Efficient maximum power point tracker based on neural network and sliding-mode control for buck converters

Abstract

Abstract This paper presents detailed design steps of an effective control system aiming to increase the solar energy harvested via photovoltaic power-generation systems. The design of an intelligent maximum power point tracker (MPPT) supported by a robust sliding-mode (SM) controller is discussed in this study. The proposed control scheme is designed to track the MPP and provide a smooth system response by removing the overshoot in the load current during any variation in the connected load. Such a system is suitable for DC–DC buck converter applications. The study begins with modelling the buck converter for a continuous current mode operation. The reference voltage of the tracking system is produced by the proposed neural network (NN) algorithm. The proposed intelligent MPPT integrated with an SM controller is simulated in a MATLAB®/Simulink® platform. The simulation results are analysed to investigate and confirm the satisfaction level of the adopted four-serially connected PV-modules system. The system performance is evaluated at a light intensity of 500 W/m2 and an ambient temperature of 25°C. Applying only the proposed NN algorithm guarantees the MPP tracking response by delivering 100 W at a resistive load of 13 Ω, and 200 W at a load of 6.5 Ω, respectively, with 99.77% system efficiency. However, this simultaneously demonstrates a current spike of ~0.5 A when the load is varied from 50% to 100%. The integrated SM controller demonstrates a robust and smooth response, eliminating the existing current spike.