Model-based torque estimator for switched reluctance motors

Abstract

High-performance torque control in switched reluctance motors (SRM) requires closed-loop control with accurate and high bandwidth torque feedback. An analytical expression for torque in terms of stator phase current and rotor position can be used as torque estimator for this purpose. Such an analytical torque expression can be derived from SRM flux-linkage model, using the coenergy principle. We have observed that most of the flux-linkage models for SRM available in the literature can map the measured flux-linkage accurately. However, the torque expression obtained from the flux-linkage model fail to predict the torque accurately. In this work, a physically intuitive nonlinear flux-linkage model reported by Torrey et al. has been adapted. The torque expression is derived from the flux-linkage expression. However, instead of using flux-linkage data to find the parameters of the flux-linkage model, static torque data have been used to obtain the parameters of the torque model directly. The torque estimator provides very good match for the static measured torque data. It has been experimentally verified that the torque estimator is accurate under dynamic conditions as well.