An Efficiency Optimized Direct Model Predictive Control of Induction Machine Drive

Abstract

Presented in this paper is a direct model predictive control (MPC) for directly controlling the rotor speed and flux of an inverter-fed induction motor, which is also applicable to induction generators. Discrete dynamic model of the induction machine is applied in stationary reference frame with pre-set present and past reference values of the speed and rotor flux for generating the stator current references The uniqueness of the MPC proposed herein is that it does not require outer cascade proportional plus integral (PI) controllers used for regulating rotor speed and flux. Elimination of the outer loop cascade is made possible in this article by the introduction of a unique four-term cost function comprising of weighted magnitudes of the errors between four measured state variables and their reference commands. The four state variables in the unique cost function are the rotor speed, the rotor flux linkage, quadrature axis and direct axis stator currents. The MPC minimizes the cost function at every switching instance by selecting the switching states that gives the least cost function. Consequently, the selected optimal switching states are used to switch optimal inverter output voltages across the machine’s stator terminals. As a precursor for obtaining an optimal rotor flux command in the MPC for every torque output, another unique efficiency improvement scheme is developed, which uniquely determines the stator angular velocity and the rotor flux that minimizes the core and copper losses while maintaining a constant slip operation. Therefore, efficacy of the proposed direct model predictive control scheme is verified by comparing its results to results obtained from equivalent vector control on equivalent machine. Results presented show lower loss regime with MPC on optimal stator angular velocity and rotor flux than vector control.