ML18067A630
ML18067A630 | |
Person / Time | |
---|---|
Site: | Palisades ![]() |
Issue date: | 07/07/1997 |
From: | BORDINE T C CONSUMERS ENERGY CO. (FORMERLY CONSUMERS POWER CO.) |
To: | NRC OFFICE OF ADMINISTRATION (ADM) |
References | |
FRN-62FR24997, RTR-NUREG-1606 62FR24997-00033, 62FR24997-33, NUDOCS 9707150295 | |
Download: ML18067A630 (15) | |
Text
{{#Wiki_filter:( I I I lJSCJ 7 *
- Jf/77 F ..
S', 7/ee/Jy /f'?7 A CMS Energy Company July 7, 1997 Chief, Rules and Directives Branch Division of Administrative Services Office of Administration U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 Comments on NUREG 1606, Proposed Regulatory Guidance Related to Implementation of 10CFR50.59 (Changes, Tests and Experiments), Draft Report for Comment The attached comments are provided in response to a request for comments published in 62 Federal Register 24997-24998. We endorse the NRC staff's goal of providing clear, definitive guidance on the interpretation of 1 OCFR50.59. We acknowledge that the NRC's motivation appears to stem from perceived differences in 50.59 program implementation across the industry. We as an industry also benefit from the resulting clarity and uniformity of NRC inspection and enforcement. Quality guidance on 1 OCFR50.59 implementation has existed since 1989 in the form of NSAC 125. This document, although voluntary, has largely became the de facto standard for most of the industry, and by the NRC staff's own admission has resulted in acceptable 50.59 programs at most plants. Some NRC staff members have expressed concern about NSAC 125 not being enforceable as an NRC requirement, but the proposals in NUREG 1606 go significantly beyond that administrative concern. It appears that the staff's desire for enforceability can be achieved by merely making compliance with NSAC 125 (or its replacement, NEI 96-07), appropriately revised, either a mandatory requirement or a formal commitment of the industry. This can be @ 9707150295 970707 ------,\ PDR llfllllflllllllllfllllJllll)lllllllllfll
- Dr f-{f Palisades Nuclear Plant
- 27780 Blue Star Memorial Highway
- Covwl, Ml 4!JU4:i
- 1e1: DtD 764 2913
- Fax: 616 764 2490
..... *
- 2 achieved in the short term without rulemaking or significant policy changes. The changes introduced in NUREG 1606, however, represent significant departures from NSAC 125 and past NRC practice.
Changes as sweeping as these must be developed carefully to assure that the resulting guidance is practical and provides recognizable benefit to justify the implementation cost. As discussed in the detailed comments, many of the proposed NRC positions as defined in NU REG 1606 are not practical for either the industry or the NRC. It is estimated that if the proposed staff positions in NUREG 1606 are adopted by the Commission, Palisades would identify approximately 200 new Unreviewed Safety Questions per year for procedure and plant equipment changes that, in the past, have been routinely performed under 50.59. The additional USQ's would result primarily from the proposed changes in the definitions of Margin of Safety and Types of Malfunctions, from the proposed new reference for determining increases in consequences, and from the impact on probabilities from the proposed restriction against packaging offsetting changes together. The staff proposals in NUREG 1606 would also have the effect of increasing the number of items subject to full 50.59 evaluations (ie, beyond screening) by hundreds of additional items per year. A cornerstone of the NSAC 125 (and Palisades) philosophy is that 50.59 evaluations need to be carefully developed, comprehensive documents. Expanding the scope of 50.59 applicability would introduce a tremendous additional burden on resources with no apparent safety benefit. As a company, Consumers Energy has participated in the efforts of the Nuclear Energy Institute, Combustion Engineering Owners Group, and the Licensing and Design Bases Clearinghouse to develop comments on NUREG 1606. We endorse the comments submitted by those groups under separate correspondence. In view of the potential impact of these proposed 50.59 implementation changes on both safety and the economic viability of Palisades, we have also chosen to submit comments directly to the NRC. We trust the attached comments will be given careful consideration.
