Zhou Pei-Yuan Center for Applied Mathematics, Tsinghua University, Beijing 100084, China.
Beijing Computational Science Research Center, Beijing 100094, China.
Phys Rev E. 2016 Mar;93(3):033310. doi: 10.1103/PhysRevE.93.033310. Epub 2016 Mar 22.
We propose using the Maxwell iteration to derive the hydrodynamic equations from the lattice Boltzmann equation (LBE) with an external forcing term. The proposed methodology differs from existing approaches in several aspects. First, it need not explicitly introduce multiple-timescales or the Knudsen number, both of which are required in the Chapman-Enskog analysis. Second, it need not use the Hilbert expansion of the hydrodynamic variables, which is necessary in the asymptotic analysis of the LBE. The proposed methodology assumes the acoustic scaling (or the convective scaling) δ(t)∼δ(x), thus δ(t) is the only expansion parameter in the analysis of the LBE system, and it leads to the Navier-Stokes equations in compressible form. The forcing density derived in this work can reproduce existing forcing schemes by adjusting appropriate parameters. The proposed methodology also analyzes the numerical accuracy of the LBE. In particular, it shows the Mach number Ma should scale as O(δ(t)(1/3)) to maintain the truncation errors due to Ma and δ(t) in balance when δ(t)→0, so that the LBE can converge to the expected hydrodynamic equations effectively and efficiently.
我们提出使用麦克斯韦迭代法从带有外部驱动力的格子 Boltzmann 方程(LBE)推导出流体力学方程。所提出的方法在几个方面与现有方法不同。首先,它不需要显式引入多时间尺度或 Knudsen 数,而这两者在Chapman-Enskog 分析中都是必需的。其次,它不需要使用流体变量的 Hilbert 展开,而这在 LBE 的渐近分析中是必需的。所提出的方法假设声速尺度(或对流尺度)δ(t)∼δ(x),因此在 LBE 系统的分析中,δ(t)是唯一的扩展参数,它导致可压缩形式的纳维-斯托克斯方程。通过调整适当的参数,本工作中推导的驱动力密度可以再现现有的驱动力方案。所提出的方法还分析了 LBE 的数值精度。特别是,它表明马赫数 Ma 应该按 O(δ(t)(1/3))缩放,以在 δ(t)→0 时保持由于 Ma 和 δ(t)引起的截断误差平衡,从而使 LBE 能够有效地收敛到预期的流体力学方程。