Investigation on aerodynamic force differences of high-speed trains with various simplified models under crosswind conditions

This is the authors' abstract. We don't add key points for this paper.

Numerical investigation of High-Speed Train (HST) aerodynamics under crosswinds plays a critical role in ensuring operational safety. However, current Computational Fluid Dynamics (CFD) simulations for HST crosswind aerodynamics widely adopt diverse simplification schemes for the underbody structure of the bogie region, whereas there still lacks sufficient research on systematic quantitative comparisons and physical interpretations of aerodynamic discrepancies caused by different bogie-region carbody simplification levels. Consequently, this study innovatively establishes CRH380A underbody carbody models with different bogie-region geometric simplifications, namely a smooth flat underbody (G1), fully enclosed bogie-region cavity (G2), open bogie-region cavity (G3), and high-fidelity benchmark model (G4), respectively. Unsteady Improved Delayed Detached-Eddy Simulation (IDDES) simulations are conducted to verify the performance of four proposed models at three typical yaw angles (β = 5°, 15°, 30°), and the numerical method is validated against wind tunnel experimental data. The results reveal that carbody simplification in the bogie region significantly affects aerodynamic prediction accuracy with distinct parameter sensitivities and unique error evolution laws. More specifically, the lift force is extremely sensitive to underbody geometric details: Model G3 produces a prominent Head Car lift relative error of 115% at the small yaw angle of 5°, while the corresponding absolute error Δij is merely 0.0777. Meanwhile, the increasing yaw angle results in a continuously rising absolute error Δij and a sharply decreasing relative error at the same time. Under moderate and severe crosswinds (15° and 30°), the maximum relative error of all aerodynamic coefficients for simplified models is limited within 15%. Besides, time-domain results show that G1 completely loses unsteady flow fluctuation features under all conditions, while frequency-domain analysis demonstrates that G2 presents lower pulsation energy and severe Tail Car drag spectral distortion compared with G3 and G4. This study clarifies the essential error mechanism of bogie-region car-body simplification and thus provides reliable model selection criteria for high-precision HSTs crosswind numerical simulation.