A model-based approach to HRA: Qualitative analysis methodology

This paper describes a model-based approach to the qualitative portion of Human Reliability Analysis (HRA). The goal of the qualitative analysis approach is twofold: 1) to incorporate salient information from the cognitive psychology literature into the analysis, and 2) to develop models and guidance to support analysis teams as they gather and organize the information needed for the follow-on quantitative portion of the HRA. A focus in the development has been to provide guidance and assistance for HRA analysts with a wide range of skill levels. This is because the growth in risk-informed applications has demanded that analysts who are not experts in cognitive science must use HRA methods to generate inputs to risk-informed decision-making. The qualitative analysis approach is also intended to be generic in the sense that it should be compatible with various quantification methods. Tools have been developed as a part of this approach to qualitative analysis, particularly the Crew Response Tree (CRT) and Fault Trees for causal delineation. The Crew Response Tree provides a structured way to identify, define, and decompose Human Failure Events in the HRA. Together with the CRT, the Fault Trees provide the structure needed to enhance consistency and traceability in the qualitative analysis. The Fault Trees are designed to guide the analyst in identifying the ways in which plant crews can fail. They represent a simplified model of human cognition in the nuclear power plant domain and have been explicitly linked to currently accepted psychological/cognitive models. This set of tools - the CRT and the Fault Trees - provides enhanced traceability of the HRA analysis since documentation is inherent in the tools.