基于数据驱动算法和LS-SVM的输电线路覆冰预测

发布时间：2018-01-13 16:21

本文关键词：基于数据驱动算法和LS-SVM的输电线路覆冰预测　出处：《电力系统自动化》2014年15期 　论文类型：期刊论文

【摘要】：输电线路的严重覆冰给电网的安全生产带来了极大的危害。针对线路覆冰与气象条件的数学描述模型不能很好地反应冰厚与微气象条件的关系和基于全局数据的智能覆冰预测算法复杂、准确度低,无法实现在线预测的难题,文中采用数据驱动的思想,以矢量的方式看待覆冰样本数据,提出一种基于数据驱动算法和最小二乘支持向量机(LS-SVM)覆冰预测模型。该方法在k均值邻近算法的基础上对覆冰历史数据进行优化选择,充分利用LS-SVM需求样本数量少、训练速度快、泛化能力强等特点对输电线路覆冰模型进行快速建模。算例表明了所提算法的有效性和正确性。
[Abstract]:The severe icing of transmission lines has brought great harm to the safety of the power network. The mathematical description model of the line icing and meteorological conditions can not well reflect the relationship between the ice thickness and the micrometeorological conditions, and based on the global data, the model can not well reflect the relationship between the ice thickness and the micrometeorological conditions. Intelligent ice prediction algorithm is complex. Because of the low accuracy, it is difficult to realize the on-line prediction. In this paper, the idea of data-driven is used to treat the ice-covered sample data in the way of vector. A prediction model of ice coating based on data-driven algorithm and least squares support vector machine (LS-SVM) is presented, which optimizes the selection of icing history data based on k-means proximity algorithm. Taking full advantage of the characteristics of LS-SVM, such as small number of samples, fast training speed and strong generalization ability, the fast modeling of icing model of transmission lines is carried out. An example is given to show the validity and correctness of the proposed algorithm.
【作者单位】：南京工程学院电力工程学院;国网浙江省电力公司运营监控中心;
【分类号】：TM752
【正文快照】： 0引言输电线路长期曝露野外,承受着自然界各种因素的侵蚀,已成为电力系统中最为脆弱的部分[1-2]。近几年来,由于极端气候的影响,在南方寒冷的冬季极易引起线路的严重覆冰,因此,如何防治输电线路的覆冰灾害已成为智能电网建设中亟待解决的问题之一[3]。国内外众多学者对电力线

【参考文献】