Automatic Recognition And Damage Mechanism Of A Shield Tunnel With Hidden Defects Behind The Concrete Lining

Hidden defects behind the concrete lining of a shield tunnel may result in apparent damage or even collapse of the tunnel during its operation.Ground-penetrating radar(GPR)is widely used for non-destructive detection of hidden defects behind railway tunnel linings,including non-compactness,air-and water-filled voids,fractures,and insufficient lining thickness.Deep learning models have achieved success in image recognition and have shown great potential for interpretation of GPR data.On the other hand,it is necessary to predict the damage modes of a shield tunnel with cavities,and accordingly reinforce vulnerable areas of the tunnel.In this paper,the intelligent detection of hidden defects and damage modes of hidden defects behind subway tunnel linings are analyzed.The main research contents and results are as follows:(1)An improved least square generative adversarial networks(LSGAN)model which employs the loss functions of LSGAN and convolutional neural networks(CNN)to generate GPR images.This model can generate high-precision GPR data to address the scarcity of labelled GPR data.We evaluate the proposed model using Frechet Inception Distance(FID)evaluation index and compare it with other existing GAN models and find it outperforms the other two models on a lower FID score.In addition,the adaptability of the LSGAN generated images for GPR data augmentation is investigated by YOLOv4 model,which is employed to detect rebars in field GPR images.It is verified that inclusion of LSGAN generated images in the training GPR dataset can increase the target diversity and improve the detection precision by 10%,compared with the model trained on the dataset containing 500 field GPR images.(2)An automatic recognition algorithm is proposed for detecting hidden defects,as well as estimating lining thickness in GPR B-scan images of railway tunnels using a You Only Look Once for Panoptic driving perception network(YOLOP)model.Firstly,a real GPR dataset of underground objects is employed for transfer learning to enhance the robustness of the prediction model.Secondly,a dataset containing 2,245 real GPR B-scan images of railway tunnel linings is created.Thirdly,the established dataset is used to retrain the YOLOP model and assess the prediction accuracy for identifying hidden defects and lining thickness in GPR images.Field experiment results demonstrate that the trained hidden defects recognition algorithm achieves an average accuracy of 80.6%,with a detection time of only 0.02 seconds for a GPR image sizing of 1024 × 512.The proposed algorithm is efficient and reliable for automatic recognition of hidden defects behind railway tunnel linings.(3)Migration can collapse diffractions and reconstruct the geometries of cavities behind the tunnel,thus aiding accurate interpretation of the GPR data.However,clutters from a segment joint could be misjudged as a cavity in the reconstructed GPR images.A band-pass filter is designed to effectively remove interfering joint clutters.Laboratory experiments were carried out in the circumference and longitudinal directions on two full-scale shield tunnel models,in which cavities were buried behind the tunnel lining.Different migration algorithms are tested on the laboratory GPR data,and the results show that the reverse time migration(RTM)algorithm outperforms traditional diffraction stack and Kirchhoff algorithms.RTM can reconstruct both the accurate geometry and location of the cavity behind the lining.The phase of the reflection signal from an air-filled cavity is opposite to that of a water-filled cavity.A portable inspection device is designed to acquire high-quality GPR data in a real shield tunnel.A field measurement in a subway tunnel verifies the effectiveness of the GPR system and the RTM algorithm,and two cavities were found behind the vault and shoulder of the tunnel,respectively.It is concluded that the RTM of GPR data can be used for detecting and characterizing cavities behind shield tunnel segments.(4)The damage modes of shield-tunnel models with cavities at different locations and sizes behind the lining constructed by a 3D printing technique.To consider the stratum-structure interaction,the tunnel models are created with grout-layers prefabricated between lining and soil.The 3D point cloud technique is then applied to observe the damage modes of the tunnel linings.Experimental results show that the damage modes of the shield tunnel with cavities contain concrete crack,concrete spalling,segment misalignment,and lining crush.The safety status of the shield tunnels can be divided into safe,dangerous,and failure stages according to the variations in internal forces.Cavities at the tunnel crown and shoulder impose a substantial impact on the lining structure.Cracks propagating across three or more segments result in mutual compression between segments,forming a crack mesh,and consequently leading to concrete spalling.The tunnel lining undergoes a failure mode of segment misalignment when the cavity angle(size)is greater than 45°.As the volume of the cavity increases,the tunnel lining transitions to a failure mode of lining crush.