中国电力 ›› 2020, Vol. 53 ›› Issue (7): 132-140.DOI: 10.11930/j.issn.1004-9649.202002062

• 智能电网状态估计及其应急仿真专栏 • 上一篇    下一篇

超大规模电网快速状态估计的实现方法

罗玉春1,2,3, 王毅1,2,3, 闪鑫1,2,3, 戴则梅1,2,3, 张磊4   

  1. 1. 南瑞集团(国网电力科学研究院)有限公司,江苏 南京 211106;
    2. 国电南瑞科技股份有限公司,江苏南京 211106;
    3. 智能电网保护和运行控制国家重点实验室,江苏 南京 211106;
    4. 国网山东省电力公司电力科学研究院,山东 济南 250003
  • 收稿日期:2020-02-14 修回日期:2020-04-17 发布日期:2020-07-05
  • 作者简介:罗玉春(1984—),男,通信作者,高级工程师,从事电网调度自动化系统电网分析及计算研究,E-mail: luoyuchun@sgepri.sgcc.com.cn
  • 基金资助:
    国家电网公司科技项目(大电网实时数据及网络分析高性能计算技术研究)

Implementation Method for Fast State Estimation of Super-large Power Grid

LUO Yuchun1,2,3, WANG Yi1,2,3, SHAN Xin1,2,3, DAI Zemei1,2,3, ZHANG Lei4   

  1. 1. NARI Group Corporation (State Grid Electric Power Research Institute), Nanjing 211106, China;
    2. NARI Technology Co., Ltd., Nanjing 211106, China;
    3. State Key Laboratory of Smart Grid Protection and Control, Nanjing 211106, China;
    4. State Grid Shandong Electric Power Research Institute, Jinan 250003, China
  • Received:2020-02-14 Revised:2020-04-17 Published:2020-07-05
  • Supported by:
    This work is supported by the Science and Technology Project of SGCC (Research on High Performance Computing Technology for Real-Time Data and Network Analysis of Large Power Grid)

摘要: 随着一体化互联大电网全局分析决策中心的建设,对实时状态估计计算速度提出了更高要求。采用多线程并行计算技术实现了快速分解状态估计信息矩阵的快速计算,在稀疏矩阵节点优化编号及其因子分解过程中采用标准模板库关联容器存储稀疏矩阵。基于新一代调控系统验证环境和实际电网拼接模型算例进行了验证。结果表明:在超大规模电网状态估计中,采用多线程并行计算信息矩阵及其因子分解具有较高的加速比,结合基于关联容器的稀疏矩阵存储格式,能够有效提升编程效率和程序品质以及状态估计的计算效率。

关键词: 电力系统, 状态估计, 稀疏矩阵乘法, 节点优化编号, 因子分解, 关联容器

Abstract: With the construction of global analysis and decision-making center for integrated interconnected large power grid, higher speed is needed for the calculation of real-time state estimation. The multi-thread parallel computing technology is used to realize the fast calculation of the gain matrix of the fast decoupled state estimation, and the STL associated container storage format is used in the process of sparse matrix bus optimal ordering and its triangular factorization. Based on the verification environment of the new generation control system and the actual grid connection models, case calculations are carried out. The results show that the multi-threaded parallel calculation of the gain matrix and its factorization have a higher speedup ratio when used for state estimation of super large power grids, and can effectively improve the programming efficiency and quality and the computation efficiency of the state estimation when combined with the STL associated container based sparse matrix storage format.

Key words: power system, state estimation, sparse matrix multiplication, bus optimal ordering, triangular factorization, associated container