The explicit solution of model-based predictive control by considering the nonlinearities in drive applications

Abstract

The main contribution of this work is reflected in the way of dealing with non-linear terms, appearing in AC machine models. Consideration of these nonlinearities in formulating the model predictive control (MPC) problem with offline optimization are normally ignored, considered as an additive disturbance, or linearized around some operating points. In this work, nonlinear terms are considered without linearization and compensated either inner or outer the MPC controller structure in different configurations. The cross-coupling and nonlinearities between state variables in the fundamental model of AC machines arise in term of induced voltages. The effect of these voltages on the MPC current controller increases proportionally by accelerating the machine. Therefore, an unloaded induction machine is introduced in this paper as a case study. Furthermore, a comparative analysis concerning the compensation procedure of these terms is introduced between the proposed methods and other ones in conventional scalar and complex PI-controllers.