2D Dynamic mesh simulation of a scroll expander

ghorrocks · May 20, 2018, 19:48

Double precision - CFX-Pre, in execution control. It is also an option in Solver Manager and a command line argument.

Equations of motion - I have no idea about the ICEM side of this, I do not do dynamic remeshing so have no experience with it. I am suspicious about the "Time This Run" variable, however.

Timestep - you are not even getting to the solver yet so time step is irrelevant. But when you do fix this motion issue and the solver starts the first thing to try when you have convergence difficulties is smaller time steps.

I have never used "Time This Run" and do not know exactly what it means. Are you sure this is doing what you expect? Why not just use the normal time variable, "t".

louisdub11 · May 21, 2018, 04:09

Quote:

Originally Posted by ghorrocks

Double precision - CFX-Pre, in execution control. It is also an option in Solver Manager and a command line argument.

Equations of motion - I have no idea about the ICEM side of this, I do not do dynamic remeshing so have no experience with it. I am suspicious about the "Time This Run" variable, however.

Timestep - you are not even getting to the solver yet so time step is irrelevant. But when you do fix this motion issue and the solver starts the first thing to try when you have convergence difficulties is smaller time steps.

I have never used "Time This Run" and do not know exactly what it means. Are you sure this is doing what you expect? Why not just use the normal time variable, "t".

Thanks a lot for all your useful answers !!

I tried my simulation with a simple x translation over time, and did everything you told me in the earlier post, and I got my simulation to work !!

However, it ran for about 3h30, and then it gave me an error message "no CFX input file exists" something like that, while remeshing. When I am recording my replay script on ICEM, do I also need to record the "cfx5 file export" step?

I also definitely have issues with my motions, because with the translation it works fine. But I have no motion in the CFX Post while trying to make an animation

even though I have my graphs showing X moving over time...

As for the double precision, I turned it on in the CFX Pre execution control, but when simulating, it gives me the warning "floating point is written in single precision only" something similar to that. What could it be?

Thanks again

!

louisdub11 · May 21, 2018, 05:13

Here the error I get in the simulation:

warning.PNG

ghorrocks · May 21, 2018, 06:36

Make sure you have the "Save Mesh" option selected in your transient results file options. You need that to see the motion of the mesh.

The error message is saying some input file is single precision. But keep the solver running in double precision while you are debugging. You don't want weird single precision round off errors making things more complex.

louisdub11 · May 21, 2018, 06:56

Quote:

Originally Posted by ghorrocks

Make sure you have the "Save Mesh" option selected in your transient results file options. You need that to see the motion of the mesh.

Could you please be more specific?

I don't see this option here:

output.PNG

ghorrocks · May 21, 2018, 07:00

Are you using an old version of CFX? If so it might be time to update

If you are on an old version of CFX without the include mesh option then your only option is to select from Options the "full save" or "all variables" option (I can't remember the exact term).

louisdub11 · May 21, 2018, 07:04

I am using the Ansys 18.1 release, I don't know if it is considered as an old version

But in any case, I can't update because I am using the academic version on the university computers

louisdub11 · May 21, 2018, 07:08

Maybe here?

output.PNG

ghorrocks · May 21, 2018, 07:08

The current version is 19.1, but I am sure that option was available in V18.1 as well. Maybe that option only comes up when you go to "Selected Variables". If that is the case then the mesh should be included in the output file. (This is confirmed in your later post)

louisdub11 · May 21, 2018, 07:11

Ok perfect, thanks !

Should I select all the variables? Wouldn't it slow the simulation too much?

ghorrocks · May 21, 2018, 07:25

To debug mesh motion you don't need any variables, just the mesh.

louisdub11 · May 21, 2018, 07:30

Ok, thanks a lot !!

I will see if I can do the entire simulation without problems

A last small question: as the motion equations seem to be problematic, would it be possible to make successive linear motion steps ?

For exemple, instead of doing a circular motion, doing hexagonal motion with 6 successive linear motions

(Hope I am clear enough

)

ghorrocks · May 21, 2018, 08:03

You can define the motion to do anything you can write a function to describe.

There are a few ways to describe the motion. The normal way is Displacement relative to "Initial Mesh", but you can change it to "Previous Mesh" and then you specify the incremental change (if I remember correctly). The big problem with Previous Mesh is that numerical errors add up and so if you define a motion and return back to where it started from it will often be displaced from the starting location.

But start simple, get some simple motions like linear and circular working before you go to something more complex.

Gert-Jan · May 21, 2018, 11:35

You can turn on double precision in the start-up menu (where you make choices regarding parallel settings, interpolation etc.) of the solver manager.

louisdub11 · May 21, 2018, 15:58

Okay, my simulation seems to work fine at the beginning.

