基于多智能体系统的电动汽车分布式调控模型

Distributed Control Model Based on Multi-Agent System for Electric Vehicles

对接入需求侧的电动汽车充电进行控制和管理,能够改善配网的负荷供需平衡,有利于提高系统可靠性和经济性。针对智能电网背景下的电动汽车需求侧响应充电模式进行研究,构建了面向电动汽车运营系统的多智能体系统体系框架,在界定各智能体的学习、适应行为的基础上,以配网负荷限制、电动汽车荷电状态和用户电费成本为优化目标,研究电动汽车响应峰谷电价的协调机制,并建立多智能体仿真模型。通过一个仿真算例对比研究了无电动汽车接入、盲充和引入多智能体系统三种情景下的电网负荷情况,研究结果表明,基于多智能体系统的电动汽车充电控制能够降低负荷峰谷差,并且能够兼顾电动汽车用户的经济利益和用电习惯,有利于实现相关主体的互利共赢。

Control and management of electric vehicle （EV） charging at demand side can achieve supply-demand bal- ance of the distribution network load, and help to improve system reliability and economy. Based on the research of EV demand-side response charging mode in the context of smart grid, a multi-agent system （MAS） framework for EV char- ging operating system is built. Following with the definition of agent learning and adapting behavior, the coordination mechanism of EV in response to the peak-valley price is proposed based on optimal goals of the distribution network load limits, EV charging state and electricity price cost. The load curve of the grid in three scenarios has been analyzed, i. e. non EV connection, blind charging and multi-agent system application. The results show that EV charging system based on muhi-agent control model can reduce peak load, and take into account the economic interests and electricity habit of the EV users, which is conducive to realizing the mutual benefit and win-win results of related subjects.