基于多源信息融合故障树与模糊Petri网的复杂系统故障诊断方法

Fault diagnosis method of complex system based on multi-source information fusion fault tree and fuzzy Petri net

针对复杂系统故障树模型构建困难且模型冗余节点多、计算复杂的问题,提出一种基于多源信息融合故障树与模糊Petri网的故障诊断方法。该方法先将多源信息进行标准化处理,从处理后的信息中提取维修元数据,同时利用数据挖掘方法得到故障关联项集。通过维修元数据、故障关联项集和系统结构关系的映射、融合,更加全面、准确地构建复杂系统故障树模型。采用模糊Petri网对多源信息融合故障树模型进行简化和改进,并利用基于模糊Petri网的动态故障推理方法和基于关联矩阵的最小割集求解方法建立复杂系统故障诊断方法,提高了故障的诊断速度与推理效率。以汽车发动机故障诊断过程为例,证明了所提方法的合理性和有效性。

Aiming at the difficulty of constructing fault tree model for complex system,and the problems such as redundant nodes and complex computation existed in the model,a new failure diagnosis method based on multi-source information fusion method and fuzzy Petri net was carried out.In this method,the multi-source was transformed into the standard form,the maintenance metadata was extracted from standardized multivariate information,and the fault correlation set was constructed by using data mining method.The fault tree of complex system was efficiently and comprehensively constructed with combination and mapping of maintenance metadata,fault correlation and system structure relation.The fault tree was simplified and improved through fuzzy Petri net,and the fault diagnosis method for complex system was established by using fuzzy Petri net based dynamic fault reasoning algorithm and correlation matrix based minimum cut set method.The rationality and validity of the proposed method was proved by an example of automobile engine fault diagnosis.