ML20153G937
| ML20153G937 | |
| Person / Time | |
|---|---|
| Issue date: | 07/29/1998 |
| From: | Shirley Ann Jackson, The Chairman NRC COMMISSION (OCM) |
| To: | Hoyle J NRC OFFICE OF THE SECRETARY (SECY) |
| Shared Package | |
| ML20153G885 | List: |
| References | |
| FACA, SECY-98-171-C, NUDOCS 9809300269 | |
| Download: ML20153G937 (6) | |
Text
_ - - . -
NOTATION VOTE RESPONSE SHEET TO: John C. Hoyle, Secretary FROM: CHAIRMAN JACKSON
SUBJECT:
SECY-98-171 - PROPOSED RULEMAKING ON 10 CFR PARTS 50,52 AND 72 REQUIREMENTS CONCERNING CHANGES, TESTS AND EXPERIMENTS AND STAFF RECOMMENDATIONS ON CHANGES TO OTHER REGULATIONS AND ENFORCEMENT POLICY Approved x in partDisapproved x in part Abstain Not Participating COMMENTS:
See attached comments Shirle Ann Jackson SIGNATURE July 29,1998 DATE Entered on "AS" Yes x No
!!!'88Ms748P5 CORRESPONDENCE PDR .,
e I Chairman Jackson's Comments on SECY-98-171 l I approve, in part, and disapprove, in pr.rt. the staff's proposal for rulemakmg. I approve the staff's proceeding with issuance of the proposed rule language for public comment in order to support the expedited finalization of a revision to these processes. I disapprove of the specific language proposed by the staff for Section 50.59(c)(2)(vii), " reductions in the margin of safety." ,
J I agree with the recent letter from ACRS on this rulemaking, in that: ,
- 1) 10 CFR 50.59 can accommodate risk-informed decisionmaking. l
- 2) the positions, as presented, on margin of safety may add regulatory burden without a conunensurate j safety benefit.
I disagree with ACRS in that I believe:
- 1) the rulemaking should go out for public comment to foster comment on this high priority issue, and
- 2) the regulatory guidance can be worked in parallel with the rulemaking.
I note that a further reason for issuing this package for public comment at this time is that the paper calls for the proper use of enforcement discretion as this rulemaking progresses, thereby providing further stability in the implementation of this rule in the industry.
Further, I propose that the SRM on this SECY, and the voting record, be placed in the FR notice to clearly inform stakeholders on preliminary positions taken by the Commission.
Giving Definition to Minimal Attached to the recent ACRS letter was "A Proposal for the Development of a Risk-Informed Framework for 10 CFR 50.59 and Related Matters." The proposal forwarded by the ACRS parallels an existing risk-informed approach described in Regulatory Guide 1.174. Regulatory Guide 1.174 describes a method for determining the level of review, based on severe accident implications, for proposed licensing actions. The proposal forwarded by the ACRS describes methodology for creating frequency-consequence curves for Class 1-8 accidents. The proposal states that existing processes could be extended to provide appropriate context for whether the results of a change are " minimal."
The proposal also notes that aspects of this type of approach are in use in the international regulatory community. The approach utilized in the proposal forwarded by the ACRS is consistent with the Commission guidance in the Staff Requirements Memorandum of March 24,1998 on SECY-97-205.
Without commenting on the specifics of the proposal forwarded by the ACRS, I am convinced that changes to nuclear plants can be evaluated in a risk-informed context. Any such approach would benefit from paralleling existing methodology. Careful consideration would be required to ensure that the " consequence" and " frequency" standards are appropriate for a 50.59 type application. For instance, " consequences" could be evaluated at one of the following levels: fractional releases, off-site or on-site doses, or challenges to fission product release barriers. " Frequency" could be evaluated for Class 1-8 accidents or for design basis accidents using existing guidelines for risk-informed regulation.
The level at which consequences and frequency of events were tracked would also impact the type of parallel, deterministic (e.g., protection of redundancy, defense in depth, etc.), considerations against which changes would have to be evaluated. For instance, evaluating consequences at the level of the loss of a single barrier, or occurrences of accident sequence initiators, might allow elimination of parallel, deterministic, considerations such as " margin."
i
T o .
' 2 6
It is of some concern to me that the while staff has pursued risk-informed approaches to issues like the l
review of TSs, the use of Graded Quality Assurance, and programs like In-service Inspection and In-service Testing, the staff appears to be more reluctant to allow risk-informed approaches if the result is the relinquishment of review and approval authority. Because prior NRC review and approval impacts the cost and schedule oflicensed activities, we must ensure that we require such prior review and approval only when justified or required by mandate. We should not limit the application of risk-informed regulation as a means to ensure continued NRC reviews and approvals oflicensed activities.
l This message is complimentary to my oft repeated message to industry that the use of risk information is
" double-edged," that is that relief and additional regulatory scrutiny may both result from its use.
