ML20055H590
| ML20055H590 | |
| Person / Time | |
|---|---|
| Issue date: | 06/04/1990 |
| From: | Taylor J NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | Michelson C Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-GENERAL, NUDOCS 9007270034 | |
| Download: ML20055H590 (17) | |
Text
.
^
y.
s
)
/p* %^\\
UMTED sT ATEs NUCLEAR REGULATORY COMMISSION
- ['
usmwoTow. o. c.aom s
-g*****
JUN 0 41990 i
Mr. Carlyle Michelson, Chairman Advisory Committee on Reactor Safeguards U.S. Nuclear Regulatory Commission i
Washington, D.C. 20555
Dear Mr. Michelson:
SUBJECT:
STAFF RESPONSE TO ACRS LETTER ON THE SEVERE ACCIDENT RESEARCH
'ROGRAM (SARP) l The staff recently met with both the ACRS Severe Accident Subcomittee and the full Comittee to discuss the elements of the Severe Accident Research Program (SARP), on April 5-7 and April 18, 1990, respectively.
In its letter dated from Carlyle Michelson to Chairman Kenneth M. Carr, " Severe April 24, 1990, Accident Research Progran.," the ACRS expressed reservations and provided coments on certain elements of the SARP.
7 The staff has evaluated the ACRS coments and its response is provided in the enclosure.
However, the following general coments are important to put the ACRS. letter in proper perspective.
By way of background, the April 24, 1990, staff met with both the ACRS Severe Accident Subcommittee and the ful1~ Committee and briefed them on the revised SARP on March 7 and 9, 1989, respectively.
In its letter dated March 15, 1989, from Forrest J. Remick to Chairman Lando W.
Zech, Jr., " Proposed Severe Accident Research Program Plan," the ACRS stated
'that the revised SARP represented a substantial-change from the previous severe accident research program and was a very positive step. Accordingly, we take the comments of the April-24, 1990, letter to apply to the. details of carrying out, or elaboration of the Severe Accident Research Plan, rather than a criticism of the Plan, which the ACRS viewed favorably last year.
The Committee stated that "there is much of deja vu in the proposed severe accident research. The same areas that were being explored at the beginning of the program almost ten years ago are still being investigated." When the staff proposed the revised SARP to the ACRS approximately one year ago, we did not 4
i l
propose any whclesale elimination of existing research programs and initiation of com.oletely riew and different ones. Rather, we explained that the major changes being nade to the previous SARP that resulted from an-extensive review were in the emahasis being placed on certain programs, the focusing of certain research to achieve near term renults, the addition cf certain new elements (e.g., scaling rationale) and the deletion of others (e.g., elimination of As discussed in NUREG-1365, the SARP was several duplicative computer codes).
revised to address in the near-term those issues pertinent to the IPE and CPI, I
such as mechanisms of early containment failure, and the long-term research is l
l being conducted to support the NRC's accident management activities and to j.
l 9007270034 900604 PDR,ACR$
p%(
OENERAL-PDC
'l e
e,'
j Mr. Carlyle Michelson 2
confirm the Commission's regulatory decisions on severe accident issues.
In that regard, we have discussed with the Committee on several occasions our IPE, CPI and accident management programs and those insights obtained from the SARP results. With respect to the Committee comment that the same areas are being explored, we agree with this statement and point out that virtually every PRA performed by either the staff or the industry has concluded that the phenomenological issues being investigated in the revised SARP remain as major contributors to risk uncertainties and that sufficient technical data are not currently available to substantially reduce the uncertainties in the likelihood
.of containment failure predictions. However, ye do agree to put more thought into defining what ranges of uncertainty shoulu be acceptable.
In its letter of April 26, 1990, from Carlyle Michelson to Chairman Kenneth M. Carr, " Evolutionary Light Water Reactor Certification Issues and Their Relationship to Current Regulatory Requirements," the ACRS made the following comments: 1) On hydrogen generation and control, the ACRS suggested that "the staff seek further technical information on possible effects, including stratification, before establishing a limit for the average hydrogen concentration," 2) On core concrete interaction, the ACRS stated that "the resolution of this issue will require engineering judgments as many of the physical processes are not fully understood," 3) On high pressure core melt ejection, the ACRS declared that "this is an extremely improbable event", and therefore, recommended only one mode of coping with the possibility.
