优化牛顿-拉夫逊算法雅可比矩阵的正交预处理方法研究  

Study on orthogonal preconditioning method for Jacobian Matrix in Newton-raphson optimization

作　　者：卓芳[1] 高仕斌[1] 

机构地区：[1]西南交通大学电气工程学院,四川成都610031

出　　处：《电力系统保护与控制》2010年第3期20-23,42,共5页Power System Protection and Control

摘　　要：预处理雅可比矩阵可以大大改善潮流计算方程的收敛性和速度,而寻找有效的预处理方法则是关键。研究了预处理的基本原理,以最低条件数为主要依据,提出正交方法预处理雅克比矩阵,相对于目前较为流行的P-Q分解法,理论上更加简单且操作方便。借助Matlab仿真工具,预处理几个典型的IEEE多节点系统,验证了正交预处理方法相对P-Q方法,计算时间更短但具备相当的有效性和适用范围。The choice of preconditioning method is a key in the process of preconditioning Jacobian Matrix, which can improve convergence and quicken calculation speed in power flow calculation. This paper, based on minimum condition number, introduces orthogonal preconditioning method for Jacobian Matrix after analysis of basic principle in preconditioning. Compared to P-Q method which is considered as the most effective way, this method is much easier to operate in theory, and is finally testified its similar availability and applicability but less calculating time using Matlab Simulink Tool in certain typical IEEE systems.

关 键 词：潮流计算 雅可比矩阵 预处理:正交 P-Q分解法 

分 类 号：TM744[电气工程—电力系统及自动化]