Research On Micro-Defects Of Carbon Fiber Composite Materials Based On Time-Frequency Analysis Of Backscattered Signal

Carbon fiber composite materials are widely used. With the development of material structure and process, traditional ultrasonic testing techniques encounter some new problems in composite non-destructive testing (NDT) areas.When ultrasonic pulse echo method is used to detect the carbon fiber composite materials, traditional back wall echo attenuation method can’t locate the defects easily and the back wall echo may miss because of the complex shape of the material and attenuation of the ultrasonic echo. This paper provides a new way for ultrasonic inspection techniques of composite materials through researching the principle of backscattered signal with some experiments based on natural science foundation project relative to composite micro-defects ultrasonic backscatter detection.In chapter 1, the background of the research and the significance of the dissertation were discussed. Moreover, main defect types of carbon fiber composites were analyzed. Defects testing technology and digital signal processing technology were analyzed. Finally, main research contents and chapters arrangement were introduced.In chapter 2, ultrasonic propagation characteristics in layered composites were analyzed. The principle of layered composite structure of noise was studied. The first-order resonance frequency of the structure noise was calculated. Low-pass filter characteristics of defects, backscattered signal components and characteristics were studied.In chapter 3, several common signal spectral analysis methods and their shortcomings were described. The advantages of wavelet time-frequency analysis of the backscattered signal processing were discussed. The characteristics of CMOR wavelet were deeply analyzed. The significance of the backscattered signal time-frequency image was described.In chapter 4, a five degree of freedom ultrasonic testing system was introduced. A method based on the regional division of the carbon fiber composite ultrasonic backscattered signal time-frequency analysis was proposed. The trends of the backscattered signal frequency and system noise frequency with the method of three-dimensional time-frequency images and textures sub-graph were analyzed.In chapter 5, a wavelet threshold method for far surface backscattered signal processing was selected. Different defect recognition models based on backscattered signal characteristics in different regions were established. The models were verified well through experiment.