ML25272A200
| ML25272A200 | |
| Person / Time | |
|---|---|
| Issue date: | 09/30/2025 |
| From: | Coyne K NRC/RES/DRA |
| To: | |
| References | |
| Download: ML25272A200 (1) | |
Text
Introduction PRA Realism Public Workshop September 30, 2025 Kevin Coyne Office of Nuclear Regulatory Research Division of Risk Analysis kevin.coyne@nrc.gov
Topics
- What is Realism?
and what is the NRC doing?
- Some Data Trends, Operational Events, and Insights
- Some Tradeoffs
- Concluding Thoughts
Realism
- Dictionary
- An inclination toward factual truth and pragmatism.
- PRA Policy Statement (60 FR 42622)
- (3) PRA evaluations in support of regulatory decisions should be as realistic as practicable and appropriate supporting data should be publicly available for review.
Realism (cont.)
- RG 1.200, Acceptability of Probabilistic Risk Assessment Results for Risk-informed Activities
- an accurate representation (to the extent practical) that reflects the expected response of the as-built and as-operated plant
- the base PRA reflects the as-designed, as-built, and as-operated plant and is as realistic as possible in assessing the risk
- estimate plant-specific, realistic fragilities: seismic, winds, floods
Strategic Plan (NUREG-1614)
- Strategy 1.2.1 - Maintain and further risk-inform the current regulatory framework using information gained from operating experience, lessons learned, external and internal assessments, technology advances, research activities, and changes in the threat environment.
Some NRC Activities
- SPAR Model Enhancements
- Level 3 PRA Project
- Fire Analysis and PRA
- Digital I&C
- Operating Experience Data Collection and Analysis
- Flood Hazard and Mitigation
- Human Reliability Analysis
- PRA Standards
- Accident Sequence Precursor Program
- Advanced Reactors
Some Data Trends Accident Sequence Precursor (ASP) Program The ASP Program evaluates:
Degraded plant conditions (CDP)
Some Data Trends (2015-2024)
The ASP index shows the cumulative plant average risk from precursors on an annual basis.
ASP results over the last 10 years are strongly influenced by just two operational events.
Operational risk is driven by the bad day, not the average day.
Recent Important Precursors
- Two events contribute significantly to the ASP index over the last 10 years.
- Duane Arnold: LOOP caused by a Derecho (8/10/20) - CCDP of 8E-04.
- Dominate sequence involves SBO, RCIC/HPCI failure, failure of early offsite power recovery (35%).
- Other scenarios involve SBO, success of RCIC, and failure of FLEX, ELAP declaration, or longer-term offsite power recovery (32%).
- Waterford: LOOP caused by Hurricane Ida (8/29/21) - CCDP of 5E-04.
- Dominate sequence involves failure of onsite EDGs and failure of FLEX diesel to charge batteries (68%).
Duane Arnold Derecho NWS estimated wind speeds were likely near 130 mph.
Resulted in all six off site power sources being damaged.
LER 331-2020-001 The high winds experienced on August 10, 2020, were not considered a beyond design basis event. The systems and components responded as designed and the overall peak wind speeds were within the Design Basis Tornado.
LAR for Surveillance Frequency Requirements to a Licensee Controlled Program (TSTF-425) the DAEC IPEEE analysis of high winds, floods, and other (HFO) external hazards was accomplished by reviewing the plant environs against regulatory requirements regarding these hazards.HFO events were screened out by compliance with the 1975 SRP criteria. As such, these hazards were determined in the DAEC IPEEE to be negligible contributors to overall plant risk.
Although the peak wind speed were within the design bases, the winds generated a significant debris load on the service water system that degraded EDG cooling water flow, contributing to the high event risk significance.
Recovery of offsite power was ~25 hours after the LOOP event.
Should PRA realism include consideration of associated impacts from external hazards (e.g., debris loading)?
Waterford Hurricane Ida LER 382-2021-001 The hurricane force winds experienced on August 29, 2021, were not considered a beyond design basis event. The systems and components responded as designed and the overall peak wind speeds were within the design basis for a hurricane event.
The damage observed coincides with the amount of damage expected for the event.
Waterford has two Class 1E EDGs, but no alternate SBO AC power source.
Offsite power recovered over two days after the initial LOOP.
A PTED generator is capable of supplying on safety-related bus, but the batteries needed for start were later found degraded.
Waterford appears to have a higher CCDP for LOOP events than comparable sites that had experienced hurricane-related LOOP events.
How effectively is operating experience from operational events being used to address risk?
Hurricane tracks in the vicinity of Waterford (1985 - 2024)
Some Tradeoffs
- Use of the failure memory approach for ECA.
- Observed successes treated probabilistically.
- Observed failures treated as failure events.
- Limitations and uncertainties associated with existing CCF data and modeling.
- Timely decision-making using best available information vs. the best answer.
- Inclusive modeling and recognition of the PRA iterative process.
- Clarity of decision-making criteria vs. robust consideration of uncertainties.
Concluding Thoughts
- Realism involves tradeoffs and a balance between accuracy and pragmatism.
- NRC engaged in numerous activities to improve realism in PRA methods, models, tools, and data.
- Broad stakeholder engagement further advances PRA realism.
Questions
SPAR Activities PRA Realism Public Workshop September 30, 2025 Michelle Gonzalez Office of Nuclear Regulatory Research Division of Risk Analysis Probabilistic Risk Assessment Branch
Advantages of SPAR/SAPHIRE Use SPAR Standardized modeling conventions - event trees, fault trees, basic events all use standard modeling and naming conventions Standardization supports efficient user use of all models, supports rapid updates and batch analysis, automatic generation of Plant Risk Information eBooks SPAR includes detailed modeling for areas that have shown to be significant for ECAs - offsite power recovery, CCF Models are benchmarked against licensee models Development approach reflect insights received during industry led peer reviews SAPHIRE Specifically designed to support event and condition assessment (delta CDF calculations, insights)
Multiple user workspaces to support novice to expert users Broad user base (e.g, NASA, universities, aerospace)
Includes modern quantification process, developed under QA program
SPAR Model Status Operating Reactors
- 69 full-power, internal events models representing all operating plants
- 69 All Hazard Models
- 8 Shutdown models
- 2 Level 2 models New Reactors
- AP1000 AHZ (seismic, internal flooding, internal fire, LPSD, Level 2)
- Vogtle 3 & 4
- NuScale
- BWRX300 - in progress
Recent advancements FLEX: All SPAR models were updated with in 2023 with PWROG Data Routine Model Updates 10 models benchmarked in FY24 12 to be benchmarked in FY25 SPAR Dash: an interactive display of Probabilistic Risk Assessment (PRA) results for all operating plants in the US, based on the NRC's Standardized Plant Analysis Risk models (SPAR Models).
