ML25034A080
| ML25034A080 | |
| Person / Time | |
|---|---|
| Issue date: | 03/12/2025 |
| From: | Walter Kirchner Advisory Committee on Reactor Safeguards |
| To: | David Wright NRC/Chairman |
| Nourbakhsh H | |
| Shared Package | |
| ML25071A261 | List: |
| References | |
| Download: ML25034A080 (1) | |
Text
UNITED STATES NUCLEAR REGULATORY COMMISSION ADVISORY COMMITTEE ON REACTOR SAFEGUARDS WASHINGTON, DC 20555 - 0001 The Honorable David A. Wright Chairman U.S. Nuclear Regulatory Commission Washington, DC 20555-001
SUBJECT:
TRIENNIAL REVIEW AND EVALUATION OF THE NRC SAFETY RESEARCH PROGRAM
Dear Chairman Wright:
During the 720th meeting of the Advisory Committee on Reactor Safeguards (ACRS), November 6 through 8, 2024, we completed our triennial review and evaluation of the Nuclear Regulatory Commission (NRC) safety research program, which is primarily conducted by the NRC Office of Nuclear Regulatory Research (RES). Our review included 11 research projects across the RES portfolio. An evaluation of each of these projects is provided in Appendices to this letter report.
In several cases, we wrote letter reports on these projects providing detailed conclusions and recommendations. These letter reports are cited in the references. Our high-level observations on the RES portfolio are provided in this letter.
Executive Summary Overall, the portfolio is well balanced considering the major regulatory challenges facing the agency over the next three to five years: subsequent license renewal, higher-burnup higher-enrichment fuels for the current fleet, and advanced non-light water reactor (non-LWR) licensing applications. The depth and breadth of the on-going safety research program continues to meet the Agencys current needs for anticipated regulatory decisions.
The research program enables staff to maintain core competencies and prepare for reviews of anticipated submittals. This was demonstrated by presentations from the RES staff. The RES staff use a systematic approach to prioritize research emphasizing enterprise risk in project selection, evaluation, and termination. The use of the Future Focused Research program, the establishment and implementation of integrated action plans, and the recent RES leadership of agency-wide initiatives are enabling the Agency to become agile and more proactive in preparing for emerging technologies associated with future licensing submittals. The result is a research program that is having greater impact on agency priorities. These activities are all signs of a healthy research organization and support the Agencys broader efforts to be a more efficient risk-informed regulator.
March 12, 2025
D.A. Wright Background Our research reviews consider the 1997 Commission direction to examine the need, scope, and balance of the safety research program, including how well RES anticipates research needs and how it positions the Agency to understand the regulatory implications of new technologies being developed by industry. In preparing this letter report, we focused our efforts on (a) determining if the research portfolio is meeting current and future agency needs and (b) evaluating the impact that the portfolio is having on the NRC mission.
NRC research activities include conducting confirmatory analyses, developing technical bases to support safety decisions, and preparing the agency to evaluate the safety aspects of new technologies. Through this process, staff competencies are improved, and risk-informed and technically sound regulatory decision-making is facilitated.
Discussion This report highlights selected high level findings from our reviews of the following research projects spanning the three RES divisions:
1.
Source Term Related Activities 2.
Digital Twins 3.
Materials Harvesting 4.
Level 3 Probabilistic Risk Assessment (PRA) 5.
Risk Assessment and Human Factors for non-LWRs 6.
Artificial Intelligence (AI) 7.
Fuel Fragmentation, Relocation, and Dispersal (FFRD) 8.
Advanced Manufacturing Technologies (AMT) 9.
Artificial Intelligence and Machine Learning in Non-Destructive Examination (NDE) and Inservice Inspection (ISI)
- 10. Computer Code Development and Validation for Non-LWRs
- 11. High Energy Arc Faults (HEAFs)
Summaries of our findings, conclusions, and recommendations are provided in Appendices to this letter report. Overall, the portfolio is well balanced. The following five themes arose repeatedly during our meetings: focus and communication, engagement, professional development, impact, and future activities.
Focus and Communication. In general, we observed good coordination between the Office of Nuclear Reactor Regulation (NRR) and RES staff, and RES is focused on NRR needs. This linkage enhances communication and allows for focused mission-centric research that will provide actionable and impactful outcomes for regulatory decision making and NRC safety review activities.
Engagement. We observed that the research personnel are well engaged with parallel activities underway in industry (e.g., Electric Power Research Institute (EPRI), American Society of
D.A. Wright Mechanical Engineers (ASME), Institute of Electrical and Electronics Engineers (IEEE)). This engagement enhances the research teams understanding of industry plans and allows industry to appreciate the corresponding regulatory needs for anticipated upcoming licensing actions.
The engagement also enables the research team to avoid duplication of industry activities and perform the confirmatory research necessary from a regulatory perspective.
Professional Development. The research portfolio is helping NRC staff become better informed about new technologies that industry is considering (e.g. AI/Machine Learning (ML) for inspection, digital twins, advanced manufacturing, and new reactor technologies) for future applications. It also builds useful staff experience and expertise and supports continuing professional development and the agencys knowledge management program.
Other projects, such as the Level 3 PRA and the work being done on non-reactor risk applications, are providing unique insights and a plethora of risk data that will serve the agency well as it becomes a more risk-informed regulator and begins to use risk in decision making beyond the realm of power reactors.
The development of reference plant models for each of the advanced reactor technologies has been invaluable for the staff to understand these systems in advance of licensing applications.
Today the staff is ready to perform confirmatory analysis for anticipated near-term advanced reactor applications.
Impact. The RES research portfolio is having real impact on regulatory decision making and reducing uncertainty in technical areas including:
Providing the technical basis for source terms from the MELCOR severe accident calculations to support an upcoming revision of Regulatory Guide (RG) 1.183; Supporting regulatory decisions by performing scoping calculations using RES developed non-LWR system analysis tools for advanced reactor applications like the Kairos Hermes reactor; Highlighting potential safety issues via synthesis of the existing database associated with FFRD in a timely manner as the industry plans for higher burnups and higher enrichments and as the staff is working on rulemaking language associated with higher enrichment fuels for the current Light Water Reactor (LWR) fleet; Informing Part 53 operator training requirements through human factors research; and Characterizing the potential safety issues associated with high energy arc faults, leading to maintenance and design enhancements that can reduce the risk of such events.
Future Activities. We offer the following observations and recommendations related to ongoing R&D activities:
Digital twins, AI, and advanced manufacturing are emerging projects, and current efforts appear reasonable. As a practical matter, these projects are limited by personnel and budget. The agency should enhance collaboration with institutions applying greater expertise and resources in these areas. The goal should be to have adequate agency technical knowledge and subject matter expertise to make sound regulatory decisions.
Insights from Level 3 PRA research should be leveraged for risk-informed decision making, as noted in Appendix D.
D.A. Wright
- The non-LWR code development activity requires additional effort, including verification and validation to support confirmatory analyses. Maintaining this capability will require continued commitment of resources by the NRC and coordination with Department of Energy (DOE) to meet current anticipated licensing timelines.
Updates to LWR source terms and the impact of FFRD at higher LWR burnups on licensing options are also anticipated as part of the increased enrichment rulemaking.
The RES Future Focused Research program is proving to be a valuable vehicle to enable the agency to meet its mission in light of dynamic changes in the range of technologies being considered in future nuclear energy systems.
We look forward to additional briefings on these projects as results become available.
Sincerely, Walter L. Kirchner Chair APPENDIX A:
Source Term Related Activities APPENDIX B:
Digital Twins APPENDIX C:
Materials Harvesting APPENDIX D:
Level 3 PRA Project APPENDIX E:
Risk Assessment and Factors for Non-Light Water Reactors APPENDIX F:
Artificial Intelligence APPENDIX G:
Fuel Fragmentation, Relocation and Dispersal APPENDIX H:
Advanced Manufacturing Technologies APPENDIX I:
Artificial Intelligence and Machine Learning in Nondestructive Examination and Inservice Inspection APPENDIX J:
Computer Code Development and Validation for Non-Light-Water Reactors APPENDIX K:
High Energy Arcing Faults APPENDIX L:
List of Acronyms Signed by Kirchner, Walter on 03/12/25
D.A. Wright REFERENCES 1.
Advisory Committee on Reactor Safeguards, Integration of Source Term Activities in Support of Advanced Reactor Initiatives, April 4, 2022 (Agencywide Documents Access and Management System (ADAMS) Accession No. ML22069A083).
