Electric Vehicles Scheduling Method for Distribution Networks Based on Distributed Proximal Policy Optimisation Considering Weather Uncertainty

ABSTRACT Large‐scale uncoordinated charging and discharging behaviours of electric vehicles (EVs) exacerbate power grid operational losses and reduce the efficiency of scheduling model solutions. To address this issue, this paper proposes an EVs scheduling method for distribution networks based on DPPO. Firstly, considering the operational conditions and security constraints of the distribution network, an optimisation scheduling model is constructed to minimise the total system operational cost. The objective function includes key parameters such as EV charging/discharging power, branch power losses and curtailment rates of wind and photovoltaic (PV) generation. Secondly, to address the computational challenges posed by the high dimensionality of variables, nonlinear constraints, and discrete variables, the optimisation scheduling model is transformed, and the DPPO algorithm is employed for efficient solution. Finally, simulation results based on the IEEE 33‐node distribution network system demonstrate that the proposed method effectively reduces the total system operational cost and network losses, and significantly enhances the distribution network's capacity to absorb distributed wind and PV generation.