混凝土坝变形Wavelet-EGM-PE-ARIMA组合预测模型

混凝土坝的总变形可以归结为由水压和温度变化引起的变形以及随时间发展的变形。其中,水压变形和温度变形体现为总变形中的周期性分量,而时效变形体现为总变形中的趋势性分量。借助复合建模思想,提出一种混凝土坝变形Wavelet-EGM-PE-ARIMA组合预测模型。首先利用小波多分辨分析功能,分解出大坝变形时间序列中的趋势性项、周期性项;其次,运用EGM模型实现对趋势性项的有效预测,采用周期外延模型实现对周期性项的有效预测,在此基础上,利用ARIMA模型实现对EGM模型和周期外延模型残差项的有效预测;最后通过某工程实例,检验所提出模型的有效性。计算结果表明:该组合模型充分考虑大坝各变形分量的变化规律,并基于此,实现对大坝变形时间序列有效的拟合和预测,且其拟合和预测精度均明显优于传统统计模型。

A Combinatorial Wavelet-EGM-PE-ARIMA Model for Predicting Concrete Dam Deformation

The total deformation of concrete dam can be attributed to the deformation caused by water pressure, temperature and time, among which the deformations caused by water pressure and temperature are reflected as periodic components, while the aging deformation as trend component. In this paper, a combinatorial deformation prediction model for concrete dam is established by integrating wavelet decomposition, Even Grey Model (EGM), Periodic Extension (PE), and Autoregressive Integrated Moving Average (ARIMA) model. Wavelet is employed to decompose the trend items and periodic items in the time series of dam deformation; EGM for the effective prediction of trend term, and PE model for periodic term; ARIMA model is adopted for the prediction of residuals of EGM and PE model. An engineering case study verifies the effectiveness of the present model. The results show that the time series of dam deformation can be fitted and predicted effectively by this combined model, in which the variation law of each deformation component of the dam is considered. The fitting accuracy and prediction accuracy of the combined model are both superior to those of traditional statistical model.