Comparison of Support Vector Machine and Radial Basis Artificial Neural Network in Fault Diagnosis Of Bearings Using Short Time Fourier and Choi-Williams Transforms

This paper focus on the fault diagnosis of bearing with two soft computing methods. Firstly, the vibration signal of a test rig is done. An intelligent method is selected to diagnose a ball bearing type 1206, using the vibration signal. This paper presents two methods time-frequency based approach for classifying the vibration signals of rolling bearing. It uses the features extracted from the time-frequency distribution (TFD) of the vibration signal segments. The results of applying the method to a database of real signals and then classifying feature by radial basis artificial neural network and support vector machine reveal that, for the given classification task, the selected features consistently exhibit a degree of discrimination between the vibration signals collected from healthy and fault machine. A comparison between the performances of the features extracted from two TFDs shows that the Choi-williams is better than STFT. Also the results show that the accuracy of SVM for classification is better than RBF artificial neural network.

Keywords:

bearing faults , support vector machine , radial basis network