Negative volume in mesh due to contact

A.Jalal · September 9, 2015, 15:28

I am simulating a ball inside a valve that its motion is determined by a UDF.

When the ball gets close to the wall I get errors about negative volumes. I tried refining the mesh, decreasing the time step, improving Flow Courant Number and relaxation factors. Also, I added the contact detection specifying the boundaries but I still get the same error.

Do I need to compile a UDF for the contact detection?

If so, any advice or push in the right direction would be appreciated.

Thanks

rsaurabh · April 4, 2017, 12:50

Hi Ahmad,
I hope you have found solution for your problem. i am new to contact detection. as i can see in contact detection you have to specify zones where possibly can happen.

in my case there are two rigid body which are rotating about their own hinge point and in opposite direction. and i need to restrict their motion between 0-55 degree. so would you guide me how can i make this setting and how to write udf for this case.

Regards
Saurabh

A.Jalal · April 4, 2017, 18:00

Saurbah,

I actually switched to Star CCM+, which made my life much simpler when it came to contacts and rigid body motion.

I am not sure about the UDF to restrict the angle. I am sure it is a simple code but I am not an expert in it.

Good luck

rsaurabh · April 4, 2017, 18:29

Thanks for reply, I will be thankful if you can provide whatever information you have regarding contact detection. Because I am not getting any.
Is it possible to do 6DOF simulation in star CCM+.

A.Jalal · April 4, 2017, 19:24

Yes and it is very simple. There are multiple modules to show how to set it up if you do a quick search.

Mainly through DFBI (Dynamic Free Body Motion) and Overset Mesh.

christoph45 · December 15, 2017, 13:30

Did anybody make some progress on that topic? I trying to simulate a valve. See this thread. And I will come across contacting faces.

I can see that star ccm can do contacting solids in this video.
I thought Fluent is THE most advanced simulation environment, so there should be way...

A.Jalal · December 20, 2017, 15:05

christoph45

I think the newer versions of ANSYS (17 and above) added overset meshing techniques. I have not tested or seen examples of it, but you can try for yourself.

When I was simulating the problem initially in ANSYS, I was advised to create a porous region where contacts would occur rather than simulate actual contact. The porous region will allow fluid to pass but create a reasonably thick "film" that resembles the actual contact of the valve with the valve seat. My advisors suggested a region of 3 - 5 cells. I have not tried this as it seemed too cumbersome at the moment.

I had access to Star CCM+ at the time and followed the example you provided as a reference.

Regards,

Christophe · December 20, 2017, 20:15

CFX user, so apologies if this isn't applicable to Fluent. I think it has problems when the mesh is forced to be zero height when the two solid pieces contact. There is no element birth/death in your case, just morphing. So yes the porous media idea sounds good, or similarly you can put in a really thin "Rigid/Solid Body" where the pinch region occurs. The solid body has an independent mesh vs. the fluid and simply overlays on top of the fluid mesh and imparts a velocity reducing source term on top of the elements it overlaps. This allows the mesh morphing to occur within the solid-fluid overlap region which helps prevent the mesh collapsing on itself and getting a negative volume.

April 4, 2017, 18:00		#3
A.Jalal New Member Ahmed Jalal Join Date: Mar 2015 Posts: 22 Rep Power: 11	Saurbah, I actually switched to Star CCM+, which made my life much simpler when it came to contacts and rigid body motion. I am not sure about the UDF to restrict the angle. I am sure it is a simple code but I am not an expert in it. Good luck Master312 likes this.

April 4, 2017, 19:24		#5
A.Jalal New Member Ahmed Jalal Join Date: Mar 2015 Posts: 22 Rep Power: 11	Yes and it is very simple. There are multiple modules to show how to set it up if you do a quick search. Mainly through DFBI (Dynamic Free Body Motion) and Overset Mesh. Master312 likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[blockMesh] Mixer mesh - negative volume problem	jadtwo	OpenFOAM Meshing & Mesh Conversion	2	November 6, 2014 16:37
Negative Volume during Mesh Motion Analysis	giov_ingr	FLUENT	2	December 13, 2013 06:09
Negative volume in moving mesh	mvee	FLUENT	5	September 30, 2011 12:56
negative volume found in es-ice mesh	lizhihua	Siemens	1	August 4, 2007 04:39
Icemcfd 11: Loss of mesh from surface mesh option?	Joe	CFX	2	March 26, 2007 18:10

April 4, 2017, 12:50		#2
rsaurabh Member saurabh kumar gupta Join Date: Jul 2016 Location: kanpur,india Posts: 53 Rep Power: 9	Hi Ahmad, I hope you have found solution for your problem. i am new to contact detection. as i can see in contact detection you have to specify zones where possibly can happen. in my case there are two rigid body which are rotating about their own hinge point and in opposite direction. and i need to restrict their motion between 0-55 degree. so would you guide me how can i make this setting and how to write udf for this case. Regards Saurabh

April 4, 2017, 18:29		#4
rsaurabh Member saurabh kumar gupta Join Date: Jul 2016 Location: kanpur,india Posts: 53 Rep Power: 9	Thanks for reply, I will be thankful if you can provide whatever information you have regarding contact detection. Because I am not getting any. Is it possible to do 6DOF simulation in star CCM+.

December 15, 2017, 13:30		#6
christoph45 New Member Join Date: Apr 2017 Posts: 10 Rep Power: 9	Did anybody make some progress on that topic? I trying to simulate a valve. See this thread. And I will come across contacting faces. I can see that star ccm can do contacting solids in this video. I thought Fluent is THE most advanced simulation environment, so there should be way...

December 20, 2017, 15:05		#7
A.Jalal New Member Ahmed Jalal Join Date: Mar 2015 Posts: 22 Rep Power: 11	christoph45 I think the newer versions of ANSYS (17 and above) added overset meshing techniques. I have not tested or seen examples of it, but you can try for yourself. When I was simulating the problem initially in ANSYS, I was advised to create a porous region where contacts would occur rather than simulate actual contact. The porous region will allow fluid to pass but create a reasonably thick "film" that resembles the actual contact of the valve with the valve seat. My advisors suggested a region of 3 - 5 cells. I have not tried this as it seemed too cumbersome at the moment. I had access to Star CCM+ at the time and followed the example you provided as a reference. Regards,

December 20, 2017, 20:15		#8
Christophe Member Join Date: Jan 2015 Posts: 62 Rep Power: 11	CFX user, so apologies if this isn't applicable to Fluent. I think it has problems when the mesh is forced to be zero height when the two solid pieces contact. There is no element birth/death in your case, just morphing. So yes the porous media idea sounds good, or similarly you can put in a really thin "Rigid/Solid Body" where the pinch region occurs. The solid body has an independent mesh vs. the fluid and simply overlays on top of the fluid mesh and imparts a velocity reducing source term on top of the elements it overlaps. This allows the mesh morphing to occur within the solid-fluid overlap region which helps prevent the mesh collapsing on itself and getting a negative volume.