基于迭代学习理论的配电网重构  

Distribution Network Reconfiguration Based on Iterative Learning Control

作　　者：徐敏[1] 吴成[1] 罗燎 林辉[2] 戴薇[1] XU Min WU Cheng LUO LiaoI LIN Hui DAI Wei(School of Information Engineering, Nanchang University, Nanehang 330031, China School of Automation, Northwestern Polytechnical University, Xi＇an 710072, China)

机构地区：[1]南昌大学信息工程学院,江西南昌330031 [2]西北工业大学自动化学院,陕西西安710072

出　　处：《中北大学学报（自然科学版）》2017年第1期66-71,共6页Journal of North University of China(Natural Science Edition)

摘　　要：迭代学习理论具有不依赖数学模型对期望轨迹进行零误差跟踪的特点,将优化目标改为对目标函数轨迹的跟踪,这样可以将其推广应用为一般的优化问题.基于该理论的跟踪轨迹机制,以网损最小为目标函数并考虑相应的约束条件,采用迭代学习理论中的开环P型学习律,通过改变网络中的开关状态进行重构,得到网损最小的配网结构,为配电网重构提供了一种新方法.采用标准IEEE33,PG&E69节点系统的算例进行验证,不仅能够求得全局最优解,而且迭代时间由原来的1.25s变为0.266s.Since iterative learning theory has the characteristic of not rely on mathematical models for the zero error expected trajectory tracking, the optimization objective function is replaced by the target trajectory. So it can be applied to the general optimization problem. Based on the mechanism of tracking the trajectory theory, a new method for the distribution network reconfiguration was presented in this paper, with the objective function of minimum network loss and the consideration of the corresponding constraints. The open-loop P-type learning law was used to get the minimum loss distribution network by changing the switch state of network for reconstruction to get the distribution network structure with the minimum network loss. Finally, the standard IEEE33 and PG＆E69 bus system were used to reconstruction which not only got the global optimal solution, but also changed the iteration time from 1.25 s to 0. 266 s.

关 键 词：迭代学习控制 配电网重构 收敛性 网损 

分 类 号：TM727[电气工程—电力系统及自动化]