GLOBAL=NO

disla

Hi,

I'm trying to set up GLOBAL=NO Keyword but is not taking any effect no matter what I try.
I have two different user orientations and Stresses Sxx, Syy, Szz,...show exactly the same result (colour map).

¿Am I missing something.?

Thanks

disla

Hi Victor,

I have check the inp file generated by Mecway and there is no *ORIENTATION Keyword.
The element sets are there ready, but I can’t find where the new local axis are assigned to the sets.

Regards

prop_design

i think you need to use an orthotropic material property as well.

prop_design

i tested your file with an orthotropic material and it seems to do the trick. everything else seems ok.

disla

@Prop_design

OOOh, , it works!!! , Nice. Thank you very much.

So, the orientation can be set for Isotropic, but the keyword is only generated for Ortho Materials.
I can confirm it happens the same for Solids. It works for orthotropic solid but not for Isotropic solid.

I can also confirm that the local axis orientation doesn't follow the mesh when there are large deformations. The Local axis orientation is respected but it keeps the user initial direction.

@Victor, ¿would it be possible to generate the Orientation Keywords also for ISOTROPIC materials? It will allow to check Stresses in the local coordinate system for them.

Thanks again both.

prop_design

that's not good about element axis not updating for nonlinear analysis

disla

I guess that must be a very hard problem to solve.
One option could be pointing to the direction you foresee your model is going to finish or work as usual with Vm or equivalent invariants.
In the pic I have rotated it only AFTER the load has finish. ( In- plane Shear Stress)

Victor

Big explanation - Orientation for shells is tricky because the way of defining it in the GUI is a vector projected onto the element but CCX doesn't (or didn't?) do that projection and allows arbitrary 3D orientation for shell materials. That means even if you apply a single orientation vector to all elements, it can balloon into a separate *ORIENTATION and associated blocks for each element in the general case where they're not parallel. To make things worse, the default orientations are also different which means this would also happen to any shell model, even without any explicitly defined orientations if it behaved so strictly. That potential explosion of *ORIENTATION blocks is why it omits them for isotropic materials where it "shouldn't" be needed.

But I'll look into this again. I don't see a reason not to include it for solid elements either.

disla

I think I catch you. That would explain why I have seen shear stresses with the same value but oposite sign one beside the other while the overall Von Misses was correct. That means the local axis is inverted in that element.
Defining only one axis in a shell is ill conditioned as there are two normal. I think this is possible because shell element can be inverted and still work. Solids do not allow that.
I liked that old “Propagating orientation to neighbors tool ”. It was deprecated for some reason. I remember asking you about that some years ago.

disla

My intuition says "That potential explosion of *ORIENTATION blocks" could be a sudden "Pop/sign reversal" of normals associated with the usage of Euler angles that may lead to known singularities at specific angles.

Victor

Yea that sign reversal on a curve does look like what you say. It should happen when the shell's normal is within 0.1 degrees of the X axis.

The "explosion" I was talking about was an explosion in the number of cards, not the numerical values. Each element might need its own unique orientation definition for CCX even if they're all defined the same way in Mecway.

The propagation of orientations tool was pretty buggy and inconsistent if I remember right.