2.
Advisory Committee on Reactor Safeguards, Interim Letter on Level 3 Probabilistic Risk Assessment - Volumes 3 And 4, Pertaining to Reactor At-Power Events, November 24, 2023 (ADAMS Accession No. ML23317A199).
3, Advisory Committee on Reactor Safeguards, Research Information Letter (RIL) 2021-13 on Interpretation of Research on Fuel Fragmentation, Relocation, and Dispersal at High Burnup, December 20, 2021 (ADAMS Accession No. ML21347A940).
4.
Advisory Committee on Reactor Safeguards, Review of Computer Code Development and Validation for Non-Light-Water Reactors, May 24, 2024 (ADAMS Accession No. ML24129A189).
D.A. Wright
SUBJECT:
TRIENNIAL REVIEW AND EVALUATION OF THE NRC SAFETY RESEARCH PROGRAM Accession No: ML25034A080 Publicly Available (Y/N): Y Sensitive (Y/N): N If Sensitive, which category?
Viewing Rights:
NRC Users or ACRS only or See restricted distribution OFFICE ACRS SUNSI Review ACRS ACRS ACRS ACRS NAME HNourbakhsh HNourbakhsh LBurkhart RKrsek MBailey WKirchner DATE 02/03/25 02/03/25 02/4/25 03/3/25 03/7/25 03/12/25 OFFICIAL RECORD COPY March 12, 2025
Appendix A 1
APPENDIX A SOURCE TERM RELATED ACTIVITIES
Background
The accidental releases of fission products (Source Term) play a critical role in the nuclear facilitys design and the agencys requirements to protect public health against radiation hazards. The regulatory basis for source term is widely dispersed among numerous documents, largely focused on light water reactors (LWRs). The original bounding source term developed in 1962 within TID-14844 was updated following extensive research performed in the decades following the accident at Three Mile Island (see NUREG-1465 and NUREG-1150). Regulatory Guide (RG) 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, developed the significant attributes that define the alternate source term and its application for both pressurized water reactors and boiling water reactors, the basis for which comes from those earlier published NUREGs. Emergency planning regulations now use source term information from the facilitys safety analysis report or probabilistic risk assessment (PRA). Guidance for establishing source terms associated with a technology-inclusive, risk-informed approach to support radiological dose assessment for emergency planning is found in Appendix B of RG 1.242, Performance-Based Emergency Preparedness for Small Modular Reactors, Non-Light-Water Reactors (non-LWRs), and Non-Power Production or Utilization Facilities, Revision 0.
A mechanistic source term is the result of an analysis of fission product release from fuel and its subsequent transport from the fuel through the reactor coolant system, through all holdup volumes and barriers, considering mitigation features, and finally, into the environment. In a July 30, 1993, staff requirements memorandum (SRM) (SRM-SECY-93-092), the Commission approved the staff recommendation that scenario-specific (mechanistic) source terms be allowed. This included a caveat such that there should be sufficient understanding of fuel performance, fission product behavior, and accident selection to bound uncertainties.
SECY-16-0012, Accident Source Terms and Siting for Small Modular Reactors and Non-LWRs, provide information for calculating a mechanistic source term and its relationship to functional containment. The RG 1.233, Guidance for A Technology-Inclusive, Risk-Informed, And Performance-Based Methodology to Inform the Licensing Basis and Content of Applications for Licenses, Certifications, and Approvals for non-LWRs, provides additional considerations related to mechanistic source terms as part of establishing the licensing basis for non-LWRs. Finally, NUREG-1537 provides guidance on source term development for a maximum hypothetical accident for non-power production and utilization facilities that may be useful for simpler, smaller microreactors. There is significant flexibility in how to establish the accident source term for reactor designs under consideration. Approaches can range from performing simple conservative bounding assessments (to support a maximum hypothetical accident approach) to more complex detailed analyses that model expected fission product releases across a spectrum of postulated events including the effects of uncertainties.
ACRS Review and Evaluation The Committee heard numerous briefings related to reactor source terms ranging from the technical basis for source terms for higher burnup LWR fuels in support of RG 1.183 to a broad discussion about integration of source term activities in support of anticipated advanced (non-LWR) reactor applications.
Appendix A 2
Establishing a defensible source term requires an integrated understanding of the physics and chemistry associated with fission product release and transport coupled with the reactor system accident response. To assure there are no gaps in the technical basis for the source term that can impact an expeditious licensing review, we emphasized the following elements for consideration by the staff:
Source terms should consider both radioactive and chemically hazardous materials as well as releases due to interactions between these materials. Because some advanced reactors use hazardous materials in their designs, the chemical source term from the facility will need to be evaluated for both workers and the public.
Source term estimates should be based on actual experimental data considering the following important effects: time at temperature, volatility of specific fission products, chemical environment effects and important fuel characteristics (e.g., burnup). When data are sparse, conservative bounding estimates may be necessary.
Understanding the relevant physics and chemistry is critical to accurate modeling of the physical form of the fission products. Aerosols can be important in many advanced reactor designs (e.g., fires associated with alkali metal coolants, aerosolization of non-water liquid coolants, dust in some gas reactor designs). In some cases where there is a lack of external aerosol sources, assuming all fission products are vapors may be more appropriate.
The necessary level of sophistication in modeling the transport of fission products through multiple barriers using a functional containment approach is strongly dependent on design details and can vary. Sometimes complex modeling using a fluid dynamics code that can capture the behavior of vapors and aerosols during transport is required. In other cases, simpler transport modeling such as a lumped parameter model can be effective. The optimal approach depends on the technology (fuel, coolant, moderator), the design of the reactor and surrounding enclosures/buildings, and the nature and progression of the postulated accident(s). But in all cases, a conservative approach to confinement barrier retention should be assumed based on the influence of relevant service conditions and accident environment on barrier effectiveness.
The Department of Energy (DOE) Handbook 3010-94 provides useful information for a wide range of potential source term constituents that could be applied to advanced reactors.1 HIGHER BURNUP/HIGHER ENRICHMENT LWR APPLICATIONS Reactors in the current fleet that wish to obtain higher burnup with their reactor cores will need to address the potential for increases in accident source term related to increased burnup. The staff briefed ACRS on the technical bases for updated LWR source terms to be incorporated into revision 2 of RG 1.183. The MELCOR simulations of a range of the early phase of loss of coolant accident ( LOCA) initiated severe accidents were evaluated to establish source terms to containment for both pressurized water reactors (PWRs) and boiling water reactors (BWRs).
The new source terms were based on improved and more recent understanding of the chemical forms of key fission products and MELCORs improved ability to calculate the trajectory of the accident, especially enhanced modeling of the reactor coolant system.
1 Department of Energy, Airborne Release Fractions/Rates and Respirable Fractions for Nonreactor Nuclear Facilities, DOE Handbook, DOE-HDBK-3010-94, December 1994.
Appendix A 3
The MELCOR calculations suggest that increased burnup and enrichment do not impact the containment source term and releases to containment are significantly reduced if the pressure boundary remains intact. While earlier studies prescribed a large number of high-pressure scenarios, examination of a broader spectrum of accidents (as required by RG 1.183) revealed a number of important low pressure accident scenarios. Early loss of the pressure boundary (via creep rupture of the pressurizer line in PWRs and thermally induced safety relief valve seizure for BWRs) as calculated by MELCOR resulted in depressurization of the primary system and a pathway for release to containment early in the accident progression. The accident progression in the low-pressure scenarios allowed more time for the core to heat up and release fission products. This longer time at temperature resulted in a greater extent of core damage and thus higher releases of halogens, tellurium, and alkali metals.
ADVANCED REACTORS: The research staff has devoted significant effort to being ready to evaluate applications across a range of reactor technologies (e.g., gas, sodium, salt, and heat pipe reactors). They have adapted their computer models (e.g., MELCOR, SCALE) and then applied them in the non-LWR demonstration project to idealized designs available in the open literature to understand the system response to accidents and the relevant physics and chemistry associated with the source term. Particularly notable is the work on analysis of accident progression in a heat pipe reactor, a concept that has not received as much attention as the other technologies under consideration. Although design-specific evaluations are needed, these reference plant evaluations have identified key phenomena, data gaps, and accident system response features that impact source terms. These insights can help inform designers and regulators where the uncertainties in source term determination are the greatest and where other aspects of the source term have less impact, focusing attention where data may be required. This activity should substantially increase the readiness of the staff and promote expeditious reviews of current and future non-LWR applications.
