Numerical studies for the thermal regime of a high-speed railway tunnel considering piston action on seasonally frozen regions

piston action describes the phenomenon that air at the train nose is pushed forward by the increased pressure and air at the train rear is drawn forward by the decreased pressure when a train passes through a tunnel. The changes of pressure can affect the thermal environment inside the tunnel, and further cause frost damage. In this paper, a fluid-thermal-solid coupled numerical model considering piston action is developed. A high-speed railway tunnel in the northeast of China is taken as an example to explore the temperature distribution laws with computational fluid dynamic (CFD). Afterwards, the effects of air temperature and train velocity on temperature distribution are analyzed. The results show that the piston action can enhance the heat transfer between cold air outside the tunnel and tunnel structure, and can cause more serious frost damage especially at the entrance and exit. The temperature distribution is characterized by three zones, including disturbed zones at two sides of tunnel and undisturbed zone at tunnel middle. The freezing length is closely related to air temperature and train velocity. And also, the lengths are different at vault and rail of tunnel portal, which indicates that the anti-freezing measure should be different at these positions considering the cost. This paper can provide some reference for determining the anti-freezing fortified length of tunnels in cold regions.

Experimental Study of the Dynamic Characteristics of a New Antidrainage Subgrade Structure for High-Speed Railways in Diatomaceous Earth Areas

The experience needed to carry out engineering and construction in diatomaceous earth areas is currently lacking. This project studies the new Hang Shaotai high-speed railway passing through a diatomaceous earth area in Shengzhou, Zhejiang Province, and analyzes the hydrological and mechanical properties of diatomaceous earth on the basis of a field survey and laboratory. Moreover, a new antidrainage subgrade structure was proposed to address the rainy local environment, and field excitation tests were performed to verify the antidrainage performance and stability of the new subgrade structure. Finally, the dynamic characteristics and deformation of the diatomaceous earth roadbed were examined. The hydrophysical properties of diatomaceous earth in the area are extremely poor, and the disintegration resistance index ranges from 3.1% to 9.0%. The antidrainage subgrade structure has good water resistance and stability under dynamic loading while submerged in water. After 700,000 loading cycles, the dynamic stress and vibration acceleration of the surface of the subgrade bed stabilized at approximately 6.37 kPa and 0.94 m/s2, respectively. When the number of excitations reached 2 million, the settlement of the diatomaceous earth foundation was 0.08 mm, and there was basically negligible postwork settlement of the diatomaceous earth foundation. These results provide new insights for engineering construction in diatomaceous earth areas.

Semi-Supervised Deep Learning in High-Speed Railway Track Detection Based on Distributed Fiber Acoustic Sensing

High deployment costs, safety risks, and time delays restrict traditional track detection methods in high-speed railways. Therefore, approaches based on optical sensors have become the most remarkable strategy in terms of deployment cost and real-time performance. Owing to the large amount of data obtained by sensors, it has been proven that deep learning, as a powerful data-driven approach, can perform effectively in the field of track detection. However, it is difficult and expensive to obtain labeled data from railways during operation. In this study, we used a segment of a high-speed railway track as the experimental object and deployed a distributed optical fiber acoustic system (DAS). We propose a track detection method that innovatively leverages semi-supervised deep learning based on image recognition, with a particular pre-processing for the dataset and a greedy algorithm for the selection of hyper-parameters. The superiority of the method was verified in both experiments and actual applications.