Fault diagnosis of rolling bearings in multiple conditions based on EMD and PSO-SVM

Survey on the application of machine learning in elevator fault diagnosis

“…How to select the best training parameters is the main problem to improve the classification performance of the model. Thus, Li et al [15], used variational mode singular value decomposition to extract features of elevator guide shoe vibration signals, effectively characterized fault feature information as model input, and optimized training parameters through sparrow search algorithm, effectively improving the accuracy and efficiency of SVM fault classification. Wu et al [16], selected Cuckoo algorithm to improve support vector regression, reduce training parameters, enhance the robustness of the model, and enhance the global search capability by using large and small steps alternately, which improved the accuracy and efficiency of SVM in finding the best parameters, and effectively avoided the occurrence of over fitting and under fitting states.…”

Analysis and Diagnosis of Rolling Bearing Faults from the Perspective of Frequency Domain

“…In the rolling bearing defect diagnostics [1][2][3][4][5], the extraction of eigenvectors is a prerequisite for accurate fault diagnosis. Accuracy of fault diagnosis can be significantly increased by effectively and accurately characterizing various fault situations.…”

“…(2) Most of the existing studies only consider the acceleration signal of the faulty bearing, ignoring the consideration of BPFI and BPFO. (3) We visualize our research to comprehensively show the process and reasoning more. ( 4) When encountering indistinguishable signals, this paper uses spectral kurtosis to select appropriate frequency bands to enhance the envelope spectral bandpass filtering.…”