SUMMARY
OF COMMITMENTS This letter contains no new commitments and no revisions to existing commitments. Thomas C. Berdine Manager, Licensing Attachment
- ATTACHMENT CONSUMERS ENERGY COMPANY PALISADES PLANT DOCKET 50-255 Comments on NUREG 1606, Proposed Regulatory Guidance Related to Implementation of 10CFR50.59 (Changes, Tests and Experiments)
_I *
- CONSUMERS ENERGY COMPANY COMMENTS ON NUREG 1606 1. Section II, second paragraph
-The purpose of the guidance should also be to define a process that does not waste resources on issues that are not significant to nuclear safety. In a typical year, the Palisades 50.59 program performs and documents 1500-2000 screening level reviews for modifications and procedure changes. Approximately 200 of those are determined to require a full 50.59 evaluation of the issue. Thousands of resource hours are expended on the program in a typical year. Hundreds more resource hours are expended on FSAR maintenance to assure the standard of reference for 50.59 evaluations is up to date. Any change in requirements that would substantially increase the number of issues requiring full 50.59 evaluation would have a significant cost in dollars and technical resources. In addition, revising definitions so that a significantly larger number of proposed changes would be defined as Unreviewed Safety Questions (USQ) which require staff review would not be an improvement in safety. Several of the staff proposals would have that effect. Except in extraordinary circumstances, the staff does not appear able to process even a routine license amendment in less than six months, or for a cost less than about 10,000 to 15,000 dollars (excludes licensee time for preparation of the application). Staff review charges for complex amendments are considerably more expensive. Lowering the threshold for defining USQs (e.g., through the proposed changes in definitions of margin of safety and types of malfunctions, and the expansion of the definition of accident) are likely to cause economic decisions not to incur the dollar or manpower cost of even worthwhile changes. This will be detrimental to safety, not an enhancement. The staff has sufficient technical inspection resources to enforce 50.59 program standards without having to perform in-line reviews of many additional proposed changes. If the staff's standards are practical and clearly defined, uniformity in licensee compliance and staff enforcement should be readily achievable without major drains on resources. With practical and clear standards, there should be few compliance issues for the staff to deal with. 2. Section 111.A.2 -The confusion over what constitutes a change does not stem from a lack of guidance. NSAC 125 provides a reasonably good discussion. Based on NSAC 125, a reasonable and sufficient definition of change is "An alteration which affects the design, function, or method of performing the function of a structure, system, or component (SSC) that is described in the FSAR." We concur with the last two paragraphs of section 111.A.4 that a proposed activity needs to be viewed broadly to assess indirect impacts on other components as well as temporary conditions which may exist while the activity is in progress.
- 3. Section 111.B.4 -The proposed definition of facility is excessively broad in that it encompasses portions of a plant which are both unnecessary and beyond the authority of NRG to regulate.
The facility under this definition would include office and storage buildings, parking lots, heating boilers, etc that are anywhere on licensee owned property within or contiguous to the plant protected area. It could conceivably apply to activities at a non-nuclear plant on the same site as the nuclear generating plant. We 1
- concur that all information contained in the FSAR would be subject to 50.59, and under 10CFR50.71(e}, licensees are required to keep FSAR information correct. As long as the scope of 50.59 is limited to the facility "as described in the FSAR", the staff's proposed definition is adequate.
If other changes are made which extend 50.59 applicability beyond the FSAR, however, this definition will require a significant rewrite. 4. Section 111.C -We concur with the discussion.
- 5. Section 111.D.4 -The proposed definition of test is excessively broad in that it would encompass many routine activities that merely involve gathering data from equipment that is operating in normal configurations.
