On the use of multibody dynamics techniques to simulate fluid dynamics and fluid–solid interaction problems
On the use of multibody dynamics techniques to simulate fluid dynamics and fluid–solid interaction problems
复制标题
DOI：
10.1007/s11044-021-09784-y
复制
发表时间：
2021-03
影响因子：
3.4
通讯作者：
Milad Rakhsha;Lijing Yang;Wei Hu;D. Negrut
中科院分区：
文献类型：
--
作者：
Milad Rakhsha;Lijing Yang;Wei Hu;D. Negrut
本文将流体动力学问题的基于多体动力学的解决方案与计算流体动力学（CFD）界使用的两种已建立的基于拉格朗日的技术进行比较。基于多体动力学的解决方案具有两个显着属性：它通过双边运动学约束强制执行不可压缩条件，并通过单边运动学约束处理与固相的耦合。基于多体动力学的解决方案，本文称为运动学约束平滑粒子流体动力学 (KCSPH) 方法，是解决 CFD 问题的拉格朗日方法。它依靠平滑粒子流体动力学（SPH）方法来离散纳维-斯托克斯方程中的空间微分算子，并依靠多体动力学的牛顿-欧拉方程来使SPH粒子及时对流。通过将基于多体动力学的方法与 CFD 社区中两种最常用的 SPH 算法（弱可压缩 SPH (WCSPH) 和隐式 SPH (ISPH) 方法）的性能进行比较，我们证明了基于多体动力学的方法是高效且准确的。比较是结合四个测试进行的：不可压缩性基准测试、溃坝测试、浮筒测试和晃动槽测试。我们得出的结论是，对于具有复杂/移动边界的流固相互作用 (FSI) 问题，KCSPH 是传统 CFD 方法的强大替代方案。本文使用的求解器和模型可在名为 Chrono 的开源软件中公开获得；该实现使用 GPU（对于 WCSPH 和 ISPH）和多核 CPU（对于 KCSPH）并行计算。
A multibody dynamics-based solution to the fluid dynamics problem is compared herein to two established Lagrangian-based techniques used by the computational fluid dynamics (CFD) community. The multibody dynamics-based solution has two salient attributes: it enforces the incompressibility condition through bilateral kinematic constraints, and it treats the coupling with the solid phase via unilateral kinematic constraints. The multibody dynamics-based solution, called herein the Kinematically Constrained Smoothed Particle Hydrodynamics (KCSPH) method, is a Lagrangian approach to solving the CFD problem. It relies on the Smoothed Particle Hydrodynamics (SPH) method to discretize the spatial differential operators in the Navier–Stokes equations, and on the Newton–Euler equations of multibody dynamics to convect the SPH particles forward in time. We show that the multibody dynamics-based approach is efficient and accurate by comparing its performance with the two most commonly used SPH algorithms in the CFD community: the weakly compressible SPH (WCSPH), and the implicit SPH (ISPH) methods. The comparison is carried out in conjunction with four tests: an incompressibility benchmark test, dam break, floating cylinder, and sloshing tank. We conclude that KCSPH is a robust alternative to conventional CFD approaches for fluid–solid interaction (FSI) problems with complex/moving boundaries. The solvers and models used herein are publicly available in an open-source software called Chrono; the implementations use GPU (for WCSPH and ISPH), and multicore CPU (for KCSPH) parallel computing.