A numerical study of particle-laden flow around an obstacle: flow evolution and Stokes number effects

Abstract

Three-way coupled numerical simulations of particle-laden fluid flow around a square obstacle are performed by a coupled finite-volume and discrete-element method, taking account of the collisions between particles. The fluid flow is described by the Eulerian formalism, while the Lagrangian formalism is used for the solid particles. The dynamics of each phase and the modulation mechanism of the flow field by the inertial particles are explored. The Reynolds number based on the side length of the obstacle is 100, which leads to a typical periodic flow pattern of vortex shedding. Under the action of particles, the flow tends to be chaotic even in the upstream region, and the vortex shedding occurs earlier. The intensity of the vortex is attenuated due to particle dissipation. At sufficiently small Stokes numbers, the particles exhibit flow tracer behavior with relatively high fidelity. However, as the Stokes number and mass loading increase, the interaction between the particles and fluid becomes important, especially in the core region of the channel. With the increase of the particle response time, the spanwise velocity fluctuation is attenuated, while the influence on the streamwise velocity fluctuation seems complicated. No particles are observed to hit the rear surface of the obstacle or deposit in the domain during the flow. When the Stokes number and mass loading are elevated from 0.0036 and 0.065 to 0.5 and 9.024, respectively, the distribution of the particles near the channel centerline varies from being filled in the shedding vortices to surrounding them. Meanwhile, the particle-free zones centered on these vortices become larger. For the Stokes number and mass loading up to 1.5 and 27.073, respectively, the particles are scattered in the downstream area, and the flow pattern of the Kármán vortex street is modified. Within the range of the Stokes number considered, the drag and lift coefficients, and the Strouhal number also show a certain trend of variation.