城市公共区域电动汽车充、换电站混合规划建模及其优化方法研究

Electric vehicle industry is developed prosperously in the past two decades, due to the rapidly shrinking of fossil fuel and the increasing of vehicle exhaust emissions. Building up sufficient changing or exchanging electric power battery facilities and taking them into function is a precondition to develop this industry in a sustainable way. The basic key problem within this field is to develop a universal optimization model for the charging or exchanging electric power battery facilities and solve it efficiently. A hybrid integral charging/exchanging station for electric power battery is built up in this study, in which the statistical population streams in public regions are analyzed to help us building up the electric vehicle energy supply prediction model, the topology of current distribution network, the charging period that the driver will take in one service of recharging his/her vehicle battery, and the harmonic wave limitation of an energy supply station for electric vehicle are considered as the constraints for the model. Besides, other factors, like driving behaviors or the financial investigation for building up a charging/exchanging station are also taken into account within the model. In order to solve this global optimization problem in an efficient way, a hybrid multi-objective evolutionary algorithm based on different hybridization coding strategy, specific designed evolutionary operators, escalating population renew strategy and meta-heuristic neighbor-hood search strategy, is proposed in the study. The employment of our proposed hybrid model and improved evolutionary algorithm in a proper way may lead to a coordinate equilibrium Pareto set solutions, which can guide the corresponding decision maker to build up an optimal electric vehicle energy supply station with his/her own preference.

随着传统化石能源日益枯竭、城市汽车尾气排放日益加剧，推动电动汽车产业迅速发展，充、换电站等基础设施的规划与建设运营，是实现电动汽车产业规模化和可持续发展的必要前提。如何建立统筹全局的整体规划模型，并进行有效求解，是该领域亟待解决的应用基础问题。本项目基于电动汽车能量补给需求预测模型，对现有配电网路空间分布、电能容量约束、谐波污染等电网侧技术约束条件与电动汽车行驶/续航里程、能量补给时间等因素对用户交通需求满足度的非技术约束条件进行融合，建立以搜索电动汽车充、换电站最佳选址-定容决策组合为目标的整体优化模型。提出以群体递进进化结构为框架、异构分层编码结构为载体、定制进化操作算子为寻优手段、启发式修复策略和交互式动态适应度评价策略为决策偏好导向的进化算法体系。在保证算法寻优效率的前提下，实现城市公共区域电动汽车充、换电站规划问题的多个目标协同改进，为电动汽车基础设施建设提供理论指导。

随着传统化石能源的日益枯竭、城市汽车尾气排放日益加剧，推动电动汽车产业迅速发展，充、换电站等基础设施的规划与建设运营，是实现电动汽车产业规模化和可持续发展的必要前提。.本项目主要工作包括：.一、建立了城市公共区域电动私家车、电动出租车的充电站规划模型，综合考虑满足充电需求及配电网约束前提下，同时优化建站投资与运营投资的经济成本最小化、设备利用率最大化、用户满意度最大等目标；建立了现有充电网络基础上进一步开展新增/扩容充电站系统的选址、定容规划模型研究。.二、提出基于分层编码表达电动汽车充电站规划的进化算法染色体方案，并提出相应的解码及进化操作算子保证解的合法性和有效性。.三、建立了基于Anylogic平台的电动汽车购买、充电、运行的仿真模型，基于该模型开展了电动汽车充电站性能评价及政府对充电电价分时补贴的策略研究。.四、以深圳、北京等城市为例，开展电动私家车、电动出租车公共充电站的规划、运行及评价的实证研究，验证方法的有效性。.研究结果表明：本项目研究建立的规划模型及求解方法可以较好地解决电动汽车充电站的选址、定容规划和二次规划/重规划问题；所搭建的仿真模型可以较好地展现电动私家车出行随机的情况下充电站满足使用需求的程度，结果更加具备实用性。.本项目建立的方法已应用于某央企和某地方电力公司十三五期间的电动汽车充电站规划专项计划地制定。后续仿真、优化、博弈综合模型的完善，可为我国大城市公共管理部门、国家部委或电网公司建立电动汽车充电网络提供决策依据和支撑，促进节能与新能源汽车产业的顺利发展。