基于自动微分和过滤线性搜索的非线性内点算法  

A Nonlinear Interior Point Algorithm Based on Automatic Differentiation and Filter Line Search Technique

作　　者：李翔 刘文涛 牛振勇 肖鸣 莫琦 范磊 

机构地区：[1]南方电网电力调度控制中心,广州510623 [2]南瑞继保电气有限公司,南京211106

出　　处：《南方电网技术》2012年第5期48-52,共5页Southern Power System Technology

摘　　要：非线性内点法已被广泛应用于求解大规模电力系统最优潮流问题。提出一种非线性内点算法,使用自动微分技术取代传统手动编程求导来计算雅可比矩阵和海森矩阵,还在寻优过程中使用过滤线性搜索技术,提高了搜索的可靠性和收敛速度。算例计算表明,该算法具有收敛性能好,可扩展性强和计算速度快的特点,具有应用于电力系统实时运行环境的潜力。Nonlinear interior point algorithm has been widely used in solving large scale power system optimization problem. An improved nonlinear interior point algorithm with automatic differentiation and filter line search technique is proposed in this paper. The proposed algorithm uses automatic differentiation and filter line search to calculate Jacobi and Hessian matrix instead of using hand writing code and to implement the optimal searching process respectively, and thus improving convergence and search speed. Numerical studies indicate that the proposed algorithm has high reliability, robust convergence, good expansibility and rapid calculation speed as to have an application potential in real time environment of power system.

关 键 词：自动微分 过滤线性搜索 非线性内点法 

分 类 号：TM744[电气工程—电力系统及自动化]