The scale effect of the wind tunnel test on the drag force of a stationary train under crosswinds.

The experimental errors of the buffeting force caused by the scale effect become more significant when the reproduced turbulence scale is not very large in comparison with the characteristic dimension of the test model. From a fluid mechanics perspective, the scale effect can be divided into the three-dimensional (3D) effect in the spanwise direction and the distortion effect in the streamwise direction. A theoretical model of the 3D drag admittance is proposed to study the influence mechanism of the scale effect. Wind tunnel tests were conducted to clarify the effects of turbulence intensities (I u) and scales (L u / D) on the spatial distribution characteristics of drag force on a stationary train under crosswinds. The results indicate that the I u mainly influences the drag coefficient. In comparison, the L u / D is the key parameter in determining the scale effect on the drag force, which becomes more significant when the turbulence scale decreases. Based on the wind tunnel tests, an improved and double-exponential coherence model of drag force is proposed, resulting in a generalized 3D AAF model to rapidly modify the scale effect on the drag force of a stationary train. 2D AAF provides us with new insight into the effect of distortion. The results indicate that L u / D plays a dominant role in determining the 2D AAF in the high-frequency domain. A decrease of L u / D results in a more rapid drop-off of 2D AAF at a higher frequency caused by the distortion of small-scale vortices in the turbulence approaching a two-dimensional train. The proposed 2D AAF model is of high accuracy in modifying the distortion effect. [Display omitted] • The scale effect on the drag of a stationary train under crosswinds is studied. • A three-dimensional theoretical framework is proposed to study the scale effect. • An improved coherence model for the drag force of a stationary train is proposed. • A closed-form formula of 3D AAF of a stationary train is proposed. • Empirical models are proposed to modify the 3D and distortion effects. [ABSTRACT FROM AUTHOR]