Seismic Models All SPAR models include seismic hazard Models utilize site specific hazard curves and a majority of the models utilize surrogate fragilities Diablo Canyon and Columbia SPAR models have plant-specific fragilities and seismic bins were expanded from 5 to 6 bins (FY25)
SPAR Internal Flood Models St. Lucie Unit 1 SPAR model was updated to include 42 dominant internal flood scenarios (FY23)
St. Lucie Unit 2 SPAR model was updated to include 53 dominant internal flood scenarios (FY24)
Shearon Harris SPAR model was updated to include 57 dominant internal flood scenarios (FY25)
SPAR High Wind/Tornado Models High winds utilize location specific frequencies based on
- Tornado - NUREG/CR-4461 Rev 2
- High Wind/Hurricane - ASCE 7-10 or NUREG/CR-7005 The high wind fragilities are surrogates for all models with the exception of Brunswick which utilized plant-specific information.
SPAR Internal Fire Models St. Lucie Unit 2 model was updated to include 122 dominant fire scenarios. (FY25)
Vogtle model was updated to include 126 dominant fire scenarios. (FY25)
Cooper model was updated to include 164 dominant fire scenarios (will be done by the end of the FY-25).
Brunswick Unit 2 was updated to include 157 dominant internal fire scenarios (FY23)
St. Lucie Unit 1 was updated to include 160 dominant internal fire scenarios (FY24)
SAPHIRE Converting SAGE Risk Database to SQLite (more modern and user-friendly format Integrating a.NET-based GraphQL application programing interface (API) for more efficient data access and model creation.
Modifications simplify workflow and software flexibility for future enhancements and move away from older legacy system Continued development of SAPHIRE 9-Scheduled release August 2026
Looking Ahead Incorporate AI for update processes Continue routine SPAR model update efforts to represent as-built-as-operated Continue updating fire models (approx. 3/year)
SAPHIRE 9 (cloud-based system)
Continue development of new/advanced reactors SPAR models
Questions
Level 3 PRA Project Modeling Advancements PRA Realism Public Workshop September 30, 2025 Alan Kuritzky Office of Nuclear Regulatory Research Division of Risk Analysis Probabilistic Risk Assessment Branch
Modeling Advancements and Insights
- Expert elicitation for ISLOCA (trial use of NRC guidance)
- Modeling of safe/stable end states
- Severe accident analysis termination time
- Stochastic combustion modeling
- Full-scope PRA (i.e., all radiological sources, all modes, and all hazards)
- Greatly enhanced staff PRA capabilities and expertise
Questions
Modeling of Digital I&C Systems PRA Realism Workshop September 30, 2025 Steven Alferink Office of Nuclear Reactor Regulation Division of Risk Assessment PRA Licensing Branch C
Areas of Consideration Licensing Post-Installation
Licensing Considerations SRM-SECY-22-0076 Expanded the policy for digital I&C common-cause failures (CCFs) to allow risk-informed approaches BTP 7-19, Revision 9 Provides review guidance for a risk-informed defense-in-depth and diversity (D3) assessment
Risk-Informed D3 Assessment Identify each postulated CCF Address the CCF using a risk-informed approach Model the CCF in the PRA Determine the risk significance of the CCF Determine appropriate means to address the CCF Determine consistency with NRC policy and guidance on risk-informed decision-making Address the CCF deterministically Justify alternative approaches
Risk-Informed D3 Assessment
- Evaluate how the CCF is modeled in the PRA and justification that modeling adequately captures the impact of the CCF on the plant
- Options for modeling the CCF in the PRA:
- Detailed modeling of the detailed I&C system
- Use of surrogate events Model the CCF in the PRA
Risk-Informed D3 Assessment
- Risk significance of a CCF can be determined using a bounding or conservative sensitivity analysis
- Bounding sensitivity analysis:
- Assumes the CCF occurs
- Provides a description of the baseline risk
- Conservative sensitivity analysis:
- Provides a technical basis for a conservative likelihood (less than 1) of the CCF demonstrating that defense in depth is addressed
- Addresses the impact of this assumption on PRA uncertainty Determine the risk significance of the CCF
Risk-Informed D3 Assessment
- Quantification accounts for any dependencies introduced by the CCF, including the ability for operators to perform manual actions
- CCF is not risk significant if sensitivity analysis satisfies two criteria
- Increase in CDF is less than 1 x 10-6 per year
- Increase in LERF is less than 1 x 10-7 per year Determine the risk significance of the CCF
Post-Installation Considerations PRA Configuration Control Maintaining or upgrading the PRA to be consistent with the as-built, as-operated plant Data Analysis Estimating the parameters used to determine the likelihood of a digital I&C CCF occurring
Likelihood of Digital I&C CCFs Licensing Post-Installation Purpose Risk-informed D3 assessment Value Bounding or conservative value for the risk-informed D3 assessment Purpose PRA update to be consistent with the as-built, as-operated plant Value Representative or realistic values for the PRA model
Likelihood of Digital I&C CCFs Timeframe Licensing Post-Installation Bounding D3 assessment
- Considered a source of uncertainty
- Sensitivity analysis determines if it is a key uncertainty
- Risk management if it is identified as a key uncertainty
- Use failure data or expert elicitation to estimate order of magnitude
- Consensus approach
- Not a key uncertainty Conservative D3 assessment
- Use failure data to determine and align on realistic values
- Consensus approach
- Not a key uncertainty
Research Activities Research Assistance Request Bounding Risk Assessments to Support Licensing Activities Associated with Risk-Informed Digital Instrumentation and Controls (DI&C) Modifications Operating Experience Data Analysis for Digital Instrumentation and Control System Reliability and Risk Assessment in Nuclear Power Plants - 2024 Report Provides a foundation for incorporating digital I&C failure data into:
Integrated Data Collection and Coding System (IDCCS)
Reliability and Availability Data System (RADS)
Proposes recommendations for modeling digital I&C reliability and CCF INL/RPT-24-82553
Next Steps Publish INL/RPT-24-82553 as a NUREG report Revise IDCCS database to add digital I&C failure events Update RADS to estimate digital I&C component reliability and CCFs Include digital I&C systems in SPAR models
Improving PRA Realism
- Modeling
- Digital I&C systems
- Data
- Share data and operating experience on digital I&C CCFs
- Include additional information on digital I&C failures in LERs How can industry help improve PRA realism for digital I&C?
Improving PRA Realism
- Digital I&C system
- Data and operating experience
- Additional information in LERs Data
Realism in Fire Research Nicholas Melly, P.E.
Kenneth Hamburger, P.E.
Adam Lee, P.E.
Topics Previous MOU Realism Research Activities Bulk Cable Tray Ignition Lithium-Ion Battery Co-located Hydrogen Methods Consolidation
NRC/Industry Workshop on Improving Fire PRA Realism - 2017 3
Objectives To develop a common understanding of NRC and industry priorities on research work for 2018-2023 to enhance realism in plant fire PRAs.
To highlight specific conservatisms driving results in plant Fire PRAs to better understand challenges.
