Artificial Neural Network-Based Control of a DC/DC-DC/AC 5-Level Inverter for Rural Power Generation

In this article, the performance of an artificial neural network (ANN)-based control for a PV stand-alone system with DC/DC Boost converter connected in cascade with a 5-level DC/AC inverter is evaluated. First, a classical digital PI controller for the Boost converter is designed to be able to make comparisons. In addition, this controller is used as a supervisor to train the ANN, which is designed with the help of Matlab/Simulink software. Once the DC-link voltage is regulated, a single-phase 5-level cascaded H-bridge inverter is utilized to obtain an AC of constant frequency at the output of the inverter. Simulated results verify that the ANN-based control scheme improves the performance of the power converter compared to the PI one, and can be used to power electrical loads in rural areas where the power grid is not available.