Preliminary quantification for human reliability analysis

A comprehensive probabilistic risk study for a nuclear power plant includes an assessment of human reliability, in which large numbers of human interactions are identified in the fault trees and quantified. Many of these will not ultimately contribute significantly to risk, so some method of preliminary discrimination or screening is usually employed to reduce the number of interactions needing detailed analysis. The objective is to apply the often limited resources available for human reliability analysis in the most efficient way. This paper describes the approach to preliminary quantification developed for one particular risk study, the Darlington Probabilistic Safety Evaluation (DPSE), currently under way in Ontario Hydro