Learning Forum

[Danmei]

Hi all,

 

I'm learning to use PBM model. From what I've learnt so far, if the concentration of particles is small enough, they won't affect the flow field of the fluid flow. 

As a starting point,I'm simulating a 2D square domain with an inlet on the left and an outlet on the right, moving walls on the top and the bottom to create a turbulent flow.

I couldn't find where to set one way coupling. In the manual, only the physics were explained.

What I did was:

1. Install PBM, Discrete method, Set bins and kernels

2. Use E-E multiphase, k-e for turbulence

3. Transient

What I'm looking for:

Particle size distribution at the outlet. I was successful with this method( I learnt this from a bubble column PBM tutorial). But I think what I did is "two way coupling"

What I tried:

I looked up some videos on youtube, but 99% of them are talking about FSI, I'm not sure how to use this for PBM. 

Can anyone who has experience with this help me out? How to perform one way coupling for CFD-PBM model?

//Notes: I referred this model (B.Wan et al. 2005)

 

Thanks a lot!

 

Danmei

[Rob]

Eulerian model is used for PBM, so particles must effect the fluid flow (check the manual). One way coupling is the DPM model but that's not part of PBM. 

[Danmei]

Hi, 

Thank you so much for pointing this out! I revised the manual. Yes, it is using Eulerian model.

However, given my case (my goal) is to be able to simulate a stirred mixing tank (eventually I'd like to use what I learnt for a 3D mechanical stirred mixing tank). I collected the following information that may be useful for me:

1. DPM uses lagrangian reference fram for the discrete phase.

2. It is capable to simulate the breakup and coalescence of droplet, even in turbulent flows.

3. works when the volume fraction of the discrete phase is less than 10%-12%.

4. DPM cannot effectively model flows such as solid suspensions within closed stirred tanks. But it can be solved by using unsteady discrete model.

I just want to confirm that, as a starting point, in order to simulate a stirred tank with, one way coupling, and track the particle size distribution, I should use the unsteady-particle discrete phase model and turn on stochastic tracking.

Appreciated.

 

Best

Danmei

[Rob]

1-3) Pretty much. The VF limit is because DPM uses a point mass so much over about 12% you lose accuracy, the model will still work so small areas with a higer VF are usually acceptable. 

4) Yes. You may also want to look up frozen flow models: especially if the solids don't affect the flow. 

Have a chat with your supervisor: the modelling you're doing is relatively simple (if you know what you're doing) and there are some additional materials on the Customer Portal which I can't share on here. Start with single phase, reference frame model on the mixing tank and see how the flow behaves. There is a tutorial in the documentation: click "Help" to have a look.