Back propagation based ANN technique for rotor position estimation of 8/6 switched reluctance motor

This paper presents a novel approach to the rotor position estimation of a switched reluctance motor (SRM). The complexity involving the conventional flux estimation techniques is eliminated by introducing an artificial neural network (ANN) based estimation method. Back Propagation based training algorithm used in this method provides adequate and accurate training which makes it possible to achieve angular position values with high precision. With a sufficiently large training data, the ANN can build up a correlation for flux, current and theta. Using these values we can deduce accurate position of the rotor which can be further used in speed control techniques, effectively obliterating the need for a conventional speed sensor. The simulation results validate the accuracy and reliability of this method.