Research On Eddy Current Detection And Quantitative Evaluation Methods For Aluminum Alloy Component Defects

Aluminum alloy is one of the most widely used non-ferrous metal structural materials in industry,which is used in aerospace,marine engineering,petrochemical and automobile manufacturing because of its low density,high strength and good corrosion resistance.During the smelting and deep processing of aluminum alloys,due to the production environment,process factors and the quality of the original billet and other problems,it is inevitable to produce various defects in the components,which will bring significant security risks to the service structure.However,the accuracy,efficiency and accuracy of defect quantification of aluminum alloy components are still to be improved.Eddy current testing technology based on electromagnetic induction principle is widely used in nondestructive testing of metal components due to its advantages of accuracy,efficiency and convenience.Based on the analysis of eddy current non-destructive testing,this thesis studies the methods of eddy current testing and quantitative evaluation of aluminum alloy component defects.The main research work is as follows:(1)In viewing of the complexity of eddy current testing of aluminum alloy components,a simulation analysis model of eddy current testing was constructed,and the influence of the selection of scanning probe,the parameters of coil diameter,excitation frequency and coil diameter on the defect detection of aluminum alloy components was analyzed on the basis of the simulation model.Through the simulation analysis,it can provide guidance for the parameter setting of the actual eddy current detection system,so as to improve the accuracy of defect detection of aluminum alloy components and the accuracy of defect edge identification.(2)Aiming at the problem of noise interference in eddy current detection of defects in aluminum alloy components,a sparse noise reduction method based on dictionary learning for eddy current detection signals was proposed to improve the signal-to-noise ratio of signals.According to the statistical characteristics of defect eddy current detection signal,the sparse representation dictionary of eddy current signal was constructed by using K-means singular value decomposition(K-SVD)dictionary learning through statistics and learning.The dilution representation dictionary constructed by learning can make the eddy current signal more accurate and sufficiently sparse.The experimental results show that compared with the wavelet de-noising method,the K-SVD dictionary learning de-noising method improves the signal-tonoise ratio by 5dB~10dB under multiple noise intensity interference,which proves the effectiveness of the method in pre-processing the signal de-noising of the eddy-current detection of defects in aluminum alloy components.(3)Focusing on the problem of nonlinearity and instability within eddy current testing signal of defects in aluminum alloy component,which causes the difficulty in feature extraction,and problem about accuracy of classification and recognition of the defect,an eddy-current detection and quantitative analysis method based on Kernel Principal Component Analysis(KPCA)and Extreme Learning Machine(ELM)was proposed.This method firstly adopts the kernel based principal component analysis method to extract the features of the eddy current signal,and secondly adopts the ELM based defect classification method while considering constructed defect feature information to realize the accurate and fast classification of defects.Finally,the influence of resistance,reactance and impedance signal on the defect quantitative analysis was studied.The resistance and reactance signals are extracted respectively and the defect quantitative analysis is carried out based on the least square linear fitting method.The experimental results show that the method can accurately identify and classify the defects of components,and the relative errors of defect length and depth are within ± 10% and ± 8% respectively in the quantitative analysis of defects.(4)To solve the problem that the detection efficiency and imaging quality of defect C-scan imaging cannot be satisfied simultaneously,a fast imaging detection method of aluminum alloy component defects based on Compressed Sensing(CS)is proposed.Under the theoretical framework of CS,the sparse representation of eddy current signal,the compression observation and the reconstruction of sparse signal are analyzed theoretically,then the sparse eddy current imaging detection of defects in aluminum alloy components is realized experimentally.The results show that when the number of compressed observations is 4 times of signal sparsity,the Root Mean Squared Error(RMSE)of the reconstructed defect image and the original C-scan image is less than 0.005,and the relative error of quantitative analysis of defect length through the reconstructed image is within ±5%,which has good robustness and applicability.The thesis in this paper is defects of aluminum alloy components,optimization analysis of detection parameters,pre-processing research of eddy current signal noise reduction,identification,classification and quantitative analysis of defect signals.Sparse fast imaging detection was completed based on eddy current nondestructive testing technology,and theoretical study on defect detection of aluminum alloy components was enriched.The application and development of eddy current testing and quantitative evaluation of aluminum alloy defects are promoted.