Abstract
The variation in wind speed outside tunnels in cold regions significantly impacts the tunnel temperature field. This study introduces the Measurement Correlation Prediction (MCP) method to establish a wind dataset for the external environment of tunnels in cold regions. The correlation model is based on the Weibull scale method. Results indicate that for the SSW (dominant wind direction) sector, the relative error between the wind data obtained from the Weibull scale model and actual wind data is 4.88%, demonstrating the high reliability of the model. According to the wind dataset, the average annual wind speed for the subsequent 30 years is predicted to be 3.46 m·s(- 1), and the maximum annual wind speed with a 30-year return period is 4.52 m·s(- 1). These findings highlight the risk of frost damage in the drainage ditch at the bottom of the tunnel, providing crucial information for the prevention and control of frost damage during tunnel operation. Additionally, this study investigates the influencing factors of frost damage in tunnels in cold regions, elucidating the relationships between the thermophysical properties of surrounding rock, ground temperature, airflow factors, and the frost depth at the bottom of tunnel (X). A neural network-based frost damage prediction model was developed using these factors as inputs. The model offers valuable insights for enhancing the operational safety and maintenance strategies of tunnel in cold region.