Dear all,
When I simulate particle interactions in a flow field, the Courant number blows up during collision and, therefore, in combination with the error "Atom sorting has bin size = 0.0", my simulation stops!
Any tips on where to pay attention on when simulating particle interactions (within a fluid flow field) in CFDEM?
Thank you in advance.
Kind regards,
Boris
Nucleophobe | Wed, 03/23/2016 - 05:23
Similar problem
Bnaar,
In general, you can keep simulations more stable by reducing the timestep. If two colliding particles penetrate one another (or a wall) by too great a distance then the forces can blow up non-physically.
You can also damp collision events by reducing the Young's modulus, although this will reduce the accuracy of the simulation. In my experience, if you only care about particle trajectories (and not the time-accurate forces) you can get away with lowering the Young's modulus by a couple orders of magnitude without altering the particle pathline results too much.
There are some other posts on this topic on the forum. csholtz recommended adding a timestep-checking fix to determine if your timestep is sufficiently small, but I never did get around to figuring out how to interpret the output. Anyway, the fix is something like:
fix ts_check all check/timestep/gran 1000 0.1 0.1
Hope this helps.
-Nuc