摘要
针对模拟电路故障诊断样本结构复杂、分类困难的问题,提出了一种基于免疫克隆聚类的故障诊断方法。在免疫克隆聚类的基础上,设计两个目标函数,既实现了类内的紧凑性,又兼顾了样本的连通性。通过核密度估计,解决了克隆方向性问题;通过混沌变异避免算法陷入局部极值。实验结果表明,该方法对复杂结构样本聚类效果好,适用于解决模拟电路故障诊断问题。
The fault diagnosis sample of analog circuit has a complex structure and is hard to be ctassttleO. Aiming at this problem, an approach based on immune clonal algorithm is proposed in this paper. Two objective functions are designed on immune clonal cluster to realize the compact category and the connectivity at the same time. The clone di- rection is determined by the kernel density estimation and the chaos variation is applied to avoid local extremum. Experi- mental result shows this method has better cluster effect in complex construction samples, and is suitable for analog cir-cuits.
出处
《仪表技术》
2017年第5期22-26,共5页
Instrumentation Technology
关键词
故障诊断
免疫克隆
聚类
fault diagnosis
immune clonal
cluster
