CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   linear mesh motion in CFX (http://www.cfd-online.com/Forums/cfx/125100-linear-mesh-motion-cfx.html)

alinik October 18, 2013 16:18

linear mesh motion in CFX
 
Hi Guys,

I want to know how I can define a linear mesh motion in CFX?
In Basic setting>Domain models>Domain motion I can only define rotating kind of motion, not simple translational motion.

Thanks

ghorrocks October 20, 2013 20:07

You need to use moving mesh and define the translation.

alinik October 20, 2013 20:14

I can only define rotary motion. I can not find translational motion in the domain motion menu. could you please more specific how I can define translational motion?

Thanks

ghorrocks October 20, 2013 20:17

You are looking in moving frames of reference. The only available option is rotating. I am talking about moving mesh - this is a different thing.

alinik October 20, 2013 20:22

thank, but I am looking in this menu and I cannot find it
Basic setting>Domain models>Domain motion

can you please tell me where I have to look to define mesh motion?
and by the way what is the difference between mesh motion and domain motion?
If the translational option is provided in CFX it will be a life saver for me.

Thanks

ghorrocks October 20, 2013 20:41

Because it is not under domain motion. It is under mesh deformation (it used to be called moving mesh, sorry about that).

alinik October 25, 2013 15:07

Glenn,

There is only one option under mesh deformation and it is called "Regions of motion specified". When I choose it, I get to choose mesh motion model and the only option available is "displacement diffusion" in which I can set the mesh stiffness model, etc.

I can not see any option by which I can introduce linear mesh motion in CFX.
could you please more specific.

I am using CFX 12.0.1

Thanks

ghorrocks October 26, 2013 06:49

The option might not be available in V12. The current version is V14.5, and the beta of V15 is already out. Time to update to the current version I think.

brunoc October 28, 2013 09:48

Quote:

Originally Posted by alinik (Post 459043)
There is only one option under mesh deformation and it is called "Regions of motion specified". When I choose it, I get to choose mesh motion model and the only option available is "displacement diffusion" in which I can set the mesh stiffness model, etc.

I can not see any option by which I can introduce linear mesh motion in CFX.
could you please more specific.

I am using CFX 12.0.1

Thanks

You're looking at the domain settings, so you just activated mesh motion. To actually create mesh motion, go to the boundary condition you want to move and you'll see a new tab there to define the motion itself.

In case you want to move the entire domain, create a subdomain that includes the entire domain you want to move, and set the motion in there.

You'll need to inform a CEL Expression that describes the motion in both cases.

BTW, do update to a newer version. The mesh smoother is a bit better, plus remeshing is a little easier (but just a little).

alinik March 24, 2014 19:06

Quote:

Originally Posted by brunoc (Post 459424)
You're looking at the domain settings, so you just activated mesh motion. To actually create mesh motion, go to the boundary condition you want to move and you'll see a new tab there to define the motion itself.

In case you want to move the entire domain, create a subdomain that includes the entire domain you want to move, and set the motion in there.

You'll need to inform a CEL Expression that describes the motion in both cases.

BTW, do update to a newer version. The mesh smoother is a bit better, plus remeshing is a little easier (but just a little).


Thank you very much Brunoc and Glenn

Now I have upgraded to R15 and have enabled beta features. Still I cannnot see the options you guys are saying. I have activated mesh motions but in neither of the boundaries I can see any tabs being added.
Brunoc do you have a similar simulation? If yes can you post the ccl here so that I can take a look at it? I am following you especially where you say I need to inform a CEL expression that...
I want to model wakes originating from a series of Rods that are moving upwards in a domain. I have tried many other options and it seems that enabling this linear sliding mesh is going to be the only solution.
So I guess I have to introduce CELs for inlet? am I right?
Is there a tutorial that explains this fully in details. I really need to do this.

