Application of Denoising Method Combining Normalized Least Mean Square and Continuous Wavelet Transform in Rolling Bearing Fault Diagnosis

Due to the coordinated operation of various mechanical components and the interference of the working environment, resulting in coupled interference of the vibration signals of rolling bearings with various other components and environments, which interferes with the subsequent feature extraction and affects the accuracy of fault diagnosis. In this paper, a new fault diagnosis method is proposed, which uses normalized least mean square (NLMS) for data preprocessing and multi-scale convolutional neural network combination. Continuous wavelet transform of the preprocessed data can highlight the feature differences and further eliminate the coupling interference. A multi-scale convolutional kernel is used to construct the parallel structure and follow the fusion structure. It is experimentally verified that the method can effectively extract high-quality features and maintain accurate classification performance in noisy environments.