Comparison of numerical methods for multiphase flows

Particulate flows are involved in applications such as fluidization, sediment transport, and hydraulic fracturing, etc. For this work, we focus with applications to fluidization and hydraulic fracturing.

We intend to solve such particulate flows using Computational Fluid Dynamics (CFD). CFD solvers can  either be used with structured or unstructured grids. Among structured Cartesian grids, there are two important methods namely, Immersed Boundary Method (IBM) and Lattice Boltzmann Method (LBM). Since these numerical methods are Cartesian based, it greatly reduces computational demand for cases like moving particles by avoiding re-meshing at each time-step and also easier parallelization. The IBM works basically with Navier-Stokes equation, whereas LBM uses discretized Boltzmann equation. Each of them have their own advantages and disadvantages.

The objective of this work is to explore the advantages, disadvantages and limitations of these solvers. Several different test cases will be performed to explore their limitations. Different metrics can be compared – computational efficiency, execution speed, grid resolution required to achieve desired accuracy, Reynolds number limitations, and code scaling, etc. The planned metrics for comparison are flexible and additional metrics can be included during the course, if needed.

In literature, both the methods are used extensively, although, concrete reasons are not cited for choosing one over other. We intend to fill this gap and these comparisons can be a good indicator for the research community on choosing appropriate solver . Hence, there is high possibility to publish the findings through journals.

This assignment can be started immediately. For more information, please write to S.K.PachaSanjeevi@tudelft.nl

*PhD researcher, TU Eindhoven