ML20296A443
| ML20296A443 | |
| Person / Time | |
|---|---|
| Issue date: | 10/21/2019 |
| From: | Patrick Raynaud NRC/RES/DE/CIB |
| To: | |
| Raynaud P | |
| References | |
| Download: ML20296A443 (15) | |
Text
Development of NRC Guidance on Probabilistic Fracture Mechanics for US Nuclear Applications Patrick Raynaud Senior Materials Engineer U.S. NRC Office of Research
Probabilistic Fracture Mechanics (PFM) at the NRC: A Little Bit of History
- Historical reliance on deterministic fracture mechanics and conservatisms
- Risk-Informed policies in place since the mid-1990s
- Gradual shift to probabilistic approaches and best-estimate with quantified uncertainties
- Recent increased use of probabilistic methods
- Unanticipated factors
- Risk informed cost benefit analyses
- Plant aging, license extensions: highly conservative safety factors impractical for plant life extension 22-24 October 2019 ISPMNA'2019 l Rockville, MD 2
NRCs Current Research on PFM: Motivations
- Increased complexity of probabilistic analyses
- NRCs PFM tools for component integrity and associated regulatory difficulties assessment
- Importance of QA and V&V - FAVOR (1994-Present)
- Cultural change in engineering approach - xLPR (2009-Present) 22-24 October 2019 ISPMNA'2019 l Rockville, MD 3
PFM Best-Practices Guidance Development
- NRC research project to develop best-practices guidance for PFM code development and analyses
- Objectives:
- Develop a robust technical basis for PFM codes and analyses
- Provide guidance on desirable attributes for PFM tools in view of regulatory acceptance
- Provide guidance on acceptable methodologies for PFM analyses in support of licensing actions
- Remediate difficulties in reviewing industry submittals using PFM
- Technical Letter Report on NRCs preliminary thoughts on increasing confidence in PFM analyses publicly available at ML18178A431
- Focus of this presentation is on developing a graded approach for PFM submittals 22-24 October 2019 ISPMNA'2019 l Rockville, MD 4
Graded Approach Overview
- Based on EPRIs white paper on minimum contents for PFM submittal
- Supplemented with additional contents and explanation from NRC
- EPRI BWRVIP 2019-016 white paper: Suggested Content for PFM Submittals to the NRC, ML19241A545
- Performed cross-walk between NRCs steps for PFM analysis and EPRI white paper
- Analyzed proposed thresholds/considerations for when additional information would be required
PFM Submittal Recommended Minimum Contents (NOT Requirements) [1/3]
- 1. Software information made available to NRC staff with PFM submittals
- Need to be able to access the code in following cases
- High safety significance
- Generic application of plant specific code
- Complex code
- Extent of differences with codes previously approved by NRC
- In person or virtual audits
- 2. Models
- May need more details in these cases:
- New failure mode
- Emergent vs. ongoing: extent of plant experience and Operational Experience for new phenomena
- Implications of unknowns 22-24 October 2019 ISPMNA'2019 l Rockville, MD 6
PFM Submittal Recommended Minimum Contents (NOT Requirements) [2/3]
- 3. Inputs
- Provide basis for categorization of uncertainty between epistemic and aleatory
- 4. Convergence
- Make sure convergence is achieved for Quantities of Interest (QoI), not necessarily for other things we dont care about
- Justify why sampling uncertainty is small enough for intended purpose
- 5. Input Importance and Sensitivity Studies
- Document following details
- Sensitivity Analysis (SA) technique, assumptions, and results
- Which QoI are the rankings based on?
- Explain how SA results influenced subsequent analysis
- Describe sensitivity studies 22-24 October 2019 ISPMNA'2019 l Rockville, MD 7
PFM Submittal Recommended Minimum Contents (NOT Requirements) [3/3]
- 6. Verification and Validation
- Allow for audit of software QA documentation
- Identify different code categories (see next slide)
- 7. Uncertainty propagation and output uncertainty characterization
- Describe output uncertainty characterization for QoI
- 8. Acceptance Criteria
- PFM outputs should be relevant for desired regulatory outcome
- Acceptance criteria are beyond scope of NRCs PFM guidance but should be derived based on risk informed decisionmaking principles
Software QA and V&V Categories
- Category 1: NRC approved code
- Category 2: commercial off-the-shelf (COTS) software
- Category 1A: NRC approved or endorsed code within validated designed for the specific purpose of the application range - Generally not Excel, GoldSim, FE software
- Demonstrate code applicability within validated range
- Demonstrate code applicability
- Category 1B: NRC approved or endorsed code outside of
- Description of the software and its pedigree validated range
- Software and documentation available for review upon request
- Provide justification for new applicability range (additional (audit)
V&V?)
- Category 3: custom code
- Category 1C: Modified NRC approved or endorsed code
- Summary of SQA program and implementation (standards?)
