A Two-Vector-Based Predictive Current Control Algorithm For Three-Level NPC Inverter

Abstract

Based on the traditional finite control set single-vector-based predictive current control (SVB-PCC), a two-vector-based predictive current control (TVB-PCC) strategy is proposed for three-level inverter in this paper. The SVB-PCC strategy of a permanent magnet synchronous motor (PMSM) drive system powered by a three-level voltage source inverter (3L-VSI) applies only single voltage vector during each fixed control period, which causes predictive current tracking control over-adjusted or under-regulated and presents high torque ripple. In this new strategy, two switching sequences are selected in one sampling period to decrease current ripple and balance neutral point (NP) voltage of 3L-VSI. The method performs two cost functions to select the dual optimal voltage vector and the duration of each voltage vector. According to the first cost function, the first optimal voltage vector is obtained. The second cost function is used to evaluate the second optimal voltage vector and their duration, and achieve the multi-objective optimization. According to the characteristic of 3L-VSI drive system, a new predictive model of NP voltage balance, which contains two switching sequences and their on-off time to calculate next period DC-link capacitor voltage, is formulated. In particular, this strategy designs multiple system constraints and control objectives to reduce switch losses and achieve better control performance simultaneously. The simulation results of PMSM system verify the effectiveness and feasibility of the proposed method.