电力系统状态估计并行加速研究  

Study on Parallel Acceleration for State Estimation of Electric Power System

作　　者：闫童 Yan Tong

机构地区：[1]同济大学电子与信息工程学院,上海201804

出　　处：《上海电气技术》2023年第1期1-5,共5页Journal of Shanghai Electric Technology

摘　　要：随着电力系统输配一体化发展,计算耗时急剧增加。根据中央处理器逻辑处理能力强,图形处理器浮点运算能力强的特点,对于状态估计算法中浮点计算最密集的线性方程组求解部分,采用图形处理器并行处理,提高求解速度。将Matpower 6.0潮流算例混合噪声作为状态估计量测,通过不同规模节点算例验证图形处理器的加速效果,体现图形处理器在状态估计问题求解方面的性能优势。With the integrated development of transmission and distribution of electric power system,the computing time increases rapidly.According to the characteristics of strong logic processing capacity of CPU and strong floating-point operation capacity of GPU,for the linear equations solution part with the most intensive floating-point calculation in the state estimation algorithm,GPU was used to parallel process to improve the solving speed.Taking Matpower 6.0 power flow case mixed noise as state estimation measurement,through different scale node examples,the acceleration effect of GPU was verified,and the advantages of GPU in solving problem of state estimation were reflected.

关 键 词：电力系统 状态估计 并行加速 研究 

分 类 号：TM744[电气工程—电力系统及自动化]