The terms test and experiment are imprecise without additional amplification. In the context of 50.59, tests or experiments may be of concern only if they could introduce some new, unreviewed risk to plant safety. The fact that a data gathering document is called a "test" does not in itself imply that the activity should be viewed a "test" in a 50.59 context which would thereby impose the burden of preparing a full 50.59 evaluation. Tests and experiments in this context should be limited to activities which place equipment or a system in an abnormal configuration which could jeopardize required safety functions. The definition should also make clear that equipment which has no impact on required safety functions is outside the scope. A definition for tests and experiments which strikes an appropriate balance reads as follows, "A test is an activity performed to verify some predicted response to known conditions or inputs, usually associated with acceptance criteria. An experiment is an activity performed to discover some unknown effect in response to known inputs, and normally does not have acceptance criteria. In the context of 1 OCFR50.59 and 1 OCFR72.48, tests and experiments involve more than merely gathering data. To be a test or experiment in the context of this procedure, the proposed activity must also have some direct or indirect impact on SSC functions which are included in plant or dry fuel storage cask Licensing Bases." 6. Section 111.E.4 -Final resolution of the legal scope of the 1 OCFR50.2 definition of Design Bases is necessary before we can fully clarify the spectrum of 50.59 related issues. In the last sentence of this section, the implied scope of the term "design basis" under 1 OCFR50.2 goes well beyond the legal intent. The term design bases has two very different meanings that must be distinguished from each other. The informal, common usage of the term design bases for a system or component encompasses anything the designer used as a basis for design decisions, whether or not there was any relevance to a safety function. These design bases include many non-safety significant details -such as equipment style and manufacturer, equipment layout, cosmetic finish, amount of operating margin to add beyond minimum required capabilities, etc. In the formal context of 1 OCFR50.2, the term is intended to mean those minimum functions and capabilities necessary to achieve compliance with NRC requirements. This interpretation of the scope of 1 OCFR50.2 Design Bases has been a longstanding NRC policy. NUREG 1397 entitled "An Assessment of Design Control Practices and Design Reconstitution Programs in the Nuclear Power Industry", quotes 2
- the 1 OCFR50.2 definition of Design Bases and then further clarifies it by saying, "The design bases, as defined here, is identical to the definition in 1 OCFR50.2 and includes only the design constraints that are included in current licensing bases and form the bases for the staff's safety judgements (see the definition of engineering design bases)." In SECY-90-365 the NRC reaffirmed its agreement with this philosophy in its endorsement of NUMARC 90-12 which distinguished the formal 10CFR50.2 Design Bases from what it termed "supporting design information", and provided a number of examples to illustrate the difference.
In short, 1 OCFR50.2 Design Bases can be defined as fundamental criteria which, in themselves, must be satisfied to comply with regulations or license conditions. They are embodied in the General Design Criteria which existed at the time of licensing, in other specific regulations (e.g., 1 OCFR50.46), and in plant-specific conditions specified in the license. 1 OCFR50.2 Design Bases, however, do not necessarily include the methods which were chosen by the designer to implement these fundamental criteria. Implementation methods chosen to achieve compliance would be "engineering design bases" as defined in NUREG 1397 or "supporting design information" in NUMARC 90-12. "Engineering design bases" and "supporting design information" contain many items that are voluntarily implemented under the sole authority of a licensee. Broadening the definition of what constitutes a 1 OCFR50.2 Design Basis is a significant policy change that has important implications for reportability under 1OCFR50.72 and 50. 73, and for 50.59 definitions of Unreviewed Safety Questions. The definitions of 1 OCFR50.2 Design Bases and Unreviewed Safety Questions are actually very closely linked. We would contend that a true 50.2 Design Basis (e.g., compliance with GDC xx) can not be relaxed or modified by 50.59 without classifying the change as a USQ. A design feature that can be changed under 1 OCFR50.59, can not be a 50.2 Design Basis. When using 50.59 as the regulatory test to determine whether a change is within the licensee's authority to make, clear definition and understanding of 1 OCFR50.2 Design Bases are needed. 7. Section 111.F -We concur with this section as written, except for the last sentence of the fourth paragraph of section 111.F.4. Specific equipment design or performance criteria for Anticipated Transients Without Scram and Station Blackouts were formally backfit to the industry as new 1 OCFR50.2 Design Bases, and therefore should be appropriately reflected in the FSAR. An inter-system loss-of-coolant event, however, was not backfit as a new Design Basis Accident. See additional discussion below regarding the definition of the term "accident".
- 8. Section 111.G -We concur with the discussion in this section. 9. Section 111.H.4 -The discussion in this section broadens the scope of the term accident significantly beyond NSAC 125 and beyond a reasonable interpretation of the rule. As proposed, the new "accident" definition could encompass any event or condition which the plant is designed to survive. The definition is so open ended that any plant upset, or even some single component failures, could be termed an accident.
For example, this proposal would cause a trip of a secondary side heater drain pump 3
- which creates a need for operators to reduce power (initiate design transient) to be classified as an accident.
