双侧互动模式下电动汽车光伏充电站的在线优化方法

Under the background of smart grid, intelligent transportation system and vehicular network, the optimal operation mode of photovoltaic (PV) charging stations has gradually developed from self-discipline to online synergetic interaction. This project mainly focuses on the online optimization method of PV charging station under the interaction with smart grid and EV users. The decomposition modeling, real-time decision-making algorithm, scheduling strategy and driving mechanism of optimal problems are set as the breakthrough points of this research. In terms of optimization model, for “station-grid” and “station-vehicle” interaction modes, online optimization models under scenarios such as automatic demand response, online ancillary service and online incentive are respectively established. The association relationship and selection mechanism among models about issues such as optimization goal, decision variables and constraint conditions is revealed. As for online algorithm, theoretical methods such as off-line optimization, cognitive rules, machine learning, data fusion and human action are synthetically considered. The framework of online algorithmic theory for optimal operation of PV charging station is proposed. In terms of scheduling strategy, scheduling framework of online energy management based on real-time state machine model is established. Different scheduling strategies under multiple scenarios are proposed. The double scheduling and driving mechanism that combines time-driven with event-driven is established. The research achievements would promote the extension and application of online optimization method in smart grid and support the development of online energy management system on the user side.

随着智能电网、智能交通系统和车联网的发展，光伏充电站的优化运行逐渐由自律模式发展为在线的协同互动模式。本课题重点研究双侧互动模式下光伏充电站的在线优化方法，以优化问题的分解建模、实时决策算法、调度策略及驱动机制等三类难点问题作为研究的切入点。在优化模型层面，针对“站网”和“站车”双侧互动模式，分别建立自动需求响应、在线辅助服务、在线用户激励等多种场景下的在线优化模型，揭示模型之间在优化目标、决策变量、约束条件等因素上的关联关系与选择机制。在在线算法层面，综合考虑离线优化、认知规则、机器学习、数据融合、人因行为等的理论方法，提出适用于光伏充电站优化运行的在线算法理论架构；在调度策略层面，构建基于实时状态机模型的在线能量管理调度框架，提出多场景下的差异化调度策略，建立时间和事件驱动结合的重调度驱动机制。研究成果将促进在线优化方法在智能电网中的发展和应用，支撑用户侧在线能量管理系统的研发。

本课题的研究基本按照申请书的计划内容及时间周期执行，重点针对三类问题展开研究，包括：（1）光伏充电站的服务策略与质量评价指标；（2）光伏充电站的站级在线优化方法；（3）光伏充电站的集群在线优化方法。针对第一类问题，建立了一套计及服务质量、环境/经济效益、对电网的影响等指标的光伏充电站有效性评价体系，提出了一种以提高服务质量和促进光伏电源的消纳为目标的光伏充电站充电策略。针对第二类问题，提出了光伏充电站的自动需求响应方法、在线辅助服务方法和在线能量优化方法。针对第三类问题，提出了基于非合作博弈和主从博弈两种框架下的光伏充电站集群在线优化方法，提出了基于李雅普诺夫优化的集群实时在线优化方法。通过上述研究，共计发表论文16篇，其中SCI论文11篇、EI论文4篇，申请发明专利5项，其中授权2项，共计培养博士生3名，硕士生3名，获得教育部自然奖二等奖1项（第一完成人）。