Different result with linear and quadratic mesh

Hello

I´m analyzing a structure with Salome version 3.2.9 and I got a huge difference on the results depending of the mesh used. The meshes were tetrahedron linear and tetrahedron quadratic, but all the others hypotheses are the same. The structure is a conical inverted pyramid, the internal faces are loaded with pressure of 0.15 MPa and the top face is fixed. The max stress found in quadratic mesh was 1.07e+03 MPa and in linear mesh was 65 MPa. I´m quite new in FEM and I don’t know what is the correct one.

1) What should I believe (linear or quadratic)?
2) Why are the results so different?

Thanks in advance.

What's the fineness of your linear mesh ? Try to diminish the mesh size and run a simulation again. I had the same problem and the answers I got were: use hexa or quadratic elements

Here are the files
- drawing: moega.stp
- tetrahedron linear: linear.med and linear.comm
- tetrahedron quadratic: quad.med and quad.comm
if I have to post others files tell me

the pyramid has square of 10650 mm on top and ~550 mm on button with 5075 mm of height. Unit of force is N and pressure is MPa
the material is steel (E = 210e3 MPa and Nu = 0.3)

for both meshes use the hypotheses of average length = 800 mm

I have problem to use a smaller mesh size, because of "lack of memory"

I wrote wrong the max stress, in fact it is 69.2 MPa in linear and in quadratic is 493 MPa

Thanks for the answers

Look at this discussion:

caelinux.com/CMS/index.php?option=com_jo...;id=2529&catid=4

I found the same difference and Joel Cugnoni answered to me to forget to use linear meshes for structural problems.

Hi Alessandro

Thanks for the attencion. I´ve already seen your topic, but in your case the error is 12% and I foud more than 7x in the results of stress.

Anyway should I use only quadratic on structural problems?
If i could get a smaller mesh the results would be more similar?