基于相空间重构理论的神经网络在大坝渗流预测中的应用  

Application of ANN Based on Phase-space Reconstruction in Prediction of Seepage in Dam

作　　者：钱镜林[1] 陈闲忠 张晔[1] 

机构地区：[1]浙江省水利河口研究院,浙江杭州310027 [2]浙江省常山县水利局,浙江衢州324202

出　　处：《水利与建筑工程学报》2006年第1期21-23,共3页Journal of Water Resources and Architectural Engineering

摘　　要：简述了混沌预测方法和原理,提出运用神经网络来模拟混沌系统的动力学模型。通过实例重构出大坝渗流观测数据相空间,将其作为神经网络的输入,对渗流观测数据进行了预测。计算结果表明,基于相空间重构理论的神经网络模型具有一定的准确性和有效性,对于处理非线性问题是一种有益的探索。The ANN is employed to simulate the chaotic dynamics model after the method and theory are described briefly. The phase-space of the observed data of seepage is reconstructed and is taken as the input of the ANN to predict the data. The calculated results show that this model is of high efficiency and of high accuracy, and it is an effective method to analyse the nonlinear phenomena.

关 键 词：相空间重构 神经网络 渗流 

分 类 号：TV139.11[水利工程—水力学及河流动力学]