Exclude single ply from composite optimization
I have a composite part that I'd like to optimize. The base ply structure is a standard 0°/45°/90°/-45°/0°-stack. I want to optimize this part and reduce the plies in areas where they're not structurally necessary. The first ply should however not be optimized to get a uniform surface of the part, so only the remaining four plies should be optimized. I can't figure out how to tell Optistruct to do that though. Things I tried:
Use an Interface laminate: I defined an interface laminate with two sublaminates. One containing the single non-optimized ply, the other one containing the other four. My idea was to only assign the free size-design variable to the sublaminate with the four plies, but after changing its entity type to STACK, it will only let me select the interface laminate, not any of the sublaminates - which is weird, because the error message I get is '"laminate1" is not of type ply laminate or sublaminate laminate', even though it is defined as one (and the interface laminate isn't?). I also tried around with composite size-design variables (don't have much experience with those though), and while I can select the correct sublaminate there, the optimization throws an error:
*** ERROR 14: Missing STACK # 1 referenced by DCOMP # 2.
When checking out the .fem-File, there indeed is no STACK 1 defined, only a STACK 2 for the interface laminate. So the error message makes sense, but I don't know how to fix it.
Is there another trick to exclude a ply from an optimization? I'm currently using a PCOMPP-property for the part, don't know if that matters.
Answers
-
if this is free-size, you just need to adjust the ply that you don't want to optimize as CONST, and assign a fixed thickness for it. CONST applies on all your plies with an specified angle, so let's say you need the first layer, 0 degree, as fixed. Make this first layer a unique value, such as 0.0001 angle, that wouldn't change your behavior, and assign CONST for this 0.0001 layer. This way it won't be changed and it will be kept constant.
1 -
That's exactly what I was looking for, thank you! Sidenote for everyone else with this question: Setting an unique angle for the constant ply isn't necessary, instead you can switch the BYANG-Keyword to BYPLY and select the ply (or plies) directly.
0