ML23010A135
| ML23010A135 | |
| Person / Time | |
|---|---|
| Issue date: | 02/09/2023 |
| From: | Alan Kuritzky NRC/RES/DRA/PRAB |
| To: | |
| References | |
| NRC-2022-0085 | |
| Download: ML23010A135 (7) | |
Text
February 9, 2023 Page 1 Response to Public Comments On April 22, 2022, the U.S. Nuclear Regulatory Commission (NRC) issued for public comment a draft report on the Level 3 Probabilistic Risk Assessment (PRA) project; specifically, Volume 3x: Overview of Reactor, At-Power, Level 1, 2, and 3 PRAs for Internal Events and Internal Floods (Federal Register [87 FR 24205], Docket ID NRC-2022-0085). Comments were received from the following individuals and organizations:
Thomas McKenna Nuclear Energy Institute (NEI)
Pressurized-Water Reactor Owners Group (PWROG) [submitted after the comment period]
A synopsis of the comments received and the Level 3 PRA (L3PRA) project team responses to these comments are provided in the table below.
Commenter Synopsis of Comment
Response
T. McKenna The Level 3 PRA should be expanded to:
Reflect the physical health attributed to protective actions (e.g., evacuation, relocations and sheltering)
Address the mental health effects that have been seen following severe nuclear power plant emergencies by putting the radiological and protective action health risks in perspective in a way that the public and decision makers can understand.
Addressing the non-radiological health effects (e.g., mental health effects or risks in evacuation) from a reactor accident are beyond the scope of the L3PRA project. The L3PRA project report (Volume 3d), however, does address numerous risk metrics, including the risk of population affected by intermediate-phase relocation. As stated in Volume 3d:
Metrics that quantify the population affected by intermediate-phase relocation were selected to serve as a surrogate metric for the adverse public health and safety consequences associated with implementing protective actions to avert population dose. Inclusion of these indirect metrics enables a more complete characterization of the adverse offsite consequences attributable to a spectrum of accidental releases from the reference nuclear power plant site by illuminating tradeoffs between radiological and non-radiological health and safety risks. There are three principal reasons for selecting such an indirect measure in lieu of direct measures of adverse non-radiological health effects: (1) estimates of this indirect measure can be calculated using available technology, (2) this indirect measure can provide insights into a range of potential
February 9, 2023 Page 2 adverse non-radiological health effects attributable to implementing protective actions, and (3) results from a recent study suggest that the number of people relocated is a good and relatively straightforward to calculate proxy measure for societal disruption caused by potential nuclear accident scenarios (Bier, et al., 2014).
Separate from the L3PRA project, the NRC continues to examine the efficacy of protective actions, including their benefits and risks.
For example, the analysis in NUREG/CR-7285, Nonradiological Health Consequences from Evacuation and Relocation, dated September 2021 (ADAMS Accession No. ML21252A104) provides insights on the nonradiological health effects attributed to displacement resulting from an emergency event.
NEI The study includes numerous simplifying assumptions that highlight that, for operating reactors, the methodologies for completing comprehensive Level 2 and Level 3 PRAs are not sufficiently mature for widespread use.
The use of simplifying assumptions during PRA development has been addressed by the Commission and PRA standard developers.
First, the Commissions policy statement on the use of PRA (60 FR 42622; August 16, 1995) notes that The Commissions safety goals for nuclear power plants and subsidiary numerical objectives are to be used with appropriate consideration of uncertainties
[emphasis added]
Second, industry consensus standards pertaining to PRA development (including approved standards and standards issued for trial use) provide requirements concerning the appropriate use of assumptions during PRA development. Consistent with the underlying purpose of these requirements, the PWROG sponsored and led peer reviews of both the L3PRA Level 2 and Level 3 PRAs for internal events and internal floods based on these standards, specifically including a review of assumptions. As well as improving the completeness and quality of the L3PRA project models and reports, this process led to further refinement of the standards themselves. In addition, the L3PRA study includes a significant effort for addressing modeling uncertainties for all PRA levels,
February 9, 2023 Page 3 including the performance of sensitivity analyses for those issues deemed to be most risk significant.
