Improved Predictive Control for Induction Motor Drives Fed by Cascaded H-bridge Multilevel Inverters

Abstract

The cascaded H-bridge multi-level inverter (CHB) is an effective solution for three-phase induction motors (IM) because the voltage has multiple levels resulting in much reduced voltage growth per winding dv/dt, low voltage harmonics. Previous studies have used model predictive control (MPC) for CHB inverters. The predictive control algorithms of traditional finite control set MPC (FCS-MPC) often have steady-state errors when operating at low sampling frequency or inappropriate parameters in the prediction model, by which, the existance of system model deviation leads to the static error. Therefore, this paper proposes a method to build a new current prediction equation with an integral component for the purpose of compensating for model error. This alternative is tested with a cascaded H-bridge multilevel inverter system for IM. The simulation results show that this improved method has eliminated the static error of the motor current on the rotational coordinate system dq, but still ensures the dynamic characteristics of the transmission system and the optimization objectives of switching and common-mode voltage.