DNS-LES pore scale simulation of thermal multi-component, multi-phase flow including radiative heating

这个子项目的重点是在模拟热多相、多组分流动的背景下研究几个重要问题，包括多孔介质中的蒸发。其中一个问题是在孔隙尺度多孔介质和大气边界层的第一个分米的交界处计算直接Navier-Stokes和大涡模拟(DNS-LES)组合模式的质量、动量和能量通量，以产生SP2中发展的更大尺度雷诺平均Navier-Stokes(RANS)模式的校准数据。这些模拟将用SP5中获得的风洞数据进行验证。另一个问题涉及对惯性效应(如海恩斯跳跃)的孔尺度模拟，以研究不同过程(渗透、吸胀、蒸发和排水)的孔尺度转变长度、前沿等待时间和前沿速度，以支持在SP5中实验进行的Richards方程对含有烧结玻璃微珠或砂粒的Hele-Shaw单元格的适用性的研究。此外，我们还将研究稳定和不稳定界面运动情况下的溶质浓度梯度。对于这些模拟，我们将开发一个新的动力学自由面模型，避免了以前的格子-玻尔兹曼模型难以解决的气水界面强密度梯度的数值离散。很大一部分模拟将在大规模并行通用图形处理单元(GPGPU)硬件上进行，以大大减少与具有超过10亿个自由度的三维瞬变计算相关的计算工作量。

This subproject focuses on the investigation of several important issues in the context of modeling thermal multiphase, multicomponent flows including evaporation in porous media. One problem is the computation of mass, momentum and energy flux from a combined Direct Navier-Stokes and Large-Eddy-Simulation (DNS-LES) model at the interface of the pore-scale porous medium and the first decimeters of the atmospheric boundary layer to generate calibration data for larger scale Reynolds averaged Navier-Stokes (RANS) models developed in SP2. These simulations will be validated with wind tunnel data obtained in SP5. Another problem is related to the pore scale simulation of inertia effects (e.g. Hains jumps) to investigate pore scale transition lengths, front waiting times and front velocities for different processes (infiltration, imbibition, evaporation and drainage) to support the investigation of the applicability of the Richards equation conducted experimentally in SP5 for Hele-Shaw cells containing sintered glass beads or sand particles. In addition we will investigate solute concentration gradients for both stable and unstable interface motion scenarios. For these simulations we will develop a new kinetic free surface model avoiding the numerical discretization of the strong density gradients at the air-water interface which were difficult to resolve with previous Lattice-Boltzmann models. A large part of the simulations will be conducted on massively parallel General Purpose Graphics Processing Units (GPGPU) hardware to substantially reduce the computational efforts associated with three-dimensional transient computations with more than a billion degrees of freedom.