ML24193A005
| ML24193A005 | |
| Person / Time | |
|---|---|
| Issue date: | 06/27/2024 |
| From: | Stephen Cumblidge, Dan Widrevitz NRC/NRR/DNRL/NPHP, NRC/NRR/DNRL/NVIB |
| To: | |
| References | |
| Download: ML24193A005 (1) | |
Text
Risk-Informing Materials Issues Risk Informed Materials Assessment Project Dan Widrevitz Stephen Cumblidge Technical Exchange Meeting June 27, 2024
Topics
- Purpose and Applicability of RIMA Project
- Defense-in-Depth
- Safety Margin
- Risk Impacts (use of risk insights)
- Performance Monitoring
- Tier List
- Sampling Considerations
- Sampling Analysis 2
RIMA - Purpose Risk-Informed Materials Assessment Project A risk-informed materials engineering forward guidance development project Leveraging the processes and guidance of RG 1.174, ADM-200, etc. to enable more efficient and effective reviews Providing applicants and reviewers guidance in utilizing risk-informed decision making for non-integrated reviews 3
RIMA - Applicability Target submittals:
LIC-206 Box 7 Type applications and reviews with non-integrated teams (e.g. materials engineers and counterparts only) 4
RIMA - Preliminary Concepts Staff has been generating a preliminary set of RIMA concepts to support potential guidance document development What it is (will be):
Clearer/broader guidance in the language of materials engineers Applicant guidance to enable high quality submittals and efficient staff review What it is not (will not be):
New policy Deviation from RG 1.174 The following slides detail current preliminary concepts 5
RIMA - Defense in Depth Further clarify the relationship between materials engineering topics and defense-in-depth considerations.
Typically, materials engineering reviews do not establish defense-in-depth characterizations, rather materials engineering supports commensurate level of assurance based on characterization.
Is treatment of subject systems commensurate with defense-in-depth functions of subject systems.
6
RIMA - Defense in Depth Key consideration:
Is there enough assurance from other four Principles of RIDM to credit subject system for defense in depth?
Risk Analysis Current Regulations Met Performance Monitoring Safety Margins Defense-In-Depth 7
RIMA - Safety Margins Further clarify the relationship between materials engineering topics and safety margin considerations.
Key consideration:
Are safety margins large enough, in concert with other Principles of RIDM, to manage uncertainties?
Material Reliability Time What is Designed What Is Built How it Ages Failure Safety Margin Design Criteria Quality Control Performance Monitoring/ISI Minimum Allowed Margin Repair/Replacement 8
RIMA - Risk Impacts Clarification and discussion of risk insights derived from qualitative or non-PRA modeling (e.g. PFM).
How insights related to one or more elements of the Risk Triplet (what can go wrong, how often, and what are the consequences) can be leveraged.
(More in a few slides) 9
RIMA - Risk Impacts PFM is often a Risk Impact insight:
Risk Triplet What can go wrong?
How often?
What are the consequences?
Frequency of potential initiating event such as LOCA 10
RIMA - Performance Monitoring Further clarify the relationship between materials engineering performance monitoring and the other Principles of RIDM.
Expanded discussion of performance monitoring and bathtub curb relationship.
Discussion of management of novel performance monitoring results.
11
RIMA - Performance Monitoring Performance monitoring adequacy rests on several pillars.
How much monitoring?
What kind of monitoring?
How often?
Are there triggers for more or less monitoring within program?
Answers to these questions must be judged in context of other Principles of RIDM (e.g. how does subject system support defense-in-depth, etc.)