A steam leak created by a steam line fitting that cracks or a relief valve that fails to reseat and causes a plant transient to isolate the affected portion could be an accident. The lack of specificity in the staff's proposal would make it impossible to determine the threshold at which a newly postulated event becomes a new accident that would have to be considered a USQ. The definition of accident should be limited to those events that could in themselves cause a degradation of one or more fission product barriers and result in radiological risk to the general public in excess of 1OCFR100 limits. Events which, when coupled with their consequential failures, do not create a potential for significant radiological risk should not be viewed as accidents, but should be addressed under the malfunction questions. Many of the conditions or events a plant must survive are more appropriately handled under the malfunction questions in 50.59. This treatment is consistent with NSAC 125. Most plants have 1 OCFR50.2 Design Bases to be able to withstand events such as a loss of offsite power, site exposure to probable maximum precipitation, earthquakes up to SSE level, fires in plant areas, losses of shutdown cooling, etc. without having those events escalate into challenges to fission product boundaries. Since the events listed would not directly cause fission product boundary challenges, they should not be construed as accidents. The appropriate way under 50.59 to evaluate proposed changes affecting these events is to assess whether the change would create a different or more probable malfunction of the plant equipment relied upon to prevent these events from escalating into accidents.
- 10. Section 111.1.2 -The staff focus on cause of a malfunction goes beyond the words of the regulation dealing with malfunction type. For example, one type of malfunction might be a failure of one RPS logic channel out of four (2/4 logic) due to a power supply failure. For an assessment of the safety impact of the malfunction, the cause of the power supply failure is immaterial.
A proposed change to install a new power supply with a different internal electronic design may create new internal failure modes that did not exist in the old design, but would not affect the power supply or RPS safety functions. Calling this change a USQ is unreasonable, but this designation would be required if the staff adopted its proposal to redefine "type" as "specific cause". Section 111.1.2 should limit its definition of malfunction types to system or function level impacts from component level failures.
- 11. Section 111.1.4 -The proposals in this section excessively expand the scope of the rule beyond what is necessary to assure plant safety. Loosely redefining the term malfunction as" ... an undesired response ... " would thereby encompass all equipment responses we (or NRC) don't like, whether or not those responses have any impact on the ability of the equipment to perform its functions.
As written, the second paragraph is also unreasonable in its attempt to expand the scope of the rule to all functions of all equipment in the facility whether or not those functions affect plant safety. We concur that any change to plant equipment, whether safety equipment or not, should be 4
- screened for any potential negative impact on equipment which has a safety function.
This is true whether the equipment being changed is safety related or not. It is not necessary, however, to do a full 50.59 evaluation on that equipment. Screening is sufficient to identify all safety functions, if any, that may be affected positively or negatively. The scope of the full 50.59 evaluations should continue to be limited to situations where a proposed activity would create some potential impact on safety functions. In the context of 50.59, a legal definition of malfunction should be limited to only those component or system problems that directly impact performance of the required safety function. Malfunction should not be construed to be any equipment problem in any safety or non-safety equipment in the plant, as this section would imply. This would create a substantial burden with no commensurate improvement in safety. A good example of an unreasonable position in this area is the position that a new filled transmitter that replaces an electro-mechanical transmitter would, by definition, be a USQ by creating the potential for a malfunction of a different kind (oil loss vs mechanical failure). In fact, from the perspective of the circuit, a transmitter failure regardless of cause is still a transmitter failure with the same effect on the circuit. Having the staff review this transmitter change (and all others like it) as a USQ would be a significant waste of staff and industry resources. A Licensee is obligated to assure a new component meets all design requirements, including environmental qualification, pressure ratings, accuracies, etc. Requiring staff review of each of the hundreds of new oil-filled transmitter installations that replace obsolete units creates a significant administrative and cost burden with no safety benefit. A similar discussion applies to replacement of old, obsolete analog controllers with state of the art digital units. If a controller change from analog-to-digital must be treated as a USQ, this position would also result in classifying replacement of digital units with analog units as USQs. The practical result of this position will be detrimental to safety if licensees attempt to make do longer with obsolete equipment to avoid incurring the NRC review cost and wasting increasingly scarce resources. It is our experience that staff review can not currently be performed in less than approximately six months on most routine items. Increasing the number of items requiring staff review will only exacerbate a situation that is already difficult. This discussion can be illustrated in the following manner. Palisades currently performs approximately 2000 Safety Reviews [screening level of 50.59 process] per year. Approximately 700 of those relate to some type of modifications to equipment. It is estimated that at least 10% of these, approximately 70, relate to instrument replacements that could have internal component changes -comparable to digital, dry-to-oil-filled, pneumatic-to-electric, etc. All seventy of these changes could be considered Unreviewed Safety Questions under the staff discussion in this section. This number does not even consider the changes to other plant equipment that could have similar minor internal changes that can not affect overall system functions, redundancy, or reliability. 5
- 12. Section 111.J.3 and 4 -Clarification is needed in the staff's position that a change may not be implemented until any associated TS changes are approved by the staff. The position as worded is vague and unduly restrictive.
