Dear all,
I want to ask a question. As I see in this forum, the mesh size of OpenFOAM should be 3-4 times of the particles diameter. If I have some kinds of particles with different diameters such as 1mm, 0.5mm, 0.1mm. Should my mesh size be 3-4 times of the diameters of the larger particles or the smaller particles? Thank you very much.
Besr regards,
Rachel
AndresMM | Mon, 11/13/2017 - 09:34
The reason the rule of thumb
The reason the rule of thumb is used is because the CFD-equations might suffer if there is too little fluid in the cells. For example if a cell is completely covered by a big particle, the CFD equations will have to work with a fluid-percentage of zero somewhere, which might create weird numerical behavior. It is also the same if the fluid percentage is just small.
paul | Mon, 11/13/2017 - 21:41
If you get into the territory
If you get into the territory of ~2 particle diameters make sure the divided voidfraction model + diffusive smoothing are used.
It is always a devil's bargain between applicability of the drag closures.
The publicly available drag closures were developed to model the influence of voidfraction, but were developed without mesh influence in mind.
Check out the work of S. Radl on the topic:
A drag model for filtered Euler–Lagrange simulations of clustered gas–particle suspensions
Executive summary of this is, that too coarse a mesh overestimates drag due to a loss of mesoscale structures.
If stable at all, too fine a mesh will cause an underprediction of drag due to a cell being covered and voidfraction assuming its lower limit in the drag law (these are usually hardcoded).
Even worse is your degree of polydispersity - I know of no drag law being applicable to a diameter ratio of 1:10.
If tasked with this, I'd choose the cell diameter according to the constituent fraction which has the largest influence on the system - not an easy choice to make.
If you have a validation experiment / case, do a parameter study and let us know how it went.
Greetings,
Paul