Hostname: page-component-7b7567c8c-tnkh4 Total loading time: 0 Render date: 2026-11-17T10:12:15.988Z Has data issue: false hasContentIssue false

Drag on prolate spheroids in wall-bounded shear flow

Published online by Cambridge University Press:  16 July 2026

Jianzhi Yang
Affiliation:
Department of Built Environment, Hefei University of Technology, Hefei 230009, PR ChinaState Key Laboratory of High Temperature Gas Dynamics, School of Engineering Science, University of Science and Technology of China , Hefei 230027, PR China
Chenghuan He
Affiliation:
Department of Built Environment, Hefei University of Technology, Hefei 230009, PR China
Luoqin Liu*
Affiliation:
State Key Laboratory of High Temperature Gas Dynamics, School of Engineering Science, University of Science and Technology of China , Hefei 230027, PR China
Chenyue Xie*
Affiliation:
State Key Laboratory of High Temperature Gas Dynamics, School of Engineering Science, University of Science and Technology of China , Hefei 230027, PR China
Xi-Yun Lu
Affiliation:
State Key Laboratory of High Temperature Gas Dynamics, School of Engineering Science, University of Science and Technology of China , Hefei 230027, PR China
*
Corresponding authors: Luoqin Liu, luoqinliu@ustc.edu.cn; Chenyue Xie, cyxie@ustc.edu.cn
Corresponding authors: Luoqin Liu, luoqinliu@ustc.edu.cn; Chenyue Xie, cyxie@ustc.edu.cn

Abstract

Content of image described in text.

The dynamics of non-spherical particles in wall-bounded flows plays a fundamental role in numerous natural and industrial processes, such as pollen dispersion, fibre suspensions and biomedical flows. Predictive simulations of such systems often employ an Euler–Lagrange approach, for which the accuracy hinges on the drag model used. Although reliable correlations exist for particles in unbounded flow or in direct contact with a wall, the intermediate regime of finite wall distance has received little attention, despite its prevalence in practical applications. For prolate spheroids – a common non-spherical particle shape – the drag force in this regime results from a complex interplay among the particle Reynolds number, the wall-normal distance and the particle’s three-dimensional orientation. In this work, we develop a drag model for prolate spheroids (aspect ratio $\lambda = 2$) that incorporates these coupled effects through a hierarchical framework based on three reference orientations and wall correction factors. The formulation recovers the exact unbounded Stokes behaviour and establishes near-wall asymptotic limits. Validation against direct numerical simulations across the parameter space considered yields a mean relative error below 2 %. Within its validated range, the proposed model accurately captures orientation-dependent wall effects for drag predictions, and is applicable in Euler–Lagrange simulations of non-spherical particles in wall-bounded flows.

Information

Type
JFM Papers
Copyright
© The Author(s), 2026. Published by Cambridge University Press
Figure 0

Figure 1.Figure 1 long description.Schematic illustration of the geometrical configuration. The horizontal azimuthal angle αH$\alpha _H$ is measured in the x$x$–z$z$ plane from the streamwise direction, and the vertical inclination angle αV$\alpha _V$ is measured between the major axis and the wall-parallel plane.

Figure 1

Figure 2.(a) Schematic of the computational domain and boundary conditions for a prolate spheroid in wall-bounded shear flow. (b) Computational mesh in the vertical plane through the particle centre (x$x$–y$y$ plane), showing the body-fitted conformal grid and local refinement near the particle surface and within the gap region. A close-up view of the mesh around the particle is provided in the inset.

Figure 2

Table 1.Parameter space considered in this study, consisting of 400 DNS that span the shear Reynolds number, horizontal and vertical orientation angles (in radians), and dimensionless wall distance.Table 1 long description.

Figure 3

Table 2.Grid independence study for the drag coefficient Cd$C_d$ at αH=αV=π/3$\alpha _H = \alpha _V = \pi /3$ and G/De=0.1$G/D_e = 0.1$. The relative error ΔCd$\Delta C_d$ is calculated as |(Cdstandard−Cdfine)/Cdfine|×100%$| (C_d^{\textit{standard}} - C_d^{\textit{fine}})/C_d^{\textit{fine}} | \times 100\,\%$.Table 2 long description.

