| Electromagnetic acoustic emission(EMAE),as an improved nondestructive testing technology,has the advantages of both electromagnetic excitation and acoustic emission testing,and is widely used in the field of nondestructive testing of metal devices.However,the long-term detection has caused pressure on data transmission and storage,and the EMAE signals collected under working conditions are mixed with complex noise,which brings great trouble to the analysis and processing of EMAE signals and affects the subsequent damage identification and life warning of metal components.In view of this,based on the denoising method and Compressed Sensing theory,we took EMAE signal as the research object and focused on the denoising analysis and compressive reconstruction of EMAE signals.Among them,the denoising study assists the compressive reconstruction of EMAE signals to improve the reconstruction accuracy.The specific work is as follows:(1)The EMAE signal collection was carried out and the microstructure changes of the specimen before and after the experiment were observed.Firstly,the technical principle and mechanism of EMAE were analyzed to make theoretical preparations for the development of EMAE experiments.Secondly,the aluminum alloy plate with preset cracks was used as the experimental object to carry out EMAE detection and signal collection,and the test results were analyzed and studied in detail.In order to observe the influence of EMAE detection technology on the detection object,the deformation structure and structure of the sample were analyzed in detail by microscope and scanning electron microscope,and the microstructural changes of the 6061 aluminum alloy specimen before and after the test were compared and analyzed.(2)A denoising method based on Cross Recurrence Theory-Variational Mode Decomposition(CRT-VMD)is proposed for EMAE signals mixed with a large amount of noise,which reduces the reconstruction accuracy of compression perception.Concretely,we used Genetic Algorithm or Grey Wolf Optimizer algorithm to optimize the number of decomposition layers and penalty factors of VMD,so as to realize adaptive decomposition,and then we utilized CRT to analyze the correlation of each modal component obtained after decomposition,and filter and reconstruct them to obtain the EMAE after noise reduction signal.The simulation and experimental noise reduction results proved that the model could effectively remove noise in EMAE signals,and could effectively assist in the realization of high-purity compression of EMAE signals.(3)The Grey Wolf Reconstruction Algorithm is introduced to realize the compression reconstruction of the signal after noise reduction.Aiming at the problem of transmission and storage of a large amount of data,on the basis of studying the theory of compressed sensing,taking the standard sparse signal as the research object,the reconstruction effects of different reconstruction algorithms are compared and analyzed,which provides a basis for the reconstruction method of the EMAE signal.Based on this,compressed sensing research is carried out on EMAE signals,and the reconstruction performance of different reconstruction algorithms and different compression ratio is compared.From the results,the Grey Wolf Reconstruction Algorithm has a higher recovery success rate and is more suitable for compressed storage of EMAE signals.In order to further verify the effectiveness of the reconstructed signal with Grey Wolf Reconstruction Algorithm,the original signal and the reconstructed signal were respectively input into the Artificial Neural Network,Support Vector Machine and Deep Echo State Network models,and the prediction results were compared.The experimental results further confirmed the reliability of the Grey Wolf Reconstruction Algorithm.This technology has important reference significance for the impairment identification,life prediction,and early-warning of subsequent metal devices. |
PDF Full Text Request