Fault Diagnosis in Induction Motor using Pattern Recognition and Neural Networks

In this paper, we propose an original approach to classify simple and combined defects in the induction motor. Such classification will be based on parameters measurements made on an induction motor with different operating scenarios. The classification strategy of the proposed approach is realized transforming the three stator currents into three images with several resolutions. Now, the classification of defects becomes a pattern recognition problem, this step is followed by the extraction of interesting features using the histogram of oriented gradient HOG algorithm to construct a robust descriptor by varying different parameters such us the Cell Size. The distinctive descriptor is used to train a multi-layer artificial neural network ANN. The evaluation results conduct on testing data not included in the training process shows the efficiency and the precision of classification of the proposed method.