Artificial Neural Network-Based Control and FCS-MPC for DC/DC and CHB Multilevel Converters for PV Water Pumping Systems

Abstract

This paper explains an artificial neural network (ANN)-based control for photovoltaic (PV) system with dc/dc boost converters connected in cascaded with a single-phase cascaded H-bridge (CHB) 9-level inverter with finite control set (FCS) model predictive control (MPC) for rural water pumping with speed control. First, a digital PI controller for each of the four boost converters is designed, then this controller is used as a supervisor to train the ANN, which is designed with the help of Matlab/Simulink software. Once each dc-bus voltage is regulated, the CHB 9-level inverter with the FCS-MPC algorithm is utilized to control the current and the speed of an induction motor water pump. The simulation results show that the ANN-based control scheme for the boost converters improves the voltage control performance compared to the PI, and an excellent behavior of the water pump is achieved with the FCS-MPC scheme developed for the CHB multilevel inverter