ML19212A732
| ML19212A732 | |
| Person / Time | |
|---|---|
| Issue date: | 07/31/2019 |
| From: | Office of Nuclear Regulatory Research |
| To: | |
| D. Stroup 415-1649 | |
| Shared Package | |
| ML19212A718 | List: |
| References | |
| Download: ML19212A732 (61) | |
Text
NRC-RES/EPRI FIRE PRA METHODOLOGY Task 12 - Fire HRA EPRI Approach to Detailed Fire HRA Quantification NRC-RES Fire PRA Workshop Module IV August 5-9, 2019 Rockville, MD
Course Overview
- 1. Introduction to HRA
- 2. Overview of the EPRI/NRC Fire HRA Guidelines
- 3. Identification and definition of fire human failure events
- 4. Qualitative analysis
- 5. Fire HRA Application Experience
- 6. Quantitative analysis a) Screening b) Scoping c) Detailed EPRI approach & ATHEANA (detailed)
- 7. Recovery analysis
- 8. Dependency analysis
- 9. Uncertainty analysis Task 12: Fire HRA - EPRI Detailed Analysis Slide 2 Fire PRA Workshop, 2019, Rockville, MD
Fire HRA Module Training Objectives 1:Be able to name the steps in the process for conducting a Fire HRA.
2:Be able to list the different categories of Fire HRA human failure events.
3:Demonstrate knowledge of ASME/ANS PRA Standard high level requirements (HLRs).
4:Be able to identify context and performance shaping factors used in the analysis of post-fire human failure events.
5:Be able to list the quantification methods available for HEPs.
6:Understand the concept and importance of addressing dependencies between post-fire HRA events.
Task 12: Fire HRA - EPRI Detailed Analysis Slide 3 Fire PRA Workshop, 2019, Rockville, MD
Outline of the EPRI Approach to Detailed Fire HRA Module Introduction/Relation to NUREG/CR-6850 (EPRI 1011989) Tasks Applicable PRA Standard High Level Requirements Overview of Quantitative Methods in the EPRI Approach:
- Cause-Based Decision Tree Overview (Cognitive)
- HCR/ORE Overview (Cognitive for Time-Critical)
- THERP (Execution)
Definition & subsequent Qualitative Analysis
- Fire Context
- Performance Shaping Factor Method Selection & Quantification Summary Task 12: Fire HRA - EPRI Detailed Analysis Slide 4 Fire PRA Workshop, 2019, Rockville, MD
What is Detailed Fire HRA?
Consists of HRA tasks that develop human error probabilities (HEPs) for the modeled human failure events (HFEs)
- HEP used in Fire PRA quantification
- HEP development provides qualitative and quantitative insights Typically done to PRA Standard Capability Category II
- Risk-significant scenarios Uses same steps in the HRA Process:
- 1. Identification & definition of HFE
- 2. Qualitative analysis - context & performance shaping factors
- 3. Quantitative analysis - method selection & quantification of HEP a) Screening b) Scoping c) Detailed HRA: EPRI approach or ATHEANA
- 4. Provides input to subsequent Fire HRA tasks Dependency analysis Uncertainty analysis Task 12: Fire HRA - EPRI Detailed Analysis Slide 5 Fire PRA Workshop, 2019, Rockville, MD
General Approaches to Quantification
- 1. Screening: Slightly modified from NUREG/CR-6850 (EPRI 1011989) to reduce the HEPs for late HFEs (after fire is out) -
covered previously
- 2. Scoping Fire HRA quantification approach - covered previously
- Less conservative than screening, but designed to be slightly more conservative than detailed approaches
- Some actions may not be able to meet some of the criteria (result in an HEP of 1.0)
- 3. Two detailed fire HRA quantification approaches, modified for application in fire scenarios
- EPRI - covered in this module Cause-Based Decision Tree (CBDT) & HCR/ORE; THERP
