Système de Gestion Énergétique Intelligent par l’Intégration de Panneaux Solaires pour une Consommation Optimisée et Durable

Abstract

Electricity consumption management is of crucial importance both environmentally and economically, especially in the context of the transition to renewable energy sources. In this regard, this work presents a comprehensive study on the use of artificial intelligence, and more specifically deep learning algorithms, to predict solar energy consumption and production based on historical and meteorological data. A neural network architecture has been proposed to estimate, on an hourly basis, both the electricity consumption and the production of a photovoltaic panel. The model was trained on time series data, and its performance was evaluated using metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and the coefficient of determination (R2

). The results

obtained confirm the system’s ability to deliver accurate and reliable forecasts. To demonstrate the system’s feasibility and effectiveness in a real-world context, hardware integration was carried out on the embedded platform Raspberry Pi 4 B+. A complete application was developed, allowing visualization of consumption and production curves, recording of results, and simulation of various energy scenarios. This implementation paves the way for the design of an intelligent energy management system capable of making optimal decisions to switch between photovoltaic energy and the power grid based on the predictions. This work was carried out as part of the SWITCH project entitled ’Stabilizing weak grids through machine learning: empowering farmers in end-of-line-communities in North Africa through artificial neural networks’. It is part of the Europe- Africa call for research and innovation on renewable energies, LEAP-RE.