多专家故障诊断系统研究  

Study of multi-experts fault diagnosis system

作　　者：胡建中[1] 许飞云[1] 贾民平[1] 

机构地区：[1]东南大学机械工程学院,南京210096

出　　处：《制造业自动化》2007年第2期1-3,51,共4页Manufacturing Automation

基　　金：国家高技术研究发展计划(863)资助项目(2003AA430010);国家自然科学基金资助项目(50475078)

摘　　要：针对已有诊断系统的知识可重用性差、问题求解能力不足的缺点,研究了多专家协同诊断系统,采用模糊积分方法进行多专家诊断的结论融合。为了寻找最优的模糊密度值,提出了利用遗传算法进行模糊密度值优化,提高基于模糊积分方法的多专家诊断结论融合精度。针对所提出的算法,结合工程实际数据进行了分析,结果表明了该方法的有效性。In allusion to the bad reusability of knowledge, deficient ability of problem solving of developed fault diagnosis systems, the collaborative diagnosis system is studied, the fuzzy integral based fusion method is adopted in multi-experts fault diagnosis system. To determine the optimal fuzzy integral measure, the genetic algorithm is used to search the optimal fuzzy integral measure values to improve the fusion accuracy of the multi-experts diagnosis decisions. For the presented algorithm, a engineering data is analyzed, the result shows the validity of the algorithm.

关 键 词：多专家 故障诊断 模糊积分 遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]