Investigation On Defect Detection Based On Ultrasonic Guided Wave Full Waveform Inversion

Ultrasonic guided wave with the character of long-distance propagation,full cross-section coverage and multi-mode nature is especially suitable for long-range detection and monitoring of elastic waveguide structures such as thin plates and pipes.In order to realize the visual detection of internal defects in the measured area,tomography algorithm has received gradually increasing attention in the field of ultrasonic guided wave detection.The existing tomography algorithm mainly includes travel time tomography or ray tomography,diffraction tomography and its combined form.Ray tomography only uses the arrival time and completely ignores the diffraction effect,resulting in a finite resolution,which is determined by the wavelength and the size of first fresnel zone.Diffraction tomography can achieve a better resolution,but due to the neglect of multiple scattering effects it is only suitable for detecting defects with small or low contrast.In order to reconstruct complex-shaped defects such as corrosion,full-wave fields containing all possible higher-order diffraction and scattering effects should be considered in tomography.Therefore,this paper proposes a full waveform inversion imaging technique for complex structures and complex defects.The main research contents are as follows:Firstly,the fundamental theory of ultrasonic guided wave propagation is introduced,including the derivation of Rayleigh-Lamb equation in thin plate structure and the solution of dispersion curve equation.Secondly,the full waveform inversion theory is studied and deduced in detail,including forward modeling,the calculation of objective function and its gradient,optimization algorithm selection.On this basis,a full waveform inversion algorithm based on L-BFGS method is programmed,and the forward modeling is verified by using the frequency domain pressure acoustic module with finite element software COMSOL.Furthermore,the ultrasonic detection simulation model is established by using the time domain module.Then,full matrix ultrasonic data is collected,and the local optimization algorithm is used to update and iterate,which realizes the defect imaging detection of the simple model and verifies the effectiveness of the full waveform inversion algorithm.Then,for the difficult problem of accurate imaging of complex structures,based on the developed full waveform inversion algorithm,the inversion imaging of different defect types of blade structure is carried out,including internal circular holes with different diameters,open cracks with different orientations,and multiple defects.Due to the poor imaging quality of single frequency inversion strategy,this paper further proposes the use of successive multi-frequency inversion strategy,and the results show that the imaging quality is effectively improved.In order to realize the quantitative detection of defects,the evaluation index of tomography results is established,and the stability of the algorithm is verified by uncertainty analysis of the initial velocity model.Finally,the simulation model of ultrasonic guided wave detection for thin plate structure is established by using ABAQUS.Based on the dispersion characteristics of ultrasonic guided wave,the single mode ultrasonic guided wave is excited in the thin plate and the full matrix detection data are collected.Using the developed full waveform inversion algorithm,the inversion imaging of defects with half a wavelength in thin plate structure is realized,and the feasibility of ultrasonic full waveform inversion is verified.