- ATHEANA - covered after this module Task 12: Fire HRA - EPRI Detailed Analysis Slide 6 Fire PRA Workshop, 2019, Rockville, MD
Fire HRA Process Steps NUREG/CR-6850 Task Fire HRA Process Step Task 2 - Component Selection Identification of previously existing HFEs & potential response to spurious actuations/indications Task 5 - Fire-Induced Risk Identification and Definition of fire Model response HFEs Task 12 - Fire HRA Qualitative Analysis - definition, context & performance shaping factors Task 7 - First/Screening Quant. Quantification -
typically screening or scoping Task 8 - Scoping Quantification Quantification -
typically scoping Tasks 11/14 - Detailed Quantification & Dependency Scenario Quantification could be screening, scoping or detailed HRA Task 15 - Uncertainty Uncertainty Task 12: Fire HRA - EPRI Detailed Analysis Slide 7 Fire PRA Workshop, 2019, Rockville, MD
Relationship of Detailed Fire HRA to Fire PRA Tasks*
Detailed Fire HRA supports Fire PRA quantification
- Developed, and typically used, for detailed fire scenarios Detailed Fire Scenarios (Tasks 11 & 14)
Uncertainty/Sensitivity (Task 15)
- But can be used at any level, such as:
Screening / First Quantification (Task 7)
Scoping (Task 8)
Detailed Fire HRA uses inputs from most, prior Fire PRA tasks
- Identification & Definition of HFEs (Tasks 2, 5, 7 & 8)
- Qualitative Analysis (Task 12 - Fire HRA)
- All task numbers refer to NUREG/CR-6850; EPRI 1011989 Task 12: Fire HRA - EPRI Detailed Analysis Slide 8 Fire PRA Workshop, 2019, Rockville, MD
PRA Standard Requirements for HRA Quantification Relevant HLRs from Internal-Events Section (Ch. 2)
HLR-HR-G (from the internal events HRA element)
The assessment of the probabilities of the post-initiator HFEs shall be performed using a well-defined and self consistent process that addresses the plant-specific and scenario-specific influences on human performances, and addresses potential dependencies between human failure events in the same accident sequence Relevant HLRs from Fire Section (Ch. 4 of Standard)
HLR-HRA-C (from the Fire HRA element)
The Fire PRA shall quantify HEPs associated with incorrect responses accounting for the plant-specific and scenario-specific influences on human performance, particularly including the effects of fire Task 12: Fire HRA - EPRI Detailed Analysis Slide 9 Fire PRA Workshop, 2019, Rockville, MD
EPRI Quantification Methods CBDTM (Cause Based Decision Tree Method)
- 8 Decision trees based on simulator experiment insights
- Default method for cognitive portion (detection/diagnosis)
HCR/ORE Correlation (Human Cognitive Reliability /
Operator Reliability Experiment)
- Used for time-critical operator actions
- Normalized time reliability correlation (function of Tavailable / Trequired)
THERP (NUREG/CR-1278) for execution Methods are implemented in EPRI HRA Calculator software, but can be quantified on paper Task 12: Fire HRA - EPRI Detailed Analysis Slide 10 Fire PRA Workshop, 2019, Rockville, MD
Post-Initiator HFE Representation:
EPRI TR-100259 Pe = Execution is quantified using THERP Pc = Cognitive is quantified using CBDTM (default)
HCR/ORE (time critical HFEs)
Task 12: Fire HRA - EPRI Detailed Analysis Slide 11 Fire PRA Workshop, 2019, Rockville, MD
EPRI Timeline for a Post-initiator HFE Tsw T0 = Start Time usually the initiating event Tavail Tsw = System time window Treqd Tdelay = Time from start until cue Tdelay Texe = Execution time (includes transit, Tcog tools, PPE & component manipulation)
