different elements, different results?

[John]

Hello everyone, I wanted to ask for information I have to calculate the pressure before arriving at plastic deformation (yielding point 260mpa) in this piece with the following data:

e = 200 gpa
poison's ratio=0.3
density =7860 kg/m^3
thickness = 10 mm

measurements in mmI created two models one with

solid>quad 8node 82

solidity

but the results obtained are different

in the first case arrive at 251e6 for a load of 2e7 pa

in the second case I arrive at 278e6 for a load of 2e7 pa

So I have doubts:

If I select different elements I get different results, I mean big differences (about 30 mpa)?

If I intensify mesh in critical points, i.e. in holes and fillet, should results change a lot?

thanks to those who will answer me:biggrin:

 

[stef_design]

Hi, I tried to do a test with ansys wb11.
max stress is lower (141 mpa) than yours.
the mesh has been fixed (size elements 2 mm).

try to see if the constraints are correct.
Hello! !

 

[John]

thanks, for the answer :d...
However I forgot to specify that the parts bound to move are the other two holes, so I think the simulation changes a little.
The importance of a good mesh is so important to me that it is fundamental, because if I break the effort, it changes clearly.

 

[cacciatorino]

I created two models one with

solid>quad 8node 82

solidity

but the results obtained are different

in the first case arrive at 251e6 for a load of 2e7 pa

in the second case I arrive at 278e6 for a load of 2e7 pa

So I have doubts:

I once had made a simulation (apposed to check this thing) loading a bending bar, meshata with:

1) linear tetra elements
2) square tetra elements
3) Linear Exhaust Elements
4) square exaggerated elements

the difference of the sigma from 1 to 4 was also 30%. in general linear tetra elements are too rigid and should not be used for structural analysis, better tetra-quad or esa lin or quad. At least that's what my fem expert friend always tells me!

 

[stef_design]

so the holes are only bound by the slots?? as in the image.

 

[cacciatorino]

I have made a simulation with code_aster + salome, I attach the images with the results for linear mesh

Consider that I've put in the two holes, not cylindrical hinges.

 

[cacciatorino]

questo invece per mesh quadratica

 

[matteoso]

I'll give you some advice again. in these cases, in reality it never forms as solid, but as plane or as shell. I suggest you do these 2 tests first and then check the values obtained from solid modeling.

 

[cugior]

I tried to perform the analysis of your meshandolo piece by plane82 elements in tension plane and assigning it the thickness 10 mm.

by infitting mesh I got the following results:
max size element --> voltage of von mises
5 mm-- 251 mpa
2.5 mm --> 269 mpa
1.25 mm --> 276 mpa
0.625 mm...> 279 mpa

I think convergence can be assumed