Parameter on which DEM number of iterations depends?

MFiX How to
#1

Hello everyone,

I am simulating a two dimensional fluid-bed with DEM approach. My aim is to actually simulate very fine particles (less than 0.1 micron) to see the effect of cohesive forces on formation of agglomerates.

Geomtery: I have done two separate DEM simulation with 1mm and 0.1mm particle in the bed of 40dp x 90dp x dp size. Meshing is done 3.33dp x 4.5dp x dp for both size of the particles. Not using any cohesion model, as i just want to compare simulation time(All other parameters are kept same)

Problem: As I am scaling down all parameters to the particles , I was expecting same simulation time.
But on reducing particles size from 1mm to 0.1mm number of DEM iteration (DEM NITs) has increased from 576 to 8900 and also DEM solver time has increased from 0.23 s to 10s.

If any of you can help me with this … Why on reducing particle size, DEM solver is taking much time in-spite of scaling down all parameters (if not pls let me know) to the particle size.

#2

I am assuming that the particle count drastically increased?

#3

Hey thanks for the reply

I am scaling down bed dimension to the particle size keeping same volume fraction. So having same number of particles (i.e 2300) in each case.

#4

The collision time scale in MFIX-DEM is a function of particle mass, spring stiffness, and damping coefficient. As particle mass decreases (or spring stiffness increases), the collision time scale decreases. See equation (32) of the DEM theory manual.

2 Likes
#5

Thanks jmusser,

Being new to this field, I got one more topic to read. Relation between collision time and material property, before that Pls clear my silly doubt of how this Collision time scale and DEM number of iterations (or DEM solver time in second) depend on each other.

#6

The DEM time step equals the collision time scale divided by 50.

1 Like