基于PARAFAC分析和SVM离心泵故障诊断方法

Fault Diagnosis Method for Centrifugal Pumps Based on PARAFAC Analysis and SVM

在多尺度平行因子分析理论的基础上,将原始信号经过多尺度小波分解得到三维时频信号,再经平行因子分析得到通道加载因子、时间加载因子和频率加载因子,通过实验分析,后二者可以明显地表征设备正常或故障状态,利用这一特征建立不同状态的离心泵与其对应的时间加载因子和频率加载因子的映射关系,并以此作为改进粒子群算法优化后的支持向量机分类器的特征向量进行故障分类。与小波包能量特征相比,所提的这种诊断方法用于离心泵故障诊断时提取特征更为简便,所提分类器的分类准确率有显著提高,而其复杂度却没有明显增加。

A new fault diagnosis method for centrifugal pumps based on parallel factor analysis(PARAFAC)and support vector machine(SVM)is proposed.Based on the multi-scale theory of PARAFAC,the original signal is decomposed by means of multi-scale wavelet transform(WT)to obtain 3D time-frequency signal.Then,by using PARAFAC,the channel loading factor,time loading factor and frequency loading factor are obtained.The experimental analysis shows that the latter two can clearly characterize the normal or fault condition of equipment.This feature can be used to establish the mapping relationship between centrifugal pumps in different states and their corresponding time loading factors and frequency loading factors.This relationship is further used as the feature vector of the SVM classifier optimized by the improved particle swarm optimization(IPSO)for fault classification.The proposed method can extract the features more easily in fault diagnosis of centrifugal pumps than the wavelet packet energy features method,and the classification accuracy of the proposed classifier is significantly raised,while its complexity is slightly increased only.