We agrea with the points in the April 26 ACRS letter, and view them as part of the justification for the research on these elements of the SARP. We have not been able to reconcile the suggestions in the April 26 letter with the doubts expressed on the same program elements in the April 24 ACRS letter.
In summary, we think that the rc y sed SARP objectives and structure are appropriate, and we are optimistic that it will provide the needed answers for the operating LWRs, and also the evolutionary LWRs. At the same time we continue to seek ways of improving and sharpening the focus of the elements of the SARP.
We are pleased with the ACRS enthusiasm about the Severe Accident Scaling Methodology (SASM), although we caution about excessive expectations for SASM accomplishments. We expand on this caution in tha enclosure.
In the April 24 letter on SARP, the Committee states in the context of the TMI-2 and Chernobyl accidents that "... little change in the regulations that govern the operation of nuclear power plants has occurred."
In actual fact, there have been a number of changes.
Examples of some of the regulations that have been promulgated are the hyJrogen, ATWS, and station blackout rules. The Commission has also promulgated policies for severe accidents and safety goals..The severe accident policy has resulted in each operating plant performing a systematic examination (the IPE program) to identify any plant specific vulnerabilities to severe accidents'and correct these vulnerabilities in accordance with the requirements of 10 CFR 50.59, 10 CFR 50.90, and 10 CFR 50.109. As recouended by the ACRS in its January 19, 1989 ietter to Chairman Zech, the Mar ( I containment recommendations have been incorporated into the IPE for Mark I plants, as have the recommendations for the other containment types, which were reviewed by the ACRS in its March 13, 1990 letter to Chairman Carr.
.c m
4 Mr. Carlyle Michelson 3
With regard to the Committee statement that '... there has been no formal recognition of severe accidents, even for new plants," we note that those same severe accident issues and requirements identified by the staff and reviewed by the Comittee in their April 26, 1990 letter on evolutionary LWRs, will be e
addressed as part of the final certification for these future designs.
With regard to the Mark-! liner failure issue, as we discussed during our meetings, we are subjecting the NUREG/CR-5423 report to extensive peer review to gain a broader spectrum of perspectives regarding the methods and the uncertainties. We are currently planning a workshop for July, at which the individual peer review group members can discuss their comments in an open forum. Accordingly, we invite representatives of the ACRS or their technical staff to attend this meeting. We would like to discuss with you the outcome of the peer review and our recommendations after the authors and the staff complete their assessments of the peer review.
We have coordinated all these efforts with the Office of Nuclear Reactor Regulation and we will continue the collaboration between the two offices in matters of severe accident research and decision making.
Finally, with regard to the Committee's specific recommendations on page 8 sf the April 24, 1990, letter, we have provided responses in the enclosure.
Sincerely, Original Signed Byt James M.Taylot '
James M. Taylor Executive Director for Operations
Enclosure:
l As Stated cc:
Chairman Carr Commissioner Roberts Commissioner Rogers l
Commissioner Curtiss Commissioner Remick m
SECY s
AEB N b AEB D
d
. 42 eroh
.pe i l
0 lUI O Myl$y l
ltawila/jms Tinkler O
Nlh
\\ NIh0 Beckjord b Ta 1r 5 lD/fo y/
z,-.
l Mr. Carlyle Michelson 4
Distribution Circ, Chron, DSR Chron OTaylor NUDOCs, AEB r/f OSniezek Eltawila r/f TMurley:
Eltawila 0Blaha Tinkler RBernero Meyer E0ordan Murphy OScinto Sheron CFiles Speis fM[
Heltames
-Beckjord i
Taylor Burda Cortez Bartlett Minners/ King Summerour File (0&M-ACRS)
+
b s
u
(.
H
i..'
Staff Response to ACRS Coment on the L
Severe Accident Research Program
~
Addina Water to a Dearaded Core With regard to the overall effort related to the issue of adding water to a degraded core, we note that the Committee is "not convinced that the codes to be used are capable of providing the information being sought with sufficient 9
validity." The Committee also seems to view schedule constraints for this In program to be dictated by the regulatory requirements for the IPE program.
order to more tightly focus this area of research the Committee has recommended l
that this program be subjected to a review of the type developed in the SASM program.
