Research On Quantitative And Visualization Methods Of Metal Surface Defects Based On Laser Ultrasound

In recent years,laser ultrasonic detection technology for metal surface defects is developing from relatively simple defect detection to nondestructive evaluation,which requires not only detecting whether there are defects in materials,but also evaluating the depth,location and size information of defects.This paper focuses on the quantitative characterization method of metal surface defect depth and the visualization method of defect location and size to realize nondestructive evaluation of surface defects.In this paper,the working principle of the laser ultrasonic detection system for metal surface defects is firstly described,and a defect detection experimental system based on thermoelastic effect is established.Then,aiming at the problem of low signal-to-noise ratio of ultrasonic signals collected during laser ultrasonic defect detection experiments,a combined denoising algorithm combining complementary total empirical mode decomposition and wavelet soft threshold is proposed and compared with other algorithms.The results show that the proposed denoising algorithm not only avoids the problem of poor denoising performance of wavelet threshold method under low signal-to-noise ratio,but also overcomes the modal aliasing problem existing in EMD decomposition.The range of IMF components to be discarded is objectively defined by using the characteristics of autocorrelation function,which effectively improves the signal-to-noise ratio and root mean square error of ultrasonic signals and avoids the problems of energy loss and signal distortion.Compared with other algorithms,it has better adaptability and stability,which lays a good foundation for quantitative characterization of surface defect depth.In order to quantitatively evaluate the depth information of metal surface defects,three analysis methods based on reflected echo center frequency,reflected echo wavelet packet frequency band energy and transmitted wave time-frequency domain distortion characteristics are adopted to extract the feature quantity that can characterize the depth information of surface defects,and a fitting expression that can retrieve the depth information of surface defects is obtained,which is verified by experimental data.The results show that the errors between the surface defect depth information obtained by the above three inversion methods and the real value are very small,which can effectively quantify the surface defect depth.In order to visualize the location and size of metal surface defects,a visual model of aluminum plate surface defects is established based on k-wave photoacoustic toolbox.Based on this model,a series of one-dimensional surface acoustic wave data are obtained by roughly scanning the region of interest of aluminum plate with array sensors.Then,the wave field visualization method is used to process the one-dimensional data to realize the visualization of the approximate location of the defects in the aluminum plate.Finally,aiming at the general distribution area of the defects,the array sensors are used to carry out fine scanning,and the energy analysis method is used to analyze all surface acoustic waves received in the scanning area,thus realizing accurate visualization of the positions and sizes of the defects.