基于改进k-means算法的有源配电网网格化规划方法

Grid Planning Method of Active Distribution Network Based on Improved k-means Algorithm

针对高渗透光伏大规模接入配电网可能引发电压越限和配电网峰谷差大等问题,提出一种考虑负荷互补特性的有源配电网网格划分方法。首先,分析各类配电网网格化规划方案和负荷峰谷互补特性;然后,以网格化规划后子网格负荷曲线与总负荷曲线的形状相似度最高为目标,充分考虑配电网运行约束条件,以改进k-means聚类算法为基础构建网格划分方案,并根据评价指标对划分结果进行评估;最后,以江苏省某县级市城区为例,验证所提网格划分方案的可行性和有效性。

Large-scale access of high-permeability photovoltaic to distribution network may cause problems such as over-limit voltage and large peak-valley difference in distribution network.This paper proposes a grid division method for active distribution network considering load complementarity characteristics.First of all,the grid planning schemes of various distribution networks and the complementary characteristics of load peak and valley are analyzed;then,taking the shape similarity of the sub-grid load curve and the total load curve after the grid planning as the goal,fully considering the distribution network operation constraints,based on the improved k-means clustering algorithm,the grid division scheme is constructed,and the division results are evaluated according to the evaluation indicators;finally,taking a county-level city in Jiangsu province as an example,the feasibility and effectiveness of the proposed grid division scheme is verified.