Analysis of the performance of an MRAS based sensorless speed estimation scheme for induction motors under fluctuating inputs

A sensorless speed estimation scheme for a three phase squirrel cage induction machine is investigated in this paper. The Model Reference Adaptive Scheme (MRAS) based sensorless speed estimator uses the error in instantaneous reactive power for speed estimation. The Q-MRAS based method is computational simple and its closed loop nature reduces steady state errors. The performance of the speed estimator under changes in load torque, fluctuations in supply voltage and frequency and under the presence of unbalance and harmonics are investigated using MATLAB simulations. The Q-MRAS based scheme is found to be quite robust under all conditions, with almost no steady state error and transient errors less than 3%.