Abstract
The molecular dynamics method (MD) and the direct simulation Monte Carlo technique (DSMC) are compared with respect to their capability of simulating vorticity distributions. The statistical assumptions underlying the evaluation of the collision term by the DSMC are analyzed. They lead to the nonconservation of angular momentum for the interaction of particles. Both methods yield equally good results for the Rayleigh–Stokes flow. Using the present parameters, however, only the MD simulation shows the generation of vortices for the flow past an inclined flat plate. This might indicate an effect on the computation of vortical flows. It is suggested that further systematic studies of the effect of the cell size and particle dimensions be carried out.