Accordingly, the L3PRA team believes that (1) its use of simplifying assumptions is consistent with Commission policy and industry consensus PRA standards, and (2) insights from the study, including insights gained about the use and impact of assumptions, will be useful in supporting risk-informed regulatory activities.
NEI NEI recommends that any reference to this study clearly state the assumptions, conservatisms, and limitations therein.
While addressing this comment is outside the scope of the Level 3 PRA project, the intent of this comment is acknowledged. Further, a goal of the project has been to more fully document the assumptions and limitations associated with the study to facilitate appropriate uses of the study in the future.
NEI The analyzed plant is from 2012 and is not fully representative the current plant design.
This fact is highlighted in the overview report (Volume 3x), where the results are provided of a sensitivity analysis that addresses the major changes in the plant design since 2012.
NEI This PRA would not meet the ASME/ANS PRA Standard.
Due to some project-specific constraints on the Level 3 PRA study (e.g., practical limits on access to the plant and plant information and the need in some instances to rely on analyses and PRA models previously performed by the volunteer licensee), it was not possible to adhere to every requirement in the ASME/ANS PRA standards. However, it is not believed that these shortcomings preclude meeting the objectives of this research study. In addition, all the initial models (i.e., the reactor at-power Level 1, 2, and 3 PRAs for internal events and floods) were subjected to a PRA standard-based peer review sponsored and led by the PWROG.
The results of these peer reviews were addressed in a manner consistent with the goals and objectives of the project.
NEI The documentation should be more readily accessible to the public.
All technical models developed as part of the Level 3 PRA study will be documented in publicly available reports.
February 9, 2023 Page 4 NEI Because the study does not comply with best practices outlined in consensus standards, this study does not fully meet the project objective to demonstrate technical feasibility and evaluate the realistic cost of developing new Level 3 PRAs. Also, the assumptions and simplifications used imply that the resources required for developing a realistic Level 3 PRA are very high.
As stated previously, the study adheres as closely as practical to the PRA standard requirements, and it is not believed that any technical gaps between consensus standards and the study preclude meeting the objectives of this research study.
Insights regarding the resources required to develop Level 3 PRAs will be provided in the summary NUREG (Volume 1) at the completion of the project. However, it is acknowledged that the cost and complexity of performing a Level 3 PRA is very project specific and it is not easy to extrapolate the experience with one project to another. The level of effort depends on many project-specific factors, including:
the scope of the project (e.g., hazards and plant operating states) the number and quality of previously completed PRA models the experience and availability of the analysts the level of familiarity of the analysts with the previous work and the plant itself the level of access to the plant and associated information NEI It does not appear that additional work on this study would contribute towards further fulfillment of the original objectives of the project because the maturity of the methodologies available to complete the work is not sufficient and the resources required would be extensive.
It was acknowledged in SECY-11-0089 that there were several areas that would need to be addressed in a full-scope, site Level 3 PRA where the state-of-practice was not fully mature. As such, the staff recommended that these areas be addressed before commencing the Level 3 PRA study. However, the ACRS recommended, and the Commission directed, the staff to immediately undertake the study and use it to help further advance the state-of-practice in these areas. The study did, in fact, advance PRA methods in several areas, though at the cost of additional time and resources. Some examples of advancements include in the quantification of interfacing systems LOCA frequency, Level 2
February 9, 2023 Page 5 (post-core-damage) human reliability analysis, and improved basis for Level 3 PRA data.
PWROG The release category frequencies (2012 and 2020 cases) for release category 1-REL-LCF seem high compared to a back of envelope estimate. Provide discussion of differences between back of envelope estimate and values calculated in the study. If appropriate, update report to remove excess conservatisms.
The relatively high frequency of release category 1-REL-LCF is primarily a function of the frequencies of station blackout sequences and the modeling of actions to mitigate or prevent releases. In the Level 2 PRA, two operator actions are credited for mitigating severe accidents for non-station blackout scenarios, one prior to vessel breach and one at or shortly after vessel breach (no credit was given for post-core-damage operator actions under station blackout conditions due to the lack of instrumentation). If the two modeled actions fail to prevent containment failure, no further actions are credited in the ensuing timeframe. In the overview report (Volume 3x) this is identified as a major modeling uncertainty and a sensitivity analysis was performed to assess the potential impact of crediting additional operator actions in the longer timeframe.
