基于多特征融合的轴承故障诊断方法OA

The rotational speed of bearings in rotating machinery equipment fluctuates with changes in the working environment,and this fluctuation can interfere with fault feature extraction.In order to more accurately identify the weak signal changes caused by bearing faults at different speeds,a method of bearing fault diagnosis based on multi-feature fusion is proposed.On the basis of the acoustic emission signal,the data of the inner ring fault,outer ring fault and rolling element fault were collected at three speeds.A one-dimensional acoustic emission time sequence signal is converted into a two-dimensional gray image by means of the wavelet transform(WT)and grayscale processing.The two-dimensional image is treated as a feature map and input into the optimized histogram of oriented gradients(HOG),local binary pattern(LBP),and deep neural network(CVGG16)for the feature extraction.An HLV model is constructed to obtain comprehensive and multi-level information from the feature map.The multi-feature serial fusion on the three types of features extracted from the HLV model conducted,and principal component analysis(PCA)is used to reduce the dimension of the fused features,so as to improve the detection rate.The support vector machine(SVM)learning algorithm is used to train the classification model,so as to realize the bearing fault diagnosis.The results show that,in comparison with other single models,the HLV feature extraction model can obtain more effective fault features with an accuracy rate of 97.50%,and PCA can improve the training speed.The proposed WHLVS bearing fault diagnosis method is superior to other methods,with an accuracy rate as high as 97.52%.The evaluation metrics P,R,F1 and mAP on three public datasets are all above 94%,which verifies the reliability and application potential of this method.