In response to the Committee's concern over the adequacy of codes, we recogni.e that the issue of adding water to a degraded core involves complex phenoment for which little or no experimental data exists. However, we do not believe this is a reason to not do some research on the subject.
In the event of a severe accident, there is a strong likelihood that operators will attempt to restore coolant injection systems sometime during the course of the accident, and proceed to inject water onto a possibly molten or degraded core. The objective of this research is not to develop detailed, validated models of the wnching phenomena, but rather to obtain a rough, bounding estimate of what (ould happen under these circumstances. As operators-inject water onto a degraded or molten core, they will most likely see increases in steam production, hydrogen generation, containment pressure, etc.
It is important that we try to understand these symptoms and assure ourselves that operators will not misinterpret these symptoms and take inappropriate actions.
Investigation of water addition to a degraded core is, as our presentation noted, related to 1
s g
accident management research end not solely to review of IPE's.
Investigation t
of accident management strategies is a longer-term issue that vill undoubtedly continue after the IPE program is completed and one in which it can reasonably be expected that our understanding will improve. However, we also believe that it is reasonable, using existing codes (with some model improvements) to establish a range of outec9 s associated with water addition to a degraded core.
Our goal is to estimate that range of outcomes on a more reasonable basis than is possible using existing analytical methods which do not show substa ntial alteration of_ accident progression due to interdiction by plant operattrs to arrest core damage in-vessel.
We also point out that at this time this research effort is not solely limited to analytical study; some testing of molten fuel-coolant interactions is L
planned. Those experiments will be subject to a systematic scaling analysis such as SASM to determine proper scale effects. Additionally, we are also contemplating larger scale testing 'nvolving molten fuel-coolant interactions in the FARO facility.
j l
Core Melt Proaression 1
The Core Melt Progression element of the program addresses core melt behavior in J
It includes hydrogen generation and lower head failure. Much of
)
the current emphasis is on BWRs beceuse of the impact on Mark-I liner failure and also to balance out a previous PWR emphasis. The ACRS letter stated that their members were unable to determine how much additional information is needed and they asked if results would be obtained in time to be used in the IPE program,
]
L
1 3
)
With regard to how much additional information is needed, the process of core melt progression can be divided into early phase and late-phase behavior, f
Early-phase behavior, including ballooning, rupture, and runaway oxidation, is fairly well known now and our knowledge is significant.
For example, I,jrogen generation is known well enough to resolve the issue on blockage effects; peak temperatures that can result from the oxidation transient (which overwhelms decay heat) are now known to reach the U0 melting point (about 2805%) locally; 2
and the types and importance of many lower temperature eutectics have been identified.
Late phase behavior is only partly understood.
For example, we know that metals and ceramics cr.n begin their relocation process at different times, that crusts can form to contain large molten pools, and that molten debris can be uncoolable under some conditions and coolable under others.
But there is additional information that we need.
For example, under what (if any) conditions would metallic debris flow into the lower head region without forming a core blockage that retain; molten ceramics.
Such a flow path, postulated for BWRs, would result in early vessel failure with low debris flow rates compared to a molten-pool scenario. We need to know the approximate rate of debris arrival in the lower plenum and the superheat of the molten debris to be able to determine the mode and timing of vessel failure. At the time of vessel failure, we need to know mass, temperature, and composition (metals and oxides) of the melt, as these control reaction rates with concrete in a core concrete interaction and the rate of hydrogen production in a high-pressure melt ejection. We also :eed to know something about temperatures, debris surface areas, and the atmostnere
I i
I 4
(oxidizing or reducing) within the vessel to utimate fission product releases during the late stages of melt progression.
Based on progress to date, we believe that all of the above can be determined, i
at least approximately, within a reasonable long term research program, and core melt progression is a key element of our long term program.
Some of the core melt results may be available in the near-term to be used in such applications as the IPE program.
