A parallel cell-centered adaptive level set framework for efficient simulation of two-phase flows with subcycling and non-subcycling
A parallel cell-centered adaptive level set framework for efficient simulation of two-phase flows with subcycling and non-subcycling
复制标题
DOI：
10.1016/j.jcp.2021.110740
复制
发表时间：
2022-01
期刊：
影响因子：
--
通讯作者：
Yadong Zeng;A. Xuan;Johannes Blaschke;Lian Shen
中科院分区：
文献类型：
--
作者：
Yadong Zeng;A. Xuan;Johannes Blaschke;Lian Shen
本文利用多层同位网格建立了一个统一的自适应水平集（LS）框架来求解不可压两相流问题。该框架允许我们使用子循环或非子循环方法逐层推进所有变量，以便每个级别上的数据推进完全解耦。一系列的同步操作旨在保持所有级别的动量和质量守恒。提出了LS函数的多级初始化方法，大大提高了两相流的质量守恒性。同位网格允许对所有变量使用一组差分格式和插值运算，这大大简化了数值实现。各种典型的问题，包括无粘剪切层，重力波，上升气泡，瑞利-泰勒不稳定性的计算框架的能力和鲁棒性进行了验证。计算结果表明，本文提出的多层次格式能够准确地求解具有重力和表面张力效应的两相流界面，同时具有良好的动量守恒和能量守恒。最后，通过对一个三维溃坝问题的仿真，验证了该框架的有效性和显著的加速效果。
We develop a unified adaptive level set (LS) framework using the multi-level collocated grid for incompressible two-phase flows. This framework allows us to advance all variables level by level using either the subcycling or the non-subcycling method such that the data advancement on each level is fully decoupled. A series of synchronization operations are designed to keep the momentum and mass conserved across all levels. A multi-level re-initialization method for the LS function is also proposed, which greatly improves the mass conservation of the two-phase flows. The collocated grid allows the use of a single set of differential schemes and interpolation operations for all variables, which greatly simplifies the numerical implementation. The capability and robustness of the computational framework are validated by a variety of canonical problems, including the inviscid shear layer, gravity wave, rising bubble, and Rayleigh-Taylor instability. It is shown that the present multi-level scheme can accurately resolve the interfaces of the two-phase flows with gravitational and surface tension effects while having good momentum and energy conservation. At last, a three-dimensional dam breaking problem is simulated to show the efficiency and significant speedup of the proposed framework.