Electric Motor Fault Diagnosis Based on Parameter Estimation Approach Using Genetic Algorithm

This paper proposes a new scheme of induction motor parameter estimation using Genetic algorithm (GA) for condition monitoring. The flux linkage model and torque model of an induction motor is adapted to the estimation. The scheme is developed to obtain all the motor parameters: stator and rotor resistance, stator and rotor leaking reactance and magnetizing reactance, which paves the way to diagnose different types of the faults. The scheme minimizes the difference between the measured and the predicted state variables: three phase currents and rotor speed. The scheme is evaluated firstly with different motor sizes and different load levels by simulation tests and then by the experimental data of the induction motors under normal operating condition at different load levels and fault conditions. The results from both tests show that the new scheme can estimate the parameters and predict the motor condition with sufficient accuracy for motor fault diagnosis. Inedex terms— Induction Motor, Parameter Estimation, Genetic Algorithm, Condition Monitoring, Fault Detection