The staff presentation on integration of all relevant source term activities contained a wealth of information (both in depth and breadth) that can serve as a starting point for potential applicants going forward. The use of a web page as a one stop shop is a good way to capture this information and keep potential applicants up to date on latest work and progress in this area.
Given the rapidly evolving nature of this topic, the use of a web page to capture relevant information is preferable to formal documented guidance. We recommended that the staff provide an overview on the web page explaining how an applicant can best use the available information in concert with pre-application consultations to be better prepared to develop high quality submittals. We also recommended that non-LWR applicants could also benefit from consolidated guidance on the acceptable attributes of the source term.
Summary The research being conducted on source term is critical to the NRCs mission and has high impact. Development of a source term that can receive regulatory approval is critical to the success of any new reactor design and licensing. The research has focused on source terms related to high burnup LWR fuel, evaluating accident progression and source terms for notional advanced reactors, and improving clarity of guidance of how to establish a credible source term for advanced reactors. The staffs use of a web page to contain all of the relevant information is an excellent communication tool for potential applicants. NRC staff has expended significant effort related to computer code development and application for non-LWR technologies. All of these NRC efforts should allow expedited staff reviews.
Appendix B 1
APPENDIX B DIGITAL TWINS
Background
A digital twin (DT) is a virtual representation of an entity, process, or system, synchronized at a frequency and fidelity sufficient to maintain state concurrence. A DT leverages various types of models, data, and frameworks to produce knowledge/insights about the represented entity, process, or system to fulfill an intended purpose.1 The potential application areas of applying DTs in the nuclear industry include design, plant construction, training simulators, predictive operations and maintenance, and autonomous operation. The Department of Energy (DOE)
Advanced Research Projects Agency - Energy (ARPA-E) has been heavily investing in innovative digital technologies research and development aimed at enabling advanced reactor development. ARPA-E has had several interactions with the NRCs Office of Nuclear Regulatory Research (RES), specifically on DT technologies. The Electric Power Research Institute (EPRI) has also explored benefits, challenges and potential applications of digital twins in advanced reactors.
The RES staff initiated efforts to assess the regulatory viability of DTs for nuclear power plants (NPPs). RES recently completed a Future Focused Research (FFR) project aimed at assessing the current state of DT technology and potential applications for the nuclear industry.
ACRS Review and Evaluation On May 4, 2022, the ACRS was briefed by the RES staff on the agencys FFR activities regarding digital twins. We also heard from representatives of ARPA-E and EPRI. The effort on digital twins is an example of how RES is using FFR projects to prepare for anticipated future submittals. A digital twin is a representation of a physical system such as a nuclear reactor. The core idea is that, rather than simply running such models offline, the digital twin is kept current with the actual condition of the physical system by using extensive real-time (or near-real-time) sensor data, to achieve improved plant analysis and control. The research is exploring challenges with the advanced sensors and instrumentation that would be used in digital twins, as well as needed guidance, standards, and potential regulatory frameworks for the use of digital twins by stakeholders.
This research program is also actively involved in communication and knowledge managementfor example, to learn from what is being done by EPRI and the Department of Energy and to communicate what the team has learned to NRC staff and industry stakeholders.
As part of this effort, the NRC held an observation public meeting with representatives from EPRI, ARPA-E, and the Nuclear Energy Institute (NEI) to discuss regulatory considerations and opportunities for digital twins in nuclear reactor applications, as well as the relevant research efforts of each organization.
The presentation by ARPA-E focused on the Generating Electricity Managed by Intelligent Nuclear Assets (GEMINA) program, focused on the potential to reduce operating expenses for advanced reactors, through extensive reliance on condition monitoring. The presentation also noted that ARPA-E is coordinating with EPRI, NEI, and the International Atomic Energy Agency (IAEA) on the development of a standard for digital twins.
1 https://www.nrc.gov/reactors/power/digital-twins.html
Appendix B 2
The presentation by EPRI focused on the question of where it makes sense to use digital twins considering factors such as technology readiness, costs, benefits, scalability, and applicability.
Their current approach emphasizes focusing on near-term use cases (i.e., starting small and then scaling up, before moving toward automation), and prioritizing standardized solutions and user-friendly interfaces.
ACRS questions that were discussed included the fidelity of the digital twin, issues related to cybersecurity, and the relative merits of using digital twins in an offline manner (e.g., in support of predictive maintenance) versus to support remote operations for future autonomous power plants. We also appreciated the fact that Division of Engineering (DE) is leveraging resources and skills through collaboration with other organizations.
Summary The use of the FFR program to study digital twins is helping the agency prepare for possible future licensing submittals. It appears that agency research on this topic is appropriately scoped to identify potential challenges associated with use of this technology and support eventual development of standards and guidance. We look forward to hearing more about this technology as the work progresses.
Appendix C 1
APPENDIX C MATERIALS HARVESTING
Background
Material harvesting involves removing irradiated materials from nuclear facilities that have ceased operation. Research conducted using harvested materials continues to be an opportunistic endeavor that requires the willingness and coordination of companies that have decided to shut down and decommission their nuclear facilities. Therefore, opportunities are limited and not necessarily systematic regarding strategic needs or priorities. It is also relatively expensive to carefully remove items while maintaining their material properties versus straight demolition. Nonetheless, the Office of Nuclear Regulatory Research (RES) has been able to participate in a few projects. Research on harvested material can provide interesting insights for the actual performance of materials in real-world operating conditions.
ACRS Review and Evaluation In October 2022, the ACRS was briefed by the staff on RES materials harvesting activities. Our review identified three comments:
1.
Staff reported that in some cases, documentation for the harvested material is not existent, difficult to retrieve or has no longer been retained by the owner. This in part may be due to the definition of maintained for the life of the facility that is generally interpreted as ending when the licensee submits their letter of license termination. Loss of the pedigree of harvested material diminishes its value as needed assumptions add uncertainty to the analysis.
2.
Harvesting of material for electrical cable aging management should consider the availability of splices in cables, because these splices can become the life limiting component in some cases.
3.
Perhaps an even more valuable output from material harvesting research would be the refinement of predictive aging models that could be used on other related materials in a proactive manner.
Summary Research conducted using harvested materials continues to be a relatively expensive, opportunistic endeavor that is not necessarily systematic regarding strategic needs or priorities. RES has however been able to participate in a few projects. Research on harvested unique irradiated materials could be leveraged to improve industry initiatives such as reactor vessel embrittlement and other life limited component databases and associated regulatory evaluations.
Appendix D 1
APPENDIX D LEVEL 3 PRA PROJECT
Background
The Office of Nuclear Regulatory Research (RES) staff is performing a full-scope site Level 3 probabilistic risk assessment (PRA) for a reference plant with two Westinghouse 4-loop pressurized-water reactors (PWRs). The staff undertook this project in response to Commission direction.
The objectives of the Level 3 PRA project are:
Develop a Level 3 PRA, generally based on current state-of-practice methods, tools, and data, that reflects technical advances since the last NRC-sponsored Level 3 PRAs (NUREG-1150), which were completed over 30 years ago and addressed scope considerations that were not previously considered (e.g., low-power and shutdown risk, multi-unit risk, other radiological sources).
Extract new insights to enhance regulatory decision-making, and to help focus limited NRC resources on issues most directly related to the agencys mission to protect public health and safety.
Enhance PRA staff capability and expertise and improve documentation practices to make PRA information more accessible, retrievable, and understandable.
Demonstrate the technical feasibility and evaluate the cost of developing new Level 3 PRAs.
ACRS Review and Evaluation The Level 3 PRA project includes Level 1, 2, and 3 PRA models and results for internal events, floods and fires, seismic events, high winds, and other external hazards, as well as other plant operating states, and other site radiological sources (i.e., spent fuel pools and dry storage casks). A potential use of the methodology and insights generated from the Level 3 PRA project is to inform regulatory, policy, and technical issues pertaining to advanced LWR and Non-LWR applicants that are using the Licensing Modernization Project (LMP) framework and are required to perform Level 3 PRA analyses.
The ACRS has been briefed numerous times since the project began in 2012. The most recent briefings presented the integrated results of the reactor at-power PRA, for Internal Events and Floods, Internal Fires and External Events.