Texe Tcog = Cognition time (consists of detection, diagnosis, & decision-T0 Cue Crew Action Action no making) received diagnosis complete longer Start complete beneficial Tavail = Time available = (Tsw - Tdelay)
Treqd = Time required for response = (Tcog + Texe )
Tdelay = Time from start of transient until cue is reached Texe = Execution time (expansion of EPRI Tm component manipulation for fire)
Tcog = Cognition time (when HCR/ORE is used = T1/2 )
Task 12: Fire HRA - EPRI Detailed Analysis Slide 12 Fire PRA Workshop, 2019, Rockville, MD
CBDTM Overview - Cognitive Method Analytical approach based on identification of failure mechanisms and compensating factors Applicable to rule-based behavior, such as when procedures are used Two high-level failure modes:
- Plant information-operator interface failure
- Operator-procedure interface failure Each failure mode is decomposed into contributions from several distinct failure mechanisms Default method, especially if not time-critical Task 12: Fire HRA - EPRI Detailed Analysis Slide 13 Fire PRA Workshop, 2019, Rockville, MD
CBDT - Summary of Failure Mechanisms Type Designator Description Failures in pc a Data not available the Operator- pc b Data not attended to Information pc c Data misread or miscommunicated Interface pc d Information misleading Failures in pc e Relevant step in procedure missed the Operator- pc f Misinterpret instruction Procedure pc g Error in interpreting logic Interface pc h Deliberate violation (not sabotage)
Task 12: Fire HRA - EPRI Detailed Analysis Slide 14 Fire PRA Workshop, 2019, Rockville, MD
CBDTM decision tree:
pc-a Data not available Warning or Indication Indication Alternative Training on pc a Available in Accurate in Indication CR Procedure (a) neg.
(b) neg.
(c) neg.
(d) 1.5E-03 Yes (e) 5.0E-02 No (f) 5.0E-01 (g)
CBDTM decision tree: pc-a Data not available Specific branch guidance Task 12: Fire HRA - EPRI Detailed Analysis Slide 16 Fire PRA Workshop, 2019, Rockville, MD
CBDTM decision tree:
pc-b Data not attended to pcb Low vs. high Check vs. Front vs. back Alarmed vs. Nominal workload monitor panel not alarmed probability Front Check (a) neg.
Alarmed Back (b) 1.5E-4 Low Not alarmed (c) 3.0E-3 Alarmed Front (d) 1.5E-4 Yes Not alarmed Monitor (e) 3.0E-3 No Alarmed Back (f) 3.0E-4 Not alarmed (g) 6.0E-3 Alarmed Front (h) neg.
Not alarmed Check (i) neg.
Alarmed Back (j) 7.5E-4 Not alarmed High (k) 1.5E-2 Alarmed (l) 7.5E-4 Front Not alarmed (m) 1.5E-2 Monitor Alarmed (n) 1.5E-3 Back Not alarmed (o) 3.0E-2 Task 12: Fire HRA - EPRI Detailed Analysis Slide 17 Fire PRA Workshop, 2019, Rockville, MD
CBDTM decision tree:
pc-c Data misread or miscommunicated pcc Indicator easy Good/bad Formal com- Nominal to locate indicator munications probability (a) neg.
(b) 3.0E-3 (c) 1.0E-3 Yes (d) 4.0E-3 No (e) 3.0E-3 (f) 6.0E-3 (g) 4.0E-3 (h) 7.0E-3 Task 12: Fire HRA - EPRI Detailed Analysis Slide 18 Fire PRA Workshop, 2019, Rockville, MD
CBDTM decision tree:
pc-e Relevant step in procedure missed pce Obvious vs. Single vs. Graphically Placekeeping Nominal hidden multiple distinct aids probability (a) 1.0E-3 Single (b) 3.0E-3 Obvious (c) 3.0E-3 (d) 1.0E-2 (e) 2.0E-3 Multiple (f) 4.0E-3 (g) 6.0E-3 Yes (h) 1.3E-2 No Hidden (i) 1.0E-1 Task 12: Fire HRA - EPRI Detailed Analysis Slide 19 Fire PRA Workshop, 2019, Rockville, MD
CBDTM decision tree:
pc-d Information misleading pcd All cues as Warning of Specific General Nominal stated differences training training probability Yes (a) neg.
