Research On Moisture Damage Identification Method Of Asphalt Pavement Based On Air-coupled Ground Penetrating Radar

Asphalt pavement moisture damage refers to the disease that occurs due to water,and most of the diseases that occur on pavements are related to moisture damage.Ground-penetrating radar is an effective method for pavement disease detection.Most of the existing moisture damage detection methods use ground-coupled ground-penetrating radar,and the detection process requires closed roads,which is inefficient,time-consuming,and not suitable for longdistance road detection.The air-coupled ground-penetrating radar maintains a certain height with the ground and can make the whole detection fast and smooth,so the detection parameters and signal identification methods of air-coupled ground-penetrating radar need to be studied in depth.Based on the principle of finite difference in time domain,a numerical simulation model of moisture damage was built by gpr Max.In order to obtain the best overhang detection height,the moisture damage echo signal under different heights is studied by using 2.3GHz air-coupled antenna,and the experimental results find that the coupling occurs between the ground reflected wave and the direct wave when the overhang height H≤8cm,when H=4cm,the relative position error between the maximum amplitude in the coupled wave and the real ground is 1.19%,H=4cm can be used as the best hovering height for 2.3GHz air-coupled ground-penetrating radar;under this height,the simulation model of moisture damage with different water content was obtained by changing the dielectric constant of water in the simulation model,and the echo signal of moisture damage with different content was obtained,The results show that the2.3GHz antenna can be used to detect moisture damage with a volumetric water content of10%-50% and depth in 5cm-10 cm at H=4cm.A model of pavement containing moisture damage was built indoors using asphalt rutting plates,and a verification experiment was conducted using a self-made carrier cart carrier ground-penetrating radar system,and the results showed that the ground could be found in the echoes obtained at the suspension height H=4cm,and the moisture damage with a volumetric water content of 10%-50% at that suspension height was significantly different from the normal data.The effectiveness of the air-coupled ground-penetrating radar for moisture damage detection was proving from the model experiments.To address the problem of large amount of manually interpreted data of moisture damage,a dataset of moisture damage was established by using indoor experimental data,and an intelligent recognition model of moisture damage was established by using time-frequency domain features as feature indicators and particle swarm algorithm as optimization algorithm.The results show that the recognition accuracy of the SVM model is 92.25%,and the recognition accuracy of the PSO-SVM model is 94.88%.It is proved that the moisture damage and normal data measured under the air-coupled condition can be effectively classified by using the time-frequency feature index and SVM modeling,and the optimization of SVM parameters by PSO algorithm can greatly improve the recognition accuracy of the model.The moisture damage identification model was further verified using actual pavement data,and the slip window method was proposed to identify the depth of moisture damage.The results show that the accuracy of SVM model is 92.51% and the accuracy of PSO-SVM is 97.67%,and the slip window method finds out the moisture damage location effectively.The practical value of PSO-SVM modeling method and slip window method in engineering applications is proved.