- SQA summary and V&V description for modified portions of
- Summary of V&V activities (data, benchmarking) the code
- Very simple applications: provide source code instead of
- Demonstration that the code was not broken as a result of standardized SQA and V&V?
changes
- Detailed documentation available for further review upon
- Normal care needs to be taken for pre and post processing request (audit) codes 22-24 October 2019 ISPMNA'2019 l Rockville, MD 9
Considerations for Submittal of Additional Depth of Information [1/3]
- Safety significance Depth of additional information
- In general, level of detail should scale with the safety significance while still taking into consideration the recommended minimum contents
- The recommended minima are satisfactory for class 1 components, and thus also sufficient for lower class components
- The extent to which some recommendations may be relaxed should inversely scale with the complexity and novelty of the application
- Failure mode
- Highly energetic failure mode, impact on other systems or on the safety of personnel, poorly understood failure mode, newly modeled phenomenon
- Impact on submittal
- Higher emphasis on model description
- Better description of model inputs
- Better documentation of sensitivities of model
- Code complexity Safety significance
- High number of I/O, large number of phenomena modeled, many code model interdependencies, complex interactions between different physics in the code Failure consequences
- Impact on submittal Code complexity
- More sensitivity studies to ensure behavior is well understood
- More robust SQA and V&V NRC would encourage pre-submittal meeting to provide direction on which areas might require additional work in the submittal 22-24 October 2019 ISPMNA'2019 l Rockville, MD 10
Considerations for Submittal of Additional Depth of Information [2/3]
- Margin to acceptance criteria Depth of additional information
- Does 95th percentile cross threshold?
- Impact on submittal:
- Higher emphasis on input distributions for low margin
- Better documentation of output uncertainty (tails) for low margin
- Potential relaxation of convergence requirements if lots of margin
- Plant specific vs. generic
- Additional information for generic applications
- Baseline level of information for plant specific applications
- Exception: if plant has unique feature requiring special software or inputs
- Impact on submittal:
- Additional proof that inputs cover wide range of generic application
- Additional proof that models cover wide range of generic application
- For plant with unique feature, additional description of specialized software and inputs
- Implications of potential unknowns Margin
- If perceived uncharacterized uncertainties are high, recommend additional sensitivity Number of plants impacted studies Implication of unknowns
- The higher the consequence, the broader the scope of investigation Difference from ASME
- Difference from ASME requirements
- The bigger the difference, the more additional information may be needed NRC would encourage pre-submittal meeting to provide direction on which areas might require additional work in the submittal 22-24 October 2019 ISPMNA'2019 l Rockville, MD 11
Considerations for Submittal of Additional Depth of Information [3/3]
- Emergent or ongoing issue Depth of additional information
- Emergent issues require more information than ongoing issues
- Impact on submittal:
License Amendment
- Better documentation of inputs and models Emergent Issue First-of-a-kind
- First of a kind vs routine applications Sole basis
- First of a kind applications require more information than routine Ongoing Issue Relief Request Supporting basis applications
- Impact on submittal:
Routine
- Better documentation of inputs and models
- Better characterization of importance and sensitivities
- Change to plant licensing basis
- More information required if plant licensing basis is impacted
- Impact on submittal:
- Better documentation and characterization of margins via uncertainty analysis as well as sensitivity analyses and studies
- PRA requirements NRC would encourage pre-submittal meeting to
- PFM sole basis vs. supporting basis of submittal provide direction on which areas might require additional work in the submittal
- If other supporting analyses, relax recommendations
- Impact on submittal
- Less need for sensitivity analyses and sensitivity studies if other analyses to show trends, etc.
22-24 October 2019 ISPMNA'2019 l Rockville, MD 12
Summary of NRCs Proposed Graded Approach for PFM Regulatory Submittals in the US
- NRC used EPRIs proposed contents for a PFM submittal as a starting point
- EPRI BWRVIP 2019-016 white paper: Suggested Content for PFM Submittals to the NRC, ML19241A545
- NRC adopted the majority of EPRIs recommendations and added recommendations where deemed necessary and appropriate
- NRC defined a categorization scheme for PFM software, and described recommended supporting information for each category
- NRC reviewed EPRIs recommended considerations for submittal of additional depth of information and provided additional guidance where deemed necessary and appropriate 22-24 October 2019 ISPMNA'2019 l Rockville, MD 13
Next Steps in NRCs Development of PFM Guidance for US Regulatory Applications
- NRC is in the process of developing final drafts of a Regulatory Guide and supporting NUREG technical bases
- Contents will follow the general principles described in Technical Letter Report on NRCs preliminary thoughts on increasing confidence in PFM analyses (publicly available at ML18178A431)
- A graded approach for PFM will be recommended (starting point is EPRI recommendations from Suggested Content for PFM Submittals to the NRC, publicly available at ML19241A545)
- Once documents are finalized and internal concurrence is obtained, NRC will hold public meeting to describe the draft guidance
- Process for publication of the PFM Regulatory Guide will include chances for stakeholder feedback 22-24 October 2019 ISPMNA'2019 l Rockville, MD 14
Acknowledgements NRC wishes to acknowledge EPRIs important contributions to the PFM guidance project, through public interactions and feedback, and especially through the development of their white paper on Suggested Content for PFM Submittals to the NRC
- NRC staff involved in the PFM Guidance project: David Rudland, Stephen Cumblidge, Mike Benson, Matt Homiack, Mark Kirk (now at PEAI and CRIEPI)
- Sandia National Labs staff involved in the project: Mike Starr, Nevin Martin, Remi Dingreville, Nevin Martin (SNL), Lauren B. Hund, Zach Stuart, John R. Lewis, Josh Mullins, Dusty Brooks, Aubrey C.
Eckert-Gallup
- Emc2 staff involved in the project: Cedric Sallaberry, Bob Kurth, Elizabeth Kurth 22-24 October 2019 ISPMNA'2019 l Rockville, MD 15