Research On Detection Method Of Interface Defect In Composite Bonded Components Based On Ultrasonic Array Technology

Composite bonding components are widely used in aerospace industry.The composite bonding components may be debonded at the interface,posing a great threat to the safety of the components.Therefore,there is an urgent need to detect the defects at the bonding interfaces effectively.Ultrasonic phased array detection technology has great development potential and application prospect in the defect detection of layered composite components,as the sound beams can be deflected and focused in any of direction and position by controlling time delay of each of array elements realizing multi-angle and large-scale scans in detection areas.In the current ultrasonic array modeling and imaging methods,the influence of acoustic directivity and acoustic effects caused by defect types on the acoustic propagation characteristics and imaging effects are not considered.These cause the problems such as that the acoustic field simulation results cannot reflect the actual acoustic propagation characteristics,the imaging signal-to-noise ratio is low,and the optimal utilization of ultrasonic array data cannot be realized.Based on the above research background,this paper carries out research on key technologies of interface defect detection in composite bonded components based on ultrasonic array,aiming at effectively improving the detection ability of interface defect in composite components.The main research contents of this paper include:(1)Research on acoustic field characteristics and modeling methods of ultrasonic array in layered media.Based on the acoustic theory and damage mechanics theory,the acoustic field models of ultrasonic array transducer in layered media with different acoustic impedance are established.The propagation characteristics of acoustic wave in the media under different parameters and the influence on the control performance and detection resolution are studied.For the directivity characteristics of each element in an ultrasonic array,a sound field calculation model that adds a directivity constraint for each element is proposed.The steering sound fields under different directivity constraints are simulated,and the results demonstrate that the directivity of each element significantly impacts the overall steering of sound beams.This demonstrates the feasibility of adding a directivity constraint when modeling sensor arrays.This brings the simulation results closer to actual sound propagation and provides necessary theoretical foundation for subsequent experimental research.(2)Experimental study on interface defect detection of composite bonded components by ultrasonic array.An experimental platform for interface defect detection of composite component based on the ultrasonic array detection system and the three-coordinate motion assistance system was built.The detection and data acquisition mode are experimentally studied.The line scan and full-matrix data acquisition of interface echo data are captured.The A-scan,B-scan and C-scan imaging of interface defects in composite bonding components are realized based on principle of ultrasonic array imaging methods.Aiming at the full matrix data unique to ultrasonic array technology,the total-focusing imaging of interface defects is also realized.(3)Study on the total-focusing imaging method of interface defects in composite bonded components under effective aperture.Aiming at the acoustic effects caused by defect types and the distortion problem of total-focusing imaging caused by redundant information in ultrasonic full-matrix data,a total-focusing imaging method based on correlation criteria is proposed.This method can ensure that the signals in the full-matrix data of at each imaging position have high correlation under the effective aperture and improve the total-focusing imaging performance.Then the ultrasonic signals under the effective aperture is converted to the frequency domain for feature extraction,and the imaging of the interface defect based on the extracted features under the effective aperture is realized,which further effectively improves the imaging effect of the interface defect.(4)Study on detection technology of composite component based on machine learning under effective aperture.Aiming at the problem that traditional ultrasonic array systems are not highly intelligent,and the determination of testing results of the defect basically relies on technicians to make decisions based on imaging results limiting the accuracy of results,an classification model for interface defect based on convolutional neural network with selecting attention area enhancement is proposed on the basis of building and studying the effects of different convolutional neural network models on the classification and recognition performance of the interface defects in the framework of deep learning.The defect detection results based on the convolutional neural network with selecting attention area enhancement model and that based on total-focusing imaging detection results under the effective aperture can be integrated for the decision-making of identification of the interface defect,which can effectively improve the detection accuracy of the interface defect in composite component and the detection performance of the imaging system.