基于免疫克隆与模糊聚类的汽轮机故障诊断方法的研究  

Study on the Method of Steam Turbine Fault Diagnose Based on Immune Clone and Fuzzy Clustering Algorithm

作　　者：刘福荣[1] 李广辉[1] 

机构地区：[1]哈尔滨电力职业技术学院,哈尔滨150030

出　　处：《汽轮机技术》2009年第1期58-60,共3页Turbine Technology

基　　金：黑龙江省教育厅科研项目(11525009)

摘　　要：提出了一种免疫克隆选择算法与模糊C-均值聚类算法相结合的混合聚类算法,是一种有监督学习和无监督学习结合的算法。该算法首先用克隆选择算法对模糊聚类中心的选取进行指导,然后进行聚类。仿真结果表明,将该算法用于汽轮机的故障诊断,能够正确地诊断出存在的故障,提高了故障诊断的准确性和有效性,其性能优于模糊C-均值聚类的故障诊断方法。This paper proposed a hybrid clustering algorithm of combine immune clone selection and fuzzy c-mean cluster algorithm. It is a kind of algorithm which combined supervised and unsupervised learning. The algorithm use clone selection algorithm to supervise selecting fuzzy clustering central, and then clustering. Result of simulation shown that used the algorithm to diagnosed the fault of steam turbine, it can diagnose fault accurately, and it improved veracity and validity of fault diagnose. The performance excelled fault diagnose based on fuzzy c-mean cluster algorithm.

关 键 词：汽轮机故障诊断 人工免疫 克隆选择 模糊C-均值聚类 

分 类 号：TP139[自动化与计算机技术—控制理论与控制工程]