Improved Model Predictive Control with Extended Horizon for Dual Inverter with Real-Time Minimization of Converter Power Losses

摘要

The paper is focused on reduction and balancing of power losses in a dual inverter. The proposed technique is based on finite control set model predictive control with two techniques which improve the behavior of the basic predictive algorithm to deal with different time constants in the system. The proposed algorithm uses approximation of the power losses by an exponential filter and furthermore, it combines this approximation with extended prediction horizon. The computational burden of the prediction horizon is reduced by the switching states preselection. The behavior of the system is then significantly improved with respect to balancing of power losses among the transistors and with respect to quality of the output current.