基于动态贝叶斯网络的高速铁路牵引变电所可靠性分析

Reliability analysis on traction substation of high-speed railway based on dynamic Bayesian network

结合GO-FLOW法的动态特性,将动态贝叶斯理论应用于高速铁路牵引变电所可靠性的分析中。首先将GO-FLOW法中的功能操作符、逻辑操作符、信号发生器、输入信号流等转换为相应的动态贝叶斯网络模块,并建立其条件概率表;然后根据牵引变电所主接线GO-FLOW图和主接线系统功能逻辑关系进行连接,得到基于GO-FLOW图的牵引变电所主接线的动态贝叶斯网络模型;最后运用动态贝叶斯算法对模型求解,得到了牵引变电所主接线的可靠性参数和可靠性变化曲线,结果表明当考虑部件随时间推移而失效的情况时更加符合实际。与其他方法相比,该方法考虑了分析对象的动态特征,减少了公式推导过程,简单清晰,便于实际应用。

Combined with the dynamic characteristics of the GO- FLOW method,the dynamic Bayesian theory was applied to the reliability analysis on traction substation of high- speed railway. Firstly,the function operator,logical operator,signal generator and input signal flow of the GO- FLOW method were converted into the corresponding dynamic Bayesian network modules,and its conditional probability table was established. Then through the link based on the GO- FLOW diagram of the main connection of traction substation and the function logical relation of main connection system,the dynamic Bayesian network model on the main connection of traction substation based on GO- FLOW diagram was obtained. Finally,the model was solved by using the dynamic Bayesian algorithm,and the reliability parameters and change curves for the main connection of traction substation were obtained. The results showed that considering the failure of components over working time is more consistent with the practice. Compared with other methods,the dynamic characteristics of analysis objects are taken into account in this method,as well as the formula derivation process is reduced,which is more simple and convenient for practical application.