6基于小波消噪和相空间重构的径流预测方法及应用(华中科技大学学报自然科学版)(待刊).docVIP

基于小波消噪的混沌径流预测模型 刘力周建中(（华中科技大学水电与数字化工程学院，湖北 武汉 430074）径流； 文献标识码：A 文章编号： Chaotic Prediction Model of Runoff Based on Wavelet De-noising LIU Li, ZHOU Jian-zhong, LI Ying-hai, ZHANG Yong-chuan (College of Hydroelectric and Digitalization Engineering, Huazhong University of Science and Technology, Wuhan 430074, China) Abstract：Considering the influence of noise and the disadvantages of traditional noise eliminating technologies erenow, the wavelet theory is applied to reduce the noise in runoff time series in this paper, of which the wavelet-transform technique is used to decompose and reconstruct noise signal, and by limiting the analyzed wavelet-coefficient to a threshold, the noise is eliminated. On this basis, the prediction model of chaos time series is built by using the radial basis function network, which is based on the phase-space reconstruction theory, and we also discuss the building method, the characteristics of the model, as well as the selecting method for the key parameters. At last, taking the Yichang station of Yangtze River as an example, the proposed model is used to recognize the chaotic feature of the monthly runoff and make prediction, and the result verifies that this model can process a complex hydrological data series better，and is of better generalization ability and higher prediction accuracy. Key words：runoff predition; chaos; radial basis function; phase-space reconstruction; wave de-noising 径流过程是一个高度非线性的不确定性动力演化系统，传统的预测模型对解决像径流时间序列预测这样的复杂问题存在着一定的困难，而混沌理论的出现为研究这种高度复杂的系统提供了新的思路。利用混沌理论的相空间重构技术，将时间序列嵌入重构的相空间中，再借助分形原理和符号动力学，可以揭示系统的复杂特性和内在规律，进而做出预测。由于径流过程对初始条件的敏感性以及内在的随机性，已经证明具有产生混沌的基本条件，因此近年来国内外众多水文学者相继将混沌理论的研究成果引入水文学中，试图对水文时间序列进行混沌识别和预测研究[1~4]。目前研究表明，径流时间序列具有一定的混沌特征，利用混沌预测方法可以对径流序列进行预测。 在进行混沌时间预测之前，有必要考虑径流序列中噪声的存在给混沌识别和预测带来的影响。虽然噪声的存在并不影响径流时间序列的混沌特性识别[5]，但是对径流序列进行消噪处理可以提高数据的可靠性和数据分析成果的精度[6]。实际上，径流时间序列常表现为非平稳随机过程，具有明显的非线性且难于建立合适的状态空间模型，因此传统消噪方法在径流时间序列消噪处理上的应用受到极大的限制，消噪范围十分有限。小波变换是基于Fourier变换而发展起来的一种全新的时频分析方法，可以在不同的尺度下观测信号不同精