Enhancing Smart Grids for Sustainable Energy Transition and Emission Reduction with Advanced Forecasting Techniques

Abstract

Smart grids are modernized, intelligent electricity distribution systems that integrate information and communication technologies to improve the efficiency, reliability, and sustainability of the electricity network. However, existing smart grids only integrate renewable energies when it comes to active demand management without taking into consideration the reduction of greenhouse gas emissions. This paper addresses this problem by forecasting CO2 emissions based on electricity consumption, making it possible to transition to renewable energies and thereby reduce CO2 emissions generated by fossil fuels. This approach contributes to the mitigation of climate change and the preservation of air quality, both of which are essential for a healthy and sustainable environment. To achieve this goal, we propose a transformer-based encoder architecture for load forecasting by modifying the transformer workflow and designing a novel technique for handling contextual features. The proposed solution is tested on real electricity consumption data over a long period. Results show that the proposed approach successfully handles time series data to detect future CO2 emissions excess and outperforms state-of-the-art techniques.

Keywords

• smart grid, short-term load forecasting, carbon dioxide, electrical consumption, deep learning model, harmful energy