基于优先级诊断树的旋转机械故障诊断专家系统

A Priority and Diagnosis Tree-based Expert System for Fault Diagnosis of Rotating Machinery

针对旋转机械故障知识的膨胀导致的专家系统诊断效率低下以及知识库难以维护的问题,提出了一种新的、基于优先级诊断树的专家系统框架。将诊断设备分解建立诊断树并划分领域知识库,再利用模糊群组多属性决策方法排序诊断树导出元知识,并在弹性匹配模式基础上改进了规则的匹配方式。诊断时,推理机从优先级最高的节点开始,按照元知识进行规则的搜索、匹配,以此建立某设备动力与传动装置的故障诊断系统。研究表明,与传统专家系统相比,新系统具有更高的效率、并且诊断结论更可靠、可维护性也更好。

To solve the problems of expert systems such as low efficiency of search and poor maintenance of knowledge base which are caused by the greatly expansion of the fault knowledge of rotating machinery, a new framework of expert system based on priority and diagnosis tree was proposed. A diagnostic tree was built according to the functions of the devices in the machine and the entire domain knowledge base was decomposed into a number of sub-bases. The diagnostic priorities of nodes in the tree are determined based on a fuzzy group multiple attribute decision making method and a meta knowledge base was generated automatically. Some improvement was made to the matching method on the base of the flexible matching model. The inference process starts from the node with the highest priority and searches for possible device at fault based on the meta knowledge base. According to the new framework, a fault diagnosis system for the power and gearing equipment was set up. The experiment result shows that the new system is more efficient, more reliable and much easier to maintain.