Model predictive current control based on a generalised adjacent voltage vectors approach for multilevel inverters

Abstract

Model predictive current control (MPCC) uses the discrete-time model of a system to predict the future behaviour of the current for all voltage vectors (VVs) generated by a power converter. In multilevel inverters, the large number of VVs imposes a long computation time for the prediction and selection of the optimal state to be applied to the converter, which increases the sampling time and decreases the closed-loop performance. An MPCC is proposed based on the idea of generalised adjacent voltage vectors (GAVVs) for multilevel cascaded H-bridge inverters with a DC-link voltage fed by photovoltaic (PV) cells. This method deals with the voltage drop and often small inter-bridge voltage imbalance and irradiance issues that occur in PV power plants. The proposed GAVV method is analytically formulated to provide three types of subsets for a given number of inverter levels. The use of the newly added subsets of four and five VVs contributes to boosting the converter output voltage and achieving acceptably balanced current and line-to-line voltage under low irradiance compared with the classical approach. Simulation and experimental results show good current response and reduced switching frequency even under a high current reference with DC-link voltage drop.