CHARGING PRICING METHOD OF ELECTIC VEHICLE CONSIDERING COORDIRATIED OUTPUT OF RENEWABLE ENERGY
陆斯悦 1及洪泉 1张禄 1徐蕙 1王培祎1
扫码查看
点击上方二维码区域，可以放大扫码查看
作者信息
1. 国网北京市电力公司 北京 100075
折叠
摘要
在一个配电网和城市交通网耦合框架中,提出一种电动汽车充电定价方法.建立以社会总成本最小为目标的电动汽车充电服务费的双层优化模型,模型上层为在含风电的配电网中求解充电服务费(Charging Service Fees,CSF)的二阶锥问题;下层为一个遵循用户均衡(User Equilibrium,UE)原则的交通分配问题.该模型考虑了风电输出和OD交通流的不确定性,建立基于深度强化学习的求解框架,对随机双层问题进行解耦和近似求解.以5节点系统和某城市交通-电力耦合网为例,验证了该模型的有效性.
Abstract
A charging pricing method for electric vehicles is proposed in a coupling framework of distribution network and urban transportation network.In this paper,a two-level optimization model for charging service charge of electric vehicles was established with the objective of minimizing the total social cost.The upper layer of the model was to solve the second-order cone problem of charging service fees(CSF)in the distribution network with wind power,and the lower level was a traffic assignment problem following the user equilibrium(UE)principle.Considering the uncertainty of wind power output and OD traffic flow,a solution framework based on deep reinforcement learning was established to decouple and approximate solve the stochastic bilevel problem.The effectiveness of the model is verified by a 5-bus system and a city traffic power coupling network.
关键词
深度强化学习/配电网/交通网/电动汽车充电费用/交通用户均衡
Key words
Deep reinforcement learning/Distribution network/Transportation network/Electric vehicle charging fees/Traffic user equilibrium
维普
万方数据