基于混沌时间序列的大坝变形短期预测
田旦;许才军;周命端;陈浩
将混沌时间序列预测理论应用到大坝变形预测中,根据非线性大坝变形时间序列,运用相空间重构理论,建立了加权一阶局域法、基于最大Lyapunov指数法大坝预测模型,对混沌的大坝变形数据短期预测模型进行了研究,对比分析了各自的特点,并结合实例完成了对大坝变形的预测。计算分析表明,该模型预测误差较小,与传统的自回归模型预测结果相比,基于混沌时间序列的预测方法在大坝变形的短期预测中具有更高的精度。
【作者单位】：武汉大学测绘学院;上海港务工程公司检修技术站
【关键词】：大坝变形预测;混沌时间序列;相空间重构;加权一阶局域法;Lyapunov指数
【正文快照】：
0引言混沌时间序列预测是20世纪80年代末发展起来的一种非线性预测方法[1],传统的预测方法主要有动力学方法和数理统计方法,其共同特点是先建立数据序列的主观模型,然后根据主观模型进行计算和预测。混沌科学的发展,使得不必事先建立主观模型,而直接根据数据序列本身所计算出来
The chaotic Time Series-based Short-term Prediction of Dam Deformation
The chaotic time series prediction theory is applied to the dam deformation.According to the nonlinear characteristics of dam deformation time series and the theory of phase space reconstruction,two prediction models of dam deformation are established.The two models are the prediction model of adding-weight one-rank method and the prediction model of Lyapunov exponents.The characteristics of the above models are compared and analyzed as well.The study indicates that the prediction errors are small.Compared with the traditional AR model,the chaotic time series prediction theory has a high accuracy in the dam deformation prediction.
【Keyword】：dam deformation prediction;chaotic time series;phase space reconstruction;adding-weight one-rank local-region method;Lyapunov exponents
