改进十字链表的稀疏矩阵技术及其在电力系统仿真中的应用  

An Improved Cross Chain Table Based Sparse Matrix Technology and Its Application in Power System Simulation

作　　者：高毅[1] 王成山[1] 李继平[1] 

机构地区：[1]电力系统仿真控制教育部重点实验室(天津大学),天津市南开区300072

出　　处：《电网技术》2011年第5期33-39,共7页Power System Technology

基　　金：国家重点基础研究发展计划项目(973项目)(2009CB219700)~~

摘　　要：十字链表具有检索方式灵活和操作方便的特点,分析了十字链表在三角分解和前代–回代计算中的应用方法及其内存分配方式对稀疏矩阵计算效率的影响,十字链表的存储结点通常散列于内存空间中,该内存分配方式下稀疏矩阵的运算效率低于存储结点连续分布在内存中的稀疏矩阵运算效率,该现象由计算机的高速缓冲存储器的工作原理造成。为此提出改进十字链表实现方法,在稀疏矩阵的运算过程中充分利用高速缓冲存储器的高速存取速度来提高计算效率。通过算例证明改进十字链表方法是一种高效的稀疏矩阵技术,在传统大规模系统和分布式发电供能系统仿真中均取得了良好的应用效果。The structure and operation manner of cross chain table and its application in the triangular factorization and forward-backward substitution are presented in this paper to illustrate that its convenience and efficiency in serching, adding and deleting matrix elements can help speeding up the calculation process of factorization and forward-backward substitution. And the impacts of the memory allocation manner of the cross chain table on the calculation efficiency are discussed. The nodes of cross chain table are a/ways stored seperately in the memory because of its memory a/location manner, which can reduce the calculation efficiency. This phenomenon is caused by the principle of the cache. Aiming at this problem, an improved cross chain table method is proposed in this paper which inherits the advantages of the conventional cross chain table and can raise the calculation efficiency by taking adavantage of the high speed access of cache. Case study proves the efficiency of the improved cross chain table, and shows that it can be well applied in both the traditional large scale power system simulation and the distributed generation system simulation.

关 键 词：电力系统仿真 十字链表 稀疏矩阵 分布式发电 

分 类 号：TM743[电气工程—电力系统及自动化]