基于随机集理论的并发故障诊断信息融合方法

Information fusion method of simultaneous fault diagnosis based on random set theory

为了诊断并发故障,提出一种基于随机集理论的信息融合方法。首先构造包含并发故障的论域,并在此论域的超幂集上定义扩展型随机集。基于该随机集和广义集值映射给出证据组合规则的随机集模型,用其构造可以同时适用于单发和并发故障诊断的新型组合规则。此外,根据传感器提供的故障信息构造故障样板模式与待检模式的模糊隶属度函数,利用模糊集的随机集表示以及随机集似然测度,获得两种模式匹配的程度作为待融合的诊断证据。最后通过在电机柔性转子平台上的试验,证明了所提方法可有效地减少单一传感器信息诊断的不确定性,显著提高转子系统故障诊断的精度。

An information fusion method for simultaneous fault diagnosis is proposed based on random set theory.Firstly,a new frame of discernment is constructed in order to include simultaneous faults;secondly,the extended random sets are defined on the hyper power set of this frame of discernment,based on these sets and the established extended mapping,a random set model of evidence combination rule is given,with which new combination rules can be generated to diagnose single fault and simultaneous faults;thirdly,membership functions are used to describe the fault templates and features respectively coming from information collected by sensors,and then,diagnosis evidence can be extracted denoting the matching degrees between fault feature and every fault templates by random set descriptions of fuzzy sets and likelihood measure of random set.Finally,the diagnosis results of a machine rotor show that proposed method can effectively reduce the diagnosis uncertainty of single sensor information and improve the accuracy of diagnosis.