Hydrodynamic forces implemented into LAMMPS through a lattice-Boltzmann fluid
Hydrodynamic forces implemented into LAMMPS through a lattice-Boltzmann fluid
复制标题
DOI：
10.1016/j.cpc.2013.03.024
复制
发表时间：
2013-08-01
影响因子：
6.3
通讯作者：
Denniston, C.
中科院分区：
文献类型：
--
作者：
Mackay, F. E.;Ollila, S. T. T.;Denniston, C.
通过创建修复程序 lb_fluid，远程流体动力学相互作用已实现到开源分子动力学包 LAMMPS 中。这些相互作用通过将 MD 粒子密度插值到离散晶格上来处理，然后将其耦合到流体。使用热晶格-玻尔兹曼算法对流体进行建模，其中包括质量和动量守恒噪声，从而为粒子和流体提供恒温器。 程序摘要程序标题：fix_lb_fluid目录标识符：AEPH_v1_0程序摘要 URL：http://cpc.cs.qub.ac.uk/summaries/AEPH_v1_0.html程序可从以下位置获取：CPC 程序图书馆，贝尔法斯特女王大学，北卡罗来纳州。爱尔兰许可条款：GNU 通用公共许可证No.分布式程序中的行数，包括测试数据等：439446No.分布式程序字节数，包括测试数据等：9579863 分发格式：tar.gz 编程语言：C++。计算机：全部。操作系统：全部。代码是否矢量化或并行化？：是。使用 MPI 指令并行化。RAM：取决于问题补充材料：“conflned_colloid”示例的数据文件可以在此处下载。分类：7.7。外部例程：LAMMPS [1] (http://lammps.sandia.gov)问题性质：将远程流体动力学效应纳入分子动力学模拟需要存在显性溶剂。目前，将这种溶剂纳入 LAMMPS [1] 模拟的唯一选择是明确包含每个单独的溶剂分子。这显然是相当计算密集型的，并且对于大型系统尺寸很快就会变得不切实际。解决方案方法：作为替代方案，我们实现了流体的粗粒度模型，简化了问题，同时保留了溶剂自由度。我们对流体使用热晶格-玻尔兹曼模型，该模型在每个流体时间步骤与分子动力学粒子耦合[2,3]。限制：虽然 LAMMPS 支持非正交模拟框，但此特定修复只能使用三维正交模拟域来执行。此外，此修复仅允许 z 方向（x-y 平面）上的外墙；始终假设模拟域沿 x 和 y 方向是周期性的。然而，用户可以在任何地方添加浸没边界。 运行时间：fix_lb_fluid 的运行时间从几分钟到几天不等，具体取决于系统大小、晶格网格点的数量以及使用的处理器数量。参考文献：[1] S. Plimpton，短程分子动力学的快速并行算法，J. Comput。物理。 117(1995)1-19.[2] S.T.T. Ollila、C. Denniston、M. Karttunen、T. Ala-Nissila、J. Chem。物理。 134(2011)064902.[3] F.E. 麦凯、C. 丹尼斯顿、J. 计算机。物理。 237 (2013) 289。(C) 2013 Elsevier B.V. 保留所有权利。
Long-range hydrodynamic interactions have been implemented into the open-source molecular dynamics package, LAMMPS, though the creation of a fix, lb_fluid. These interactions are treated by interpolating the MD particle density onto a discrete lattice, which is then coupled to the fluid. A thermal lattice-Boltzmann algorithm is used to model the fluid, which includes mass and momentum conserving noise, providing a thermostat for both the particles and the fluid.Program summaryProgram title: fix_lb_fluidCatalogue identifier: AEPH_v1_0Program summary URL: http://cpc.cs.qub.ac.uk/summaries/AEPH_v1_0.htmlProgram obtainable from: CPC Program Library, Queen's University, Belfast, N. IrelandLicensing provisions: GNU General Public licenseNo. of lines in distributed program, including test data, etc.: 439446No. of bytes in distributed program, including test data, etc.: 9579863Distribution format: tar.gzProgramming language: C++.Computer: All. Operating system: All.Has the code been vectorized or parallelized?: Yes.Parallelized using MPI directives.RAM: Depends on the problemSupplementary material: The data file for the "conflned_colloid" example can be downloaded here.Classification: 7.7.External routines: LAMMPS [1] (http://lammps.sandia.gov)Nature of problem:The inclusion of long-range hydrodynamic effects into molecular dynamics simulations requires the presence of an explicit solvent. Currently, the only option for incorporating such a solvent into a LAMMPS [1] simulation is the explicit inclusion of each of the individual solvent molecules. This is obviously quite computationally intensive, and for large system sizes can quickly become impractical.Solution method:As an alternative, we have implemented a coarse-grained model for the fluid, simplifying the problem, while retaining the solvent degrees of freedom. We use a thermal lattice-Boltzmann model for the fluid, which is coupled to the molecular dynamics particles at each fluid time step [2,3].Restrictions:While LAMMPS supports non-orthogonal simulation boxes, this particular fix can only be performed using a three-dimensional, orthogonal simulation domain. In addition, this fix allows for external walls in the z-direction (x-y plane) only; the simulation domain is always assumed to be periodic along the x and y directions. However, immersed boundaries can be added anywhere by the user.Running time:The run time for fix_lb_fluid varies from minutes to days depending on the system size, the number of lattice mesh points, and the number of processors used.References:[1] S. Plimpton, Fast parallel algorithms for short-range molecular dynamics, J. Comput. Phys. 117 (1995) 1-19.[2] S.T.T. Ollila, C. Denniston, M. Karttunen, T. Ala-Nissila, J. Chem. Phys. 134 (2011) 064902.[3] F.E. Mackay, C. Denniston, J. Comput. Phys. 237 (2013) 289. (C) 2013 Elsevier B.V. All rights reserved.