Problem regarding convergence of results

[Showkat Ahmad]

I am doing non linear analysis of girders and the same is not converging it is showing error after 37th sub step.
please recommend some remedies.

 

[karachun]

Attach file with problematic model.

 

[klaus.]

there are 1000 possible things that can make a nonlinear calculation not to converging....
I spent many days or months or maybe years of my life searching for such 'errors'

 

[FEA way]

Which software do you use ? Please describe settings of your simulation (details are necessary) and attach pictures, if you can.

As klaus says there are many potential reasons of nonconvergence. Incorrect contact and boundary condition definitions are among the most common.

 

[Showkat Ahmad]

I am Using ANSYS APDL. I am doing non linear analysis and the first buckling mode is used as initial imperfection (using a scaling factor).
The settings are like this:
1) Large deformations on.
2) time at the end of load step 1000
3) automatic time stepping on.
4) Number of substeps 100.
5) Max number of iterations 40.

 

[Showkat Ahmad]

I also did the modal analysis to check whether the boundary conditions are good. the results were consistent. The first mode shape is attached below

 

[FEA way]

So it’s nonlinear buckling analysis. Convergence issues are very common for this type of simulations. Do you use load or displacement control ? The second one is much better for convergence. Please try again to share the picture showing mode shape since it’s not attached to your post. Also plot force vs displacement to examine behavior of the system prior to error (and take a look at contour plot of deformation). Try using stabilization or arc-length method. These may help.

 

[rb1957]

@klaus … "I spent many days or months or maybe years of my life searching for such 'errors' " … time well wasted ?

another day in paradise, or is paradise one day closer ?

 

[Showkat Ahmad]

I use load control. the curve is shown below:

 

[klaus.]

… time well wasted ?

no not at all...I learned lots by searching for errors...
Nowadays when my younger colleagues come to me with problems saying the program is not running i usually find the error very fast