or the Federal Government have standards gov!rning such emissions. Rather, the primary effects of CAA and State air-quality programs under that Ict are through restrictions on the siting of and missions from fossil-fueled plants, which may in:rease the attractiveness of the nuclear option for lectric utilities.

State Siting Activities

Twenty-five States currently have siting laws. hese include "multistop" regulation by a vari

ety of State agencies, each concerned with a separate aspect of the construction or operation of a plant; State licensing through a "one-stop" agency charged with determining the suitability of all aspects of a particular site on behalf of all State regulatory bodies; or State ownership of the site, with a single agency empowered to administer the terms of a lease with the utility or consortium that owns the plant.

For the last decade, nuclear plant regulation as been slow, unpredictable, expensive, and ustrating for many involved in licensing. Morever, it has failed to prevent accidents such as lose at Three Mile Island and Browns Ferry as ell as construction problems like those experinced at Diablo Canyon and Zimmer. Even in e case of the Byron plants where the OL was enied by the ASLB, the problems were not acted n until the two plants were nearly complete. The ustrations, costs, and uncertainties have resulted extensive criticisms of the regulatory process nd a variety of proposals for changes in that rocess. The focal points for such criticisms are ackfitting,* hearings and other NRC procedures, e current two-step licensing process, NRC reonsibilities not directly related to safety, the use rulemaking, and safety goals.

The principal concerns about nuclear plant regation expressed by utilities and the industry are at neither the criteria nor the schedules for sitg, designing, building, and operating nuclear ants are predictable under the current licensg scheme. The industry and some regulators so complain about the extensive opportunities r public participation in licensing, arguing that

Although "backfitting" technically refers only to design or regtory changes ordered by NRC during plant construction, and cheting" to changes imposed after a plant goes on line, "backfit"' usually is used in the literature to refer to both types of nges. Modifications requested by the permit or license holder termed "amendments" or "variances."

such participation prolongs hearings unnecessarily without adding to safety. They believe that these factors have contributed to higher costs and longer construction times and may have reduced safety by requiring the applicant and the regulators to focus more of their efforts on the procedural aspects of licensing to the detriment of substance.

Nuclear critics, on the other hand, argue that the lack of predictability and construction difficulties were due to the immaturity of the technology and a "design-as-you-go" approach. The critics feel that many of their safety concerns would not have arisen had it not been for the rapid escalation in plant size and number of orders that occurred in the 1960's and 1970's, utility and constructor inattention to quality assurance, and inconsistent interpretation and enforcement of regulations within NRC. While some critics say that nuclear plants will never be safe enough, others believe that the current regulatory process could ensure safety if it were interpreted consistently and enforced adequately. Most critics agree that limiting the opportunities for interested members of the public to participate in licensing will detract from safety.

This section will describe in detail the various parties' views (as determined by OTA) on NRC regulation-what they believe works, what they believe doesn't, and why they hold their viewsand how they think the regulatory process could be improved. These views were solicited by OTA

at workshops and panels involving a broad spectrum of interested parties, including nuclear critics and representatives from utilities, vendors, and AEs. These meetings were supplemented by a survey conducted for OTA in which a small sample of qualified individuals responded to an extensive questionnaire(15). Suggestions for revisions to current NRC regulations and procedures have been made by a number of interested parties. They will be presented in the following text as originally proposed and then evaluated on the basis of the information available to OTA. The discussion of regulatory revisions will focus on two legislative packages currently before Congress: The Nuclear Powerplant Licensing Reform Act of 1983, submitted by NRC (23), and the Nuclear Licensing and Regulatory Reform Act of 1983, proffered by the U.S. Department of Energy (DOE) (21).

