Program received signal SIGABRT: Process abort signal.

Numerical methods and mathematical models of Elmer
Post Reply
Gary R
Posts: 162
Joined: 12 Apr 2012, 07:23
Antispam: Yes
Location: Long Beach CA, USA

Program received signal SIGABRT: Process abort signal.

Post by Gary R »

Operating System: Debian GNU/Linux 11
KDE Plasma Version: 5.20.5
KDE Frameworks Version: 5.78.0
Qt Version: 5.15.2
Kernel Version: 5.10.0-13-amd64
OS Type: 64-bit
Processors: 4 × AMD FX(tm)-4350 Quad-Core Processor
Memory: 15.6 GiB of RAM
Graphics Processor: AMD CAICOS

Hi all,
Every time I try to run this program (included in the attached .tar file) it aborts with the SIGABRT signal. Why is this happening? Also, I've never worked with this solver combination before. Any comments on the .sif file would be appreciated.

Gary R.
ArcTest.tar.gz
(637.54 KiB) Downloaded 48 times
Attachments
Screenshot_20220805_130956.png
Screenshot_20220805_130956.png (37.25 KiB) Viewed 728 times
Rich_B
Posts: 421
Joined: 24 Aug 2009, 20:18

Re: Program received signal SIGABRT: Process abort signal.

Post by Rich_B »

Hello,

After downloading the archive and opening the contents, I tried loading the step file in ElmerGUI. It ran for quite a while before I stopped it. Likewise, I tried loading the brep file, it also ran for quite a while before I stopped it. The log files started with these messages:
Intersecting Triangles: trig 61 with 12676!
Intersecting Triangles: trig 62 with 12676!
Intersecting Triangles: trig 62 with 12681!
.... and many more like this.
Then I loaded the msh file, it loaded right up, no problem. I hid the sphere and zoomed in to one of the electrodes, see attached screenshot.
electrode-mesh.png
electrode-mesh.png (118.05 KiB) Viewed 719 times
I'm guessing this may be contributing to the fault.

Rich.
Gary R
Posts: 162
Joined: 12 Apr 2012, 07:23
Antispam: Yes
Location: Long Beach CA, USA

Re: Program received signal SIGABRT: Process abort signal.

Post by Gary R »

Hi Rich,
Thanks for the reply. This is strange. My files show no notch in the electrodes as shown below.

Also the top of the solver log shows just the standard header information. I am using Elmer version 9.0 built from scratch. Is there a difference between our versions? I think I am going to download the latest version an recompile. Maybe that will solve my problem.

Gary R.
[attachment=0]RightElectrode.png[/att/home/gary/FreecadFiles/Propulsion/ArcTest/RightElectrode.pngachment]/home/gary/FreecadFiles/Propulsion/ArcTest/RightElectrode.png
Attachments
RightElectrode.png
RightElectrode.png (9.43 KiB) Viewed 713 times
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Program received signal SIGABRT: Process abort signal.

Post by kevinarden »

I have the same experience as Rich when loading the mesh. Also it doesn't appear that the electrodes share any nodes with the sphere.
Rich_B
Posts: 421
Joined: 24 Aug 2009, 20:18

Re: Program received signal SIGABRT: Process abort signal.

Post by Rich_B »

Hello,

The gmsh generated msh file is the one that was in the archive in this thread. I also loaded the same msh file in gmsh, and saw the same distortion, as shown in this screenshot.
electrodes-gmsh.png
electrodes-gmsh.png (28.68 KiB) Viewed 706 times
In some cases reducing the maximum element size will help with distortion.

Rich.
Gary R
Posts: 162
Joined: 12 Apr 2012, 07:23
Antispam: Yes
Location: Long Beach CA, USA

Re: Program received signal SIGABRT: Process abort signal.

Post by Gary R »

I really appreciate the participation on this post but I can not duplicate the notch result that you all are getting. I'm exporting the freecad file to a .brep file, loading into gmsh, meshing in 3D and splitting mesh. I have tried it both with and without negen. I then run the .msh file through "ElmerGrid 14 2 ... -autoclean" to produce the Elmer mesh set. Is this different than you are doing?

It just occurred to me that there may be a difference in the way the electrodes are being produced. I use part design and rotate my sketch to produce a single entity. The electrodes in the pictures seem to be made from parts fused together. I have no idea if that is true or if it would make any difference. Just a thought.

Another thought. When I first tried this I had the same notch problem. I cured it by increasing the length to redius ratio. If the ratio gets too great, the code in the solver runs out of adequate precision. Increasing the radius may solve the notch problem. This still doesn't solve the error problem.

Gary R.
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Program received signal SIGABRT: Process abort signal.

Post by kevinarden »

The mesh in my pictures and Rich's just comes from loading the msh file posted. When I load the brep into gmsh and mesh then I get the mesh you are showing. If I remove the potential setting in the boundary condition the error goes away, but the solution is zero.
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Program received signal SIGABRT: Process abort signal.

Post by kevinarden »

I used Tutorial 12 to do just Electrostatics and obtained a solution where the potential was solve as being 1.0
I add the potential = 1000 BC, and the potential solved for was still 1.0.
I added the stat current solver and received the error again.
I remove the potential BC and it solved but the potential was 1.0.

The error only occurs if the electrostatic is coupled with statcurrent solver and there is a potential BC.
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Program received signal SIGABRT: Process abort signal.

Post by kevinarden »

I changed the variable in the electrostatic solver to epotential and it solved successfully. In the results I obtained epotential as 1.0 and potential as 1000.0. It seems both solvers are trying to solve for potential and are conflicting each other. I think usually you would solve for potential in one solver and pass that variable to the other solver if it was needed.
mcase.sif
(3.43 KiB) Downloaded 36 times
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Program received signal SIGABRT: Process abort signal.

Post by kevinarden »

I built my own mesh in Salome. embedded the electrodes completely in the sphere with shared boundaries. Built sif starting from tutorial then added statecurrent solver. A solution was obtained with the 1000 potential BC.
ArcTest.zip
(375.95 KiB) Downloaded 46 times
Post Reply