ML24052A343
| ML24052A343 | |
| Person / Time | |
|---|---|
| Issue date: | 02/20/2024 |
| From: | Christopher Hunter NRC/RES/DRA/PRB |
| To: | |
| References | |
| Download: ML24052A343 (1) | |
Text
Status of the U.S.
NRCs ASP Program Chris Hunter U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research Division of Risk Analysis Performance and Reliability Branch
Program Status and Challenges The number of precursors identified remain at historical low levels.
- The 81 precursors identified in the past decade is the lowest 10-year period total since the ASP Programs inception.
- The number of LERs and potential precursors identified also remain at historical low values.
The potential for missing risk significant degraded conditions not reported remains.
- The NRC is planning to evaluate the use of AI to explore this issue.
Evaluating nonmodeled hazards (e.g., internal fires) remains a challenge.
Increased use of IDHEAS-ECA HRA method.
- Including the IDHEAS dependency method.
February 20-21, 2024 26th PSAEA Technical Meeting 2
LER Data February 20-21, 2024 26th PSAEA Technical Meeting 3
ASP Results and Trends
2023 Precursors Plant/Description LER/IR Event Date Exposure Time CCDP/
CDP Summer, Failure to Correct a Condition Adverse to Quality Results in an EDG Failure (ML23342A000) 05000395/2023002 (No LER issued) 11/2/22 162 days White Finding Farley 1, Failure to Identify Nonconforming Work Instructions Results in EDG Lube Oil Leak (ML23263B166) 348-23-001 2/26/23 115 days White Finding Calvert Cliffs 1, Failure to Establish and Implement Adequate Maintenance Practices Contributes to the Failure of EDG 1A (ML23297A192) 05000395/2023050 (No LER issued) 4/24/23 216 days White Finding Calvert Cliffs 2, Failure to Establish and Implement Adequate Maintenance Practices Contributes to the Failure of EDG 1A (ML23297A192) 05000395/2023050 (No LER issued) 4/24/23 196 days 2x10-6 February 20-21, 2024 26th PSAEA Technical Meeting 5
Potential Precursors Still Being Analyzed February 20-21, 2024 26th PSAEA Technical Meeting 6
Plant/Description LER/IR Event Date Exposure Time Braidwood 2,Train B AFW Pump was inoperable due to Degraded Oil in the Crank Case 457-23-001 9/22/23 127 days Farley 1, RHR Pump Inoperable for Longer Than Allowed by TS 348-23-002 10/2/23 35 days
Grand Gulf Loss of Feedwater Transient Loss of reactor feedwater with RCIC unavailable due to maintenance.
- Feedwater was recoverable.
Used IDHEAS-ECA to evaluate two HFEs.
- Operators fail to recover reactor feedwater.
- Operators fail to initiate emergency depressurization.
Used IDHEAS dependency method to determine if additional dependency considerations were needed.
- ASP Programs first use of this method.
Final mean CCDP was calculated to be 5x10-6.
- Below the plant-specific CCDP for a nonrecoverable loss of condenser heat sink and feedwater (1x10-5).
February 20-21, 2024 26th PSAEA Technical Meeting 7
Ginna Concurrent UAs of Both EDGs EDG B failed to load during sequence testing during refueling outage due to breaker failure to close.
Last successful operation was during monthly testing while in Mode 1.
- Exposure time of ~4 days.
EDG A was unavailable due to testing for 4.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> during this time.
Implemented appropriate changes to FLEX modeling (e.g., revised FLEX hardware reliability data).
Final mean CDP was calculated to be 3x10-7.
- Below the ASP threshold for degraded conditions (1x10-6).
- Qualitatively evaluated internal fires.
Used licensee importance measure for EDG B as confirmatory calculation.
February 20-21, 2024 26th PSAEA Technical Meeting 8
Precursor Trends (2014-2023)
Precursor Group Trend p-value All Precursors Decreasing 0.02 Important Precursors (i.e., CCDP/CDP 10-4)
No Trend 0.3 Precursors with CCDP/CDP 10-5 No Trend 0.2 Initiating Events Decreasing 0.004 Degraded Conditions No Trend 0.3 LOOPs No Trend 0.2 EDG Failures No Trend 0.6 BWR Precursors Decreasing 0.02 PWR Precursors No Trend 0.3 February 20-21, 2024 26th PSAEA Technical Meeting 9
All Precursors Trend February 20-21, 2024 26th PSAEA Technical Meeting 10
EDG Failure Precursors Trend February 20-21, 2024 26th PSAEA Technical Meeting 11
Building an IDHEAS-ECA Knowledge Base
Introduction The NRC has begun applying Integrated Human Event Analysis System for Event and Condition Assessment (IDHEAS-ECA) in various risk-informed activities.
