Aerodynamic aspects of high-speed railway underground station with adjoining tunnels

This dissertation started with a simple question: "Can a high-speed railway underground station be operated safely in term of air pressure fluctuations if an express train passing it at full speed?" This was a real life problem that needed to be resolved and train speed reduction was not an option.;A comprehensive literature research had been conducted to review the aerodynamic aspects related to high-speed train and tunnel interactions. The pressure comfort and safety criteria were established based on the findings. It was also noted that there was no precedent at all. The underground station with adjoining tunnels analyzed in this dissertation was the first in the world.;The required numerical method for Computational Fluid Dynamics (CFD) simulations was investigated first. Axi-symmetrical approach was adopted and the results were compared with published full scale data of train portal entry. The findings indicated that CFD simulation could be used for subsequent detailed analysis.;A simplified method was developed for the initial assessment on the pressure inside the station and also the required mitigation measures for the pressure control. Once the mitigation measures were fixed by using the proposed simplified method, two other methods were adopted to assess the resulting pressure inside the public areas of the underground station, namely Three Dimensional (3D) CFD simulations and full scale site measurements. The findings of both methods confirmed that the underground station was safe and comfortable in term of air pressure. It also confirmed that the proposed simplified method was suitable for the initial assessment of pressure relief mitigation measure for the current station and tunnel layout.;Detailed assessment on the differences between the 3D CFD simulations and full scale site measurements indicated that the 3D CFD simulation approach could predict the pressure wave propagation and also with reasonable good accuracy on the occurrence of peaks and troughs. But their magnitudes were over estimated or offset. Reasons for the over estimations were identified based on the result comparisons of different approaches. And further works to refine the 3D CFD method were suggested.