In particular, three tasks (2.5, 2.6, and 3.1) in the revised SARP plan call for near-term results on lower head failure and on debris conditions at the time of failure. Since this information, which describes conditions at the very end of the complex meltdown process, will be the most difficult to obtain, an alternate method to get approximate information is being useJ, as described below.
Examination of TMI-2 Lower Head and Lower Head Failure Analysis Plan One of the ACRS concerns on this area of work was related to the staff response when asked by the ACRS subcommittee how the information being collected under the TMI-2 lower head examination would be used by the NRC. We regret that we were unable to satisfactorily respond to your question and the following response might alleviate your concern regarding this program.
l During the course of a severe accident, the mode and. timing of the reactor vessel lower head failure could have a strong influence on the containment loadings.
In particular, the outcome of events associated with direct
~
containment heating and Mark-I containment shell meltthrough are strongly
(
I
l f
5 affected by the mode and timing of lower head failure, and this is the reason why it has been identified as an important near-term task in the revised SADp.
1 Several evaluations of lower head failure have already been made for some vessel types and assumed core debris conditions.
If likely debris conditions were well i
known, it is believed that the mode and timing of failure of the lower head could be predicted with reasonable confidence. While considerable progress has been made in the last 2 years in understanding late-phase melt progression, it is unlikely that debris conditions will be known well enough, and soon enough for ME's near-term needs. Consequently, another approach is being tried here.
In this lower head failure analysis activity, we are looking at different modes of failure for a wide range of debris and coolant (wet or dry) conditions, to try to determine what conditions lead to each failure mode.
Existing experimental data and analytical methods will be used for this assessment whenever possible, with new analyses performed only to fill the gaps; this is not a modeling or code development program. The concept of this program stens from earlier ACRS discussions, and an interested ACRS member (Paul Shewmon) participated in the review of documents and related discussions. As a result of these discussions, the staff determined the work scope and selected the l
contractor for this program.
The TM1-2 lower head examination will play an important role in our lower head failure program. Because the TMI-2 vessel did not fail, and even appears not to have been challenged severely, results from THI-2 will hopefully shed light on
-what debris conditions do ng.t lead to vessel failure. This should be very l,
l
f
~
6 l
(
1 helpful in better identifying those conditions that would lead to failure.
i For example, U b.9 that during the course of the THI-2 accident, a metallic blockage formed in the core region, creating a crucible of metallic core debris, in which a pool of molten ceramic debris collec.ted. We also know that the metallic crust first failed on the side, leading to a coherent pour of ceramic melt into the water filled lower plenum. Hence, by the examination of the debris in the lower head, coupled with an examinttion of the lower head metal, we can infer how the melt interacted with the lower head, and therefore infer the conditions that may preclude vessel failure and allow in vessel coolability
~
of the debris.
If you recall, previous analyses of the THI-2 accident piedicted 4
that, based on known thermodynamic conditions, the lower head should have failed. This discrepancy has been attributed to the inability of NRC's severe accident codes to correctly estimate the heat removed by the water in the lower plenum. By knowing the temperature of the lower head (from the THI 2 sampling program), we will be able to infer the total energy transferred to the lower head. Hence, we can determine the total energy transferred to the water based on the composition of the debris bed and the amount of decay heat. This information is not only important from the standpoint of understanding lower head failure but also from the stanapoint of accident management.
Removal of THI-2 vessel specimens and their examination is part of a joint OECD-NRC international project. Our agreement with OECD states that we will establish to what extent properties of the lower head were affected during the j
accident and the margin of structural integrity that remained. While there seems to be some general interest in margin to failure, as reflected by. this
,4r
+.-
w
~
7 l
agreement, the NRC's goal is to understand the process of vessel failure with regard to current issues rather than to determine what was the remaining margin at TMI-2.
Finally, we do not share your pessimism "that the number of variables likely to enter into a determination of lower head failure is so large and t
largely unpredictable that the predictions of the likelihood of the various possibilities may be subject to very large uncertainties." The object of the research is to account for these variables by categorizing the failure modes.
Hetce, even though large uncertainties may exist, we may be able to show they have little effect on our ability to predict the lower head failure mode.
Severe Accident Scalino Methodoloav We are pleased that the ACRS is enthusiastic about the SASM. We do think it appropriate to point out what SASM can and cannot accomplish.