But I have a problem, when I get to the remeshing step, the solver stops running and gives me this error:

error1.PNG

Here you can see the solver if it can help you see something I don't see

cfxpre.PNG

I also have another weird thing, I'm supposed to just have an X translation of one of the scroll profile, and I get this strange deformation, and I still don't have a motion even if I have activated the "include mesh" option:

scroll.jpg

Any clue?

ghorrocks · May 21, 2018, 19:21

You should do the remeshing tutorial examples on the ANSYS customer webpage for how to run remeshing simulations.

Your image shows you have not set up the motion definitions correctly. I suspect you want the impeller walls to have defined displacement, but the outer boundary to be stationary.

louisdub11 · May 22, 2018, 04:28

Quote:

Originally Posted by ghorrocks

You should do the remeshing tutorial examples on the ANSYS customer webpage for how to run remeshing simulations.

Just to be sure, the tutorials are a PDF called "Ansys CFX Tutorials", and not a set of videos?

Quote:

Originally Posted by ghorrocks

Your image shows you have not set up the motion definitions correctly. I suspect you want the impeller walls to have defined displacement, but the outer boundary to be stationary.

It is strange, because I created 2 different boundaries, the stationary part I called it "wall1" and the moving part is called "wall2". I defined a motion for one and the other one is stationary in the boundary parameters in CFX Pre

I really don't understand why there is this deformation...

louisdub11 · May 22, 2018, 05:52

I have this tutorial, but don't know if it is the one you're talking about:

Screen Shot 2018-05-22 at 11.51.11.jpg

ghorrocks · May 22, 2018, 07:36

Dynamic remeshing is not included in the normal CFX tutorials on the link you show. Dynamic remeshing involves both CFX and ICEM so two separate packages. For this tutorial you need to go to the ANSYS Customer site and search the tutorial examples there - that is where you will find the dynamic remeshing tutorials.

Regarding the unexplained deformations: Have a look at the whole mesh in the post processor. In particular look at the intersection of boundaries. At edges where a symmetry boundary and a wall meet you have the wall with a stationary motion and the symmetry with an unspecified motion. Make sure that the solver choose the stationary motion and not the symmetry motion in these cases.

louisdub11 · July 1, 2018, 18:03

Hello again,

As I couldn't have access to the customer portal for dynamic meshes, my objectives have changed: I need now to do the simulation of a journal bearing.

My first question is: when doing the mesh, should I generate a mesh for all the different parts, or should I generate a mesh only for the fluid region?

Thanks

May 20, 2018, 19:48		#41
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,705 Rep Power: 143	Double precision - CFX-Pre, in execution control. It is also an option in Solver Manager and a command line argument. Equations of motion - I have no idea about the ICEM side of this, I do not do dynamic remeshing so have no experience with it. I am suspicious about the "Time This Run" variable, however. Timestep - you are not even getting to the solver yet so time step is irrelevant. But when you do fix this motion issue and the solver starts the first thing to try when you have convergence difficulties is smaller time steps. I have never used "Time This Run" and do not know exactly what it means. Are you sure this is doing what you expect? Why not just use the normal time variable, "t". __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 21, 2018, 06:36		#44
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,705 Rep Power: 143	Make sure you have the "Save Mesh" option selected in your transient results file options. You need that to see the motion of the mesh. The error message is saying some input file is single precision. But keep the solver running in double precision while you are debugging. You don't want weird single precision round off errors making things more complex. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 21, 2018, 07:00		#46
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,705 Rep Power: 143	Are you using an old version of CFX? If so it might be time to update If you are on an old version of CFX without the include mesh option then your only option is to select from Options the "full save" or "all variables" option (I can't remember the exact term). __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 21, 2018, 07:08		#49
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,705 Rep Power: 143	The current version is 19.1, but I am sure that option was available in V18.1 as well. Maybe that option only comes up when you go to "Selected Variables". If that is the case then the mesh should be included in the output file. (This is confirmed in your later post) __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 21, 2018, 07:25		#51
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,705 Rep Power: 143	To debug mesh motion you don't need any variables, just the mesh. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 21, 2018, 05:13		#43
louisdub11 Member Louis Dubail Join Date: May 2018 Location: Brussels, Belgium Posts: 52 Rep Power: 7	Here the error I get in the simulation: warning.PNG

May 21, 2018, 07:04		#47
louisdub11 Member Louis Dubail Join Date: May 2018 Location: Brussels, Belgium Posts: 52 Rep Power: 7	I am using the Ansys 18.1 release, I don't know if it is considered as an old version But in any case, I can't update because I am using the academic version on the university computers

May 21, 2018, 07:11		#50
louisdub11 Member Louis Dubail Join Date: May 2018 Location: Brussels, Belgium Posts: 52 Rep Power: 7	Ok perfect, thanks ! Should I select all the variables? Wouldn't it slow the simulation too much?

