I want to simulate water flowing into a vase, like this : http://img.photobucket.com/albums/v4...pnet/mykem.jpg But the problem is I don't know how to set up a variable to simulate the flow !!! I think there must be a variable, which is called "water" or something like that, otherwise, when using CFX-Post, we just have : pressure, temperature, velocity, …. But we don't know how water flows. Therefore, I use a variable called "water", type of which is "Volumetric" and Units is "kg m^-3" but I think it's not correct. The simulating result looks very bad. This is my result : http://img.photobucket.com/albums/v4...t/myresult.jpg So, can you tell me how to set up a variable to simulate water flowing into a vase ? I need the simalation to be as real as possible ! Like this picture : http://img.photobucket.com/albums/v4...et/mychuan.jpg

Did you use VOF method?

I don't know ! In Ansys Multiphysics, we can use VOF, but limit with 2D model ! We can't use VOF method with 3D model in Ansys Multiphysics ! In CFX, I don't know how to use VOF ??? thanks for reading !

Hi,

You have to chose 2 fluids (multiphase flow): Water and Air (I suppose the vessel is previously full of Air). The presence of water will be characterized by its volume fraction (Water.vf in CFX-Post)

In CFX-Pre, Domain, Fluid Models, chose the Standard Free-Surface Model (take a look at the CFX-5.7.1 Documentation, Solver Theory, Multiphase Theory, page 161). You may chose the Homogeneous or the Inhomogeneous Model, I don't know which one will work better in your case, but I think you may start with the Homogeneous Model and if you cannot get good results change to the Inhomogeneous Model (take a look at Solver Modelling, Multiphase Flow Modelling, pages 154-155).

Take a look at Tutorial 7 (Free Surface Flow over a Bump), and if you have access to the CFX-Community web site, take a look at this paper: http://www-waterloo.ansys.com/cfxcom...CFXTP00341.pdf . Both deal with steady-state situations, but they deal with water-air free surface flows. However, in transient simulations you are not allowed to use mesh adaption.

About your model:

1- If it is axissimetric, I suppose you may model just a slice of the domain and apply symmetry boundary conditions.

2- There has to be an outlet for the air to leave the vessel. For what I saw in one of the .jpg there isn't such boundary.

3- You used the same time-steps as in Tutorial 4. Do the equations converge? I'm nearly sure they don't. For this kind o processes you will need quite smaller time-steps

Regards,

Rui

Usually for free surface flows, we use Volume Of Fluid (VOF) method. If you don't know the answer about my previous question, it means that you don't use VOF method because for that with CFX you need to activate some options to use VOF method.

Following your geometry, why do you use a full 3D model. You should perhaps use a slice of your geometry with only 1 cell layer in the thickness. It is a way to model a 2D geometry with a 3D software.

You should read the manual to know how to set-up a VOF model and do some tutorials about this.

Good luck.

HV

PS: I am actually trying to learn vietnamese and it was funny to see your boundary conditions or fluid name written in vietnamese.

To HVN : You're right ! I'm very suprised !!! You knew me from Vietnam ? Why you're trying to learn Vietnamese ? I have just read example No.7 of CFX, so, I think I knew the way to set up VOF method in CFX 5.7. Anyway, can you tell me more clearly about what you say : "Following your geometry, why do you use a full 3D model. You should perhaps use a slice of your geometry with only 1 cell layer in the thickness. It is a way to model a 2D geometry with a 3D software." PS : Anyway, where are you come from ? thanks everbody !!!

See tutorail 7, 11 or 19 for an example of 2D simulations with a 3D CFD software. In other words, take one section of your geometry with a small thickness and put symmetry conditions on both sides of your section. But depending of your meshing software, it would be easy or not to mesh efficiently (for myself, I would use Gambit or Prostar for that).

HV

PS: I am a vietnamese french.

To HVN : thanks for helping ! But I think what I need is 3D model and transient ! I understood your meaning : "2D simulations with a 3D CFD software", but I think in this case, it's not enough ! First, now, I am trying to do example No.7, in transient ! But I don't know what I did right or wrong, hic ! Anyway, can you show me the way to do Ex.No.7 in transient ??? You know, my english is bad, so, you should be sympathetic towards me, trying understand what I said !!!

PS : to HVC : can you give me your e-mail ? My name is TRUNG

thanks everybody !