Model predictive control with non-linear feedback for permanent magnet synchronous motor drives

A linearized model of permanent magnet synchronous motor drives is developed in this paper by using non-linear feedback. The conventional motor model is transformed, and new state variables are used for decoupling the mechanical and electromagnetic parts. The developed model is then incorporated into the control law of a finite-state model predictive controller (MPC), which enables it to predict the new state variables and select the switching combination that satisfies all the control objectives with minimum cost. When combining MPC with non-linear control, the decoupling is more effective due to the response of controller that allows fast transient characteristics. The performance of the developed controller is compared with that of a cascaded PI control structure by considering reference speed and load torque changes.