MFiX 24.1: DEM Hopper Tutorial – Remove the walls, allow particles to fall off?

MFiX How to
,
#1

Hello, I am a new user of MFiX 21.4. I am experimenting with the first tutorial hopper_dem_3d.mfx, Tutorial 3.1.

I am trying to modify the simulation so that when the particles reach the walls of the bin, they may just fall off; I am trying to simulate a sand-pile, as in this literature (Fig 2, right: pictures of experiments), so when the sand-pile starts avalanching, the particles can fall down and off.

I tried to just increase the radius of the bin past the domain, but then the particles just stop at the walls of the domain (which has a square shape), so it’s essentially still a bin. Is there a setting within the geometry or regions tabs that allows for this?

I have attached the modified files. I also changed the friction coefficients and the outlet radius. The simulation runs in less than 10 minutes on my little laptop.
hopper_dem_3d_nowalls.zip (2.9 MB)

#2

Not sure I fully understand. If you want to first settle the particles in the hopper so they are packed and then let them drop to forma pile, you can run 2 simulations:

  1. First simulation with a wall at the bottom of the hopper. Save the particles into the particle_output.dat file at the end of the simulation.
  2. Second simulation without the wall at the bottom of the hopper. Copy and rename particle_output.dat from 1) into particle_input.dat so you an use it as initial conditions.
1 Like
#3

Hi Jeff, thanks for your response. To provide a little more detail, I want the particles to form a pile, yes, but one that is unsupported by side walls. I am trying to recreate plots of the sand-pile pressure dip.

After writing my post, I discovered the MFiX Tutorial 3.6, which describes allowing particles to exit through the bottom of the domain; I was able to modify these steps to allow the particles to leave through the sides of the domain, specifically through the sides of the bin.

I do appreciate knowing that it’s that straightforward to set the output data as the input data. Note that the bin I refer to is the lower portion of the entire hopper. If I wanted to let the particles settle before falling into the bin, I think an easy process would be to just adjust the initial position of the particles – does that accomplish what you describe? Or is it different, due to the particles randomly falling and settling in step 1., but being assigned positions in what I propose?

#4

Thanks for the additional details. Yes you can use a pressure outlet on the side instead of a solid wall. You can specify the coordinates for the initial positions of particles, but it is not easy to insure they are packed and randomly distributed. You will likely need to write a script to achieve this. This is why I was proposing to run a simple simulation first and use this as the starting point for the production run.