采用改进最优模糊C均值聚类和改进和声搜索算法的配电网动态重构  

Dynamic reconfiguration of distribution network based on improved optimal fuzzy C-means clustering and improved harmony search algorithm

作　　者：周洁洁[1] 阙凌燕[2] 王良毅[3] 卢敏[2] 郭创新[1] 

机构地区：[1]浙江大学电气工程学院,浙江杭州310027 [2]国网浙江省电力公司,浙江杭州310000 [3]国网重庆市电力公司,重庆404100

出　　处：《机电工程》2015年第4期531-536,543,共7页Journal of Mechanical & Electrical Engineering

基　　金：国家自然科学基金资助项目(5117143);国家重点基础研究计划资助项目(2013CB228206);浙江省自然科学基金资助项目(LZ12E07002)

摘　　要：针对配电网络动态重构中的开关操作次数约束难以处理与和声搜索算法寻优效率低的问题,分别提出了改进的最优模糊C均值聚类法和改进的和声搜索算法,利用前者对研究时间区间的负荷进行聚类从而将配电网动态重构转为以聚类中心代表各分段负荷的多个静态重构子问题,利用后者对各静态重构问题进行求解。每个时间点的优化网络结构由相应的聚类中心的重构结果决定,由此得到配电网络重构的操作时间点和操作开关。在IEEE 33节点配电系统负荷数据的基础上构建1天的负荷数据并对该系统进行动态重构。研究结果表明,该方法能够控制开关操作次数,提高静态重构算法的寻优效率。Aiming at the constraint of operation times of all switches and the inefficiency of harmony search algorithm,the improved optimal fuzzy C- means clustering technology and the improved harmony search algorithm were proposed. The former was applied to classify the operation state in a time interval according to load data that the dynamic reconfiguration of distribution network was converted to the multiple static reconfigurations that take the corresponding cluster center as the representative load. The latter was applied to solve the static reconfiguration. The optimal network structure at any time was determined by the corresponding cluster center＇s reconfiguration result,at the same time,the reconfiguration time and the operated switches were known. On the basis of IEEE-33 bus system＇s load data,one day＇s load data was constructed and the dynamic reconfiguration was carried out. The results indicate that the proposed method can strictly limit operation times of switches and improve the efficiency of harmony search algorithm.

关 键 词：配电网 动态重构 改进的最优模糊聚类 改进的和声搜索算法 

分 类 号：TM727[电气工程—电力系统及自动化]