In multiphase flow the choice of approach is more decisive than the choice of turbulence model is in single-phase flow. A problem set up with the wrong approach may converge and still not give the right answer, because the physics you meant to capture is not in the equations [1] [2].
Two questions decide it
1. What is the volume fraction of the secondary phase? 2. Is the interface large enough to be resolved by the mesh, or smaller than a cell?
Free-surface approach
The interface is resolved and tracked. The phases are separated and the interface is large and continuous.
Where: tank filling and draining, sloshing, weir and channel flow, waves, liquid level inside a separator.
Requires: a mesh fine enough to resolve the interface and a small time step. The Courant number in the interface region is usually kept below one.
Eulerian-Eulerian approach
The phases are solved as interpenetrating continua. The interface is not resolved but modelled.
Where: bubble columns, fluidised beds, dense solids transport — anywhere the secondary phase volume fraction exceeds roughly ten percent.
The price: closure relations are needed for interphase drag, lift and turbulence interaction. The result is sensitive to the choice of those relations, and that sensitivity belongs in the report.
Particle tracking (Lagrangian)
The secondary phase is tracked as discrete particles through the continuous phase.
Where: collection efficiency in cyclones and separators, sprays and evaporation, erosion, droplet transport. The volume fraction should typically be below ten percent.
Watch out: the efficiency curve depends directly on the particle size distribution. Running with a single mean diameter erases the most critical quantity — the escape fraction of the small particles.
Cavitation
This is a subject of its own: the phase change is driven by pressure. When cavitation inception in a pump or valve is the question, vapour pressure, nucleation and dissolved gas have to be considered together. A transient solution is often necessary, because the instantaneous minimum pressure is well below the mean.
A rule that holds throughout
Whichever approach is chosen, a multiphase result should not be quoted as an absolute number without validation data. Comparative use — is design A better than design B — is far more reliable.
References
- Hirt, C. W. & Nichols, B. D. (1981). Volume of fluid (VOF) method for the dynamics of free boundaries. Journal of Computational Physics, 39(1), 201–225. DOI: 10.1016/0021-9991(81)90145-5
- Morsi, S. A. & Alexander, A. J. (1972). An investigation of particle trajectories in two-phase flow systems. Journal of Fluid Mechanics, 55(2), 193–208. DOI: 10.1017/S0022112072001806
- Singhal, A. K., Athavale, M. M., Li, H. & Jiang, Y. (2002). Mathematical basis and validation of the full cavitation model. Journal of Fluids Engineering, 124(3), 617–624. DOI: 10.1115/1.1486223