Learning Forum

[Mepal]

Hello all,

Could you please suggest me what should I do.

I want to make a sand layer in a close domain. I understand that sand is granular, so I used an Eulerian granular multiphase model.

To make my model simple, In a close domain, all boundaries are “wall”, there are only air and sand (Silicon(s)). I retained many the default setting. Please see attached pictures below for my reference.

 

I think it is simple, not complicated. However, the results look weird.

By iterating with a steady way: sand spread into the air. I don’t know by heart what force makes sand spread.

Then I just "only changed to iterate in a transient way", the results look logical. The sand layer looks stable. However, the height of the sand layer looks to decrease over time.

My question is 1) Why don’t the results between using steady and transient similar?
2) Do you have any general suggestions to fix it to create a steady sand layer?
3) Should I use steady or transient in this case? (I think I should use steady because it is close boundary and there is not any forces in a boundary, but the result I did look weird, that is why I ask you this question.)

Thank you for any help you may provide.

[Rob]

What are you trying to find out?  

10micron particles will tend to follow the air flow, so think flour rather than sand when it's moving. 

In the two cases, check packing limit at the initial condition, and monitor mass. Check convergence is acheived: that's the most likely cause of weird things happening. 

[Mepal]

Hello Rob,

Thank you for your reply. I've followed your suggestion by modifying the granular properties. Then it worked

I am finding out the effect of air velocity to the movement of fine particle.

I think I should make a stable granular region before including the effect of air movement.

Below is my setting. The other I retained the defaults. (I simulated in a close boundary)

Now, I've faced the problem about the stable granular

The granular region is not stable when I implement steady iteration. Its height decreases until vanishes.

However, I justed changed to iterate with transient, then the granular region was stable.

 

Do you have any general suggestion why iterating with transient, the granular layer is stable than that with steady?

[Rob]

Check the convergence on the steady model, chances are the solver is struggling and will then lose mass: remember the solution is the equilibrium state which means a 0.1% mass loss at inifinite time is nothing is left.

The transient model may also not be fully retaining the mass, but it's a slower process. 

[Mepal]

Many thanks Rob for your suggestion,

 

For the steady, I checked the convergence, mass flux, and hest flux. I think nothing strange. Below is their pictures.

Moreover, for the transient, I checked the convergence, mass flux, and hest flux. I still think nothing strange. Below is their pictures.

That is the reason I am still curious why iterating with steady, the height of granular decreases and vanishes (although no mass and heat loss!...the pictures above). But with transient, the granular layer is stable after many iterations.

[Rob]

You're monitoring boundary flow: if they're walls they should be zero! Have a look at the volume monitors. 

[Mepal]

Thank you for your suggestion,

 

Did you mean monitering the result by using volume integrals?

Before calculating with a steady, the granular mass is 0.064 kg.

Then after 7,000 iteration, the granular mass reduced to be 2.51e-5 kg.

Does it mean the granular mass vanish after iterations?

[Rob]

It's likely the solver isn't converging well and then throws mass away.