The evaluation of proposals for changes in NRC regulation must depend first on an assessment of the goals to be served by regulation and by the individual changes. The primary goal of NRC regulation as defined in the Atomic Energy Act is to ensure that the utilization or production of special nuclear material will be in accord with the common defense and security and will provide adequate protection to the health and safety of the public. Therefore, in analyzing proposals for changes in licensing, the first consideration must be whether they are necessary to further the fulfillment of this goal. If changes would further this

health and safety goal-or at least not detract from it-then secondary policy goals might be:

⚫ to provide a more predictable and efficient licensing process in order to assure license applicants that a plant, once approved, can be built and operated as planned;

• to increase the effectiveness of public par ticipation in licensing; and

⚫ to improve the quality of NRC decisions ir order to increase public confidence in plan safety.

Achieving these secondary policy goals prob ably is a necessary condition in ensuring (whethe for national security, economic, or other reasons that nuclear power remains a viable option fo utilities in choosing their mix of generating tech nologies. However, it should be emphasized tha these goals cannot be accomplished through li censing changes alone. Rather, they also will re quire a commitment to excellence by all partie in the management of plant licensing, construc tion, and operation, as well as a commitment t resolving outstanding safety and reliability issue

Another consideration in evaluating proposal for change in the licensing process is whethe amendment of the Atomic Energy Act is necessar to accomplish a particular change, or whethe it can be accomplished through rulemaking c even simply more effective implementation of th existing regulations.

The utilities' and the nuclear industry's complaints about lack of predictability in reactor regulation focus principally on the potential for changes in regulatory and design requirements during plant construction and operation ("backfitting'').

The present NRC regulations define backfitting as"... the addition, elimination or modification of structures, systems or components of the facility after the construction permit has been issued" (16). Under the present regulations, the standard NRC may use (the language in the regulations

is discretionary) in ordering a backfit is whethe it will "provide substantial additional protectio which is required for the public health and saf ty or the common defense and security."

NRC never has invoked the backfit definitio formally to amend a permit, license, rule, regr lation, or order. Rather, it has changed its requir ments through a variety of less formal procedure such as bulletins, circulars, regulatory guides, ar informal meetings. While NRC has justified th changes on a safety basis, the decisionmakir process has not been as transparent nor as pr

dictable as desired by the industry or its critics. The industry would like to have the backfit rule ewritten and the procedure for invoking it revised so that it would provide greater certainty and stability in terms of costs and schedules and greater flexibility in implementation. The critics would like to see NRC follow an established and documented procedure in ordering backfits to facilitate evaluation of safety considerations.

Specific Concerns

Until recently, NRC's Office of Nuclear Reacor Regulation has been responsible for reviewng and coordinating backfit proposals. The nulear industry has perceived that review to be unystematic, haphazard, and without reference to ny regulatory standard. The industry does not elieve that all of the changes made to plants over he years have contributed significantly to safey. In fact, it considers some of these changes to ave made plant design and operations more omplicated, less predictable, and possibly less afe (8). Moreover, these changes have absorbed large share of the utilities' financial and techical resources. For example, after a decade of peration, there were still hundreds of people working to make changes at the Browns Ferry uclear plants (5). At another utility, the backfits 1980 alone required $26 million and 10 to 12 aff-years of engineering. In addition, the longrm expenditures associated with Three Mile land backfits are estimated to cost $74 million the same plant (26). Thus, there are powerful centives for the nuclear industry to try to have e backfit rule and its implementation changed. Nuclear critics counter that the rule would be lequate if it were implemented consistently and nderstandably. They contend that backfits serve important safety function, since many probms arise only after construction or operation as been initiated. The critics, however, agree ith the industry that it would be more approiate to allocate resources to the design phase ther than using them to satisfy safety concerns Ith backfits. Unfortunately, this is not an option r existing plants, but can be applied only to the ext generation of nuclear reactors.

To review these claims about backfitting, OTA undertook a survey of people of all viewpoints connected with nuclear power, including industry representatives, regulators, and critics. The results of this survey form the basis for the analysis presented here.

