ML20213F762
| ML20213F762 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 11/06/1986 |
| From: | Capra R NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | |
| Shared Package | |
| ML20213F753 | List: |
| References | |
| CIV-PEN, EA-84-137, NUDOCS 8611170145 | |
| Download: ML20213F762 (17) | |
Text
-_
4 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ADMINISTRATIVE LAW JUDGE In the Matter of
)
)
Docket No. 50-320 GPU NUCLEAR CORPORATION
)
(Civil Penalty)
)
License No. DPR-73 (Three Mile Island Nuclear Station,
)
EA 84-137 Unit No. 2)
)
AFFIDAVIT OF ROBERT A. CAPRA k
I, Robert A. Capra, being duly sworn, state the following:
1.
I am a Senior Program Manager, on the Regional Operations and Generic Requirements Staff, Office of the Executive Director for Opera-tions.
A copy of my statement of professional qualifications is attached (Enclosure 1).
The following information is provided in response to In-terrogatories 44 and 45 of GPU Nuclear Corporation's First Set of Interrogatories.
2.
On November 22, 1983, I was involved in an interview with Mr. Richard D. Parks, a former Site Engineer at TMI-2.
The purpose of this memorandum is to explain the circumstances surrounding that interview.
3.
During 1983/1984, the Division of Systems Integration (DSI) within the Office of Nuclear Reactor Regulation (NRR) was involved in reviewing the technical analyses developed by EDS Nuclear Inc. and B&W associated with High Pressure Injection (HPI) actuation at or around 5:41 a.m. on the morning of the TMI-2 accident.
Who terminated HPI, if such an actuation occurred, became known as the " mystery man issue. "
8611170145 861106 PDR ADOCK 0500 0
G
'O.
~
During 'the Office of Investigations' (OI) investigation of the " mystery man issue," Mr. Parks made allegations that the analysis presented in the GPU v. D&W trial by EDS Nuclear Inc. associated with HPI actuation at 5:41 a.m.
was not correct.
OI documented these allegations in.
4.
Because of NRR's involvement in the technical review of this subjcet, an interview with Mr. Parks was arranged to discuss these alle-gations.
The interview took place at NRC Headquarters in Bethesda, Maryland on November 22, 1983.
To the best of my knowledge the per-sonnel present during the interview included:
Mr. Brian Sheron (then the Pcactor Systems Branch Chief), Mr. Norman Lauben (then a Section Leader in the Reactor Systems Branch and principal technical reviewer of this issue) and me (then Technical Assistant to the Director DSI/NRR).
Although I do not recall for certain, it is also possible that Mr. Parks l
was accompanied by his personal attorney (name not recalled) during the interview.
5.
After reviewing our files on October 22, 1986, we did not dis-cover any formal " summary" of Mr. Parks' interview.
The only surviving document that may represent " notes" from this meeting is a single page of handwritten notes written by Mr. Lauben.
See Enclosure 3.
Based on l
my conversation with Mr. Lauben, it cannot be confirmed that these notes were in fact taken during the meeting with Mr. Parks.
If other " notes were taken, they no longer exist.
To the best of my knowledge, the i
meeting lasted approximately one hour. All of the questions asked during the interview were related to either Mr. Parks' concerns regarding the validity of the EDS technical analysis or to the basis for Mr. Parks' l
l
aliegation that he " knew" there was a " mystery man" and who it was.
The results of what Mr. Parks told us during the interview were consid-ered in NRR's final technical evaluation of the issue (Enclosure 4).
6.
I have discussed this matter with Messrs. Sheron and Lauben and the foregoing information is true and accurate to the best of my knowledge and belief.
Y f~-
Robert A. Capra
/
Subscribed an sworn to beforn me this M day of November,1986 k'
Sk" s
t Fotary Public My commission expires:
//f/90
l Enclosure 1
ROBERT A. CAPRA My name is Robert A. Capra, I presently reside at 1202 River Bay Road, Annapolis, Maryland 21401 and am employed as a Senior Program Manager on.the Regional Operations and Generic Requirements Staff in the Office of the Executive Director for Operations, U. S. Nuclear Reguhtory Commission, Washington, D.C. 20555.
