基于小波抗体记忆克隆算法的电机故障诊断  

Fault Diagnosis of Motor Based on Wavelet Analysis and Antibody Memory Clonal Algorithm

作　　者：吴洪兵[1,2] 楼佩煌[1] 唐敦兵[1] 

机构地区：[1]南京航空航天大学机电学院,南京210016 [2]淮安信息职业技术学院电气工程系,淮安223003

出　　处：《振动．测试与诊断》2013年第5期741-745,907,共5页Journal of Vibration,Measurement & Diagnosis

基　　金：国家自然科学基金资助项目(61105114);江苏省科技支撑计划资助项目(BE2010189)

摘　　要：针对电机定子、转子、轴承偏心、气隙偏心等故障,提出了一种将小波分析和抗体克隆算法相结合的故障诊断新方法。使用小波技术对电机定子电流监控数据进行预处理,对采样信号进行小波分解,提取各频段的能量,归一化后将能量作为故障诊断的特征向量。将得到的故障特征向量作为抗原,由算法建立的聚类中心作为免疫系统的抗体,然后利用抗体记忆克隆算法对故障样本进行故障识别分类。试验和应用结果表明,用小波记忆克隆算法能很好地分类出电机的各种工作状态,使电机故障诊断具有较高的正确率和浓缩率。A new approach based on a wavelet analysis and antibody clonal algorithm is proposed for stator,rotor,bearings eccentric,and air gap eccentric fault diagnosis of motor.Stator current monitoring data are pre-processed and decomposed by wavelet technique.The energy of the band that extracted by wavelet analysis is normalized as eigenvectors of the fault diagnosis.Antigens are the fault eigenvectors,and antibodies are clustering center established by immune algorithm.Then antibody memory clonal algorithm begins to identify fault classification.The experiment and application results show that the wavelet analysis and memory clonal algorithm may classify the working condition of motor,and it has higher accuracy and enrichment rate.

关 键 词：故障诊断 小波分析 电机 人工免疫 抗体记忆克隆算法 

分 类 号：TH17[机械工程—机械制造及自动化] TM3[电气工程—电机]