CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > OpenFOAM Running, Solving & CFD

Segmentation Fault (Core dumped)

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   March 27, 2013, 01:38
Default Segmentation Fault (Core dumped)
  #1
Senior Member
 
starter
Join Date: Sep 2012
Posts: 109
Rep Power: 6
sihaqqi is on a distinguished road
Hi

In my case, flow enters a pipe in +Y direction and turns at 90 degrees in +X direction. Flow is turbulent and Reynold's number is 3x10e5. Medium is air. When I type simpleFoam, after 50 iterations, terminal gives message "segmentation fault (core dump)". I have also tried to change value of inlet and outlet in nut from 0 to 0.0001 but to no avail. Please advise how to rectify this error.

Regards
sihaqqi is offline   Reply With Quote

Old   March 27, 2013, 12:24
Default
  #2
Member
 
Eric R
Join Date: Jul 2012
Location: Mississippi State University
Posts: 75
Rep Power: 5
msuaeronautics is on a distinguished road
Did you check to see if there may be any meshing errors?
msuaeronautics is offline   Reply With Quote

Old   March 27, 2013, 18:26
Default
  #3
Senior Member
 
starter
Join Date: Sep 2012
Posts: 109
Rep Power: 6
sihaqqi is on a distinguished road
it is not giving me the type of error. i made my geometry and meshing in ansys. in fluent it was working fine, but this is my MS project and according to my supervisor, i have to work in openfoam. how can i determine the type of error. does ubuntu use any particular software for this or are there any special commands to track the source.

this is my first time with openfoam so i am a bit struggling.

thanks for response.
sihaqqi is offline   Reply With Quote

Old   March 27, 2013, 18:28
Default
  #4
Senior Member
 
starter
Join Date: Sep 2012
Posts: 109
Rep Power: 6
sihaqqi is on a distinguished road
Thanks for the response.

It is not giving me the source of error. I made my geometry and meshing in ansys. In fluent it was working fine, but this is my MS project and according to my supervisor, I have to work in openfoam. How can i determine the type of error. does ubuntu use any particular software for this or are there any special commands to track the source.

This is my first time with openfoam so i am a bit struggling. error image is attached. complete case is at this address

http://www.4shared.com/archive/U267rBPa/90mitretar.html

Regards
Attached Images
File Type: jpg error.jpg (79.7 KB, 76 views)

Last edited by sihaqqi; March 28, 2013 at 01:27.
sihaqqi is offline   Reply With Quote

Old   March 28, 2013, 09:23
Default
  #5
Member
 
Jim Knopf
Join Date: Dec 2010
Posts: 60
Rep Power: 6
JimKnopf is on a distinguished road
Maybe the problem is due to the bounding of epsilon. What BC did you use for the epsilon equation?

Did you already make a checkmesh run? Do you have non-orthogonal or higly skewed cells in your mesh. Fluent has much more "dark magic" in its solvers so they are more robust to bad meshes.

have you already tried to start the simulation from a potentialFoam run?


Greetz,
Jim

Maybe you should reproduce your mesh with BlockMesh, this ensures a better mesh quality

Last edited by JimKnopf; March 28, 2013 at 09:24. Reason: Got another Idea
JimKnopf is offline   Reply With Quote

Old   March 28, 2013, 10:04
Default
  #6
Senior Member
 
starter
Join Date: Sep 2012
Posts: 109
Rep Power: 6
sihaqqi is on a distinguished road
Hello Jim

checkMesh results are attached. blockMesh is valid only if there is a blockMesh dict. I used Ansys DesignModeler for geometry and Meshing. Anyways, checkMesh results are good as in the image.

My complete case is at http://rapidshare.com/files/83104833/90mitre.tar.gz should you want to check it.

I have only used simpleFoam
epsilon value is 862.132 as mean free stream velocity is 75.5m/s

What do you mean by bounding of epsilon. For BC, please refer to my calculation sheet. I am a beginner level openFoam user so I am still learning basic stuff.

Thanks for your response. I appreciate it.

Regards
Attached Images
File Type: jpg checkMesh.jpg (41.5 KB, 36 views)
Attached Files
File Type: xls calc sheet.xls (8.0 KB, 10 views)
sihaqqi is offline   Reply With Quote

Old   March 29, 2013, 19:32
Default
  #7
Member
 
Jim Knopf
Join Date: Dec 2010
Posts: 60
Rep Power: 6
JimKnopf is on a distinguished road
Hi Sihaqqi

The checkMesh looks OK.

But in the screenshot of the log (first post), there is a line, Bounding of epsilon. This might be the cause of the segmentation fault.

If your case is a pipe, then you should give potentialFoam a try for a starting solution. Just run potentialFoam and afterwards simpleFoam.
This might reduce Gradients and there make it more easy for the solver to iterate a steady and stable solution.

I tried download your case but you have denied download permission, so I couldn't. If you grant download access i can have a look at your case.

