基于贝叶斯方法的智能诊断系统中的故障树排序

Optimize the Fault Tree Sort in the Fault Tree Intelligent Diagnosis System with Bayesian Methods

主要解决了在以故障树为基础的数控机床智能诊断系统中,面对多棵故障树如何确定各故障树的诊断顺序的问题。在故障树诊断方法中融合贝叶斯算法定位已发生故障所在的故障树,利用已有的机床异常状态信号来计算各故障树被触发的后验概率,概率高的故障树优先被诊断,其后对优先级高的故障树采用最小割集的方法进行诊断。这种方法提高了诊断速度,同时避免了贝叶斯网络进行故障诊断时故障诊断系统可靠性低这一弊端。

This paper mainly solved how to determine the diagnosis order of each fault tree when facing many fault trees in the CNC machine intelligent diagnosis system based on fault tree. The Bayesian method was integrated into the fault tree diagnostic method to locate the fault tree where the fault had occurred. The existing machine abnormal status signal would be used to calculate the posterior probability triggered by each fault tree. The fault tree with high probability was first diagnosed, and then the high priority fault tree was diagnosed with the minimum cut set method. This method improved the diagnosis speed, and avoided the low reliability of using Bayesian network method alone.