PROFESSIONAL QUALIFICATIONS I received a Bachelor of Science Degree in Marine Engineering from the U.S.
Naval Academy in 1971.
I have also taken courses in PWR and BWR Systems Operations and at the B&W plant simulator in Lynchburg, Virginia.
I enlisted in the U.S. Navy in July 1964 and served in that capacity for three years. During that time, my duties included attending the Naval Enlisted Nuclear Power School, Mare Island, California followed by subsequent study and qualification as an operator and staff instructor on the Navy's
" DIG" reactor located in West Milton, New York.
Following my enlistment, I attended the U. S. Naval Academy where I graduated in June 1971 and was comissioned as a line officer in the U. S. Navy.
~
Additional graduate level studies in nuclear reactor theory, thermodynamics, electrical engineering, radiation control and other related engineering courses were completed in 1972 at the' Navy's Officer Nuclear Power School, Bainbridge, Maryland.
I subsequently returned to West Milton, New York where I studied and qualified as Engineering Officer of the Watch.
From 1973 to 1976, I served aboard an operating nuclear submarine, during atch as Officer of the Deck, which time my duties included standing w#ecting, training and supervising Engineering Officer of the Watch, and d technicians in the operation, maintenance and repair of various equipment and systems primarily associated with the ship's nuclear reactor. During this period, my assignments included supervision of the Operations Department, Reactor Controls Division, Electrical Division, Main Propulsion Division and the Chemistry and Radiological Control personnel. In addition, I qualified as Engineer Officer for the supervision of operation and maintenance of Naval y
Nuclear Propulsion Plants.
%From 1976 to 1978, I was assigned as a Company Officer at the U. S. Naval Acad,,my where my duties included supervising, directing and evaluating the e
training and activities of 130 Midshipmen Since July 1978, I have been employed with the NRC as a Licensing Project Manager; Operating Reacter Project Manager; Standardization Project Manager; Technical Assistant to the Director Division of Systems Integration; and as a Seni'or Program Manager. During this period, my duties have included:
coordinating the licensing review of two Construction Permit applications and two Operating License applications; serving as the Project Manager for all B&W operating plants on the post-THI accident Bulletins & Orders Task Force; serving as the Project Manager of the B&W Reactor Transient Response Task Force: serving as the Technical Coordinator of the GPU vs B&W lawsuit review; providing staff evaluation of THI management integrity issues; providing testimony for the THI-1 and Rancho Seco Restart and the Shoreham licensing proceedings, and serving as a member of the joint 01/NRR team investigating reactor coolant system leak rate practices at TMI-2.
1 Enclosure 2
[
\\g UNITED STATES E'
~h NUCLEAR REGULATORY COMMISSION M
g WASHINGTON. D. C. 20555
'. Si.'.... /
y '4 October 28, 1983 g\\g,
6L fiEMORANDUM FOR:
Harold R. Denton, Director j'.
Re p j% g lition j
FROM:
Be aeY,^01 Office of Investigations
SUBJECT:
EDS NUCLEAR INC. ANALYSIS OF THE HIGH PRESSURE INJECTION PUMPS Laurence P. King, former Site Operations Director at Three Mile Island, Unit 2 (TMI-2), and Richard D. Parks, former Site Engineer at THI-2, have made allegations that the analysis presented in the GPUN/B&W trials by EDS Nuclear on the TMI-2 High Pressure Injection (HPI) Pumps activity is not correct.
Messrs. King and Parks have stated that they would be willing to explain to NRC the weaknesses in the EDS analysis and show from the analysis that the HPI pumps were turned on and subsequently turned off at critical time periods.
Mr. King did not provide further infomation or details on the EDS analysis.
However, a summary of Mr. Parks' assertions on the EDS analysis is included as an attachment to this memo.
