Free BC with body force problem

[MaddogMcCree]

Hey guys (and girls ,

i d like to simulate thermal stresses in a hybrid material consisting of several different materials. Static Analysis is all fine and yields nice results.
Now for a thermal analysis i d like do some research on the thermal expansion and on the resulting stresses.

To check out whether everything would work i set up a generic example. A cube with a body force of 100K in each direction.
CTE is isotropic and 14*10-6. The problem is that with any kind of boundary condition, the Temperature results in unreasonable high stresses (where there should be none).
Normally a 3,2,1 - 3,2 - 3 bearing should do the job.

Please help me on this. I don t have a clue of what could be wrong. Nevertheless it has to be some silly fault.

Best regards,
Maddog

 

[BAretired]

What program are you using?

BA

 

[MaddogMcCree]

Hah, sorry.
I m using ABAQUS 6.8-4.

 

[Elabbasi]

If CTE is the same for all materials and you have a cube with symmetry boundary conditions on 3 perpendicular sides and free on the other 3 you should get zero stresses everywhere in a static analysis (assuming also the applied temperature is the same everywhere on the cube).

Nagi Elabbasi
Veryst Engineering

 

[rb1957]

i don't know abaqus, but could it be that the proram thinks you're heating one side of the material ?

 

[BAretired]

What about the 100k force in each direction? That ought to produce a few stresses.

BA

 

[paddingtongreen]

If you are heating a piece of material that is between a rock and a hard place, you are going to see some significant forces.

Does/do the restraint(s) on a face prevent perpendicular movement?

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[MaddogMcCree]

The 100K should lead to a volumetric expansion, i.e. linear expansion in all dimensions. And the boundary conditions
are chosen (in my opinion) in a way that free expansion without stresses should be possible (given a homogeneous material).

 

[hokie66]

You may have thrown some by describing the temperature change as "a body force of 100K". 100 degrees Kelvin is not thought of as a force by folks in this forum, but 100k is 100,000 lbf. Anyway, I agree that free expansion should not create stress.

 

[rb1957]

i think the BCs are set up right, to minimally constrain the panel ... as described they'll take out the six degrees of rigid body motion (correct) or not (incorrect, and the model won't run).

when you say you've done the statics and they work out well, was that with a model or hand calc ? (i'm uessing it the model). temperature loadings are a funny thing ... the temp only creates strains and it's only when the strains are prevented that loads build up. i think too that they're quite code dependent ... try running an abaqus example or call their helpdesk.

 

[MaddogMcCree]

Thermal Simulations are always a funny tihng, you are right.
I need to get my hands on some examples.
I thought that a simple generic example would do it.

And 100 Kelvin are not a force, that s right. I don t know why ABAQUS calls it "body force".
Thanks for your help.

Yet maybe someone already did a similar simulation and would like to supply a Python-file, or an Input-file.

Best regards,
Maddog

 

[BAretired]

If ABAQUS is calling it a body force of 100K in each direction, perhaps they are interpreting it in the same way as I did, namely a body force of 100,000# in each direction.

BA

 

[Elabbasi]

I don't see the naming "body force" applied to Temperatures in Abaqus, and even if this description is used it shouldn't be "in each direction" since it's a scalar. There is a "body heat flux" which is a scalar too.

I am sure you'll find some input files in the Abaqus documentation that can serve as good starting points for you.

One more thing. You described the boundary conditions as "Normally a 3,2,1 - 3,2 - 3 bearing should do the job.". I did not understand that description, so please elaborate on your boundary conditions maybe there is something there that's causing those high stresses.

Nagi Elabbasi
Veryst Engineering

 

[MaddogMcCree]

Never trust a chinese =)
A chinese gave me the advice with the body force.
It turned out to be totally wrong. I now use "predefined field" as load and it works all fine.

Thanks for your help,
this thread can get closed.

Best regards,
Maddog