No (b) 3.0E-3 (c) 1.0E-2 (d) 1.0E-1 (e) 1.0 Task 12: Fire HRA - EPRI Detailed Analysis Slide 20 Fire PRA Workshop, 2019, Rockville, MD
CBDTM decision tree:
pc-f Misinterpret instruction Standard, pcf unambiguous All required Training on Nominal wording information step probability (a) neg.
(b) 3.0E-3 (c) 3.0E-2 Yes (d) 3.0E-3 No (e) 3.0E-2 (f) 6.0E-3 (g) 6.0E-2 Task 12: Fire HRA - EPRI Detailed Analysis Slide 21 Fire PRA Workshop, 2019, Rockville, MD
CBDTM decision tree:
pc-g Error in interpreting logic pcg Not And or or Both and Practiced Nominal statement statement and or scenario probability (a) 1.6E-2 (b) 4.9E-2 (c) 6.0E-3 (d) 1.9E-2 (e) 2.0E-3 Yes (f) 6.0E-3 (g) 1.0E-2 No (h) 3.1E-2 (i) 3.0E-4 (j) 1.0E-3 (k) neg.
(l) neg.
Task 12: Fire HRA - EPRI Detailed Analysis Slide 22 Fire PRA Workshop, 2019, Rockville, MD
CBDTM decision tree:
pc-h Deliberate violation Belief in Adverse Policy of adequacy of consequence Reasonable verbatim Nominal pch instruction if comply alternative compliance probability Yes (a) neg.
(b) 5.0E-1 No (c) 1.0 (d) neg.
(e) neg.
Task 12: Fire HRA - EPRI Detailed Analysis Slide 23 Fire PRA Workshop, 2019, Rockville, MD
CBDTM - Recovery Factors Recovery Factor Time Effective Self Review At any time there is a subsequent cue, other than the initial cue that would prompt the operator to revisit the decision OR Is there a procedural step that either returns the operator to the initial step where the error was made, or that repeats the initial instruction?
Other (Extra) Crew At any time that there are crew members over and above the minimum complement present in the CR and not assigned to other tasks Shift Technical 10 to 15 minutes after reactor trip.
Advisor Emergency 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> after reactor trip - if constituted Response Facility/
Technical Support Center Shift Change 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> after reactor trip given 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> shifts 9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br /> after reactor trip given 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> shifts Task 12: Fire HRA - EPRI Detailed Analysis Slide 24 Fire PRA Workshop, 2019, Rockville, MD
Post-Initiators: CBDTM Recovery Factors With an HFE Self- Extra STA Shift ERF Tree Branch Review Crew Review Change Review Pca all NC 0.5 NC 0.5 0.5 Pcb all X NC X X X Pcc all NC NC X X X Pcd all NC 0.5 X X 0.1 Pce a-h X 0.5 NC X X Pce i 0.5 0.5 X X X Pcf all NC 0.5 X X X Pcg all NC 0.5 X X X Pch all NC X X NC NC NC - No credit X - Apply recovery credit following THERP dependency Task 12: Fire HRA - EPRI Detailed Analysis Slide 25 Fire PRA Workshop, 2019, Rockville, MD
Dependency Analysis THERP Dependency Formulas Dependence Approximate Value Equation Level for HEP < 0.01 Zero (ZD) HEP HEP Low (LD) (1+19 X HEP) / 20 0.05 Medium (MD) (1+ 6 X HEP) / 7 0.14 High (HD) (1 + HEP) / 2 0.5 Complete (CD) 1.0 1.0 Task 12: Fire HRA - EPRI Detailed Analysis Slide 26 Fire PRA Workshop, 2019, Rockville, MD
HCR/ORE Overview - Cognitive Method Cognitive modeling of time-critical operator actions
- For example, less than 30 minute time window Empirical method, a time-reliability curve Fitted to successful response times Data points in which crews were totally on the wrong path not included in the fitting (outliers)
Pc therefore conditional on a correct decision, or the initial error was discovered in a timely manner Normalized time to be limited to time windows on which observations were made. Extrapolation not valid Guidance in EPRI-TR100259:
- If Pc < 1E-02, use the CBDTM