Per the suggestion of. Mr; William T. Russcll, Deputy Director, Division of Human Factors Safety, this memo and its attachment are being sent to you so that it can be pa's' sed to your group considering the EDS Contact point in my office is Ronald A. Meeks, 492-7246.
matter.
The 01 investigation on the " mystery man" issue is ongoing; therefore, the contents of this memo and the attachment should be disseminated on a need to know basis among those assigned to the EDS review.
- ?
Attachment as ststed z.
i 9
.e RICHARD D. PARKS' OVERVIEW OF THE EDS NUCLEAR INC. ANALYSIS OF THE HIGH PRES 5URE INJECTION ACTIVITY DURING THE TMI-2 ACCIDENT Richard D. Parks raised the " mystery man" issue, in part, because he was con-cerned that the causes of a key event during the accident may not have been fully disclosed to the public, the NRC, and later in the GPUN/B&W trial.
Mr. Parks knew that at the trial an anlaysis from EDS Nuclear, Inc. was intro-duced to conclude that the HPI pumps were not turned on at 5:41 A.M.
According to Mr. Parks, the EDS analysis had unproven, incomplete and inaccurate assumptions, which make it especially questionable as a basis to effectively rewrite the history of Sequence of Events (SOE) already published by GPU and, to Parks' knowledge, not yet amended.
Nr. Parks' own review of the EDS report suggests that its conclusions are pre-mature but access to the relevant supporting data is necessary for more definitive conclusions. Mr. Parks' evaluation of the EDS report is based on the adequacy of its premise. The analysis basically rests on the premise that the level of the makeup tank also determined whether the second HPI pump, makeup pump "c",
was initiated at 5:41.
EDS analyzed and prepared graphs for the makeup tank levels surrounding two times that it concluded HPI pumps were initiated, at 4:02 and'7:20 as well as for 5:41. The' graph for 5:41 did not exhibit the same characteristics.
Although the EDS analysis is reasonable, it's signifi-cance is limited due to the selective use of data and unjustified assumptions, all biased in favor of the EDS premise.
i According to Mr. Parks there are a series of(questionable factors with respect to the EDS claim that turning on the HPI purnps at 4:02 accounts for the graph in its report.
In some intstances the facts as reported are suspect. For s
example, EDS' asserts that an operator-started MVP 1-A and opened MU-VI6b at the.
y".
time of the reactor trip. But the SOE reports that the evolution did not occur until 41 seconds after the reactorTrip, or 4:01:31 A.M.
Similarly EDS offers no citation for its claim that the BWST suction valve, DS-VSA, opened at 4:02:13., This assumption also is not verified in the SOE.
u t o qui-q In fy t, the Emergency Core Cooling System (ECCS), which' includes the HPI pumps, did not have to be turned on at all at 4:02 in order to produce the characteAstics on the EDS graph. Alternate sources better explain the Q
/s 1
~
increase in. pressure.
EDS did not analyze whether the increase in pressiire at j
4:03 could have~come from opening the MUT Level Control Value (MU-V '9) to refill the tank from the Reactor Coolant Bleed Tanks (RCBT), instead of from the HPI pumps.
This explanation would be consistent with the requirements in i
emergency. procedure 2202-1.3 for operators.
In other words. EDS failed to consider the possibility that the operators followed the procedures at 4:02.
Similarly, EDS failed to consider whether at 4:02 water was drawn automatically through the makeup level control valve, again as an alternative to the HPI assumption.
l These alternative explanations are even more persuasive when the EDS assumptions are compared with the capacity of the HPI pumps and the data provided by EDS.
1 8
l
'g JLu ep/
The EDS graph discloses a four inch rise (approximately 280 grams per minute) in the makeup tank during the 26 seconds from 4:02:13 - 4:02-39. That is a substantially higher rise than is achievable solely from HPI pump recirculation flow.
Finally, even if EDS' assumptions and facts were accurate, the results demon-strate that conditions were not analogous at 4:02 and 5:41.
