Project "Smart Agent For Efficient Charging" (SAFEC) aims to revolutionise the electric vehicle (EV) charging process by deploying advanced AI techniques to make charging decisions in real time, thereby reducing costs and supporting environmental sustainability. This project is a collaboration between Gridicity, an EV energy management software company, and TechnoQuest, an AI development firm. The project's novelty resides in the comprehensive optimisation of EV charging in real time using AI. The primary challenge SAFEC addresses is to manage EV charging times efficiently, accounting for variable energy tariffs and multiple other relevant factors. These range from recurring seasonal and daily patterns in grid demand, day-ahead and real-time energy market prices, fleet size, driving patterns, to current and required states of charge for each vehicle. Optimising these factors could result in substantial cost savings and productivity gains for EV fleets. While smart energy management solutions exist in the domestic EV sector, they are not yet widely applied in the commercial EV fleet and public charging network sectors. SAFEC not only represents an economic opportunity but also aligns with the UK government's net zero emissions goals. By shifting power demand to off-peak times when cheaper, often renewable energy is available, SAFEC prevents excessive fossil fuel usage, contributing to a greener, more sustainable future. Additionally, by reducing the operating costs of EV fleets, SAFEC encourages faster EV adoption, further supporting the shift towards sustainable transportation. Overall, SAFEC harmoniously integrates economic and environmental objectives through innovative AI applications in EV charging.

This project investigates the feasibility of a functional vehicle-to-X operating framework that minimises the energy costs of the microgrid using EVs as storage for temporary mismatches between demand and supply. The relevance of the project resides in the increasingly higher frequency of these mismatches because of the intermittent nature of renewables and the high variability of the demand, as well as the new potential revenue stream from provision of ancillary services using V2X technology. To investigate this issue, data will be collected to accurately predict EV plug-in times and energy demand in the day-ahead horizon using Artificial Intelligence. Optimal load balancing, minimising energy cost and maximising revenues from the provision of ancillary services for the next day, will then be evaluated using these accurate predictions. V2X DC microgrid infrastructure will be the basis for testing of use cases and demonstration of the technology. The combination of predictions, management platform, hardware infrastructure and testbed allows feasibility testing of an end-to-end V2X solution for fleets in a microgrid.