您的位置:首页  > 论文页面

水文时间序列长度对混沌识别的影响分析

发表时间:2012-10-15  浏览量:1299  下载量:585
全部作者: 于新伟,刘东
作者单位: 东北农业大学水利与建筑学院
摘 要: 采用建三江分局七星农场1957年至2010年648个月的降水观测资料,基于相空间重构技术、G-P算法等分别对72个月、216个月及648个月等不同长度序列的关联维数及Kolmogorov熵进行计算。结果表明:1) 七星农场月降水时间序列存在混沌特性;2) 延迟时间τ=3个月,嵌入维数m=7,序列长度对延迟时间τ及最小嵌入维数m的选取影响较小;3) 饱和关联维法不适合小数据量时间序列的混沌性识别,且序列长度越长,关联维数越容易达到饱和;4) Kolmogorov熵法可以对不同长度时间序列进行处理,且序列长度越长,Kolmogorov熵越容易趋于稳定。研究为混沌理论的进一步完善及其在水文学方面的应用提供了科学依据,同时对于建三江分局降水时间序列的分析具有一定的指导意义。
关 键 词: 农业工程;水文时间序列;相空间重构;关联维;Kolmogorov熵;降水�
Title: Analysis on influence of hydrological time series length on chaotic identification
Author: YU Xinwei, LIU Dong
Organization: College of Water Conservancy and Civil Engineering, Northeast Agriculture University
Abstract: According to the observed precipitation data from Qixing farm of Jiansanjiang during 1957-2010 and based on the phase space reconstruction technique, G-P algorithm, the correlation dimension and the Kolmogorove entropy of different length sequence for 72 months, 216 months and 648 months were calculated respectively. The results showed that: 1) Chaotic characteristics of monthly precipitation time series of Qixing farm was existed. 2) The delay time was τ=3 months, the smallest dimension was m=7, the sequence length had little influence on the selection of the delay time and the smallest dimension. 3) Saturation correlation dimension method was unsuitable for chaotic identification for a small amount of data. With the longer sequence length, the correlation dimension reached saturation more easily. 4) The Kolmogorove entropy method could deal with different lengths of time series, and with the longer sequence length, the Kolmogorov entropy was more easily stabilized. This research provided the scientific basis for the further improvements and applications of chaos theory in hydrology, and it had a certain significance for precipitation time series analysis of Jiansanjiang.�
Key words: agricultural engineering; precipitation time series; phase space reconstruction; correlation dimension; Kolmogorov entropy; precipitation
发表期数: 2012年10月第19期
引用格式: 于新伟,刘东. 水文时间序列长度对混沌识别的影响分析[J]. 中国科技论文在线精品论文,2012,5(19):1871-1876.
 
0 评论数 0
暂无评论
友情链接