ML20059F973
| ML20059F973 | |
| Person / Time | |
|---|---|
| Issue date: | 02/22/1993 |
| From: | Taylor J NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | Rogers, Selin I, The Chairman NRC COMMISSION (OCM) |
| References | |
| NUDOCS 9401190004 | |
| Download: ML20059F973 (10) | |
Text
,
foR 77 f **
.x O
UNITED STATES n.
3-
'E NUCLEAR REGULATORY COMMISSION E
WASHINGTON, D.C.20556 ~
\\
,/
'**d February 22, 1993 MEMORANDUM FOR:
The Chairman.
Commissioner Rogers i
Commissioner Curtiss Commissioner Remick Commissioner de Planque FROM:
James M. Taylor Executive Director for Operations
SUBJECT:
RISK-BASED REGULATION In a March 26, 1992 Staff Requirements Memorandum (SRM), the staff was asked to provide their views on the practicality of risk-based regulation and on the feasibility of developing a transition strategy from deterministic based regulations. This memorandum responds to that request.
.l In a broad sense, risk based' regulation involves the use of probabilistic risk
-)
assessment insights.to focus licensee and regulatory attention on design and operational issues commensurate with their impact on risk to the public. As such, there are many ways in which risk information can be used in the regulatory process (many of which are already in place) and many options for further exploration.
Accordingly, in response to the March 26, 1992 SRM the staff has considered a broad range of topics related to risk based regulation. These topics included:
o the current structure, and nature (e.g. prescriptive vs.
performance oriented) of the regulations and supporting guidance o
the evolution in the use of quantitative risk-information in regulation development (e.g. regulatory analysis, backfit rule, safety goals) o the state of the art in quantitative risk analysis and potential technical and policy issues associated with the further use of quantitative risk information in regulation o
staff and industry initiatives / suggestions for the further
[
incorporation of risk insights into regulation
?
A discussion of these topics and planned actions is provided as an enclosure to this memorandum.
In summary, I believe that further incorporation of risk insights into the regulatory process is practical and should be pursued. As discussed in my December 4,1992 memorandum to the Commission,. I have established a Regulatory Review Group to examine the current regulatory framework with_the purpose of identifying where changes should be made to reduce unnecessary burden-with f
9401190004 930222 PDR: ORG NE E (i t w.
1
4
_ the emphasis on public health and safety. Other related activities, such as the marginal to safety program, will be recognized by the examination as will the further use of performance and risk-based alternatives to existing requirements and practice.
The results from this task group will provide additional insight regarding how and where to pursue changes.
In addition, we encourage and are receptive to initiatives die nted toward the further use of risk based information in regulation.and will keep abreast of those industry trial or pilot programs 'that have the potential to improve safety and/or regulatory efficiency. Where successful, full implementation would be considered.
I believe that these actions meet the intent of the Commission's March 26, 1992 SRM and will result in a feasible transition strategy to a more flexible, focused, and coherent regulatory process improving both safety and/or regulatory efficiency.
I also believe that these actions will result in evolutionary changes to the regulatory process, not radical changes. The ACRS was provided an early version of this paper and in a November 16, 1992, letter to Chairman Selin, expressed concerns about the integration and. coherence of the many risk related programs.
They also stated that they expected to see a later and improved version of the paper.
I fully expect the Regulatory Review Group to have input related to the use of risk information in the regulatory process which will facilitate better integration and coherence. Since the outcome of this review group ' ill play a key role in developing a transition w
strategy, I do not believe further interactions with ACRS on this paper are useful. However, as the work of the review group proceeds and the elements of a transition strategy are identified, we plan further interactions with ACRS.
Originst signed by James M. Taylor James M. Taylor Executive Director for Operations
Enclosure:
As stated cc: SECY OGC ACRS DIST:
Subj-DSIR c/f Circ Chron Distribution List
- SEE PREVIOUS CONCURRENCE Ek RES/DSIR RES/DSIR RES:DD RES:D KING /dpv MINNERS HELTEMES*
BECKJORD*
TAtt0R 1/29/93*
1/29/93*
2/8/93 2/8/93 2/p/93
l ENCLOSURE RISK-BASED REGULATION
==
Introduction:==
The use of probabilistic risk assessment (PRA) in NRC regulatory activities up to now has consisted primarily of the use of generic insights regarding weaknesses of plant design and operation.
