摘要
随着高性价比可扩展集群并行系统的逐步成熟和应用 ,大规模电力系统潮流并行计算和分布式仿真成为可能。该文首先简要介绍了电力系统的潮流计算模型及基本算法 ,分析了算法中存在的困难。然后分析了 4类较重要的电力系统潮流并行算法 :分块法、多重因子化法、稀疏矢量法和逆矩阵法 ,同时讨论了这 4种潮流并行算法的基本原理和实用效果 ,比较了各种并行算法的优点和局限性 ,并指出基于集群系统的粗粒度区域分解潮流并行算法最具发展潜力。
The wide use of scalable cluster parallel systems, which have a high performance -price ratio, has enabled the power flow parallel computing and distributed real time simulation of large scale power system. This paper outlined the computational models and the basic algorithms of power flow computation, and analyzed the main difficulties in the algorithms. Then, starting from the basic principles and practical effects of the algorithms, it presented the research progress of parallel power flow algorithms based on partitioning, multiple factoring, sparse vectors and inverse matrices. Finally, comparing the advantages and limitations of these parallel algorithms, the coarse granularity domain decomposed power flow parallel algorithm on cluster parallel systems is the most promising method.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2002年第9期1192-1195,1198,共5页
Journal of Tsinghua University(Science and Technology)
基金
国家自然科学基金资助项目 (6993 3 0 2 0 )