Marcin of safety The staff proposes to provide a specific definition of " Reduction in margin of safety associated with any technical specification," and to revise the current provisions of 10 CfR 50.59(a)(2)(iii) to explicitly refer to this definition. While I commend the staff on its efforts to provide clear, definitive, requirements in this proposed rulemaking, I am concerned that the proposed rule is not consistent with policy direction established by the Commission in the SRM dated March 24,1998. I concur that it is important that the staff has the independence to (and,I believe, has the responsibility to) inform the Commission when there are concerns with Commission guidance (as it did in COMSECY 98-013).
However, I believe that when the staff proposes to take action that is inconsistent with Commission direction, it is obliged to provide a clear and complete rational for the proposed departure. I do not feel that the staff has met that obligation for the " margin of safety" aspect of this proposed rule. However, this said, I do not disagree with the staff's conclusion that we should be careful to understand, and maintain, a consistent regulatory basis on " margin of safety." We must proceed in a manner that does not call into question the existing deterministic basis for " reasonable assurance" of public safety embodied in plants' Technical Specifications (TSs).
l My previous discussions with the staff have indicated that it is extremely difficult (and probably not legally defensible) to allow decreases in the " margin of safety" when the upper and lower limits between which " margin" may ' exist are not defined in relation to the regulatory requirements for safe operation.
' Based upon these discussions, I can only assume that the staffis hesitant to allow direct reductions in margin within the " basis" for TSs because some such changes could create a de-facto change in the TSs themselves. He staff may also be concerned by the lack of consistency in the " margin of safety in the basis for TSs" associated with the different generations of existing licenses (e.g., older customized TSs compared to improved standardized TSs), and associated with the different methods utilized in the ,
technical review and approval of the TS (e.g., some TSs might be based on maintaining margin between accident analysis results and acceptance limits, while other TSs might be based on margin which was built into analytical techniques and methodologies used in the accident and safety analysis, with no
" margin" between the results and the acceptance limits, etc.).
The staff s proposed method of requiring prior agency approval to changes ofinput assumptions, l analytical methods, etc., for those parameters which affected the selection of TSs, results in the newly controlled parameters being treated essentially the same way as values in the TSs. It also appears that implementation of the staffs proposed control over a broad range of parameters used in the safety !
analysis would effectively prevent any change to the facility that would result in a " minimal change in consequence," a condition allowed elsewhere in the proposed rule. In other words, it is not clear what type of changes would successfully pass the 10 CFR 50.59 test for allowed " minimal mereases m .
l l
l l
J
3 consequences," without failing the test for "no reductions in the margin of safety." I do not believe that the potential safety significance of all the parameters to be covered under the proposed definition of a reduction in the margin of safety always justify the requirement of prior NRC approval.
The staff should continue to work to establish a technically sound method for allowing licensees to make plant changes where there is only " minimal" impact on safety. If fundamental conflicts exist with allowing reductions in some " margins of safety," especially those on which the validity of TSs are based, then staff should provide a clear explanation of this, and should address how other changes to the structure of the regulation, which do not create fundamental conflicts, can be made in a manner which achieves the Commission's objective of removing unnecessary burdens from licensees.
Attachment "A" to this vote describes one alternate method for addressing the issue of" margin of safety." This alternative would maintain existing margins of safety (associated with TSs), while providing greater flexibility to licensees in implementing changes to their facilities. This alternative is based on methodology similar to that described in NEl 96-07. This methodology requires evaluating the effect of proposed tests and changes on the accident analysis results (rather than inputs, as proposed by the staff), in cases where TSs are based on accident analysis considerations. Prior NRC approval of changes, tests, and experiments would be limited to those cases where there was a Et effect on the accident analysis results. The alternative also recognizes the significance of the analytical techniques used in the safety or accident analysis, and would require some form of prior approval for analytical '
methods used to support changes when the change did not have prior NRC approval. This approach could provide staff reasonable assurance that the assumptions made by the license reviews are not invalidated. The staff should evaluate this option, along with other comments in this area, during the comment period.
In considering the technical and regulatory underpinning of this clause of Section 50.59, I have become concerned that we are evaluating incremental changes to a provision which is not well suited to such changes. I am concerned that the result may be the addition of yet another layer of regulatory process rather than the elimination of any unnecessary layers. For this reason, the staff should be receptive to internal or public comments on feasible alternatives which eliminate the discussion of"the margin of safety in the basis of TSs," while maintaining the integrity of the plant's licensing basis. I envision that it may be possible to eliminate the rule language criteria on " margin of safety" if evaluations of
" frequency" and " consequences" are performed at a level of significance which bounds allowable
" minimal" reductions in margin.
ACCIDENT OF A DIFFERENT TYPE In determining the effect of any proposed change to Section 50.59, it will be necessary to more clearly understand what an " accident of a different type" is. The staff should provide a more definitive definition of an accident than was included in COMSECY-98-013. The information provided by the staff should address, as a minimum, the following: l
- 1) What is an " accident" under this Section, and is it consistent with other existing regulations (e.g., i Section 50.92, Section 50.34, Appendix A of Part 50, etc.)?
