Study On Vibro-acoustic Modulation For Detecting Debonding Defects Of CFRP-steel Hybrid Structure

Carbon Fiber Reinforced Plastic(CFRP)has excelent corrosion resistance,fatigue resistance and damage safety.As a reinforcement material,it is widely used in aerospace,civil engineering,wind power blades,automobile trains and other fields to form CFRP-steel hybrid structure.The effective bonding of CFRP and the reinforced structure is a necessary condition to ensure the reinforcement effect.However,in the reinforcement construction and later use,it is easy to have debonding defects,which seriously affect the structural strength,or even cause structural failure and accidents.The debonding defect of CFRP-steel hybrid structure is harmful and difficult to be observed.Therefore,it is of great significance to realize effective nondestructive testing of CFRP-steel hybrid structure to ensure its service safety.Vibro-acoustic Modulation(VAM)is a nonlinear ultrasonic detection method with strong recognition ability and adaptability,especially for closed defects,which has attracted extensive attention from scholars at home and abroad.In this paper,starting from three aspects of finite element simulation,actual experiment and deep learning algorithm,the vibration tone detection technology for the non-destructive testing of the debonding defect of CFRP-steel hybrid structure is studied,so as to provide support for the safe and healthy service of CFRP-steel structure.The main research contents of this paper are as follows:(1)Numerical simulation study on Vibro-acoustic Modulation for detecting debonding defects of CFRP-steel hybrid structureBased on the "open and close" theory of the Vibro-acoustic Modulation technology,the mechanism of the "open and close" phenomenon in the Vibro-acoustic Modulation process of CFRP-steel hybrid structure was revealed by ABAQUS.The Vibro-acoustic Modulation detection effect of debonding defects of different sizes when the thickness of the adhesive layer is 0.0002 m,0.0004 m and 0.0006 m is explored,and the influence of the thickness of the adhesive layer on the detection effect is analyzed.The effect of vibration tone system on the detection of defects in different positions and sizes was compared and studied,and the influence of excitation relative position of high frequency ultrasonic signal and low frequency vibration signal on the detection effect was explored.The relevant research results provide guidance for subsequent experimental research.(2)Experimental study on Vibro-acoustic Modulation for detecting debonding defects of CFRP-steel hybrid structureAn experimental system for Vibro-acoustic Modulation was built,and the test pieces of CFRP-steel hyrid structure with intact specimen and debonding defects were made to carry out Vibro-acoustic Modulation experiments.Based on the experimental results,the optimization setting principles of key detection parameters such as low-frequency excitation vibration signal frequency,low-frequency excitation vibration signal amplitude,high-frequency excitation ultrasonic signal frequency and high-frequency excitation ultrasonic signal amplitude are studied.The influence of the size of debonding defect and the relative position between the high-frequency excitation ultrasonic signal and the low-frequency vibration signal on the detection results was explored to verify the validity of the numerical model and lay a necessary foundation for the formation of the Vibro-acoustic Modulation technology for the debonding defect of CFRP-steel hybrid structure.(3)Deep learning detection method of Vibro-acoustic Modulation for debonding defects of CFRP-steel hybrid structureThe test data obtained from the experiment were preprocessed as input samples,and an intelligent recognition model suitable for the Vibro-acoustic Modulation of CFRP-steel hybrid structure was built based on two deep learning algorithms,convolutional neural network(CNN)and stack self-encoding(SAE),to realize the intelligent detection and classification of the debonding defects.Based on the detection results,the important parameters of the two intelligent recognition models are optimized and improved.The optimized model is used for comparative analysis,and it is found that the SAE model has higher detection accuracy and efficiency than the CNN model for the vibration and acoustic modulation detection of CFRPsteel hybrid structure debonding defects,and is more suitable for application in engineering practice.Therefore,the effective detection of debonding defects of CFRP-steel hybrid structure by using Vibro-acoustic Modulation technology is of great significance for the safe service of CFRP-steel hybrid structure and the realization of structural health monitoring.