An Improved Duty-Cycle-Based Model Predictive Flux Control for Induction Motor Drives

Model predictive control with duty-cycle optimization has been widely adopted for induction motor drives. However, a fixed combination of a null vector and an active vector fails to guarantee the global optimal steady-state performance in the full speed range. Furthermore, recent study reveals that the pulse pattern also has influence on the current harmonics and switching frequency, even if the same voltage vectors are applied. Aiming at solving these issues, an improved duty-cycle-based model predictive flux control is proposed in this paper. Firstly, to avoid the high complexity caused by the introduction of duty cycle control, the reference voltage vector is directly calculated by deadbeat flux control with space vector modulation (SVM). The optimal vector combination can be easily obtained from SVM, which includes the combination of two active vectors, thereby reducing the number of vector combination evaluations and enhancing the steady state performance at high speeds. Secondly, to achieve minimum harmonic current and flux ripples, an optimal pulse pattern is employed, which places the active vector at the beginning and end of a vector sequence, and the middle vector may be a null vector or an active vector. Thirdly, the distribution ratio of the first vector is further optimized to minimize the flux ripple, leading to further performance improvement. Finally, the dynamic adjustment of vector sequences ensures a lower switching frequency. The experimental results confirm that the proposed control strategy has lower current THD than that of prior arts under similar switching frequency. The current THD can be reduced by nearly 60% at the rated speed with rated load.