The first step is to build a knowledge base of IDHEAS-ECA application examples.
Completed six initial evaluations of base SPAR model HFEs identified as being the most risk significant.
February 20-21, 2024 26th PSAEA Technical Meeting 13
IDHEAS-ECA Examples February 20-21, 2024 26th PSAEA Technical Meeting 14 HFE HEP Notes Operators fail to initiate feed and bleed cooling 5x10-3 A larger HEP of 4x10-2 was calculated for scenarios where feedwater or condensate is potentially recoverable. In addition, some plants may have a reduced time window, which would result in higher Pt and, therefore, higher overall HEP (as compared to the base case).
Operators fail to initiate HPR or LPR 5x10-3 0.1 The base case HEP is not expected to change for scenarios other than medium and large LOCAs. The IDHEAS-ECA HEP for low-pressure recirculation is higher due to relatively short time window and, therefore, it is dominated by the Pt.
Operators fail to trip RCPs 1x10-3 The IDHEA-ECA base case HFE is for loss of service water scenarios that occur prior to the reactor trip. For scenarios where seal injection and cooling are lost after the reactor trip, the time window will be reduced resulting in a more significant influence of Pt and, therefore, a higher overall HEP.
Operators fail to emergency depressurize the reactor 4x10-3 Bounding case of medium LOCA will likely result in larger HEPs due to reduced time window resulting in dominant Pt contribution to the overall HEP (e.g., 9x10-2).
The SPAR HEP is lower than can be calculated using IDHEAS-ECA without recovery credit.
Operators fail to vent containment 6x10-3 This HEP is not expected to increase significantly for other scenarios due to large Tavail estimates even for bounding scenarios such as a large LOCA. Pc contributions could increase the HEPs for plants with a lot of execution steps outside the MCR.
Operators fail to initiate SPC 4x10-3 Bounding case of ATWS will likely result in larger HEPs due to reduced time window resulting in dominant Pt contribution to the overall HEP. The SPAR HEP is lower than can be calculated using IDHEAS-ECA without recovery credit.
Key Observations February 20-21, 2024 26th PSAEA Technical Meeting 15 The HEPs for the same HFE calculated for scenario and technological variabilities show significant increase or decrease from the base case HEP.
If the Understanding or Decisionmaking cognitive failure modes are selected, IDHEAS-ECA calculations result in a practical minimum HEP of 1x10-3.
The procedure for evaluating Pt and the selection of performance influencing factors can be enhanced.
Next Steps An evaluation of the SPAR model FLEX HFEs is being performed.
- The evaluations will leverage the information and previous work from the IDHEAS-ECA FLEX pilot study documented in RIL 2020-13, Volume 2 (ML21032A119).
A better understanding of the impact of potential changes to the HEPs in the SPAR models is needed.
Development of an IDHEAS-ECA Desktop Guide.
February 20-21, 2024 26th PSAEA Technical Meeting 16
Recent Significance Determination Process (SDP) Issues & Changes John David Hanna U.S. Nuclear Regulatory Commission Office of Nuclear Reactor Regulation Division of Risk Analysis PRA Oversight Branch
Program Status and Challenges Since the start of ROP (CY 2000), the NRC averages 15 finalized Greater-than-Green (GTG). Total = 364.
CY 2018 saw a substantial decrease in the number of issued GTGs.
CY 2023 saw a return nearer to the average with 14 GTGs issued across all cornerstones.
Some challenges have been:
- Number of potential GTG issues has been a workload concern.
- Licensee use of testing to demonstrate operable/functional.
- Licensee exercising option to NOT share PSA & other information
- External events, specifically fire modeling.
