Predictive Torque Control Algorithm for a Five-Phase Induction Motor Drive for Reduced Torque Ripple With Switching Frequency Control

Abstract

This article presents a progressive comparison of predictive torque control (PTC) algorithms for a five-phase induction motor fed with a two-level five-phase inverter. The extension of three-phase PTC to the five-phase motor, introduces current harmonics of the order 10n $\pm$ 3 (n = 0, 1, 2.. .), which are not responsible for torque/flux production. The inherent disadvantage of PTC is being a variable frequency algorithm, which adds complexity to the design of magnetic components. This disadvantage of PTC can be overcome by applying dwell time ( $t_{s}$ ) to the voltage vectors obtained through a minimum torque ripple condition, which is determined based on the ripple equation. In the proposed algorithm, a set of synthetic voltage vectors are generated and used for reducing flux and torque error through cost function. The use of synthetic voltage vector simplifies the cost function and further reduces the computation time. A two-step delay compensation is adopted to improve the precision of the control algorithm further. In each control cycle, an optimal switching time is calculated to reduce the torque ripple and to maintain the constant switching frequency. In this article, a modified PTC is proposed, which targets to achieve constant switching frequency, eliminate the xy-subspace current harmonics, reduce the torque ripple, and minimize the computational burden involved in the conventional predictive torque control algorithm. The experimental study is carried out on a laboratory prototype. The results show compliance with all the experimental study with different PTC algorithms with different sampling frequencies, and the modified cost function.