12
RIMA - Tier List The materials staff wanted a risk ranking of important systems to help risk-inform materials reviews The NRC Staff used the SPAR-Dash tool to rank important systems For this work we have decided to focus on broad systems rather than components 13
Tier List - Fleetwide System Importance 14
Tier List - PWR and BWR System Importance 15
Tier List - Final Tier List Fleetwide Tier X Reactor Protection Emergency/Aux Feedwater Tier 1 High Pressure Injection Reactor Coolant Residual Heat Removal Service Water Component Cooling Water Tier 2 Low Pressure Injection Main Steam Tier 3 Main Feedwater Low Pressure Core Spray PWR Tier X Reactor Protection Emergency/Aux Feedwater Tier 1 Service Water High Pressure Injection Reactor Coolant Component Cooling Water Main Steam Tier 2 Low Pressure Injection Residual Heat Removal Tier 3 Main Feedwater BWR Tier X Reactor Protection Residual Heat Removal Tier 1 Service Water Tier 2 Main Feedwater Low Pressure Injection Tier 3 High Pressure Injection Low Pressure Core Spray Main Steam Component Cooling Water 16
Tier List - Takeaways The Tier List is not directly applicable to individual plant reviews as there is large variability in risk significance from plant to plant for the same systems The risk rankings are helpful to focus attention when looking at fleet-wide issues The list is a tool to teach new staff about the importance of different systems 17
RIMA - Sampling Considerations Expanded discussion of performance monitoring including framework to help identify target concepts supporting optimization of performance monitoring Includes discussion of qualitative factors as well as a sample statistically driven sampling calculation Leverage bathtub curve terminology to create common language for discussion 18
RIMA - Sampling Considerations The following tables are initial thoughts regarding the impact of various considerations on necessary sampling.
- Means a consideration likely indicates a particular column applies
- Means a consideration increases emphasis
- Means a consideration decreases emphasis Color vs. implies a stronger or weaker association between a consideration and a particular column.
19
Sampling Considerations - Generic Life Stage Generic life-stage determination table Burn-in Maturity Wear-out Novel material, process, or design Novel repair Repair Novel degradation mechanism identified Novel degradation parameters (CGR, etc.)
Degradation threatening function PSI only PSI + 1 interval of ISI PSI + more than 1 intervals of ISI
- Checks in multiple columns are ors 20
Sampling Considerations - Qualitative Factors Qualitative factors affecting sampling intensity table Component level sampling Population Level sampling Burn-in Maturity period Wear-out Safety related RISC-2 (50.69 approved designation, system designation)
Consequence significant Aging management program Failure tolerant (LBB, etc.)
Low impact on other safety significant systems Redundant Isolable
- Gray marks indicate that column should be considered but is not a priori necessary21
Sampling Considerations - Emergent Events Qualitative factors affecting sampling due to emerging events table Component level sampling Population level sampling Site sampling expansion Population sampling expansion Site-specific event or chemistry issue
Novel indications identified at a single site
Novel indications identified at multiple sites
OE limitations (e.g. low coverages or other issues)
Extensive OE demonstrating no degradation
Extensive OE demonstrating limited degradation
Extensive OE demonstrating unmodeled degradation
Extensive OE demonstrating modeled degradation
- Marks in multiple columns indicate row indicates multiple columns are applicable or should be considered 22
RIMA - Sampling Analysis Quantitative sampling calculation can be derived from statistical calculations For example, NRC staff leveraged this in support of review of PROMISE Code submittals Detailed discussion of approach in PVP2023-105203, Statistical Approach to Developing a Performance Monitoring Program 23
Binomial Distribution
- The binomial distribution is frequently used to model the number of successes in a sample of size n drawn with replacement from a population of size N
- Can be used to find # of inspections needed to find a crack
- Only a function of the number of inspections and the % cracked
- Very easy to use (beware of limitations)
,, =
1
=
k= number of successes (cracks found) n=number of trials (inspections) p= probability of success on an individual trial
(% of population cracked)
If k=0 then this is the probability of no successes is:
, = 1 and therefore, the probability of at least one success is:
1,
24 24
Monte Carlo Analysis
- Same idea can be developed through a MC analysis
- Allows maximum flexibility in analysis
- Binomial response can be recreated Sample a weld population with x%
cracked Loop on weld population cracked?
Sample Inspection of y% of welds inspected?
count=count+1 MC Loop, n realizations p=count/n yes yes no no Done?
yes no Done?
yes no For large populations For small populations 25 25
Sampling Combining insights from Sampling Consideration slides with Sampling Analysis approaches allows for high quality proposals in performance monitoring space 26
Take Aways RIMA Project aims to build forward from RG 1.174 and similar guidance in materials engineering specific language Focus is on non-integrated (e.g. NRC materials engineer reviewer only) submittals Guidance on all five Principles of RIDM to be translated and extended Tier List and Sampling Considerations provide increased domain specific granularity of guidance 27