Thermal stress

gemada

Hi.
I hope you are all fine.
I'm trying to calculate thermal stresses in a concrete beam fixed at both sides, caused by heating (120 minutes of fire duration). After thermal transient analisys I run the static analisys (loads: transfer temperatures from solution and new thermal stress with reference temperature of 23C). No thermal stresses are generated...this should happen if the beam was allowed to along with heating but if it's double fixed there should be thermal stresses, right?
Best regards

Victor

That sounds like it should work. Is there any strain or displacement or is it completely ignoring the thermal load? Perhaps post the model.

If it's completely ignored, some things to check:

• That the material has a thermal expansion coefficient.
  
• That the temperatures really were transferred from the solution - select the node temperature item it created and make sure the numbers shown in the graphics area are reasonable. Beware that for transient analysis, it takes temperature from the current time step, which is the initial one by default, where they might be all at the initial value.
  

Stoyan

I do a fair bit of thermal work with Mecway, if you upload a test case I'd be happy to take a look for you

gemada

Thanks a lot for the replies.
Victor, I've checked everything but it doesn't generate thermal stresses.
Stoyan, I'm uploading the simple test case (thermal and static files). Thanks a lot.

Victor

Add stress to the solution with Solution -> New stress and strain -> stress.

gemada

Thanks a lot Victor. That worked!

gemada

I'm attaching another example: simply supported beam exposed to ISO 834 fire only at the bottom surface: 1 file with thermal transient analisys for conduction, convection and/or radiation, 1 file with thermal stresses just for heat conduction and 1 file with temperature comparison. In the file with temperature comparison case A is convection and radiation and case B is only heat conduction and (usually this is the combination used in scientific papers).

My doubts above hehehe:
1) Temperature values in case A are strange because of the mesh and transient analysis parameters? (I believe they should be more refined).
2) I'm not sure about the stress values...they seem to high. Are these the expected results?
3) Is there any thermal validation cases in here for beams or slabs or columns (transient temperatures x thermal stresses)?

Victor

1) Yes, those wiggles look like a typical result of time steps or elements being too big.

2) What makes them seem high? Check for mesh convergence. That might go very slowly with linear elements. Quadratic ones are much better on thermal stress problems with non-constant temperature fields.

3) I don't know, sorry. You could get some confirmation by using the CCX solver too.

gemada

Thanks a lot once again Victor.
I must learn a bit more about mesh convergence.
I'll try all your tips.