We certainly concur that prior NRC approval of a modification to remove equipment required by Technical Specifications (e.g., removal of containment spray additive hydrazine system) is necessary before the change can be implemented. We do not concur that the rule prohibits installation of new equipment that should eventually be incorporated into Technical Specifications until after the Technical Specification change is approved. We believe the use of the word "involve" in the rule encompasses only those changes which must be made to avoid violating a Technical Specification as written. The staff position should clearly state that if the new installation may be performed without creating a USQ, and the new installation would not conflict with any provision of existing TS, the rule does not prohibit its installation. It is unreasonable to interpret the rule to say that any plant improvement for which a TS change might eventually be pursued can not be installed, and a literal reading does not require such an interpretation. For example, a number of years ago, Palisades voluntarily made a significant design improvement to its off-site power arrangement to enhance the reliability and redundancy of power sources for its vital AC busses. This change did not conflict with the TS as written, so it was installed under 50.59. Later, a Technical Specifications Change was proposed to the NRC to make the Technical Specifications requirements more stringent and require more of the power sources to be available during plant operation. It is quite possible that this significant safety improvement would have been delayed for years if prior NRC approval had been pursued, and may never have been made. The post-TM! action items offer numerous other examples of improvements (e.g., AFW upgrades) that were installed under 50.59 prior to their eventual incorporation into the Technical Specifications.
- 13. Section 111.K.4 -We partially concur with the staff-position.
It is possible for an NRC directed change to create a safety hazard. It is the licensee's responsibility to assure that any NRG-directed change is safe and prudent to implement. The real test for safety and prudence, however, is not provided by the 50.59 review process. The 50.59 evaluation is a test to determine whether a change meets the regulatory threshold for staff review prior to implementation. The existence of an NRC SER on a plant's docket which directs or approves a change at that plant is generally a sufficient basis for concluding there is no USQ. A USQ means that staff review is required. Since the SER provides the evidence that review is complete and the proposed activity has been approved, the activity can not be a USQ. It is not acceptable, however, for a licensee to take an SER for another plant and assume it also provides authorization for the same activity at the licensee's plant. A 50.59 evaluation should, therefore, be performed even for staff-directed-changes to assure the SER, if one exists, is -applicable and sufficiently comprehensive for the individual facility.
- 14. Section 111.L -We concur with this section. 15. Section 111.M.4 -We concur with the NRC position.
It should be noted that some recent enforcement actions have appeared to use PRA numerical results as the basis 6
- for citing a violation.
Plant licensing bases, as aptly noted in NSAC 125, are based on broad categories of events grouped by perceived event frequency. The categories are then broadly used to define acceptance standards for plant responses. The frequency categories are not based on precise numerical estimates of probability. PRA calculated frequencies that show little or no change provide strong evidence that the event under study still falls within its original frequency category. This is true whether the PRA numerical value shows an increase or decrease. The meaning of the term probability as used by the rule must be viewed in the context of the tools which existed when the rule was published. Expanding the term now to some other use is a significant policy change subject to 1OCFR50.109. In addition, for NRC to issue violations on the basis of PRA numerical results would be inappropriate in nearly all cases. We do encourage the staff to pursue rulemaking or other publicly observable means to define requirements or policies for both industry and NRC staff on the relationship of PRA with existing regulations.
- 16. Section 111.N.4 -The staff position that nothing can be removed from an FSAR is unreasonable and has no apparent legal basis. This position creates a one-way funnel that limits the flexibility of those plants that have made good-faith efforts to maintain their FSARs as comprehensive reference documents.
