. C. 20555
,;.U */j E
g, AUG 311979 MEMORANDUM FOR: Roger J. Mattson, Director Lessons Learned Task Force FROM:
Wayne D. Lanning NRC/TMI Special Inquiry Group
SUBJECT:
COMMENTS ON CHRONOLOGY OF TMI-2 HYDROGEN BUBBLE I have reviewed the chronology dated August 30, 1979. My revisions are enclosed.
g.
Wa ne D. Lanning NRC/TMI Special -
iry Group
Enclosure:
As stated above i
l l
I L
l 1m,c2
[fE s Y ** I d
f 1
CORRECTIONS TO 8/30/79 CHRONOLOGY OF HYDROGEN BUBBLE P-2, 3/30, 1600: - Lanning (memory) - requested by Ron Scraggins to join himself, Stan Fabic, and Pete Anderson to go to MNB8 concerning accident at TMI-2.
Briefed by R. Budnitz concerning existence of bubble in reactor vessel. Presented data of pressure versus change in pressurizer level and change in incore temperature measurements. Bubble size was approximately 1000 cubic 1
feet 0 1000 psi and 2800F and increasing. Concerned that core would uncover due to growth. Participated in per-fonning heat-up and boil off calculations for core.
P-2, 3/30, 2000: - T. Murley (INEL notes) requested L. Ybarrondo and others at EG&G, Idaho to ascertain the detonation potential of hydrogen gas in reactor vessel (details of request. unknown).
P-4, 3/30, after 2000:
- W. Lanning and T. Murley initiated Semiscale Test to explore venting of bubble through pressurizer relief valve. Obtained system volumes and pipe sizes from Ed Kane at B&W/Lynchburg.
Obtained current TMI-2 conditions from Tad Marsh at IRC.
Bubble volume in TMI-2 was 1000 ft, pressurized pressure 1100 psi and hot leg temperature 2800F. Noted that pressure surge line had potential for water seal due to geometry layout provided by B&W. Recollect there was a lot of uncertainty concerning size of relief valve opening.
After 2200:
- Reviewed system description and piping diagrams concerning ways to vent bubble. Reviewed make-up and purification system in attempt to resolve blockage of letdown line.
P-6, 3/31, 0200: - Received call from W. Minners who requested a calculation of the explosive force of a stoichiometric mixture of hydrogen and oxygen inside vessel. Assume mixture con-tents for a 1000 cubic foot bubble at 1000 psi which would produce the largest force.
l P-6, 3/31 l
After 0200:
- W. Lanning (memory) requested L. Ybarrondo (EG&G) to perfonn calculation requested by Minners.
P-6, 3/31 l
After 0800:
- W. Lanning (memory) assisted R. Sudnitz in evaluating l
various methods for scavenging hydrogen from primary coolant system. Discussed Minner's recuest with S. Levine, R. Budnitz, and T. Murley.
l l
l l
(
2 P-6, 3/31, 1100: - Monitored progress of Semiscale test.
V. Noonan contacted B. Saffel (EG&G) requesting hydrogen detonation calculation (notes from EG&G details unknown).
P-6, After 1100: - Transmitted preliminary results of Semiscale Test to B&W (BobJones).
I recall S. Levine discussing results with someone at IRC.
Discussed results of test with J. Cudlin (B&W) and calcula-tion performed by B&W.
P-17, 4/1, a.m.:
- Bob Jones (B&W) requested second test in Semiscale.
Suggested doubling size of bubble and include HPI flow based on B&W's contingency plan for depressurization.
Remainder of day spent primarily on coordinating Semiscale test results and second test.
Provided answers to various Labs and individuals who were working on hydrogen explosion, degassing, radiolysis and recombination of hydrogen and oxygen (no notes available).
P-26, 4/2 Midday:
- Coordinated results of second semiscale tests to B&W and IRC.
B&W indicated that they were going to revise HPI flow rates in contingency emergency plan to vent the bubble as the results of tests. Discussed analytical predictions of time period to vent bubble in conparison to Semiscale results.
Relocated from MNBB to East-West Towers. Perfomed calcu-lations of time period to degas using pressurizer spray.
