摘要
提出利用原对偶非线性变尺度(primal-dual nonlinear rescaling,PDNR)算法求解电力系统最优潮流问题。在满足2阶最优条件时,PDNR方法具有1.5Q的超线性局部收敛速度。为实现全局收敛的最优潮流,将线性搜索算法与PDNR算法结合,形成一个全局收敛的原对偶非线性变尺度最优潮流(global PDNR optimal power flow,GPDNR-OPF)算法。该算法的特点是具有较好的全局收敛性和快速的局部收敛性。对多个IEEE测试系统进行数值仿真分析,结果显示GPDNR-OPF算法的快速收敛性得到了验证,并且在寻优解过程进入最优解邻域后,海森矩阵的病态条件与PDNR算法无关。
An approach using the primal-dual nonlinear rescaling （PDNR） method for optimal power flow （OPF） was proposed. The PDNR algorithm has a 1.5Q superlinear rate under the standard second order optimality conditions. For implementing the global convergence OPF, a global primal-dual nonlinear rescaling method for optimal power flow （GPDNR-OPF） was constructed through combining the algorithm of backtracking line search and the PDNR method. The main characteristic of this method has good global convergence and fast local convergence. The GPDNR-OPF was numerically implemented and tested on OPF problem from some IEEE testing systems. Numerical results show that the convergence rate of GPDNR-OPF method mentioned above is verified, and the ill-conditioning of the Hessian of minimized function becomes irrelevant for the PDNR method once primal-dual vector enter into the neighborhood of the OPF optimum solution.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2009年第31期47-52,共6页
Proceedings of the CSEE
基金
国家自然科学基金项目(50577017)~~
关键词
电力系统
优化潮流
原对偶非线性变尺度方法
收敛性
power system
optimal power flow （OPF）
primal-dual nonlinear rescaling method
convergence
