Development Of Quantitative Detection Device For Interfacial Debonding Of Steel Beams Reinforced By CFRP Plates Based On Percussion Method

As a lightweight,high strength,corrosion resistant and low cost structural reinforcement material,carbon fiber reinforced plastic(CFRP)plate has been widely used in the field of structural reinforcement.In the process of reinforcement and repair of steel structure,the reinforcement with CFRP plate can effectively improve the bearing capacity and service life of steel structure.However,due to the influence of various factors in the construction process and service process,the adhesive layer between CFRP plate and steel structure interface will have a debonding defect,which will seriously affect the reinforcement effect of CFRP plate on steel structure and shorten the service life of the reinforced steel structure.Among the traditional interface debonding detection methods,there are mainly infrared thermal imaging,X-ray analysis imaging,ultrasonic imaging,piezoelectric active sensing,impedance detection and percussion detection.Among them,percussion detection method is widely used because of its simple principle,convenient operation and low cost,but hammering detection method mainly relies on human ears’ subjective experience to judge,and can only realize the location identification of debonding position,but can not realize quantitative identification.In recent years,due to the rapid development of technologies and methods such as deep learning,the CNN(Convolutional Neural Network)with local perception,parameter sharing and powerful feature extraction capabilities has been widely used.In this paper,the Welch PSD(Power Spectral Density)estimation method,MFCC(Mel Frequency Cepstral Coefficients)combined with CNN are mainly used to realize the initial location of interfacial debonding position and the secondary quantification of debonding area,and the above methods are used to develop and test the quantitative detection device of interfacial debonding.This paper mainly carried out the following research:(1)The feasibility of percussion combined with PSD estimation,MFCC and CNN in the location and quantification of interface debonding detection was explored.Based on the traditional frequency damage location method,this paper uses PSD estimation method to extract the frequency features,and realizes the preliminary location of interface disadhesion detection.In addition,the MFCC features of sound signals are extracted and converted into two-dimensional feature maps,and then a two-dimensional CNN model is built.The quantitative identification of interface debonding area is realized through the trained CNN model.Based on the above research method,this paper carried out an exploratory test on the interfacial debonding detection of steel beams reinforced by CFRP plate,which verified the effectiveness of the above method.In addition,the property influence of the hammer head of the percussion tool on the test results was also explored in the exploratory test.(2)A quantitative detection device for percussion interface debonding based on the above exploratory results was developed.On the basis of the above research,the trained CNN model is used to develop a quantitative detection device for percussion interface debonding.Based on the different tapping methods,the automatic percussion type interface debonding detection device and manual percussion type interface debonding quantitative detection device are developed.In addition,in the debugging process,two sets of percussion type detection devices are optimized and improved to improve the use effect of the detection device.(3)The developed interface debonding detection device is tested and verified.After the research and development of two percussive interface debonding quantitative detection devices,the testing and verification of the two detection devices were carried out.The test conditions based on different testing devices are designed,and the interface debonding test and verification test are carried out on the steel beam reinforced with carbon fiber plate.Through testing and verification experiments,the effect of the two detection devices is evaluated.The test results show that the two detection devices can realize the quantitative identification of the interfacial debonding area of the steel beam reinforced by CFRP plate.In addition,based on the problems in the test,the shortcomings of the existing two sets of detection devices and the next improvement direction are explained in the paper.The test and verification test is an important step to promote the application of the detection device to the practical engineering.The detection device developed in this paper provides an important reference and reference significance for the interface debonding detection in the practical engineering.