Generalized Two-Vector-Based Model-Predictive Torque Control of Induction Motor Drives

Abstract

Conventional single-vector-based model-predictive torque control (MPTC) has been widely studied owing to its intuitive concept and quick response. To improve the steady-state performance, recently, the concept of duty cycle control was introduced in MPTC by inserting a null vector along with an active vector during one control period. However, this still fails to reduce the torque error to a minimal value due to the imposed restriction on vector combination and the cascaded processing of vector selection and vector duration. This paper proposes a generalized two-vectors-based MPTC (GTV-MPTC) by relaxing the vector combination to two arbitrary voltage vectors. By evaluating the vector combination and their durations simultaneously in the predefined cost function, global minimization of torque error can be obtained in theory. However, the computational burden is also significantly increased. By choosing a proper method to determine the vector durations, the redundant vector combinations can be eliminated, which makes the proposed GTV-MPTC suitable for real-time implementation. Both simulation and experimental results were carried out to verify the effectiveness of the proposed method. The presented results show that, compared to prior MPTC with or without duty cycle control, the proposed GTV-MPTC achieves much better performance with lower sampling frequency over a wide speed range. Furthermore, the average switching frequency is even lower than that of conventional MPTC in the medium speed range.