Bounding the torque ripple in switched reluctance motors using polyhedral invariant set theory

For a switched reluctance motor (SRM) torque control we propose an approach based on constrained predictive control. In particular, the highly nonlinear characteristic of the motor torque is approximated in a piecewise affine (PWA) form. Based on this approximation a simple PWA model of the SRM is constructed and used for model-based torque controller design. The user-specified maximum torque tracking error allowed is used to determine the set in the space of motor states and torque references within which we compute the polyhedral maximum controlled invariant set for control.