In the related letter, the ACRS recommended that additional attention should be placed on insights from the work in the following areas: technical limitations of the work, the limited scope of FLEX strategies assessed (only Phase 1 and 2), the innovative level 2 human reliability analysis, technical rationale for the very low early fatality and latent fatality risks, the technical basis for various mission timelines used in the assessment, and a discussion of key uncertainties (both considered and not considered in the study) and their influence on the overall projects conclusions.
Not all the Level 3 PRA expected analyses and results could be completed due to unforeseen circumstances. Nevertheless, the project is expected to meet its original objectives; in the case
Appendix D 2
of FLEX, it continued the analysis beyond the original plan. At the completion, the value of the project will be enhanced by providing a clear statement of insights gained in the performance of the study. The Volume 1 Summary Report, which will be published after all technical work for the Level 3 PRA project has been completed, will provide valuable information on general PRA related issues, including an exposition of major assumptions, sensitivity analyses, treatment of uncertainties, modeling limitations, and possible enhancements for future PRA efforts.
Finally, an important ancillary benefit of this work is the training of a team of next generation risk analysts that should serve the agency well as it continues to integrate risk information in its regulatory decision making.
Summary When completed, the Level 3 PRA study will be the most comprehensive full-scope PRA performed by NRC. The coverage of the PRA subject matter, including risks associated with severe accidents, is extensive. It applies experience gained over the 30 years since NUREG-1150, providing valuable insights related to regulatory decision-making and PRAs documentation, technical feasibility, and cost.
As such, the Level 3 PRA project provides a substantial source of very useful information in support of licensing advanced reactor designs that are using the LMP framework and are required to perform Level 3 PRA analyses. The results of the Level 3 PRA, particularly regarding integrated site risk, treatment of uncertainties and model limitations, application of Level 3 PRA sequences, etc. should provide ways to enhance the safety focus of reviews for future reactor designs. In addition, these Level 3 PRA results for an operating reactor site could be used for testing the risk informed framework for future reactor designs.
Given that results from this work will provide important risk insights for regulatory decision-making, resources should be prioritized to ensure that the remaining documentation is issued without significant delays.
Appendix E 1
APPENDIX E RISK ASSESSMENT AND HUMAN FACTORS FOR NON-LIGHT WATER REACTORS
Background
Non-Light Water Reactors (Non-LWRs) (also known as advanced reactors) offer important improvements in safety through their use of passive safety systems and inherent safety characteristics. However, these attributes also challenge historical approaches to risk analysis and risk-informed decision making that are needed as part of the NRCs review and approval of licensing applications.
The Division of Risk Assessment in the Office of Nuclear Regulatory Research (RES) has embarked on new research related to risk analysis for advanced reactors to continue to grow the agencys risk informed decision-making (RIDM) activities. Efforts are focused in three major areas: (a) fire risk evaluation methods, risk tools and data for advanced reactors, (b) methods and tools for evaluating the risk of advanced reactors, and (c) scalable human factors reviews including remote, autonomous and multi-unit operations.
Fire Risk Assessment The Non-LWRs could introduce new risks related to fires. There is limited specific guidance available, especially for low power shutdown (LPSD) plant operating states (POS). During LPSD POSs, breaching, opening or removing fire barriers may occur leading to fire propagation.
Furthermore, historical results for LWRs show the risk during LPSD is similar to risks when a unit is at power. The staff is evaluating the current state of knowledge and determining how current guidance and knowledge can be leveraged to assess impacts of fire during LPSD POSs.
They anticipate a research information letter on the topic to support NRR licensing reviews.
RIDM for Advanced Reactors As part of the agencys approach to RIDM for advanced reactors the research staff is considering the following topics: non-LWR probabilistic risk assessment (PRA) acceptability, treatment of uncertainties, non-LWR RIDM guidance, appropriate non-LWR risk metrics, operating data for non-LWRs and how to deal with passive systems in a risk analysis.
Trial use RG 1.247 has endorsed the non-LWR PRA standard and the Nuclear Energy Institute (NEI) peer review guidance. Staff is working on finalizing RG 1.247 and establishing a public web site for staff positions on the Regulatory Guide (RG). They are also evaluating if updated guidance is needed for non-LWRs. They are also considering if there are alternative risk metrics for non-LWRs that better capture the risk of these advanced reactors. These activities are leading to an Interim Staff Guidance (ISG) to support of Office of Nuclear Reactor Regulation (NRR) licensing reviews.
The staff has also noted the need for advanced reactor operational experience data to inform risk modeling. While there are databases on most of the advanced technologies (e.g., sodium, gas and molten salt), they differ in the depth and breadth of the information that is available. The risk staff also rightly notes the specific challenges for estimating passive system reliability that will be needed in a PRA. Given the challenges anticipated, they conclude that a mixture of the limited operating data, expert/engineering judgement and simulations coupled with safety margin will be needed to make RIDMs meaningful for advanced reactors.
Appendix E 2
Human Factors for Advanced Reactors Current human factors practice for LWRs is based on guidance found in NUREG 0800, -0711 and -0700. The draft rule 10 CFR Part 53 for non-LWRs notes the need to reflect state of the art human factors principles for safe and reliable performance. As a result, additional consideration of the role of human factors in advanced reactor operation is warranted. The simpler design, passive features and inherent characteristics of advanced reactors will affect the associated concepts of operations and the degree of reliance on human actions. As a result, the near-term research is focused on establishing an approach for scalable human factors engineering (HFE) reviews of non-LWRs. In the longer term, the staff wants to develop insights about HFE for remote and autonomous operation. To guide their thinking, they plan on developing a taxonomy of human factors considerations for ensuring safe operation.
Another major part of the research is to establish a scientifically rigorous nuclear human factors technical basis for guidance development through the use of targeted empirical human factors simulation-based work at unique facilities around the world. These facilities include Halden in Norway, the Idaho National Laboratory Human System Simulation Laboratory and the University of Central Floridas Human Performance Test Facility. These state-of-the-art facilities will help evaluate new technologies for next generation of human-machine interactions.
ACRS Review and Evaluation We find the research program appropriate for the challenges associated with risk assessments and risk informed decision making associated with non-LWRs. The staff have identified good milestones for their work (e.g. RIL, ISG) to assure the results will influence licensing reviews.
The research staff is well engaged with NRR staff, PRA standards groups, and the broad HFE research community, which should provide a high probability of success on the current research.
Summary The staff is performing anticipatory risk-related research in support of advanced reactors that should help the staff in future licensing applications.
Appendix F 1
APPENDIX F ARTIFICIAL INTELLIGENCE
Background
Artificial Intelligence (AI) is defined as a machine-based system that can go beyond defined results and scenarios and can emulate human-like perception, cognition, planning, learning, communication, or physical action. For a given set of human-defined objectives, AI can also make predictions, recommendations, or decisions influencing real or virtual environments.1 As a result of recent exponential acceleration in AI development, both the public and private sectors are noticeably evaluating and implementing AI technology to improve organizational efficiency and extend or enhance their capabilities. The nuclear industry has also expressed a growing interest in using AI technologies to enhance its decision-making processes during the design and operation of nuclear facilities.
The NRCs AI Strategic Plan, Fiscal years 2023-2027 (NUREG-2261), establishes the vision and goals for the NRC to continue to improve its skills and capabilities to review and evaluate the application of AI to NRC-regulated activities, maintain awareness of technological innovations, and ensure the safe and secure use of AI in NRC-regulated activities.2 The Office of Nuclear Regulatory Research (RES) is sponsoring AI research through its Future Focused Research (FFR) program to explore how AI can support the NRC mission and build foundational knowledge across the agency.
ACRS Review and Evaluation On November 15, 2023, a joint meeting of the ACRS Human Factors, Reliability, and Probabilistic Risk Assessment and the Digital I&C Subcommittees was held to discuss the NRC activities related to AI. During this meeting, our Joint Subcommittees had the benefit of discussions with the NRC staff and representatives of Idaho National Laboratory. Our Joint Subcommittees also benefited from discussions with a speaker invited at the initiative of the ACRS, Dr. Mary (Missy) Cummings, director of the Autonomy and Robotics Center at George Mason University, an expert on the uses and risks of AI in transportation.
The nuclear industry has expressed interest in using AI for its NRC-regulated activities. The NRC has recognized that it must keep pace with AI technological innovations and be ready to assess the potential applications of AI. The NRCs AI Strategic Plan presents the vision and goals for the agency to cultivate an AI-proficient workforce, keep pace with AI technological innovations, and ensure the safe and secure use of AI in NRC-regulated activities.
