[Tobi]

Hi all,

can someone tell me why the elasticThermalFoam in the foam-extend project is wrong? I already checked a few things and the following stuff is not correct:

• The T setup is wrong compared to the analytical one
• If we change that, we get the analytical temperature (that is not really difficult to achieve)
• Why we have some pressure force applied in the U file?
• The results from sigma - analytical and numerical are not really equal or even a bit similar

Furthermore do someone know a benchmark case for thermal stresses? I was thinking (even my colleague) to make a beam that is cooling down and fixed on both sides (1D). Analytically I would calculate the stress as:



And compare it to the foam calculation.

Finally, can someone tell me if the tractionDisplacement BC is for free surfaces? So for example, no stress if it could expand free and stress if the expansion would lead to compression (like in the plateHole tutorial) - the exansion around the hole should produce stress due to the fact that the hole is not able to move freely.

Thanks in advance.

[Tobi]

In addition I would like to know why we have 3K instead of K. Some literature is welcomed. As far as I understand, K is only in one direction and hence we have a volume we need 3 times K (x, y and z). Can someone proof this?

[Tobi]

Hi guys,

now I got the point (thanks to Philip Cardiff). First the displacement is based on the thermal stress:



It is clear that in the momentum equation we have the divergence of that term and hence we get the gradient. So the displacement is based on grad(T) but the stress is related on (T-T0).

The tutorial in the foam extend is wrong (Philip confirmed this).

• T_outside = 0
• U_outside and inside are traction free (p = 0)
• E = 200e9
• nu = 0.3
• alpha = 1e-5
• d2dt2Schemes and ddtScheme = steadyState (mentioned by Philip)

Then you should get the same results than the analytical one. About the traction BC, Philip refer to his thesis https://www.researchgate.net/publica...tress_Analysis

[Tobi]

Hi all,

I forgot a very important thing. Philip mentioned that the bug that he realized Nov-14 still exist... see here: http://www.cfd-online.com/Forums/ope...tml#post519288

To get the analytical solution, you need to change that too.