There are certain ways in which backfits have the potential to adversely affect the safety of nuclear plants. First, additional equipment can impair normal operations or safety functions; backfits related to seismic protection often are cited as examples of these problems. As discussed in chapter 4, requirements for additional pipe hangers and restraints increasingly have constrained the layout of the piping systems in nuclear powerplants. This could contribute to thermal stresses in normal operation and make the system more prone to failure in accident situations. Another adverse consequence associated with seismic backfits is that they can lead to overcrowding, making some equipment virtually inaccessible for inspection or maintenance.

Backfits also can affect safety by disrupting normal plant operations while new equipment is being installed. While this potential problem is less a result of NRC's management of backfits than of utility planning and expertise, it is important to recognize that there are safety implications associated with installation. Such an incident occurred at the Crystal River plant in 1980 when the utility attempted to install a subcooling monitor while the unit was still operating. This action triggered a series of unplanned events, eventually followed by safe shutdown of the plant.

It should be noted that while examples can be found in which backfits have adversely affected safety, this does not imply that the overall impact has been negative. In fact, one recent study indicates that modifications made to plants after Three Mile Island may have reduced the probability of a large-scale accident by as much as a factor of six at some plants (13). However, the overall gain or loss in safety due to backfits has not been analyzed in any comprehensive fashion.

OTA concludes that, while most backfits represent safety improvements, they can have a negative impact when deployed in a manner

that does not allow for sufficient analysis of the consequences of installing or modifying equipment and its interaction with other systems. A more rational and less hurried approach could improve this situation for current plants. If the next generation of reactors is an evolutionary development of today's light-water reactors (LWRs), new plants should be even less troubled by backfits. New LWR designs will incorporate the lessons learned from Three Mile Island and the accumulated experience of current reactors, and they will address unresolved safety issues with state-of-the-art technology. However, if an alternative reactor design is selected for commercial deployment, it may be impossible to avoid extensive backfitting until the technology is fully

mature.

NRC backfit requirements also have been criticized by the nuclear industry as being overly prescriptive to the point of being incompatible with practical design, construction, or operating techniques. Rather than establishing general guidelines or safety criteria and allowing individual utilities some flexibility in applying them, NRC generally issues detailed and specific requirements. Nuclear powerplant designers must conform to the regulations and appendices in 10 CFR Part 50 as well as 10 other major parts to title 10, over 150 regulatory guides, three volumes of branch technical positions, numerous inspection and enforcement circulars and bulletins, proposed rules, and over 5,000 other voluntary codes and standards that may be invoked at any time by regulatory interpretation. During construction, these standards and codes often are interpreted in the strictest sense possible, with no allowances for engineering judgment. For example, the fillet weld, which is commonly used in field construction, varies in width along the length of the weld. Plant designers recognize that some variation will occur and set the design requirements according to an average width. An inspector, by strict interpretation of an industry code, may not look at the average width, but reject an otherwise acceptable weld if it is slightly less in width than called for by the designer at any point along the length of the weld. Constructors compensate for such anomalies by overwelding, which entails considerable time and expense (19).

It is OTA's conclusion that the requirements associated with the design, construction, and operation of nuclear powerplants are prescriptive and, in some cases, internally inconsistent or in conflict with other good practices. However, while the inconsistencies and contradictions are problematic, the prescriptive nature of the rules should not pose insurmountable difficulties for plant owners and designers. Some utilities have been able to accommodate to the same prescriptive requirements that govern all nuclear construction and still complete their plants efficiently and with few regulatory difficulties. Moreover, NRC is not wholly responsible for prescriptive regulation. The nuclear industry has developed a large and growing set of voluntary standards to provide guidance in interpreting NRC criteria. These standards were expanded greatly in the mid-1970's to match the growth in NRC requirements and often were written with little consideration of their impact. In addition, many of the early standards were written too rapidly to reflect field experience and a convergence of accepted practices. NRC magnified these problems by invoking the standards precisely as written rather than allowing them to evolve gradually (19).

