电工技术学报2016，Vol.31Issue(15)：109-116,8.
电气化铁路牵引网节点方程组数值解法比较
Comparison of Numerical Methods for Nodal Analysis in Electric Railway Traction Network
摘要
Abstract
The traction power supply network in electric railways is a complicated single-phase unbalanced network with the earth as a part of the current return loop, which is different from the three-phase power system. The nodal analysis is a method in common use to describe the network. With the train regarded as a power consumption model,it needs iterations to solve the nonlinear power flow equations. The Picard iteration is the most popular iteration method. However,if the loads are set to be very heavy in the calculation,it may fail to converge. It has not been discussed whether the failure of the Picard iteration means the nonexistence of the power flow solution or not. In this paper,the converging characteristics of the Picard iteration,the Newton-Raphson method and the Levenberg-Marquardt algorithm are compared with an example. The relation between the convergency of the algorithms and the existence of the power flow solution is discussed. The Picard iteration is also discussed in the aspects of the iteration time, the change of the train voltage during the iteration, and the influence of the initial value.关键词
电气化铁路/牵引网/潮流计算/节点分析法/数值算法Key words
Electric railway/traction network/power flow/nodal analysis/numerical method分类
信息技术与安全科学引用本文复制引用
张俊骐,吴命利..电气化铁路牵引网节点方程组数值解法比较[J].电工技术学报,2016,31(15):109-116,8.基金项目
中国铁路总公司科研开发计划重点课题（2015J008-F）资助项目。 （）
