智能电网经济运行的分布式控制方法及其收敛速度优化

发布时间：2018-09-09 18:47

【摘要】：分布式控制算法的实现只依赖于本地信息以及局部通信,在主动配电网广域保护与控制中具有良好的控制鲁棒性以及灵活的系统可扩展性。基于一致性理论的分布式控制算法中,各受控对象只需使用局域网络,获取自身以及邻近受控单元的状态信息来更新自身的状态,最终使各受控对象的状态值一致收敛。智能电网是一个电力一次系统和二次系统高度融合的信息物理系统(CPS),电力CPS的快速发展给电力系统中负荷的经济分配问题提供了新的解决思路。文章首先对智能电网进行电力CPS建模,并针对集中控制下的电力系统经济分配问题进行分析求解。其次,文章对电力系统主网的经济分配问题的集中解进行了两种不同形式的化解,使之分别适用平均一致和渐进一致两种算法进行分布式计算。同时考虑到分布式控制对算法收敛速度的要求,文中引入一种只利用智能体中记录的历史迭代值的多项式滤波技术分别对两种分布式经济分配算法进行收敛速度优化。仿真结果验证了所提的分布式经济算法及其收敛速度优化算法的有效性。最后,文章综合考虑智能配电网(主动配电网)中的电源种类多样、可再生电源波动性大以及配电网调度的实际要求,提出一种适用于主动配电网在恒联络线功率和最优经济运行两种模式下的分布式实时调度策略。仿真结果表明,所提的控制策略能够在主网峰谷电价、新能源出力波动、负荷波动等不同情况下分布式地实现对有功功率的实时经济最优分配。
[Abstract]:The implementation of distributed control algorithm only depends on local information and local communication. It has good control robustness and flexible system scalability in wide area protection and control of active distribution network. In the distributed control algorithm based on consistency theory, each controlled object only needs to use local network to obtain the state information of itself and adjacent controlled units to update its own state, and finally make the state value of each controlled object converge uniformly. Smart grid is a highly integrated information physical system, (CPS), power CPS, which is highly integrated with primary and secondary systems. The rapid development of smart grid provides a new way to solve the problem of economic load distribution in power system. In this paper, the power CPS model of smart grid is first built, and the economic distribution problem of power system under centralized control is analyzed and solved. Secondly, in this paper, the centralized solution of the economic distribution problem in power system is solved in two different forms, which make it suitable for the distributed computation of the average consistent algorithm and the asymptotic consistent algorithm respectively. At the same time, considering the demand of distributed control for the convergence speed of the algorithm, a polynomial filtering technique based on the historical iterative values recorded in the agent is introduced to optimize the convergence speed of the two distributed economic allocation algorithms. Simulation results verify the effectiveness of the proposed distributed economic algorithm and its convergence rate optimization algorithm. Finally, considering the variety of power sources in smart distribution network (active distribution network), the volatility of renewable power supply and the actual requirements of distribution network dispatching. This paper presents a distributed real-time scheduling strategy for active distribution networks under the two modes of constant tie-line power and optimal economic operation. The simulation results show that the proposed control strategy can realize the real-time and economic optimal allocation of active power under different conditions such as peak and valley price of main network, fluctuation of new energy output, load fluctuation and so on.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM76

【参考文献】