Predictive Control of Grid-Connected Multilevel Inverter with Output LCL Filter

The paper deals with a control of a multilevel grid connected photovoltaic inverter with an LCL output filter. A proposed control technique uses the dynamical model of the LCL filter and ensures inherent active damping of the LCL filter oscillations. The inverter is a one-phase cascade H-bridge inverter with three galvanically separated DC sources and the model predictive control is used as a control technique. A Kalman observer is used to minimize the number of required sensors. Only the grid current and the grid voltage measurements are needed. The model predictive control ensures active damping for the LCL filter. This removes the need for a passive damping of the LCL filter and increases overall system efficiency. The design of the model predictive controller as well as the design of the Kalman observer is presented. The proposed control technique is verified by simulations and measurements on the laboratory model.