Date of Award


Publication Type

Master Thesis

Degree Name



Electrical and Computer Engineering

First Advisor

Saif, Mehrdad




Electric motor condition monitoring can detect anomalies in the motor performance which have the potential to result in unexpected failure and financial loss. This study examines different fault detection and diagnosis approaches in induction motors and is presented in six chapters. First, an anomaly technique or outlier detection is applied to increase the accuracy of detecting broken rotor bars. It is shown how the proposed method can significantly improve network reliability by using one-class classification technique. Then, ensemble-based anomaly detection is utilized to compare different methods in ensemble learning in detection of broken rotor bars. Finally, a deep neural network is developed to extract significant features to be used as input parameters of the network. Deep autoencoder is then employed to build an advanced model to make predictions of broken rotor bars and bearing faults occurring in induction motors with a high accuracy.