Mathematical Modelling of Particle-Laden Turbulent Flows

点击翻译按钮获取中文摘要

This research program is concerned with the development of mathematical models to predict the motion of particulates in gaseous turbulent flows. Particle-laden turbulent flows occur in a broad range of engineering problems including pollution dispersion in atmospheric flows, spray combustion, ventilation, and medicinal aerosol drug delivery systems. Computational fluid dynamics (CFD) provides the engineering designer or analyst with information on the fluid flow and particle velocities and concentrations within the domain. This allows for development of enhanced engineering designs. There is therefore a need for robust mathematical models of particulate motion that are able to capture the complexities of the fluid-turbulence/particulate interaction, while also maintaining reasonable computational effort. Fluid turbulence influences a vast array of particle-flow characteristics including the dispersion and drift of particles, the rate of deposition on solid surfaces, and the rate of mixing of fuel droplets or solid particles with air for combusting flows. Spatial variations in turbulence intensity can result in particles accumulating in regions of low turbulent kinetic energy. Even within homogeneous turbulent flows, local clustering of the particles can occur due to the influence of underlying turbulent structure; particles with a time constant similar to the smallest turbulent time scale can be ejected from vortical turbulent structures and tend to concentrate locally in regions of high strain and low vorticity.  Moreover, the presence of the particles can influence the underlying fluid velocity field.    Two-way coupling between the gas and particulate phase is complex. For example, depending on the particle time constant and size relative to the important turbulent scales, particles can enhance turbulence production through vortex shedding or can act to modulate the turbulent kinetic energy as a result of a transfer of momentum from the fluid to the particles. The objective of this research program is to develop particle-laden flow models that accurately capture the complex physics of particle-fluid turbulence interactions. This includes the ability to predict the impacts of unresolved turbulence structure, turbulence inhomogeneity and anisotropy, and two-way coupling.  This research will build on the previous work that we have done on model development and assessment for large eddy simulation and Reynolds-averaged Navier-Stokes turbulence modelling frameworks for CFD simulation of particle-laden turbulent flows.  Direct numerical simulation is used as a benchmarking tool.  This research will lead to improved models for incorporation into commercial or open-source CFD codes used in industry.  In addition, the rigorous training that the highly qualified personnel will receive will provide industry with CFD analysts with a strong theoretical foundation which is critical for successful CFD simulation and analysis.