Understanding FEA results

[drodrig]

Hi there,

I'm a casual FEA user with a basic knowledge. The programs I use are Ansys workbench and the FEA module of CatiaV5.

Putting loads and meshing can be achievable, the problem is understanding the results.

We have two aluminium alloy plates. They are 20 mm thick; check the attached pictures. The round holes are the supports (fixed surfaces in the analysis) and then we apply a difference of 45ºC of temperature.
- p18 01 is the small plate, ~125 mm the longest direction
- fa 01 is the big plate, ~750 mm of "diameter"

We want to study if we make a big plate or a combination of small ones. The advantage is the combination of small plates won't suffer much for the thermal shrinking but it needs more supports (two per plate)

The results of the simulation:
- p18 02: stresses in the small plate: 347 MPa which is unacceptable. But this happens on a edge next to the fixed support. Is this realistic? The stresses gradient if very big, I can't tell if this is really true.
- p18 03: here are the deformations

- fa 02: stresses in the big plate. Again here we see stresses concentration in the fixed supports (round holes) and big gradients
- fa 04: detail of the stresses
- fa 03: deformation

What do you think? can one believe these results?

Thank you

http://files.engineering.com/getfile.aspx?folder=4352c78f-facc-4118-bb12-cd64d49150f2&file=fa_01.png

http://files.engineering.com/getfile.aspx?folder=c4fb0b60-a0ad-42cf-a658-99a33535ea80&file=fa_02.png

http://files.engineering.com/getfile.aspx?folder=060aacd7-3fa5-4da6-b8bc-be056b1238d1&file=fa_03.png

http://files.engineering.com/getfile.aspx?folder=30cae623-fa0e-4806-bdc6-cb0944f22f43&file=fa_04.png

http://files.engineering.com/getfil...9-dd71-4796-8b6d-428b8e6f44be&file=p18_01.png

http://files.engineering.com/getfil...e-eb36-4228-9028-92c44c8d08be&file=p18_02.png

http://files.engineering.com/getfil...e-1c1d-484a-bd94-e3f33ad75165&file=p18_03.png

http://files.engineering.com/getfil...4-ba4e-4e84-b131-1b19f63fb133&file=p18_04.png

 

[rb1957]

i have trouble understanding how fa_01 fits with p18; or are they two different plates ?

p18_02 looks odd, for the max stress to be at the interface between your two plates.

p18_03 shows a constraint on the bore of the large holes ... is this meant to be ? how ??

fa_02 shows high stresses in unradiused corners ... radius the corners, else live with the local plasticity ... it might be a problem if the loading is cyclic.

for my money, i think this isn't really the right place for the extensive discussion required to properly answer the questions. FEA is as good as the tool and the person using it ... if either is suspect then the results are usually pretty pictures.

Quando Omni Flunkus Moritati

 

[jhardy1]

A couple of thoughts:

In reality, the supports are almost certainly NOT rigid, and in thermal expansion analysis, the developed stresses are HIGHLY dependent upon the rigidity of the supports. Try "relaxing" the supports just a little (put in stiff springs rather than fixed supports), and you will probably find the stresses drop dramatically.

Non-linear effects may also be very significant in a problem like this - especially "out of plane" deflections. A linear analysis will assume that the in-plane expansion of the plate will generate in-plane (compressive) stresses acting between the supports. However, if the plate deflects just a little out of plane (think of it as "buckling"), the plate can expand by bending without generating such large in-plane stresses - even if the supports ARE in fact truly "rigid".

Hope this heps!

http://julianh72.blogspot.com

 

[drodrig]

Rb1957,
They are two different plates. The idea is making the big one (fa) of many of the small one (p18)
We have two feet (two rods of glass fiber or kevlar holding the plate) in the two round holes. That’s why they are fixed
The temperature from -25ºC to room temperature will change with some frecuency.

Jhardy1,
Nice point. We have these supporting feet which of course will move. I will make a combined simulation with the feet (bonding them to the Al plate)

Thank you all