基于模糊聚类分析的无功电压控制分区

Power Network Partitioning Based on Fuzzy Clustering Analysis

无功电压控制研究中的系统分区是一个非线性的大规模组合优化问题,使用常规方法常难以得到理想的结果。运用模糊聚类分析法将系统分区,并基于电压幅值对无功功率的灵敏度定义了电力系统各节点间的电气距离,对待分类对象的全体作适当的标定,运用传递闭包法求出动态分类,通过计算统计量F最后得出最优分类。该方法所占内存小,计算速度快。经IEEE39节点测试系统对该方法进行了验证,其结果表明:该方法有效、可行。

The partitioning of electric power system for reactive voltage control is a nonlinear large scale combinatorial optimization problem. Conventional methods are not normally appropriate for such a problem. This paper first defines the electrical distance between the two nodes in power system based on the sensitivity of voltage to reactive power injection then all objects classified should be properly demarcated. Passing bundle method is implemented to get the dynamic clustering. After calculating the statistic F, we can get the final optimal clustering. The EMS memory of this method is small and the calculate speed of it is rapid. Finally the feasibility and efficiency of the proposed method are verified on the New-England 39 bus system.