基于贝叶斯网络和AMESim仿真的液压系统故障诊断方法

摘要

为考虑多属性信息和验证故障分析结果,提出基于贝叶斯网络和AMESim仿真的液压系统故障诊断方法.根据转向沉重故障征兆建造T-S故障树,求得基本事件的T-S关键重要度;建造故障搜索决策的贝叶斯网络拓扑结构,考虑基本事件的搜索成本、维修综合代价和T-S关键重要度等属性,根据求得的综合评价值进行故障诊断排序;按故障诊断的顺序利用AMESim进行故障仿真,将仿真得到的压力流量特性与实际故障征兆对比,指导完成故障诊断.%In order to consider the multi-attribute information and verify fault analysis result,the fault diagnosis method of hydraulic system based on Bayesian network and AMESim simulation was proposed.T-S fault tree was established by hydraulic understeering symptoms so as to calculate T-S criticality importance of basic events.Bayesian network topology structure of fault searching determination was established,the multi-attribute of searching costs,comprehensive maintenance costs and T-S criticality importance of basic events were considered,and the sequence of fault diagnosis was obtained by the comprehensive evaluation value.Fault simulation was completed by using of AMESim according to the sequence of fault diagnosis.Therefore fault diagnosis is accomplished under guidance by comparing the pressure and flow characteristics obtained from simulation with the actual fault symptoms.