A Simplified Voltage Vector Preselection-Based Multivector Predictive Current Control for Improved Torque Performance of PMSM Drive

Abstract

Predictive Current Control (PCC) is a pioneering approach for constrained non-linear systems demanding dynamic performance in industrial and transportation applications. This article proposes a multivector PCC scheme based on a simplified voltage vector (VV) preselection for improved torque performance of PMSM drives. The prediction and optimization operations in the basic PCC (BPCC) scheme increase the computational burden of the processor. The proposed method aims to reduce the complexity of the control scheme with a VV-preselection based on an effective position estimation using the gradient concept. This helps to directly determine the optimal VV without evaluating cost-function using all the VVs. However, the application of a single VV in BPCC leads to large undulations in the torque and flux responses. This necessitates the application of multiple VVs in each control period to ensure improved torque response. To achieve this, the duration of application of the VVs is determined using the cost-function ratio of the different VVs chosen in the proposed scheme. This effectively reduces the steady-state torque ripples of the drive while retaining its dynamic response. The performance of the proposed scheme is evaluated through experimentation and compared comprehensively with BPCC and recent literature.