In addition, the NRC has encouraged the use of PRA-based tools to enhance our ability for objective decision making. These tools focus attention on the safety objective associated with the regulation or guidance and provide a means to systematically assess the importance and the need for various regulatory actions. The continuing development of PRA tools for regulatory activities is expected to lead to its increased use on a plant-specific basis and to additional flexibility and efficiency for licensees in meeting the intent of the regulations.
The staff believes that the evolving use of risk-based information in the regulatory process will and. should continue. The Regulatory Review Group
~
proposed by the ED0's January 4,1993 memorandum will address as one of its objectives how PRA can be used to provide better focus and more flexibility in the regulatory structure. This paper summarizes the consideration of risk in current staff activities and the staff plan for moving toward a more systematic and extensive use of probabilistic risk-based information and insights in regulatory activities. The planned staff actions represent evolutionary changes to the current regulatory structure and practices, not radical changes.
Backaround:
Risk-based regulation can be viewed as using probabilistic risk assessment insights to focus utility and regulatory attention on design and operational issues commensurate with their impact on risk to the public.
- However, assessing the potential risk to public safety has always been a fundamental part of regulatory decision making.
In the early development of regulations, this assessment was based on qualitative analyses, simple reliability principles and practices (such as worst case or enveloping analysis), defense in depth, and the single failure criterion. The likelihood or probability of the hazard was not always an explicit factor, primarily because the overall state-of-the-art of PRA technology was not sufficiently advanced and accepted.
Currently, due to the advancements made and our increased confidence in PRA, regulatory activities have progressively relied upon risk based insights derived from probabilistic assessments. This is evidenced by development and j
implementation over the past decade of formal processes which rely upon quantitative risk information in decision making (e.g. generic issue prioritization process, regulatory analysis guidelines).
In addition, the safety goals, which are expressed in terms of a quantitative risk, establish top level guidance to support objective regulatory decision making.
4
~
. However, the degree to which quantitative (i.e., probabilistic) risk information has been employed in regulatory activities has been and still is dependent on the state of development of PRA techniques and data bases.
That is, as the uncertainties associated with the techniqu'es and data bases have decreased and their credibility increased, the use of probabilistic risk information in the regulatory arena has expanded.
The Current Reaulatory Process Other than those dealing with processes and procedures, each rule promulgated by the Commission is a result of the identification of a safety concern.
Safety concerns derived from qualitative risk insights have primarily been the bases for most rules because they were developed prior to the availability of PRA. Since its advent, many safety concerns have been identified or assessed by PRA; e.g., PRA generically indicated station blackout as a dominant contributor to residual risk. A safety concern, once identified, is examined to determine whether a rule is needed. This determination is based on adequate protection or enhancement of safety considering cost-benefit and the Safety Goals.
PRA, where practical, is used to evaluate a proposed rule's effect on safety; e.g., the change to the core damage frequency'. However, if adequate PRA information is not available, other quantitative and qualitative risk information is used to the extent it is available and pertinent in making such regulatory decisions.
A rule type can be either prescriptive or performance-oriented in nature.
In a prescriptive rule, the acceptable programs, analyses or hardware are specified.
For a performance rule, the goal or results to be obtained are specified with less emphasis on prescribing the program, analyses or hardware to accomplish these results. A performance rule needs to be supported by acceptance criteria and an evaluation method.
These criteria and methods, which may be defined in the rules or may be part of regulatory guidance (such as a regulatory guide, the Standard Review Plan or Standard Technical Specifications), provide the basis for determining the acceptability of a licensee's response to the regulatory action and are extremely important to assure a common understanding by the NRC, licensees, and the public regarding how the results will be judged.
Risk is reflected in the regulations by developing either prescriptive or i
performance requirements based on either quantitative or qualitative defined risk estimates.
For example, the ATWS rule which is prescriptive in nature was based on quantitative risk insights; that is, the potentially significant contribution of ATWS to core damage and offsite consequences identified the i
need for the ATWS rule (10CFR50.62). The ECCS rule, which is a performance-oriented rule, was based and justified on a qualitative risk assessment; that is, the need for a required defense in depth response (i.e., ECCS) to design basis accidents was identified based on engineering judgement.. In concept, a quantitative risk measure, such as a core damage frequency, can become a performance requirement and be reflected in a regulation or regulatory i
1 guidance.
