I have tried tackling this problem in kangaroo but I cannot seem to get anywhere…
I am trying to simulate a manufacturing process whereby a flat plastic component is heated up and pressed into a metal mold by a press with silicon pads. This should cause the silicone pads (which are softer than the plastic component) to wrap around the piece and press it onto a metal mold, thereby imparting into the component a specific type of curvature. You can see a rough diagram below:
Ideally, I’d like to be able to simulate both the silicone pad and the component in order to see how both of them deform… If this is not possible, I’d at least would like to see the component conform to the mold.
If anyone could provide some guidance or a good online tutorial to follow (I haven’t found anything tackling a similar problem) I’d appreciate it a lot!
I think this is a complex bit of finite element analysis that isn’t really well suited to Kangaroo. Especially considering the resolution you’d need to accurately represent the “wrapping” of the silicone around the thin and tightly curved glasses frame. Just doing some napkin maths here, but I reckon you’d need about a 0.5 million points and ~3 million springs per silicone block to get a .5mm resolution at the face and enough depth to distribute the forces. Those sort of numbers make my blood run cold!
Ultimately, this doesn’t look too hard to prototype in the real world, if you simplify the geometry and scope.
I don’t have a 3D scan but the result doesn’t require that high of a precision to the real life result. My complain whilst using flow or splop was that it seems to perform some kind of very strong normal mapping of the mesh/surface UVs, this would cause the side walls of the frames to be way more angled (matching the normals of the mold surface) than how it turns out in reality. I understand the reasoning for why these tools behave in this way, I also understand that in a perfect world the outcome from these tools would mirror the real life result. I am trying to find a solution which gives me enough control to mimic the real world scenario and achieve a similar result.
As I said, simulating the silicon pads is not necessary, I simply wondered if it was possible. I am more focused on the frame deformation at this point.
One issue I had with the silicone pads in the kangaroo simulation is that the sharp edge does not push down the center part of the frame. I think it could work a little better if the silicone pad was one piece.