To identify potential approaches for prioritized research To identify approaches to documentation of Fire PRA method improvements
Improvements for Modeling Bulk Cable Tray Ignition 4
Fire PRA Realism Failure at Collar Failure at Overlap Series of fire tests conducted with EPRI under the MOU to develop guidance for identifying the conditions necessary for cable tray ignition and propagation through an arrangement of cable trays This research effort will improve on the guidance developed in 2016 FAQ 16-0011, "Cable Tray Ignition To objective is to evaluate cable fire behaviors, combustion characteristics, flammability properties such as time to ignition, heat release rate, cable fire spread and ignition thresholds for both single cable and bulk cable tray ignition
Lithium-Ion Battery Fire Hazards 5
- Lithium-ion batteries (LIBs) are increasingly found in the facilities
- Lack of guidance on how to treat hazard
- Thermal runaway and growth rate likely different from current guidance
- Provide basis for development of guidance
- Joint NRC-RES / EPRI project
Project Objectives and Plan Phase I
- EPRI : Whitepaper
- Analysis of hazard with focus on emerging and existing applications
- Outlines barriers, detection, suppression, and emergency response
- Recommends hazard screening framework, performance-based modeling, and risk-informed mitigation analysis Phase II
- NRC/EPRI : hazard characteristics and risk method
- Use literature and focused experimental series to characterize LIB heat release rate (HRR) profile
- Develop risk screening /
quantification method (frequency, transient, etc.)
Co-Located Hydrogen Risk 7
Substation / Power Conditioning Reboiler Steam Electricity
- Develop a fire risk framework to characterize NPP risk from co-located hydrogen generation facilities
- Focus on screening approach developed for external hazard
- Detailed modeling and fragility analysis to provide robust technical basis
Co-Located Hydrogen Approach focuses on hydrogen jet fire (thermal heat flux hazard) and unconfined explosion (overpressure hazard)
Conceptual production facility designs used in the hazard estimation Damage limits determined for hazard type and vulnerable structures or components Quantitative validation of hazard model HyRAM+
Methods Consolidation
- Consolidated tool to summarizing the state-of-the-art fire PRA methods available for assessing nuclear power plant (NPP) fire risk.
- High-level synopsis of the existing methods and how the methods fit into the Fire PRA framework (NUREG/CR-6850).
- The update will include a roadmap of how the methods fit into the fire PRA (NUREG/CR-6850) structure.
Accessibility using search functions and accessible links will be incorporated to the extent feasible.
Purpose
Questions
Realism in Operating Experience Data Collection & Analysis PRA Realism Public Workshop September 30, 2025 John C Lane PE Division of Risk Analysis Office of Nuclear Regulatory Research
Topics
OpE Milestones
Where do we ensure realism in U.S. OpE data?
Who uses OpE data?
How is OpE data gathered, coded, analyzed?
Where is OpE data housed?
How do we ensure realism going forward?
2
Even Presidents Love OpE!
President Carter 1977-make mandatory the current voluntary reporting of minor mishaps and component failures at operating reactorsto develop the reliable databaseto improve reactor design and practice
NPRDS-Nuclear Plant Reliability Data System (1974)
Edison Electric Institute (1933) trade association initiative managed by Southwest Research Institute NUREG-0161-LER Data Entry Guidance for Tech Spec Reportable Occurrences (1977)
Lewis Committee Report (NUREG/CR-0400) on WASH-1400 (Reactor Safety Study)
Calls for better treatment of uncertainty and consequences (issued prior to 79 TMI-2)
NRC expressed support for PRA in regulatory decision-making Accident Sequence Precursor Prog 1st report (1982)
Formalized in LER Rule (1983, 10 CFR 50.73)
LERs managed at ORNL (Sequence Coding Search System (1984) 3
Realism has been Fundamental to OpE IREP (Interim Reliability Evaluation Program)
Assess reliability of components/systems to identify risks Contractors and licensees participated and conducted surveys/studies Maintenance Rule-1991 (10CFR50.65)
Required additional OpE capability to track component failures in safety and non-safety systems Use OpE to help set performance goals One of the reasons INPO created EPIX reporting system INPO/EPIX-1998 (Equip Performance & Information Exchange)
Voluntary Upgrade and expansion of NPRDS Partially funded by NRC contract (to this day) 4
OpE Realism--Today MSPI-Mitigating Systems Performance Index (2005)
Monitor(s) the reliability of key safety systems based on their ability to perform risk-significant functions Uses OpE to balance maintenance (availability) with equipment performance (reliability)
Consensus reliability methods NUREG/CR-6823 HOPE Statistics Handbook (2003)
NUREG/CR-6928 SPAR models use EPIX data (2007)
Components/Systems, IE and LOOP updated every 2 years New SPAR model parameters every 5 years at https://nrcoe.inl.gov Widespread public availability of LERs and OpE reliability and failure rates via LER-Search and internet 5
6
Impactful OpE Results Require Realism NRC/NRR (Near-Term OpE)
Significance Determination Process evaluations following an inspection
Operating Experience Clearinghouse
MD 8.3 NRC Incident Investigation Program
Quick turnaround assessment in the immediate aftermath of potentially risk significant initiating events and conditions,
SPAR models used to help inform decisions about the appropriate level of NRC inspection response
Regional inspectors gain access to INPO IRIS for their plants events
Risk-Informed Treatment of SSCs (10CFR50.59)
Fire PRA NRC/RES (Long-Term OpE)
SPAR Model Parameters-IE, failure rates, common cause, LOOP
Accident Sequence Precursor Program U.S. Industry/Public
Reference material for licensee plant-specific PRA models; Bayesian inference updating
LERSearch https://lersearch.inl.gov/Entry.aspx
Generation of Mitigating System Performance Index 7
Realism in OpE Data Consistency, Attention to Detail and Peer Review NRCs data collection and coding program follows well-established guidance
- Handbook of Parameter Estimation for PRA, NUREG/CR-6823 (HOPE, circa 2003, currently being updated)
- IDCCS Coding Manual
- users guide describes each study and provides guidance for completing each field in the IDCCS
- Data Quality Assurance Program Plan for NRC Div of Risk Analysis Programs at INL (INL/EXT-09-16156)
- Initial coding at INL via experienced nuclear industry professionals (industry or navy experience) followed by independent second person review
- Use of dropdown pick lists to standardize data entry
- Randomly select a sample of records for an independent quality review 8
Where does OpE reside?
