ML20028C944
| ML20028C944 | |
| Person / Time | |
|---|---|
| Site: | Indian Point  |
| Issue date: | 01/12/1983 |
| From: | Holahan G, Rowsome F Office of Nuclear Reactor Regulation, NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | |
| Shared Package | |
| ML20028C925 | List: |
| References | |
| ISSUANCES-SP, NUDOCS 8301140332 | |
| Download: ML20028C944 (17) | |
Text
UNITED STATES OF AMERICA NUCLEAR REGULATORY C0tm!SSION ATOMIC SAFETY AND LICENSING BOARD In the Matter of Docket Nos. 50-247-SP m
50-286-SP CONSOLIDATED EDISON COMPANYOF NEW YORK (Ind POWER AUTHORITY OF THE STATEOFNEWYORK(India R0WSOME,III TESTIMONY OF GARY M. HOLAHAN AND FR ON BOARD ObESTION 2.2.1 m
ith the NRC.
Mr. Holahan, state your name and positions wAssessment Q.1 Gary M. Holahan, Section Leader, Safety i i n of Licensing, NRR Operating Reactor Assessment Branch, Div s o A.I f ssional qualifications?
Have you prepared a statement of your pro e is attached.
i Yes, a copy of my professional qualificat ons Q.2 A.2 ding the generic study Q.3 What are your review responsibilities regard the India of steam generator tube rupture events an982 which iden and 3, Board Order of November 15, 1 for this proceeding?
ff's on-going generic I am a principal contributor to the NRC Staand tube ruptu A.3 i
study of steam generator tube degradat on i k analysis assoc I am the principal author of the generic r sadapted the gen with the above-mentioned study and I have
[
analysis to Indian Point Units 2 and 3.
1 0301140332 83011205000g PDR ADOCK T
- _ - _ _ -. _ _ _ _. l Mr. Rowsome please state your name and position with the NRC.
f Q.4 A.4 My name is Frank H. Rowsome, III.
I am Deputy Director of the 7
Division of Risk Analysis in the Office of Nuclear Regulatory I
Research.
I Have you prepared a copy of your professional qualifications?
f Q.5 Yes, a copy of my professional qualifications is attached to this t
A.5 I
testimony.
Q.6 What is the purpose of this testimony?
j The purpose of this testimony is to address Board Question 2.2.1.
f A.6 This question provides:
Should any of the requirements proposed at the July 29, 1982 l
I meeting of the NRC Staff and members of the SGOG be required for Indian Point Units 2 and/or 3, considering the risk of a steam generator rupture in this high population area?
[
l In addition, this testimony describes the basis for the Staff I
evaluation of the contribution to core melt likelihood provided by l
l steam generator tube ruptures in the Staff's testimony on f
Commission Question 1, and addresses the safety significance of a f
tube rupture as it pertains to the Staff's response to Contention l
2.2(a).
j What is the concern relative to steam generator tube rupture?
Q.7 The concern relative to steam generator tube degradation stems from f
A.7 the fact that the steam generator tubes are a part of the reactor t
coolant system (RCS) boundary and that tube failures result i In addition, the steam generator tubes loss of primary coolant.
i constitute a particularly important part of the RCS boundary s ;
their failure allows primary coolant into the steam generators a
d The where its isolation from the environment is not fully assure.
release of primary coolant into the environment has two major s
l The first is the direct release of safety implications.
s f cooling radioactive fission products, and the second is the loss o An extended loss of water which is needed to cool the core.
l i
cooling water outside of containment would result in the dep e ility of the initial RCS inventory and ECCS water without the capab A inside to recirculate the water as would be the case for any LOC The major safety concerns are, therefore, related the containment.
s to events in which a loss of steam-generator tube integrity occu There in combination with a loss of secondary system integrity.
are two classes of events with these characteristics:
tube ruptures followed by a loss of coolant through the (1) secondary system due to failure of the SG safety or relief valve (due to valve failure or operator error) or main steamline failure; and a secondary system failure (steam line break, feedwater (2) line break or stuck open valve) with a consequential failure of previously degraded steam generator tubes.
Have you analyzed the probability of a core melt associa Q.8 steam generator tube ruptures?
_4-
~ Yes, I am working on a generic study of core melt probabilities A.8 from events involving steam generator tube ruptures and have adapted the results to the Indian Point plants.
