Evaluating the effectiveness of CFD-DEM and SPH-DEM for complex pipe flow simulations with and without particles

Abstract

We investigate the effectiveness of two computational fluid dynamics (CFD) approaches: mesh-based CFD and meshfree particle-based smoothed particle hydrodynamics (SPH) for simulating pipe flows of varying complexity. The study covers laminar and turbulent flows, different fluid rheologies (Newtonian, power-law, Bingham plastic, Herschel-Bulkley), and different particle-laden scenarios, validated using experimental Lagrangian measurements obtained by positron emission particle tracking or available theoretical solutions, as appropriate. We assess these methods based on their ability to predict radial profiles of local phase velocity and concentration, as well as computational cost. In single-phase flows, CFD aligns well with experimental data and theoretical models. SPH exhibits boundary discrepancies due to no-slip condition approximations and limitations in turbulent flow simulation which need further development. Integrating the discrete element method (DEM) significantly enhances both techniques for particle-liquid flows. Mesh-based CFD is computationally efficient, while particle-based SPH can offer more insights into Lagrangian fluid dynamics.