A consistent multiphase flow model with a generalized particle shifting scheme resolved via incompressible SPH
A consistent multiphase flow model with a generalized particle shifting scheme resolved via incompressible SPH
复制标题
DOI：
10.1016/j.jcp.2022.111079
复制
发表时间：
2022-02
期刊：
影响因子：
--
通讯作者：
Lijing Yang;Milad Rakhsha;Weihua Hu;D. Negrut
中科院分区：
文献类型：
--
作者：
Lijing Yang;Milad Rakhsha;Weihua Hu;D. Negrut
这一贡献概述了一个多相流的解决方案，结合不可压缩的光滑粒子流体动力学（SPH）离散化方法。我们提出了（a）一个广义的多相粒子移位技术，保持一个规则的颗粒分布，并防止界面粒子混合;和（B）一个基于颜色函数的排斥力的压力梯度项在投影方法的校正步骤，提高了界面的光滑度，同时最大限度地减少了人工力量的贡献。该方法采用一致的离散梯度和拉普拉斯算子，这提高了空间微分算子的精度。所采用的表面张力模型依赖于连续表面应力（CSS）公式。所提出的求解方法是广泛适用的，因为i）广义粒子移位方案适用于2D和3D模拟，并且i i）界面平滑度和求解鲁棒性可以在密度比高达1：1000和粘度比高达1：100的情况下保持。该解决方案的方法已被用于与一组基准问题，以证明其准确性，通过比较模拟结果对解析解，其他数值方法的结果，和实验数据。该方法已经被实现为利用GPU使能的并行性，并且在本文中用于通过包括超过一百万个SPH粒子的模拟来解决在3D气泡上升和聚结问题中遇到的复杂物理现象。
This contribution outlines a multiphase flow solution method developed in conjunction with the incompressible smoothed particle hydrodynamics (SPH) discretization. We present (a) a generalized multiphase particle shifting technique that maintains a regular particle distribution and prevents interface particle mixture; and (b) a color-function-based repulsive force for the pressure gradient term at the correction step of the projection method, which improves the interface smoothness while minimizing the contribution of artificial forces. The solution approach employs consistent discretizations for the gradient and Laplacian operators, which enhances the accuracy of the spatial differential operators. The surface tension model embraced relies on the continuum surface stress (CSS) formulation. The solution methodology proposed is broadly applicable since i) the generalized particle shifting scheme applies to both 2D and 3D simulations, and i i) the interface smoothness and the solution robustness can be preserved with density ratios up to 1: 1000, and viscosity ratios up to 1: 100. The solution method has been used in conjunction with a set of benchmark problems to demonstrate its accuracy by comparing simulation results against analytical solutions, results from other numerical methods, and experimental data. The method has been implemented to leverage GPU-enabled parallelism and used herein to solve complex physics phenomena encountered in 3D bubble rising and coalescence problems via simulations that include over one million SPH particles.