并行稀疏系统直接求解库SuperLU_MT在状态估计中的应用  

Application of Parallel Sparse System Direct Solver Library Super LU_MT in State Estimation

作　　者：陈荃韡 龚成明[3,4] 赵晋泉[1,2] 王毅[3,4] 邹德虎[3,4] CHEN Quanwei GONG Chengming ZHAO Jinquan WANG Yi ZOU Dehu(Research Center for Renewable Energy Generation Engineering of Ministry of Education, Hohai University, Nanjing 210098, China College of Energy and Electrical Engineering, Hohai University, Nanjing 210098, China NARI Group Corporation （State Grid Electric Power Research Institute）, Nanjing 211106, China NARI Technology Co. Ltd., Nanjing 211106, China)

机构地区：[1]河海大学可再生能源发电技术教育部工程研究中心,江苏省南京市210098 [2]河海大学能源与电气学院,江苏省南京市210098 [3]南瑞集团公司(国网电力科学研究院),江苏省南京市211106 [4]国电南瑞科技股份有限公司,江苏省南京市211106

出　　处：《电力系统自动化》2017年第3期83-88,共6页Automation of Electric Power Systems

基　　金：国家电网公司科技项目"智能电网调度技术支持系统数据稽查关键技术研究";江苏省普通高校研究生科研创新计划资助项目(SJZZ15_0064)~~

摘　　要：利用基于超节点LU分块分解算法、列消去树和多核多线程并行计算原理的并行稀疏系统直接求解库Super LU_MT实现电力系统状态估计,主要在求解迭代方程、不良数据检测辨识环节调用Super LU_MT实现加速。通过对电气与电子工程师协会(IEEE)测试系统以及实际电网算例验证,表明通用的并行稀疏系统直接求解库适用于电力系统状态估计,能够获得较显著的并行计算效果,在多核计算机上缩短了状态估计的计算时间。The library of SuperLU_MT based on the use of the super node LU decomposition block algorithm, column elimination trees, multi-core and multi-threaded parallel computing theory, is used to develop electric power system state estimation. To achieve acceleration, the interface of SuperLU_MT is called mainly in the solution of the iterative equation, the bad data detection and identification. The validity of the algorithm is verified by the Institute of Electrical and Electronic Engineers （IEEE） test power system and an actual power system, which indicates that the parallel sparse system direct solver library is suitable for state estimation of power systems. It will significantly reduce state estimation computation time on multicore computers.

关 键 词：电力系统状态估计 并行矩阵计算库 超节点 LU分块分解 多核多线程并行计算 

分 类 号：TM711[电气工程—电力系统及自动化]