电力系统分布式模型预测控制方法综述与展望

Review and Prospect on Distributed Model Predictive Control Method for Power System

模型预测控制作为一类约束优化控制策略,在电力系统控制领域得到了广泛的应用,传统的集中式模型预测控制难以满足复杂互联电力系统的多目标优化、可靠性、实时优化等要求。分布式模型预测控制能够协同多个局部控制器实现全局优化或纳什优化。文中全面综述分布式模型预测控制在电力系统中的研究现状和发展方向。首先,按照控制结构对模型预测控制方法进行了分类和对比分析,接着重点阐述了非协作式和协作式2类分布式模型预测控制方法。然后,围绕分布式模型预测控制在不同应用场景分别进行综述和分析。最后,指出了分布式模型预测控制在电力系统控制应用的研究方向。

Model predictive control(MPC),as a constraint optimization control strategy,has been widely used in the power system control.Traditional centralized MPC is difficult to satisfy the requirements for multi-objective optimization,reliability and real-time optimization in complex interconnected systems.Distributed MPC can coordinate with multiple local controllers to achieve global optimization or Nash optimization.This paper summarizes the research status and development trends of distributed MPC in the power system.First,the MPC methods are classified and compared according to the control structure,and the two kinds of noncooperative and cooperative distributed MPC methods are emphasized.Then,the different application scenarios of distributed MPC are summarized and analyzed,respectively.Finally,the research areas of distributed MPC in power system control are pointed out.