Rolling Bearing Fault Diagnosis Method Based on Multilayer Noise Reduction Technology and Improved Convolutional Neural Network

： Aiming at the problem that the weak fault characteristics of rolling bearings are difficult to achieve effective diagnosis under strong noise, a new method of bearing fault diagnosis based on multilayer noise reduction technology and improved improved convolutional neural network(ICNN) is proposed. First, the vibration signal of the rolling bearing is pre-processed to obtain labeled data samples, which are divided into a training set and a test set. Then singular value decomposition(SVD) is used to process the training samples, the number of valid singular values is selected by the singular value mean method to obtain the original noise-reduced signal and the noisy signal. In order to avoid losing the details of weak faults, the noisy signal is further denoised by SVD to eliminate modal aliasing and input empirical mode decomposition to obtain the intrinsic mode function(IMF). The IMF component is selected according to the variance contribution rate and superimpose it with the original noise reduction signal to obtain the final signal. The processed training set data is input into the improved ICNN that introduces attention mechanism for learning. Finally, the obtained diagnostic model is applied to the test set, the fault category diagnosis results are output. Through the rolling bearing fault diagnosis simulation test, the test is performed in a strong noise environment, the results show that the proposed method a strong