Regression Based Performance Analysis and Fault Detection in Induction Motors by Using Deep Learning Technique

The recent improvements related to the area of electric locomotive, power electronics, assembly processes and manufacturing machines have increased robustness reliability induction motors. Regardless availability, application motors in many fields alleges need for operating state supervision condition monitoring. In other words, fault identification at initial stage helps make appropriate control decisions, influencing product quality as well providing safety. Inspired by these demands, this work proposes a regression based modeling analysis performance approach, feature extraction process is combined with classification efficient detection. Deep Belief Network (DBN) stacked multiple Restricted Boltzmann Machine (RBM) exploited robust diagnosis faults adoption training process. influences harmonics over are identified losses mitigated. simulation suggested approach its comparison traditional approaches executed. An overall accuracy 99.5% obtained which turn proves efficiency DBN detecting faults.