基于改进模糊C均值聚类时段划分的配电网动态重构

Dynamic Reconfiguration of Distribution Network Based on Improved Fuzzy C-means Clustering of Time Division

随着大量光伏和风电等间歇性电源接入配电网,传统的静态重构方案不再适用于动态变化下的网络。在此背景下,提出了一种基于改进模糊均值聚类的动态重构策略。首先依据分布式电源(distributedgenerator,DG)和负荷的时变性建立确定性等值负荷预测曲线,通过改进的模糊C均值聚类算法进行时段划分,并且利用损失函数确定最优时段划分方案。其次采用区间数描述DG和负荷预测的不确定性并建立以网损区间值最低为目标函数的动态重构模型,并引入仿射泰勒扩展的潮流计算法求解区间潮流方程。最后采用基于回路搜索的十进制粒子群算法对重构模型进行求解,实现不确定因素下的配电网动态重构。通过IEEE33节点系统仿真验证了所提方法的有效性和优越性。

With increasing intermittent power sources, such as photovoltaic and wind power, connected to distribution network, traditional static reconfiguration scheme is no longer suitable for dynamic network. In this context, a dynamic reconfiguration scheme based on improved fuzzy C-means clustering is proposed in this paper. Firstly, the deterministic equivalent load forecasting curve based on the time-varying property of DG and load is divided into segments using the improved fuzzy C-means clustering algorithm. And the loss function is applied to determine optimal time division scheme. Secondly, the interval value describing the uncertainty of DG and load forecasting is adopted to establish a dynamic reconfiguration model with minimum network loss. Then the power flow method based on affine Taylor expansion is used to solve the interval power flow equation. Finally, the reconfiguration model is solved with decimal particle swarm optimization algorithm based on loop search. The simulation analysis of IEEE33 distribution system shows that the proposed method is effective and superior.