SMART Crack Growth Error (SIG$SEGV )

  • Last Post 12 April 2019
xavierBAM posted this 08 April 2019


I´m currently running a simulation where a Double Cantiliver Beam with a pre-crack is being fractured in a Mode I Fracture. I have a pre-meshed crack and I´m using the SMART Crack Growth feature on a 3D Model. 80% of the times I get this error:

An unexpected error ( SIG$SEGV ) has occurred...  ANSYS internal data  
 has been corrupted.  ANSYS is unable to recover and will terminate.    
 Previously saved files are unaffected.  Please send the data leading   
 to this operation to your technical support provider, as this will     
 allow ANSYS, Inc to improve the program

It doesn´t follow any pattern, sometimes it´s at the substep 41 out of 75 or at 14 out 90 (for example). I´ve tried to reduce the number of nodes, the number of contours or play with the number of substeps I´m using. Also, depending on the number of subteps, the length of my cracks changes quite a bit (given the overall length of the beam, which is 110 mm). Here is a graph showing Length VS SIFS. They have all been done with 6 contours except for one; you can see a clear tendency but nothing too conclusive. I´ve done a real life experiment of the case I´m studying and It breaks all the way up until the end. 

Attached Files

Order By: Standard | Newest | Votes
DavidW posted this 08 April 2019


Which version are you in? This may be related to a defect that has been fixed in the current version, 2019 R1.

xavierBAM posted this 09 April 2019

I´m currently running on the 2019 R1 version.

SandeepMedikonda posted this 09 April 2019

What school are you from? Are you able to reach out to your Account Manager/ANSYS Student Coordinator to escalate this?

xavierBAM posted this 11 April 2019

No I´m not, what use would that be? If it´s an internal problem from the software because it´s a new feature, when do you believe that it will be fixed? (when does a new release come out?) or may be it´s a common thing and you know a way to avoid this problem?

SandeepMedikonda posted this 12 April 2019

Segmentation errors can be caused by a multitude of reasons and it would not be possible to debug something like that on the forum.

You can try and see if this is fixed in the next release but it could also be model specific or machine specific (you can try a coworkers machine or a simple model to eliminate these possibilities). My guess is it is something specific to your model in which case, we would have to inspect each object or generate a dmp file and proceed as needed.

Reach out to your Account Manager, ASC or if you are student ask your professor to escalate this directly to ANSYS.