模糊支撑半径变量贝叶斯网络多态系统故障概率分析

Failure probability analysis of multi-state system based on Bayesian network with fuzzy support radius variable

针对现有隶属函数描述系统及部件故障状态的不足,构建了一种含模糊支撑半径变量的隶属函数来描述部件故障状态,提出了一种基于模糊支撑半径变量的贝叶斯网络多态系统故障概率计算方法.在隶属函数的构造中,用变量代替精确值描述隶属函数模糊支撑半径,建立了含模糊支撑半径变量的隶属函数;并将其引入贝叶斯网络,利用桶消元法对多态系统叶节点故障概率进行分析,得到叶节点故障概率变化曲线.将所提出的方法与文献中基于模糊支撑半径为定值的隶属函数贝叶斯网络方法相比较,验证了方法的有效性.最后,将该方法应用到数控机床主轴系统故障分析实例中,对叶节点故障概率进行计算.结果表明,该方法能够有效解决多态系统故障状态隶属函数选择中的主观性问题.

In order to overcome the shortage of describing the fault states of system and compo-nents with existing membership functions ,a membership function with fuzzy support radius vari-able is constructed to describe the fault states and a failure probability calculation method of multi-state system based on Bayesian network with variable fuzzy support radius is proposed .The variable was introduced to describe fuzzy support radius of membership function and used to con-struct the membership function .Introducing the established membership function with variable into Bayesian network ,the leaf node failure probability curve was obtained by analyzing the fail-ure probability of leaf nodes on the multi-state system using bucket elimination method .The pro-posed method was compared with Bayesian network method based on membership function with constant fuzzy support radius to verify the validity and generality of the method .Finally ,the method was applied to analyze the CNC machine tool spindle system ,and calculated the leaf node failure probability .The results show that the method can solve the subjectivity of fault states membership function selection in multi-state system effectively .