Have you developed an estimate of the probability of a core melt I
i.
Q.9 from accidental sequences associated with steam generator tube
{
ruptures?
!~
The total core melt probability from all causes involving steam A.9 generator tube failures is estimated to be 4 x 10-0
/RY. Fifteen These sequences address different accident sequences were studied.
single and multiple tube ruptures, steamline failures, steam f
generator safety valve failures, large Loss of Coolant Accidents, loss of all feedwater, loss of RHR, Anticipited Transients Without Fault-trae analyses were Scram and various operator errors.
performed to establish the sequences leading to core melt and to In this analysis three estimate the probability of those sequences.
important areas needed to be addressed; the initiating event p bility; the probability of component failures; and the probability The event probabilities and component failure of operator errors.
probabilities that would lead to a core melt involving SG tube failures were chosen based on, or extrapolated from, the operati The operator error experience of all PWRs and are generic values.
probability for the most important operator actions (failure to isolate the faulted steam generator and failure to depressurize th reactorcoolantsystem)andsystemreliabilityassessmentsare i
,____.l._
[.
drawn from NUREG/CR-2934 on which testimony will be prov Comission Question 1.
?
Q 10 Describe your contributinn to the staff analysis of the
. m by steam generator tube rupture at Indian Point Units 2 an
[
6 A.10 I provided an estimate of core melt probability of 2 x 10 /R t
e This infoma-associated with steam generator tube rupture events.
t m tion was then used in estimating the risk associated with s ea As a result of continuing review generator tube rupture events.
his this estimate has been modified to the values presented in t testimony, that is, 4 x 10-6/Ry.
Q.11 Explain why this analysis is applicable to Indian Point it A.11 To check the applicability of the generic analysis to Indian Units 2 and 3, the important elements in the analysis were For the generic analysis, the largest contributors to reviewed.
risk were found to be:
multiple tube ruptures combined with failure of a steam (1) generator safety valve; and an Anticipated Transient Without Scram resulting in (2)
\\
multiple tube failures.
l The important elements in the multiple tube rupture and safe the multiple tube rupture probability; ECCS failure scenario are:
t capacity;RefuelingWaterStorageTank(RWST) size;steamg
6-size; and safety valve failure probability. These parameters have been checked and found to be applicable to the Indian Point plants.
Although there has never been a tube rupture at either of the Indian Point units, the steam generator design and operating experience are typical of other plants. The generic value for the probability of a tube rupture is therefore applicable. The generic analysis assumed the RWST contains 300,000 gallons of water while the minimum value for Indian Point 2 is 345,000 gallons. This difference would result in an increase of approximately 15% in the time available for operator action. However, the capacity of the high pressure and low pressure injection pumps is about 10% greater for Indian Point than in the generic analysis; and this would result in approximately a 10% decrease in the time available for operator action time. The net difference of approximately 5% in the time available for operator action is insignificant in assessing the probability of operator The Indian Point units use the same steam generator type errors.
(Westinghouse Model 44) as was studied in the generic analysis.
Finally, the Indian Point plants use spring loaded steam generator safety valves which are comparable in design to other plant safety valves. The generic failure probabilities are therefore applicable.
For the ATWS plus multiple tube rupture scenario', the probability of an ATWS (i.e. reactor protection system failure) and the probability of tube failures during an overpressure event are the important elements. The generic values studied are appropriate for Indian Point 2 and 3 for this scenario as well.
2- - - -. ----- -
--m----
,-Q-
,-p r---
.,-,m._-,,-w
,m---,-,
,,--mn.,,an-,,,.,-g r-
- I Q.12 Mr. Rowsome, how important is steam generator tube rupture to the overall risk posed by severe reactor accidents at Indian Point Units 2 and 3?
A.12 Core melt accidents entailing steam generator tube rupture represent This estimate roughly one' percent of the offsite radiological risk.
is applicable to both Indian Point Units 2 and 3.
Q.13 How did you arrC'e at this estimate?
A.13 The overall core melt frequency for each Indian Point unit, as they are designed and are (or will be) operated in 1983, we estimate to be roughly 3.6 x 10~4 per reactor year. The basis for this estimate will be presented in the staff testimony on Commission Question 1.
