关键词: 复杂系统, 故障诊断, 多源信息, 故障树, 模糊Petri网

Abstract: 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.

Key words: complex system, fault diagnosis, multi-source information, fault tree, fuzzy Petri net