Learning Forum

[lomss]

Hi all,

I tried to simulate anisotropic creep behavior using Hill Yield Criterion + Generalized Garofalo.  I was able to do the simulation under mechanical (displacement) loading, and the result matched the analytical solution. 

However, when I changed the loading condition to thermal loading only (also applied anisotropic CTE to the model), the model showed error messages:
many lines of "Creep algorithm did not converge for element number XXX with material number 1 and Creep model number 8."
and followed by "The unconverged solution (identified as time XXX substep 999999) is output for analysis debug purposes.  Results should not be used for any other purpose".

I tried with smaller time steps but it did not solve the problem.  Later, I set the temperature in each time step to be 20°C which is identical to the initial temperature.  The error messages remained the same.

Does anyone have a similar experience and knows how to solve it?  Any help will be greatly appreciated.

Thank you.

[Ashish Kumar]

Hi,

It might be helpful to post-process the last successfully converged solution to assess what the stress-strain state is when trouble begins. Does the model converge without creep? Are you using temperature-dependent creep coefficients? Are the creep coefficients using consistent units? Sometimes it is also insightful to take this information and predict (via hand calculation) what the expected creep strain will be at the next iteration based on the stress state temperature and time increment. Also, can you plot a force vs. defection curve to assess the stiffness of the structure? Is the referenced element losing all its stiffness? If so what is the cause?

 

Regards,

Ashish Kumar

 

[lomss]

Hi Ashish, Thank you for the reply.  Yes, it converged without creep.  I attached more information about the model below.  I turned on the creep function after step 1.  The convergent issue only occurred when creep is turned on.  What would you check in the next step?

Material properties:

Anisotropic elasticity:

Anisotropic CTE:

Mesh:

Boundary conditions:

Temperature loading:

Time step setting for step 2 and 3:

Error messages:

[lomss]

Update: after I turned off the anisotropic CTE material property, the error messages disappeared. I hope this information may help.

[Ashish Kumar]

Hi,

I do not have any comments. Please wait for other members to comment.

Regards,

Ashish Kumar

[John Doyle]

Maybe it would help to apply the temperature first (with creep effects turned off) and then turn on creep in a subsequent load step and use smaller time increments.  With creep off, can you verify how much creep strain you expect for a given stress state and time increment (by hand calc).  Maybe that would he insightful.

[lomss]

Hi John,

My description might not be clear.  I did not turn on creep in the first step.  I only turn on creep for step 2 and 3.  The temperature loadings I applied are all 20°C at every step which is identical to the ambient temperature.

The simulation worked fine in step 1 but encountered error messages in step 2 and stopped.  I checked the stress and displacement in step 1.  They are virtually zero (less than 1E-10).  Hope this information helps.  Please let me know if you need more information.