ML20055D006
| ML20055D006 | |
| Person / Time | |
|---|---|
| Issue date: | 04/20/1990 |
| From: | Rogers K NRC COMMISSION (OCM) |
| To: | Chilk S NRC OFFICE OF THE SECRETARY (SECY) |
| References | |
| NUDOCS 9007030070 | |
| Download: ML20055D006 (6) | |
Text
pt f ~
y'".
N'O T A T I 0 N V g.yo.o o o o o o o o o o oo o o o 4
@ LEASED TO THE PDR L
RESPONSE SHEET {
6/>4/90 I
.{'
s.
Idale 1
................ini$ tis 4
.T0:
SAMUEL.j. CHILK, SECRETARY OF'THE COMMISSION FROM:
' COMMISSIONER ROGERS
SUBJECT:
SECY-89-102 - IMPLEMENTATION OF SAFETY GOAL POLICY APPROVED /" Par et DISAPPROVED W par im s ABSTAIN NOT PARTICIPATING REQUEST DISCUSSION COMMENTSi tcc arruascuT l
}
SIGNATUREC 9007030070 900420 h 3,
[
yg PDR COMMS NRCC l~
L CORRESPONDENCE pdc DATE YES g
)
L(
S ENTERED ON "AS"
/
/_ /
/
E NRC SECY_ FORM MAY, 1987
i COMMISSIONER ROGERS'S COMMENTS ON SECY-89-102 A.
General Imolementation conceot This section contains a number of useful concepts.
Clearly, a great deal of hard thinking has gone into it.
With the incorporation of the suggestions offered below, I approve it.
Considerable clarity uould be brought about by adopting and consistently following a few basic definitions and conventions for the use of certain terms when discussing Safety Goals or safety goals objectives.
The key terms are: " safe enough", " adequate protection", " reasonable assurance", and "backfit(s)".
I suggest these definitions and conventions:
1.
Adopt the two qualitative and two quantitative Safsty Goals as defining " safe enough".
(This is in agreement with an ACRS recommendation of February 16, 1989).
2.
Adopt the convention that the term " safe enough" always i
signifies a result but never a means.
3.
Restrict the use of the phrase " reasonable assurance" so that it is never coupled to " safe enough" but might be coupled to
" adequate protection".
Thus " reasonable assurance of adequate protection" would be a permitted usage, but " reasonable assurance of safe enough" would be unacceptable.
4.
Restrict the use of " safe enough" to associations with the Safety Goals; never associate the phrase with
" adequate protection".
The entire regulatory program is the means of "providing assurance of adequate protection".
New generic regulatory requirements should be compatible with " safe enough".
5.
Adopt the convention that the term "backfit(s)" is never to be coupled with " safe enough" or to the Safety Goals when an individual plant backfit is at issue.
For example, we would not say that because the Safety Goals are met further "backfits" are not justified.
We could say that, because ther e already is
" reasonable assurance of adequate protection" and additional safety improvements are not substantial, further "backfits" are not Justified.
This usage is consistent with the position NRC took in defending the Backfit Rule.
2 The fundamental principle that Safety Goal objectives should be targets for generic regulatory requirements, but not be requirements or standards to be net or be criteria for individual plant licensing decisions is correct.
This is consistent with points 1.
and 2.
The recommendation of the staff that none of the Safety Goal objectives be construed as a generic and quantitative adequate protection standard is also acceptable.
The definitions and conventions above help to assure clarity in this regard.
Safety Goals should be viewed as aspirational targets which if proximately achieved and maintained should be
" safe enough".
However, there must be some way of assessing whether " proximate achievement" has been reached.
This requires some form of i
measurement.
(The IPEs appear to offer several ways to carry out such a measurement.)
l The four primary Safety Goal objectives should be regarded as benchmarks for the overall collective average of all plants.
Subsidiary Safety Goal objectives can be regarded as benchmarks against which proposed changes in regulatory requirements can be judged, or they can be used to assess future plants.
The Safety Goals large release guideline is expressed as an overall mean frequency of a large release from a reactor accident.
This guideline should be interpreted as an overall or collective average for all plants. The same interpretation should be used for any additional subsidiary objectives such as core damage frequency.
The question remains open of what connection, if any, should be made between " adequate protection" and " safe enough" or the Safety Goals.
Staff says in this SECY, "Furthermore, it'is essential that such an understanding, or definition, of adequate protection have a clear relationship to Safety Goal Policy.
The staff believes that such a relationship can be established by reference to the two quantitative health objectives (QHSs) defined in the Safety Goal Policy." They also say "Some quantitative and generally applicable definition of adequate protection may eventually emerge... which i
I take a partly quantitative form" and, "The specification of a bench I
mark standard along these lines could be viewed as an initial step in defining adequate protection...."
I concur with this point of view.
While "PRAs are not up to the task of providing definitive findings on which narrowly differentiated individual plant licensing findings can be made", I feel that they can be used to ascertain " proximate" achievement of the Safety Goals.
Because several years may be required to bring to completion a program for defining " adequate protection" based on a plan that includes a benchmark relationship to the Safety Goal Policy, it is l
[
~_._.._.___.
3 my view that the Commission should now direct the staff to begin to prepare such a plan.
B.
PrinciDal Elements of ImDIementation First element.
I approve the general approach taken by the staff as portrayed in Figure 1., but not necessarily with the definitions suggested by the staff for safety Goal Levels Three and Four as described in Part C.
and Enclosure 1.
I agree with the ACRS that defining Level Three in terms of off-site fatalities is inappropriate, since Level Two already provides an objective in terms of fatalities.
I will comment further on this subject in my responses to Part C.
Second element.
I approve of the staff's proposal to review existing and future PRAs and to assess the proximity of the results to the safety Goals objectives. Through thic process and a careful choice of plants, a representative sample of the entire U.S.
population of plants could be created.
Comparison of this sample mean to the objectives of the Safety Goal hierarchical Level Two could then be made.
This would be a major step towards the program of defining " adequate protection" that I suggested in my comments on Part A.
The staff's intention to fold this work into the Systematic Review of Light Water Reactor Requirements has my endorsement.
However, how they plan to do this and particularly the staff and other resources requirements for carrying out their j
proposal have not been made clear in this SECY.
Addressing the
)
list of nine topical areas in Table 2. will be quite demanding of resources, but otherwise I have no reservations about the list.
Therefore I approve of this element subject to the availability of resources in the existing Five Year Plan.
Third element.
The development effort sketched here appears to be important in providing a very useful tool for evaluating the risks involved in modifications considered by the NRC or by the licensees.
However, the expansion and maintenance of a data base capable of being used as proposed here might require sonsiderable resources.
My approval of this development and maintenance effort is contingent upon the required resources already being provided in the Five Year Plan. If this condition is satisfied, then the use of this tool for estimating the overall proximity of individual and collective plant performance to Safety Goal objectives has my strong endorsement.
My previous comments on the relevance of such an effort to a program of defining " adequate protection" also apply here.
I am concerned however about the use of this tool (s) for NRC new regulatory requirements initiatives.
Therefore, my approval of this element is further contingent upon a commitment of the staff to discuss with the Commission exactly how they propose to use such a tool before they proceed to develop any new requirements which are based upon its use.
I am opposed to the use of Safety Goal 1
. ~.
r 4
objectives in any decisions on individual plants without Commission concurrence.
Fourth element. This element, the use of subsidiary quantitative targets through a partitioning of Safety Goal objectives would appear to have merit, but it is somewhat unclear as to how far ranging the staff's intentions are.
I interpret partitioning to imply a horizontal (of equal significance) grouping of objectives as. distinct from a vertical (hierarchical) grouping.
Such an interpretation immediately gives rise to the question of how the elements which are to be partitioned are to be selected.
The information in this SECY is insufficient for me to approve this element, so I must disapprove it at this time.
1 l
C.
Structure of Safety Goal Obiectives t
1.
Proposed Safety Goal Obiectives I approve the staff proposal for the hierarchy of objectives and the association of these objectives with the hierarchy of PRA's as in Figure 1.
Nevertheless, I do not believe that a Level Five is needed.
It is clear that the Safety Goals must be reached through our rules and regulations.
However, the rules and regulations are not cast in term of probability or risk and therefore are not appropriately considered as part of a hierarchy of probabilistic objectives.
(a) Level Three I disapprove the staff's Level Three proposed safety Goal objective definition.
I could accept the ACRS position that the definition should be in terms of the release itself; that is, in terms such as the quantity or fraction of material released, and should not l
be site or meteorologically dependent.
I could also accept a definition patterned on th2 stcIf's option on page 7.
of Enclosure 1.
"A large release is any release from an event involving severe core damage, primary system pressure boundary failure, and early containment failure".
I would encourage the staff to test this type of definition but suitably modified so that it can also be applied to designs other than conventional light water reactors.
If Level Thrgg is defined along either of these ways, a major objective is accomplished of relating the previously defined health objectives to the reactor itself.
Defining Level Three in terms of near-term offsite fatalities is inappropriate, since Level two already provides an objective in terms of fatalities.
I believe that the Commission should request the staff to further develop a large release definition that relates to the reactor rather than to health effects, and to return it to the Commission for approval.
4 i
5 (b) Level Four l
I approve the staff's Level Four proposed Safety Goal objective j
that the overall mean frequency of coes damage events should not exceed 1 x E-4 per reactor year of operation for current reactors, i
plus the separate objective of 1 x E-5 per reactor year of operation for future advanced standardised plants.
I recognize that this latter objective may be somewhat more stringent than the Level Two Safety Goal, but I would be more troubled by failure to make such a
statement now than I
would be by the small inconsistency that it would introduce in the overall scheme.
This goal is currently attainable, and should further growth in U.S.
nuclear generating capacity occur with the construction of advanced
- plants, a more stringent goal will be necessary to avoid an l
integrated U.S. probability of a core damage accident that exceeds the current value.
I approve the staff's intention not to employ a quantitative containment performance objective expressed as a
conditional containment failure probability, since such an objective could thwart efforts to maintain a balance between accident prevention l
and accident mitigation.
l 2.
Comoarina PRA Results to Safety Goal Obiectives I approve the staff's intention that each quantitative safety Goal objective in Levels Three and Four of the hierarchy is to encompass all contributing initiating events (internal and external) except sabotage.
I also support the staff proposal to define the plant performance objective in terms of mean values of a distribution, and I agree with their comments with regard to comparison of PRA results with safety Goal objectives on pp 14-15.
I approve the staff's omission of an ACRs recommendation for quantification of the quality of reactor plant operation as a
" plant performance" objective at Level Four.
This is not feasible at present.
3.
Subsidiary Obiectives I approve the staff's proposed approach outlined here. I also believe that the staff should develop a
plan by which a
representative sample of the existing ensemble of licensed plants as characterized by twenty to thirty IPEs or existing Level II PRAs are used to measure the extent to which the Level Two Safety Goals are met.
I continue to believe that this could constitute a major step forward in demonstrating closure of the regulatory process to the public and the congress.
D. Guidelines for Conduct of Cost-Benefit Analyses I approve this section as it stands.
.-