Research On Deformation Monitoring And Prediction Of BP Neural Network Based On Wavelet Denoising

Wavelet analysis is a signal analysis tool that has a good locality in both the time domain and the frequency domain.This feature makes it ideal for handling unsteady-changing signals.At present,GPS technology is widely used in deformation monitoring project,but because of the deformation monitoring technology for the measurement of high precision and GPS technology in data acquisition,signal transmission and results processing is more complex,the resulting deformation signal that deformation signal There is also a noise signal.Therefore,in the process of using GPS deformation monitoring technology,the distortion signal is subjected to wavelet denoising to eliminate the error and get the real deformation.Through different wavelet functions,threshold function,threshold determination rule and wavelet decomposition layer number can be combined into a variety of different wavelet denoising model.In the study of different models,using MATLAB to control the other three factors are the same,only one of the factors Experimental comparison,which analysis of which combination model of the experimental signal denoising effect of the most ideal.In addition,in order to ensure the safety of engineering construction,to prevent the occurrence of engineering accidents,but also the deformation of the project building to monitor the forecast.At present,one of the more commonly used methods of deformation prediction is BP neural network.However,the deformation of engineering buildings is affected by geological conditions,climate change,surrounding construction environment and so on.Therefore,the BP neural network of single model can not meet the requirement of high precision forecasting.In this paper,the wavelet denoising technique is combined with BP neural network,and the wavelet basis function in wavelet denoising is used as the transfer function of hidden layer.A new prediction model is obtained,which is based on BP neural network deformation prediction model based on wavelet denoising,The model can be used to make deformation prediction of the deformation.The main contents and results of this paper are as follows:1?Fourier transform,window Fourier transform and wavelet transform are introduced.The application limitations of Fourier transform and window Fourier transform and the superiority of wavelet transform in signal processing are pointed out.2?The decomposition and reconstruction of discrete wavelet and discrete wavelet are described,and the mathematical characteristics of wavelet base are described.The relationship between wavelet and wavelet is also described.3?This paper introduces the working principle of BP neural network,the key problems in BP neural network application and the construction method of BP neural network model based on wavelet denoising and the steps of its application.4?Combine the specific engineering case,the use of wavelet to do the GPS deformation monitoring data denoising.According to several key problems of wavelet technology application,combined with the evaluation index of wavelet denoising,a way to deal with the ideal denoising of the signal is summed up.Finally,when the signal is processed,the wavelet adopts db10 wavelet function,Threshold determination method,the hard threshold function and the decomposition level is three layers,the denoising effect is the best.5?The BP neural network based on wavelet denoising is used to forecast the deformation of the dam.The experimental results show that the BP neural network forecasting model combined with wavelet technology is much better than the prediction accuracy of a single BP neural network.