For instance, different pumps would have been involved.
At 4:02 the A&B pumps would have been used.
At 5:41 the A&C pumps would have been running.
The C pump can only draw from the BWST.
Even more revealing, the 4:02 incident would have involved switching on two pumps. The 5:41 event would only have involved switching on the C pump.
7,-
}
Mr. Parks relates that the EDS analysis of events at 5:41 is no more definitive.
To illustrate, in some cases the analysis is incomplete.
On page 12 of its report, EDS states thatiupon ECCS actuation DH-VSA was opened.
On page 13. EDS states that the injection valves MU-V-16a and B also opened.
Unfortunately, the report failed to identify when they were shut after 4:02 and plot the effects on a graph.
In addition, the EDS analysis concerning 5:41 concludes that the HPI pumps could not have been turned on at 5:41, since the makeup tank level was only at 59".
However, EDS failed to analyse the special circumstances then in effect, such as increased letdown flow, lowered reactor coolant system pressure, and higher temperatures than normal.
Most persuasive, EDS' conclusions about 5:41 contradict other data in its report.
On page 16 of the report EDS stated that a makeup tank pressure of 39 PSIG would have been necessary to achieve the 59" level.
But figure 14 in the same report indicates that the 59" level as pchievable with 31.5 PSIG.
Finally, the EDS analysis of conditions at 7:20 assumes an initial makeup tank pressure of 27 PSIG only nominally higher than nonnal.
This assumption fails.
to analyze the severe effects of the accident; which between 5:30 and 7:20 had led to superheated steam and significant portions of the reactor coolant system in a steam-void condition by 6:00 A.M., to two-thirds of.the reactor core uncovered by around 6:20, to hydrogen generation and half the reactor coolant system free volume in a steam hydrogen mixture by around 7:00 A.M.
These 5
condRipns could well have led to abnormal overpressurization due to steam and/or hygrogen in the top of the makeup tank, a possiblity not considered by EDS.
Fnclosure 3
~
9v A
'22.0l-3 2.2..H ET))
IkM unch tvud ed ak Ad4P s~uce
-L c-+A<l N
.{.)
-Q_
bleed hsILp LA<
led M ?
on prav qn-erJ k r k uwb Le LJ w>
e +).
hoiJic' 1< H t4-
~ la~p sv L H~
ua l<s p (J W
,J.u y
I;% w% -
om
%U G
h.J L
+u;a l* L
.A4~ h 'l NLO.,p bk 3
l u sg L m tal
~L-g Lk i
u o n< > m.A o 4 -
d v.A /
/
P;4
$ 6%
d HP,L u6,.h wu IG vrim
+4 alhd f u fd.,P2 L Weu M y S-)ph a l e.>3 I/ b's )
$ U of /.
W9 h Ac usflc4 uds k
4 ks)Qo
- 0. v % h w p ~h l.'
e pas W J v A d l+di n,L,ck yl&
cJ pdJL.
84nT /d e
kew To6 os9
0' APR 2 4 1984 1,.
HENORMDUH FOR: Ben 8. Hayes. Director Office of Investigation FRfti:
Harold R. Dentnn. Director Office of Nuclear Reactor Regulation
SUBJECT:
NRR EVA!.UATION OF HP! ACTUATION AT THI-2 ABOUT 5:41 A.H. ON HARCH 28. 1979
Reference:
1)
Analysis of Reactor Coolant System Hake-up During the TNI-2 Event, by EDS Nuclear, Dececher 29. 1982.
k 2)
Response to GPUN Questions Concerning HPI Actuation at Tit!-2 about 5:41 a.m. on March 28. 1979 B&W UPGD.
3) liigh Pressure Injection Hear 5:41 a.m. During the TNI-2 Accident by C.B.. Davis EGaG Idaho. Inc..
Decedxtr 1983.
