Fault Diagnosis Method of Rolling Bearing Based on Adaptive Cyclostationarity Blind Deconvolution

The vibration signals collected by sensors can be regarded as the convolution result of multiple interference factors and fault source signals, so the deconvolution method has certain advantages in signal processing. A fault diagnosis method for rolling bearings based on adaptive maximum second-order cyclostationary blind deconvolution (ACYCBD), improved arithmetic optimization algorithm (IAOA) and optimized support vector machine (SVM) is proposed. Firstly, the optimization criterion based on the least square envelope spectral dispersion entropy is constructed, and the optimal inverse filter parameters are solved by improved arithmetic optimization algorithm. Then, the optimal inverse filter is used for blind deconvolution of the fault signal, and the fault features of the filtered signal are extracted by multi-scale dispersion entropy to form feature vectors. Finally, the feature vector is input into the classification model. The hyperparameters of SVM are optimized by IAOA and train the best classification model and test it. The test results show that the proposed method can accurately and effectively identify the faulty kinds of bearings, and can maintain more than 90% accuracy even under noise pollution.