小波包分解法与粗糙集神经网络在电机故障诊断中的应用  

Application of Instantaneous Power Decomposition via Wavelet Packet and RS-RBF in Induction Motor Fault Diagnosis

作　　者：林茂[1] 李孝全[1] 张兴[1] 

机构地区：[1]空军工程大学导弹学院,陕西三原713800

出　　处：《大电机技术》2012年第5期31-34,共4页Large Electric Machine and Hydraulic Turbine

摘　　要：针对感应电机故障特征复杂、特征提取方法不足,提出了瞬时功率小波包分解的方法。分析电机单相瞬时功率,滤波后进行小波包分解,求取故障特征对应子频带小波包分解系数的均方根值及其变化率,并用以表征故障特征,以此作为电机故障的依据,运用粗糙集理论进行约简,将约简结果作为特征向量输入到RBF网络中,进行故障诊断。结果表明该方法诊断灵敏度高,可用于电机的故障诊断。To solve the problems in obtaining bearing fault characteristics of induction motors,a method based on instantaneous power decomposition via wavelet packet is processed.By analyzing the induction motors and decomposing the signal by wavelet packet after it is filtered,the root mean square of node coefficients and its change rate used as the symptom of bearing fault was calculated.Using fault data as value attributes to build the decision table.And use of rough sets of theoretical calculation of the reduction decision.On this basis,the RBF net work can be used to make fault diagnosis.The result shows that the actual diagnosis can effectively improve the faults diagnosis accuracy.

关 键 词：感应电机 RBF网络 粗糙集 小波包分解 

分 类 号：TM346[电气工程—电机]