4
=
~
. Potential for Further Use of Risk Information in the Reaulatory Process The direct use of quantitative risk requirements or performance measures in NRC regulations was considered.
The staff generally believes that such an approach is neither practical nor desirable at this time.
In theory, formulation of such rules could provide greater flexibility to the industry in achieving safety, reduce regulatory burden, facilitate decision making, and allow development of rules that would apply generically to many different designs, including revolutionary designs.
For these goals to be achieved
'without compromising safety, however, the' state of the art for PRA needs to be commensurate with such decision making. The availability of data, the degree to which systems and components are modeled, the understanding of the accident progression, and potential phenomena are examples of the uncertainties and limitations impacting the accuracy of PRA. The PRA methods that would be used to evaluate plants against quantitative probabilistic criteria would need to be specified.
In addition, careful attention must be given to the implementation of probabilistic risk criteria in rules to ensure defense-in-depth and an appropriate balance between accident prevention and mitigation are maintained.
Strict " regulation by the numbers" is not likely to be practical from a licensing and enforcement perspective.
Some current rules do make qualitative statements that systems or components are required to be reliable, but do not specify in the rule what constitutes reliability (e.g., the General Design Criteria). However, most of the current regulatory structure is the result of an evolution which did not have the benefit of a systematic process such as PRA. Due to the large number of suppliers and licensees, technological improvements, accumulated experience, and increased public expectations for higher levels of safety over the years, regulatory requirements and guidance have become more prescriptive. As a result, the staff agrees that the issue of identifying and eliminating unnecessary regulatory burden resulting from overly restrictive requirements should be addressed and the use of the latest PRA techniques should be explored.
The use of PRA techniques as a tool to identify overly restrictive requirements is considered by much of the industry t'o be now quite practical, since most nuclear power plants have or are developing PRAs as a result of their response to the Generic Letter 88-20, " Individual Plant Examinations.
They believe that the state-of-the-art of PRA, in the form of a continually updated (living) PRA, should allow the direct use of quantitative risk measures on a plant-specific basis.
That is, oversight could be increased on the most risk significant structures, systems, and components (SSCs) and decreased for the less significant, as dictated by a living PRA. A living PRA would provide updated assessments of the risk of the plant based on the actual performance of the systems and structures derived from operating experience.
This would result in improving both plant safety and plant efficiency. By reducing the attention and resources on the less safety significant SSCs and focusing attention on the more safety significant, the overall safety of the plant should be improved.
Industry may submit proposals to the staff in this regard.
Overall the staff believes that the use of risk-based regulation could be an effective means for reducing unnecessary regulatory burdens and increasing regulatory effectiveness while also having.the potential to enhance and/or maintain the current level of safety. - Both the NRC and the industry must provide leadership to accomplish this objective. The NRC has the responsibility. and.the means to change. the regulations and regulatory practices while the industry has both the incentive and the more detailed knowledge needed to develop the technical basis upon which changes to current practices could be based. Accordingly, the staff will consider regulatory changes sought by the industry when they are technically justified and do not adversely affect existing plant safety. Further, the NRC will continue to grant exemptions to the industry when ap'propriate requests are submitted in the form of topical reports or plant-specific requests.
Planned Staff Actions following the briefing to the Commission on March 10, 1992, by representatives from NYPA and NUMARC on a risk-based regulation transition strategy, the staff has continued to receive information from the industry regarding their plans and interest in this area. Most recently, on October 14 and December 18, 1992 members of the staff were briefed by representatives from NUMARC and industry on the status of their activities on risk-based regulation.
Based on these briefings, it appears that the staff and industry agree in concept with the following:
(1)
The current structure of a mixture of performance and prescriptive requirements is generally effective, but needs to be re-examined to identify those elements of the regulatory structure which are overly restrictive and where it is both appropriate and practical to use PRA to develop appropriate revisions and assure a proper focus on the most safety significant SSCs.
(2)
Pilot programs need to be conducted to confirm the feasibility of and to establish a standard structure for an expanded use of PRA for regulatory purposes.
(3)
A number of technical and policy issues (see attached Table) need to be addressed as part of the on-going staff activities to provide the basic 4
structure and criteria for using PRA for regulatory activities.
