标题:
基于小波消噪的混沌神经网络月径流预报模型The Chaotic Neural Network Model of Monthly Runoff Forecast Based on Wavelet De-Noising
作者:
周建中, 张娟娟, 郭俊, 张勇传
关键字:
径流预报, 小波消噪, 饱和嵌入维数, 混沌神经网络Runoff Forecast; Wavelet Do-Noising; Saturated Embedding Dimension; Chaotic Neural Network
期刊名称:
《Journal of Water Resources Research》, Vol.1 No.3, 2012-06-27
摘要:
受天气系统和流域下垫面系统综合作用的影响，径流过程具有高度的非线性特征。针对径流时间序列强相关性和复杂特性，本文综合运用小波变换、混沌理论和神经网络非线性理论对水文时间序列进行分析和预测。首先通过小波变换对月径流序列进行消噪处理，然后推求出大于零的李雅普诺夫指数，证实了宜昌站的月径流序列具有混沌特性，为此引入混沌理论中的相空间重构方法计算出宜昌站1882~2008年月径流序列的最佳延迟时间和饱和嵌入维数，进而以相空间重构后的时间序列作为神经网络的输入进行网络训练得到最佳的混沌神经网络径流预报模型。实例研究结果表明，该模型能较好地处理复杂非线性径流序列，预测精度高，具有实际工程应用价值。
Runoff process is highly nonlinear characteristics under the synthetic action of weather system and underlying surface system. Considering the strong correlation and high complexity of runoff time series, the wavelet transform is applied to eliminate noise in the monthly runoff time series in this paper, of which the Lyapunov index method is used to recognize the chaotic feature of the monthly runoff time series. On this basis, the phase space restructure of chaos theory is used to calculate the best delay time and saturated embedding dimension of the runoff time series from 1882 to 2008 of the Yichang station. At last, taking the times series computed by the phase space restructure as the input of chaotic neural network model to get the proposed model by network training. Prediction results show that this model can process a complex hydrological data series better, and is of higher prediction accuracy and good prospect of engineering application.