Fault diagnosis of multi-state gas monitoring network based on fuzzy Bayesian net

The monitoring system is a critical function in modern underground mine gas accidents prevention. As a complex dynamic critical system, its fault diagnosis is generally based on the traditional two-state fault tree. In order to solve the problem that the model can only deal with two-state problem, a fuzzy Bayesian network (BN) is used to deal with the polymorphic fault tree. A Bayesian model of the fault of the multi-state gas monitoring system is constructed for machine learning and optimization. On this foundation, the model is validated and applied to a real system. Combined with the fuzzy fault tree importance algorithm, the reliability analysis which has guiding significance for improving the reliability of the gas monitoring system is carried out. The top event, the failure probability of each node, and the fuzzy importance and status importance of each factor are obtained respectively through calculating. Finally, the troubleshooting order of all nodes and key troubleshooting nodes are obtained. When the system is in various states, different parts of the system were diagnosed.