Effective model predictive current control for a sensorless IM drive

Abstract

This paper introduces an innovative sensorless model predictive current control (MPCC) technique for an induction motor (IM) drive. The proposed technique handles the state of the inverter as a discrete control variable in contrast to the classic vector control in which the discrete model of the inverter is not considered. An effective formulation of cost function is used, in which there is no need for tuning the value of a weighting factor that is usually used in conventional forms, and thus saving computational time. In addition, the formulation introduces a way to investigate deeply the operation base of model predictive current control procedure. A sensorless approach is used, based on an improved Luenberger observer in which stator current, rotor flux, speed and load torque are estimated. Firstly, the dynamic performance of the drive is tested by simulation using Matlab/Simulink software. The experimental validation for the new procedure is performed utilizing a dSpace1104 control prototyping board. The feasibility of the proposed control technique is realized through the obtained measurements, which prove a precise transient dynamic performance of IM drive for the given speed commands.