复杂系统故障诊断中的模糊聚类方法

为了提高复杂系统故障的诊断能力，采用模糊Ｃ－均值聚类算法对原始采样数据进行聚类，并通过模糊传递闭包法和绝对值指数法得到模糊Ｃ－均值法的初始迭代矩阵．用划分系数、划分熵和分离系数来评价聚类的结果是否最佳．采用模糊聚类方法可避免研究复杂系统的内部特性，比仅依据其外部输出的数据进行故障诊断方法简便．通过某飞行器测试系统的应用，表明采用模糊聚类方法后，提高了判别故障的准确率．

Fuzzy Clustering for Complicated System Diagnosis

To improve the ability of fault diagnosis in complicated system, a fuzzy clustering method is introduced on the basis of the original fault diagnosis. It can distinguish complicated faults more efficiently and accurately. This paper uses fuzzy C mean method to classify original sampling data and uses the other methods to get the original iterative matrix. Finally, the three coefficients are used to evaluate the result of fuzzy clustering. The experimental results indicate that the accuracy of this fuzzy fault diagnosis is improved.