Adaptive Particle Discretization Methods for Multiscale Non-equilibrium Flows
摘要
Highly accurate simulation results for complex flow fields with strong nonequilibrium and noncontinuum effects, such as present during atmospheric entry processes, can be obtained by the Direct Simulation Monte Carlo (DSMC) method. However its use is limited by the increasing computational cost in denser regions. In these cases, efficiency-increasing methods are required. Alternative particle methods, such as Fokker-Planck (FP) or Bhatnagar-Gross-Krook (BGK) can achieve similar accuracy to DSMC in regions with lower Knudsen numbers, but with a lower computational demand. Additionally, these methods allow for a straightforward coupling to DSMC. Several optimization techniques have been implemented in the simulation code PICLas, including adaptive particle weights and a mesh refinement. In addition, various equilibrium breakdown criteria have been implemented for an efficient coupling between DSMC and other particle-based methods. All implemented adaption techniques are evaluated in terms of their accuracy and efficiency, in comparison to a non-adapted DSMC simulation.