Abstract:

This paper aims to study on prediction method of dam deformation for GNSS automatic monitoring system.According to the GNSS deformation monitoring data's characteristics of large sample,high sampling rate and continuous,a new method combining wavelet analysis with BP and NAR Neural network to predict dam deformation is presented.Firstly,we use multi-scale wavelet analysis to decompose and reconstruct the deformation monitoring data sequences.Then,the BP neural network is used to forecast the low frequency approximation sequences,and the NAR neural network is used to forecast the high frequency detail sequences.Finally,the forecasting results of each scale are accumulated to get the prediction results of dam deformation.The application shows that the proposed dam deformation forecasting model in this paper has high predictive accuracy and great generalization performance,which can be widely used in dam deformation prediction for GNSS automatic monitoring system.

Key words: GNSS automatic monitoring system, wavelet analysis, BP neural network, NAR neural network, dam deformation prediction