A THOROUGH EVALUATION OF THE FOKKER-PLANCK KINETIC MODEL IN THE COUETTE FLOW

A THOROUGH EVALUATION OF THE FOKKER-PLANCK KINETIC MODEL IN THE COUETTE FLOW

In this paper, the evolution of Fokker-Planck (FP) approach in various Knudsen numbers and number of particles per cells (PPC) was investigated. A canonical argon Couette flow has been chosen. However, there are some other direct simulation approaches to model rarefied gases such as DSMC and Lattice Boltzmann; Ideally, one would like to have an accuracy of DSMC method with computational efficiency of mesoscale methods such as LB. In this paper, it was pointed out that another possible computationally attractive option is to work with the Fokker-Planck kinetic model. The results show that this method is acceptable up to moderate Knudsen ranges. Additionally, with about 1000 PPC the computational efficiency is really approximate in comparison with DSMC. Also, the dependence of FP method to the computational grid is lower than other approaches. Therefore, this method can be work with a few computational grids. However, the Fokker Planck model was used and was developed; there is no evaluation and validation due to cell, PPC and time step.

In this paper, the evolution of Fokker-Planck (FP) approach in various Knudsen numbers and number of particles per cells (PPC) was investigated. A canonical argon Couette flow has been chosen. However, there are some other direct simulation approaches to model rarefied gases such as DSMC and Lattice Boltzmann; Ideally, one would like to have an accuracy of DSMC method with computational efficiency of mesoscale methods such as LB. In this paper, it was pointed out that another possible computationally attractive option is to work with the Fokker-Planck kinetic model. The results show that this method is acceptable up to moderate Knudsen ranges. Additionally, with about 1000 PPC the computational efficiency is really approximate in comparison with DSMC. Also, the dependence of FP method to the computational grid is lower than other approaches. Therefore, this method can be work with a few computational grids. However, the Fokker Planck model was used and was developed; there is no evaluation and validation due to cell, PPC and time step.