Fire PRA Methodology

This wiki is intended to serve the needs of a fire risk analysis team by describing a structured framework for conduct of the overall analysis, as well as providing references to specific recommended practices to address each key aspect of the analysis. This information follows the layout of EPRI 1011989 / NUREG/CR-6850. The methodology described here addresses the processes for the performance of a Fire PRA that focuses on a Level 1 PRA (core damage frequency - CDF) with consideration of large early release frequency (LERF). The wiki pages for each fire PRA task describe the task overview (including objective, purpose, and scope), related element of the ASME/ANS PRA Standard, related appendices in EPRI 1011989 / NUREG/CR-6850, and any supplemental guidance. The two most recent versions of the PRA Standard include ASME/ANS RA-Sa-2009 and ASME/ANS RA-Sb-2013. ASME/ANS RA-Sa-2009 was endorsed by the NRC in RG 1.200 Revision 2.

Fire PRA Process Overview

The following considerations are important in the use of this figure.

• A Fire PRA is iterative. Certain tasks may need refinement after conduct of one or more of the subsequent tasks. It may also be appropriate to incorporate only limited detail in the first pass through an analysis task, deferring the pursuit of additional detail pending the results of a later task. For example, the number of components and circuits credited in Tasks 2 and 3 is likely to be revised after attempts at screening in Tasks 4 and 7. The flow chart does not attempt to incorporate potential feedback loops. Analyst judgment is needed to ensure that an appropriate overall analysis process is followed consistent with study objectives.

• Even though the process flow illustrated should work for predominant cases, users may find other analysis task sequences to be more appropriate for their objectives. Task sequence choices may, for example, be influenced by plant-specific fire protection features as well as the availability and depth of plant information supporting the Fire PRA. Each analysis task incorporates added detail into a given aspect of the Fire PRA. Task ordering is subject to practitioner judgment. Indeed, the task ordering applied to one set of fire scenarios may differ from that used in another set of scenarios within the same overall analysis.

The following is a short description of each technical task for the overall Fire PRA methodology.

Technical Tasks

Plant Boundary Definition and Partitioning (Task 1)

The first step in a Fire PRA is to define the physical boundary of the analysis, and to divide the area within that boundary into analysis compartments.

Fire PRA Component Selection (Task 2)

The selection of components that are to be credited for plant shutdown following a fire is a critical step in any Fire PRA. Components selected would generally include any and all components credited in the 10 CFR 50 Appendix R post-fire SSD analysis. Additional components will likely be selected, potentially including any and all components credited in the plant’s internal events PRA. Also, the proposed methodology would likely introduce components beyond either the 10 CFR 50 Appendix R list or the internal events PRA model. Such components are often of interest due to considerations of combined spurious operations that may threaten the credited functions and components.

Fire PRA Cable Selection (Task 3)

This task provides instructions and technical considerations associated with identifying cables supporting those components selected in Task 2. In previous Fire PRA methods (such as EPRI FIVE and EPRI's Fire PRA Implementation Guide [TR-105928, no longer available on epri.com]) this task was relegated to the SSD analysis and its associated databases. EPRI 1011989 / NUREG/CR-6850 offers a more structured set of rules for selection of cables.

Qualitative Screening (Task 4)

This task identifies fire analysis compartments that can be shown to have little or no risk significance without quantitative analysis. Fire compartments may be screened out if they contain no components or cables identified in Tasks 2 and 3, and if they cannot lead to a plant trip due to either plant procedures, an automatic trip signal, or technical specification requirements.

Plant Fire-Induced Risk Model (Task 5)

This task discusses steps for the development of a logic model that reflects plant response following a fire. Specific instructions have been provided for treatment of fire-specific procedures or preplans. These procedures may impact availability of functions and components, or include fire-specific operator actions (e.g., self-induced-station-blackout).

Fire Ignition Frequency (Task 6)

This task describes the approach to develop frequency estimates for fire compartments and scenarios. Significant changes from the EPRI FIVE method have been made in this task. The changes generally relate to use of challenging events, considerations associated with data quality, and increased use of a fully component based ignition frequency model (as opposed to the location/component-based model used, for example, in FIVE).

Quantitative Screening (Task 7)

A Fire PRA allows the screening of fire compartments and scenarios based on their contribution to fire risk. This approach considers the cumulative risk associated with the screened compartments (i.e., the ones not retained for detailed analysis) to ensure that a true estimate of fire risk profile (as opposed to vulnerability) is obtained.

Scoping Fire Modeling (Task 8)

This step provides simple rules to define and screen fire ignition sources (and therefore fire scenarios) in an unscreened fire compartment.

Detailed Circuit Failure Analysis (Task 9)

This task provides an approach and technical considerations for identifying how the failure of specific cables will impact the components included in the Fire PRA SSD plant response model.

Circuit Failure Mode Likelihood Analysis (Task 10)

This task considers the relative likelihood of various circuit failure modes. This added level of resolution may be a desired option for those fire scenarios that are significant contributors to the risk. The methodology provided in this document benefits from the knowledge gained from the tests performed in response to the circuit failure issue.

Detailed Fire Modeling (Task 11)

This task describes the method to examine the consequences of a fire. This includes consideration of scenarios involving single compartments, multiple fire compartments, and the main control room. Factors considered include initial fire characteristics, fire growth in a fire compartment or across fire compartments, detection and suppression, electrical raceway fire barrier systems, and damage from heat and smoke. Special consideration is given to turbine generator (T/G) fires, hydrogen fires, high-energy arcing faults, cable fires, and main control board (MCB) fires. There are considerable improvements in the method for this task over the EPRI FIVE and EPRI's Fire PRA Implementation Guide (TR-105928, no longer available on epri.com) in nearly all technical areas.

Post-Fire Human Reliability Analysis (Task 12)

This task considers operator actions for manipulation of plant components. The analysis task procedure provides structured instructions for identification and inclusion of these actions in the Fire PRA. The procedure also provides instructions for estimating screening human error probabilities (HEPs) before detailed fire modeling results (e.g., fire growth and damage behaviors) have been developed. Estimating HEP values with high confidence is critical to the effectiveness of screening in a Fire PRA. This report does not develop a detailed fire HRA methodology. There are a number of HRA methods that can be adopted for fire with appropriate additional instructions that superimpose fire effects on any of the existing HRA methods, such as SHARP, ATHEANA, etc. This would improve consistency across analyses i.e., fire and internal events PRA.

Seismic Fire Interactions (Task 13)

This task is a qualitative approach to help identify the risk from any potential interactions between an earthquake and fire.

Fire Risk Quantification (Task 14)

The task description provides recommendations for quantification and presentation of fire risk results.

Uncertainty and Sensitivity Analyses (Task 15)

This task describes the approach to follow for identifying and treating uncertainties throughout the Fire PRA process. The treatment may vary from quantitative estimation and propagation of uncertainties where possible (e.g., in fire frequency and non-suppression probability) to identification of sources without quantitative estimation, where knowledge of a quantitative treatment of uncertainties is beyond the state-of-the-art. The treatment may also include one-at-a-time variation of individual parameter values to determine the effect on the overall fire risk (sensitivity analysis).

Fire PRA Documentation (Task 16)

This provides suggestions for documenting a Fire PRA.

Technology Transfer

Reports documenting past fire PRA trainings (including power point slides and video recordings of past trainings) are available here.