Finite-size particles in homogeneous turbulence: a numerical study

我们建议通过直接数值模拟（DNS）研究在引力场作用下稀分散相（由固体颗粒组成）和均匀湍流（或自诱导）流之间相互作用的基本机制。所考虑的两相流型可以用0（100）的颗粒雷诺数和与最小流动尺度相当或更大的颗粒直径来表征。因此，一般的点-粒子近似不能直接适用。因此，我们完全解决了相界面以及所有相关的流动尺度。所提出的流动结构在统计上是均匀的，由最初各向同性的湍流场或加入重颗粒的环境流体组成。这种情况是在大气（云）或化学工程过程中遇到的流动条件的理想化情况。本项目生成的数据（流场和粒子运动）的分析将以以下基本问题为指导：湍流背景流中的沉降速度如何受到有限尺寸、有限雷诺数和集体效应的影响？粒子诱导湍流增强/衰减的机制是什么？分散相的空间分布如何受到问题参数的影响？重粒子如何改变现有的流动结构/产生新的结构？可以预期，所提出的研究结果将进一步促进我们对颗粒流动动力学的理解。此外，所获得的洞察力将有利于未来的努力，以改进现有的此类流的工程目的模型。

We propose to study via direct numerical simulation (DNS) the basic mechanisms involved in the interaction between a dilute dispersed phase (consisting of solid particles) and a homogeneous turbulent (or self-induced) flow subjected to a gravitational field. The two-phase flow regime under consideration can be characterized by particle Reynolds numbers of O(100) and particle diameters comparable to or larger than the smallest flow scales. Therefore, the common point-particle approximation is not directly applicable. As a consequence, we fully resolve the phase interfaces as well as all relevant flow scales. The proposed flow configuration is statistically homogeneous, consisting either of an initially isotropic turbulent field or of ambient fluid to which heavy particles are added. This situation is an idealization of flow conditions encountered e.g. in the atmosphere (clouds) or in chemical engineering processes. The analysis of the data generated in this project (flow field and particle motion) will be guided by the following fundamental questions: How is the settling velocity in turbulent background flow affected by finite-size, finite-Reynolds-number and collective effects? What are the mechanisms of particle-induced turbulence enhancement/attenuation? How is the spatial distribution of the disperse phase influenced by the problem parameters? How do heavy particles modify existing flow structures/generate new structures? It can be expected that the results of the proposed research will further promote our understanding of particulate flow dynamics. Moreover, the insight gained will benefit future efforts to improve existing engineering-purpose models of such flows.