ML20056C521

Comments on 930106 Meeting Re Review of NRR Development of Key Policy Issues Re Preapplication Review of Candu 3,MHTGR, Pius & Prism Designs

ML20056C521
Person / Time
Issue date:	01/15/1993
From:	Ward D Advisory Committee on Reactor Safeguards
To:	Wilkins J Advisory Committee on Reactor Safeguards
References
ACRS-CT-2058, NUDOCS 9306240412
Download: ML20056C521 (7)

Text

_ _-

f 67'o705ff 1

Poe @ve 77 cc: Medhat El-Zeftawy 1/15/93 To:

J. Ernest Wilkins, Chairman ACRS Advanced Reactors subcommittee From: Dave Ward, ACRS consultant g u

\\

Subject:

Comments on subcommittee meeting of 1/6/93 INTRODUCTION The meeting was a review of NRR's development of key policy issues related to the preapplication review of CANDU 3, MHTGR, PlUS, and PRISM designs. The staff has singled out certain issues as important for the Commission to consider early in the review program.

The draft Commission paper was well written and presentations to the subcommittee were competent. However, the staff-proposed program is deficient in several areas. The most important deficiency is that there seems to be no intent by the staff to develop any generic design requirements. instead, staff review will be almost entirely reactive. I suggest that ACRS should recommend that the paper not be published for Commission and public comment at this time, but that the staff undertake a more generic approach to top-level requirements for advanced reactors. I'll expand on this point and on other items below.

DISCUSSION SECY recommendations The draft SECY paper makes five recommendations to the Commission.

[

These are repeated below in italics, with my comment following each.

/

Approve the staff recommendations in Enclosure 1 for conduct of preapplication reviews.

l suggest that the Commission should not approve the present

@k recommendations, but should request further development along the y

i lines I discuss below.

Sy

{

Approve the staff's conclusion that, based on the position that the

=

if CANDU 3 design is an evolutionary heavy-water design deriving fro QU CANDU designs operating in Canada and elsewhere, a prototype fy

?

CANDU 3 is not required for design certification.

\\

1 93o6240412 930115 L

PDR ACRS 0

CT-205B PDR

l T

2 I agree that a prototype should not be necessary for CANDU 3, but should be for the other three designs. The staff wants to classify CANDU 3 as an " evolutionary" rather than " advanced" design. I tried to find out in the meeting whether this had implications beyond the absence of a requirement for a prototype. But, I could not.

Note that positions which change as preapplication review experience is obtained will be communicated to the Commission and that as the staff identifies new issues it will inform the Commission.

This will make applicants unhappy, but I agree with the staff that it might be necessary to change some positions as knowledge increases. Ideally there will be a hierarchy of requirements, with the top of the hierarchy least likely to change. See my comment below on the need for some generic requirements.

Note that the Commission is requested to provide highest priority attention to those issues identified in the enclosure as being applicable to the PRISM design.

I do not agree with this priority assignment. Perhaps justice demands a first in first out program, but that should be the only reason for PRISM taking first claim on staff resouices. It may not be appropriate for the NRC to unilaterally establish priority based on national need, but what NRC decides to take first will certainly influence outcomes. If one considers the greatest need, if the benefits of nuclear power are to be practically available to the nation, it is not stretching the uranium resource -- which is PRISM's forte. The highest priority should go to designs which have the potential for permitting lower costs and reduced uncertainty and obscurity in safety performance.

1

• Note that due to the preliminary nature of the design information on j

the advanced reactor and CANDU 3 designs, and the preliminary nature of the staff's preapplication reviews, the staff does not j

recommend proceeding with generic rulemaking on any of j

the policy issues (my emphasis) identified in this paper. The staff will consider generic rulemaking, as appropriate, as the reviews progress, and the staff gains greater confidence in the final design certification.

4 l'

3 The staff still has this backwards. After a generation of experience with licensing nuclear power they still ask for more information on what designers are proposing before telling designers what the NRC believes is necessary.