Another concern about backfitting is that there are no clear and consistent priorities. Permit and license holders argue that they have not always been given consistent and stable criteria by which to construct and operate a plant and, as a result, some less important backfits have been imposed before more critical ones. The prime example of this cited by utilities and the industry is the Three Mile Island action plan, in which the NRC gave the utilities no guidance on the relative priorities among approximately 180 requirements of varying importance. The action plan was developed with little comprehensive analysis. As discussed above, if the next generation of plants incorporates a clean-sheet design based on past experience with LWRs, backfits should not be as serious a problem as they have in the past. While a lack of priorities has been troublesome for plants currently under construction or in operation, it is unlikely that future LWRS will experience the same degree of difficulty.

A final concern about backfitting is its potential contribution to increases in construction lead

times and plant costs. These issues are discussed in greater detail in chapters 3, 4, and 5 and are only summarized here. The most recent plants to obtain CPs from NRC required 30 to 40 months after docketing (i.e., not including the preliminary utility planning phase) to obtain their permits, compared to 10 to 20 months between 1960-70. Similarly, construction (the time beween issuance of the construction permit and he operating license) typically takes 100 to 115 months, up from the 32 to 43 months in 1960-70 18). Backfitting has been suggested as one of the ources of delay, along with deliberate delays due o a decrease in the need for power and difficulies in financing construction.

In order to examine the impact of regulation ›n nuclear powerplant construction leadtimes, >TA analyzed case studies of the licensing procss, which are detailed in volume 2 of this report 1). Because it is difficult to separate the effects of backfitting from other regulatory activities, they vere considered in the context of the entire liensing process. Based on these case studies, on ublished analyses of the causes of increases in uclear plant construction leadtimes, and on exensive discussions with parties from all sides of he nuclear debate, OTA has concluded that the egulatory process per se was not the primary Durce of delay in nuclear plant construction. ather, during the 1970's (when leadtimes escaated the most), utilities delayed some plants deiberately because of slow demand growth and nancial problems. Plant size was being scaled p very rapidly and construction was begun with complete design information. The increasing! complex plant designs meant that more mateals-concrete, piping, electrical cable—were reuired, and constructors often experienced deys in delivery of equipment and materials. At e same time, worker productivity declined subantially, at least in part because plants were ore complicated and thus more difficult for the tilities to manage and build (3).

Backfits did lead to delays in some plants, esecially those subject to the extensive regulatory hanges that followed the accidents at Browns erry and Three Mile Island, but in others the efcts of regulatory changes were moderated

through strong management. All plants had to accommodate to some backfits that resulted from the immaturity of the technology and the overly rapid scale-up of plant size. In these cases, regulatory delays must be considered positive. Moreover, in some plants that have experienced regulatory delays, such as Cincinnati Gas & Electric Co.'s Zimmer plant, regulatory actions were an appropriate response to evidence of improper construction practices(9). NRC should not be arbitrarily limited from imposing backfit requirements that lead to long delays in such cases, since interest in the public health and safety should supercede concern for minimizing leadtime and

cost.

In general, OTA concludes that, as in other aspects of quality control, skillful management by the utility, its contractors, and NRC is the key to avoiding delays that otherwise might result from the licensing process. Thus, licensing is most likely to proceed without hitches with experienced, committed utility and contractor management personnel; a clear need for power from the plant; and a constant and open dialogue among NRC staff, nuclear critics, and utility and construction managers. Since skillful management has not been a hallmark of NRC administration, changes to make the organization more responsive and efficient should enhance the licensing process and reduce unnecessary delays. However, such changes cannot substitute for good utility management and a commitment to safety in construction and operation.

Proposals for Change and Evaluation

In 1981, NRC created the Committee for Review of Generic Requirements (CRGR) to respond to some of industry's concerns and to reduce some of the burdens that the utilities felt backfitting had imposed on them. The CRGR review should guide the industry in assigning priorities, even if it does not solve some of the more fundamental problems with backfitting. The NRC and DOE proposals for reform attempt to address the larger issues.

In evaluating the proposals outlined below, it is important to recognize that backfitting cannot