Primary Domestic OpE Databases
INPO IRIS Industry Reporting and Information System Primary location for industry-reported failure events Available to NRC staff and industry (Proprietary)
IDCCS-Integrated Data Collection & Coding System at INL NRC Staff and Industry availability via https://nrod.inl.gov/ (proprietary)
NRR ROE (Reactor Operating Experience)
ENs, Morning Reports, Inspection Activities, Plant Status Reports NRC staff 9
INPO IRIS NRC acquires proprietary OpE data via a multi-year contract & MOU IRIS collects
- Failure, reliability and availability information IRIS Supports
- Risk-informed operational decisions
- Maintenance Rule compliance (10CFR50.65)
Realism maintained by varied input from:
- Utility Managers
- Systems Engineers
- Maintenance Rule coordinators
- PRA practitioners 10
IDCCS (INL)
Supports the development of risk models for all U.S. operating nuclear power plants Integrates information contained in IRIS LER coding & repository LER-Search-https://lersearch.inl.gov/LERSearchCriteria.aspx Monthly operating reports from licensees (Generic Letter 97-02)
Supports NRC studies
IE frequencies e.g. LOOP, Loss-of-main-feedwater
Component unreliability e.g. fail-to-start and availability (test/outage)
CCF factors 11
Input
- LERs - Licensee Event Reports
- Industry IRIS Reports
- Monthly Operating Reports
- NRC Inspection Reports Process
- 1. Collect / download data
- 2. Categorize data
- 3. Perform Data Quality Assurance (QA) checks
- 4. Evaluate data
- 5. Input categorized data into appropriate databases Output
- Current LERtrk Database with categorized data
Publicly Available Data Summaries promote public involvement http://nrcoe.inl.gov/resultsdb/
13
Overview of the NROD Web Site
- Restricted access to NRC, INL & NRC licensees Searchable for specific INPO/IRIS data used in models PRA CalculationsRADS/CCF OutageView outage information based on criteria RADS Calculator Reliability, Availability, Initiating Events, LOOP. CCF 14
International OpE Databases International Cooperation provides another layer of peer review
ICDEInternational Common Cause Data Exchange (OECD/NEA/CSNI)
Working Group on Operating Experience (OECD/NEA/CNRA)
Working Group on Risk (OECD/NEA/CSNI)
IRS-Incident Reporting System (IAEA)world-wide proprietary database of event reports
INES-International Nuclear and Radiological Event Scale (IAEA)
CODAP (Component Operational Experience, Degradation and Ageing Programme) for piping failures and stress corrosion (OECD/NEA) 15
Nuclear Industry Safety Improves with Cooperation on OpE Data
- NRC and industry have a good track-record of working together to improve PRA parameters:
- BWR Turbine Driven Pump failures/Misclassified events (2025)
- NRC Audit of PWROG FLEX parameters, (ML22014A084)
- EPRI Update of LPSD study (2022)
- PWROG OpE review, (2021, ML21242A029,30)
- EPRI revisions to LOOP recovery times (2020)
- Westinghouse peer review of CCF events (~2015) 16
Can we avoid excess Conservatism in PRA Parameters?
Unavoidable Facts:
True values of risk parameters can only be estimated
Parameters vary over time, sometimes increasing, sometimes decreasing Recommendations:
INPO IRIS needs to be thorough and concise Inaccurate/incomplete data in IRIS makes the process more expensive to administer and may lead to less accurate/more uncertain result NRC and the industry benefit from thorough and comprehensive reporting of all relevant OpE events to INPO/IRIS and to NRC via LERs Final Thoughts New analytical insights (AI) may contribute towards even better parameter estimation Data uncertainty needs to be accounted for (ASME/ANS RA-S-1.1-2024) HLR-DA-D
- SECY-97-101 (Data Rule) Staff should continue to work with industry to improve the content of the voluntary data
- Minimize the need for compensatory measures to derive needed parameter estimates
- Staff should periodically update the Commission on its efforts to work with industry 17
Questions 18
External Hazards: Flooding Research and Process for Ongoing Assessment of Natural Hazards Information(POANHI)
PRA Realism Public Workshop September 30 - October 1, 2025 Joseph Kanney Timothy Eichler Office of Nuclear Regulatory Research Division of Risk Analysis Fire and External Hazards Analysis Branch
Topics
- Flooding
- Probabilistic Flood Hazard Assessment (PFHA)
Research
- Flooding Regulatory Guidance
- POANHI
- Overview
- Recent Assessment Activities
PFHA Research Phase 1 Technical Basis Projects Phase 3 Guidance Phase 2 Pilot Studies Application Inform &
Update Objective: Transition flood hazard assessment from deterministic maximum credible event basis to probabilistic basis to better inform risk and improve realism Phase 1 - Technical Basis Research - Complete Climate and precipitation Mechanistic, statistical and probabilistic modeling of flooding processes Reliability of flood protection features and procedures Modeling Frameworks Natural Hazard Information Digest (NHID)
Phase 2 - Pilot Studies Local Intense Precipitation (LIP) Flooding - Complete Riverine Flooding - Complete Coastal Flooding - In Progress (report in review)
Completion target: December 2025 Phase 3 - Guidance - In Progress Draft RG Completion target: CY2025 Research Partners:
USGS, USACE, USBR, PNNL, INL, ORNL, UMD EPRI, ASNR
Flooding Guidance Updating existing guidance (in progress)
- Improve realism by incorporating advances in data, models and methods
- Incorporate lessons learned from post-Fukushima flooding reevaluations and impact assessments RG-1.59 Design Basis Floods for Nuclear Power Plants RG-1.256 Guidance for Assessment of Flooding Hazards Due to Water Control Structure Failures and Incidents RG-1.102 Flood Protection for Nuclear Power Plants New guidance (in progress)
- RG-1.XXX Probabilistic Flood Hazard Assessment for Nuclear Power Plants Treatment of uncertainty Local intense precipitation flooding (LIP)
Point precipitation estimates Site-scale hydrologic and hydraulic modeling Riverine flooding Flood frequency analysis Watershed-Scale precipitation estimates Watershed-Scale hydrologic and hydraulic modeling Water control structures Coastal Flooding Storm surge Tsunami Associated Affects
POANHI Overview Scope Hydrology Meteorology Geology and Seismology Components Knowledge base Active technical engagement and coordination Assessment Activities Partners
- Office on Nuclear Reactor Regulation, External Hazards Center of Expertise (EHCOE)
Assessment activities
- Office of Nuclear Regulatory Research Knowledge base activities Technical Engagement and coordination
Knowledge Base
- External Hazards Information Digest (EHID)
- Internal tool for NRC staff use
- Site-specific hazard information for current operating fleet
- design basis, license basis, cumulative information Natural hazards
- Flooding
- Seismic & Geologic
- Meteorological Human-related hazards
- Nearby transportation facilities and routes
- Nearby military and industrial facilities
- Onsite storage facilities
Technical Engagement
- Leverage ongoing periodic interactions with internal and external organizations to facilitate identification of new data, models, and methods.
- Federal partner agencies
- USGS, NOAA, USACE, FERC, USBR, FEMA, NIST
- Industry stakeholders
- Professional Societies
- ANS, ASCE, AMS, AGU, NGWA, SSA
- International counterparts
- IAEA, OECD/NEA, Canada (CNSC), France (ASNR), Korea (KINS, KAERI), Japan (NRA, JAEA),
Assessment Activities
- Performed by NRC subject matter experts
- SMEs aggregate new information with previously collected information
- Evaluate the change in the hazard represented by the aggregated information
- Limited scope quantitative or qualitative initial screening
- Consider available risk insights to determine whether the change in the hazard has a potentially significant effect on plant safety.
