Emission-Aware RL for Sustainable EV Charging and CO2 Reduction

A new research paper proposes an emission-aware reinforcement learning (RL) strategy for electric vehicle (EV) charging that prioritizes carbon reduction. The method uses the Soft Actor Critic (SAC) algorithm with a multi-objective reward penalizing carbon emissions, curtailed renewables, and unmet demand. Trained on the EV2Gym platform with behind-the-meter solar and wind profiles and EirGrid carbon intensity data, the approach aims to reduce CO2 while managing grid stability. Existing methods like MPC and standard RL rarely treat real-time carbon intensity or renewable availability as primary objectives, leaving decarbonisation potential untapped.