NRR has evaluated references 1 and 2 to detemine the extent to which they support or refute whether full or partial HPI actuation occurred on or about 5:41 a.m. during the THI-2 accident. We contracted with EGAG Idaho to assist us in this effort. The result's of that work are contained,in rgference 3.
We also interviewed Mr. Richard D. Parks forner Sit'e En acer at THI-2 to detemine if he could identify specirte technical info tion which could be used to assist us in our
~
evaluation.
It is ourgonclusion that actuation of HPI issnedtately after the last l
reactor coolant pusios were tripped at 5:41 a.m. is extremely unlikely.
[
Analysts and evaluation of the information contained in references 2 and 3 indteate that cooling was in progress and _ Partial _ actuation prior to that tine:-(at about 5:34 a.m.) for a short time (about 6 minutes) is possible. However due to (1) the unavailability of the alarm printer at i
thetine.(2)uncertaintyinsystemboundarycondttionsand(3) uncertainties in analytical modelling techniquest it is not possible to conclusively affirm or reject limited HP! actuation teveediately prior to 5:41 a.m.
0405iAn'A q q W-3 ADOCK 0,noco3pg CF 1
l I
)
em, l
6
}
._.__--___--__m-
Ben 8. Hayes 02 We have enclosed refersnee 3 for your information. Our evaluation is also enclosed.
This completes what we can learn from technical analysis of the data from the accident. Further insight concoming HPI actuation. if needed.
would have to come fmm the ongoing investigation of the testimony of individuals involved in the accident. This view was previously discussed with the Cocaission on March 23. 1984 The April 2.1984 memo from Mr. Chilk confIrus the Cosetsston's intentions in this regard.
06 mal s e s er t
e W. 4. hetae Harold R. Denton. Director Office of Nuclear Reactor Regulation Enclosures :
As stated cc:
D. Eisenhut R. Meeks. O!
K. Christophers. O!/RI W. Dircks V. Stello Ols tribution I. Central, Filet
.J RSS R/F' RSD S/F TMI N. t.auben
- 8. Sheron R. Capra R. W. Houston l
R. Mattson 7
W. Russell 4 G. Case it. Denton n
1
- 0RIGINAl. CONCURRENCE TO BE RETAINED BY
/ RETYPED IN R. MAjTSON'S OFFICE 4/23/84
- 0..
- 0$l:RS8
- 051 I *AO:R$...............%.f..DilFS :NR....... I
..N.R.R........
....Y.......
- DSI : DI p
NLauben/S$hgr00,,8Capre..
. PWous t00.... D84itson..... WRus s ell D,.dC41t....d...
1 3/23/84 3/24/84 3/ 27f 84 3/28/84 3 4/2/84 / W l ' o i n
.o. >
f EVALUATION REPORT ON EDS NUCLEAR REPORT AND B&W REPORT f
CONCERNING HPl ACTUATION AT THI-2 ABOUT 5:41 A.M. ON HARCH 28, 1979 f
1 h
i 1.0 Introduction NRR with the support of EGaG Idaho has reviewed references 1 and 2 I
i to determine the extent to which they support or refute full or partial HPl actuation on o* about 5:41 a.m. during the TM!-2
{
3 accident. EG&G (Ref. 3) also performed independent sensitivity
?
calculations to evaluate uncertainties in analyzing the available i
1
- data, j
i I
2.0 HP! Actuation Near 5:41 a.m.
i t
i e
{
The CDS report (Ref.1) examined mkeup tank level behavior at 5:41 i
a.m. to determine if the response exhibited characteristics of ECCS 3
actuation.
EDS neither addressed actuation at any other time, nor i
/
did they examine other data which might be affected by ECCS
- i
+-
actuation.
Two hydraulic models were used to predict makeup tank 4) 1 level based on certain assumptions regarding the makeup systems and j
the Borated Water Storage (BWST).
The two models were benchm rked
,b i
against, data taken at times when ECCS actuation was known to occur.
i T
Because alarm printer data was not available between 5:14 a.m. and f
S 6:48 a.m. it cannot be known with certainty if an ECCS actuation cccurred around 5:41 a.m..
