Method for Vehicle Roller Bearings Fault Diagnosis Based on Data-driven

By analyzing the rough set and support vector machine theory, an approach for the data-driven design of roller bearing fault diagnosis was introduced in this paper. The fault diagnosis scheme proposed consists of data reduction, Support Vector Machine parameters optimization. The attribute kernel parameters in the rough set were taken into antibodies coding, which was used to vaccinate the population stochastically through a bacterin extraction algorithm. The minimum rule set got from the data reduction algorithm was regarded as the fault of characteristic vectors, which was given as input parameters of SVM classifier to train and classify the fault sample. The experiment results of the rolling bearing fault diagnosis have illustrated the approach proposed has outstanding advantages in precision and efficiency, especially for a small number of samples and the atomization of the roller bearing fault diagnosis can be applied.