Figure 4

Figure 3.Comparison of the normalised streamwise wall shear stress Cf,x=τw,x/((1/2)ρUc2)$C_{f,x} = \tau _{w,x} / (({1}/{2})\rho U_c^2)$ on the particle surface for the case G/De=0.1$G/D_e = 0.1$, Res=1$\textit{Re}_s = 1$, (αH,αV)=(π/3,π/3)$(\alpha _H,\alpha _V)=(\pi /3,\pi /3)$, obtained with the standard mesh (Δmin/De=0.03$\varDelta _{\min }/D_e = 0.03$) and the fine mesh (Δmin/De=0.02$\varDelta _{\min }/D_e = 0.02$): (a) windward side, (b) leeward side.

Figure 5

Table 3.Domain size independence study for the drag coefficient Cd$C_d$ at αH=αV=π/3$\alpha _H = \alpha _V = \pi /3$ and G/De=0.1$G/D_e = 0.1$. Both meshes have minimum grid spacing Δmin/De=0.02$\varDelta _{min}/D_e=0.02$. The relative error ΔCd$\Delta C_d$ is calculated as |(Cdfine−Cdlarge\ domain)/Cdlarge\ domain|×100%$| (C_d^{\textit{fine}} - C_d^{\textit{large domain}})/C_d^{\textit{large domain}} | \times 100\,\%$.Table 3 long description.

Figure 6

Figure 4.Comparison of drag coefficient Cd$C_d$ as a function of Reynolds number Re$\textit{Re}$ for a wall-attached prolate spheroid (λ=2$\lambda = 2$, αV=0$\alpha _V = 0$) in wall-bounded shear flow. Present DNS results (circles) are validated against the reference data from Fillingham et al. (2021) (triangles). Different colours correspond to varying horizontal azimuthal angles αH$\alpha _H$: blue for 0$0$, red for π/6$\pi /6$, green for π/4$\pi /4$, purple for π/3$\pi /3$, and black for π/2$\pi /2$.

Figure 7

Figure 5.Validation of the unbounded-flow drag formulation for prolate ellipsoids of aspect ratio λ=2$\lambda = 2$. Comparison of the empirical fits from (2.9) and (2.10) (red solid line) with DNS data (symbols) from Fröhlich et al. (2020) and Sanjeevi et al. (2022) at orientations: (a) α=0$\alpha =0$ and (b) α=π/2$\alpha = \pi /2$. The DNS data for axisymmetric rod-like particles in unbounded flow from Chéron et al. (2024) are also included for comparison.

Figure 8

Figure 6.Wall correction factors Kwi$K_{wi}$ as functions of dimensionless wall distance G/De$G/D_e$ for the three reference orientations: (a) streamwise-aligned (0,0)$(0,0)$, (b) transverse-horizontal (π/2,0)$(\pi /2,0)$, and (c) transverse-vertical (0,π/2)$(0,\pi /2)$. Solid lines represent the model predictions from (2.16); symbols denote DNS data. The DNS data denoted by the star symbols (five-pointed stars and six-pointed stars) for the rod-like particle for the streamwise-aligned configuration from Chéron & van Wachem (2024) are also included for comparison.

Figure 9

Table 4.Empirical coefficients for the wall correction factors Kwi$K_{wi}$ defined in (2.16).Table 4 long description.

Figure 10

Figure 7.Figure 7 long description.Streamwise velocity contours and streamlines at G/De=0.1$G/D_e = 0.1$ and Res=10$\textit{Re}_s = 10$ for the three orientations (a,d) streamwise-aligned (0,0)$(0, 0)$; (b,e) transverse-horizontal (π/2,0)$(\pi /2, 0)$; and (c, f) transverse-vertical (0,π/2)$(0, \pi /2)$, for (a–c) x$x$–y$y$ plane, (d–f) x$x$–z$z$ plane. All planes are taken through the geometric centre of the ellipsoid.

Figure 11

Figure 8.Normalised drag coefficient as a function of azimuthal angle αH$\alpha _H$ for different wall distances G/De$G/D_e$ and shear Reynolds numbers Res$\textit{Re}_s$: (a) αV=0$\alpha _V = 0$, (b) αV=π/4$\alpha _V = \pi /4$, (c) αV=π/3$\alpha _V = \pi /3$. The normalised drag is defined as (Cd−Cd,αH=0)/(Cd,αH=π/2−Cd,αH=0)$(C_d - C_{d,\alpha _H=0})/(C_{d,\alpha _H=\pi /2} - C_{d,\alpha _H=0})$. Solid lines denote the sin2⁡αH$\sin ^2\alpha _H$ scaling; symbols denote DNS data.

