Performance improvement of induction motor drives with model-based predictive torque control

Authors: FATİH KORKMAZ

Abstract: One of the most important advantages of using modeling and simulation software in design and control engineering is the ability to predict system behavior within specified conditions. This paper presents a novel error vector-based control algorithm that aims to reduce torque ripples predicting flux and torque errors in a conventional vector-controlled induction motor. For this purpose, a new control model has been developed that envisages flux change by applying probabilistic space vectors' torque and flux control. In the proposed predictive control algorithm, flux and torque errors are calculated for each candidate voltage vector. Thus, the optimal output voltage vector that minimizes the error is determined for the next sampling time. A MATLAB/Simulink model of the proposed control algorithm was formed and experimental studies were conducted to test the applicability and effectiveness of the proposed method using a dSpace 1104 controller board. The results obtained by comparing the effects of the conventional torque control and the proposed predictive direct torque control methods on the motor performance are presented under various speed and loading conditions. The experimental results prove that the proposed algorithm reduced the torque ripples of the motor remarkably when compared to the conventional torque control. The proposed method stands out with its simple structure due to simplified cost function with the flux model prediction approach and easy applicability with satisfactory results.

Keywords: Variable speed drive, predictive control, induction motor, machine vector control, direct torque control

Full Text: PDF