Smart Forecasting of Photovoltaic Panel Power Output: A Neural Network Approach with an Interface Based on Arduino Board

Abstract

The use of photovoltaic solar energy seems to be an essential asset and a necessity for the future with advantages such as abundance and availability. However, the outdoor exposure of photovoltaic modules brings together a complex combination of factors which cause their degradation. This results in a negative impact whether on the efficiency or on the characteristics of the panels, which means that even for the optimal temperature and radiation conditions, we will have electrical characteristics different from those of the panel nameplate, the case maximum power for example. The objective of this paper is to realize a system capable of forecasting the maximum power delivered from the photovoltaic field, knowing only the value of the temperature and the value of the solar radiation, whatever its state of degradation. We propose an approach based on an artificial intelligence technique. The acquisition circuit, implemented cost-effectively, will be detailed in this work. The results obtained will be also shown, validating the proposed approach. This integration of AI with renewable energy systems promises to enhance their efficiency and reliability.