The AI Attributes Working Group, formed in May 2023, includes members from multiple agency offices. The purpose of this interoffice working group is to leverage data science, legal, licensing, and oversight expertise to comprehensively document the attributes and characteristics of the notional AI and autonomy levels identified in the AI Strategic Plan.
1 https://www.nrc.gov/ai.html#ai 2 U.S. Nuclear Regulatory Commission, Artificial Intelligence Strategic Plan, Fiscal years, 2023-2027, NUREG-2261, May 2023, (ML23132A305).
Appendix F 2
RES has hosted a series of Data Science and AI Regulatory Applications Public Workshops to provide a forum for the NRC, nuclear industry, and stakeholders to discuss the state of knowledge and research activities related to data science and AI, and their application in the nuclear industry. RES was also considering how best to align the agencys efforts with the AI Risk Management Framework of the National Institute of Standards and Technology (NIST) and other frameworks being developed.
RES has completed a four-month FFR project intended to explore whether unsupervised machine learning could be applied to data from past inspection findings to identify safety-related patterns and inform inspection planning. The work was carried out by contract to SphereOI, an AI engineering company. The results of this project were inconclusive, perhaps because of the limited scope of this effort. However, the project was a good example of how the agency is using small exploratory projects to identify possible promising directions and increase staff capability and knowledge.
With the sponsorship of the U.S. Department of Energy (DOE) Light Water Reactor Sustainability Program, the Idaho National Laboratory is conducting research with a focus on the use of AI and machine learning to support plant modernization, sustainability, and efficiency through the carefully targeted use of automation. The emphasis of one such project was on safety-related uses of computer vision. In particular, the Joint Subcommittees were briefed on the results of an elaborate effort to automate fire detection using image recognition.
During our Joint Subcommittees meeting, Dr. Cummings also discussed the potential advantages and disadvantages of AI in safety-related fields such as transportation and nuclear power. The talk was highly thought-provoking and provided an excellent opportunity for the subcommittee members (and agency staff in attendance) to think proactively about AI issues that may affect the nuclear industry in the future.
Summary The agencys public workshops, small projects, and coordination with NIST, DOE, and other agencies are helping the agency prepare for and evaluate possible uses of AI in both agency and industry activities in future. The field of AI is vast and rapidly evolving, but these exploratory activities will help the agency keep track of developments in the field, both the potential benefits to nuclear safety and possibly also potential challenges. We look forward to hearing more about this technology as the work progresses.
Appendix G 1
APPENDIX G FUEL FRAGMENTATION, RELOCATION AND DISPERSAL (FFRD)
Background
Under reactor accident conditions, fuel pellets may fracture due to expanding fission gas bubbles, thermal stress, loss of mechanical constraint, and/or mechanical loading. The results of experiments conducted in United States and international laboratories demonstrated that the higher the burnup of the fuel during accidents, the smaller each fuel fragment becomes.1 Industry is currently looking to receive approval to increase fuel burnup limits from 62 gigawatt days per metric ton of uranium (GWd/MTU) to 75 or 80 GWd/MTU. A major technical issue to extending burnup is fuel fragmentation, relocation, and dispersal (FFRD) during loss of coolant accidents (LOCAs). To support the review of fuel designs, methodologies related to fuel burnup extension and associated licensing reviews, the Office of Nuclear Regulatory Research (RES) has provided technical reviewers with interpretations of over 10 years of research on this complex technical issue important to high burnup fuel safety. The results of this review were documented in Research Information Letter (RIL) 2021-13.2 In late 2021, The Committee reviewed the RIL 2021-13 on FFRD during LOCA. The RIL evaluates the publicly available experimental data related to FFRD at high burnup during a LOCA with the objective of seeking to gain a better understanding of fuel performance, under LOCA conditions, in particular for fuel with burnup greater than the current license limit of approximately 62 GWd/MTU.
To this end the RES staff sponsored or participated in experimental programs to determine the potential impacts of FFRD on fuel performance at and beyond the current approved limits.
These include the Studsvik SCIP-3 program, and research conducted at U.S. national laboratories and the Halden reactor. Experimental results from the literature as well as other international laboratories were also considered in the assessment.
The results from a number of research programs and fuel evaluations suggested that, at burnups > 55 GWD/MTU (pellet average burnup), and cladding strain >3%, fuel failure can result in FFRD. Since the current burnup limit is 62 GWd/MTU, these results suggested that FFRD needs to be accounted for in fuel performance evaluations during a LOCA.
ACRS Review and Evaluation In our December 20, 2021, letter evaluating the RIL, we acknowledged that the phenomena associated with FFRD are complex and offered a number of cautions that should be considered going forward:
1.
The conditions of the experiments often differed significantly from conditions that would exist at PWR operating conditions. Depending on the specific test, key variables that were not always prototypic include: linear heat generation rate (low), terminal temperature (high) and heatup rate (low).
1 https://www.nrc.gov/reactors/power/atf/technologies/burnup.html 2 U.S.NRC Interpretation of Research on Fuel Fragmentation, Relocation, and Dispersal (FFRD) at High Burnup, RIL 2021-13, December 2021 (ML21313A145)
Appendix G 2
2.
The fuel/thermal variables influencing fragmentation and relocation phenomena are many. While fuel burnup is easily calculated, and verified by measurement, the cladding strain must be calculated - a calculation fraught with uncertainty. Fuel pellet cracking occurs almost immediately on initial startup. The evolution of fuel microstructure during burnup, including the rim, also introduces uncertainty.
3.
Operational variables, including flow-induced vibration (normal and accident) can strongly influence the fragmentation and relocation process as well as dispersal after cladding breach.
4.
It is one thing to determine an observed burnup at which fuel fragmentation begins to occur. In this case it is essential to determine the difference between conditions of a test/examination and the conditions that would exist under actual operating conditions. A careful evaluation of uncertainty will be key. It is quite another thing to determine the point at which FFRD actually influences LOCA performance. The RIL made no claims in this area, one that is critical to determination of practical consequences. There are a number of experimental programs that will produce data that more closely approach PWR conditions including the TREAT tests. These results will likely provide more prototypic data and hence reduce uncertainty.
5.
Lastly, the larger picture should be considered. The document would benefit from additional context by identifying the entire scope necessary to resolve the safety issues related to FFRD and describing the role of the RIL as a specific piece of that overall scope. Because this RIL suggests a numerical threshold burnup for the onset of fragmentation that is different than that used presently for licensing actions, the document should identify the purpose for transmitting this different burnup limit to NRR and identify plans to provide additional information relative to its potential impact on any future regulatory activity.
Our letter also provided the following conclusions and recommendations:
1.
The current data set on FFRD has been expanded. However, there remains a significant degree of uncertainty in large part because the problem is multivariate and the experiments from which the data were developed did not always represent actual light water reactor conditions.
2.
The staff recognized, and we agree, that this document would benefit from additional context by identifying the entire scope necessary to resolve the safety issues related to FFRD and describing the role of the RIL as a specific piece of that overall scope.
3.
A risk informed approach should be undertaken that examines both the likelihood of expected event conditions combined with a more complete modeling evaluation of FFRD consequences. This activity could add substantial value to future research program development and to the regulatory decision-making process.
Appendix G 3
Recent Staff and Industry Activities since the RIL During the past year, the staff performed an extensive evaluation of the performance of fuel across a reactor core during a postulated LOCA (or similar) event. The results showed significant FFRD in the high burnup (>55 GWd/MTU) portions of the core.
The industry is currently proposing that the burnup limit for fuel be increased to ~ 75 GWd/MTU to enable longer fuel cycles and better economics. Several LARs have been submitted and approved that allow increased burnup under limited conditions by requiring novel changes to fuel management loading patterns, for example. However, the industry would like these restrictions to be relaxed. Additionally, an increase in burnup will require an increase in enrichment beyond the current limit of 5 weight-percent uranium-235. The Commission has directed the staff to develop rulemaking for increased enrichment. However, a key requirement is that FFRD must be addressed.
Given the requirements for LOCA analysis that are currently in place, accounting for FFRD has proven to be a very difficult task. FFRD is an extremely complicated phenomenon. Doing relevant experiments that would reduce uncertainty and provide actionable results, would be extremely time consuming, expensive, and might still not be definitive enough.
