Fault Diagnosis of Vehicle Gearbox based on Support Vector Machine Optimized by Improved Beetle Antennae Search

Abstract

Aiming at the fact that the performance of support vector machine （SVM） in vehicle gearbox fault diagnosis is greatly affected by parameters，a new method of vehicle gearbox fault diagnosis based on improved SVM is proposed based on the research of beetle antennae search （BAS）. Compared with BAS，IBAS modified the step size formula and added Gaussian variation behavior，so that the search ability of the algorithm in the early and late period is balanced and has the ability to jump out of the local optimal，and better SVM parameters could be obtained. The results of gearbox fault diagnosis show that the proposed method can effectively improve the accuracy of fault diagnosis，and the results are better than those of other methods.

Keywords

• Beetle antennae search;Support vector machine;Gearbox;Fault diagnosis