Gary Holahan's estimate that core melt events associ-ated with steam generator tube rupture have a frequency of 4 x 10-6 This represents roughly one percent of the over-per reactor year.
all core melt frequency. Neither the overall core melt frequency nor the portion attributable to steam generators tube rupture events is known with precision, but it is my judgment that the fraction of all core melt events attributable to steam generator tube rupture probably lies between 10% and one tenth of one percent of the total, with 1% our best estimate.
The magnitude of the radiation release associated with core melt accidents involving steam generator tube rupture is neither one of I
the most severe nor one of the most benign. There is not one but a On average the spectrum of releases possible for such accidents.
l l
w l of radiological severity of such accidents I believe to be typica n
Therefore I. infer that the spectrum of other core melt accidents.
the proportion of offsite radiological risk attributable to steam 1%
generator tube rupture / core melt accidents is the same -- abo k
The subject of uncertainties
-- as the proportion of the accidents.
[
in the risk analysis will be taken up again in the testimony on 2
1 Commission Question One.
Q.13 please elaborate on how the estimates are ma.de of radia attributable to core melts entailing steam generator tube rupture, for use in answering Commission Question One.
A.13 The principle path by which fission products can escape from plant in core melt accidents entailing steam generator tube rup is through the ruptured tube (s) and out to the atmosphere throug lief or the affected steam generator, the main steam line and its re No formal, computerized analysis has been made of safety valves.
timing or quantity of fission product releases for such scenarios.
Instead, a qualitative analysis has been made of the course of these accident scenarios, and the release category assignment h In the more likely scenarios, the pressure i
been made by analogy.
in the reactor coolant system drives reactor coolant through the rupture (s) in the steam generator tubes into the faulted steam The emergency core cooling system replenishes the lost generator.
reactor coolant until the tanks upon which the ECCS system draw I
Usually the operators will have cooled and are depleted.
l depressurized the reactor so that the residual heat removal s l
m v.--'
g
.can be employed to sustain core cooling. If, however, the operators have been unable to achieve this, the decay heat in the reactor core will begin to boil off the remaining reactor coolant until the core uncovers and begins to overheat.
In the event that the faulted steam generator remains full of displaced reactor coolant, the steam, hydrogen, and fission products emerging from the overheating reactor will percolate through this water in the steam generator on its way' to the atmosphere. This percolation is expected to be an effective
- filter, so that the atmospheric release would be quite modest. The release is roughly what one would expect of a core melt in a containment equipped with a controlled, filtered vent.
In the event that the faulted steam generator drains back into the reactor, then we lose the benefit of the water as a filter, but gain the benefit of the water as reactor coolant.
It would take several extra hours for decay heat to boil off the water draining back into the reactor before core melt would commense (or resume).
The extra time would be available to the operators to establish long term core cooling.
Among the possible outcomes of such a sequence of events remains the possibility of a late core melt with the faulted steam generator dry. Even dry, the steam generator provides fairly good filtration.
The steam gent.rator tube bundle and the moisture separators provide
, le surface area upon which a long, convoluted release path with amp mv fission products will plate out.
t d of these scenarios is The release to the atmosphere to be expec ef a core melt w t
roughly analogous to that expected o fan coolers and/or sprays to the containment,with containment In other words, the tion.
.j working, but an open containment penetra ect of the " leaking co i
release is similar to what one would exp d t tainment" model of a core melt acci en.
d for the " leaking A fission product release model was prepareStudy, and it has t
containment" case in the Reactor Safe y However,because Point case.
adapted for the specifics of Indianrupture release is by b
our analysis of the steam generator tu e cenario-specific,calcula-analogy, rather than by plant-specific, se in a more pessimistic l
tion, we have chosen to model the re easrtantly underestimate d
way, to be sure that we have not ina vein o Commission of core melt sup-consequences.
Question One, the early version of the frequencyr) haf 0
plied by Gary Holahan (2 x 10' / unit yeadeveloped for late contributers to the release category ontainment leakage h
failure of containment, rather than to t e ctly gives substa t
release cateogry, as this choice consis >n This release estimate s
l greater offsite radiological consequence. t steam generat l
is consistent with my earlier statement tha i
a
,p g
7-y-,
,e-w--g
- 11 rupture accidents leading to core melt represent roughly one percent of the overall offsite radiological risk.
i The risk analysis to be supplied in the Staff testimony on Connission Question 1 has not been revised to reflect Gary Holahan's current estimate of 4 x 10-6 rather than the original estimate of 2 x 10-6 core melt accidents per unit year associated with steam generator We have not revised the calculations because (1) the tube rupture.
difference is not significant,.and (2) we had passed the deadline for revising the computer analysis of risk and the Staff could not have completed recalculations by the filing time. The difference between the two frequency estimates is small compared with the uncertainty (roughly a factor of ten) associated with the accident frequency estimates.
