Towards Advanced Diagnosis Recognition for Eccentricities Faults: Application on Induction Motor

Abstract

Artificial Intelligence (AI) is expected to be a large driver in industrial applications competitiveness in the not-so-distant future. Induction motors (IMs) are used worldwide as the “workhorse” in industrial applications. The paper reviews the possibility of integrating artificial intelligence techniques for condition monitoring and fault diagnosis of induction motors so-called advanced diagnosis. The paper focuses on advanced diagnosis method related on the recognition, classification and prognostics of eccentricities faults in induction motor drives. Rotor eccentricity has been the aim of many researchers. However reliably detection and accurate prediction of eccentricity fault is still not possible and difficult task if appear individually. To face this situation, an intelligent diagnosis system merges Neural Network and Hidden Markov Model together (NN-HMM) into a common framework to overcome the deficiencies of eccentricity diagnosis. Current measurements based on non-parametrical Time-Frequency Representation (TFR) are used for features extraction. Then, a features selection method using Fisher's Discriminant Ratio (FDR) is applied to select an optimal number of the extracted features associated with polynomial approach to track, recognize of various eccentricities faults types and degree precisely. An experimental study on a 7.5h induction motor prove the reliability and the efficiency of the proposed method in condition monitoring of eccentricities with different degree 0%, 20%, 40%, 60, 80% precisely independent of load or motor type.