MSc project: A minimalistic look at re-suspension of finite-size particles in turbulent channel transport by means of DNS

Particle-laden flows are everywhere around us… From the red blood cells flowing through our veins to the sand transported by the sea waves that gets stuck in our hair whenever we go for a swim. In industry one can think about the sediment transport in the dredging industry, or in fluidized bed reactors. In many cases, the flow is turbulent and the particles are finite-sized (i.e., their characteristic size is of-the-order or bigger than the smallest scales of the flow). There is a big discrepancy between the amount of applications where we can find these flows and the fundamental knowledge that we have about them. The simple question “what causes a particle settled in a sediment bed to get (re-)suspended into the turbulent bulk?” does not have a precise answer.


With the continuous improvement of numerical methods and computational resources it is now possible to simulate the flow conforming ~O(10^5) finite-sized particles. These fully resolved simulations give us an immense detail on the physical mechanisms taking place at the flow, as we know, for every particle, the distribution of pressure and velocity at its surface. The goal of this work is giving insights on the answer to the following question:


In a turbulent channel flow with Reynolds number Re, laden with N finite-sized particles of diameter D and q times heavier than the fluid, how likely is it that a particle resting in a sediment bed will be re-entrained into the flow?


To try to answer this question we will analyze with detail two reduced problems:

  1. Interaction between an averaged turbulent structure obtained from conditionally averaging a simulation of single-phase turbulent channel transport with one single particle – we will study the interaction between the particle and this averaged structure and analyze the induced lift and drag forces acting on the particle. This way we can see how ‘strong’ do the coherent structures have to be to generate a lift force that balances the particle submerged weight.

  2. Simulating the turbulent flow in a sediment bed with O(10^3-10^4) particles, and computing a representative averaged coherent structure of the flow over the sediment bed. Then we will study the interaction of this structure with the sediment bed and try to answer our research question.


The numerical tool that will be used is a standard pressure-correction, finite-volume Navier-Stokes solver, combined with an Immersed Boundary Method (IBM) to impose no-slip/no-penetration boundary conditions at the particle surface and a recently developed collision model for particle-particle/-wall interactions. The attached figure shows a typical result obtained from a simulation of turbulent sediment transport in an open plane channel.


For more information/references contact Pedro Costa,


Involved People:

Facilities used: