An Advanced Model Predictive Controller for Grid-Tied Four-Leg Multilevel Inverters

An advanced Model Predictive Control (MPC) for three level diode clamped four leg Inverter is Introduced by exploiting its space vector representation in natural coordinates. The advanced MPC predicts reference inverter voltage (RIV) vector at every sampling instant to simplify the implementation of MPC. In this procedure, every switching state is isolated from each other by its optimal volume to control grid current. The optimal volume of the switching state is identified by predicting the RIV vector. The identification procedure employs the simple coordinate geometry concept. It does not require the calculation of the objective function. A generalized advanced MPC for any four leg grid-tied multilevel inverter is developed as well. Simulations are carried out for the advanced MPC. A detailed comparison with the conventional MPC is included. Preliminary experimental results of the advanced MPCC are also presented.