In addition, the conservatism associated with using the late overpressure failure model of release severity (which exaggerates the consequences by roughly a factor of ten or more) is ample to cover the difference. We remain confident that the resulting risk analysis does not underestimate the contribution to risk posed by steam generator tube rupture.
i f
Q.14 Should any of the requirements proposed at the July 29, 1982 l
meeting of the NRC Staff and members of SG0G be required for Indian I
Point Unit 2 and/or 3?
A.14 No, not at this time. The risk reduction associated with these items, even if they were 100% effective, is small relative to other core melt risks at the Indian Point plants. Therefore, no immediate i
However, there appears to b2 no' reason to actions are required.
exempt these plants from the generic program to resolve the co ts, related to steam generator tube degradation and tube rupture eve including the generic resolution of the related Unresolved Safet 8i Issues.. Although the risk reduction associated with the pote f
higher requirements may not be very significant, in light of other
'[
y the core melt probabilities, the Staff is continuing to evaluate If the most costs and benefits of these potential requirements.
j effective of these potential requirements can be implemented little or no cost or at a net savings, or if they are effective at reducing occupational radiation exposures, then they may be implemented even though the risk reduction is estimated to Q.15 Does your analysis of steam generator tube rupture 2
the contention that the cooling system at Indian Point Units longer and/or 3 should be changed so that Hudson River Water is no i
used for cooling?
i A.15 The estimated probability of tube ruptures in the generic S
The was based on the operating experience of all PWRs in the U..
specific steam generator problems experienced at the In h
site have also been reviewed in response to the contention d
cooling system at Indian Point Units 2 and/or 3 should be cha Neither of the Indian Point plants has experienced a steam i
tube rupture,but both have experienced some tube degradat on The tube degradation at these including a few sniall tube leaks.
The plants has been similar to that found at many other PWRs.
.-.c 4
s._
,.-y
-.a
--,w,
.--,,,-..,,,,eww.yn,p-ig-,, - -, -y,. -,, - - -m
l 13 -
Indian Point plants are neither among the group of plants with the
?1 most severe problems nor among the group of plants with the fewest problems. ' Up to 1981, the tube degradation experience at both units has been quite typical of the problems experienced at other 1
The steam generator tube pitting Westinghouse-designed plants.
recently found at Indian Point Unit 3 is uausual and has led to a However, this type of degra-program of extensive tube sleeving.
dation, which may be associated with the intrusion of Hudson River Water, is not likely to lead to tube ruptures since the pits are very limited in terms of the surface area of the tube thst is A tube with severe pitting, which is not removed from degraded.
service or sleeved, can lead to through wall leakage but is very The core melt analysis there-unlikely to lead to a tube rupture.
fore adequately addresses the Indian Point Units 2 and 3 including t
the influence of their present cooling system.
The change in the cooling system at Indian Point Units 2 and/or 3 f
would not have a significant effect on the risk of core melt.
However, as discussed in the response to Question 1, the Indian Point plants are included in the generic assessment of steam 9enerator degradation and tube rupture events and secondary system water chemistry is specifically being addressed in that program.
l l
-r-r
,--w.,
+,,
,r r
m--.,
-,,e
---, - -r
,c, - - --,,
--nn-
-a-
,a
.w---
n,.,
PROFESSIONAL QUALIFICATIONS FRANK'H. R0W50ME, 3rd U.S. NUCLEAR REGULATORY COMISSION i,
I am Frank H. Rowsome, 3rd. Deputy Director of the Division of Risk A I have served in that capacity since the Office of Nuclear Regulatory Research.
The work entails planning, budgeting, managing-
. joinirg the NRC in July 1979.
Much of the work of the Division is devoted to and staffing the Division.
The remainder antails risk research in reactor accident risk assessment.
assessment applied to non-reactor aspects of the nuclear fuel cycle and to standards development related to system reliability or risk.
