Railway subgrade defect is the serious threat to train safety. Vehicle-borne GPR method has become the main railway subgrade detection technology with its advantages of rapidness and nondestructiveness. However, due to the large amount of detection data and the variety in defect shape and size, defect recognition is a challenging task. In this work, the method based on deep learning is proposed to recognize defects from the ground penetrating radar (GPR) profile of subgrade detection data. Based on the Faster R-CNN framework, the improvement strategies of feature cascade, adversarial spatial dropout network (ASDN), Soft-NMS, and data augmentation have been integrated to improve recognition accuracy, according to the characteristics of subgrade defects. The experimental results indicates that compared with traditional SVM+HOG method and the baseline Faster R-CNN, the improved model can achieve better performance. The model robustness is demonstrated by a further comparison experiment of various defect types. In addition, the improvements to model performance of each improvement strategy are verified by an ablation experiment of improvement strategies. This paper tries to explore the new thinking for the application of deep learning method in the field of railway subgrade defect recognition.
This study presents an innovative, intelligent obstacle avoidance module intended to significantly enhance the collision prevention capabilities of the robotic arm mechanism onboard a high-speed rail tunnel lining inspection train. The proposed module employs a fusion of ORB-SLAM3 and Normal Distribution Transform (NDT) point cloud registration techniques to achieve real-time point cloud densification, ensuring reliable detection of small-volume targets. By leveraging spatial filtering, cluster computation, and feature extraction, precise obstacle localization information is further obtained. A fusion of multi-modal data is achieved by jointly calibrating 3D LiDAR and camera images. Upon validation through field testing, it is demonstrated that the module can effectively detect obstacles with a minimum diameter of 0.5 cm, with an average deviation controlled within a 1–2 cm range and a safety margin of 3 cm, effectively preventing collisions. Compared to traditional obstacle avoidance sensors, this module provides information across more dimensions, offering robust support for the construction of powerful automated tunnel inspection control systems and digital twin lifecycle analysis techniques for railway tunnels.
As the number and length of high-speed railway tunnels increase in China, implicit defects such as insufficient lining thicknesses, voids, and poor compaction have become increasingly common, posing a serious threat to train operation safety. It is, therefore, imperative to conduct a comprehensive census of the defects within the tunnel linings. In response to this problem, this study proposes a high-speed railway tunnel detection method based on vehicle-mounted air-coupled GPR. Building on a forward simulation of air-coupled GPR, the study proposes the F-K filtering and BP migration algorithms based on the practical considerations of random noise and imaging interference from the inherent equipment. Through multi-dimensional quantitative comparisons, these algorithms are shown to improve the spectrum entropy values and instantaneous amplitude ratios by 4.6% and 11.6%; and 120% and 180%, respectively, over the mean and bandpass filtering algorithms, demonstrating their ability to suppress clutter and enhance the internal signal prominence of the lining. The experimental results are consistent with the forward simulation trends, and the verification using the ground-coupled GPR detection confirms that air-coupled GPR can meet the requirements of high-speed railway tunnel lining inspections. A comprehensive GPR detection model is proposed to lay the foundation for a subsequent defect census of high-speed railway tunnels.
22 November 2018 in Journal of Low Frequency Noise, Vibration and Active Control
A vehicle-track rigid flexible coupling-dynamic model is established. The first five vertical vibration natural frequencies of system are solved. The medium wave track irregularities are taken as the excitation. The car body vertical harmonic response vibration characteristics are analyzed. It could be concluded that the second-order natural frequency is the sensitive response frequency. The car body vertical vibration simulation calculation is analyzed under vertical medium-wave track irregularity with influence factors of ballast bed, including the ballast bed vibration mass, ballast-subgrade linear and nonlinear stiffness. Through a large number of simulation analysis and Da Cheng railway test verification, it is observed that when the ballast bed thickness is between 0.35 m and 0.65 m, the amplitude of car body vertical vibration acceleration is maximum, and the stronger the bed thickness suddenly changes, the greater the acceleration of vibration increases. When the ballast bed thickness is not enough, the car body vertical vibration acceleration increases obviously. The physical correlation between two tests is established for the first time through the correctness of car body vertical acceleration data of track inspection that is verified by ground penetrating radar ballast detection.
1 January 2016 in Mathematical Problems in Engineering
A three-dimensional deformable finite element model of wheel-rail system is established. The vertical vibration natural frequencies are calculated by modal analysis. The improved central difference method by ABAQUS explicit nonlinear dynamics module is chosen to calculate the vertical vibration of wheel axle. Through the vibration measurement experiment ofwheel by China Academy of Railway Sciences, the contact parameter is optimized and the model is modified. According to the nonlinear simulation results, the important influence of vertical track irregularity on the wheel set high-frequency vibration is discovered. The advantage frequencies through spectrum are close to the vertical vibration natural frequencies of seventh, eighth, ninth, and tenth mode. When the velocity is 350 km/h, system’s nonlinear dynamic characteristic is higher than the results at 200 km/h. The critical wavelength of vertical track irregularity on the wheel vertical vibration is about 0.8 m. In addition, a particular attention should be paid to one situation that the main dominant frequency amplitude is more than 50% of the acceleration amplitude even if the peak acceleration is low.
28 April 2020 in Advances in Mechanical Engineering
In order to overcome the great interference of many ancillary facilities in the process of tunnel lining detection, a vehicle-borne high-speed railway tunnel lining detection device was designed based on the arc rotation mechanism to improve the detection efficiency and quality. The device could automatically maintain the antenna–lining distance and carry three radar antennas at one time, which detected the tunnel from the side of the catenary frame without stopping. Through the establishment of the mechanism three-dimensional finite element model, the simulation and analysis of natural frequency and vibration characteristics were carried out, and the results were confirmed by the field dynamic test. It is concluded that the dynamic characteristics can meet the technical requirements under the radial and the longitudinal vibration at 5 km h−1, but the detection data quality of the three radar antennas under the radial vibration is quite different. After optimizing the bent arm material to carbon fiber, the dynamic stability of the mechanism is significantly improved through simulation analysis and the detection speed can be increased to 10 km h−1 with a substantial leap.