- Recent/Ongoing Assessment Activities
- NGA-East
- GAO Audit Response Additional information (including Annual Reports) on NRCs Public Website:
Process for the Ongoing Assessment of Natural Hazard Information (POANHI)
Questions
External Hazards - SEISMIC PRA Realism Public Workshop September 30, 2025 Sunwoo Park Office of Nuclear Reactor Regulation Division of Risk Assessment PRA Licensing Branch C
Topics
- Why Seismic PRA Realism Matters
- Challenges for SPRA Realism
- Enhancing Realism in SPRA o Seismic Hazard Characterization o Seismic Fragility Analysis o Systems and Consequence Analysis
- Conclusion 2
Why Seismic PRA Realism Matters
- Balanced risk profile: Prevents seismic risk from appearing overstated compared to internal events or other hazards.
- Supports licensing and oversight: Provides credible inputs for risk-informed regulatory decisions.
- Better prioritization: Identifies truly risk-significant SSCs and seismic vulnerabilities.
- Enhanced Credibility: Realistic results increase regulator and public confidence in seismic safety.
3
Challenges for SPRA Realism
- Rare but potentially catastrophic events: Limited earthquake experience makes calibration of seismic models challenging.
- Large epistemic uncertainties: Hazard and fragility estimates rely heavily on limited data and expert judgments.
- Significant correlated effects: A single earthquake can simultaneously challenge multiple SSCs across the plant.
- High regulatory impact: Seismic PRA results often drive licensing decisions, making realism essential.
4
Enhancing Realism in SPRA:
(1) Seismic Hazard Characterization
- Apply advanced PSHA: Use robust SSHAC processes with multiple expert models and structured logic trees.
- Adopt modern GMMs: Apply updated ground-motion models (e.g., NGA-East) to improve seismic hazard estimates.
- Site-specific studies: Develop local seismological and geotechnical models instead of relying solely on regional or generic data.
- Update with new data: Periodically revise hazard models with seismic monitoring and advances in seismology.
5
(Example) POANHI Seismic NRC Updates to Seismic Hazard Characterization based on o CEUS Seismic Source Characterization o NGA-East Ground Motion Models o Updated Site Response Analysis o Latest Geologic Site Profiles Industry and NRC staff may use the updated seismic hazards in licensing application and safety evaluation.
o (Case 1) Updated seismic hazard allowed EPRI Tier-1 alternative seismic approach in 10 CFR 50.69 categorization (downgraded from Tier-2).
o (Case 2) Updated seismic hazard produced measurable reduction in SCDF.
6
Enhancing Realism in SPRA:
(2) Seismic Fragility Analysis
- Develop plant-specific fragilities: Use as-built data, walkdowns, and configuration details instead of relying on generic data.
- Consider nonlinear soil-structure interaction (SSI) models:
Capture more realistic response of SSCs, especially under beyond-design-basis earthquakes.
- Use updated test data: Incorporate results from shake-table and component qualification testing.
- Regularly update with operating experience: Integrate insights from actual earthquake events at nuclear or comparable industrial facilities.
7
(Example) Updating Plant-Level Fragility NRC GI-199 provides plant-level fragility data for operating reactors (including HCLPF), largely derived from IPEEE results.
Some values are considered unrealistically conservative (low HCLPF), especially for plants with reduced-scope Seismic Margin Assessments (SMA).
Several licensees increased plant-level HCLPF values published in GI-199 to support their risk-informed license amendment requests (e.g., TSTF-505, 10 CFR 50.69) o HCLPF capacity enhanced in proportion to SSE/GMRS ratio over the 1-10 Hz frequency range.
o HCLPF capacity enhanced by upgrading the reduced-scope SMA (IPEEE) to a focused-scope SMA.
8
Enhancing Realism in SPRA:
(3) Systems and Consequence Analysis
- Model correlated failures realistically: Use physics-based correlations rather than extreme zero or 100% correlation.
- Treat uncertainty explicitly: Separate aleatory and epistemic uncertainties and carry them through to risk metrics.
- Credit non-safety and FLEX equipment realistically: Reflect true availability and limitations without under-or over-crediting.
- Integrate HRA under seismic conditions: Account for stressors (degraded environment, aftershocks), workload, and realistic operator response times.
9
Conclusion
- Seismic PRA is a key contributor to overall plant risk.
- Realism in SPRA is essential:
o Too conservative => seismic risk may be overstated.
o Too optimistic => vulnerabilities may be overlooked.
10
ACRONYMS EPRI: Electric Power Research Institute FLEX: Diverse and Flexible Coping Strategies GI-199: Generic Issue 199 GMM: Ground Motion Model GMRS: Ground Motion Response Spectrum HCLPF: High Confidence of Low Probability of Failure HRA: Human Reliability Analysis IPEEE: Individual Plant Examination for External Events NGA-East: Next Generation Attenuation for Central and Eastern North America NPP: Nuclear Power Plant POANHI: Ongoing Assessment of Natural Hazard Information PSHA: Probabilistic Seismic Hazard Analysis SCDF: Seismic Core Damage Frequency SMA: Seismic Margin Assessment SPRA: Seismic Probabilistic Risk Assessment SSE: Safe Shutdown Earthquake SSHAC: Senior Seismic Hazard Analysis Committee TSTF: Technical Specifications Task Force 11
Questions 12
IDHEAS - status and current activities PRA Realism Public Workshop September 30, 2025 Jing Xing, Sr. Human Performance Engineer Office of Nuclear Regulatory Research Division of Risk Analysis Human Factors and Reliability Branch
What is IDHEAS?
