基于离散时间贝叶斯网络的复杂机械系统重要度计算方法

A Complex Mechanical System Importance Calculation Method Based on Discrete-time Bayesian Networks

为了对复杂动态系统部件进行有效的重要度分析,在构建基于事件树-动态故障树(ET-DFT)的概率安全评价模型的基础上,把ET-DFT模型映射为离散时间贝叶斯网络(DTBN),给出各静态和动态逻辑门向DTBN转化的方法以及各逻辑门条件概率表的计算方法。利用DTBN节点的独立性和双向推理功能,给出ET-DFT分层模型FV、RRW、BM和RAW等重要度的计算方法。数控机床液压系统应用实例的分析验证结果表明,基于离散时间贝叶斯网络的复杂机械系统重要度计算方法既能有效得到元件在各时间区间内的重要度,又能准确求出系统故障时各元件在各时间区间的故障概率以及某元件在某时间区间故障时各节点的故障概率。

In order to analyze importance of components of complex dynamic system effectively, the probabilistic safety assessment model ET-DFT was established and it was mapped to Discrete-time Bayesian Networks （ DTBN ）. Methods were given for converting static and dynamic logic gates to DTBN and computing conditional probability table of each logic gates. The computing methods of FV, RRW, BM and RAW importance of ET-DFT layered model were presented by using the independence of the DTBN node and bidirectional reasoning function. A certain numerical control machine hydraulic system was taken as an application example to illustrate the proposed method. The results show that not only the importance of components in a certain time interval can be obtained effectively, but also the failure probability of each element in each time interval when the system is failure and the failure probability of each node when an element is failure in a certain time interval can be solved.