A multi-output deep learning model for energy demand and port availability forecasting in EV charging infrastructure

Electric vehicles (EVs) play a crucial role in the transition towards sustainable urban mobility. Nevertheless, the swift escalation in EV adoption necessitates accurate forecasting of energy demand and port availability to optimize charging infrastructures and bolster grid reliability. This study unveils a multi-output deep learning framework that concurrently anticipates energy consumption and charging port availability in realistic public EV charging contexts. The research amalgamates open data, advanced feature engineering, and multi-output learning methodologies to realize precise short-term predictions. A comparative analysis with single-output deep learning and transformer models reveals the superior efficacy of the proposed framework, which achieves notable reductions in forecasting inaccuracies and improved classification precision. The model's robustness is further tested under unexpected conditions, such as the COVID-19 pandemic, and across diverse spatial and operational contexts. The findings underscore the potential of this approach to bolster urban energy systems, facilitating informed decision-making for urban planners, policymakers, and grid operators. A conceptual framework is proposed to integrate EV charging infrastructure with broader urban energy systems, offering a scalable and interdisciplinary strategy for optimizing energy demand and improving system resilience. These findings showcase the role of predictive analytics in driving low-carbon urban development and effective energy management.