Fault bayesian network based method for risk assessment of slope collapse accident

This paper is aimed at converting fault trees to fault Bayesian networks,and solving the main causes and probability of slope collapse accidents.For slope excavation and support system in which the causal relationship is uncertain,description with probability is more reasonable.The fault tree logic gates do not have this capability because it is described in the logic of certainty.Bayesian network is regarded as uncertain knowledge representation and reasoning of the most effective theoretical model.In this paper,referring to the qualitative analysis of slope collapse accident with Fault Tree Analysis(FTA) done by previous researchers,a Bayesian network conversion from FTA is carried out.Based on the fault tree of slope collapse,a corresponding fault Bayesian network model of the slope collapse is established.To establish Bayesian network,the causal relationship between all variables are analyzed based on their prior probability.When new evidence becomes available,the posterior probabilities of a set of variables of structure condition can be updated,which has practical value for evaluation of the structure.Applying the fault Bayesian network model,the probability of risk event of slope collapse are calculated.The basic events are rearranged based on the importance according to the importance analysis to find out the most influential potential factor for the occurrence of slope collapse accident.The result shows that the fault Bayesian network based method could obtain more additional information.Furthermore,the network can help to make account of the changing conditions of nodes induced by the variation of any other nodes of networks,which the fault tree approach failed to do so.Therefore,it can be said that the Bayesian network approach can be taken as a good substitute for fault tree approach for hazard assessment with promising perspective of application.