- If Pc believed to be conservative, use CBDTM Task 12: Fire HRA - EPRI Detailed Analysis Slide 27 Fire PRA Workshop, 2019, Rockville, MD
HCR/ORE - Equation TW ln( T )
PC = 1 1/ 2 PC = Probability of cognitive non-response
= Logarithmic standard deviation (Determined based on cue response structure - next slide)
= Standard normal cumulative distribution TW = TSW - Tdelay - TM = time window available for cognitive response T1/2 = Crew median response time Task 12: Fire HRA - EPRI Detailed Analysis Slide 28 Fire PRA Workshop, 2019, Rockville, MD
HCR/ORE - Sigma Values based on cue-response structure Values for Cue-Plant Response Type Structure Average Upper Lower Bound Bound BWRs CP1 0.70 1.00 0.40 CP2 0.58 0.96 0.20 CP3 0.75 0.91 0.59 PWRs CP1 0.57 0.88 0.26 CP2 0.38 0.69 0.07 CP3 0.77 *
Categorization of Type CP Actions TSW Tdelay T1/2 TM CP1 First Cue Execution Starts Execution Ends Undesired Consequence IF t=0 time TSW WHEN Tdelay T1/2 TM CP2 Execution Execution Undesired First Cue Second Cue Starts Ends Consequence t=0 time TSW Tdelay T1/2 TM CP3 Execution Execution Undesired BEFORE First Cue Second Cue Starts Ends End State t=0 time Task 12: Fire HRA - EPRI Detailed Analysis Slide 30 Fire PRA Workshop, 2019, Rockville, MD
THERP: Technique for Human Error Rate Prediction (NUREG/CR-1278, 1983)
This is the most extensively documented and the most widely used (and misused) HRA technique. The handbook has four main sections:
- Basic concepts.
- Method for analysis and quantification of human performance.
- Human performance models and HEPs.
- Tables of HEPs and examples.
Simplified version developed as Accident Sequence Evaluation Program Human Reliability Analysis Procedure in NUREG/CR-4772, 1987
- Referred to as ASEP For fire HRA THERP is used to model the execution HEP Task 12: Fire HRA - EPRI Detailed Analysis Slide 31 Fire PRA Workshop, 2019, Rockville, MD
THERP Overview Applicable to pre- and post-Initiator HFEs Execution modeled as task analysis
- Tasks reviewed to identify critical steps
- Each critical step has two failure modes Error of omission Error/s of commission HFE can be represented in a HRA event tree
- Branch probabilities from Chapter 20 tables
- Performance shaping factors applied
- Recovery and dependencies are addressed.
Task 12: Fire HRA - EPRI Detailed Analysis Slide 32 Fire PRA Workshop, 2019, Rockville, MD
THERP Execution Analysis The THERP approach develops a functional logic model of the human interaction execution failure by reviewing the procedure to identify:
- Critical steps, which are those steps that if carried out incorrectly would fail all or part of the function that is to be achieved
- Recovery steps, which are those steps, which can recover previous, failed critical steps primarily through revisitation
- Alternative (redundant) steps, which are steps along an alternate success path functionally in parallel with the critical steps Task 12: Fire HRA - EPRI Detailed Analysis Slide 33 Fire PRA Workshop, 2019, Rockville, MD
THERP Execution Errors Execution Error Omission Commission Task 12: Fire HRA - EPRI Detailed Analysis Slide 34 Fire PRA Workshop, 2019, Rockville, MD
THERP Execution Errors: Omission Procedural Items (Table 20-7)
Stress (Table 20-16)
HEP for EOM Task 12: Fire HRA - EPRI Detailed Analysis Slide 35 Fire PRA Workshop, 2019, Rockville, MD
THERP Execution Errors: Commission Display Selection (Table 20-9)
Interactions with Displays Read/Record Quantitative Data (Table 20-10)
Check/Read Qualitative (Table 20-11)
Task 12: Fire HRA - EPRI Detailed Analysis Slide 36 Fire PRA Workshop, 2019, Rockville, MD