Ensuring that large, expensive experiments are applicable to the-issue under study using scaling methods is not a new technique, but rather a generally well known engineering practice used extensively in the thermal hydraulic program carried out by the NRC in the 1970's and early 1980's. We Lgree that major experimental 2
programs should be based on sound scaling methods and it is our intent to do this. That is the reason we undertook the development of SASM.
i The ACRS letter suggests that the SASM should be used to resolve issues and should be applied to all programs. This is not possible or tractical in all cases.
For example, while the SASM does provide a systematic method to develop a scaling rationale for experiments, it may not be applicable to all issues.
For some issues, the processes may involve complex processes involving
r s
simultaneous heat and mass transfer, along with multiple competing chemical reactions (i.e., core melt progression). While an assessment using a SASM type approach is beneficial, it may be that an issue is not completely amenable to a scaling analysis, and other approaches need to be taken.
Secondly, SASM does not resolve issues. For example, the issue of DCH involves t
not only high pressure core melt ejection, but also the competing processes of natural circulation-induced failures and concomitant depressurization of the primary system.
In general, SASM can tell us whether or not there is reasonable assurance that the phenomena observed in a scaled experiment are the same as l
that which would be observed in a reactor.
Issue resolution, on the other hand, involver an integrated look at the complete processes and relative probabilities of these processes in order to determine an overall likelihood of containment failure. The method used by Theofanous of UCSB for the Mark I liner issue.s a prime example of this integrated approach, and one which we plan to apply to the DCH issue.
The Probability of liner Failure in a Nark-I Containment NUREG/CR-5423 The ACRS commented that "... conclusions about many important phenomena...are L
t supported primarily by the authors' judgment." We disagree with this view.
j l
What the authors tried to do was pull together all of the work done by various experimenters and analysts and, in general, reflect their conclusions.
l-
~ Unfortunately, trying to integrate the work of others and draw conclusions from it will undoubtedly lead one to believe that the conclusions are only those of L
I
.s C
9 the authors. While the staff believes that the UCSB effort is consistent with and responsive to the previous ACRS comment on factoring the results from l
previous work or expected results from existing research programs of U.S.
industry or foreign organizations into the NRC program; the NUREG/CR 5423 effort went well beyond this mere collation and use of existing information. New and significant contributions were also made in a number of technical areas. They include:
(a) quantification of hydrodynamically controlling spreading, (b) quantification of the melt superheat transient, (c) quantification of the melt-to liner heat transfer coefficient, and (d) quantification of liner thermal The new insights gained in each of these areas were instrumental to response.
' the successful integration of this issue into a coherent, quantifiable picture of liner failure. As we discussed at length at our meetings with the ACRS, we do not intend to accept or reject the report based on the authors' judgments alone. We specifically stated that we were subjecting the report to an extensive peer review, including 15 internationally known experts, National Laboratory research managers, and other NRC program effices and organizations.
We think that this process of peer review of the Ka+ 1 Containment Liner Failure Report is appropriate and will help in the resolution of this issue.
With regard to the specific observations listed on page 4 of your letter, we find these constructive and have instructed the authors to address them in the final report. We are also planning to conduct a peer review workshop in July of
[
this year. At this workshop, we intend to discuss each reviewers comments to fully understand the comments and the underlying bases. We invite representatives from the Committee or the ACRS techr.ical staff to attend this workshop. We would like to discuss with you the outcome of the peer review and I
u
1, 10 our recommendations after the authors and the staff complete their assessments of the peer review.
Continuina Code Develooment i
I On the subject of continuing code development the letter stated that the Committee was unable, based on the information provided in the meeting, to make any recommendation at this time. This is an important area which should be thoroughly addressed since the CONTAIN and MELCOR codes are the NRC's principal analytical tools for severe accident sequence analysis. We regret that we were unable to address all of your questions and propose to revisit this matter and fully answer your questions at a future meeting should you desire.
