Moving Wall
 

Hi there,

Does anybody knows how to prescribe a displacement for a wall? I would like to impose a sinusoidal motion to a boundary to observe how the fluid around this wall behaves. I am trying to use the option "specified displacement" with a CEL sinusoidal function, but I dont know if there is something wrong. What happens is, when I post-process the results, for some steps I have some distortions in the wire frame, exactly in the region where the mesh motion was specified.

Thanks in advance

Re: Moving Wall
 

Hi, I had this problem too, I don´t know if it´s exactly the same as yours, but maybe this will help you. The problem was that I specified the motion of a piston inside a cilinder, but I forgot to set the cillinder wall as "Unspecified" instead of "Stationary", so the piston moved but the cilinder wall stood still, and then the mesh became really weird a few steps after I startes simulating. Maybe your problem is that you specified the motion of the wall but forgot to set the mesh that is in contact with this wall as "Unsepecified".

Re: Moving Wall
 

Hi, I also thought that, but I changed all the walls around to unspecified motion and nothing changed. About this approach, when I configure a wall with unspecified motion, what does it means physically? Is it a closed wall with some compliance?

thanks in advance

Re: Moving Wall
 

Sorry, I didn´t understand what you´re trying to say really well, but I guess what you´re asking is if when you set unspecified motion for a wall, and it moves, if the wall is being comprised or it diminishes. If you´re dealing with fluid the wall simply diminishes, some part of it cease to exist, or is created, according to the movement of the specified movement part. Sorry if I didn´t help much, but I really didn´t understand what you tried to say with "closed wall with some compliance".

Re: Moving Wall
 

Hi,

Mesh movement is done by defining the motion of the wall boundaries and using a Laplacian mesh smoother to generate the position of the internal nodes. When you set a wall to "unspecified" then rather than explicitly defining the location of the wall then the positions of the nodes in this wall will also be found using the mesh smoother.

Glenn Horrocks

Re: Moving Wall
 

Ok, I understood how it works, but if I simulate a closed tube, where the closing part is set to wall with unspecified motion, what am I doing? Is it working like a tube rigidly closed, a closing with some compliance or what?

I tried to simulate a closed tube where the input is a sinusoidal prescribed mesh motion. I would expect some wave behavior because of the fluid compressibility, but I couldn't understand how this wall at the end works physically. In a real situation, depending on the frequency of the input displacement I would have different pressure waves propagation. Also according to the end stiffness.

Can you help with this question?

Thanks..

Re: Moving Wall
 

ok, now I see where you´re trying to get. In your tube case, even if you set the closing part as unspecified motion, it will not move! I´m simulating a similar case, it´s a compressor cilinder, and the valves open accordingly to the increasing or decreasing pressure, caused by the sinusoidal movement of the piston, but for that to happen I had to set the valves as specified motion, accordingly to a CEL that calculates the fluid force on the valves and diminishes it from the spring force acting on the valve. So if you want your closing part to move you´ll have to write a CEL telling it when to move according to the pressure acting on it, but it will move as a rigid part, if you want it to have some elasticity it´s more complicated and unfortunatelly I can´t help. There´s a really good tutorial on CFX that you can read, it´s the fluid-structure interaction. Hope I helped....

Re: Moving Wall
 

Hi Gustavo,

Sorry for the long time without answer. I was not here in the last days. About your last answer, I did exactly what you said. About rigid body motion, it is not a problem, because is my first approach.

when I use this idea of force equilibrium, something doesn't work as well. I used the idea of the tutorial 20 but it doesn't work.

I set the mesh region to "specified displacement", and used as CEL expression for the displacement of this mesh the equation:

displ = force_x()piston / k

where piston is the region where the fluid forces act, and k is the elasticity of the spring.

thanks in advance

Re: Moving Wall
 

hummm ok. I really don´t know what´s happening there, but let me ask you something that maybe will solve this problem, did you remember to put in the cel expression a "max" funtion? for example: displ = max (0, force_x()piston / k ) so that the top of the cilinder can´t go down, because of the spring force. Hope I helped.

Gustavo

Re: Moving Wall
 

Ahhh... no! I will try it... thanks!