模糊Petri网在故障诊断知识表示和推理中的应用  

Application of fuzzy Petri net to knowledge representation and inference in fault diagnosis

作　　者：潘兴隆[1] 贺国[1] 王三龙 

机构地区：[1]海军工程大学船舶与动力学院,武汉430033 [2]海军驻某军事代表室,广东湛江524003

出　　处：《海军工程大学学报》2012年第1期52-56,共5页Journal of Naval University of Engineering

基　　金：国家部委基金资助项目(415117123);海军工程大学博士研究生创新基金资助(HGBSJJ2011003)

摘　　要：针对传统的模糊产生式规则知识表示复杂、推理效率低、单一阈值设定不能满足复杂故障诊断需求的问题,给出了模糊产生式规则转换为模糊Petri网的一般方法和故障诊断模型,以及基于双阈值设定的两级故障诊断推理算法,并将其应用于某型柴油机遥控系统故障诊断之中。通过设置双阈值提高了故障诊断的推理速度和推理效率,同时又能对系统的准故障状态进行检测,及时发现潜在故障,排除故障隐患。The conventional fuzzy production rule is complicated and lowly efficient in knowledge representation and inference,and the single threshold value can not meet the requirement of fault diagnosis.Therefore,a general method and its corresponding fault diagnosis model were established for transforming fuzzy production rule to fuzzy Petri net.And meanwhile the two-level algorithm was presented based on setting double threshold value.The typical method can be used not only to analyze and diagnose the three kinds of states of the system and complex faults effectively,but also to handle faults in the remote control system of a certain diesel.The results indicate that the speed and efficiency of inference can be improved by setting the double threshold value,thereby detecting and removing potential faults.

关 键 词：故障诊断 产生式规则 模糊PETRI网 阈值 

分 类 号：TK427[动力工程及工程热物理—动力机械及工程]