Weightless Model Predictive Torque Control of Induction Motor with Simplified Candidate Voltage Vectors Set

Abstract

In order to reduce the torque ripple, flux ripple and calculation burden of the conventional model predictive torque control (MPTC) of induction motors (IM), and avoid the design of weight coefficients, a weightless MPTC method with simplified voltage vectors (VVs) based on deadbeat (DB) is proposed. The duty cycle of the VVs is calculated through DB method, and then the MPTC link is performed after VVs are corrected to select the optimal VV. Based on the DB-MPTC, the symmetrical distribution of the duty cycle with the VVs is used to reduce the number of non-zero VVs traversed by half. Secondly, the cost function in the algorithm only contains the flux term, and has almost no effect on the control performance, thus eliminating the weight coefficient and further simplifying the MPTC model. Finally, the influence of extending VVs on DB-MPTC is briefly studied. The experimental results prove that through the method, the torque ripple is improved by 64.92%, flux ripple is improved by 52.24%, current total harmonic distortion (THD) is optimized from 12.68% to 5.60%, the design of weight coefficient can be effectively avoided, and the calculation burden is also reduced by 48.61%, so the control and real-time performance is improved.