An intelligent approach for cascaded multi-level inverter (CMLI) with grid-connected hybrid system

A novel hybrid control method is proposed for cascaded multi-level inverters (CMLIs) in grid-connected hybrid systems. The photovoltaic (PV) and wind turbine (WT) sources are combined in the hybrid system. Each is connected to the cascaded multi-level inverter (MLI)-isolated DC links through its own DC-DC converter. This proposed method combines the ladder spherical evolution search algorithm (LSE) and the recalling-enhanced recurrent neural network (RERNN) method to make the CMLI control signal better. From this, the LSE method dynamically generates an optimal control signal dataset for the CMLI, while RERNN enhances the adaptability and robustness of the control system. The simulation of the proposed LSE-RERNN method is implemented and evaluated using MATLAB software. From that, the simulation outcome demonstrates that the proposed method achieves significant improvements in efficiency and power quality (PQ) compared to existing approaches. Specifically, the switching power losses are reduced to as low as 0.05 W, and the system efficiency reaches 99.6%.