[Mu_CFD]

Hello everyone,

My simulation consists os a venturi scrubber with multiphase model to predict the momentum balance. The approach used is eulerian-lagrangian to obtain the track profile of particles and droplets.
The three components are air - continous phase-, water injected as 4 point sources near the throat in the form of droplets - particle transpor liquid - , and dust injected in the mixture with air through the inlet - particle transport solid.

I have used the SST model because of the shape of the geometry and one-way approach to air-dust, because they have a much smaller diameter ( 2 micron) and consequently follow the air trajectory, and two-way coupling to air-water because they are large (5mm rosin ramler) and consequently affect the continous phase more intensively.

I'm interested in simulate the position profile of water droplets and dust to calculate collection of dust particle along with the pressure drop across the geometry.

So far I've been able to attain good results in the trajectory of water droplets and dust particles, speeding up in the throat section and slowing down in the divergent section. I'm also activate the droplet breakuo model to refine the droplet size distribution and the mean diameter of the droplet was reduced due to drag forces in the surface of the droplet.

I've searched for some literature and here, in other threads, and couldn't find exactly a way to do this.

What I've found out is that there is a collision model capable of consider the particle shocks and set perpendicular and normal coeficients to define the elasticity rate of the collision. The problem is, lookinf ta CFX manuals, when i set the collision model for water, it will consider the collision between droplets, as well as if I set the collision model for dust, it will consider only between itsef.

If somehow I could model the collision between droplets and dust and set that, for dust, to interrupt the particle tracking, there would be a less mass flow rate at the outlet that, compared with the load at the inlet, the efficiency may be calculated.

The questions are:

1) Are my consclusion right about the the collision model?
2) The interaction between particles is a four-way coupling. In this case in CFX the set in two-way coupling in fluid pair model and activate the collision, eriosion models in the case of a four-way?

I would like to thank you anticipated.

[Mu_CFD]

The problem seems unsolvable...

[ghorrocks]

Your proposed simulation goes beyond the built in models of CFX (collisions between different particle species) so you are going to have to develop some of your own models to handle these physics. This is going to increase the degree of difficulty.

Also your application is very specialised, so you are not going to find anybody with experience doing what you propose to do. You are going to have to work a lot of it out yourself.

I do not understand your two questions - what conclusion are you talking about? I don't understand what your four way coupling question is asking.

[Mu_CFD]

When I say a four-way coupling, it means that, as well as, Crowe, the interaction between particles.

My question was if particle collision model is capable of simulate only particles of the same specie I think you answered my question in your first explanation since there is no model capable of deal with different particles interaction in CFX.

When you say that I'll have to develop my own model the question that comes to my mind is how? It means that I'll have to write a code? Is there some material that explain the step-by-step? Not in my specific case of course because I understand that it's not ordinary but in theorical ways or examples in other simpler events for learning.

Because only knowing how to configure I'll be able to work on the theory behind the physical event.

Thank you and I hope my words were clear.

[ghorrocks]

You might want to consider a discrete element model (DEM) like EDEM or Rocky. I think both of these can be coupled to CFX. You can model far more sophisticated particle interactions with those softwares.

But before you do, why can't you do this model using the Eularian particle model? It would be much simpler to implement this particle model in a Eularian framework than a Lagrangian one. In my experience people assume that particle tracking means Lagrangian models, and are not even aware that it can be done in CFX as a Eularian model as well.

[Mu_CFD]

I've tried to use the eulerian-Eulerian model but I had some problems in convergence because of the specific geometry.

The water is injected by a set of four orifices, tangencially to the circular cross section area in the converging section of the equipment. Because of these, I had to have the four inlets added to the geometry, which needed a short lenght of tube. There were some swirl inside this tube that jeopardize the whole results.

In addition, the eulerian-lagrangian model is recommended to systems with low water volume fraction, which is my case. At least that is what I've found in the literature. Since my interest is the contact, when the occur, I thought that the eulerian-lagrangiam model is more apropriate.I'm not interested in the interface between the phases.

About de EDEM or Rocky, they are like suplements of the workbench or, like, other program, that the results are imported somehow to CFX? I'll look formmore details aswell.

Thank you for your help.

[ghorrocks]

In addition, the eulerian-lagrangian model is recommended to systems with low water volume fraction, which is my case.

Why do you say that? I think you will find that the Lagrangian model in CFX is ONLY applicable at low volume fraction. The Eularian model in CFX can work both at low and high volume fraction. But for a low volume fraction flow you will find both Lagrangian and Eularian approaches are suitable so the decision on which to go with depends on other factors. So don't use the Lagrangian model just because you have low volume fraction.

EDEM (https://www.edemsimulation.com/) and Rocky (https://rocky.esss.co/) are not ANSYS software, they are other providers. So they are not supplements to Workbench, but I know they both can couple with CFX.

Regarding the contact between the particle types, have you thought about what is happening in detail when a water and dust particle either collide or near miss? I think you will find that it is not simply a matter of whether the trajectories intersect. I suspect you will find things like electrostatics have a role (which will cause particles to collide more) and the aerodynamics of two particles approaching each other might have a role as well (which will cause particles to collide more). Detailed physics like this is not included in either the DEM, Lagrangian or Eularian model, you will have to include it yourself as a custom model.