Regarding the specific back of envelope estimate provided in the comment, three points are germane:
As stated in Volume 3x, most of the LOOP CDF in the 2020-FLEX case arises from operator failure to restore systems after AC power is recovered following an SBO or from failure to successfully implement FLEX strategies.
Also, in cases where offsite AC power is recovered following an SBO, no ELAP is declared, so no credit is given for FLEX.
At the reference plant, recovery of offsite power requires the operation of switchyard breakers that get their control power from non-safety batteries that have a 2-hour lifetime. Since no procedures exist to extend the life of these batteries and the offsite power recovery procedure has a pre-requisite for DC control power to be available to operate the breakers in
February 9, 2023 Page 6 question, there is only a 2-hour window to recovery offsite power.
Specifying the severe accident analysis timeframe to 7 days after accident initiation also contributes to the release category frequency. If additional operator actions could be credited to terminate the accident over this timeframe, then a portion of the frequency could be shifted from the late containment failure end state (1-REL-LCF) to the intact containment end state (1-REL-NOCF).
PWROG As part of a future effort, consider incorporating FLEX failure probabilities developed in PWROG-18042-NP [accession number ML22123A259].
This suggestion will be added to the final summary NUREG (Volume 1) as a candidate for future work.
PWROG Please expand on whether the NRC has come up with a process to adjust the L3 PRA results for the reference plant to reflect other sites to obtain qualitative risk insights, or if the intent of this sentence in the Abstract is to say it would, in theory, be possible to do so.
The intent of this sentence in the Abstract was that it would, in theory, be possible to adjust the L3PRA results for the reference plant to reflect other sites to obtain qualitative risk insights.
However, NUREG guidelines necessitate a very brief Abstract, so this sentence (and many others) does not appear in the revised Abstract.
PWROG Section 2 states that reducing modeling time to approximately 2 days after accident initiation reduces late containment failure to less than 20 percent of CDF. Should this sentence say to less than 20 percent of LERF?
The statement has been corrected to state, Reducing modeling time to approximately 2 days after accident initiation reduces large release frequency to less than 20 percent of CDF.
PWROG Early fatality risk was shown to be almost negligible, and latent fatality risk was shown to be well below the The fact that the reference plant risk was shown to be well below the QHOs associated with the Commissions safety goals is a major insight of the study and will be highlighted in the final summary
February 9, 2023 Page 7 QHO associated with the safety goals. Is it reasonable to think that, assuming other (simplified?) L3 PRAs showed similar results for other plants, such assessments would provide [at least partial] justification for relaxation of CDF/LERF risk targets (e.g., the 1.0E-04/yr target in RG 1.174)?
Consider including a discussion on the impact of risk thresholds in other regulatory documents.
NUREG (Volume 1). The extent to which the impact of this insight is discussed in Volume 1 will be the subject of future internal deliberation. It should also be noted that in SECY-12-0123, the staff identified this as a potential use of the L3PRA project:
Gain insights regarding the use of a core damage frequency (CDF) of 10-4 per reactor year and a large early release frequency (LERF) of 10-5 per reactor year as surrogates for the quantitative health objectives (QHOs) regarding individual risk of latent fatalities and early fatalities, respectively."
PWROG One of the largest contributors to LCF risk is long-term reoccupation of contaminated land. This may have been discussed in other volumes, but were assumptions regarding reoccupation static, or do they change with time? For example, is the decision to allow the population to return based on current population and land-use data, or does it reflect expected changes to population and land-use within the time between the accident and the end of the intermediate phase? List assumptions used to determine reoccupation of contaminated land.
As stated in a footnote at the end of Section 3.3.2 of the overview report (Volume 3x), [t]he use of the EPA intermediate-phase PAGs (2 rem in the year of the accident and 500 mrem in subsequent years) is assumed as a surrogate for decisions on cleanup and reoccupation.