Testing Subtask Quantification Assumptions for Dynamic Human Reliability Analysis in the SPAR-H Method

Abstract

Human reliability analysis (HRA) to date has relied almost entirely on static methods. To provide a more precise model of human performance, dynamic HRA has been developed. Dynamic HRA must model a range of human actions typically at a finer resolution than is accounted for by overall tasks or human failure events (HFEs) used by static HRA. Parsing HFEs into finer units of analysis requires consideration of the mathematical underpinnings of the HRA methods that will be used for quantification. This paper serves to test how conventional HRA methods scale to this level of precision. SPAR-H, a static HRA method, was evaluated as part of this research. SPAR-H, which is based on HFEs, may require further refinement before its quantification approach can be employed in dynamic HRA.