Welcome to the PSAM 18 Abstract Status page.
Abstract KG128Full Paper + Presentation
Completion of EPRI Credible Dependent Failure Mechanisms (CDFM) Methodology
In 2025, EPRI released a new HRA Dependency method called Credible Dependent Failure Mechanisms (CDFM). This work began in 2021, with a workshop that used guided brainstorming sessions aimed at identifying novel and viable alternatives to address the challenges with HRA dependency analysis. Building on the results of this initial workshop, EPRI developed CDFM which is a new methodology to HRA dependency focused on identifying those realistic (that is, credible) causes (failure mechanisms) of human actions that can fail multiple (dependent) accident mitigation responses. In 2025 the complete methodology was published and documented in EPRI Report 3002032023.
The CDFM process consists of seven steps and uses PRA scenarios (Event tree and Fault Tree logic) and available information related to existing HFEs to identify, define and quantify Dependent Combination Events (DCEs).
• Task 1: Identify Event Tree Functions (ETFs)
• Task 2: Define and assign ETF Dependent Combination Events (DCEs)
• Task 3: Identify and Define Support System DCEs
• Task 4: Quantify the DCE probabilities
• Task 5: Incorporate the DCEs into the PRA model
• Task 6: Review and finalize the PRA Model
• Task 7: Document the analysis
In 2025, EPRI hosted a hybrid workshop which introduced members to this new draft methodology and provided them with opportunities to apply selected portions of this approach to their models and provide feedback to EPRI. Following the workshop, the guidance was updated and documented in EPRI Report 3002032023.
The methodology includes two approaches to quantification. Approach 1 is the selection of the lowest cognitive HEP (Pcog) within the HFEs of the DCE under consideration. In this approach, it is important to ensure the individual HFE quantifications are capturing the same common failure mechanism. This approach is only applicable if all HFEs within the DCE are homogenous. When Approach 1 is not appropriate, Approach 2 will be required. Approach 2 uses the same methodology used to quantify individual HFEs in the DCE to evaluate the DCE probability. Expanded examples and guidance for these approaches can be found within the 2025 Report. This paper will focus on the new quantification approach developed in 2025.
← Check another abstractThe CDFM process consists of seven steps and uses PRA scenarios (Event tree and Fault Tree logic) and available information related to existing HFEs to identify, define and quantify Dependent Combination Events (DCEs).
• Task 1: Identify Event Tree Functions (ETFs)
• Task 2: Define and assign ETF Dependent Combination Events (DCEs)
• Task 3: Identify and Define Support System DCEs
• Task 4: Quantify the DCE probabilities
• Task 5: Incorporate the DCEs into the PRA model
• Task 6: Review and finalize the PRA Model
• Task 7: Document the analysis
In 2025, EPRI hosted a hybrid workshop which introduced members to this new draft methodology and provided them with opportunities to apply selected portions of this approach to their models and provide feedback to EPRI. Following the workshop, the guidance was updated and documented in EPRI Report 3002032023.
The methodology includes two approaches to quantification. Approach 1 is the selection of the lowest cognitive HEP (Pcog) within the HFEs of the DCE under consideration. In this approach, it is important to ensure the individual HFE quantifications are capturing the same common failure mechanism. This approach is only applicable if all HFEs within the DCE are homogenous. When Approach 1 is not appropriate, Approach 2 will be required. Approach 2 uses the same methodology used to quantify individual HFEs in the DCE to evaluate the DCE probability. Expanded examples and guidance for these approaches can be found within the 2025 Report. This paper will focus on the new quantification approach developed in 2025.