February 20-21, 2024 26th PSAEA Technical Meeting 18
SDP Results and Trends
SDP Data February 20-21, 2024 26th PSAEA Technical Meeting 20
SDP by Cornerstone Historical since Reactor Oversight Program inception February 20-21, 2024 26th PSAEA Technical Meeting 21
SDP by Cornerstone Historical over the last 5 years February 20-21, 2024 26th PSAEA Technical Meeting 22
2023 Emergency Diesel Generator SDP Issues Plant/Description LER/IR Event Date Exposure Time CCDP/
CDP Summer, Failure to Correct a Condition Adverse to Quality Results in an EDG Failure (ML23342A000) 05000395/2023002 (No LER issued) 11/2/22 162 days White Finding Farley 1, Failure to Identify Nonconforming Work Instructions Results in EDG Lube Oil Leak (ML23263B166) 348-23-001 2/26/23 115 days White Finding Calvert Cliffs 1, Failure to Establish and Implement Adequate Maintenance Practices Contributes to the Failure of EDG 1A (ML23297A192) 05000395/2023050 (No LER issued) 4/24/23 216 days White Finding Calvert Cliffs 1, Failure to Establish and Implement Adequate Maintenance Practices Contributes to the Failure of EDG 1A (ML23297A192) 05000395/2023050 (No LER issued) 4/24/23 196 days 2x10-6 February 20-21, 2024 26th PSAEA Technical Meeting 23
Emergency Diesel Generator Trend February 20-21, 2024 26th PSAEA Technical Meeting 24
Change Items Directly Related to the Significance Determination Process
Changes with Direct Impacts to the SDP Revision to IMC-0308, Appendix J and IMC-0609, Appendix J - Steam Generator Tube Integrity - updates in process 10CFR 50.69 - Risk Categorization of SSCs -
how licensees implement the process MAY affect how we inspect under IP-37060 and other Baseline Inspection Procedures Baseline Security Working Group - revisions to make the process more consistent & risk-informed IMC-0609, Appendix M Significance Determination Process Using Qualitative Criteria to be revised this year February 20-21, 2024 26th PSAEA Technical Meeting 26
Change Items with Indirect Impacts to the Significance Determination Process
These items are related to SDP but are more tangential PRA (or PSA) Configuration Control -
Operational Experience Smart Sample has been developed - roll out in CY2024 Shutdown generic models for four different plant designs are in the works Group for Risk Evaluation and Assessment Tools Review (GREATR) - formerly known as Risk Users Group IDHEAS-ECA - new human error probability tool A Performance Monitoring Strategy to Enhance Consistency in Risk-Informed Decision Making -
i.a.w., a Noise Audit February 20-21, 2024 26th PSAEA Technical Meeting 28
Some More Detail about GREATR Group for Risk Evaluation and Assessment Tools Review (GREATR) - formerly known as Risk Users Group
- Completed: update of SPAR models with 3 changes related to FLEX equipment: 1) PWROG data update, 2) turning off N+1 equipment nominally, and 3) truncating mission times for ELAP from 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> to 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />
- Current work: two working groups on EDG/Offsite Power Recovery and Event Assessment under various processes (e.g., SDP vs. MD8.3)
- Upcoming work: complete update to RASP Manual Volume 1, Internal Events in CY 2024 Question: is there a community of practice for updating your models, procedures, guidance, etc.
beyond what is strictly required in the ASME/ANS PRA standard?
February 20-21, 2024 26th PSAEA Technical Meeting 29
30 Inspiration for a Noise Audit in NRC Decision Making Inspired by the book Noise by Daniel Kahneman, et. al.
Some professional areas which are noisy:
- Medicine, e.g., reading of chest X-rays
- Child custody decisions, - who goes into foster care?
- Forensic science - do these fingerprints match?
- Patent decisions - should I grant this patent or not?
- Criminal sentencing - if 50% of robbery convicts receive 5 years
& 50% receive 15 years, is a 10-year average reflective of good decision making?
If these areas are noisy given smart, well educated, experienced professionals, then how are we doing?
Preliminary Indications of (Unacceptable?)
Variability in NRC Decision Making Some hesitancy in our organization to review past decisions, including high risk ones Evidence from some limited scope self-assessments that there is noise in NRC decision making on Event/Condition assessments
- Region 1 Audit of prior MD8.3 decisions
- Cross regional Audit of Losses of Offsite Power, specifically whether the MD8.3 process was entered & whether Baseline or Special Inspections were launched Some data points where NRC response vs. risk/safety significance was not proportional 31
NRC responds to issue?
Risk +/or safety significance is high?
No Yes No Yes Proportional response Over reaction Proportional response Under reaction Pascals Wager - A Modified Approach
NRC responds to issue?
No Yes No Yes 1E-6 1E-8 1E-4 1E-2 SIT for reactivity control issues Grid related LOOP &
challenge to SW cooling, BI & EP In the majority of cases, we SHOULD be landing here, b/c risk is often low But if not, we SHOULD be landing in this quadrant Focused Baseline or nothing additional Special Inspection Augmented Inspection Incident Investigation 1
2 3
4 Risk +/or safety significance is high?
Pascals Wager - A Modified Approach
34 Differences Between Noise & Bias Noisy decisions are represented by B and biased decisions by C Other interesting aspects:
-If you turn the bulls eye targets at left around, you can still determine when noise is unacceptably high, but
-we are fortunate in the nuclear industry to have objective measures (PRA) that tells us where the center of the target is.
-This FFR project should be able to measure: 1) within person variability, 2) cross-person noise, and 3) variability across regions and possibly others.