Research On Quantitative Detection Of Composite Materials Based On Lock-in Thermography

Composite materials are widely used in aerospace,aircraft manufacturing,automotive and other fields due to their excellent characteristics.Because of the heterogeneity and anisotropy,composite materials are relatively easy to produce defects in the manufacturing process and application.Therefore,it is extremely important to carry out the research on nondestructive testing of composite materials.Traditional detection method s mainly rely on ultrasonic testing and X ray detection,and there are many problems,such as long cycle,low efficiency and high cost.Infrared nondestructive testing technology is better than the above two detection methods in detecting speed and display effect.In this paper,Lock-in infrared nondestructive testing technology is studied.The one-dimensional heat transfer process of a single layer and multi layer medium with sinusoidal heat flow are analyzed.The three-dimensional heat conduction model is analyzed by the finite element method,and the influence of modulation frequency,defect depth and defect size on the surface temperature signal and phase difference are obtained,which provides a theoretical basis for the later experiment and quantitativ e evaluation of defects.The experimental platform of the optical lock-in infrared nondestructive testing system is built,and the acquisition module is developed.The traditional denoising algorithm and Savitzky-Golay filter are used to process the image sequence to filter the infrared image sequence noise.The result shows the superiority of Savitzky-Golay filtering denoising.The image sequences are processed by Four-point correlation method,Fourier transforms algorithm to obtain the amplitude and phase image.Principal component analysis is used to process the image sequence to extract the signal.Compared the effect of three kinds of signal extraction methods,the experimental results show that Four-point correlation method has better extraction effect on phase and amplitude,while the image obtained by principal component analysis has higher signal to noise ratio.The quantitative evaluation of the size and depth of the defects in infrared nondestructive testing is studied,and the phase shearing algorithm is used to judge the size and position of defects.Due to the poor performance of phase shearing algorithm in irregular defect detection,the region growing algorithm is proposed.Aiming at the judgment of defect depth,the popular blind frequency d etection technology is introduced,and the method of neural network learning prediction is put forward to predict the depth.The hardware and software platform of lock-in infrared nondestructive testing system is built and the experimental study of composi te material detection is carried out.Two most common defects of composite materials(skin honeycomb debonding and carbon fiber delamination)are made and quantitative detection is carried out for the sample species.The experimental results show the feasibility and superiority of lock-in infrared nondestructive testing platform.