Research On Ultrasonic Detection And Full Matrix Imaging Of Multilayer Discontinuous Impenance Structure

Multilayer discontinuous impedance structure(hereinafter also referred to as multi-layer structure)is usually formed by special processes such as coating and cementing of different materials.Due to the different material properties of each layer,its acoustic impedance has the characteristics of discontinuity.This kind of structure is generally used as the protective shell of airtight vessel,which is made of multi-layer thin plate structure.It is widely used in aerospace,special equipment,petrochemical industry and other industrial fields.The quality of products directly affects its safety in the process of use.As a conventional nondestructive testing method,ultrasonic testing is widely used in the industrial nondestructive testing of homogeneous materials because of its advantages of strong penetration,high detection sensitivity and low detection cost.However,for the heterogeneous multi-layer composite structure component,due to the discontinuity of acoustic impedance at the bonding interface,multiple reflection and scattering phenomena will be formed,As a result,serious signal aliasing and dispersion will occur,which further leads to the inability to effectively separate ultrasonic signals carrying structural state information.In addition,due to the discontinuous impedance structure,especially from the high acoustic impedance into the low acoustic impedance structure,the ultrasonic signal is severely reflected,resulting in weak information carrying the state of the deep structure,which affects the application of ultrasonic detection technology in the detection of multi-layer composite structures and promotion.Especially,if ultrasonic imaging is employed,it is difficult to perform imaging detection of multi-interface defects due to the weak and severe aliasing of the signal.Aiming at the above problems,this paper studies the ultrasonic detection and full matrix imaging method of discontinuous acoustic impedance multilayer structure defects.Aiming at the complex ultrasonic propagation mechanism in the multilayer discontinuous acoustic impedance structure,the frequency response model and dispersion model of ultrasonic signal transmission are established.Based on the characteristics of the multilayer composite structure and the characteristics of ultrasonic transmission,this paper focuses on the ultrasonic characterization of the interface debonding state from the aspects of signal attenuation and frequency diffusion.Based on the signal transmission theory of linear systems,the structure is regarded as a cascade of different linear systems,and the states of different interfaces are described by solving the frequency response function.The amplitude-frequency characteristic curve,phase-frequency characteristic curve and echo spectrum of the four-layer composite structure SRRR(steel-rubber-rubber-rubber)board at different interfaces are obtained by using the frequency response model;In this paper,the ultrasonic signal characteristics of metal-non-metal composite structural component are studied.The research results show that under the normal incidence of ultrasonic waves,for the metal-non-metal composite structure,if the attenuation factor of the material is ignored in the numerical simulation solution process,when the composite structure is debonded at different interfaces,its amplitude-frequency characteristics,phase-frequency characteristics and the frequency spectrum will change slightly;if the attenuation factor of the material is considered,the amplitude-frequency characteristics,phase-frequency characteristics and frequency spectrum are basically the same when different interfaces are debonded,that is,it is difficult to solve the change characteristics of interface performance by using the frequency response model.For the multilayer thin-plate composite structure,due to the selective effect of the thin plate on the wavelength,a guided wave will be formed after the ultrasonic incident.Based on the characteristics of guided waves,this paper uses the semi-analytical finite element method to construct the ultrasonic propagation equation of the multilayer bonded structure,solves the propagation characteristics of the guided wave in the multilayer plate structure,and derives the frequency dispersion characteristics of the multimodal guided wave.The change in dispersion properties was used to evaluate the degradation of the interface quality of the multilayer bonded structure.This paper takes the detection of the solid rocket motor casing multilayer structure as an example,and uses the dispersion model to obtain the dispersion characteristic curve when the interface quality of the structure decreases.Aiming at the problem that the detection rate of deep interface debonding defects in multilayer structures is low in the time domains and frequency domains,a defect identification method based on the combination of wavelet packet transform and machine learning algorithm is proposed.According to the different energy loss of interface debonding at different depths,the method first uses wavelet packet transform to extract the corresponding energy characteristic signals of different interface debonding defects,and then uses the obtained energy characteristics as the input vector of the machine learning algorithm for classification and identification.where these energy 70% of the data is used as the training set,and the remaining 30% is used as the test set.In this paper,three machine learning algorithms,K-neighbor(KNN),random forest(RF),and support vector machine(SVM),are used or signal classification.Through the experimental verification of ultrasonic detection of SRRR board,the classification accuracy of KNN,RF,and SVM reached 85.62%,90.66%,and 95.33%,respectively.The above method can accurately identify different interface debonding defects,but the detection accuracy of tiny hole defects in the layer below the surface layer is low.The ultrasonic full matrix capture(FMC)technology collects and superimposes all the information of the inspected medium.Using full matrix imaging inspection can reconstruct the image of deep hole defects,but full matrix imaging method has two main problems in multilayer structures,its low imaging efficiency and its inability to directly image multilayer structural defects.Aiming at the low efficiency of ultrasonic phased full matrix imaging,a full matrix capture ultrasonic focusing imaging method based on finite element model is proposed,which includes half matrix focusing imaging method and simplified matrix focusing imaging method.Based on the principle of acoustic reciprocity,the half matrix focusing method extracts the upper triangle or lower triangle data from the whole matrix data for computational imaging.Simplified matrix focusing method is based on the different energy weights of different ultrasonic array elements to the imaging region,part of trapezoidal data adjacent to the main diagonal data is extracted for imaging.These two methods simplify the calculation process of full matrix data capture and full focus imaging.Through the simulation experiment on the imaging detection of monolayer steel medium,it is verified that the defect SNR of the half matrix imaging method and the simplified matrix imaging method has little difference with the defect SNR of the full matrix imaging method,but the imaging efficiency is improved.Aiming at the problem that ultrasonic phased full-matrix focused imaging cannot be directly detected in multilayer discontinuous impedance structures,this paper proposes a total focusing imaging method based on the root-mean-square(RMS)velocity by referring to the method in seismic data processing.The principle of root mean square velocity in seismic data processing is studied,and the root mean square velocity model in different formation media is introduced into ultrasonic signal processing of discontinuous structure,and the total focus imaging,half matrix focus imaging and simplified matrix focus imaging of multilayer structure are realized,then compared with conventional total focus imaging of multilayer structures.Through the experimental verification of the multilayer structure(organic glass-aluminum)imaging detection,the post-processing imaging method based on the root mean square velocity can reconstruct the defects.The imaging results show that The defect signal-to-noise ratio of the full-matrix imaging method based on the root mean square velocity and the defect signal-to-noise ratio of the traditional multi-layer full focus are both higher and the signal-to-noise ratio of the two is similar.The defect signal-to-noise ratio of the half-matrix imaging method based on rms velocity and the defect signal-to-noise ratio of the simplified matrix imaging method are both poor.However,the half-matrix imaging method and the simplified matrix imaging method based on the rms velocity have relatively high computational efficiency.The full-matrix imaging based on RMS velocity has the lowest imaging efficiency.Computational efficiency of f traditional multilayer total focus method is in the middle.