J-Integral

General discussion about Elmer
DePurpereWolf
Posts: 22
Joined: 13 Mar 2013, 23:46
Antispam: Yes

J-Integral

Post by DePurpereWolf »

Hello,
To understand fracture mechanics, a j-integral can be used which would eliminate the need to make a very fine mesh near a crack initiation site.

Does anyone have any experience in implementing the J-integral in Elmer?

If you have any advice. It is much appreciated if you could tell us how you implemented it.
saxonfahrer
Posts: 38
Joined: 27 Jan 2011, 13:47

Re: J-Integral

Post by saxonfahrer »

I'm also very interessted on fatigue prediction (and so on J-integral implementation in Elmer).
Is there any Progress or new tipps/hints on this topic how to solve in elmer (with programmers beginner Level)?
raback
Site Admin
Posts: 4868
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: J-Integral

Post by raback »

Hi

Don't know anything about the J-integral but I found this in my repository. Probably based on some earlier question where I did some minor coding. Maybe you can make some sense on it.

-Pete
Attachments
Jintegral.tar.gz
J-integral example
(13.47 KiB) Downloaded 678 times
DePurpereWolf
Posts: 22
Joined: 13 Mar 2013, 23:46
Antispam: Yes

Re: J-Integral

Post by DePurpereWolf »

I've tried out the code but it's giving me a segmentation fault.
The code runs and even outputs something interesting. I've included the elmer output.

I doesn't give me any hint whay might be wrong except for: Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Regards,
Attachments
output.txt
output.elmer
(18.48 KiB) Downloaded 430 times
raback
Site Admin
Posts: 4868
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: J-Integral

Post by raback »

Hi

Sorry, cannot replicate your problem. For me the last few lines are:

Code: Select all

Jintegral: Jintegral:    4.872E+01
Jintegral: Line length:    2.800E+00
ElmerSolver: *** Elmer Solver: ALL DONE ***
ElmerSolver: The end
SOLVER TOTAL TIME(CPU,REAL):         0.30        0.36
ELMER SOLVER FINISHED AT: 2014/09/15 17:43:35
and the case is run by

Code: Select all

>elmerf90 -o Jintegral.so Jintegral.f90
>ElmerSolver
-Peter
DePurpereWolf
Posts: 22
Joined: 13 Mar 2013, 23:46
Antispam: Yes

Re: J-Integral

Post by DePurpereWolf »

I think this is related to the fact that I'm running Elmer on Ubuntu 14.04

When I run elmerf90 -o Jintegral.so Jintegral.f90
I get:

Code: Select all

gfortran -I/build/buildd/elmerfem-6.1.0.svn.5396.dfsg2/debian/tmp/usr/include -fPIC -m64 -fPIC -fPIC -I. -Ibinio -I/usr/include/mpi -I/usr/include -m64 -fPIC -I/usr/share/elmersolver/include -shared -o Jintegral.so Jintegral.f90 
Warning: Nonexistent include directory "/build/buildd/elmerfem-6.1.0.svn.5396.dfsg2/debian/tmp/usr/include"
Warning: Nonexistent include directory "binio"
Any input on this?
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: J-Integral

Post by mzenker »

Hi,

just an observation: Your Elmer seems rather old. Maybe an update would improve things...

Matthias
DePurpereWolf
Posts: 22
Joined: 13 Mar 2013, 23:46
Antispam: Yes

Re: J-Integral

Post by DePurpereWolf »

The simulation works on a 13.10 machine I have in the office. So it seems to be specific to my ubuntu 14.04 install.

I'm a little apprehensive of building Elmer 7 on ubuntu 14.04 I have seen a lot of issues with it here on this forum.

viewtopic.php?f=3&t=3585&p=12184&hilit=compile+elmer

Anyway, thanks for the J-Integral code, it is very insightful. We may look into this in the future to expand it to 3 dimensions. Currently we don't have the resources for that. But it looks vey interesting.
raback
Site Admin
Posts: 4868
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: J-Integral

Post by raback »

Hi

Generalization into 3D should be pretty straight-forward. Few minutes of coding if you know what to do - and of course indefinite time if coding with Elmer is new ;-)

-Peter
saxonfahrer
Posts: 38
Joined: 27 Jan 2011, 13:47

Re: J-Integral

Post by saxonfahrer »

Hi Peter, Hi DePupereWolf,
thank you for the f90 code. I'm using Windows and i tried to make it useable with ElmerGUI also.
First i compiled the Jintegral.f90 files to dll.
Additionally (for the GUI): I tried to edit a xml-file to the J-Integral. In principle it works, but i struggle with the "logical true" checkbox in the Boundary Condition Section. It only become "true" in the SIF at the end.
What do i have to do to make it from "true" to "logical true"? If this would work, it is very cheap but it would work in ElmerGUI also.

Code: Select all

<?xml version='1.0' encoding='UTF-8'?>
<!DOCTYPE edf>
<edf version="1.0" >
   <PDE Name="J Integral" >
      <Name>J-Integral</Name>
      <Solver>
         <Parameter Widget="Edit" >
            <Name> Procedure </Name>
            <DefaultValue> "Jintegral" "Jintegral" </DefaultValue>
         </Parameter>
      </Solver>

      <Equation>
      </Equation>

      <InitialCondition>
         
      </InitialCondition>

      <BodyForce>
          
      </BodyForce>

      <Material>
         
      </Material>

     <BoundaryCondition>
         <Parameter Widget="CheckBox">
            <Name> J Integral </Name>
			<Type> Logical </Type>
			<Whatis> If checked then J-Integral will be calculated </Whatis>
        </Parameter>
      </BoundaryCondition>
   </PDE>
</edf>
Attachments
J-Integral.xml
(781 Bytes) Downloaded 415 times
Post Reply