CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Simple Impinging Fountain problem (http://www.cfd-online.com/Forums/cfx/84625-simple-impinging-fountain-problem.html)

MarkK February 3, 2011 21:29

Simple Impinging Fountain problem
 
1 Attachment(s)
Hi, I'm new to CFX (and CFD in general). I've gone through most of the tutorials and thought I'd play around with a simple impinging fountain problem (i.e. fluid being shot at a wall) to get a feel for CFX. The tutorials already have most of the work done for you, so I thought it would help to do some of my own tutorials.

Essentially, I've simply got a small pipe with an inlet inside, shooting air (at 25 C) out at 500 m/s (the distance from the inlet to the bottom surface is 15 mm, just for reference). My problem is that in CFX Post, the 'vector' tool shows the air moving very fast, but it just ends (not even stops; the air doesn't slow down, there just is suddenly no velocity). You can see this in the picture I've attached. It makes it a little past the opening and into the fluid domain, but suddenly is no more.

I've ensured that the opening at the end of the pipe has been set as a boundary 'opening', I've made sure the vector tool is set for all domains, I've set the global initialization (no idea if that effects it), I've ensured the three parts are separate sketches and each rotation is frozen. I could've sworn I'd got all of the basics done, but clearly I haven't.

It's probably something very simple that I've missed. I've checked the tutorials, tried googling it (though I'm not sure what search words I should use) and looked through the user guide, but I haven't been able to find what's going wrong. Would anyone happen to know what I've missed?

Thanks in advance for any answers.

(Also, the reason there's a line through the center of the model is because I'm trying to figure out the symmetry boundary condition, so I've reflected one half to the other side. This isn't the reason the velocity ends; the same thing happened when I tried it with a fully rotated pipe which doesn't need any symmetry boundary conditions)

ghorrocks February 4, 2011 06:36

By the way - air at 500 m/s will probably be very compressible and you are currently running an incompressible model. It will converge easily but be totally wrong.

What are you showing in the image? Draw a plane which cuts through the domain and draw vectors on that.

Is your mesh connected? If not this can cause havoc.

MarkK February 4, 2011 07:56

1 Attachment(s)
Quote:

Originally Posted by ghorrocks (Post 293612)
By the way - air at 500 m/s will probably be very compressible and you are currently running an incompressible model. It will converge easily but be totally wrong.

Yeah, I'm not going for accuracy currently, I'm just fiddling around with CFX so I can learn how to use it properly for a project in the near future.

Quote:

Originally Posted by ghorrocks (Post 293612)
What are you showing in the image? Draw a plane which cuts through the domain and draw vectors on that.

Sorry that it wasn't clear. The picture shows a plane that cuts right through the center of all of the domains (they're all rotated about the same point and axis). So, what you're seeing is the cross section of three cylinders. The vectors have been drawn onto this plane.

Quote:

Originally Posted by ghorrocks (Post 293612)
Is your mesh connected? If not this can cause havoc.

I'm not sure that I have. I created the 3 parts at once (made them as separate sketches, rotated them, froze them individually) and then when I got to the mesh creation stage I simply clicked on the 'generate surface mesh' and 'generate volume mesh' buttons. I thought this would mean there'd only be one mesh, therefore there wouldn't need to be any connections between meshes. Is this wrong? Does this actually create multiple meshes (or otherwise require me to still join meshes or domains)?

How do I connect the meshes? Using the domain interface tool? I tried it, but when I saved and moved on to solver, solver wouldn't load saying something about 'user input required'. I was sure to enter all of the values required in the domain interface settings (though I'm unsure of which region I should connect to the outlet).

I've included another picture (please forgive the horrible Paint job!). Hopefully that makes it a little more clear. It's only one quarter of the full revolution (as I said, I'm trying to use the symmetry boundary conditions to speed up the solution).

Thanks very much for taking time out to reply, I really appreciate it.

ghorrocks February 5, 2011 06:02

Your simulation is 2D, so model a small (maybe 5 degree) wedge with 1 element thickness. Again the answer is almost certainly not 2D but you already said you don't care about accuracy you just want it to run.

You have not connected the meshes correctly. Looks like a pretty basic mistake. Have you done the meshing tutorials? There are some on the ANSYS Community website.

Do not connect them with interfaces. Fix the mesh properly and you will not need interfaces.

MarkK February 5, 2011 06:30

Quote:

Originally Posted by ghorrocks (Post 293766)
Your simulation is 2D, so model a small (maybe 5 degree) wedge with 1 element thickness. Again the answer is almost certainly not 2D but you already said you don't care about accuracy you just want it to run.

You have not connected the meshes correctly. Looks like a pretty basic mistake. Have you done the meshing tutorials? There are some on the ANSYS Community website.

Do not connect them with interfaces. Fix the mesh properly and you will not need interfaces.

OK, I'll try that soon. I've done all of the tutorials in CFX; I never realised there were tutorials for meshing on the ANSYS website! Thanks very much for that, I'll go do them now.


All times are GMT -4. The time now is 15:48.