A lot of uricertainty existed concerning flow rates in pressurizer surge and spray lines and makeup and purifi-cation system (coordinated with K. Parzewski).
l l
'h O
crw NOTE T0: Roger Mattson
SUBJECT:
HYDROGEN BUBBLE CHRONOLOGY l
Saturday, March 31 1800 hrs. - Berlinger (Recollection) - Lauben and Berlinger were asked by Stello to provide Stello and Matt Taylor with general informatica on hydrogen / air flammability and detonation limits. Information was transmitted to Taylor and Stello including 4". flammability limit and 8% detonation limit (most probably 16% at TMI conditions).
These figures were off the top of the head best estimates based on our previous experience in combustion processes.
Stello was advised to contact Dr. Bernard Lewis (Pittsburgh, Pa.)
as a known. authority in this area.
2200 hrs. - Berlinger (Recollection) - Lauben and Berlinger (at dinner) discussed the possibility for a hydrogen explosion in the Reactor Vessel. We concluded, based on available information, that the probability of a hydrogen explosion in the Reactor Vessel was very remote since oxygen levels in the RV would have been depleted during cladding oxidation. The effects of radiolysis were unknown at that time.
Sunday, April 1 0900 hrs. - Berlinger (Recollection) - Berlinger was asked by Stello/Vollmer to calculate the change in containment hydrogen concentration which would occur jf the RCS bubble, assumed to be all hydrogen at 1000 to 1500 ft, 875 psia and 300*F, was vented to containment.
The analysis results were transmitted to Stello/Vollmer and left in the NRR Trailer / files. The results indicated a change in hydrogen containment concentration from s1.9". to s3.5%. Since the hydrogen concentration in containment was stabilizing at s2". the possibility of a hydrogen explosion appeared remote, unless conditions changed dramatically due to radiolysis.
__________J
s UNITED STATE 5 ja assoq*
NUCLEAR REGULATORY COMMisslON
[(
wasmcfon. o. c. :osss v.<
g o
t
%....,/
AUG 3 01979 MEMGRANDUM FOR:
File FRG4:
Saul Levine, Director Office of Nuclear Regulatory Research PRINCIPAL CONTACT MADE WITH EXTERNAL CRGANIZATIONS
SUBJECT:
DURING TMI 2 ACCIDENT r
This memo replaces my June 7,1979 meno to File witn corrected and updated information.
I spoke with Robert Ritzmann of Science Applications concerning H 2 generation rates in TMI 2 vessel in the period March 31 2
1.
and 0 He informed me that, although one could calculate an April 1.
increase of 1",0 per day without considerations of a reformation 2
rate of H, and 0, due to bubble back pressure, that the 1 percent rate was Drobably too hign. He also said that he felt the rate was probably no higher than 0.1 percent per day and could be zero, but that he did not have the data to calculate an explicit rate.
- l 2.
I spoke with Dr. Kouts of SNL several times on Saturday and Sunday concerning the possibility of a hydrogen explosion in the reactor vessel. His view as that this would not be likely. He referred me to Harolc Schwar 2 as a source for precise information.
IalsospokewithHaroldSchwark:ofSML,IbelieveonSunday 3.
afternoon, April 1.
He said that he believed that not only wculd tubble, but that whatever H7 and 0, not be accumulating in the H C's were in the ouoble woulo be $epletec at f airly rapid rates.
Ha said that he would calculate tne rates and call me later on Sunday. On Sunday evening he told us tnat he had completed his could not accumulate in calculations which confirmed tnat 02 l
the bubble and tnat it would in fac. be depleted.
I scoke eith James Proctor of the Naval Surface Weaccns Cantar t
l 3
l acout the ef ~ec:s of a nydrogen explosion on vessel integrity.
He saic :ne cylindrical portien of ' te vessel wculd be su::jected I
to acout 5 percent str31n, whicn should not break it, anc that it woulo also be suojected to a lifting force of accut i.5 x ICC ibs.
I He ccuid not calculate wnether tne main lecc :iping could ncid the vessel down when subjectad to this fcrce, since he did not have detailed.tnformation on lant layout.
A copy of Mr. Prec00r's memo is attacned.
4(
W m
Saul Levine, Direc:ar
<- 2 i a
-A-I7WMYM Cffice of Nuclear Regulatory Research