It is likely that industry initiatives directed toward the further use of quantitative risk information will be forthcoming and the staff will consider these.
F T
4 In this regard, the two areas that could have the most near term results are:
(1)
The use of risk insights to improve technical specifications, particularly for surveillance intervals, allowed outage times, identifying the most risk significant SSCs, and assuring a consistent risk perspective.
(2)
The use of risk insights to support changes in existing rules where requirements are not commensurate with safety and to develop new rules and the associated needed guidance.
Staff activities are already underway in both of these areas and we will continue to keep apprised of industry initiatives in these and other areas.
In addition to the above, the Regulatory Review Group, consisting of NRC senior staff and established by the ED0's December 4, 1992 memorandum to the Commission, will examine selected portions of the regulatory structure to.
identify overly restrictive elements. Through public meetings, ACRS briefings, and senior NRC staff oversight, the Group will defin.e the problems and identify the process for correcting them. The Group will also suggest near term solutions based on information and practices currently available.
Inherent in this assessment process will be an examination of the ongoing programs such as the marginal to safety program and the improved standard technical specifications program. Also, the use of pilot programs to evaluate industry initiatives will be pursued.
The results of the Regulatory Review Group will provide recommendations that will serve as input into the staff activities for the transition strategy for the further use of risk information in the regulatory process.
Pilot programs can be used to test the viability and practicality of proposed changes in the regulatory process, as necessary.
The results of these actions are expected to result in evolutionary changes to the current regulatory structure, not radical changes.
TABLE Examples of Key Technical and Policy Issues Associated with Imolementina Risk-Based Reaulati2D The PRA provides a systematic method of studying risk significant contributors in a plant. The practicality of implementing risk-based regulation is dependent upon the practical utilization of PRA. Therefore, it is important that issues associated with utilization of the PRA be evaluated. The following topics are examples of key relevant issues that may need to be addressed depending r, the issue at hand.
Technical Issugi a)
Adequacy and use of analytical tools:
1.
Level of analysis necessary (i.e., Level 1, 2, 3) 2.
Adequate scope level of detail modeled (e.g., should I&C be included? What about SSCs not modeled in PRA?)
hardware failures (passive, active, partial) system interactions human interactions common cause failures effect of aging recovery actions adequacy and reliability of data base external events uncertainty analysis 3.
Plant states needed (full power, low power, shutdown, etc.)
4.
Adequacy of the importance ranking methods 5.
Effect of uncertainty on the ranking process 6.
Effect of human errors on the ranking process 7.
Effectiveness of PRAs to quantify the risk significance of constant changes in plant configuration (configuration control) 8.
Adequacy of methods for failure monitoring and updating the plant PRA 9.
Adequacy of methods for accounting for maintenance, training, management effectiveness 10.
Lovel of review necessary 11.
Adequacy of IPEs for this purpose b)
Availability of Decision Criteria and Methods l.
Availability of decision criteria and analysis methods 2.
Implications of selecting only the high ranking SSCs for regulation (e.g., address the overall safety implications of not regulating low ranked components and systems) 3.
Role of uncertainties in decision criteria and methods
c
_2_
4.
Dependency of PRA results on existing prescriptive regulations, criteria, guidance and how changes to be more risk based ensure safety is not unduly compromised 5.
Threshold of safety significance below which no attention need to be ensured Policy Issues 1.
Consistency of risk-based regulation with standardization, backfitting, severe accident and safety goal policies 2.
Consistency of risk-based regulation with the NRC's inspection and enforcement activities. Adequacy of methods to test compliance or noncompliance 3.
Implications of risk-based regulation for traditional NRC deterministic practices, such as:
defense-in-depth single-failure-criterion 4.
How to best implement risk-based regulation in selected areas:
are license amendments required?
can 10 CFR 50.59 be used?
-}
L
DISTRIBUTION LIST E. Lois M. Drouin P. Norian J. Mazetis T. King W. Minners C. Johnson J. Murphy B. Sheron C. Heltemes E. Beckjord F. Gillespie M. Wolf C. Craig T. Gody T. Murley D. Rasmuson P. Baranowsky D. Ross S. Fonner G. Mizuno l
J. Fitzerald EDO J. Sniezek M. Bridgers (9200084)
A. Summerour l
l I
I i
1 I
.