Hostname: page-component-7b7567c8c-2n9c9 Total loading time: 0 Render date: 2026-11-16T08:28:53.681Z Has data issue: false hasContentIssue false

Grid criteria for numerical simulation of hypersonic aerothermodynamics in transition regime

Published online by Cambridge University Press:  24 October 2019

Xiang Ren
Affiliation:
School of Astronautics, Beihang University, Beijing 100191, China
Junya Yuan
Affiliation:
School of Astronautics, Beihang University, Beijing 100191, China
Bijiao He
Affiliation:
School of Astronautics, Beihang University, Beijing 100191, China
Mingxing Zhang
Affiliation:
Beijing Institute of Space Long March Vehicle, Beijing 100076, China
Guobiao Cai*
Affiliation:
School of Astronautics, Beihang University, Beijing 100191, China
*
†Email address for correspondence: cgb@buaa.edu.cn

Abstract

Grid is an important factor in numerical simulation of hypersonic aerothermodynamics. This paper introduces three criteria for determining grid size in the transition flow regime when using the computational fluid dynamics (CFD) method or the direct simulation Monte Carlo (DSMC) method. The numerical relationship between these three criteria sizes is deduced according to the one-dimensional fluid theory. Then, the relationship is verified using the CFD method to simulate the flow around a two-dimensional cylinder. At the same time, the dependence of simulation accuracy on grid size in the CFD and DSMC methods is studied and the mechanism is given. The result shows that the simulation accuracy of heat flux especially depends on the normal grid size next to surfaces, where the $Re_{\mathit{cell},w}$ criterion and the $\unicode[STIX]{x1D706}_{w}$ criterion based on local parameters are applicable and equivalent, while the $Re_{\mathit{cell},\infty }$ criterion based on the free-stream parameter is only applicable under the assumption of constant viscosity coefficient and constant temperature wall conditions. On the other hand, the trend of the heat flux changing with grid size obtained by CFD and DSMC is exactly the opposite. Therefore, the grid size must be strictly satisfied with the grid criteria when comparing CFD with DSMC and even the hybrid DSMC with Navier–Stokes method.

Information

Type
JFM Papers
Copyright
© 2019 Cambridge University Press 
Figure 0

Figure 1.Computation field and grid settings.

Figure 1

Table 1.Free-stream flow settings.

Figure 2

Table 2.The VHS molecular model parameters.

Figure 3

Figure 2.Correctness of the two solvers. (a) Surface distribution of pressure. (b) Surface distribution of heat flux.

Figure 4

Table 3.The free-stream velocity.

Figure 5

Figure 3.For different gases, $l_{\mathit{cell},\infty }$ and $\unicode[STIX]{x1D706}_{w}$ change with $Ma_{\infty }$: (a,c) Ar and (b,d) $\text{N}_{2}$. In (a,b), transport parameters are relayed on the VHS model, and in (c,d), these are constant.

Figure 6

Table 4.Wall-normal grid settings.

Figure 7

Figure 4.Comparison of flow fields between DSMC and CFD: (a) pressure; (b) temperature; (c) $Ma$ number; (d) density.

Figure 8

Table 5.The normal grid size based on three criteria (units: $10^{-4}~\text{m}$).

Figure 9

Figure 5.(a) Surface pressure and (b) surface heat flux computed by the two solvers under different grid sizes.

Figure 10

Figure 6.According to the two solvers under different grid sizes, (a) the pressure and (b) the heat flux at the stagnation point.

Figure 11

Figure 7.Axis temperature distribution obtained by the two solvers under different grid sizes: (a) DSMC; (b) CFD.