As a result of these difficulties, the industry has taken several approaches to demonstrate that fuel failure may not occur. One approach uses specific operational fuel management strategies that limit where the highest burnup fuel assemblies are located in the core. Another approach is to conduct analysis to demonstrate that a large break LOCA is a beyond design basis event, which would allow for a best estimate analysis to be performed instead of the traditional conservative approach. If fuel cladding failure does not occur, then dispersal is impossible. This approach has, in some cases, required that a no burst criteria be established. Current results and analysis suggest that meeting a no burst criteria can provide for extensive margin-in some cases 500-700°F below the 2200°F limit. In short, the requirement to address FFRD has required a change in the accident analysis paradigm.
Summary The results from a number of research programs and fuel evaluations have suggested that, at burnups > 55 GWD/MTU (pellet average burnup), and cladding strain>3%, fuel failure can result in FFRD. Since the current burnup limit is 62 GWd/MTU, these results suggested that FFRD needed to be accounted for in fuel performance evaluations during a LOCA.
During the past year, the staff performed an extensive evaluation of the performance of fuel across a reactor core during a postulated LOCA (or similar) event. The results showed the likelihood of significant fragmentation, relocation, and dispersal in the high burnup (>55 GWd/MTU) portions of the core during a LOCA.
While early Commission policy indicated that FFRD was not a significant issue, new data has suggested that reevaluation of the potential for FFRD is warranted. In response to new Commission directive, the staff has proposed that FFRD be accounted for as a part of the current increased enrichment rulemaking. More recently, on February 22, 2025, the ACRS issued its letter report on the increased enrichment rulemaking to the Commission.
Appendix H 1
APPENDIX H ADVANCED MANUFACTURING TECHNOLOGIES
Background
Advanced manufacturing technologies (AMTs) refer to those techniques and material processing methods that have not been traditionally used in the U.S. nuclear industry and have yet to be formally standardized by the nuclear industry (e.g., through nuclear codes and standards, through a submittal, or other processes resulting in NRC approval/endorsement).
AMTs can be applied to new and replacement components, repair activities of existing components, and specific fabrication elements (i.e., welds, coatings, etc.) of a component to provide a performance or operational benefit.1 Various stakeholders are working towards more widespread use of AMTs in both existing and future nuclear applications. Vendors and licensees/applicants are identifying candidate applications and are developing technical basis for gaining regulatory acceptance. The Nuclear Energy Institute (NEI) has developed a roadmap to understand industry needs/interests and assist with regulatory acceptance. Meanwhile, the Department of Energy (DOE) and the Electric Power Research Institute (EPRI) are sponsoring research at various levels of maturity to improve the understanding and demonstrate AMTs for nuclear applications.
Since 2017, NRC has been taking a proactive engagement strategy to prepare for implementation of AMT components. In 2020, NRC published an action plan for AMTs that identified objectives, regulatory paths, key considerations and near-term tasks to help NRC prepare for AMTs. The office of Nuclear Regulatory Research (RES) has been an integral part of development of technical information, knowledge, regulatory guidance, and tools to prepare the NRC staff to review AMT applications.
ACRS Review and Evaluation On July 6, 2022, RES staff briefed the Committee on NRC activities on AMTs.
The NRC is conducting research to support license application review and approval of AMTs.
RES has a two-part approach to the research. The first task involves technical preparedness and focuses on technical information, knowledge and tools to prepare the NRC staff to review applications. The second task is focused on regulatory preparedness in developing guidance and tools from a regulatory perspective to prepare the staff for the efficient and effective review of additive manufacturing components that may be submitted to the NRC for review and approval.
Additive manufacturing technology involves a process where material is added in a pattern, layer by layer, to form a part. There is growing interest in the nuclear industry to use this technology to make certain reactor components that are more effective due to greater sophistication (like fuel assembly debris filters) and highly machined components such as thimble plugs. Staff is reviewing the wide range of additive manufacturing techniques to ensure there is the right level of technical competence within the agency to review license applications for the application of additive manufacturing. Based on industry interest and the potential of near-term applications, the NRC focused on five major additive manufacturing processes that 1 https://www.nrc.gov/reactors/power/amts.html
Appendix H 2
are: laser powder bed fusion, laser directed energy deposition, cold spray electron beam welding and powder metallurgy hot axial static pressing. Research has completed the technical activities that focus on developing the technical information needed to identifying gaps, knowledge and tools to prepare staff for additive manufacturing reviews.
Staff is also developing the content of information necessary to develop regulatory guidance such as non-destructive examination and quality assurance requirements. It is anticipated that this guidance will be in the form of Regulatory Guides.
The fundamental goal of this research is to ensure the agency is prepared to efficiently and effectively review and approve license submittals that are expected soon.
Summary Applicants will soon be submitting license applications for a variety of components created by additive manufacturing. The NRC has been taking a proactive engagement strategy to prepare for implementation of AMT components. The office of Nuclear Regulatory Research has been an integral part of development of technical information, knowledge, regulatory guidance, and tools to prepare the NRC staff to review AMT applications.
Appendix I 1
APPENDIX I ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING IN NONDESTRUCTIVE EXAMINATION AND INSERVICE INSPECTION
Background
Nondestructive examination (NDE) has played a critical role in detecting component degradation in nuclear energy systems since the early days of the industry. 10CFR50.55(a)(b) incorporates by reference the ASME Code Section III, Rules for Construction of Nuclear Facility Components, and Section XI, Rules for Inservice Inspection of Nuclear Power Plant Components. NDE is essentially a means for identifying and characterizing initial as-built flaw distributions as well as service-induced flaws. Plant life extensions from 40 to 60 and now up to 80 years, add to the critical importance of NDE. Flaw growth due to environmental degradation (environmentally assisted cracking, fatigue crack growth) is generally an initiation and growth process both of which are non-linear in initiation time as well as growth rate.
While NDE plays a critical role in assessing performance, the process itself often requires evaluation of data that is subject to interference from non-flaw related features that in some cases makes calling a defect difficult. False positive and false negative calls are a constant reminder of the difficulties. Extensive training is required for analysts. Additionally, actual in-service defects are often difficult to reproduce for training. As a result, inspection times during outages often increase the outage time, and radiation exposure is an issue resulting in increased costs. In addition, the industry is projecting a shortage of NDE technicians with the proper skills to meet future needs. Automated Data Analysis (ADA) and Machine Learning (ML) have the potential to address the above concerns. NDE using ADA and ML is widely used in other industries and there are several open-source ML algorithms available as well as numerous commercially available data analysis tools. The nuclear industry is also funding work to use these tools for automated data analysis algorithms to analyze NDE data.
The Office of Nuclear Regulatory Research (RES) has initiated projects to provide the technical basis to support regulatory decisions and Code actions related to ADA/ML for NDE.
ACRS Review and Evaluation On March 6, 2024, the staff briefed the Committee on RES research projects related to use of AI and ML/ADA for NDE. The purpose of the RES projects are to evaluate the existing commercially available techniques for potential use in Section III and XI analysis. Two methods of using ADA were evaluated: ADA-Assisted Examination and Fully Automated Examination. In the first case there is a human (fully qualified inspector). In the latter case the ADA algorithm analyzes the data without human input.
The RES program seeks to provide a technical basis describing current capabilities of machine learning and automated data analysis for NDE.
The ORNL was contracted to evaluate ML for ultrasonic examinations and Pacific Northwest National Laboratory (PNNL) was contracted to evaluate commercially available data analysis platforms including rule-based and ML-based systems. Rule-based systems make decisions based on explicit set of rules. Learning-based systems make decisions based on training data.
Appendix I 2
Expected outcomes included identification of: 1) capabilities and limitations of ADA, 2) factors that influence ADA performance/reliability, 3) validation approaches and methods for qualification, and 4) any gaps in codes and standards.
The assessment of rule-based ADA suggests that rule-based ADA is not likely to be appropriate on its own for nuclear system inspections, but rule-based methods have the potential to complement learning-based methods on a case-by-case basis. The assessment of ML suggests that this technique is capable of high True Positive detection rates (TPR), and low False Positive (FP) and False Negative (FN) rates. In short ML, if used with care, can be used for NDE data classification. However, training data must be representative of the types of data expected during testing.
In conclusion, the results suggest that elimination of the human from the loop will not be appropriate for nuclear inspections, but ADA and ML can provide for guidance for the inspector.
Lastly, none of these techniques will do well at identifying new or emerging issues.
Summary The ADA has the potential to improve detection of flaws as an adjunct to the analyst.
