Improvement of induction motor fault diagnosis performance by using genetic algorithm-based feature selection

In this study, features are extracted from time vibration signals for the purpose of diagnosing motor faults. On the basis of the specific distance criterion, a simple genetic algorithm (GA) is employed to evaluate and select the optimized features for induction motor fault classification. The selected features are applied to the decision tree and the k-nearest neighbour (k-NN) algorithm in order to show the efficiency of the proposed feature selection method. The diagnostic results show that the optimal feature selection is useful to improve the fault diagnosis performance.