THERP Execution Errors: Commission Manipulation of manual controls in Control Room Control/Selection/
Use (Table 20-12)
Task 12: Fire HRA - EPRI Detailed Analysis Slide 37 Fire PRA Workshop, 2019, Rockville, MD
THERP Execution Errors: Commission Valve Selection Local (Table 20-13) manipulation of manual valves Stuck Valve Detection (Table 20-14)
Task 12: Fire HRA - EPRI Detailed Analysis Slide 38 Fire PRA Workshop, 2019, Rockville, MD
THERP Execution Errors: Total HEP and Recovery HEP for HEP for EOM EOC Total HEP (unrecovered)
Recovery Task 12: Fire HRA - EPRI Detailed Analysis Slide 39 Fire PRA Workshop, 2019, Rockville, MD
THERP Task Analysis: Identification of procedural steps (PWR feed and bleed example)
- 12. Actuate SI Critical
- 13. Verify RCS Feed Path Recovery
- 14. Reset SI Not critical
- 15. RESET Containment Isolation Phase A and B Not critical
- 16. ESTABLISH Instrument Air to Containment Not critical
- 17. ESTABLISH RCS Bleed Path:
17.a Check Power to Pressurizer PORV Block Valves Not critical Available 17.b Verify Prsszr PORV Block Valves OPEN Not critical 17.c Open all Pressurizer PORVs Critical
- 18. VERIFY Adequate RCS Bleed Path Recovery Task 12: Fire HRA - EPRI Detailed Analysis Slide 40 Fire PRA Workshop, 2019, Rockville, MD
THERP - Logic model Fail to actuate SI AND Fail to verify RCS feed path OR Fail to open all PORVs AND Fail to verify RCS bleed path Task 12: Fire HRA - EPRI Detailed Analysis Slide 41 Fire PRA Workshop, 2019, Rockville, MD
THERP - Quantification Each critical step has two basic failure modes:
- Error of omission (EOM)
- Error of commission (EOC)
The various tables in Chapter 20 of THERP are used in determining the HEPs for each failure mode Performance shaping factors can increase the HEPs Dependencies impact recovery Task 12: Fire HRA - EPRI Detailed Analysis Slide 42 Fire PRA Workshop, 2019, Rockville, MD
THERP - PSFs and Stress The ASME Standard requires the evaluation of PSFs:
Quality [type (classroom or simulator) and frequency] of operator training or experience Quality of the written procedures and administrative controls Availability of instrumentation needed to take corrective actions Degree of clarity of the cues/indications Human-machine interface Time available and time required to complete the response Complexity of the required response Environment (e.g., lighting, heat, radiation) under which the operator is working Accessibility of the equipment requiring manipulation Necessity, adequacy, and availability of special tools, parts, clothing, etc.
Task 12: Fire HRA - EPRI Detailed Analysis Slide 43 Fire PRA Workshop, 2019, Rockville, MD
THERP - PSFs and Stress PSFs are not necessarily independent As the only vehicle to adjust nominal HEPs in the THERP methodology is the stress multiplier, all PSFs are considered to culminate in the stress level Therefore, stress decision tree developed for HRA Calculator Task 12: Fire HRA - EPRI Detailed Analysis Slide 44 Fire PRA Workshop, 2019, Rockville, MD
THERP - Stress Decision Tree Plant Response as Workload Execution PSFs Stress Expected Optimal Low Low Negative Moderate Yes Optimal Moderate High Negative High No High Task 12: Fire HRA - EPRI Detailed Analysis Slide 45 Fire PRA Workshop, 2019, Rockville, MD
THERP - Stress - Plant Response as Expected Yes
- Applies to tasks directed by the emergency operating procedures (EOPs), or in the initial parts of the Functional Restoration Procedures (FRPs) or Emergency Contingency Actions (ECAs). In these scenarios, the procedures are apparently working and the operators feel in control.