Additionally, humans and computers ADA make different kinds of mistakes. Complementarity has the potential to improve both the analyst and produce an updated algorithm, which is an overall benefit. However, ADA also has the potential to introduce common cause failures of common inspections across the industry. This possibility must be guarded against. As with any new technology, ADA may be difficult to implement, and training can be complicated. The ADA/ML algorithms may require a new class of experts to support the analysis. Lastly, ADA/ML can be trained to handle known problems but may not identify new or emerging issues.
Appendix J 1
APPENDIX J COMPUTER CODE DEVELOPMENT AND VALIDATION FOR NON-LIGHT-WATER REACTORS
Background
In 2016, the NRC staff published NRC Vision and Strategy: Safely Achieving Effective and Efficient Non-Light Water Reactor (non-LWR) Mission Readiness and subsequent near-, mid-and long-term implementation plans. These documents describe the objectives, strategies, and contributing activities necessary to achieve technical and regulatory readiness to ensure staff preparedness and capability to perform efficient reviews.
The Office of Nuclear Regulatory Research (RES), approach to developing non-LWR codes is structured around the 2016 "Vision and Strategy" report. The principal emphasis of RES code development activities involves the Comprehensive Reactor Analysis Bundle (BlueCRAB) code suite. The BlueCRAB suite includes key tools such as SAM (for plant systems analysis),
PRONGHORN (for subchannels), and GRIFFIN (for core neutronics). These codes are rapidly maturing and validation against benchmarks has been ongoing. Substantial results have been realized in areas like plant systems and fuel performance for specific non-LWR designs such as high-temperature gas reactors (HTGRs), liquid metal fast reactors (LMFRs), molten salt-cooled reactors (MSRs); and heat-pipe-cooled microreactors. The staff has also progressed in validating the MELCOR and MELCOR Accident Consequence Code System (MACCS) codes for severe accident analysis, focusing on radionuclide transport, source terms, and consequence analysis for advanced reactors.
ACRS Review and Evaluations Since 2016, as part of our periodic review of NRC Safety Research programs, we have reported on our assessment of RES progress on code development and validation for non-LWRs, emphasizing capability and limitations specific to identifying and resolving safety issues associated with non-LWRs.
On April 3, 2024, we were briefed on the status of computer code and model development for non-LWRs. We observed that the RES staff has made significant strides in implementing the mid-term strategies, with a focus on technical readiness for design reviews.
Informed by lessons-learned from their near-term phase activities, specific activities led by RES have emphasized the maturation of their non-LWR codes. The collaboration between RES staff, U.S. Department of Energys (DOEs) Nuclear Energy Advanced Modeling and Simulation (NEAMS) program, and other external organizations has been crucial to the success of this effort. These partnerships not only expedite code refinement but also help identify gaps in data and experimental needs.
The alignment of code development with the specific design needs of HTGRs, LMFRs, etc. has been well-coordinated. A core strength of the RES staffs approach has been their regular engagement with stakeholders, particularly through workshops and public presentations. These engagements have been critical in advancing the capability and usability of these codes, ensuring that they meet the practical needs of NRC staff responsible for topical and safety analysis report reviews.
Appendix J 2
Continuous training and knowledge management are essential for the NRC to maintain its technical expertise. Staff training on new reactor designs and associated phenomena during code development ensures a robust knowledge base. The committee recommends continued efforts in this area to preserve the knowledge acquired over years of development.
Sustainability of success in this area will depend upon the establishment of a formal knowledge management processes, which will ensure the scrutability of the NRC codes relied upon in regulatory design reviews.
The development of BlueCRAB, which includes tools for plant systems analysis, fuel performance, accident progression, and dose consequences, has substantially improved the NRC's ability to independently assess non-LWR designs. Staff have shown an ability to adapt and refine their technical reviews based on practical feedback and real-world applications, as demonstrated in the successful review of the Kairos HERMES design. This enhances regulatory efficiency and technical competency.
While significant progress has been made by RES, particularly in deterministic analysis capability, there remain challenges to achieving the original 2016 vision for agency capability.
For example, incompleteness in the datasets used for code verification and validation still hampers full confidence in the accuracy of these codes. This is a historic challenge, not unique to the development of non-LWR analysis capability. Experimental programs, particularly in collaboration with international and DOE partners, are encouraged to fill these gaps.
Additionally, the agency is urged to formalize its data preservation processes to safeguard the extensive knowledge embedded in its codes through their validation, thus ensuring they remain valuable tools for regulatory analysis and industry engagement.
Attention by RES is also needed to facilitate the integration of these computer codes into modern evaluation models. Such evaluation models rely on explicit uncertainty and fault/event tree characterizations required by best-estimate plus uncertainty analyses and probabilistic risk assessments. This capability will enable agency staff in their review of novel analyses methods arising from the Part 53, Risk Informed, Technology-Inclusive Regulatory Framework for Advanced Reactors, rulemaking.
These computer codes are the consequence of substantial domestic and international investments in experiments and model development. As such, maintaining strong partnerships with DOE and international entities will be critical to advancing the NRCs capabilities. Moving forward, the NRC should prioritize further collaboration and development in areas where critical gaps still exist. The committee anticipates that these efforts will ensure the NRC's preparedness for the anticipated influx of non-LWR design applications, contributing to the safe, efficient, and timely review of advanced reactors.
Summary ACRS recognizes the RES staffs significant progress in implementing its non-LWR code and model development plans. In doing so, they have acted on our prior recommendations to preserve simplicity in the capability and to focus development efforts based on the expected hazards. In our assessment, codes in the analysis suite are complete (i.e., with respect to the more important design-specific phenomena and processes). It is our opinion that the NRC is in a good position to support technical review of advanced reactor design applications that are anticipated in the near future.
Appendix K 1
APPENDIX K HIGH ENERGY ARCING FAULTS
Background
As shown by the Browns Ferry fire in 1975, fires can present a significant risk to reactor safety.
High Energy Arc Faults (HEAFs) are a category of electrical fires characterized by a potentially explosive source of energy. One definition of a HEAF is:
High Energy Arc Faults (HEAFs) are energetic or explosive electrical equipment faults characterized by a rapid release of energy in the form of heat, light, vaporized metal and pressure increase due to high current arcs between energized electrical conductors or between energized electrical components and neutral or ground. HEAF events may also result in projectiles being ejected from the electrical component or cabinet of origin and result in fire.1 Due to the potential impact of HEAF events, modeling their frequency and consequence is important when conducting a fire probabilistic risk assessment (PRA). At the time of publication of fire PRA guidance document NUREG/CR-6850 (EPRI/NRC-RES Fire PRA Methodology for Nuclear Power Facilities) in 2005, the HEAF phenomena was known, but the state of knowledge of its potential hazard in switchgear and load centers had not progressed to the point where it was possible to estimate the distance in which a HEAF can cause damage or failure of a target.
Instead, the zone of influence (ZOI) for switchgear and load centers in NUREG/CR-6850, Appendix M, was developed primarily from a single catastrophic event at San Onofre involving a medium-voltage switchboard.
In 2011, following the Tohoku earthquake in Japan, an arcing fault occurred in the No. 7 and No. 8 sections of the non-emergency 6.9-kV switchgear at the Onagawa nuclear power plant.
The arcing fault led to a fire in all ten vertical sections of the switchgear. Control cables for non-emergency components, such as feedwater pumps, condenser pumps, etc., directly above the cabinet were affected by the heat generated by the fire.2 Following the Onagawa HEAF event, the Secretariat of Nuclear Regulation Authority of Japans Regulatory Standard and Research Department performed a series of experiments to characterize and understand HEAF phenomena. The NRC Office of Nuclear Regulatory Research (RES) was invited to observe and support the testing that occurred between 2013 and 2015. One observation from this test series was a greater-than-expected thermal energy release, which was hypothesized to result from exothermic oxidation of aluminum bus bars instead of copper bus bars.3 In its April 28, 2016, report on review and evaluation of the NRC 1 OECD / Nuclear Energy Agency (NEA), Committee on the Safety of Nuclear Installations (CSNI), OECD Fire Project Topical Report No. 1 - Analysis of High Energy Arcing Fault (HEAF) Fire Events, NEA/CSNI/R(2013)6, 2013 (https://www.oecd-nea.org/nsd/docs/2013/csni-r2013-6.pdf).
2 ibid 3 U.S. Nuclear Regulatory Commission, International Agreement Report - Nuclear Regulatory Authority Experimental Program to Characterize and Understand High Energy Arcing Fault (HEAF) Phenomena, NUREG/IA-0470, August 2016 (https://www.nrc.gov/reading-rm/doc-collections/nuregs/agreement/ia0470/index.html).