It rewards those plants that may not have conscientiously added information. The dilemma now faced by the staff and industry is caused by the staff's recent compliance oriented focus on even trivial information in the FSAR. This is motivating plants to defend themselves from unfounded FSAR interpretations. One unreasonable position that some staff members appear to espouse is that any fact mentioned in an FSAR section entitled "Design Bases" must, by definition, be 1 OCFR50.2 Design Basis information. Another unreasonable position that some staff have discussed is that the entire contents of the FSAR should be treated as 1 OCFR50.2 Design Bases. The threat that some staff members will act on interpretations such as these provide powerful incentive for utilities to correct and clarify their FSARS. A staff position that nothing can be removed from an FSAR could prevent many of the needed corrections and clarifications from being made. 17. Section 111.0.4 -We generally concur with the staff's discussion. Several comments are in order, however. First, the staff must take care not to confuse the 50.59 evaluation with a justification for equipment operability or continued plant operation (various definitions of JCO exist in GL 91-18 and other NRC guidance). A 50.59 evaluation is not designed to provide the basis for determining that continued operation is safe or prudent. The timing of a 50.59 evaluation is not a factor in assuring safe plant operation, and one should not be required as part of the decision process. Second, the discussion in (1) is imprecisely written and would have the effect of prohibiting worthwhile compensatory measures until a 50.59 evaluation is prepared. For example, if an automatic control feature were to fail, a prompt compensatory response might be to station an operator at the manual station while troubleshooting 7
- and repairing the automatic feature. Compensatory measures are routinely implemented promptly upon discovery of fire protection deficiencies.
We concur that formalizing compensatory measures to provide a basis for continuing long term plant operation while compliance is restored should be accomplished under normal processes for temporary modifications and procedure changes; and as such, be subject to 50.59 reviews. Existing requirements already have the force the staff seeks, and do not need to be more comprehensive. More restrictive requirements about compensatory measures as proposed in this section could be detrimental to plant safety. It should also be noted here that completing a 50.59 evaluation for a proposed compensatory measure is not the same as completing a 50.59 evaluation of the initiating issue that made a compensatory measure necessary. A 50.59 review of a proposed compensatory measure may be limited to whether or not it is acceptable to implement that measure without regard to why the measure is desirable. An operability justification unrelated to 50.59 would be where the necessary documentation should be found to justify why the compensatory measure is sufficient to maintain equipment operable. It appears that this proposed staff position goes beyond the original intent of GL 91-18 and attempts to backfit a more restrictive policy than defined in the GL. . The discussion in this section appears not to recognize the relationship between an Unreviewed Safety Question, a component that lacks Full Qualification under GL 91-18, and conditions that are reported under 50. 72 and 73 as being Outside the Design Basis. GL 91-18 clearly permits continued operation with nonconforming conditions, even when reported as being Outside the Design Basis. In fact, nonconforming conditions that are outside the design basis are likely to also be USQs. For example, we can postulate a discovery that pipe supports in a seismically designed system were discovered to be damaged such that the system function would still be satisfied in the event of an SSE, but FSAR allowable stress limits would not be met. Under GL 91-18, the degraded system would be considered operable, so plant operation could continue. This discovery that FSAR allowable stress limits were not met would be reportable as being Outside the Design Basis under 50. 72 and 50. 73 even though system operability is not impaired. A 50.59 evaluation of this condition as a proposed change would conclude that intentionally modifying the plant such that FSAR allowable stresses would be violated would result in the proposed change being classified as a USQ. These examples illustrate how it is inappropriate to create artificial distinctions which assume Unreviewed Safety Questions are unrelated to 1 OCFR50.2 Design Bases, reportability decisions, and decisions about continued operation under GL 91-18. In the last paragraph of the section, it is stated that licensees may be permitted to continue operation with a USQ, but would not be allowed to start up with a USQ prior to staff approval. This position is inconsistent with the NRC's responsibility to protect the health and safety of the public. It is implicit in a decision to allow continued operation with certain USQs that the condition may not place the plant in an unsafe condition. Allowing a plant to start up with the same condition, therefore, logically can not place the plant in an unsafe condition. There is no difference in the staff's legal basis for 8
- such a decision whether the plant is operating already or shut down. This policy, therefore, is arbitrary and not related to safety.