Any help is much appreciated

PeMo March 25, 2014 05:27

Have a look at this tutorial
http://www.edr.no/index.php/blogg/an...cfx_re_meshing
It explains the basic function of moving mesh and automatic remeshing.
Also you see the additional tabs and CEL Bruno was referring to.

alinik April 1, 2014 12:44

PEMO,

Thanks for this tutorial. This is exactly what I am looking for. the problem is the simulation ends before it starts. and I receive this message:



--> Final synchronization point reached by all partitions.

+--------------------------------------------------------------------+
| An error has occurred in cfx5solve: |
| |
| The ANSYS CFX solver exited with return code 1. |
+--------------------------------------------------------------------+

End of solution stage.

+--------------------------------------------------------------------+
| The following transient and backup files written by the ANSYS CFX |
| solver have been saved in the directory |
| /scratch/nikp11/WD/remeshing/remeshing_001: |
| |
| 0.trn |
+--------------------------------------------------------------------+


+--------------------------------------------------------------------+
| The results from this run of the ANSYS CFX Solver have been |
| written to /scratch/nikp11/WD/remeshing/remeshing_001.res |
+--------------------------------------------------------------------+


My domain consists of two domains, one of them is stationary and the other one is the one that is supposed to have moving rods in it. I do every thing step by step according to this example but I do not know why it ends before it starts. It is not even giving me a good description of error.

any help is much appreciated.

Thanks

alinik April 1, 2014 21:00

any idea PEMO?

PeMo April 2, 2014 05:59

Difficult to say what caused the error depending on this snippet. Have a look at the trn and res file for any wrong initialization and boundary setting. Also I would switch off any solver equation (expert parameter), set the coefficent loop to 1 and an appropriated time step. So you can have a look only at the moving mesh, to be sure there is no folding and negative cell volumes during the expected simulation time.

alinik April 7, 2014 14:50

PEMO

I fixed the previous problem but now I have this new error:


+================================================= ===================+
| ****** PROBLEM REPORT ****** |
|--------------------------------------------------------------------|
| Subsystem: Input and Output |
| Subroutine name: ErrAction |
| Severity level: Fatal Error |
| Error message number: 001100279 |
|--------------------------------------------------------------------|
| Message: |
| |
| REDHDR: locating dataset failed: what=G/NAMEMAP where=EVERY |
| |
| |
| |
| |
| |
+================================================= ===================+
WARNING: CFXSTP, Regression check error.

+--------------------------------------------------------------------+
| An error has occurred in cfx5solve: |
| |
| Error interpolating results onto the new mesh: |
| /g/software/ansys_inc/v150/CFX/bin/linux-amd64/double/solver-pcmp- |
| i.exe exited with return code 1. |
+--------------------------------------------------------------------+

It seems that it is due to lack of enough memory. I have increased the memory allocation factor but still have this issue.
Have you ever had such error before?

Thanks,

Ali

alinik April 10, 2014 18:35

Guys,

apparently it is not a memory issue. I checked the error log file and it says that one of the boundaries(the one that is moving) is not valid and hence it is not recognized. I do not know what to do?

ghorrocks April 10, 2014 18:37

Yes, it is unlikely the error is a memory issue. Messages like this tell you very little about what the error actually is. All you can do is simplify your model until it runs, then add the complex bits back in one at a time until you get the error again. Then you know where the error comes from.

alinik April 10, 2014 18:48

I can not simplify it any more. it is now a very small domain in which two rods are moving upwards and all surrounding boundaries are no-slip wall.
It says that :
the location of boundary "ROD1" is invalid
Region "ROD1" is not recognised.
the location of boundary "ROD2" is invalid
Region "ROD1" is not recognised.

Any ideas?

ghorrocks April 10, 2014 19:34

Looks like you have found it - your boundaries ROD1 and ROD2 are not correctly defined. Delete them and redefine them.

alinik April 10, 2014 20:00

Did that and ran it again.
Still the same error. :(
any other ideas?


All times are GMT -4. The time now is 19:09.