“1. Solution exception: Could not find node at (2.20210841232375 / -0.828296184285726 / 0) where support number 2 is attached.”
The beams are created from Grasshopper curves using LineToBeam and the model assembles correctly without supports. The support points are taken from curve end points filtered by Z-value (ground supports). Visually, the support point coincides with a beam end, but Karamba still reports that no node exists at that coordinate. If I clean duplicate support points using tolerance-based methods, the error appears; if I don’t, Karamba warns about duplicate supports. Snapping supports to nodes via Disassemble Model causes recursive data stream errors. It seems Karamba requires an exact numerical match between support points and beam nodes, and very small coordinate changes break this link. Has anyone faced this or knows a reliable way to define supports in such cases?
the line to beam and assemble model components have tolerances which snap points within a certain threshold (0.05 as a default). Therefore I suggest you use the Point output of the Line To Beam component and snap your selected support points to them before inputting them into the Support Pos/Ind input.
Thanks for the suggestion. I tried snapping the support points to the Point output of the LineToBeam component exactly as you described and then used those points as input for Support Pos/Ind. Unfortunately, the issue persists and Assemble Model still throws the same error:
“Could not find node at (…) where support is attached.”
I am attaching a screenshot of the script showing how the support points are generated and connected. It seems that even when using the LineToBeam Point output, Karamba still fails to associate some supports with valid nodes in this highly connected parametric structure.
If you do, you would see you have a DataTree of lines going into your remove duplicate lines.
This only removes duplicates per branch of DataTree.
I believe a quick fix is to flatten the curves.
But you are losing any underlying structure to your data. You might not be explicitly controlling that as of now (e.g. edge curves, grid curves, etc.). However for now, flattening the input to remove duplicate lines gets rid of a lot of duplicates and allows the model to run.
