Optimum multi-fault classification of gears with integration of evolutionary and SVM algorithms

In the present work, the multi-fault classification of gears has been attempted by the support vector machine (SVM) learning technique using frequency domain data. The proper utilization of SVM is based on the selection of SVM training parameters. The main focus of the paper is to examine the performance of the multiclass ability of SVM technique by optimizing its parameters using the grid-search method, the genetic algorithm (GA) and the artificial-bee-colony algorithm (ABCA). Four different fault conditions have been considered. Statistical features are extracted from frequency domain data. The prediction of fault classification has been attempted at the same angular speed as the measured data as well as innovatively at the intermediate and extrapolated angular speed conditions. This is important since it is not feasible to have measurement of data at all speeds of interest. The classification ability is noted and it demonstrates the excellent performance.