Figure 12

Figure 9.Figure 9 long description.Drag coefficient as a function of inclination angle αV$\alpha _V$ for various wall distances G/De$G/D_e$ and shear Reynolds numbers Res$\textit{Re}_s$: (a) Res=1$\textit{Re}_s = 1$, αH=0$\alpha _H = 0$; (b) Res=10$\textit{Re}_s = 10$, αH=0$\alpha _H = 0$; (c) Res=10$\textit{Re}_s = 10$, αH=π/2$\alpha _H = \pi /2$. Solid curves show the model predictions from (2.19); symbols denote DNS data.

Figure 13

Figure 10.Viscous and pressure drag coefficients as functions of the inclination angle αV$\alpha _V$ for different wall distances G/De$G/D_e$: (a–c) the viscous drag coefficient Cd,v$C_{d,v}$; (d–f) the pressure drag coefficient Cd,p$C_{d,p}$. The cases correspond to: (a,d) Res=1$\textit{Re}_s=1$, αH=0$\alpha _H=0$; (b,e) Res=10$\textit{Re}_s=10$, αH=0$\alpha _H=0$; (c, f) Res=10$\textit{Re}_s=10$, αH=π/2$\alpha _H=\pi /2$.

Figure 14

Figure 11.Surface pressure coefficient distributions at Res=10$\textit{Re}_s=10$ and G/De=0.1$G/D_e=0.1$. Each panel shows the frontal (left) and rear (right) views for a given orientation. The configurations correspond to (a) (αH,αV)=(0,0)$(\alpha _H,\alpha _V)=(0,0)$, (b) (αH,αV)=(0,π/4)$(\alpha _H,\alpha _V)=(0,\pi /4)$, (c) (αH,αV)=(π/2,0)$(\alpha _H,\alpha _V)=(\pi /2,0)$ and (d) (αH,αV)=(π/2,π/4)$(\alpha _H,\alpha _V)=(\pi /2,\pi /4)$.

Figure 15

Figure 12.Comparison between drag coefficients obtained from DNS (Cd,DNS$C_{d,\textit{DNS}}$) and predictions of various drag models (Cd,Pred$C_{d,\textit{Pred}}$). The solid line indicates perfect agreement. Symbols denote predictions from the models of Ouchene et al. (2016), Fröhlich et al. (2020), Fillingham et al. (2021), Sanjeevi et al. (2022) and Chéron et al. (2024); red stars represent the present model. The correlations of Fillingham et al. (2021) and Chéron & van Wachem (2024) are applied for G/De=0.1$G/D_e = 0.1$, while the unbounded-flow correlations are used for G/De=1.5$G/D_e = 1.5$.

Figure 16

Figure 13.(a,b) Lift coefficient Cl$C_l$ and pitching torque coefficient Cm$C_m$ as functions of the wall distance G/De$G/D_e$ for the orientation (αH,αV)=(0,0)$(\alpha _H,\alpha _V)=(0,0)$. (c,d) Variation of Cl$C_l$ and Cm$C_m$ with the horizontal azimuthal angle αH$\alpha _H$ at fixed αV=π/4$\alpha _V=\pi /4$ for Res=1$\textit{Re}_s=1$. (e, f) Variation of Cl$C_l$ and Cm$C_m$ with the vertical inclination angle αV$\alpha _V$ at fixed αH=π/4$\alpha _H=\pi /4$ for Res=1$\textit{Re}_s=1$.

Figure 17

Figure 14.Comparison between drag coefficients obtained from DNS (Cd,DNS$C_{d,\textit{DNS}}$) and predictions of various drag models (Cd,Pred$C_{d,\textit{Pred}}$) for prolate spheroids with aspect ratio λ=4$\lambda = 4$. The solid line indicates perfect agreement. Symbols denote predictions from the models of Ouchene et al. (2016), Fröhlich et al. (2020), Fillingham et al. (2021), Sanjeevi et al. (2022) and Chéron et al. (2024); red stars represent the present model (calibrated for λ=2$\lambda =2$ and used without recalibration). The correlations of Fillingham et al. (2021) and Chéron & van Wachem (2024) are applied for G/De=0.1$G/D_e = 0.1$, while the unbounded-flow correlations are used for G/De=1.5$G/D_e = 1.5$.