Decisions about a plant's configuration or operation should be focused on safety, and should not be made on arbitrary compliance-only bases. 18. Section 111.P.4 -The staff position implies that uncertainty about probabilities shall cause a default to a conclusion that the probability has increased. The position is impractical. It is unreasonable to apply a contemporary definition of probability to a thirty year old rule. We agree that some clarification is appropriate now to reflect advances in the art, but those advances do not supersede the intent and basis of the original rule. It should be sufficient to limit consideration of probability under 50.59 to the event frequency classifications which are aptly described in NSAC 125 and other earlier standards.
- 19. Section 111. Q -We concur with this section. 20. Section 111.R.4 -It is inappropriate in the first paragraph to threaten enforcement action for this new staff definition of an increase in consequences, particularly since the position has not been approved for implementation by the Commission.
In the second paragraph, the basis for this position appears to be that past staff reviews may not have been comprehensive enough or well enough documented in SERs for licensees to understand what factors may have been important; and that an NRC review method based on alternate calculations does not actually grant approval of the method that the licensee happened to have chosen. We believe this position is an unwarranted indictment of the performance of NRC staff. In our experience the bases for staff* approval as documented in an SER have been sufficient to identify the review standards. Where a staff reviewer departed from the Standard Review Plan or other standards for judging acceptability, the alternate acceptance criteria for critical parameters are generally documented to the extent needed. We would think it unlikely that individual staff members would depart substantially from established NRC review criteria (e.g., SRP) without documenting the new criteria in the SER It would also seem unlikely that the extensive management concurrence process within the NRC would have permitted that to happen. In short, the staff has no basis to redefine an increase in consequences to reference a previous analysis rather than the acceptance criteria defined by staff SERs and logic. It should also be noted that it is impractical to assert that no change can be allowed to current predicted values of consequences without specific staff review and recalculation. Al Palisades, of approximately 2000 Safety Reviews [screening level of 50.59 evaluation process] performed each year, about 900 are for procedure changes. Up to one third of these could be related to procedures that might have some dose implications. These would include procedures for radwaste processing, radwaste storage, radwaste releases, operating and emergency procedures relating to accident mitigation, etc. This staff position could, therefore, result in several hundred procedure changes per year being declared USQs requiring NRC dose consequence reviews. It is 9 ,. *
- likely that a plant would choose to accept mediocre procedures long before it would expend the dollars and manpower to process these changes through the NRC. 21. Section 111.S.2 -It is incorrect to state that the staff position in 111.S.4 is consistent with the industry position.
The industry position on a definition of Margin of Safety is clearly defined in NSAC 125 as the margin between NRC's acceptance criterion and the point at which the results become unacceptable (e.g., unacceptable health effects, pipe ruptures). It is the margin inherent in the limits established by the regulator, and under the sole control of the regulator. NRC, by establishing design codes and other requirements (e.g., seismic qualification), is actually establishing a minimum margin of safety to be incorporated into the design. This margin may not be well defined numerically, since the ultimate failure point may not be well defined, but this margin does exist. The regulatory Margin of Safety has not been interpreted by the industry, and should not be construed by the staff, to include the operating margin between a predicted or estimated response and the acceptance limit. For example the design pressure of a piping system as approved by NRC is selected so that the piping system can perform its function under all licensing bases events and conditions. The selection of how much operating margin between expected system operating pressure and design pressure to design in to the system is an entirely voluntary decision of the licensee. It is not accurate on page 26 to characterize the NSAC 125 definition of acceptance limit as being equivalent to failure point. The acceptance limit is the staff's acceptance criterion. One would expect that this criterion would be well below the point of actual failure. The apparent NRC contention that any increase in system operating pressure, even if still below the design pressure, could lead to a significant increase in failure probability is a technically unreasonable position. We concur with the staff position in paragraph
- 2) on page 26 that both the SAR and relevant SERs should be reviewed to define the staff's acceptance criteria.
We strongly disagree with the discussion in paragraph
- 3) on page 26 that the reference value for determining margin of safety should be the actual value of the parameter reported in the SAR, regardless of the acceptance criteria.
