Based On EEMD-WNN Combination Model On Dam Deformation Forecasting Methods

During the operation of the dam,the deformation law of dam deformation can be grasped and the accurate prediction is made,which is of great significance to dam safety and disaster control.At present,BP neural network model is widely used in non-linear,non-stationary time series prediction,but its local minimization and slow convergence rate and other issues,resulting in prediction accuracy is not high.The wavelet neural network after wavelet neural network combines the wavelet time-frequency local analysis ability and the neural network self-organization and self-learning ability,which can effectively solve the problem that the BP neural network algorithm is easy to fall into the local minimization and the convergence rate is slow Problem,greatly improving the accuracy of deformation prediction.On the other hand,there are many ways to analyze the original deformation data directly,and when the data is non-stationary,it is difficult to obtain satisfactory results.Therefore,if a certain data processing method can be used to fully excavate the feature information implied in the original sequence,the non-stationary signal is transformed into a stationary signal,and then the prediction model is established.It is expected to improve the prediction accuracy,and the empirical mode decomposition method Is a data processing method that can convert non-stationary signals into stationary signals.Based on the above research,this paper combines the empirical mode decomposition method and the wavelet neural network to analyze and forecast the dam deformation information.Firstly,the deformation time series is decomposed into deformation components with different physical scale features by using the empirical empirical modal decomposition method in order to reduce its nonstationarity.Then,in order to reduce the number of modeling and improve the prediction efficiency,the run-length method is used to reconstruct the three components with high,medium and low frequencies,and the WNN prediction model is established.Finally,the predicted values are Final prediction results.The comparison between GM(1,1)model,BP neural network model and WNN model shows that the algorithm has high accuracy and can be used for dam deformation prediction.