I studied I received a bachelor's degree in physics from Ha'rvard in 1962.
theoretical physics at Cornell, completing all requirements for a Ph.D except From 1965 to 1973. I taught and engaged in research for the dissertation in 1965.
in theoretical physics at several colleges and universities.
j In 1973 I joined the Bechtel, Power Corporation as a nuclear engineer.. M assignment was to perform accident analyses for nuclear plant license a After six months in that job. I was transferred to a newly formed group of sys engineers chstged with developing for Sechtel a capability to perform ris ments and system reliability analyses of the kind the NRC was then develop l
l In that capacity I performed reliability analyses of the Reactor Safety Study.
nuclear plant safety systems, developed computer programs for system re analyses, perfonr.ed analyses of component reliability data, human relia analyses, and event tree analyses of accident sequences. 'I progressed fr nuclear engineer, to senior engineer, to grcup leader, to Reliability Group In this last position Supervisor before leaving Bechtel to join the NRC in 1979.
l at Bechtel, I supervised the application of engineering economics, reliability l
4
,s----
--e.
m,e
._,--h,-
.,. -. +, _
_.,,.e,
+,__.,+,ew,__,_
_%,,,,eg.gp,_yrywy-.-,~yp,-.9-,w-,,,,s
J Page 1 Frank H. Rowsome, 3rd Professional Qualifications (Cont.)
engineering, and analysis techniques to power plant availabidty optimi as well as nuclear safety analysis.
8 Risk Analysis (and its
, While serving as Deputy Director of the Division of
[
anticedent, the Probabilistic Analysis Staff). I also served as Acting Directo j
b i
(7 months), acting chief of the Reactor Risk Branch (9 months) and I.
['
of the Risk Methodology and Data Branch (4 months).
This experience has given me the practitioner's. view as well as the ma view of those facits of reactor risk assessment entailing the classificatio reactor accident sequences, system reliability analysis, human reliability I
analysis, and the estimation of the likelihood of severe reactor accidents.
have the manager's perspective but not the practit'ioner's experience with those facits entailing containment challenge analysis, consequence analys and risk assessment app 1ted to other parts of the nuclear fuel cycle.'
My role in the development of testimony for this hearing has been as of the preparation of testimony on risk and one of the coordinators of the l
i Safety Study."
, technical critique of the licensee's " Indian Point Probabi ist c I am not an expert on the design or operation of the Indan Point plants.
i i
9
, ~.,.. -
. -.. m.
e List of Publications "The Role of System Reliability Prediction in Power Plant Design,"
1.
j F.H. Rowsome, III, Power Engineering, February 1977.
3 "How Finely Should Faults be Resolved in Fault Tree Analysisf' by 2.
F.H. Rowsome, III, presented at the American Nuclear Society / Canadian
.[
Nuclear Association Joint Meeting in Toronto, Canada, June 18, 1976.
4 "The Role of IREP in NRC Programs" F.H. Rowsome III. U.S. Nuclear 3.
20555.
Regulatory Commission Washington, D.C.
" Fault Tree Analysis of an Auxi'iiary Feedwater System," F.H. Rowsome. III, 4.
77 805-5.
Bechtel Power Corp., Gaithersburg Power Division, F l
~'
Gary M. Holahan Professional qualifications l
I have a Bachelor of Science degree in Physics (Manhattan Colle Master of Science degree in Nuclear Engineering (Catholic thivers 1.
I have five years of experience in reactor and safety analys Combustion Engineering.
I position.
I have six years of NRC experience as a technical reviewer and Four of these years were in technical areas 2.
core thermal hydraulic design; ECCS design analysis; Section Leader.
Reactor Systems analysis; and post-accident analysis of TMI-2.
including:
For two years. I have been the Section Leader for*the Systems Section. Operating Reactors Assessment Branch, responsible for 3.
review in all of the systems areas for operating reactor assessment This includes responsibilities for all of the following reactor core, reactor coolant system, in NRR.
areas for operating reactors:
reactor protection system, instrumentation and con electric power systems, containment systems.
includes day-to-day review of the safety of operating reactors.
technical support on short term licensing changes.
I an assignment as group leader, Recent activities have included:for the plant systems analysis, 4.
the Ginna Steam Generator Tube Rupture event; and an assignment on the development of an integrated NRC program to address steam generator tube degradation and tube rupture events.
]
+
e
_y.
--,s w-
_