Abstract: To improve the reliability of complex systems,the modeling and reasoning methods of Improved Fuzzy Fault Petri Net (IFFPN) was proposed based on fuzzy Petri net and fault Petri net.To describe machinery failure information more intuitively and clearly,the coloring rules of place,token and transition were defined by combining with the fuzzy production rules in IFFPN model.The forward reasoning of IFFPN model reflected the failure flooding characteristics.The forward deduction algorithm based on the intelligent judgment of firing matrix was applied in realizing machinery failure condition evaluation.The backward reasoning of IFFPN model was the process of machinery failure diagnosis,and the backward reasoning matrix was proposed by combining with a minimal cut sets diagnosis based on the incidence matrix of Petri net to realize automatic diagnosis with any place in the model.The minimal cut sets priority could be acquired according to the machinery failure frequency,which could contribute to the efficiency of the failure diagnosis.A feed system of CNC machine tools was taken as the example to validate the correctness and efficiency of the proposed approach.

Key words: improved fuzzy fault Petri net, machinery diagnostics, Petri net reasoning, minimal cut sets, feed system

摘要： 为提高复杂系统的可靠性,在综合模糊Petri网和故障Petri网优点的基础上,提出了改进模糊故障Petri网的建模及推理方法。该模型定义了库所、托肯及变迁的着色规则,并结合模糊产生式理论,应用于复杂系统的故障推理,从而更加直观、明确地描述故障状态及信息。该模型的正向推理反映了故障传播的固有特性,采用基于智能点火判断矩阵的正向演绎算法,实现了故障状态快速、准确的智能评价。该模型的逆向推理为故障诊断过程,引入智能的逆向推理矩阵,并结合基于Petri网关联矩阵的最小割集诊断法,实现了模型中任意库所故障的逆向自动诊断,同时根据故障易发率得出最小割集优先顺序,提高故障诊断的效率。以数控机床进给系统的故障分析为例,验证了所提模型和方法的正确性与高效性。

关键词: 改进模糊故障Petri网, 故障诊断, Petri网推理, 最小割集, 进给系统