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.