Fu Shaotong, Hao Zikang, Su Weite, Zhang Huahai, Wang Limin
State Key Laboratory of Mesoscience and Engineering, <a href="https://ror.org/03j4x9j18">Institute of Process Engineering</a>, Chinese Academy of Sciences, Beijing 100190, China School of Chemical Engineering, <a href="https://ror.org/05qbk4x57">University of Chinese Academy of Sciences</a>, Beijing 100049, China.
Phys Rev E. 2024 Oct;110(4-2):045309. doi: 10.1103/PhysRevE.110.045309.
A lattice Boltzmann (LB) scheme for a level-set equation is proposed to capture interface and is coupled with the LB model for incompressible fluid to simulate immiscible two-phase flows. The reinitialization of a level-set field is achieved directly by adding a source term to LB equation, which avoids solving an additional partial differential equation as required in traditional level-set methods. Compared to the classical phase-field lattice Boltzmann method, the proposed approach demonstrates significantly reduced errors in solving interface motion and deformation. Furthermore, GPU parallel computation is implemented for the level-set lattice Boltzmann method (LS-LBM) to enhance computational efficiency. To validate the LS-LBM, it is employed to simulate four benchmark problems: static droplet, layered Poiseuille flow, rising bubble, and Rayleigh-Taylor instability. Numerical results show that LS-LBM exhibits good stability, accuracy and high efficiency, demonstrating its feasibility for accurate simulations of immiscible two-phase flows, even with large density ratios or high Reynolds numbers.
提出了一种用于水平集方程的格子玻尔兹曼(LB)格式来捕捉界面,并将其与不可压缩流体的LB模型耦合,以模拟不混溶的两相流。通过在LB方程中添加源项直接实现水平集场的重新初始化,这避免了像传统水平集方法那样求解额外的偏微分方程。与经典的相场格子玻尔兹曼方法相比,该方法在求解界面运动和变形时的误差显著降低。此外,对水平集格子玻尔兹曼方法(LS-LBM)进行了GPU并行计算,以提高计算效率。为了验证LS-LBM,用它来模拟四个基准问题:静态液滴、分层泊肃叶流、上升气泡和瑞利-泰勒不稳定性。数值结果表明,LS-LBM具有良好的稳定性、准确性和高效率,证明了其对于精确模拟不混溶两相流的可行性,即使在密度比大或雷诺数高的情况下也是如此。