Research On Detection Of Tunnel Lining Disease And Safety Evaluation Method By Ground-Penetrating Radar And Deep Learning

With the increase of the number of tunnels put into operation in our country,the safety and reliability of tunnel lining structure are becoming more and more prominent.During the operation period,the tunnel lining structure is easy to produce cracks,voids,seepage and other diseases due to the complex geological environment,bad natural climate and changeable internal and external loads.At present,there are some problems in the research of tunnel lining structure disease detection and safety assessment in our country,such as low automation efficiency,large artificial subjective factors and imperfect nondestructive testing technology.In order to solve the above problems,this paper proposes a tunnel lining disease detection and safety evaluation method based on depth learning and ground penetrating radar.The research contents are as follows:First,this paper carries out the research of lining disease intelligent inversion based on deep learning and ground penetrating radar data,analyzes the characteristics of lining structure disease data collected by ground penetrating radar,and develops the intelligent inversion technology of disease data based on convolution neural network.After that,the research of lining disease detection technology based on deep learning and intelligent inversion results of ground penetrating radar is carried out,and 30 different structures Faster R-CNN models are designed for parameter and structure optimization of the model.Finally,this paper studies the tunnel lining safety evaluation method based on fuzzy evaluation method,and puts forward the comprehensive fuzzy evaluation model suitable for tunnel lining structure safety evaluation.The results show that the intelligent inversion method realizes accurate and stable inversion of lining-wall rock interface,steel bar position,lining crack,lining leakage,and lining voiding based on GPR data.The optimized faster R-CNN model has the ability to identify and locate lining cracks,lining leakage,lining voiding and lining structure reinforcement.The results of stability analysis show that the disease identification of faster R-CNN is not affected by lining structure.The engineering practice results show that thetunnel lining safety evaluation method based on fuzzy evaluation method can carry out quantitative evaluation and safety prediction of lining structure.