Maximum Power Point Tracking Controller of PV System Based on Two Hidden Layer Recurrent Neural Network

Abstract

Solar energy is one of the most well-known and cutting-edge energy sources in the age of renewable energy. However, because of fluctuating meteorological factors like solar insolation and temperature, the output of a solar photovoltaic system varies greatly. For the effective use of solar energy harvested using solar PV units under different climate factors, the Maximum Power Point Tracking (MPPT) technique is a crucial component that needs to be present. The MPPT system regulates the PV system's output (current and voltage) to give maximal power to the load. Conventional approaches may not efficiently use available electricity and may fail in partial shade conditions. This study describes how to build MPPT for a photovoltaic system utilizing a two-hidden-layer recurrent neural network (THLRNN). The system comprises a photovoltaic module linked to a boost DC-to-DC converter, and the THLRNN algorithm is used in this work to produce the duty cycle to the boost converter that drives the PV voltage to the optimal value. Using the MATLAB/Simulink tools, the suggested algorithm's effectivity has been verified. Furthermore, the outcomes that have been obtained have been compared with other MPPT methods (like improved grey wolf optimization algorithms and artificial neural networks), and from the results that have been obtained it was shown that the proposed technique is superior to other methods and increase the efficiency of PV system by 96.6%. Also, this method has been tested under various environmental conditions (variable irradiation and variable temperature) and found that the photovoltaic system with the proposed MPPT continuously traces the highest power point of the PV module. Additionally, the implementation of this algorithm is simple and can predict the output in a highly efficient way.