Different meshing different results ???

[Lucherino]

Good morning,

I am tring to meshing the attached geometry using Abaqus. Originally the part come from Inventor. I receive it as a .sat file. It is a filter, loaded only by mechanical loads on the nozzle.

I am speaking of tridimensional solid geometry. If I use tets no problem, it will be automatically mesh and the solver show me around 150 Mpa.

If I wont to use brick it is more complicated. Before I have to split geometry and after apply different size on each zone. I mesh it ( but model become very heavy ) but the risult is completely different 650 Mpa.

I think this solution it is wrong because displacement are about 150 mm and this is impossible.

The main point now is : I wont use brick because they are better to solve stress problems but in this case it appear not correct / no good. I don't understand where I do mistake.

Any suggestion?

Thanks

 

[corus]

You can't tell from the picture what your results are showing as all you can see is the mesh. Show the results from both models for comparison, but remove the mesh first. It may be that the maximum stress is just due to a discontinuity in the loading or restraints, but who can tell.

 

[Mustaine3]

Maybe it's hourglassing or shear locking. Try again with parabolic hex elements.

 

[gonzo42]

Since you are splitting geometries and applying brick elements separately, discontinuity issues would arise. To start off, I would seed the entire part with an approximate size (selecting all the geometries) and gain some insight. Later you should be able to refine your meshing schemes. It is unclear in your picture, to pick out the issue specifically.

Also, different meshing schemes will definitely give you different results. Be careful with tets. Unless you run convergence tests, you can never be sure of your result.