Research on bearing fault diagnosis based on SGA-optimized ICEEMDAN

Aiming at the problems of traditional bearing fault diagnosis methods, such as the selection of modal decomposition parameters depending on experience and the insufficient extraction of nonlinear features, an intelligent fault diagnosis method based on improved adaptive noise complete ensemble empirical mode decomposition (ICEEMDAN) and slope entropy (SE) optimized by Snow Geese Algorithm (SGA) is proposed in this paper. Firstly, SGA is used to optimize the key parameters of ICEEMDAN to improve the completeness and noise immunity of modal decomposition. Secondly, the slope entropy is introduced to quantify the dynamic nonlinear characteristics of the fault signal. Finally, a hybrid diagnosis model is constructed by combining a convolutional neural network (CNN) and support vector machine (SVM) to achieve accurate classification of bearing fault categories. Experiments based on bearing data from Jiangnan University show that compared with ICEEMDAN-MSE (sample entropy) and ICEEMDAN-PE (permutation entropy), the average diagnostic accuracy of the proposed method reaches 85%, which verifies the robustness and engineering application value of the method under complex working conditions.