A HRA methodology suite that is:
- Practical
- State-of-the-art
- Strong scientific basis
- Tied to data
- Human-centered - Can be applied to all NRC domains and those outside of nuclear
- Gives detailed, practical predictions on why there is human error and how to prevent/correct it
IDHEAS Suite Scientific and Data Basis Methodology Method and guidance HRA practices Cognitive basis for HRA (NUREG-2114)
IDHEAS-DATA
- Human Error Database (RIL 2025-01)
Human Error Handbook (RIL 2025-02)
Analysis of time uncertainty in simulator data (RIL 2024-03)
IDHEAS-G -
IDHEAS General Methodology (NUREG-2198)
IDHEAS-DATA Methodology (NUREG-2257)
IDHEAS At-Power -
Internal At-power Application (NUREG-2199)
IDHEAS-ECA -
Event and Condition Assessment (NUREG-2256)
IDHEAS-DEP -
Dependency Between Human Actions (RIL 2021-14)
IDHEAS-REC -
Human Error Recovery (NUREG-2259)
IDHEAS-ECA in DI&C work environment IDHEAS-ECA application for radioactive material handling IDHEAS-ECA Application in SPAR model (RIL 2024-17)
IDHEAS-ECA Desktop Guide
IDHEAS Suite Scientific and Data Basis Methodology Method and guidance HRA practices Cognitive basis for HRA (NUREG-2114)
IDHEAS-DATA
- Human Error Database (RIL 2025-01)
Human Error Handbook (RIL 2025-02)
Analysis of time uncertainty in simulator data (RIL 2024-03)
IDHEAS-G -
IDHEAS General Methodology (NUREG-2198)
IDHEAS-DATA Methodology (NUREG-2257)
IDHEAS At-Power -
Internal At-power Application (NUREG-2199)
IDHEAS-ECA -
Event and Condition Assessment (NUREG-2256)
IDHEAS-DEP -
Dependency Between Human Actions (RIL 2021-14)
IDHEAS-REC -
Human Error Recovery (NUREG-2259)
IDHEAS-ECA in DI&C work environment IDHEAS-ECA application for radioactive material handling IDHEAS-ECA Application in SPAR model (RIL 2024-17)
IDHEAS-ECA Desktop Guide
IDHEAS Suite Scientific and Data Basis Methodology Method and guidance HRA practices Cognitive basis for HRA (NUREG-2114)
IDHEAS-DATA
- Human Error Database (RIL 2025-01)
Human Error Handbook (RIL 2025-02)
Analysis of time uncertainty in simulator data (RIL 2024-03)
IDHEAS-G -
IDHEAS General Methodology (NUREG-2198)
IDHEAS-DATA Methodology (NUREG-2257)
IDHEAS At-Power -
Internal At-power Application (NUREG-2199)
IDHEAS-ECA -
Event and Condition Assessment (NUREG-2256)
IDHEAS-DEP -
Dependency Between Human Actions (RIL 2021-14)
IDHEAS-REC -
Human Error Recovery (NUREG-2259)
IDHEAS-ECA in DI&C work environment IDHEAS-ECA application for radioactive material handling IDHEAS-ECA Application in SPAR model (RIL 2024-17)
IDHEAS-ECA Desktop Guide
IDHEAS Realism - Data
- IDHEAS Human Reliability Database (IDHEAS-DATA) o 1,000+ human error datapoints - evaluated, generalized, and integrated o HEP calculation in IDHEAS methods are based on the data o Continuously adding data and periodically updating HEP calculation
- A Human Error Handbook o Verification, summary, and analysis of the data sources in IDHEAS-DATA o Helps HRA analysts understand and model human events in IDHEAS-ECA o Improving consistency of applying IDHEAS-ECA
7 IDHEAS-DATA Example - datapoints from SACADA database SACADA collects operators task performance data in simulator training.
The unsatisfactory performance rates (UNSAT) for training objective tasks were calculated from the SACADA data before April 2019 and used for IDHEAS-ECA.
For example, SACADA characterizes operators scenario familiarity as three options: Standard, Novel, and Anomaly. The datapoints are used for the base HEPs of PIF Scenario Familiarity (SF3.1) for CFMs Failure of Understanding (U) and Failure of Decisionmaking (DM), as shown in the table below SACADA data IDHEAS-DATA Task (Training Objectives)
Situation factors Error rates CFM PIF Uncertainties Operators diagnose in simulator training Anomaly scenario 1.2E-1 (8/69)
U SF3.1 Other PIFs may exist Operators make decisions in simulator training Anomaly scenario 1.1E-2 (1/92)
Where are we now on IDHEAS?
- Obtaining users feedback:
- RIL-2024-17: IDHEAS-ECA WG applied IDHEAS-ECA on 14 SPAR HFEs, 7 internal event HFEs, and 7 FLEX HFEs and provided feedback
- EPRI feedback (workshop, comments)
- Developing IDHEAS-ECA Desktop Guide
- Completing the development of the detailed guidance
- Seeking public comments on IDHEAS-DEP and plan to update the report to NUREG-2258
- Working on the guidance of crediting human error recovery
Where is IDHEAS-ECA heading?
- Complete/Practical/State of the Art HRA Method
- Improves the current state of practice at the NRC
- Program for periodic updates based on user feedback and data
- Exploring uses for material safety and physical/cyber security
Questions
Example: IDHEAS-ECA HEP Quantification Parameters from IDHEAS-DATA 11 PIF Attribute D
U DM E
T MT0 No impact 1
1 1
1 1
MT1 Distraction by other on-going activities that demand attention Weak - 1.2 Moderate - 2 High - 2.8 1.1 1.1 Weak - 1.2 Moderate - 2 High - 2.8 Weak - 1.2 Moderate -
2 High - 2.8 MT2 Interruption taking away from the main task Weak - 1.1 Moderate - 2.8 Frequent or long - 4 Weak - 1.1 Moderate -
1.5 Frequent or long-1.7 Weak - 1.1 Moderate -
1.5 Frequent or long-1.7 Weak - 1.1 Moderate - 2.8 Frequent or long
- 4 Weak - 1.1 Moderate -
2.8 Frequent or long - 4 MT3 Concurrent visual detection and other tasks Low demanding -2 Moderate demanding
- 5 High demanding - 10 NA NA NA NA Demonstration: A portion of Table B-13 PIF Weights for Multitasking, Interruption, and Distraction (From IDHEAS-ECA report NUREG-2256 Appendix B)
Use of Simulator Data to Inform Human Error Probability Estimates Y. James Chang, Ph.D.
Sr. Reliability and Risk Analyst Public Workshop on NRCs Approach to Improving PRA Realism September 30 - October 1, 2025 1
=
Background===
- Objective: Use operator performance data in simulator training to inform human error probability (HEP) estimates
- Activities:
- Collect simulator data
- Exchange simulator-based HEP data with other organizations
- Current state
- Limited simulator-based HEPs were used in IDHEAS-ECA
- Analyze simulator data to support the development of IDHEAS-ECA desktop guide 2
Activity Highlights
- NRC-STPNOC MOU (2011-2023)
- Developed SACADA software tool to collected STPNOCs operator performance data in simulator training (2013 - 2020)
- Used SACADA to collect other simulator data
- The STPNOC, the international and US HRA empirical studies, and Halden simulator experiments
- Organized four international HRA data workshops to exchange simulator-based HEP data (2018-2020, 2024)
- Published two journal papers in Nuclear Engineering and Technology on comparing SACADA and HuREX methods and data (2022) 3
- SACADA: Scenario Authoring, Characterization, and Debriefing Application
Recent Activities Support IDHEAS-ECA desktop guide development
- Issue: NRC and EPRIs limited exercises of IDHEAS-ECA identified about 10 IDHEAS-ECA Performance Influencing Factor Attributes (PIFA) that need additional guidance to improve user consistency
- Project: Develop IDHEAS-ECA desktop guide to enhance the descriptions of nuclear operations context of the PIFAs Revisit the references cited in IDHEAS-DATA and enhanced the data basis with additional simulator data if data are available Jointly conducted two IDHEAS-ECA workshops with EPRI, supported by PWR/BWR OGs, to provide operational context for PIFAs and SACADA data (8/2024 and 3/2025)
Two-phase development: Phase-1 focuses on the PIFAs identified by the NRC and EPRI; NRC develops and EPRI reviews; Estimated to complete in early 2026. Phase-2 covers the remaining PIFAs Explore SACADA data sources
- Presented SACADA to X-energy, Nuclear Regulation Authority (NRA) and Central Research Institute of Electric Power Industry (CRIEPI) of Japan (2024 -
2025) 4
A SACADA Screenshot 5
SACADA data include context information needed by the new generation of HRA methods (e.g., IDHEAS-ECA) to estimate HEP.s.
