Introduction of Model Predictive Control Strategies for Control of Power Converters

This article provides an overview of previous studies involving the application of Model Predictive Control (MPC) to power converters. In the reviewed MPC designs, cost functions penalizing state tracking errors are utilized, and the optimal weights of this function are systematically determined by solving a linear matrix inequality (LMI)-based optimization problem. Two different types of applications are examined, with distinct methods employed to address uncertainties in each scenario. Firstly, a robust offset-free tracking control scheme is presented for a three-phase DC-AC inverter equipped with an output LC filter, intended for uninterruptible power supply (UPS) applications. The uncertainties associated with the LC filter are represented as potential ranges of capacitor and inductor values. Secondly, a sensorless model predictive control (MPC) strategy is outlined for a grid-connected inverter featuring an inductive-capacitive-inductive (LCL) filter, utilizing only grid-side current measurements. State estimators and disturbance observers are devised based on Lyapunov stability theory to minimize sensor count and eliminate steady-state errors.