The existing body of regulations, especially the GDC at the top, is a reasonable starting place for the review of advanced reactor designs. With these, applicants know, by and large, what is expected of them if they are to get a design certified. And, the NRR review staff knows, by and large, what they should be looking for and what they should accept cr not accept when they begin to examine the details of plant designs. However, because the existing body of regulations was developed for LWRs, some parts of the existing body of regulations don't fit and other parts aie inadequate.

The NRC should be able to distill the essence of what has been learned in the past generation about safety and regulation and cast that essence into a set of general criteria for advanced reactors which will modify or replace the GDC for LWRs. They have made a start by identifying ten areas where the LWR regulations are not adequate. Now they should bite the bullet and-stats what they believe i.s required for adequate advanced reactor designs. It is really not an impossible task.

Recommendations on individual issues in the " enclosure" The enclosure to the draft SECY makes recommendations for each of ten policy issues. Eight issues [A through H] are classified by the staff as

" Category 1", meaning that the staff agrees with applicant contention that a departure from current [ LWR] regulations snould be considered.

Two issues [I and J) are classified by the staff as " Category 2" meaning that the staff does not believe departure from present regulations is warranted. I will not repeat the staff recommendations here, but will comment on each of the ten issues in the paragraphs below.

First, two general comments:

j

1) The staff proposes that all of the ten issues are not applicable to every one of the four designs. There are cases where the proposed design presents nothing that would fall outside of present practice.

For example, " Source Term" does not apply to PlUS because whatever is worked out in this area for the current generation of LWRs will apply equally to PlUS. What they have proposed seems reasonable to me.

2] is the list of issues complete?

I believe the staff has done a good job in identifying the important

4 I'

4 areas where the existing regulations are not good enough, i.e., the key licensing issues. The areas identified are similar to those identified at the time of preliminary reviews back in 1987-8. Four were identified then:

Accident selection Siting source term selection and use Adequacy of containment systems Adequacy of off-site emergency planning ACRS commented on those in a series of letters:

6/9/87 draft NUREG-1226 7/20/88 key licensing issues 10/13/88 MHTGR 11/22/88 PRISM 1/19/89 SAFR The 7/20/88 letter recommended that some additional issues be considered:

1) How safe should these plants be?

2) To what extent should regulatory requirements accommodate public perception?

3) Extra capacity in decay heat and scram systems

4) Need for prototyping

5) Use of cost-benefit analysis 6] Design for resistance to sabotage

7) Operation and staffing i

i The new staff list covers these, except as follows:

1) The safety goal policy lurks in the background, but there is no j

specific mention in the policy paper that it [the safety goal) is the applicable standard for the new designs. Staff responses to questions about this at the subcommittee meeting did nothing to clarify-ambiguity in the written staff position.

2) There seems to be no intention on the part of the staff to insist that the designs accommodate public perception of safety needs as separate from technically-justified safety needs. That's good.

3) These subjects, scram and residual heat removal systems, are identified as key issues. The specifics are not exactly as discussed in the ACRS letter, but the staff is looking at them.

4) The staff proposal addresses this. I agree with their positions --

not necessary for CANDU 3, but necessary for the others.

5) There is no indication the staff intends to use cost-benefit analysis as a test of acceptability of a design feature in their review. That's good.

I' 5

6] There is no guidance for incorporating resistance to sabotage in the review. I still think this is doable, and should be attempted.

7) This is on the staff list, as ACRS suggested.

A. Accident Evaluation The recommendations say that selection of events against which designs are to be evaluated will be " supplemented with insights from PRA", based on " expected frequency of occurrence", and "with appropriate conservatisms factored in to account for uncertainties".

These are all good words but eventually the words will have to be translated into quantitative guidelines. I believe these quantitative guidelines should not be left to the stage where there is negotiation between the NP.R project manager and working-level people with the designer. The quantification is, to a large measure, the policy making.

That should be done up front and the values should be understood and debated by policy makers.