Overview of PRA Standards Engagement, Current Focus Areas, and Guidance PRA Realism Public Workshop October 1, 2025 Matt Humberstone Office of Regulatory Research Division of Risk Analysis Performance and Reliability Branch
Topics
- PRA Acceptability and Framework
- PRA Standard Priorities and Endorsements
- Conclusions
Applications specific guidance relies heavily on RG 1.200 and the PRA acceptability determination.
dance PRA Standards and Regulatory Guidance
PRA Acceptability
- PRA standards
- Staff Positions and Endorsements
- Trial RG 1.247
- Peer Review Guidance
NRC PRA Standard Priorities and Status PRA Standard Endorsement Status NRC Priority Level 1/LERF RG 1.200 update Completed Level 2 RG 1.200 update Completed Medium Level 3 None Completed Low Non-LWR RG 1.247 Completed Advanced LWR RG 1.200 update In Progress High LPSD RG 1.200 update In Progress High Multi-unit None In Progress Low
LWR PRA Standards Endorsements
PRA Standards Endorsement Plan PRA Standard Endorsement Level 1/LERF RG 1.200 rev 4 Level 2 RG 1.200 rev 4 Level 3 No Plans Non-LWR RG 1.247 Advanced LWR RG 1.200 rev 5 LPSD RG 1.200 rev 5 Multi-unit No Plans
RG 1.200 Revisions
Conclusions NRC representatives on JCNRM Working Groups, Sub Committees, Executive Committee, and Main Committee Continuing our efforts in supporting JCNRM activities 2024 Level 1/LERF PRA standard(Rev 4 of RG 1.200) 2024 Level 2 PRA standard (Rev 4 of RG 1.200)
Advanced LWR Appendix to Level 1/LERF PRA standard (Rev 5 of RG 1.200)
LPSD PRA standard (Rev 5 of RG 1.200)
Moving forward with plan to endorse select PRA standards Leveraging risk-insights in more regulatory activities Focusing on most safety significant issues NRC continues to evolve into a more risk-informed regulator Used to establish PRA acceptability Provides NRC staff confidence in the results of a PRA JCNRMs PRA standards have served an important role in the NRCs RIDM programs
Questions
Accident Sequence Precursor (ASP) Program PRA Realism Public Workshop September 30, 2025 Christopher Hunter Office of Nuclear Regulatory Research Division of Risk Analysis Performance and Reliability Branch
Topics
- Revised ASP Analysis Review Process.
- Inclusion of Hazards Beyond Internal Events
- SPAR Model Feeback
- Common-Cause Failure (CCF) Modeling and Treatment
Revised ASP Analysis Review Process
- In 2024, all precursor analyses started being sent to licensees for a 30-day review period.
- Prior to 2024, only precursors with a CCDP/CDP 10-4 were sent for licensee review.
- This change was made largely due ASP analyses now evaluating hazards other internal events.
- This evaluation includes applicable hazards that may not be included the SPAR model.
- Licensee response is optional.
- Provides an opportunity to ensure that analysis assumptions are appropriate for the event/degraded condition and the results are reflective of the as-built, as-operated plant.
Inclusion of Additional Hazards
- Historically, the ASP analyses of degraded conditions only included hazards that were included in all SPAR models.
- Not including applicable hazards in the ASP evaluations, especially internal fires, could have significant impacts on the results and insights.
- Seismic hazards started being considered in 2018.
- High winds (including hurricanes and tornadoes) started being considered in 2020.
- However, internal fires and floods are not included in all SPAR models.
- In addition, some SPAR models that include these hazards are based on older information.
Inclusion of Additional Hazards (cont.)
- If the SPAR model does not include a specific hazard, analysts will leverage all available risk information.
- License amendment requests.
- Discussion with SRAs.
- Importance measures from licensee PRA.
- IPE/IPEEE submittals.
- Risk from these hazards can be treated quantitatively or qualitatively.
- Given the high uncertainties with this approach, focus is to determine if unmodeled hazards can results in risk impacts result in degraded conditions exceeding precursor thresholds.
SPAR Model Feedback
- Issues identified during analyses are considered for base SPAR model changes.
- Technical issues are identified and forwarded to the GREATR Group for consideration.
- Examples
- FLEX equipment mission time
- Safe/stable end state requirements
- AC power recovery requirements
CCF Modeling and Treatment
- Focusing work in reducing uncertainties with the current data collection and modeling approaches.
- Developed CCF parameters using component-specific priors.
- Use of the generic prior influence CCF impacts of components with dissimilar components that may have a stronger/weaker coupling.
- Developed revised guidance on common cause component group modeling.
- New guidance will reduce overcounting of CCF impacts.
- Developed interim approach to treatment of cross-unit CCF.
- Use of existing parameters will overestimate CCF across units in most cases.
- Initiated research to calculate cross-unit CCF parameters for key components.
Questions
PRA Licensing and Risk-Informed Applications PRA Realism Public Workshop x
October 1, 2025 Adrienne Brown Office of Nuclear Reactor Regulation Division of Risk Assessment
2 Agenda
- Background - Realism, How Did We Get to This Point?
- Insights, What Are You Using this Information for?
- Cross-Cutting Initiatives
3
Background:
Realism, How Did We Get to This Point
- TMI, 1979 - Desire to achieve realistic understanding of plant risk by both NRC and Industry
- 1980s IDCOR, Industry Degraded Core Rulemaking
- Generic Letter 88-20, Individual Plant Examination for Severe Accident Vulnerabilities
- NUREG 1560: Individual Plant Examination Program, Perspectives on Reactor Safety and Plant Performance
- NUREG 1742, Perspectives Gained From the IPE
C.1 Regulatory Position, RG 1.200 PRA Acceptability: Modeling PRA Realism ATTRIBUTES OF PRA ACCEPTABILITY
- 1. SCOPE OF BASE PRA TECHNICAL ELEMENTS OF BASE PRA LEVEL OF DETAIL OF BASE PRA PLANT REPRESENTATION AND PRA CONFIGURATION AND CONTROL (3) REALISM IS INCORPORATED INTO THE ANALYSES THAT REFLECT THE EXPECTED PLANT RESPONSE.
PRINCIPLE 3 OF NATIONAL CONSENSUS PRA STANDARD: TO FACILITATE THE USE OF THE STANDARD FOR A WIDE RANGE OF APPLICATIONS, CATEGORIES CAN BE DEFINED TO AID IN DETERMINING THE APPLICABILITY OF THE PRA FOR VARIOUS TYPES OF APPLICATIONS.
