ANN based RF-MRAS speed estimation of induction motor drive at low speed

In this paper, artificial neural network assisted MRAS based speed estimator is presented for the poly-phase induction motor drive. The adaptive model of conventional RF-MRAS is replaced by a neural network in order to immune the system from rotor resistance variation and for the reduction of complex mathematical computational efforts. To train the network, Levenberg-Marquardt back-propagation algorithm is used. The result is simulated at different operating conditions, particularly at low speeds.