面向综合效益的城市充电网点规划研究
2021-03-22张家浩李国庆刘琳余彬龙莉娟陈杰
张家浩 李国庆 刘琳 余彬 龙莉娟 陈杰



摘要:随着国家大力发展新能源汽车的政策,电动汽车和充电基础设施规模不断扩大,如何科学地规划充电网点成为亟待解决的问题。本文根据分期规划的原则,采用基于注意力机制的长短期记忆网络进行充电需求预测,使用量子粒子群优化算法进行近期充电网点选址定容,然后根据Voronoi图来划分远期充电站服务区域,从而建立了充电网点精准规划模型。通过对杭州市钱江世纪城区域进行验证,表明本文模型在科学规划充电网点的同时,能够实现电网、企业、用户的多方共赢。
关键词:充电网点规划;长短期记忆神经网络;量子粒子群优化;Voronoi图
中图分类号:TP183 文献标识码: A
文章编号:1009-3044(2021)06-0215-03
Abstract: As the country vigorously develops new energy vehicles, the scale of electric vehicles and charging infrastructure continues to expand. How to scientifically plan charging outlets has become an urgent problem to be solved. According to the principle of staged planning, this paper adopts an attention mechanism-based Long Short-Term Memory Neural Network to predict charging demand, uses Quantum Particle Swarm Optimization algorithm to select the location and capacity of charging network, and then divide the long-term charging station according to the Voronoi diagram. The service area has thus established a precise planning model for charging network points. The verification of the Qianjiang area in Hangzhou shows that this model can achieve a win-win situation for power grids, enterprises, and users while scientifically planning charging points.
Key words: Charging Network Planning; Long Short-Term Memory Neural Network; Quantum Particle Swarm Optimization; Voronoi Diagram
隨着电动汽车用户数量不断增加,如何科学地规划和管理充电网点亟待解决的问题。传统的电动汽车充电网点规划主要从电动汽车用户充电的便利性或者是充电基础设施运营商的效益角度出发。文献[1-2]考虑用户出行需求,以最小化电动汽车用户空驶距离为目标进行选址。文献[3]以建设投资和运行成本最低为目标进行规划,文献[4]考虑车网互动技术(Vehicle to Grid,V2G)和分时电价,以充电站年总收益最大为目标进行选址定容。文献[5-8]兼顾用户和运营商利益,规划结果不仅可满足用户出行需求,还能提高运营商的利益。但文献[1-8]都没有考虑到电动汽车充电站与电网之间的关系。
因此,本文对多源异构数据进行价值挖掘。首先利……
