The impact of karst collapses on railway engineering spans the entire lifecycle of railway construction and operation, with train loads being a significant factor in inducing such collapses. To study the dynamic response characteristics of subgrades in karst areas and to select appropriate monitoring points and indicators for long-term effective monitoring, a numerical simulation method was employed to analyze the vibration response characteristics of the subgrade. A three-dimensional finite element model coupling the high-speed train, ballastless track, and subgrade foundation was established to study the vibration responses of subgrades when the train passes over a subgrade with an underlying soil hole and one without a soil hole. The results indicate that when there was a soil hole, both the dynamic displacement amplitude and vibration acceleration amplitude decreased, while the dominant frequency slightly increased, with the dominant frequency being higher at locations closer to the soil hole. The vibration response at the soil hole location showed significant attenuation, with the attenuation coefficient of dynamic displacement amplitude being higher than that of the vibration acceleration amplitude. Monitoring points were arranged at positions 0 m to 10 m from the toe of the slope, with vertical dynamic displacement, vertical vibration acceleration, the dominant frequency of vertical vibration acceleration, and corresponding amplitude selected as monitoring indicators. These indicators effectively reflect whether soil holes exist within the subgrade and help identify the locations of defects. This study summarizes the dynamic response characteristics of subgrades in karst areas under different conditions, providing a basis for the design and monitoring of railway subgrades in regions prone to karst collapse.

The detection of subgrade distresses in ballastless track railways poses a formidable challenge due to the presence of steel interference caused by the unique characteristics of high-speed rail track slabs and the dense arrangement of the steel reinforcement mesh within them. Here, we aim to examine the influence of varying distribution patterns of steel reinforcement in ballastless tracks on the detection of subgrade distresses using ground-penetrating radar. Through a combination of on-site testing and forward modeling, this paper analyzes the interference of steel reinforcement on the detection of voids beneath the steel using electromagnetic waves. The research findings reveal that incident electromagnetic waves from the ground-penetrating radar experience attenuation near steel reinforcements, with only a fraction able to penetrate the surface layer and propagate into the subsurface through interstitial gaps between the reinforcing bars. Furthermore, this influence diminishes as the spacing between the reinforcing bars increases and the bar diameter decreases. When steel bars are distributed on the upper and lower layers, the detection results of the lower void are most significantly influenced by the interlocking of the steel bars in the two layers. These research results can offer theoretical and technical support for the detection of ailments in high-speed railway subgrades.

This study compares three detectors—Single Shot MultiBox Detector (SSD), Faster Region-based Convolutional Neural Network (Faster R-CNN), and Only Look Once v11 (YOLOv11)—for detecting subgrade distress in GPR radargrams. SSD converges fastest but shows weaker detection performance, while Faster R-CNN achieves higher localization accuracy at the cost of slower convergence. YOLOv11 offers the best overall performance. To push YOLOv11 further, we introduce three enhancements: a Multi-Scale Edge Enhancement Module (MEEM), a Multi-Feature Multi-Scale Attention (MFMSA) mechanism, and a hybrid configuration that combines both. On a representative dataset, YOLOv11_MEEM yields a 0.2 percentage-point increase in precision with a 0.2 percentage-point decrease in recall and a 0.3 percentage-point gain in mean Average Precision@0.5:0.95, indicating improved generalization and efficiency. YOLOv11_MFMSA achieves precision comparable to MEEM but suffers a substantial recall drop and slower inference. The hybrid YOLOv11_MEEM+MFMSA underperforms on key metrics due to gradient conflicts. MEEM reduces electromagnetic interference through dynamic edge enhancement, preserving real-time performance and robust generalization. Overall, MEEM-enhanced YOLOv11 is suitable for real-time subgrade distress detection in GPR radargrams. The research findings can offer technical support for the intelligent detection of subgrade engineering while also promoting the resilient development and sustainable operation and maintenance of urban infrastructure.

To elucidate the evolutionary patterns and influencing factors affecting the stability of the surrounding rock during tunnel excavation in ductile shear zones, this study takes the tunnel section within the ductile shear zone of the mountain area as a case study. Numerical simulation methods were employed to analyze vertical deformation, plastic zone development, and stress redistribution characteristics of the surrounding rock under a three-step excavation approach. Furthermore, this research investigates how burial depth and surrounding rock grade impact stability. The findings indicate that during the tunnel excavation process, the vertical deformation, plastic zone, and extrusion deformation at the face exhibit significant phased characteristics. Notably, the weak zone of the fractured surrounding rock serves as a concentrated area of deformation. Upon completion of the excavation, the maximum settlement observed at the arch top reached 246.07 mm. The plastic zone primarily experienced shear failure and showed a tendency to stabilize after reaching section Y = 35 m during excavation. The burial depth exerts a significant influence on the stability of the surrounding rock. As the burial depth increases from 400 m to 550 m, there is an observable upward trend in the settlement at the top of the arch, uplift at the bottom of the arch, and maximum principal stress; notably, the rate of increase for maximum principal stress remains stable. The instability mechanism of the surrounding rock, primarily characterized by shear failure, has not altered. The classification grade of the surrounding rock serves as a critical factor influencing stability. The vertical deformation scale, extent of plastic zones, and values for maximum principal stress in Grade IV surrounding rock are considerably smaller than those observed in Grade V. Enhanced mechanical properties and integrity within the rock mass can significantly improve stability conditions for surrounding rocks. These research findings provide a theoretical foundation and engineering reference for optimizing support systems in tunnels traversing ductile shear zones.