论文标题
流体驱动的颗粒动力学通过一致的多分辨率粒子方法
Fluid-driven granular dynamics through a consistent multi-resolution particle method
论文作者
论文摘要
由流体流动驱动的颗粒动力学在许多工业和自然过程中无处不在，例如河流和沿海沉积物传输。然而，它们复杂的多物理性质挑战了数值模型的准确性和效率。在这里，我们研究了基于增强的弱弱压缩运动粒子半平移（MPS）方法，研究快速流体驱动的颗粒侵蚀的动力学。为此，我们开发并验证了一个新的多分辨率多相MPS公式，以一致且保守的方程式（包括粒子稳定技术）的一致和保守形式。首先，我们通过两个数值基准情况讨论了所提出的近似运算符的数值准确性和收敛性：多粘度POISEUILLE流量和多密度静水压力。然后，将开发的模型与广义流变方程相连，我们研究了可移动床上的水坝断裂波。粒子收敛研究证实，所提出的多分辨率配方以可接受的准确性和收敛顺序预测分析溶液。验证多相颗粒流程表明，这种流体驱动问题的机械行为对水染色密度比非常敏感。较浅的谷物的床经历了极端侵蚀和界面变形。对于具有较重材料但几何设置的床的床，浪涌速度和运输层厚度几乎保持不变（远离门）。此外，尽管多分辨率模型可以准确估计全局沉积物动力学，但单分辨率模型低估了流动的演变。总体而言，对结果的定性和定量分析强调了多尺度多密度相互作用在流体驱动建模中的重要性。
Granular dynamics driven by fluid flow is ubiquitous in many industrial and natural processes, such as fluvial and coastal sediment transport. Yet, their complex multiphysics nature challenges the accuracy and efficiency of numerical models. Here, we study the dynamics of rapid fluid-driven granular erosion through a mesh-free particle method based on the enhanced weakly-compressible Moving Particle Semi-implicit (MPS) method. To that end, we develop and validate a new multi-resolution multiphase MPS formulation for the consistent and conservative form of the governing equations, including particle stabilization techniques. First, we discuss the numerical accuracy and convergence of the proposed approximation operators through two numerical benchmark cases: the multi-viscosity Poiseuille flow and the multi-density hydrostatic pressure. Then, coupling the developed model with a generalized rheology equation, we investigate the water dam-break waves over movable beds. The particle convergence study confirms that the proposed multi-resolution formulation predicts the analytical solutions with acceptable accuracy and order of convergence. Validating the multiphase granular flow reveals that the mechanical behavior of this fluid-driven problem is highly sensitive to the water-sediment density ratio; the bed with lighter grains experiences extreme erosion and interface deformations. For the bed with a heavier material but different geometrical setups, the surge speed and the transport layer thickness remain almost identical (away from the gate). Furthermore, while the multi-resolution model accurately estimates the global sediment dynamics, the single-resolution model underestimates the flow evolution. Overall, the qualitative and quantitative analysis of results emphasizes the importance of multi-scale multi-density interactions in fluid-driven modeling.
