基于混沌的大坝监测序列小波RBF神经网络预测模型

本文结合混沌理论、小波分解与重构,以及径向基函数(RBF)神经网络的优点,提出了一种基于混沌的大坝监测序列小波RBF神经网络预测模型。该模型主要利用小波分析将大坝监测序列分解为趋势项和细节时间序列,并利用RBF神经网络和基于RBF神经网络的混沌理论对两种时间序列进行预测,最后通过小波重构得到预测值。实例分析表明,本模型能够克服监测序列中的噪声干扰,反映大坝监测序列的多尺度特性,对监测数据的预测精度较高,可应用于实际工程。

Chaos-based dam monitoring sequence wavelet RBF neural network prediction model

Combined with the advantages of the chaos theory，wavelet decomposition and reconstruction and the RBF neural network，a prediction model of dam monitoring sequence wavelet RBF neural network based on the chaos is proposed； in which the wavelet analysis is mainly used to decompose the dam monitoring sequence into the trend item and the detailed time series，and then the two kinds of time series are predicted with the RBF neural network and the chaos theory based on RBF neural network. Finally，the prediction value is obtained through the wavelet reconstruction. The actual case analysis shows that the noise interference in the monitoring sequence can be overcome by this model along with the reflection of the multi-scale characteristics of dam monitoring sequence； and the prediction precision of the monitoring data is higher，thus can be applied to the actual project.