Application Research Of Wavelet Transform In Deformation Monitoring Data Denoising And Information Extraction

Because of its advantages of multi-resolution analysis and time-frequency localization,wavelet transform is suitable for denoising deformation monitoring data of non-stationary series,which contains a plurality of frequency components,and extracting deformation information.In the use of wavelet transform to denoise the deformation monitoring data,different kinds of deformation monitoring data have different characteristics(such as sampling rate,the degree of noise pollution,etc.),therefore it should select different denoising parameters,such as wavelet basis,optimal decomposition order and the threshold value.This problem has been the focus of the study.At present,many experts and scholars have done a lot of research on the choice of the optimal decomposition order and threshold value,while there is not a systematical and normative selection standard for optimal wavelet basis.On the other hand,the deformation information of engineering body is often manifested as changes in frequency components of monitoring data.Based on this fact,the deformation information extraction in deformation monitoring data is to get the position of signal singularity of signal frequency,however in the process of data analysis using single-band reconstruction algorithm,frequency aliasing which rooted in the Mallat algorithm would lead to a wrong extraction of characteristic frequency,and the length of reconstructed data will also be changed due to the convolution with the filter.Aimed at the two problems,based on consulting a large number of literature at home and abroad,this thesis attempts from the energy and entropy theory to put forward an fusion index for selecting optimal wavelet basis,which can to a certain extent guide different deformation monitoring data to choose suitable optimal wavelet basis.At the same time,thesis studied the reason of frequency aliasing in the Mallat algorithm,and put forward an improved anti-aliasing single sub-band reconstruction algorithm based on improved single-band reconstruction algorithm to help solve the problems of frequency aliasing and signal length variation.Using the method proposed in this paper,the simulation data is processed by wavelet threshold denoising and feature information extraction.Through the analysis,it verifies the reliability and practicability of the fusion index of wavelet selection,and it proves that improved anti-aliasing single sub-band reconstruction algorithm can overcome frequency aliasing and ensure that the signal length is not affected by the convolution with the filter.Finally,the two methods are applied to the measured data processing of deformation monitoring,and the satisfactory results are obtained.