free shape optimization without result influence

Rahul_P1

Hi guys,

in my size - free shape optimization is something going wrong. I'm optimizing a U-section (shell elements) and i created 2 layers inside this U-section with elm-offset. The first layer is the base layer and the second should be optimized with free shape, shrink only, grid con on the base layer. A size optimization works on the U-sections shell elements. The Layer material got a lower density then the U-section. Objective is to reduce mass.

But instead of strengthen the critical radius and reduce the shell layer size, the optimization moves the free shape layer near zero and doesn't change anything on the shell element size.

If i create a size - free shape optimization with only one layer the optimization runs normal and strengthens the radius.

Does somebody have an idea where my error is coming from?

Rahul_P1

By creating those layers the stresses values in the shell elements were lowered so there has to be some optimization potential to reduce the mass.

The .out data shows: (aren't the exact values but the course is the same)

start size value : 5 mm
Iteration 1 : prop size: 4.90mm
Iteration 2 : prop size: 4.85mm
Iteration 3 : prop size: 4.80mm
Iteration 4 (now the shape optimization shrinks the design layer) -> prop size: 5mm
and end of optimization.

The mvfactor influences the initial shrink value doesn't? If i lower it could it solve my problem?
Maybe the factor is set to high so that the first shrink step ruins my optimization.

Rahul_P1

Tensa,

That is logical, the default value for MV factor is 0.5 did you change this, or is 0.5 too high for your model dimensions?

Rahul_P1

Hi,

yes i changed it now and made it a slice smaller. Now it works. I didn't know that even a small change of the mvfactor could ruin the whole optimization.

Thanks for the help Rahul !!!