Double Vectors Model Predictive Torque Control Without Weighting Factor Based on Voltage Tracking Error

Abstract

In order to reduce the computation burden and eliminate the weighting factor in conventional model predictive torque control (MPTC), this paper proposes an improved MPTC algorithm without the use of weighting factor. Based on the deadbeat direct torque and flux control, the reference voltage vector, which would be applied in the next period, is predicted to avoid testing all feasible voltage vectors, hence reducing the computation burden of conventional MPTC. Moreover, the torque and flux error-based cost function of conventional MPTC is replaced by the voltage vector tracking error-based cost function; thus, the weighting factor of stator flux in conventional MPTC, which is always necessary because of the different units between torque and flux, is eliminated. Furthermore, two voltage vectors, which include an active vector and a null vector, are applied during one control period to improve steady-state performance. The durations of the selected voltage vectors are determined based on the principle of voltage vector tracking error minimization. Moreover, to further improve the steady-state performance, a different vector selection way, in which the second voltage vector is not fixed as a null vector but selected in a border range, is introduced. The selection process of candidate voltage vectors is analyzed, and the duration is calculated. Simulation and experimental results both show the validity of the proposed control approach.