基于小波与谱分析的大坝变形预报模型

传统大坝变形回归分析模型不具备降噪功能,同时人为拟定环境量周期项,存在一定的不足,对此,引入小波与谱分析方法,提出了基于小波与谱分析的大坝变形预报模型,以提高预报精度。首先对实测的位移数据进行分段三次Hermite插值,获取均匀采样数据;然后对采样数据进行小波分解,做降噪处理;最后通过谱分析,根据功率谱图,确定位移量的主要周期项,结合传统回归方法构建合理的拟合模型,实现大坝变形的预报。以2010年三峡大坝175 m蓄水后的实测数据为样本,采用文中方法对大坝变形值预报,结果表明预报精度较高,新方法对提高大坝安全评估和类似变形分析有重要价值。

Forecast model for dam deformation based on wavelet and spectral analysis

The traditional regression model for dam deformation is not equipped with denoising function, and has certain defects because the cycle items of environment are artificially determined. In order to improve the accuracy of the model forecast, the method of wavelet and spectral analysis is used, and a forecast model is proposed. The new method uses the observed data interpolated by the piecewise cubic Hermite polynomial to get the uniform sampling data. Then the noise of the uniform sampling data is reduced by the wavelet transform. Finaly, the dominant periodic terms of the data are found by using the spectrual analysis. Combining the traditional regression model, the appropriate model is established. On this basis, the deformation of Three Gorges Dam with impoundment of 175 m is predicted. The calculated results show that the proposed new method can effectively forecast the dam deformation with high precision and has the reference value for dam safety evaluation and similar deformation analysis.