Molten Core-Concrete Interaction l
[
The Committee, in response to our presentation on research related to molten L
core concrete interactions (MCCI), has properly identified the significance of the ex vessel core debris coolability issue as it relates to the proposed advanced / evolutionary reactor designs. The Committee also observed that estimating the contribution of MCCI to late containment failure requires more inforniation than establishing an adequate floor area for spreading and cooling i
of debris. We agree and we sought to identify in the presentation that MCCI phenomena do impact other issues of interest, in particular the generation and release of refractory radionuclides, specifically the less volatile fission product species of strontium, barium, and lanthanum. We did emphasize the importance of our research related to melt coolability because we believe that
s i
[
11 it.is appropriate for this area of research to focus on mechanisms and I
strategies by which accidents can be terminated and the plant restored to a stable controlled condition.
With regard to the importance of MCCI to the Mark 1 liner failure issue, we call your attention to Chapter 5 of NUREG/CR-5423, *The Probability of Liner Failure l
in Mark-! Containment,' where MCCI plays an important role in determining the transient aspect of melt superheat which, in turn, plays an important role in the liner heat loading.
Lono-Term Research s
The areas of research, such as core sielt progression, hydrogen behavior, molten fuel-material-coolant interactions, etc., are broad phenomenological areas involved in severe accident progression and these, of course, have not changed with time. Within these broad areas, many specific phenomena have been understood well enough for our purposes. For example, the 10COR issue of blockage and hydrogen generation has been resolved and work has shifted to the l
i i
later stages of melt progression; hydrogen combustion has been understood well enough to issue a final rule and resolve USIs for Mark-!!Is and PWR Ice i
Condensers, and work is now focused on detonation limits under prototypic high.
l temperature conditions in steam environments; the alpha-mode steam explosion issue was resolved and emphasis is now on more likely interactions that might
)
affect accident management.
In the area of codes, this prop am produced an L
analysis tool, the Source Term Code Package, that was the first to analyze severe accidents in an integrated way not possible with WASH-1400 technology; l.
+
12 code development is now applied to a different architecture in MELCOR that wH!
facilitate uncertainty analysis.
SECY-88 147, the Integration Plan for Closure of Severe Accident Issues, includes the concept of a long term severe accident research program. This program will continue beyond " closure" to provide for decision confirmation, development of information in areas of most significance to risk, and general maintenance of a severe accident analysis capability. A detailed plan for the long term part of the SARP is under development and was not presented to the ACRS.
In this detailed plan, you can expect that a fairly large number of sub.
areas will be completed (e.g., gas production and erosion rates in core concrete interactions, iodine chemistry assumptions for Regulatory Guides) and Research Information Letters will be prepared.. A reduction in budget within five years to about one half of the current level is also expected for this planned work, but there will be some minimum funding level below which a maintenance effort i
could not be sustained. We also anticipate that severe accident research in areas unique to advanced reactors will increase in the coming years, s
The Committee has posed questions on how much uncertainty is acceptable, and on how much the proposed research will reduce the uncertainty. These are important but very difficult questions to answer. We have, for example, uncertainty estimates for risk in NUREG-IISO. We do not expect the agency to take any 1
action on the 5 plants analyzed in NUREG-ll50. Accordingly one might conclude I It should be noted in this connection that utility management took action to correct deficiencies uncovered by the initial NUREG-1150 analysis.
In the case of Zion, modifications were subsequently undertaken. Our statement refers l
to the status after the modifications.
13
~.
that these uncertainty estimates are acceptable. A more important question might be how much confidence we have in current uncertainty estimates (e.g., a large fraction.of the uncertainties cited in NUREG 1150 are based on expert elicitationratherthandata). We will attempt to address these questions in the long term research program being developed. However, it is anticipated that some phenomena will need to be developed further than others depending on the application of the results and the overall contribution to risk.
For example, issues that are likely to lead to an early containment failure would need more emphasis than those that could lead to late containment failure. Also, if operator interdiction could change the course of an accident, we must have a
' good understanding of the phenomena involved to make a rational judgment on that 5
strategy.
I With regard to the ACRS comment on the risk of seismic events and human performance relative to that of vessel failure, there is no disagreement. The expertise of the staff that gave this presentation is in severe accident phenomena, and not in seismic loadings and human factors.
i l
We welcome the ACRS' thoughts on this subject and we will be glad to discuss L'
this with you at a future meeting.
1 i
L l
1
~
m.
c,-
-