May 21, 2018, 07:30		#52
louisdub11 Member Louis Dubail Join Date: May 2018 Location: Brussels, Belgium Posts: 52 Rep Power: 7	Ok, thanks a lot !! I will see if I can do the entire simulation without problems A last small question: as the motion equations seem to be problematic, would it be possible to make successive linear motion steps ? For exemple, instead of doing a circular motion, doing hexagonal motion with 6 successive linear motions (Hope I am clear enough )

May 21, 2018, 08:03		#53
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,705 Rep Power: 143	You can define the motion to do anything you can write a function to describe. There are a few ways to describe the motion. The normal way is Displacement relative to "Initial Mesh", but you can change it to "Previous Mesh" and then you specify the incremental change (if I remember correctly). The big problem with Previous Mesh is that numerical errors add up and so if you define a motion and return back to where it started from it will often be displaced from the starting location. But start simple, get some simple motions like linear and circular working before you go to something more complex. louisdub11 likes this. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 21, 2018, 11:35		#54
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,827 Rep Power: 27	You can turn on double precision in the start-up menu (where you make choices regarding parallel settings, interpolation etc.) of the solver manager. louisdub11 likes this. Last edited by Gert-Jan; May 22, 2018 at 07:33.

May 21, 2018, 15:58		#55
louisdub11 Member Louis Dubail Join Date: May 2018 Location: Brussels, Belgium Posts: 52 Rep Power: 7	Okay, my simulation seems to work fine at the beginning. But I have a problem, when I get to the remeshing step, the solver stops running and gives me this error: error1.PNG Here you can see the solver if it can help you see something I don't see cfxpre.PNG I also have another weird thing, I'm supposed to just have an X translation of one of the scroll profile, and I get this strange deformation, and I still don't have a motion even if I have activated the "include mesh" option: scroll.jpg Any clue?

May 21, 2018, 19:21		#56
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,705 Rep Power: 143	You should do the remeshing tutorial examples on the ANSYS customer webpage for how to run remeshing simulations. Your image shows you have not set up the motion definitions correctly. I suspect you want the impeller walls to have defined displacement, but the outer boundary to be stationary. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 22, 2018, 05:52		#58
louisdub11 Member Louis Dubail Join Date: May 2018 Location: Brussels, Belgium Posts: 52 Rep Power: 7	I have this tutorial, but don't know if it is the one you're talking about: Screen Shot 2018-05-22 at 11.51.11.jpg

May 22, 2018, 07:36		#59
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,705 Rep Power: 143	Dynamic remeshing is not included in the normal CFX tutorials on the link you show. Dynamic remeshing involves both CFX and ICEM so two separate packages. For this tutorial you need to go to the ANSYS Customer site and search the tutorial examples there - that is where you will find the dynamic remeshing tutorials. Regarding the unexplained deformations: Have a look at the whole mesh in the post processor. In particular look at the intersection of boundaries. At edges where a symmetry boundary and a wall meet you have the wall with a stationary motion and the symmetry with an unspecified motion. Make sure that the solver choose the stationary motion and not the symmetry motion in these cases. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

July 1, 2018, 18:03		#60
louisdub11 Member Louis Dubail Join Date: May 2018 Location: Brussels, Belgium Posts: 52 Rep Power: 7	Hello again, As I couldn't have access to the customer portal for dynamic meshes, my objectives have changed: I need now to do the simulation of a journal bearing. My first question is: when doing the mesh, should I generate a mesh for all the different parts, or should I generate a mesh only for the fluid region? Thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
decomposePar problem: Cell 0contains face labels out of range	vaina74	OpenFOAM Pre-Processing	37	July 20, 2020 05:38
[mesh manipulation] Importing Multiple Meshes	thomasnwalshiii	OpenFOAM Meshing & Mesh Conversion	18	December 19, 2015 18:57
radial turbine blade simulation with dynamic mesh 6dof(fluent)	mamyjooooon	FLUENT	0	April 7, 2011 14:28
[snappyHexMesh] snappyHexMesh won't work - zeros everywhere!	sc298	OpenFOAM Meshing & Mesh Conversion	2	March 27, 2011 21:11
Icemcfd 11: Loss of mesh from surface mesh option?	Joe	CFX	2	March 26, 2007 18:10