Bearing fault diagnosis method based on improved convolutional neural network

The invention relates to a bearing fault diagnosis method based on an improved convolutional neural network, and the method comprises the steps: inputting an original vibration signal of a to-be-detected rolling bearing into an improved CNN model for fault diagnosis, and the improved CNN model mainly consists of an input layer, a feature extraction layer and a classification output layer. The feature extraction layer comprises a plurality of convolution layers, activation layers and pooling layers which are stacked in sequence; a global mean pooling layer is adopted to replace a full connection layer after the convolution layer of the last group of feature extraction layers, and then a Softmax classifier is connected; the output of the Softmax classifier is used as a test result; and the Softmax classifier outputs the maximum probability value of each row, namely the test result. According to the bearing fault diagnosis method, the original fault data are directly used as model input, the fault classification