國立清華大學 動力機械工程學系 林昭安所指導 周維紀的 Numerical simulations of microflow by lattice Boltzmann method with different lattice models and wall functions (2010),提出GWF18DB關鍵因素是什麼,來自於晶格波茲曼、微流道。
Numerical simulations of microflow by lattice Boltzmann method with different lattice models and wall functions
為了解決GWF18DB 的問題,作者周維紀 這樣論述:
Microchannel flows study has been focused due to MEMS applications recently. In this thesis, we employ kinetic Lattice Boltzmann method(LBM) to simulate microchannel flows. In our simulation, considering a long microchannel with pressure boundary conditions at both inlet and outlet, using three dif
ferent models, D2Q9, D2Q13, and D2Q21 to simulate Poiseuille flow, respectively. In order to predictthe accurate slip velocity at the wall and pressure distribution along streamwise direction, it is essential to apply modification to these models. There are two keypoints, one is correction of wall f
unction, the other is boundary condition. Firstly, we use three different wall functions to test, which are Lockerby’s wall function(LWF),Stop’s wall function(SWF), and Guo’s wall function(GWF). LWF,SWF,and GWF not only lower the slip velocity but also predict a nonlinear behavior in nearwall region
. Here, wall function is applied to the modification of relaxation time.Secondly, boundary condition is discussed. The traditional boundary conditions were implemented for walls, such as bounceback scheme, but it can not generate enough slip velocity on walls. However, kinetic boundary condition lik
e diffuse scattering boundary conditions(DSBC) [18], may over predict the slip velocity on wall. For capturing the slip velocity correctly, we introduce β-weighted diffusive-bounceback boundary condition, which combines the bounceback and diffuse-scattering boundary condition. β is a function of Knu
dsen number and it ’s obtained by fitting the linearized Boltzmann solutions at wall. In addition, we utilize two different schemes to calculate the unknown distribution function at inlet and outlet after streamingstep. All present results are compared with Direct Simulation Monte Carlo (DSMC).