Hi there, I am trying to simulate a diaphragm in a complex pressure vessel in CFX. I mean, I was wondering, how can we model a piston with very small displacements. I tried in these to ways:

1- I used a opening to impose the pressure inside the volume and to one of walls I gave a prescribed displacement, one described simply as the displacement of a spring with known stiffness. (displacement = Pressure in wall * Area of wall / Stiffness);

2- An other tried possibility was, to use also a wall with prescribed displacement as input, instead of the opening.

The idea is to simulate a experiment in which the input varies with frequencies and to understand the state of pressure of the compressible fluid in different geometric configurations of the vessel.

Thanks in advance

Hi,

Use whatever approach is physically more realistic (in this case it sounds like the moving wall approach). Only simplify things after you have checked the accuracy implications of the simplification.

Glenn Horrocks

Hi,

Thanks for your answer. You are right, I used the moving wall approach and I found some results. Actualy, I feel some fault in my knowledges about meshes theory, and how it works in cfd. I saw that you are always interested in moving meshes problems posted in this forum. Do you have advices about literature in this area?

In my first approach, I used the same idea presented in the tutorial 20, what helped a lot. But now, for the next steps, I was wondering: Are there some way to do a kind of modal analysis for fluids in tubes (like analytical wave theory), but in CFX?

Thanks in advice.

Hi,

The tutorial examples are good and there are many more tutorial examples for moving mesh coming in V12. I did lots of beta testing in the early days for the moving mesh stuff in CFX and used to use it extensively but not so much now. I don't have any other references, although my PhD thesis may be of general interest - though it used CFX4 which is quite different. http://hdl.handle.net/2100/248

Modal analysis - the only way in CFX is to excite the fluid and do a full transient simulation. CFX cannot work in the frequency domain like structural solvers can.

Glenn Horrocks

Hi,

About moving wall approach am trying to get more knowledge about that. I am still in this piston idea and I tried to prescribe a displacement function for a wall, to see how the fluid reacts. Do you know how to prescribe a displacement of a wall to observe the pressure distribution in the volume? I mean, set a surface as wall, then I set the boundary wall as prescribed displacement. I saw that the mesh around the prescribed moving mesh should not be set as stationary. There are more settings to take care?

Thank you in advice.

So what is your question?

Have a look at the tutorial examples for the basics of how to set moving mesh simulations up.

Glenn Horrocks

You might have better luck doing this as an ideal buildup graph of the cylinder pressure using Td=Ts(Pd/Ps)^((n-1)/n) and pd=ps(Pd/Ps)^(1/k) instead of trying to model with CFX.

http://portal.ptcuser.org/p/fo/do/download=1&fid=1283 if you have a ptcuser login and/or Mathcad.