Fault Diagnosis of On-board Signaling Equipment Based on Bayesian Network

Aiming at the problems of excessive dependence on expert experience and low diagnosis accuracy rate of high-speed railway train-controlled on-board equipment, a fault diagnosis method based on expert knowledge and Bayesian network as the core is proposed. Firstly, taking the on-board equipment of the CASCO U888 CBTC signaling system as the research object, a Bayesian network based on expert knowledge was established. Secondly, according to the fault record dataset, the K2 algorithm is used to learn the structure, and the learned Bayesian network structure is appropriately parsized. The maximum likelihood estimation algorithm is used to learn the parameters to obtain the optimal Bayesian network for fault diagnosis. Finally, through diagnostic reasoning, the diagnostic accuracy rate of the fault diagnosis model is 89.6%, which verifies the feasibility and effectiveness of the Bayesian network model.