No
- Applies to tasks directed later in the FRPs or ECAs. In these scenarios the last procedural options are attempted before entry into the SAMGs would be required. The response is not as expected as the procedures are apparently not leading to success due to e.g. multiple equipment failures Task 12: Fire HRA - EPRI Detailed Analysis Slide 46 Fire PRA Workshop, 2019, Rockville, MD
THERP - PSFs and Stress The nominal THERP HEPs for step-by-step procedural actions should be adjusted for the impact of stress per THERP Table 20-16:
Table 20-16 Modifiers Table 20-16 Table 20-16 HRA Calculator Item Stress Level Stress Level Skilled Novice (2) Optimum Low 1 1 Moderately (4) Moderate 2 4 High (6) Extremely High High 5 10 No unique fire specific stress multipliers Task 12: Fire HRA - EPRI Detailed Analysis Slide 47 Fire PRA Workshop, 2019, Rockville, MD
Quantification:
Fire HEPs for HFEs from the Internal Events PRA If HFE has been quantified using EPRI HRA Approach for internal events, quantification for fire is a relatively simple modification in following areas:
- Timing
- Cue and indications impacts
- Increase in stress
- Increase in workload
- Use of multiple procedures
- For local actions, consider alternate routes if fire impacts the normal or ideal travel path Task 12: Fire HRA - EPRI Detailed Analysis Slide 48 Fire PRA Workshop, 2019, Rockville, MD
Fire Impacts on Timing Tsw Tavail Treqd Tdelay Tcog Texe T0 Cue Crew Action Action no Initiating received diagnosis complete longer Event complete beneficial T = 0 is considered the start of the fire - For existing HFEs T=0 is typically reactor trip. In most cases, the FPRA assumes the fire and reactor trip coincide.
Tdelay = Time from start of transient until cue is reached. If the cue is considered to be procedure step the fire may cause delays in the procedure implementation.
Tcog = If the fire impacts some but not all of the instrumentation Tcog will be increased from the internal events case to account for the time required for the operators to asses the situation &
determine which instrumentation is correct or diagnose based on secondary cues.
Texe = For main control room actions in which there is no fire in the control room, Tm is considered to be the same for the internal events case and the fire case.
For local actions, Tm will account for any detours caused by the fire. Tm must also account for PPE & tools.
Task 12: Fire HRA - EPRI Detailed Analysis Slide 49 Fire PRA Workshop, 2019, Rockville, MD
Fire Impacts on Timing (contd)
If time available for recovery is reduced due to fire impacts on timing, then the recoveries previously credited in the internal events PRA within the CBDTM are to be revisited If time-critical action and cues/indications are impacted, then consider using upper bound for sigma when applying HCR/ORE Task 12: Fire HRA - EPRI Detailed Analysis Slide 50 Fire PRA Workshop, 2019, Rockville, MD
Fire Impacts on Instrumentation If all instrumentation is impacted and there are no cues for diagnosis then HEP =1.0 Partial instrumentation impacted is modeled in decision tree Pc-a & Pc-d (HEP range 1E-2 to 1.0)
If the fire causes no impact on instrumentation then Pc-a and Pc-d typically evaluate to Negligible Task 12: Fire HRA - EPRI Detailed Analysis Slide 51 Fire PRA Workshop, 2019, Rockville, MD
CBDT Example - Fire Impacts on Workload (Pc)
Increased workload:
- modeled explicitly
- decision tree Pc-b
- if fire causes increase in workload
- select High workload
- part of the cognitive phase (detection & diagnosis)
- potentially recover if have additional staff Task 12: Fire HRA - EPRI Detailed Analysis Slide 52 Fire PRA Workshop, 2019, Rockville, MD
CBDT Example - Fire Impacts on Workload (Pe)
Increase in workload is reflected by an increase in stress Task 12: Fire HRA - EPRI Detailed Analysis Slide 53 Fire PRA Workshop, 2019, Rockville, MD
Fire Impacts on Procedure Usage If EOPs are implemented in parallel to fire procedures, then multiple procedures are used If EOPs are suspended while fire procedures are being used, then only one procedure is credited and any time delays are accounted for in the timeline Task 12: Fire HRA - EPRI Detailed Analysis Slide 54 Fire PRA Workshop, 2019, Rockville, MD
