About UNCOUPLED TEMPERATURE-DISPLACEMENT

1

Hi.

I’m looking to implement UNCOUPLED TEMPERATURE-DISPLACEMENT, but my specific use case requires a static structural analysis. Does CalculiX support a combined mode for statics and transient heat transfer? I’ve heard that thermo-mechanical coupling is currently limited to dynamics, but I noticed logic for static thermo-mechanical models in the prediction.c source file. If there is no native support for this, what is the best way to implement a static-transient thermal coupling?

thanks

2

This procedure is the right choice if you need one way coupling:

You could also check its (partially) two way coupled equivalent: Mechanical-Thermal analysis - sequence - #2 by Calc_em

I just wonder why they include the ALPHA parameter.

Abaqus, unlike CalculiX, also has dynamic explicit thermomechanical procedure.

3

Thanks for your response. I noticed the post describes uncoupled temperature-displacement as a combination of heat transfer and static mechanical analysis. However, the code sets nmethod=4, which triggers a dynamic calculation. I’m a bit confused by this discrepancy

4

I took a really heavy (80 kg) element and moved it without introducing any thermal effects. I tested 4 steps:

  • static - stress close to zero
  • dynamic - large stress due to inertia
  • uncoupled temp-disp
    • steady-state - stress close to zero
    • transient - large stress due to inertia
  • coupled temp-disp
    • steady-state - stress close to zero
    • transient - large stress due to inertia

So apparently the steady-state/transient setting affects also the mechanical part which is not the case in Abaqus:

In Abaqus/Standard a fully coupled thermal-stress analysis:

  • neglects inertia effects; and

  • can be transient or steady-state.

5

Thank you for your response. What you mean is that when set to steady state, Calculix performs calculations for statics and steady-state temperature fields. However, the steady-state temperature field calculation only provides a final temperature, not a temperature that changes over time, which makes this calculation less meaningful. If I want to achieve calculations for statics and transient temperature fields, would I have to modify the source code?