用于故障诊断的新型多专家系统  

A New Multi-Expert System Applied in Fault Diagnosis

作　　者：熊小龙[1] 王洁[1] 牛天林[1] 杜燕波[1] 

机构地区：[1]空军工程大学导弹学院,陕西三原713800

出　　处：《电光与控制》2009年第7期58-61,共4页Electronics Optics & Control

摘　　要：为了能在复杂系统发生故障时,更快更好地诊断和排除,结合支持向量机、模糊积分和专家系统的各自特点,提出了最小二乘支持向量机(LSSVM)与模糊积分(FI)的多专家系统整体结构和诊断模型。LSSVM具有模型结构简单、推广能力较好和全局最优解等优点,将其用于专家系统中,实现知识自动获取和快速逻辑推理;FI方法的运用,则提高了多专家协同诊断的精度;专家系统完成知识库管理与维护、符号推理诊断以及解释等工作。该方法克服了传统的诊断专家系统缺少有效的自学习、自适应机制,难以解决非线性系统故障问题的不足。因此,这对于解决复杂系统多征兆、多原因的故障问题,是一条方便快捷的途径。In order to make a fast fault diagnosis when a complex system fails, we studied the characteristics of Support Vector Machine（ SVM）, Fuzzy Integral （FI） and expert system, and put forward an integral structure and diagnosis model of multi-expert system based on Least Square SVM （LSSVM） and FI. LSSVM has the advantages of simple structure, good generalization and globally optimal solution etc. , which can realize the functions of obtaining knowledge automatically and logic reasoning quickly when used with expert system. The using of FI can improve the cooperation diagnosis precision of multiple experts. The expert system implements management and maintenance of the repository, symbol reasoning diagnosis and explanation. This method overcame the shortcomings of traditional expert system, such as lack of the mechanism of effective self-study and self-adapt, and difficult to resolve the nonlinear system fault problems. Therefore, it is a convenient way to deal with the multi-sign and multi-reason fault problems of a complex system.

关 键 词：复杂系统 故障诊断 最小二乘支持向量机 模糊积分 专家系统 

分 类 号：V267[航空宇航科学与技术—航空宇航制造工程] TP18[自动化与计算机技术—控制理论与控制工程]