短期负荷预测最大李亚普诺夫指数法的改进.pdfVIP

29 卷 7 期 电 网 技 术 Vol. 29 No. 7 2005 年4 月 Power System Technology Apr. 2005 文章编号：1000-3673 （2005 ）07-0031-05 中图分类号：TM734 ；O415.5 文献标识码：A 学科代码：4704054 短期负荷预测最大李亚普诺夫指数法的改进 杨正瓴，田勇，张广涛，林孔元 （天津大学电气与自动化工程学院，天津市 南开区 300072 ） IMPROVEMENT OF MAXIMAL LYAPUNOV EXPONENT METHOD TO SHORT TERM LOAD FORECASTING YANG Zheng-ling，TIAN Yong，ZHANG Guang-tao，LIN Kong-yuan （School of Electrical Engineering and Automation，Tianjin University，Nankai District，Tianj in 300072 ，China ） ABSTRACT: Three technical improvements for the existing 大李亚普诺夫指数；延时时间 load forecasting methods based on the maximal Lyapunov 1 引言 exponent are presented. The first is use of the maximal Lyapunov exponent estimation in the time window 张步涵等先生的“基于混沌时间序列的负荷预 corresponding to load at the time point to be forecasted; the 测及其关键问题分析”一文在混沌预测的最大李亚 second is to improve the rule of acceptance and rej ection by 普诺夫指数l1 方法中的预测方程求解方面取得了重 use of the correlation coefficient between atmospheric [1] 要进展 。该文提出的“取舍规则”既适合于考虑天 temperature and load; the third is adopting multi-neighboring vectors in the forecasting to enhance the ability of anti-noise. 气因素、又可以在不考虑天气的负荷预测中使用， On the basis of above-mentioned improvements, two kinds of 并且有较高的预测准确率。本文探讨的重点则是对 forecasting methods are as following: (A) the forecasting 最大l1 方法“取舍规则”的进一步改进，即如何加 method using varying maximal Lyapunov exponent, improved 强抗噪声能力、与相空间邻近等距模式的关系（回 rule of acceptance and rej ection and multi-neighboring vectors; 避l 估计）以及对延时时间t嵌入维数m 的优化等。 1 (B) the forecasting method of weighted valu