Volume 38: Energy Transitions toward Carbon Neutrality: Part I

Battery thermal-health jointly concerned charging scheduling for Solar PV penetrated Energy-Transportation Nexus: a DRL-based approach empowered by a Cyber-Physical system Xuyang Zhao, Hongwen He, Jianwei Li, Zhongbao Wei, Ruchen Huang, Hongwei Yue

Abstract

Using effective vehicle-to-grid (V2G) strategies, the onboard batteries in grid-connected electric vehicles (GEVs) can be leveraged to alleviate the impact of solar photovoltaic (PV) systems and provide grid support. Nevertheless, the abuse of batteries during V2G is inevitable owing to balancing the fluctuation of solar power while ensuring charging effectiveness, resulting in risks on battery rapid degradation and thermal safety. Regarding this, a multi-physics-constrained charging scheduling strategy is proposed in this study, enabled by a novel deep reinforcement learning (DRL) technique to mitigate solar PV impact while minimize the expected customer’s charging cost, including energy cost and battery aging cost as well as satisfying the customer service quality and battery operation safety constraints. The proposed strategy is further performed within a cyber physical system-based framework, where the complicated training is carried out in the cloud, while the trained low-complexity policy is executed in the onboard controller to mitigate high computing burden. The effectiveness of the proposed strategy is verified by hardware-in-Loop tests and practical battery charging/discharging experiments combined by a real distribution system in Australia.

Keywords renewable energy resources, grid-connected electric vehicles (GEVs), deep reinforcement learning (DRL), battery health management, thermal safety