EDS applied the benchmarked model to i
the 5:41 time period assuming a constant letdown flow of 62 gpm.
T:
5
Two different makeup tank pressures were used.
Both calculations indicated a sub'.tantial surge in flow from the makeup tank would occur with an eventual balancing of the make up tank and the BWST hydrostatic heads. Actual makeup tank data did not indicate this behavior.
Instead a continuous drawdown of makeup tank level was observed from Sj35 a.m. to 5:53 a.m..
Based on this, E05 concluded that no ECCS actuation occurred at 5:41 a.m..
Mr. Richard Parks, former THI-2 site engineer, suggested other possible sources of water to the makeup tank but was not aware of other specific technical date that was available.
In particular, he suggested that the letdown wy have been circulating through the bleed hold up tanks and that omergency boration couH have been occurring throug.. the makeup tank.
Ne do not believe that changes ik the makeup or letdown flow prior to actuation would alter the conclusion reached by EDS. A distinctive signature of makeup tank behavior would be expected during HPI actuation uhless compensating
[
changes in makeup and letdown flow were made at the same time.
That would seem to be all too complex a maneuver in light of everything else going on in the control room at that time.
In. referenct 2, B&W perfomed mass and energy balances on the system af ter 5:41 a.m., assuming various modes of HP! actuation, in order to predict system pressure.
I For all,HP1 modes analyzed. the B&W calculations indicate that a r
more rapid primary system depressurization would have occurred then was observed.
Since the calculated pressure behavior did not match the data. B&W concluded the HPI pumps were not actuated after 5:41 a.m. when the last reactor coolant pumps were tripped. After reviewing references 1 and 2. EG&G also concurred with a high degree of certainty that HPI was not actuated at or inmediately after 5:41 a.m..
NRR concludes there is no basis to conclude that i
HPI actuation occurred at or immediately af ter 5:41 a.m..
t 2.0 Analysis of Coo'.down Between 5:34 and 5:40 a.m.
I t
Prior to the reactor coolant pump trip from 5:34 a.m. to 5:40 a.m.
a reactor system cooldown was observed, characterized by falling reactor coolant system temperatures and pr, essure. EDS Nuclear did not address this time period. Based on an analysis of the EDS report. EG&G has concluded that EDS would probably have concluded j
that no HPI actuation occurred during this cooldown either. B&W.
however, did analyze this cooldown by examining three possible e
ve.
scenar;jos to account for this behavior.
The three scenarios assumed the cooldown was caused by (1) HP! only. (2) Emergency feedwater (EFW) only, or (3) a combination of HPI and EFW, B&W used much of the available data to evaluate the three scenarios.
8 3
t
' 2.1 Scenario 1: HP1 Only One important feature of this cooldow, is that it began about li minutes prior to an observed decrease in makeup tank level. Thus, if HPI only (scenario 1) was responsible for the cooldown, HP! was not aligned to the makeup tar.k at the beginning of the cooldown (5:34 a.m.).
To account for this, B&W initiated HP! in scenario 1 at 5:34 a.m. with HPI pump "C" not aligned to the makeup tank. At 5:35:30, HP! pump "A" was allowed to start and draw from both the BWST and the makeup tank. Makeup and letdown flow were adjusted frequently during the cooldown analysis, so that the resulting makeup tank level response matched the data. The required letdown and makeup flows show a large variation but these flows would not be impossible to achieve. EG&G noted that to match the data the B&W analysis assumed HPI pump "A" actuation and simultaneous increase in letdown flow at 5:35 a.m..
T.hese would normally be conflicting operator actions.
B&W performed an energy balance calculation to determinb what HP!
[
flows would be required to match the observed cooldown. The l
At*
- analysis required more HPl flow between 5:34 and 5:35:30 than was required to match the makeup tank data.
