ML20289A694
| ML20289A694 | |
| Person / Time | |
|---|---|
| Issue date: | 10/08/2020 |
| From: | Michael Montecalvo NRC/NRR/DRA |
| To: | |
| Kichline M | |
| Shared Package | |
| ML20289A715 | List: |
| References | |
| Download: ML20289A694 (11) | |
Text
NRCs audit results for the PWROGs FLEX OpE data collection and analysis Mike Montecalvo, U.S. NRC NEI FLEX Summit September 2020 1
Objective
- Brief History
- Audit Team
- PWROG approach
- Audit Results
- Path Forward
- Questions 2
Project History
- One of the two remaining challenges to crediting FLEX
- Highlighted from the beginning of crediting FLEX project
- Credit for FLEX has been provided
- Data is needed to support efficient decisions
- PWROG took over the effort (2017)
- Access to the PWROG report (March 2020)
- Audit - March 24-25 extended to May 4th 3
Commissions PRA policy statement:
PRA evaluations in support of regulatory decisions should be as realistic as practicable and appropriate supporting data should be publicly available for review.
Audit Team
- Team
- Matt Humberstone (NRR/DRA)
- John Lane (RES/DRA)
- Mike Montecalvo (NRR/DRA)
- Frank Arner (RI SRA)
- Zhegang Ma (INL)
- Idaho National Lab Support
- Shawn St. Germain
- John Schroeder
- Andrea Mack (Statistics)
- Cynthia Gentillon (Statistics) retired
- Cory Atwood (Statistics) retired
- Tom Wierman 4
PWROG Approach PWROG-18043-P rev.0, FLEX Equipment Data Collection and Analysis GOAL: Consistent with NUREG/CR-6928, Industry-Average Performance for Components and Initiating Events at U.S. Commercial Nuclear Power Plants Did not address CCF or unavailability Data Compilation Data from all nuclear sites Condition reports (CRs), preventative maintenance (PMs),
and EPRI PM database Component Boundaries Common FLEX equipment (589 pieces of equipment, 16 categories) 32 total parameters (16 categories with FTS and FTR) 5
PWROG Approach Continued PWROG-18043-P rev.0, FLEX Equipment Data Collection and Analysis Used INLs Nuclear Reliability and Operating Experience Database (NROD) and the RADS calculator to support their calculations Data Analysis Approach (3 methods)
Empirical Bayes (EB) (8 parameters)
Jeffreys Noninformative Prior (JNI) (13 parameters)
Weakly Informed Prior (WIP) (11 parameters)
Parameters with weak operating experience
< 50 demands or <100 hrs operating Represents 54 out of 582 pieces of FLEX equip (~ 9% )
Uses permanently installed equip as priors Engineering judgement 6
PWROG Results PWROG-18043-P rev.0, FLEX Equipment Data Collection and Analysis Table 6-1, Generic Failure Rates for FLEX Equipment Failure Mode Distribution Mean Method (EB, JNI, or WIP)
Most failure probabilities are greater than similar permanently installed equipment FLEX DGs ( 5x FTR, 3x FTS)
Combustion Turbine Generator (1.5x FTS, 0.2x FTS)
Motor-Driven PD pump (80x FTR, 8.7x FTS) 7
Audit Result/Observations
- PROCESS
- Need for explicit failure definition
- Formal update process needed
- Component Boundary definitions DATA COLLECTION
- Overall data pedigree (PM frequencies aligned with actual starts)
- Data collected from before FLEX was officially established
- Does run data reflect loaded conditions
- DATA ANALYSIS
- WIP method relies heavily on engineering judgement
- Potential errors in WIP results
- CCF and unavailability 8
Path Forward
- PWROG to review the NRCs audit summary
- Potential adjustments to process
- Finalize FLEX OpE data analysis report
- Publicly available
- NRC audit team plans to review PWROGs final report
- NRC will establish a position on final report
- The process/results could have significant changes 9
Conclusions
- NRC Audit Summary (ADAMS Accession No. ML20155K827)
- INL/EXT-20-58327, Evaluation of Weakly Informed Priors for FLEX data, (ADAMS Accession No. ML20155K834)
- Audit Conclusion If these observations and concerns are addressed in the PWROGs updated approach, this will provide a robust basis for FLEX equipment failure probabilities.
10
Questions?
1 1
11