Appendix K 2
safety research program, the ACRS recommended continued NRC support for this HEAF research project.4 From 2014 to 2019, RES collaborated in a full-scale HEAF experimental program with the Nuclear Energy Agency (NEA), the National Institute of Standards and Technology (NIST), and various international partners through the Organization for Economic and Cooperative Development (OECD). The testing showed the possibility that the HEAF event damage may be more severe than modeled in Appendix M of NUREG/CR-6850, leading to a larger ZOI. Given the apparent significance of these observations, a possible generic issue concerning the vulnerability of current-carrying aluminum components subject to HEAFs was initiated in May 2016. The NRC exited the generic issue process in August 2021. The closure memo identified that additional long-term research was necessary to determine the risk significance of the issue.5 Simple approximation of ZOI in NUREG/CR-6850 was insensitive to several factors that influence the potential of hazard of a HEAF, including the fragility limits for targets. In 2020, the NRC and Sandia National Laboratories (SNL) conducted fragility testing to investigate the physics and failure modes of cables exposed to a HEAF. These tests subjected thermoset and thermoplastic jacketed cables to high-heat-flux short-duration exposures.6 A follow-on effort between the Electric Power Research Institute (EPRI) and the NRC analyzed the available data and proposed target fragilities for electrical cables subject to the effects of a HEAF, with recommendations for the treatment of other equipment that may be impacted by these phenomena. The conclusions and recommendations resulted from this effort represent the consensus of the joint NRC-RES/EPRI HEAF working group.7 In parallel, efforts were underway to pursue modeling options that could leverage experimental data and provide ZOI information for configurations that were not subject to full-scale testing.
NISTs Fire Dynamics Simulator (FDS) was ultimately chosen as the modeling tool. The FDS computer code includes a hydrodynamic model capable of solving the low-Mach number, thermally driven flow equations using a large eddy simulation turbulence model. Development of FDS to model HEAFs began in 2019 as a proof of concept. Benchmarking against previous testing began in 2020. A November 2022 joint EPRI/NRC research report documents the methodology, validation, and results of simulation of relevant HEAF FDS scenarios.8 In 2023, RES and EPRI completed an effort to combine recent HEAF-related research and provide methods and data to calculate plant risk due to HEAFs more realistically. The result of 4 ACRS letter, Review and Evaluation of the NRC Safety Research Program, April 28, 2016 (ADAMS Accession No. ML16125A376).
5 U.S. Nuclear Regulatory Commission, Closure of Proposed Generic Issue PRE-GI-018, High-Energy Arc Faults Involving Aluminum, August 31, 2021 (ADAMS Accession No. ML21237A360).
6 U.S. Nuclear Regulatory Commission, HEAF Cable Fragility Testing at the Solar Furnace at the National Solar Thermal Test Facility, Experimental Results RIL 2021-09, 2, SAND2021-11327, September 2021 (ADAMS Accession No. ML21259A256).
7 Electric Power Research Institute, Target Fragilities for Equipment Vulnerable to High Energy Arcing Faults, NRC-RES, RIL 2022-01 and EPRI 3002023400, May 11, 2022 (ADAMS Accession No. ML22131A339).
8 Electric Power Research Institute, Determining the Zone of Influence for High Energy Arcing Faults Using Fire Dynamics Simulator, NRC-RES RIL 2022-09 and EPRI 3002025123, November 2022 (ADAMS Accession No. ML22322A100).
Appendix K 3
this effort is documented in NUREG-2262 (EPRI 3002025942).9 One important insight is that the risk from copper and aluminum bus bars was found to be comparable relative to ZOI determination, although aluminum enclosures for bus ducts were found to present a larger ZOI.
NUREG-2262 includes several recommended changes to the HEAF modeling guidelines of NUREG/CR-6850 Appendix M.
ACRS Review and Evaluation The Committee was briefed on March 7, 2024, on recent work performed to understand high energy arcing faults (HEAF) and their impact on reactor safety risk. This research has roots in the investigation of the Browns Ferry fire in 1975 and has been noted over the last 20 years to be significant in fire probabilistic risk assessments (PRAs). We were briefed on the overall history of this research as well as the more recent work done to investigate operational experience implying that the risk of HEAF in switchboards containing aluminum components was higher than previously thought.
The NRC staff expended significant effort responding to operational data that implied aluminum busbars led to larger safety risk due to HEAF than was previously understood. This effort resulted in a better understanding of the HEAF phenomenon, allowing improvements in safety analysis techniques, plant maintenance practices, and design objectives for future plants. The following aspects of this effort are worth highlighting:
As part of developing techniques to estimate risk, EPRI developed a set of maintenance practices that, if followed, can justify reducing the modeled likelihood of a HEAF event.
While the motivation may have been to support justification of better fire risk metrics in light of modeling uncertainties, these improved maintenance practices will result in reduced incidence of HEAF events in real life and hence safer and more effective plant operation. In this way, the research provides a tangible benefit that is independent of the details of a particular safety analysis. We encourage the staff to look beyond development of safety analysis methods for ways that research results can translate into improved, safer plant operation.
The NRC staff performed a risk-informed evaluation of the implications of HEAF events after the generic issue was closed in 2021. This risk-informed evaluation brought in numerous stakeholders and used the best available information to determine the urgency to resolve the potential HEAF safety implication. This application of a risk-informed approach to problem resolution was effective and we encourage a similar approach when other issues arise.
The RES evaluation revealed that the design of the electrical power distribution system is a major factor in the likelihood of HEAF. Specifically, a significant fraction of the HEAF events occur in plants that do not have a generator output circuit breaker, such that there is significant energy from the spinning-down generator to feed the arc fault until the generator spin-down is nearly complete. We suggest that staff consider incorporating this insight into guidance for review of the electrical distribution for new and advanced reactors.
9 U.S. Nuclear Regulatory Commission and Electric Power Research Institute, High Energy Arcing Fault Frequency and Consequence Modeling, NUREG-2262, April 2023 (ADAMS Accession No. ML23108A113).
Appendix K 4
Summary The research being conducted on HEAF is vital to resolving the issues raised by evaluation of operational experience in the form of actual HEAF events. This research led to both improved fire risk modeling techniques and recommendations for improved maintenance practices that should reduce the incidence of HEAF events.
Appendix L 1
APPENDIX L LIST OF ACRONYMS ACRS Advisory Committee on Reactor Safeguards ADA Automated Data Analysis ADAMS Agencywide Documents Access and Management System AI Artificial Intelligence AMT Advanced Manufacturing Technologies ARPA-E Advanced Research Projects Agency - Energy ASME American Society of Mechanical Engineers BLUECRAB Comprehensive Reactor Analysis Bundle BWRs Boiling Water Reactors DOE Department of Energy DT Digital Twin EPRI Electric Power Research Institute FDS Fire Dynamics Simulator FFR Future Focused Research FFRD Fuel Fragmentation, Relocation, and Dispersal GEMINA Generating Electricity Managed by Intelligent Nuclear Assets GW/MTU Gigawatt Days Per Metric Ton of Uranium HEAFS High Energy Arc Faults HFE Human Factors Engineering HTGRs High-Temperature Gas Reactors IAEA International Atomic Energy Agency IEEE Institute of Electrical and Electronics Engineers ISI Inservice Inspection LMFRs Liquid Metal Fast Reactors LMP Licensing Modernization Project LOCA Loss of Coolant Accident LPSD Low Power Shutdown LWR Light Water Reactors MACCS MELCOR Accident Consequence Code System ML Machine Learning MSRs Molten Salt-Cooled Reactors NDE Non -Destructive Examination NEA Nuclear Energy Agency NEAMS Nuclear Energy Advanced Modeling and Simulation NEI Nuclear Energy Institute NIST National Institute of Standards and Technology NON-LWR Non-Light Water Reactor NPPs Nuclear Power Plants NRC Nuclear Regulatory Commission NRR Office of Nuclear Reactor Regulation OECD Organization for Economic and Cooperative Development PNNL Pacific Northwest National Laboratory POS Plant Operating States PWRs Pressurized Water Reactors PRA Probabilistic Risk Assessment RES Office of Nuclear Regulatory Research RIDM Risk Informed Decision-Making
Appendix L 2
RIL Research Information Letter RG Regulatory Guide SNL Sandia National Laboratories SRM Staff Requirements Memorandum TPR True Positive Detection Rates