Bayesian network based fault diagnosis and maintenance for high-speed train control systems

High speed train control systems are complex, realtime, and distributed systems. Failure of any of such subsystems can have heavy impact on the service itself, leading to obvious deterioration of performance, reduction of perceived quality and increment of costs. This paper proposed a Bayesian network based fault diagnosis and maintenance for high-speed trains control systems. Firstly, a Bayesian network based fault model was generated by Bayesian learning from fault table. Then, the maximum possible fault cause through the reverse reasoning ability of the Bayesian network was deduced. Finally, a Dynamic Bayesian Network (DBN) based maintenance model was presented and the real-time maintenance results of high-speed train control systems was used to verify the efficiency of the proposed algorithm.