Kangaroo - output delay

Hi everyone,

For my research I am working on robotic elastic bending of glulam and I would like to use kangaroo to approximate curve geometries for fabrication.

Since I am integrating galapagos to evaluate the output from kangaroo, I would like to know if is there a way to get kangaroo to spit out only the equilibrium state of the out_geometry or to somehow delay the output.

in doing so, galapagos will test the solution that has a shape that is as close as possible to the equilibrium state instead of all the 'false' intermediate states.

Thanks in advance.

Attachments:

bending_output.gh, 24 KB

Replies to This Discussion

Permalink Reply by Anders Holden Deleuran on July 19, 2016 at 4:52am

Use the Zombie solver (available in both Kangaroo 1 and the last two versions of Kangaroo 2).

Permalink Reply by Fabio on July 19, 2016 at 7:16am

Hi Anders. Thanks a lot for the input. It looks like it could work greatly in my case. Although I haven't managed to get an output from the Zombie component. I have been alterating the iteration and kinetic energy values but everytime I change the position of my anchor points, the geometry is not updating.... what am i doing incorrectly?

Attachments:

bending_output_2.gh, 24 KB

Permalink Reply by Anders Holden Deleuran on July 20, 2016 at 2:55am

It's because you are essentially resetting the whole solver system (the zombie solver that is) each time you change any of the inputs. So in your case, moving the anchor points means they are no longer coincident with the vertices on the polyline and thus these vertices are no longer constrained by the anchors. A way around this would be to pull the vertices to the anchor points instead (maybe using a zero length spring with high strength from each polyline vertex to each anchor point). Finally, I would also recommend upgrading to Kangaroo 2 :)

Permalink Reply by Fabio on July 20, 2016 at 7:04am

Hi Anders,

sorry to bother you again but I am not completely grasping what's behind the zombie solver. It works somehow in a counter-intuitive way.

I understood that the anchor pts need to coincide with the polyline vertices (they already do) but I don't understand what do you exactly mean with 'pull the vertices to the anchor points instead'?

Although all the vertices and the anchor pts are connected through springs, why do the anchor pts cease to belong to the input polyline once they are moved? and why the normal kangaroo solver will execute the simulation correctly instead?

Attachments: