A human reliability analysis method based on cognitive process model for risk assessment

Abstract

This paper presents a human reliability analysis method based on cognitive process model. Human cognitive processes are modelled in different forms under different risk scenarios respectively. The cognitive process model consists of an event tree based on human cognitive process and Bayesian networks modelling the cognitive phases. The Bayesian networks (combining with evidence theory) use hierarchical structure and take organizational factors, situational factors and individual factors into consideration in a more explicit way than the traditional methods do. This paper also discusses the applicability of this method for the risk assessment of on-board ATP (Automatic Train Protection) of high-speed railway. With the help of this human reliability analysis method, the human reliability of the train driver and accident rate can be calculated accurately. The weakest cognitive phase in particular cognitive process can be identified and enhanced by improving the key organizational factors.