Effect of collision‐partner selection schemes on the accuracy and efficiency of the direct simulation Monte Carlo method |
| |
Authors: | M. A. Gallis J. R. Torczynski |
| |
Affiliation: | Engineering Sciences Center, Sandia National Laboratories, Albuquerque, NM 87185‐0346, U.S.A. |
| |
Abstract: | The effect of the collision‐partner selection scheme on the accuracy and the efficiency of the Direct Simulation Monte Carlo method is investigated. Several schemes that reduce the mean collision separation, including the fixed sub‐cell scheme, the transient adaptive sub‐cell scheme, and the virtual sub‐cell scheme, are evaluated. Additionally, a new scheme is proposed that limits the population from which collision partners are selected based on the distance traveled by a simulator and performs near‐neighbor collisions using this population. These collision‐partner selection schemes are assessed for Fourier flow (heat conduction between parallel plates) and a standard hypersonic benchmark problem (Mach 15.6 nitrogen flow over a 25–55° biconic). The new limited‐selection near‐neighbor scheme has superior performance compared to the other schemes for both flows and reduces both the spatial and temporal discretization errors relative to random‐selection and nearest‐neighbor collision‐partner selection schemes. Copyright © 2010 John Wiley & Sons, Ltd. |
| |
Keywords: | DSMC algorithm convergence rarefied gas dynamics |
|
|