This is a fundamental change in philosophy that would greatly increase the number of proposed changes defined as USQs requiring staff approval. In fact, many of the numbers in an FSAR are nominal design values that do not represent limits at all. The practical result of this new definition would be to predispose plants to avoid change entirely by making existing equipment and plant configurations survive until the end of their operating licenses. To . illustrate the.impact on Palisades, we estimate that there are tens of thousands numbers contained within the FSAR. Of approximately 2000 Safety Reviews [screening level of the 50.59 evaluation process] performed at Palisades per year, approximately 200 result in performance of full 50.59 evaluations. Most of these result from some proposed change in a number which appears in the FSAR. If this staff position were adopted, most of these full 50.59 evaluations would result in classification as USQs which would require staff review and approval. 10
- < ,. *
- We would also reiterate comments made to 111.R. This section, like 111.R, appears to contain an unwarranted indictment of the professionalism and competence of the staff in performing past reviews. It is inappropriate to design a severely restrictive, expensive process to correct a problem that doesn't appear to have been widespread, and did not contribute significant safety concerns.
It should also be noted that the position published in Inspection Manual chapter 9900 in April of 1996 should not be viewed as the process currently being implemented by the industry or enforced by the staff. That staff position departed significantly from the de facto industry guidance in NSAC 125, and does not appear to have been evaluated by the staff in accordance with 1OCFR50.109. This IM chapter has not necessarily been implemented by many licensees, and is unlikely to be implemented voluntarily because it is impractical. If this were to become the formal staff requirement after commission review, it would result in a large increase in issues submitted for staff review as USQs. 22. Section 111.T -We agree generally with the staff approach. Since there is no formal definition of the term bases as used in 50.59, however, it is not unreasonable to interpret it as encompassing the theoretical source of the summary information provided by the Bases and typically referenced in the Bases, namely the FSAR. 23. Section 111.U -We generally concur with this section. 24. Section 111.V.4 -We acknowledge that the concept of offsetting probability increases from one source with decreases from another could be abused. Claiming that an increase in accident probability can be offset with greater reliability of mitigating systems may have validity in PRA applications, but it is not always consistent with. nuclear plant design bases compliance. We do not believe, however, that this is a widespread problem that warrants an immediate broad change in staff policy or significant new restrictions on industry. The proper treatment of PSA in the 50.59 context is a subject which warrants a rulemaking proceeding.
- 25. Section IV.A.3 -The discussion in the second paragraph can be correct in that new accident types may be identified by NRC that have not been incorporated into selected licensee FSARs. The examples, however, are written too broadly, and include several events which are not accidents.
For example, a station blackout is an event for which each plant is required to be able to withstand without fuel damage, but would not be defined as an accident. The discussion provided in NSAC 125 attempted to deal with this ambiguity by distinguishing accidents from other design basis non-accident events (e.g., station blackout, loss of shutdown cooling, flooding, severe winds, etc). See additional discussion on this subject in response to Section lll:H.4. Any documents issued by the staff should carefully define the terms used so that all readers can understand and apply the staff positions. It is also emphasized that means other than rulemaking already exist for the staff to achieve more uniform treatment of 50.59 and FSAR updating. Formal staff endorsement of NSAC 125 (NEI 96-07) and issuance of generic correspondence to that 11 ' y *
- effect would in itself provide a measure of consistency as utilities voluntarily refine their programs.
This could be coupled with an NEI initiative or formal NRG/licensee correspondence to make compliance with those guidelines a licensee commitment. Implementation of reasonable staff positions need not wait for completion of rulemaking. Staff positions which depart significantly from NSAC 125, however, including some of those in NU REG 1606, are such fundamental changes that voluntary adoption by industry prior to completion of rulemaking may be unlikely.
- 26. Section IV.B.3 -This section, as written, implies that the published NRC positions are already established as requirements for 50.59 compliance.
In fact, these positions represent significant changes from the principles of NSAC 125, the de facto industry standard since 1989, and past NRC practice. Before these positions are adopted by NRC as established policy, they require backfit analysis under 1OCFR50.109 and formal approval of the commissioners. We believe these positions are extreme and should not be pursued by the staff. If the staff chooses to pursue them, however, the appropriate vehicle would be with significant public involvement in a rulemaking proceeding. 12}}