HI Sepideh, 

The dust is other phase, other material that you need to add to simulation in the setup. Once you select more than one material e.g. air  and water, the  interface detects that simulation will be a two-phase flow. In your case, you have a gas solid flow. There many models available. VOF, Two-Fluid Model, Mixture model, Lagrangian with one, two and four-way coupling, Quadrature Methods, Populational Balances, and some variations in each one. 

I advise you start with a tutorial about solid-gas flow to help to understand as software works. 

Hi Ricardo

Thank you for responding. I'm very grateful for the advice.

I have read through the multiphase flow theory guide, and I determined that the best method to solve this problem is to use the mixture model, however, I'm not too sure how to apply the model.

I have attached a summary of what I am trying to achieve and the geometry that I am trying to configure.

Another problem I am having is trying to get the geometry right. I thought that I had to subtract my cylinder from my fluid domain, but when I did this, I would get no flow of air and particles into my cylinder. Then I was told to make the cylinder have zero thickness, which then caused problems with the meshing (I'm very new to Ansys and not so good with meshing). Another suggestion I got was to leave the cylinder and the fluid domain as separate bodies, and then name the outer surface of the cylinder as 'wall' and the top surface as 'interior'. I haven't had a successful run with this method because my meshing is causing problems.

Hi Sepideh, 

 I've read your description and it seems ok, it is an possible approach. To my understanding, do you want to consider the  cylinder's interior also? If YES, the cylinder wall can be made as a single surface (interior to Fluent), because its thickness I think it is very thin compared to de domain scale. These interior walls must be set as  regions in the mesh. Once do it, you can set boundary conditions in the pre-processor as no-slip, temperature, etc.  The given name to the  mesh parts is not important. The goal is to discriminate each surface with a boundary condition, because the solver need this information to set the boundaries conditions. Two bodies is possible, but not needed. 

These simplifications of geometry are named 'defeaturing' , because we  neglected some geometric entities or  features that are not important to the net result.

I think that there is no significant fliow inside your cylinder due to turbulence model. k-epsilon is dissipative, and the eddies that will carry the dust into cylinder are damped. try uso none turbulence model - ( is named laminar in te solver) , or SAS or LES turbulence models. 

Hi Ricardo

Thank you very much for the response.

I understand what you are saying, but the problem is that I don't really know how to do it. I kept the geometry that I had made, and defined the top surface of my cylinder as 'interior', and now I am getting the error message "WARNING: Unassigned interface zone detected". I checked on the Ansys forums to find out what this means, but I did not find the responses very helpful. 

I have uploaded an archive of the program I am working on if you would like to take a look: https://drive.google.com/open?id=0B9wuwoOGjZhIejJxS09qaURraE0

Again, thank you very much for all the help you have given me. I really appreciate it. 

Hi!!   I've some  troubles to download. I cant  download the directories. May you compact in  a one sigle file?   whats is the ANSYS release? The mine one is 12.1 and rarely I work with a workbench project. I do prefer each module in stand alone. I use ICEM instead ANSYS Mesh. I like to control each step of my mesh generation.  If your version is higher, I 'll not able to open the project. 

regarding error: have you close the cylinder with a 'interior' surface? A interface is a surface that separate two  distinct fluid regions, and you have ONE, or would to have one single fluid regions with some  wall and surfaces. You need to remove it or assign some condition   as interface to assure the flow between the two zones. Otherwise, no fluid or dust  will flow to cylinder and from it. 

See the attached  figure 

May you post some print screens to help me to help you?

If is an interface, the solution is only inform to fluent that type of interface  it is. 

Hi Ricardo

I tried to implement your advice, and I think I may have managed to do it. I now have two different interfaces that separate my fluid domain volume from my cylinder volume. (I have attached a document with some images).

I am not sure if I should join these two interfaces, or if it is necessary to do this to get the flow inside the cylinder. 

I did a run and I got some results, but I couldn't get them to display very nicely. 

Thank you

Hi Ricardo

I finally managed to get some results. I don't know if they are right though. I have attached two of the images I obtained.

I'm not too sure if there is flow inside my cylinder or how to see the inside of my cylinder.

Thank you for your help so far. I really appreciate it.

HI Sepideh, you have a good evolution.

The interface must be assigned as an interface. An interface has two side, one to each domain. YES you have two fluid domains! Once a interface is assingned, the solver will put a default condition - conservative flux.  ( it can be changed, but let the default) 

I was off line this weekend, sorry.  Fisrt , yourmesh processs is very complicated, it can be more easy. But it is common at begging in CFD . Do not worry about it. The flow around cylinder it seems Ok! . To see inside the cylinder, there are  some options:

IN CFD post, 

1) whan you generate the streamlines, select ALL DOMAINS. If is not able to do it. Open the [...] button side of the edit box, and all regions will open. CONTROL+ left Button of mouse will select any regions you want. 

1) select the cylinder wall, double click, and a edition windows will open at left in the botton. Select Render Tab, and set some transparency such as 0.5 ( it means 50%) . If all right,  you will be able to see the streamlines inside cylinder.

2) With all surfaces turned of f -  see only wireframe - create a plane: in Location  Menu>Plane. Select  a plane that a cuts your domain. The same way as before, in render tab you can change the color of plane according to same variable value, as velocity. Or, as a plane is now a location, you can operate any plot operation over it. Create vectors, surface streamlines, contour. Try all these possibilities and you will see if your data has some meaning. 

IF NOT, your interface is not working. But as I've told you. You do not need an interface. Generally we choose to have more than one domain if we have  different physics such, a solid inside a fluid, interaction with structure and the fluid. Porous and fluid domais. In these cases we need more than one domain. You have ONE physics : a solid-gas flow outside and inside from cylinder. 

Hi Dear, 

I did a simple tutorial to you use ICEM instead WORKBENCH. ICEM is part of the ANSYS suit. Is is fast and  better.

Technical Report Example of Mesh Generation on ANSYS ICEM

Good morning Ricardo

I just wanted to let you know that I have managed to successfully complete my project and get the desired results.

I want to thank you for all the help and advice that you offered, and apologise for all the trouble that I caused you.

I am extremely grateful for the help you offered. I would not have been able to complete my project without your help. If I can assist you in any way in the future, please do not hesitate to let me know. 

I hope that you are having a pleasant week, and again, thank you!