As the operating speed of high-speed trains increases to 300–400 km/h, trains entering enclosed tunnels generate a significant piston effect, producing strong compression and expansion waves. These pressure waves can cause substantial pressure differences between the interior and exterior of the train, thereby reducing passenger comfort. Shock waves at tunnel portals can also generate intense noise and threaten the stability of surrounding slopes. Meanwhile, the substantial increase in aerodynamic resistance leads to higher energy consumption. To address these issues, this study focuses on fundamental fluid mechanics theories, the Navier–Stokes (N–S) equations, one-dimensional pressure-wave propagation, and wave transmission and reflection. Based on these principles, the aerodynamic phenomena occurring throughout the entire process of a train passing through a tunnel are examined, with particular attention to the coupled variations in airflow velocity and pressure. Drawing on these theories and previous tunnel optimization studies conducted in China and abroad, this study summarizes structural optimization approaches for railway tunnels, including tunnel cross-sectional designs and portal buffer structures. The results show that increasing the tunnel clearance and reducing the blockage ratio can weaken air compression when a train enters the tunnel. Gradually varying tunnel portals and buffer hoods can reduce the pressure gradient at the front of the compression wave, while perforated buffer structures can further mitigate micro-pressure waves at the tunnel portal by diverting compressed air. This study can provide a theoretical reference for the aerodynamic design of ultra-high-speed railway tunnels and the standardization of buffer structure design for high-speed railway tunnels.

Under the background of digital transformation in the construction industry, Building Information Modeling (BIM) and artificial intelligence (AI) have become an important technical path to promote the intelligent upgrading of the construction industry. AI-BIM integration improves design efficiency, project control quality and building full lifecycle management capability. Fragmented data and poor interoperability hinder its widespread use. This paper conducts a literature review to summarise AI-BIM applications in automated design, construction management, operation and maintenance optimisation, and cross-stage collaboration. This paper further looks into how key technologies like machine learning, computer vision and NLP can be integrated. It also sums up current challenges and future research directions. Studies show that AI can improve BIM's ability in data processing, scheme design and smart decision-making. In the future, combining AI with digital twins and large language models will break existing limits and support the high-quality development of intelligent construction.