基于克隆聚类的特征提取与多传感器故障诊断  

Feature Extraction and Multi-Sensor Fault Diagnosis Based on Clonal Clustering

作　　者：侯胜利[1] 王威[1] 乔丽[1] 史霄霈[1] 周根娜[1] 

机构地区：[1]徐州空军学院,江苏徐州221002

出　　处：《电光与控制》2010年第6期69-72,88,共5页Electronics Optics & Control

基　　金：国家自然科学基金资助项目(60672179)

摘　　要：基于人工免疫理论中的克隆选择算法,结合聚类分析方法,提出了基于克隆选择聚类分析的故障特征提取方法。该方法通过删除对分类无关的特征以及压缩类间相关特征,得到最有利于分类的子特征集,提高了分类器的分类性能。并且该算法具有本质上的并行性、计算效率高和聚类能力强等优点。多传感器故障诊断的实验表明,经过克隆选择聚类分析提取的特征对发动机的故障具有更好的识别能力,为发动机的状态监测与故障诊断提供了依据。In order to improve the rapidity and validity of fault diagnosis,a novel approach is proposed for aeroengine fault feature extraction based on clonal selection algorithm and combined with clustering analysis.This data analysis approach can not only reduce the dimension of features by getting rid of the correlation among them but also remove the duplicated or proximately similar data.The obtained subset of features can reduce the cost of computation during the classification process,while improving classifier efficiency.And the method has the essential advantages of high parallel,high efficiency of computing and good clustering ability.Experiments of multi-sensor fault diagnosis were carried out to test the performance of this method.The results showed that the extracted features based on clonal selection clustering analysis have better recognition ability for aeroengine faults.Therefore,it lays a foundation for aeroengine condition monitoring and fault diagnosis.

关 键 词：多传感器 故障诊断 特征提取 聚类分析 航空发动机 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TP212.9[自动化与计算机技术—检测技术与自动化装置]