Nonlinear Resudial Stresses

[Staszkinson]

Hello,
Take a look here

https://dl.dropbox.com/u/3745405/residuals.png

The image shows my problem. I need to set up residual stresses in the flange of my I-shape beam along the element.

1) As You can see, stresses are nonlinear. Is there any way to apply residual stresses only by *INITIAL CONDITIONS command without subroutine SIGINI(need fortran compiler and scripting, that's a little bit complicated).

If not - what do You think about approximating stresses in finite elements like that:

https://dl.dropbox.com/u/3745405/approx.JPG

Is it correct approach?

2)How do I know what stresses I have to apply to get the stress distribution wnich is on the first image?

Thanks!

 

[mecheng1234]

i have the exact same issue that i need resolved, i have a residual stress profile through the thickness of a specimen.

it would be interesting to see what you come up with.

 

[mvp23]

Ok I faced a similar issue a while back where I had to input residual stresses from a static analysis into a fully coupled thermal stress model. I figured that there isn't a way to specify initial residual stresses in Abaqus withouyt using the subroutine. this is what i did

Step1 : Fully coupled thermal stress step
( instead of a static general step) This step is to determine the initial stresses .

Step2 : Fully coupled thermal stress step

This step will read initial stress values from Step1.

Finally in the odb, to verify, the final state of step1 will be equal to the initial state of step2. This is what I did and it worked.

Note :
You dont have to specify a temperature just because you are running a fully coupled thermal stress problem in Step1 as the ultimate goal is to determine stresses. Just make sure u use fully coupled thermal stress elements.
I compared the stresses from a purely static analysis with the stresses resulting from a fully coupled thermal stress analysis (STEP1) and the numbers are within 0.3 % of each other.

I dont know if you guys are running a second step, but I thought it was worth sharing my approach as I had to input residual stresses into my model as well.