Probabilistic network-based approach to infrastructure safety assessment with human factor consideration

Abstract

The combinations of low probability events (hardware and software faults, anomalous nature events, human operator errors) cause the infrastructure accidents and disruptions. There are different approaches for evaluation of human operator's reliability. Multi-factor analysis is the essential step for obtaining the trustworthiness estimations of infrastructure's safety. The application of Bayesian Belief Networks (BBN) as a basis of multi-factor safety analysis is suggested in the paper. Two approaches for integration of probabilistic estimations in different qualimetric scales are proposed. The example of using of BBN for assessment of human factor in NPP Fukushima-1 disaster is considered.