Particles ending up in geometry (again)

Clemenslischka

Hi all,

i have again problem with particles ending up in my mixer geometry. I tried a low timestep (10% of Rayleigh) and have increased the Young's Modulus value to the max 1e14. Still particles go into the mixing tool, which rotates at a velocity of 25 m/s. (~15.000 rpm). I know this value is probably high but shouldn't that be doable nonetheless? I also tested a higher resolution mesh but this seemed to only slow down the calculation.. Also tested double instead of single precision. Nothing helps..

Ignacio Diez Alonso_21259

Hello Clemens,

In addition to what @Stephen Cole has mentioned above.

What I would recommend if you have to run the simulation with the current set-up, would be combining multiple approaches. Using a high geometry shear/young modulus, a low time-step value (I would try a lower one than 10% as due to the rotational velocity the solver may not be calculating the updates fast enough to ensure that the contacts with the geometry are solved before the particle has reached a region that is fully within the geometry elements). Lastly what I would also recommend doing is to create 3-5 copies of the geometry and offset them towards the inside so that you have multiple walls and each has a gap to the previous one of 1/10 of the smallest particle diameter or smaller. The mesh resolution should not make too much of a difference, so I would not worry about it much.

I hope that this helps.

Kind regards

Ignacio

Stephen Cole

Hi Clemens,

 

The particles getting stuck in the geometry is unlikely to be precision rather that force applied.

 

Is there an area in the mixer where the blade is trapping the larger particles against a wall.  When we set motions in EDEM they typically follow a fixed path regardless of opposing force, if a particle gets trapped between a wall an moving blade the blade will always go through the particle regardless of properties if the particle has no where to go.  In reality the particle may get crushed or the blade stop rotating.

 

If you scaled the particles up in the system this would likely be the cause of the particles getting trapped, I'd recommend stepping back in the simulation to check the scenario which causes the particles to be trapped and then this should help identify a route forwards such as modifying the geometry to create a larger gap, introducing breakage or force-feedback for the blade.

Regards

Stephen

Ignacio Diez Alonso_21259

Hello Clemens,

In addition to what @Stephen Cole has mentioned above.

What I would recommend if you have to run the simulation with the current set-up, would be combining multiple approaches. Using a high geometry shear/young modulus, a low time-step value (I would try a lower one than 10% as due to the rotational velocity the solver may not be calculating the updates fast enough to ensure that the contacts with the geometry are solved before the particle has reached a region that is fully within the geometry elements). Lastly what I would also recommend doing is to create 3-5 copies of the geometry and offset them towards the inside so that you have multiple walls and each has a gap to the previous one of 1/10 of the smallest particle diameter or smaller. The mesh resolution should not make too much of a difference, so I would not worry about it much.

I hope that this helps.

Kind regards

Ignacio

Clemenslischka

Thanks guys,

 

that is very helpful indeed. I will try check your suggestions and come back to you.

It makes sense to me now that particles might get stuck between the wall and mixing tool.

@Stephen Cole 

You mentioned e.g. force-feedback for the blade. Can you point me to direction where this is explained?

 

Cheers!

Stephen Cole

Clemens Lischka_20358 said:

Thanks guys,

 

that is very helpful indeed. I will try check your suggestions and come back to you.

It makes sense to me now that particles might get stuck between the wall and mixing tool.

@Stephen Cole 

You mentioned e.g. force-feedback for the blade. Can you point me to direction where this is explained?

 

Cheers!

Hi Clemens,

With regard to force-feedback there is the option of coupling to MotionSolve for a more complex control of the motion:

 

https://learn.altair.com/course/view.php?id=157

 

Or for a simpler approach EDEM has in-built force feedback via the Motion Control option:

 

https://learn.altair.com/course/view.php?id=147&section=5 (lesson 5.4)

Regards

Stephen

Clemenslischka

Dear @Ignacio Diez Alonso ,

 

i tried your approach by off-setting the geometries which is a bit tricky to position everything exactly but seems to work, i dont see any particles stuck in the geometry anymore. BUT it seems i'm losing now mass/particles in my system at a constant rate. Could it be that the particles still get stuck in the mesh but simply get deleted out of the system?

 

Thanks!

 

Clemenslischka

Ok, i seem to have solved the mass loss issue by reducing the timestep to 5% Rayleigh time!

 

Thanks guys for the help!