B. Source Term It is not clear to me that the R&D programs for CANDU 3, MHTGR, and PRISM are sufficient to develop the sort of mechanistic information on fission product release and migration that is needed. The staff was vague about what the Canadians claim to have available. The MHTGR program without the NPR piece is probably not enough. A couple of elaborate tests in TREAT will divulge a little, but probably not enough, about PRISM fuel behavior.

The third recommendation [this is, of course, closely related to the accident selection issue] uses the phrases " bound credible severe j

accidents" and " uncertainties". Once again, quantifications will eventually have to be substituted for these phrases. Those j

quantifications will, in effect, become the policy, and should not be left buried at low levels of the review process.

C. Containment Performance The paper tells us that the MHTGR preapplicant assumes a leak rate of 100% volume per day and the CANDU 3 preapplicant uses 5% That's interesting, but at this stage we should have discussion what the staff thinks is necessary, not just what the applicant proposes.

One recommendation is that the containment should be required to hold (to ASME Level C criteria) for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following the onset of " core Jamage, the specified containment challenge event". Interesting, but what is the core damage event to be? That's the big question. Any dummy can agree that holding to Level C stresses for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> is pretty

I 6

reasonable. Maybe it ought to be 12 or 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br />, or even one week. But, controversy over the holding time pales to insignificance against the controversy introduced by quantitatively specifying the core damage event to be considered. That's where policy makers should get involved.

l D. Emergency Planning The staff could not answer our questions about how much, by quantitative measures, present EP is estimated to contribute to risk management. Makers of policy about EP for advanced designs can't make intelligent decisions unless they have this sort of information available. It would include quantitative estimates of the benefit of EP to risk reduction with advanced designs.

i don't propose basing a decision for or against EP solely on the risk

]

reduction numbers. I believe EP should be regarded as defense in depth, as we regard containment. ACRS has said a containment should provide l

a CCFP of 0.1 no matter the claim or estimate for core melt probability.

Similarly, NRC could insist on some conditional mitigative benefit from EP no matter core melt and containment claims. Is that what NRR proposes? Can they get quantitative about it?

E. Reactivity Control The debate should not be about whether there have to be " rods" or not, but whether there should be a requirement for some sort of independent i

system which can plunge, squirt, drop, or whatever some amount of J

negative reactivity into appropriate parts of the core within some

-l appropriate time.

-l F. Operator Staffing Particular attention needs to be given to multiple-module designs.

Individuals modules could be in various states of power operation, refueling, accident, etc. at a given time. Staffing should be able to cope with that.

G. Residual Heat Removal The designs with residual heat removal systems that are always on line are much superior to existing LWRs which all share the flaw that a scram is not a straightforward default to an always safer condition.

NRR should not legislate away this major advantage for advanced designs because of lack of vision.

H. Positive Vold Reactivity l think this is a really sticky point. Personally, I don't like flirting with a positive void coefficient for any reason. Maybe for a good one; but, I'm not sure we've heard one. There is some allusion in the report

c L

7 that PRISM might be able to get rid _of the positive void coefficient, but only by trading for the safety disadvantage of higher worth in individual rods. I don't understand that, but there could be a reason for accepting some negative void coefficient in there somewhere. For CANDU 3, the reason for accepting a positive void coefficient seemc to be economics. I am not sure that is a good reason.

I. Control Room Design The staff position that there must be safety grade shutdown equipment in the main control room, even if there is such equipment in accessible alternative shutdown areas, seems pretty arbitrary to me. If the Commission is to endorse this, I think they need to be given a better argument than the staff has mustered so far.

J. Safety Classification Blanket insistence of the full-range of safety-grade features for certain systems is archaic, I believe. Applicants should be permitted to justify the extent of " safety-gradedness" [ resistance to earth, fire, air, and water) depending on the mission of a particular system and the accident sequences for which that mission is important in controlling i

ris k.

1 D

f i.--.

- -~