- 2) is an '..ccident of a different type" better described as an " initiating event (e.g., loss of feedwater, loss of ofTshe power, new common mode failure mechanism, etc.) of a different type?"
- 3) What are the bounds which limit those " accidents" which are the subject of this Section (e.g., only those initiating events which, when evaluated using approved analytical techniques, result in transients i
l l
- . - - .-. -_ _ _ -. - . - - - . - ~ . -
4 with the potential to challenge fission product barriers, etc.)?
PROCEDURES I commend staff on inserting a definition for the term " Procedures as described in the final safety analysis report (as updated)." However, I am concerned that the definition provided may cloud the distinction between: (1) those procedures which must be screened, or evaluated, under Section 50.59, and (2) the criteria which necessitates a full safety evaluation. I believe that staff seeks to indicate that all procedures which are described as being required in the FSAR are subject to a 50.59 screening. The screening would identify the need for a full safety evaluation only if a proposed procedure change created a change to the "information in the FSAR regarding how structures, systems, and components are operated and controlled .. ." Staff should solicit comment on this definition and clarify the proposed definition, as required, in the final rule.
MAKING THE RUI E RISK INFORMED I note with interest that members of the ACRS believe that there are substantial barriers in the existing deterministic framework of 10 CFR Part 50 to the concept of allowing " minimal" changes in accident probabilities or consequences. In my previous vote on SECY-97-205, " Integration and Evaluation of .
Results from Recent Lessons-learned Reviews," I approved the staff's proposal to develop the framework for risk-informed regulatory processes. In particular, I called for the staff to develop a series of milestones by which the Commission could " chart its course in its move to more risk-informed regulatory processes." Additionally, I promoted the idea of promulgating a new regulation in 10 CFR Part 50, that would make clear how the Commission uses risk information in its decision-making. In proceeding with the "short-term" changes to 10 CFR 50.59 (and related regulations; "short-term" actions from SECY-97-205), and in responding to the ACRS, the staff should re-evaluate whether the Agency should initiate action to provide for a risk-informed framework that would allow for the efficiencies to be gained through use of risk-informed, performance-based revisions to our regulatory processes.
Attachment "A"
" STRAW MAN" ON MARGIN OF SAFETY Regarding margin:
. The margin between regulatory limits and the failure of physical barriers is protected in the regulations (and also in the portion of the Technical Specifications (TSs) called " safety limits").
. The margin, as reflected in approved safety and accident analyses, between the protection afTorded by the TSs (e.g., the limiting safety system settings and limiting conditions of operations) and the associated regulatory limits is "the margin of safety as defined in the basis for any TS."
a The margin between normal plant or system operation and the " bounding" assumptions used in accident analysis is below the threshold of safety significance that requires NRC prior approval for changes.
. The results of safety and accident analyses are subject to significant variance, depending on the analytical techniques and methods used in the analy,is. Where a licensee wishes to make a change in their facility without prior NRC approval, the effects of the change must be evaluated using analytical techniques and methods which are NRC approved for the application, or which are reviewed and vetted (but not subject to specific NRC approval) in a NRC approved manner.
Direct changes to technical specifications require prior NRC approval. Before changing other operational characteristics described in the UFSAR, a safety evaluation must be performed to determine, among other things, if the change results in a reduction in the level of protection afforded by the TS
[ margin of safety as defined in any TS). Such a reduction would typically occur only if the operational characteristic had been used as a bounding condition in the analysis upon which the selection of TS was based, or in analysis where the acceptability of selected TS values was demonstrated. Licensees can make desired changes to operational characteristics without prior NRC approval, provided that the change does not result in accident analysis results that are nearer the regulatory, or safety, limits than the corresponding results that the NRC used in evaluating the acceptability of the TS during licensing of the facility.
This regulatory position could be codified by adding the following footnote to Section 50.59(a)(2)(iii):
The ' margin ofsafety as defined in any technical specification " (margin of safety) is the amount (quantitative or qualitative) ofmargin between the operation of thefacility as described in the technical speceptations and the exceedance ofsafety limits listed in the technical specipcations or other regulatory limits. In relation to accident analysis, the margin ofsafety is typically the diference between calculatedparameters (e.g., peakfuel clad temperature, maximum RCS pressure, etc.) and the associated regulatory or safety limit. The margin ofsafety is a product of spectpc values and limits contained in the technical specipcations (which cannot be changed without NRC approval) and other values, such as assumed accident or transient initial ;
conditions or assumed safety system response times, which are not specipcally contained in the technical specifica: ions. Any change to the values not specipcally contained in technical l specipcations must be evaluatedfor impact on the margin between the calculated result of an accident or transient and the safety or regulatory limit. Changes, or the net efect of multiple changes, which result in a reduction in the margin ofsafety require prior NRC approval.
Changes, or the net efect ofmultiple changes, which do not cause a reduction in margin of safety do not require prior NRC approval. All evaluatory work in assessing the impact of proposed changes must be performed using methodology and analytical techniques which are either reviewed and approved by the NRC or which are reviewed and vetted in a manner approved by the NRC.
1