Composites shells

Sture

Hi,

Having more comprehensive laminate analyses possible would be nice.
Say:
-More than one material in a laminate.
-Sandwich formulation.
-More than Tsai-Wu failure criteria. (Tsai-Hill, max strain, max stress, Hashin, Puck, etc)
-Linear buckling of composite shells (native solver)
-Materials and laminates (of those materials) defined separately.
-Stresses and strains in the layer local coordinate system
-And more

Victor

Thanks Sture. I've added a few of these to Github issues https://github.com/victorkemp/Mecway/issues

I'm a bit hesitant on failure criteria, in part, because you should be able to do write them as formulas in the solution.

You can already have more than one material in a laminate.

What do you mean by sandwich formulation? Is that something different from a 3 ordinary laminate?

Sture

Hi,

This is the laminate definition and one can clearly have different material properties for the different layers. However, how do you define failure criteria for each material? Is it necessary to write a separate formula for each layer?

Victor

Oh I see. No, the failure criterion is only for all of them, so I suppose it assumes they're all the same material. Good point.

Ropa

Hi Victor,

It would be great if the points highlighted by Sture could be implemented in the near future—it would make calculating composite structures in Mecway much more enjoyable! 🙂

If I may share my two cents on the topic: in the past, I was able to derive the stresses in the layer directions (s11, s22, t12) using formulas. However, I got stuck when defining the failure criteria, as I wasn’t able to differentiate between compressive and tensile states using formulas. I assume some kind of “if” check should be implemented—am I correct?

Below is a screenshot of the max stress/strain criteria, which might help clarify what I mean.

Victor

Understood.

For now, you can use the heaviside() function in a formula like an if statement:

if x > 0:
 y
elif x < 0:
 z

would be:

heaviside(x) * y + heaviside(-x) * z

Sture

Hi,

If I calculate Findex, can I plot the results in the postprocessor?

/Sture

Ropa

@Victor, thanks for the hint on the formula, I did a quick test and it works

@Sture, I assume you are referring to a FI colour plot, then the answer is yes