Particles passing throug wall --> When rotating wall is close to stationary wall

Submitted by tuks123 on Wed, 12/09/2015 - 15:00

Hi,

I am facing problem of particles passing through wall. I am simulating case where, particles are rotated in a container using rotating impeller. This impeller (rotating wall) is very close (5mm) to the stationary wall of container. There are two types are particles (with 1.5mm diameter and with 7mm).

When impeller is rotated, particles pass through wall. Smaller particles escape through wall more than larger particles.

The Young's modulus (YM) is 1e7 and timestep is 9% of Rayleigh time. I tried with increased YM upto 1e9 and also with smaller timestep (2% of Rayleigh time). Also tried with different default neighbor distance from 1.4mm to 9mm.

When I increase gap between impeller and stationary wall, sufficiently then I don't see particles passing through wall. But moving impeller is not desirable.
Are particles escaping because of the high force exerted?

Can somebody help, how to overcome this problem?

Thanks and regards
Tuks.

ckloss's picture

ckloss | Thu, 12/10/2015 - 22:10

Hi Tuks,

this is an intrinsic problem - if the particle diameter is larger than the gap, then the particle will be forcibly pushed out

Best wishes
Christoph

tuks123 | Fri, 12/11/2015 - 05:19

Hi Christoph,

Thanks for reply. But why smaller particles escape? why not bigger particles? The properties of both particles are similar except difference in density (bigger has higher density).

Is there any thumb-rule for having minimum gap to avoid this? Because when I made gap 2mm more than largest particle diameter, then also I see particles passing through wall. But when the gap is 2*Dparticle_large, then particles do not passing through wall.

Really appreciate your help.

Thanks
Tuks

anandmds's picture

anandmds | Tue, 01/19/2016 - 11:48

hi, I am facing a similar problem, I am oscillating the base of my container containing particles, and when the particle diameter is 1mm, all hell breaks loose, and all the particles just pass through the walls of my container immediately after the simulations start. When I increase the diameter to 2mm, all the particles stay in the container. In both cases, all conditions are the same, except for the diameter of the particles. What should I change in my code to avoid this ? I am not able to locate the source of this problem. Please help.

Thanks & regards

Anand.M
M.S. Mechanical Engineering,
I.I.T. Madras

esmaeilyazdani | Mon, 05/02/2016 - 18:05

Dear All
I have same problem, is any solution for this problem except increasing the diameter?

Regards
Esmaeil

cstoltz | Wed, 05/04/2016 - 14:23

Basically in these cases you're running into a situation where you're pinching the particles in the gap, and as they can't fracture, the forces get enormous and cause them to suddenly displace to a large degree in the span of a single timestep. This can push them straight through the walls of the geometry. The contacts with other smaller particles can also lead to large forces being transferred to those particles, blowing them out of the system in turn.

To avoid this, you need particles that are either sufficiently larger or smaller than the gap size so you avoid the pinching. Reducing the timestep should also help some, but it won't fully solve the problem. I don't know of a good rule of thumb, but I would guess keeping a 2-1 ratio in particle-to-gap or gap-to-particle size would probably suffice. I'm also guessing that this is a conservative estimate, so perhaps 1.5-1 or even less might work... Need to keep in mind that things like particle stiffness will also play a role.

Regards,
Chris

Matteo | Wed, 03/22/2017 - 08:53

Hi all,

I add a similar problem in the past, I solved reducing the time step. It slows down the simulation but you don't lose particles...

mardussa | Thu, 10/17/2019 - 00:21

I would also like to confirm that smaller timesteps also worked for me when I was having particles pass through walls and cause my CFDEM simulation to diverge.

Robert.