Fire safety assessment using Bayesian causal network

摘要

Bayesian causal network consisting of chance, decision and utility nodes is used to analyze a fire protection system. Typical chance nodes of the network include fire start, detection, tampering, sprinklers, smoke detection, fire brigade, fire flashover, and structural collapse. The decision node and utility nodes are implemented to enable risk assessment under fire design situation. Appropriate causal links, input conditional probabilities and utility data characterize performance of the system. It appears that Bayesian causal networks provide an effective tool to analyze efficiency of fire protection systems as well as to make risk assessment related to fire design situation.