Evaluation of the SBT-TAS collision scheme on treating unsteady flows

Abstract

The aim of the current work is to evaluate the accuracy and efficiency of a newly suggested collision scheme in the Direct Simulation Monte Carlo (DSMC) method; Simplified Bernoulli Trials implemented over transient adaptive subcells, SBT-TAS, for two different unsteady flow test cases. First, the performance of the SBT-TAS scheme is evaluated via simulating the shock-tube test case through a broad range of Knudsen numbers. The shock-tube investigations indicated that the SBT-TAS scheme could predict solutions as accurate as the nearest neighbor (NN) scheme. We show that the implemented SBT-TAS scheme requires at least 3–4 particles per each subcell to predict intermolecular collision frequency accurately. Then, a two-dimensional cylinder at Mach number of 0.6 is considered to investigate the accuracy and efficiency of the SBT-TAS in the prediction of the unsteady vortexes shedding. On the same grid and using the same number of particles per cell (PPC), the SBT and No Time Counter (NTC) collisions schemes fail to predict the correct vortex shedding behind the cylinder accurately and the structure of vortices are smeared out while the SBT-TAS and NN collisions schemes predict the strength of the vortices and their oscillation frequency with the same level of accuracy. Our results indicate that SBT-TAS preserves its accuracy in prediction of the aforementioned phenomenon while it uses relatively the same sample sizes and number of particles compared with its NN counterpart.