含大规模电动汽车参与的智能电网最优双向协调调度机理研究

Wide-scale deployment of Vehicle to Grid (V2G) technology will play an important role in the future smart grid. However, the traditional power dispatch method cannot eliminate the negative effects of plug-in electric vehicles on the power grid. Therefore, new power dispatch methodologies need to be developed to accommodate large-scale integration of plug-in electric vehicles. Under the new dispatch, fleets of plug-in electric vehicles are expected to actively participate in grid operation through their flexible, bidirectional power flow control capabilities. However, due to the mobile, stochastic, and bidirectional nature of plug-in electric vehicles, the mathematical model for the new power dispatch will have gigantic dimensions as well as stronger nonlinearity, complementarity, and randomness, when compared with the traditional power dispatch models. Most of the existing studies focus on modeling, and then simplifying the model and solving it by using some traditional optimization method. This project proposes to create an optimal bidirectional coordinating dispatch model for the next-generation smart grid, study the solvable condition of the model, and develop efficient solution methodologies using Stochastic Semidefinite Programming, Canonical Duality Theory, Convex Optimization and Message-passing decentralized method. Specifically, this research proposes to develop statistical models for the charging and discharging behaviors of plug-in electric vehicles and study the temporal and spacial charge and discharge aggregation of large numbers of electric vehicles and the optimal dispatch mechanism, while taking into account the operating characteristics of the power grid. The findings of the proposed research will provide strong theoretical and technical support for wide-scale deployment of V2G technology and will have positive impact on the development of the next-generation smart grid.

大规模电动汽车是未来智能电网的重要组成部分。传统电力调度方法很难消除电动汽车入网带来的负面影响。因此有必要研究考虑电动汽车参与的电力调度新模式，以发挥其移动分布式电源参与双向调节的作用。但由于其提供的电源具有移动性、随机性和双向性等特点，考虑电动汽车参与的电力调度数学模型不仅维度巨大，且非线性、互补性和随机性更强。现有相关研究侧重于建模，再采用传统优化方法对其简化后求解。本项目将根据电动汽车充、放电行为的统计学模型，研究电动汽车最优充、放电时空聚集粒度的规模和调控机制，结合电力系统运行特性，运用最优化理论最新成果，构建合理的电动汽车大规模接入电网的最优双向协调调度数学模型和定解条件，开展基于随机半定规划、正则对偶理论、锥规划和基于消息传递的分散优化算法的研究。本项目力图从电力系统最优调度的角度出发，为大规模电动汽车接入智能电网提供理论技术支持，并对智能电网理论体系的完善起一定的作用。

从电网角度出发，大规模电动汽车既是负荷，也能作为有限储能装置为电网运行提供辅助服务。本项目主要解决大规模电动汽车与电网双向协调调度问题。由于电动汽车具有移动性、随机性和双向性的特点，考虑电动汽车参与的电力调度数学模型不仅维度巨大，且非线性、互补性和随机性更强。我们根据电动汽车充、放电行为的统计学模型，研究了电动汽车最优充、放电时空聚集粒度的规模和调控机制，以有功网损最小为目标，构建计及日充电需求、线路和配变容量约束的考虑系统潮流及安全运行约束的电动汽车与电网双向协调调度模型，并采用互补内点算法高效求解。考虑电动汽车在线充电数量、充电起始时间及时长、行驶消耗等不确定因素，采用盒式不确定集表示不确定参数，所得决策能保证电力系统的安全稳定运行，提高了系统的安全性。针对电动汽车与电网的双向协调调度模型的采用集中式优化求解算法效率低的问题，我们提出了交替方向乘子法将原优化问题转化为系列子问题进行分散式求解，克服电动汽车充电集中式优化带来的缺陷。进而引入滚动时域控制理论，结合电动汽车的充电特性，考虑各相电压约束、三相平衡约束、配变和线路负载约束，提出电动汽车充放电实时调度优化方法，能对负荷和电价曲线具有很好的跟踪性能，达到计算效率和优化效果的有效平衡，应用前景广阔。综上，本项目通过对电动汽车与电网双向协调调度机理的研究，通过电能调度或动态电价等策略引导并优化大规模电动汽车的充放电过程，在满足电动汽车出行电能需求的同时，利用其储能能力对电网运行提供辅助服务，提高电网运行的灵活性和互操作性。本项目所提电动汽车与电网双向协调调度理论和方法为大规模电动汽车接入智能电网提供了新的理论技术支持，并对智能电网理论体系的完善起一定的作用。