新息图状态估计分块算法  

Partitioning algorithm for innovation graph state estimation

作　　者：张艳军[1] 施毅斌 周苏荃[1] 

机构地区：[1]哈尔滨工业大学电气工程及自动化学院,黑龙江哈尔滨150001 [2]辽宁电网公司,辽宁沈阳110006

出　　处：《哈尔滨工程大学学报》2008年第11期1166-1171,共6页Journal of Harbin Engineering University

基　　金：国家自然科学基金资助项目(50177006)

摘　　要：针对大型电网分区管理需要,提出了新息图状态估计分块算法.该方法采用两级调度协调配合方式,在上一级调度构建原整体网络的简化网络,计算子区域边界支路条件,并下发给每个子区域;子区域根据上一级调度下发的边界支路条件实现新息图状态估计.该方法具有子区域间不需要数据交互、两级调度之间交换数据量小、计算时间短、能够实时在线运行的优点,采用国内某省360节点,550条支路的实际网络验证了算法的有效性.A partitioning algorithm for innovation graph state estimation was proposed to meet the management re- quirement of large-scale power grid. With two-level cooperative control centers, this method established a simplified network in the upper-level control center to replace the original whole grid. Then in the simplified network, it cal- culated the innovation vector of boundary branches and sent them to every sub-network control center. Using the calculated innovation vectors of boundary branches, the innovation graph state estimation was implemented in the sub-network. This method doesnt need to exchange data between sub-networks and the amount of data exchanged between two-levels of control centers is very small, so the computation time is less. This allows the method to be used in real-time operations. The effectiveness of the proposed method was demonstrated on a practical network with 360 nodes and 550 branches on a provincial electrical network.

关 键 词：新息图 分块算法 不良数据 拓扑错误 

分 类 号：TM76[电气工程—电力系统及自动化]