RIDM Risk Insights What are you using this information for?
Licensing
6 On-Going Cross-Cutting Initiatives Risk-Informing 10 CFR 50.59 Rulemaking Risk-Informing 10 CFR 50.36, Technical Specifications, PRM 50-126 Joint Committee on Nuclear Risk Management (JCNRM), Code Case on Risk Significance Executive Order 14300, Section 5(h)
7 Discussion and Q&A
PRA Configuration Control (PCC) Update PRA Realism Public Workshop x
October 1, 2025 Reinaldo Rodriguez Office of Nuclear Reactor Regulation Division of Risk Assessment
2 Agenda
- Background
- 2025 OpESS Status
- Acts and Orders
- Path Forward for PCC
3 Background on the PCC Effort
- Working Group (WG) established to identify suitable oversight of PRA and PCC given the approval of risk-informed programs.
- WG identified a balanced approach (OpESS) within the existing reactor oversight process (ROP) inspection program to monitor implementation of PCC programs of PRA models.
- OpESS 2023/02 on PCC (ML23255A006)
- PCC Cross Regional Panel (CRP) Charter (ML24081A131):
SDP Guidance (Enclosure 2)
PCC Examples of Minor Issues (ML24152A029)
4 OpESS - Summary OpESS has been performed 9 times:
8 - Team Inspections across 4 Regions (3 - CETI, 4 - FPTI, 1 - 50.69)
1 - Resident PI&R Minor program gaps have been identified:
Conducted 5 Cross Regional Panels (All determined to be Minor)
Documentation/Record Keeping Issue Process gaps between plant changes and PRA updates OpESS is still ongoing and available for inspection use.
Utilization has decreased
No planned changes to current OpESS and SDP
PCC specific SDP has not been applied (no more-than-minor issues have been identified).
5 Advance Act and Executive Orders Inclusion of PCC Oversight into the ROP was targeted for 2027 to align with the new ROP cycle.
The ROP is getting a full Re-baselining in Jan 2026.
Current proposal is to include some of the OpESS guidance into the new ROP inspection procedures.
6 Path Forward New ROP baseline recommendations will seek Commission approval sometime in October.
New inspection procedures expected to be implemented by January 1, 2026 to align with the new ROP cycle.
Frequent public meetings are being conducted to keep all stakeholders informed. Next public meeting is expected in October.
7 Discussion and Q&A
PRA REALISM IN ADVANCED REACTOR APPLICATIONS PRA REALISM PUBLIC WORKSHOP SEPTEMBER 30, 2025 Marty Stutzke Office of Nuclear Reactor Regulation Division of Advanced Reactors and Non-power Production and Utilization Facilities
Commission Policy Statements Licensing Pathways PRA Acceptability Industry PRA Consensus Standards Lack of Operating Experience Regulatory Decision Making PRA Realism Advanced Reactors Overview: How Do These Topics Relate?
Commission Policies and PRA Realism
- The Commissions Advanced Reactor Policy (73 FR 60612; October 14, 2008)
- Minimize the potential for severe accidents and their consequences severe accident policy (50 FR 32138; August 8, 1985)
- Expect advanced reactors to comply with the safety goal policy (51 FR 30028; August 21, 1986)
- Use PRA as a design tool PRA policy (60 FR 42622; August 16, 1995)
- PRA evaluations in support of regulatory decisions should be as realistic as practicable
- Long-standing consensus that realistic and useful risk insights may be developed for advanced reactors.
Reference site Site-specific Licensing Pathways and PRA Realism Part 50 Construction Permit (CP)
Part 50 Operating License (OL)
Part 52 Early Site Permit (ESP)
Part 52 Manufacturing License (ML)
Part 52 Combined License (COL)
Part 52 Standard Design Approval (SDA)
Part 52 Standard Design Certification (DC)
§ 50.71(h)(1)
Fuel Load PRA for COL Holders Increasing PRA scope, level of detail, and realism no PRA requirement no PRA requirement no PRA requirement
PRA Acceptability and PRA Realism PRA Acceptability Area Definition Nexus to PRA Realism Scope of the PRA Metrics used to characterize risk
- Use of risk surrogates?
- Radiological sources
- Plant operating states
- Causes of initiating events (hazard groups)
- Screened out vs. set aside for later?
Level of Detail of the PRA Resolution of the modeling used to represent the behavior and operations of the plant
- Detailed vs. simplified modeling?
Elements of the PRA Fundamental technical analyses needed to develop and quantify the PRA model for its intended purpose
- Realistic vs. bounding analyses?
- Alternative methods (e.g.,
seismic margins, EPRI FIVE)?
- Approximations?
Plant Representation and PRA Configuration Control How closely the PRA represents the plant as it is designed, built, and operated
- Analysis lags design?
Sources: RG 1.200 and RG 1.247
Non-LWR PRA standard Technical Elements and Subelements Programmatic Elements Technical High-level Requirements Documentation High-level Requirements Supporting Requirements (see notes)
Supporting Requirements Same SR for both Capability Categories Unique SR for each Capability Category Notes:
- Some SRs only apply to certain lifecycle stages
- Apply risk assessment application process in the standard to determine applicable SRs 239 (22%)
838 (78%)
PRA Configuration Control Peer Review The Non-LWR PRA Standard and PRA Realism Capability Category impacts PRA scope, level of detail, plant-specificity, and realism Source: ASME/ANS RA-S-1.4-2021
Lack of Operating Experience and PRA Realism Much operating experience from the current fleet may be applied to new and advanced reactors.
- SSC reliability data exchangeability
- Identical design, and
- Same operating conditions
- NUREG-2201: The assumption of exchangeability is approximately valid at the component level.
Expert elicitation
- NUREG-2255, Guidance for Conducting Expert Elicitation in Risk-Informed Decisionmaking Activities, 2024
- NUREG-2213, Updated Implementation Guidelines for SSHAC [Senior Seismic Hazard Analysis Committee] Hazard Studies, 2018 statistics, as a subject, is the study of frequency type information. That is, it is the science of handling data. On the other hand probability, as a subject, we might say is the science of handling the lack of data.
Stanley Kaplan and B. John Garrick, On the Quantitative Definition of Risk, Risk Analysis, Vol. 1, No. 1, 1981
Regulatory Decision Making and PRA Realism
- Consider the role of the PRA in the application.
- Use importance measures and sensitivity analyses to focus on what is important.
- Use defense in depth to compensate for uncertainties and lack of PRA realism.
Regulatory decision making must be based on the current state of knowledge:
The current state of knowledge regarding design, operation, and regulation (as reflected in the PRAs) is key.
The current state of knowledge is informed by science, engineering, and operating experience, including past incidents.
PRAs evaluate and assess potential accident scenarios to inform the decision makers current state of knowledge.
Commissioner George Apostolakis PSAM 12, Honolulu HI, June 23, 2014