Fire Impacts on Execution Stress is often increased from internal events case
- Except for control room actions when operator actions occurring more than 70 minutes after the fire started, because
- 1. 99% of fires are extinguished within 70 minutes per NUREG/CR-6850 Suppl. 1 (EPRI 1019259, Sept 2010)
- 2. On average, a fire is extinguished in 13 minutes For local actions, an additional factor can be applied
- Account for smoke, communication impacts, or
- Additional equipment required by fire Examples: SCBA, ladders, keys, tools Task 12: Fire HRA - EPRI Detailed Analysis Slide 55 Fire PRA Workshop, 2019, Rockville, MD
Fire Response HFEs Method selection depends on timing
- CBDT approach to quantification applied first
- HCR/ORE for time critical fire response actions May use upper bound based on sigma value Ex-control room actions required due to loss of control are not substantially different from other local actions (e.g.,
during SBO) provided that local actions are not credited in close proximity to fire location No separate guidance for MCR abandonment
- MCR typically is completely abandoned due to uninhabitability, not due to loss of control/functionality initial results show that frequency is low enough to not be a concern If required, additional decision trees may be developed to model locus of control moving outside the control room Task 12: Fire HRA - EPRI Detailed Analysis Slide 56 Fire PRA Workshop, 2019, Rockville, MD
Fire Response HFEs Same considerations as internal events actions and the following additional considerations
- Ambiguously worded procedures: Fire procedures are not standardized like EOPs. Modeled in decision tree Pcf. For internal events HFEs Pcf typically evaluates to negligible.
- Local controls may not be as easily accessible and as well trained on as for internal events actions. In this case, higher Error of Omission is selected from THERP
- No base case from which to build the analysis, so entire analysis must be developed Task 12: Fire HRA - EPRI Detailed Analysis Slide 57 Fire PRA Workshop, 2019, Rockville, MD
Undesired Response to Spurious Indication or Actuation The following can be screened from consideration during identification:
-Actions for which multiple indications are available for different parameters or via redundant channels
-Actions that have a proceduralized verification step, if verification will be effective given the fire scenario Task 12: Fire HRA - EPRI Detailed Analysis Slide 58 Fire PRA Workshop, 2019, Rockville, MD
Quantification of Undesired Operator Responses to Spurious Signals HEPs for actions that do not screen from consideration are initially to be set to 1.0 (failed)
EPRI approach to quantification
-Assume the Error of Commission has occurred, then
-Identify, define and quantify a recovery action Task 12: Fire HRA - EPRI Detailed Analysis Slide 59 Fire PRA Workshop, 2019, Rockville, MD
Detailed Fire HRA Summary Consists of HRA tasks that develop human error probabilities (HEPs) for the modeled human failure events (HFEs)
- HEP used in Fire PRA quantification
- HEP development provides qualitative insights on results drivers Uses most of the steps in the HRA Process:
- 1. Identification and Definition of HFE
- 2. Qualitative analysis - context and performance shaping factors
- 3. Quantitative analysis - method selection and quantification of HEP a) Screening b) Scoping c) Detailed HRA a) EPRI approach (CBDTM or HCR/ORE & THERP) b) ATHEANA
- 4. Provides input to subsequent fire HRA tasks Dependency analysis Uncertainty analysis Task 12: Fire HRA - EPRI Detailed Analysis Slide 60 Fire PRA Workshop, 2019, Rockville, MD
Course Overview
- 1. Overview of the EPRI/NRC Fire HRA Guidelines
- 2. Identification and Definition of post-fire human failure events
- 3. Qualitative analysis
- 4. Quantitative analysis a) Screening b) Scoping c) EPRI approach (detailed) a) EPRI Examples (See handouts) d) ATHEANA (detailed)
- 5. Recovery analysis
- 6. Dependency analysis