基于SVR回归的电力系统暂态稳定隐式梯形算法的改进  

Improvement in Hiding-Trapezium Method for Electric Power System Transient Stability Based on the SVR Technology

作　　者：丁卫[1] 王建全[1] 

机构地区：[1]浙江大学电气工程学院,浙江杭州310027

出　　处：《江南大学学报（自然科学版）》2008年第2期187-190,共4页Joural of Jiangnan University (Natural Science Edition)　

摘　　要：针对用传统隐式梯形积分法分析电力系统暂态稳定时存在计算速度慢的问题,提出了一种基于回归支持向量机技术的改进方法.改进后的方法不仅保持了原有的功角预测精度,而且明显减少了总的迭代次数,有效地提高了计算速度.这种优化结果有利于现代大电网的实时控制.To solove the problem of slow calculating speed in analysing electric power system transient stability by the traditional hiding-trapezium method, the paper comes up with an improved method based on the SVR technology. The optimized method can not only maintain the intrinsic forecasting precision of power angle, but also largely reduce the iterative times and effectively accelerate the calculation. The improved method is very useful for real-time control of transient stability in modem large grid.

关 键 词：暂态稳定 功角预测 回归支持向量机 隐式梯形积分法 实时控制 

分 类 号：TM712[电气工程—电力系统及自动化]