Greetz
Jim
JimKnopf is offline   Reply With Quote

Old   March 30, 2013, 01:21
Default
  #8
Member
 
Eric R
Join Date: Jul 2012
Location: Mississippi State University
Posts: 75
Rep Power: 5
msuaeronautics is on a distinguished road
sihaqqi,

I wonder if using linearUpwind for epsilon will help this, or any other bounded scheme. I have had a similar issue before and a scheme change always helped in this situation.
msuaeronautics is offline   Reply With Quote

Old   March 30, 2013, 01:54
Default
  #9
Senior Member
 
starter
Join Date: Sep 2012
Posts: 109
Rep Power: 6
sihaqqi is on a distinguished road
msuaeronautics

thanks, i am trying to find where and how to change schemes as I am only a beginner with OpenFoam. I shall go through some documentation in OpenFoam and if I can find a tutorial, I shall run and advise you on the result

Jim, I am also trying to implement your advise from motorbike tutorial and as a result I have made a new omega file and made some modifications in fvsolution file but still I cannot run first potentialFoam and then simpleFoam. In fact it is giving strange errors which means I have not implemented conditions properly. The files I have uploaded on links I have given above does not contain omega and fvschemes changed file. It still contain the same files. Perhaps you may want to have a look at them as I have discovered how to give public access on both links. If you have any tutorial in pdf which advise the whole thing, I can try at my end.

Many thanks to you both for your precious time.

Regards
sihaqqi is offline   Reply With Quote

Old   March 30, 2013, 08:35
Default
  #10
Member
 
Jim Knopf
Join Date: Dec 2010
Posts: 60
Rep Power: 6
JimKnopf is on a distinguished road
Hi there,

i got your case running, although bounding of epsilon is still active. Anyways,
what I did was to initialize with potentialFoam - if you like to call it this way. You need to add a potentialFoamdict to your fvSolutions

Code:
potentialFlow
{
  nNonOrthogonalCorrectors 10;
}
and also a laplacian term to fvSchemes
Code:
laplacian(1,p) Gauss linear corrected;
additionally I reduced relaxation factors for k and epsilon to 0.5 (you find them in the fvSolutions as well)

then it was running on my computer, or is still running

I have another hint on your case, I think you should definitly refine the mesh around the bend. This is a sharp corner and you definitly get separation behind it.

Maybe you can think of a scaled Version of your case, then it wouldn't need so much computational power.

HTH, greetz,
Jim

p.s.: I didn't use the powerlaw BC, just a fixedValue, but I think it should work anyway
JimKnopf is offline   Reply With Quote

Old   March 30, 2013, 20:45
Default
  #11
Senior Member
 
starter
Join Date: Sep 2012
Posts: 109
Rep Power: 6
sihaqqi is on a distinguished road
thanks Jim but it is the powerLawVelocity profile which is creating trouble. I intend to refine mesh considerably once I see that this profile which is 1/7th power Law is working alright as the next step is LES. I shall give you a background of the project.

My project is tee junction and turbulence around it. The reason why my supervisor has advised me to do the elbow is because it is half and has symmetry. Once I will do tee junction, another 4m leg shall be added on the other side and geometry will not be symmetry but full circular pipe which means cells will be 4x what they will be in a refined mesh. You are correct and I shall have to go for very refined mesh but at this stage I need to get this velocity profile working in smaller time as compared to what it would be in a refined mesh. Right now I have 125000 cells. My other refined mesh which I made for same geometry has 850,000 cells but it requires more time to get these errors or results. Once I make a tee with circular pipe it shall be more than 2 million. This is the first phase of my MS by research project. I must get velocity profile working. I shall reduce vertical upstream leg to reduce computations as I can have developed flow in smaller dimensions.
Second phase of the project include frequency measurement of forces and vibrations those forces are creating. Excessive vibrations at tee junctions are a constant issue and sometimes hairline cracks happen.
I have again tried with what you have said and result is still same except for the fact that bounding of epsilon has removed but segmentation failure is there. Thanks for all your support.
sihaqqi is offline   Reply With Quote

Reply

Tags
core dumped, segmentation

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
solving a conduction problem in FLUENT using UDF Avin2407 Fluent UDF and Scheme Programming 1 March 13, 2015 03:02
Superlinear speedup in OpenFOAM 13 msrinath80 OpenFOAM Running, Solving & CFD 18 March 3, 2015 06:36
Segmentation Fault w/ compiled OF 2.2.0 - motorBike example sudo OpenFOAM Running, Solving & CFD 3 April 2, 2013 17:27
segmentation fault when installing OF-2.1.1 on a cluster Rebecca513 OpenFOAM Installation 9 July 31, 2012 15:06
KIVA and Segmentation Fault Felix Main CFD Forum 2 January 18, 2006 02:24


All times are GMT -4. The time now is 10:53.