BLW also reviewed other data (RC flow, SRM signals, BWST volume, and RCS inventory) and concluded that while the scenario involving only the HP! is possiblv, it is unlikely.
a
- - - - - - ' - ' - - ~ - ~ ' ' ^ ' ^
~~
~^ ^
^
^^
L
~
Based on'the evidence presented, NRR has concluded that the scenario of HPI only is the least likely explanation for the RCS cooldown between 5:34 and 5:41 a.m..
2.2 _ Scenario 2: EFW Only Based on heat removal requirements, B&W concluded that the capacity of one motor driven EFW pump would be more than sufficient to cause the observed cooldown. Flow at 360 g.p.m. would be required, whereas the capacity of one EFW pump is about 500 gpm. B&W compared the steam generator level and pressure behavior during the cooldown to a similar condition when EFW was initiated into a dry steam generator at 4:08 a.m..
The behavior in both cases was simi.lar, indicating that significant heat removal can occur before o
the, secondary level begins to rise iri the steam generator. Thus, the cooldown comencing at 5:34 a.m. is consistent with EFW initiaticn at that time and with the delay in secondary pressure and the level rise in the "A" steam generator.
W.
t.
B&W analyzed the response of the "A" steam generator using the AUX-!! code. The same delay in pressure and level rise was ob-served, at 4:08 a.m. and 5:34 a.m..
In order to match the pressure a.1d 'evel data reasonably well, an EFW flow of only 200 gpm was l
required. However this would not be enough to cause the cooldown andwouldrequiretheatmosphericdumpvalve(ADV)tobealmost e
5
wide open. B&W suggests that this indicates that a more likely scenario is that in addition to EFW, some HPI was being added between 5:34 and 5:40 a.m. (Scenario 3).
EGaG determiied that B&W had not considered the cooling effect of makeup and letdown flow.
Based on a series of RELAPS calculations with simplified nodalization. EG&G estimated that an EFW flow rate as low as 220 gpm would be ad1guate;to cause the cooldown if a reasonable amount of makeup and letdo n we e assumed. Since this is fairly close to r
9 the 200 gpm detemined b3 the AUX-II analysis EG&G concluded that the EFW only scenario u somewhat more likely than indicated by B&W. EG&G's analysis of calculated uncertainties suggests that a fully open ADV may also not be needed to support this scenario as suggested by B&W. B&W also concluded that the assumption of only EFW only is not inconsistent with most of the other available data.
NRR has, reviewed both the B&W and EG&G ana'1'ysis and has detemined f
that the 'EG&G analysis accounts for makeup and letdown in a
[
reasonable way. We therefore conclude that the EFW only scent.rio is somewhat more Itkely than indicated by B&W, w.
h 5
2.3 Scenario 3: Combination of EFW and HPI Neither B&W nor EG&G perfomed specific analyses for scenario 3.
B&W believes that this scenario addresses the main questions raised by the other two. That is, the added inventory of scenario 1 6
seemed somewhat inconsistant with the data, and the apparent requirement for a wide open ADV in scenario 2 is eliminated in scenario 3.
EG&G agreed that scenario 3 is the most likely scenario.
3.0 Additional Analysis Because of the uncertainties in boundary conditions and in analyt-ital models. EG&G did not recommend additional analysis. NRR agrees that additional detailed analysis would riot be worthwhile.
I 4.0 Conclusions l
l
+
Based on the material presented in th'e three reports, there is no clear evidence of HPI actuation at any time between 5:30 and 5:50 a.m..
However some partial HP! actuation cannot be ruled out between 5:35:30 and 5:41 a.m. while the RC pumps "where" still M* tunning. The evidence is strong that af ter 5:41 a.m. EFW Flow to steam generator "A" was the only major heat removal mechanism causing the cooldown until 5:50 a.m. when hot leg temperature began to increase. EFW may aise have exclusively been responsible for the kooldown prior to reactor coolant pump trip. However NRR believes, that it is equally likely that some l'PI, although throttled, may have contributed to the cooldown between the 5:34 and 5:40a.m.
7