摘要
研究在潮流迭代求解过程中雅可比矩阵方程组的迭代求解方法及其收敛性。首先利用PQ分解法进行潮流迭代求解,并针对求解过程中雅可比矩阵对称且对角占优的特性,对雅可比矩阵方程组采用高斯置信传播算法(GaBP)进行求解,再结合Steffensen加速迭代法以提高GaBP算法的收敛性。对IEEE118、IEEE300节点标准系统和两个波兰互联大规模电力系统进行仿真计算后结果表明:随着系统规模的增长,使用Steffensen加速迭代法进行加速的GaBP算法相对于基于不完全LU的预处理广义极小残余方法(GMRES)具有更好的收敛性,为大规模电力系统潮流计算的快速求解提供了一种新思路。
An iterative algorithm and its convergence of the Jacobian matrix equations for load flow iterative solution were researched. First, the P-Q decoupled method was used to solve load flow equations, and according to the feature that the Jacobian matrix of correction equations is sysmmetric and diagonally dominant, the Gaussian belief propagation （GaBP） algorithm was proposed for solving the Jacobian matrix equations. The Steffensen＇s iteration was used to speedup GaBP con- vergence. Numerical simulation tests on four systems including IEEE 11S-node system, IEEE 300-node system and two Po- land test systems indicate that, with the scale expanding, contrasting to the generalized minimal residual （GMRES） method with incomplete LU decompostion preconditioner, the convergence of GaBP with Steffensen＇s iteration is remarkable. The method provides a new idea for the fast power flow calculation in power systems.
出处
《计算技术与自动化》
2016年第4期76-80,共5页
Computing Technology and Automation
