Advanced Forecasting Techniques for Smart Grids to Enhance Energy Efficiency and Sustainability

The rapid increase in global electricity demand, driven by technological advancements and the electrification of various sectors, demands effective energy forecasting to ensure sustainable and proactive energy management. This study marks the first stage in the development of a smarter energy management solution, which is based on cooperative intelligence distributed at the smart grid levels, involving the entire chain from producers to consumers. Consequently, we conducted a comparative analysis of three advanced forecasting models to determine their effectiveness in predicting energy consumption. Using a real dataset, collected in a student residency in Bologna, Italy, which includes weather and energy consumption data, the models were evaluated using metrics such as mean absolute error (MAE), mean squared error (MSE), root mean squared error (RMSE) and R-2. The results indicated that the LSTM model outperformed the others, demonstrating superior accuracy in capturing temporal dependencies. The Transformer model also showed robust performance, highlighting its potential in scenarios that require the capture of long-range dependencies. This study sets the stage for further research and development to leverage advanced forecasting techniques to optimize energy distribution and ensure the sustainability of energy systems.