Creep properties interpolation problem?

Dear Calculix users

i am submitting here a problem that i can not solve properly in Calculix (2.20, Windows).
I am using the Norton Bailey creep law within a VISCO step (no time dependence), after elastic and plastic definitions.
I am trying to reproduce experimental creep data, with the main goal being to correctly predict the time up to a creep strain of 0.01.
The simulation involves a single hexaedral element of edge size 1mm, all BCs are correctly defined, and the reference ‘constant’ stress is imposed via a node force and an equation to link the master node to the slave nodes of the face on which i apply a traction.

When i define the creep parameters of the material at a single temperature, ie the reference temperature set in initial conditions, the results are perfect and as expected.

When i define temperature dependent creep parameters, even if the previously mentionned creep parameters definition point is in the list, the results are totally off. I assume this has something to so with some interpolation or fitting ?

Any idea on that ?

My final goal is to use this Norton law in a part with a temperature map and a complex stress distribution. But because of the inconsistency, i can not.

Thank you in advance for your help !

Regards

Frederic

OK i finally found out that Calculix does not like to interpolate the temperature stress exponent data. Whenever i use a temperature dependent stress exponent, results are off when the temperature map is not uniform. So FYI, it is better to fit you rexperimental data only assuming a temperature dependent creep coefficient, and a best fitted but constant stress exponent.

Sounds like a bug. Perhaps you could report it on Github? Issues · Dhondtguido/CalculiX · GitHub

Thank you for your reply !