Simulation Research on Thermal Effect of Urban Subway Spatial Layout Based on Machine Learning Algorithm

In recent years, due to urban overexploitation and traffic congestion on the ground, underground railway traffic has become a major mode of public transportation because of its safety and speed. Machine learning algorithm is used to iteratively recover missing data and extract features. Combined with a small amount of temperature measurement data, the temperature distribution is reconstructed in real time. Using computational fluid dynamics, the thermal environment of subway station public area is numerically simulated. Including the simulation of single-train inbound and outbound working conditions and double-train inbound and outbound working conditions, the distribution of velocity field and temperature field in the public area of subway station personnel is determined according to the simulation, and the piston air ventilation rate basically meets the ventilation demand of five air changes per hour after being checked hourly throughout the day. The research in this paper provides theoretical analysis basis and test scheme for the engineering application of controllable tuyere screen door system in subway.