Torque and Speed Estimator for Sensorless Induction Motor Drive Using Cascade Neural Networks

Many electronic drivers for the induction motor control are based on sensorless technologies. The proposal of this work is to present an efficient estimator of torque and speed, from transient to steady state, using artificial neural networks. The method proposed is based on off-line training considering different types of load and wide range of voltage applied to the induction motor. The inputs of the network are the RMS voltage and current of the induction motor and the control processing effort is reduced to simple matrix solving after the neural network is trained. Simulation results are also presented to validate the proposed approach