Dynamo workflow to recreate complex geometry from imported sat file

Hi there, I’m working on a project to create some furniture families for a manufacturer. My target is to have a good balance between file size and model geometry so that these families can be used to output nice 3d renderings as well.

So far, I’ve been able to receive the STEP files, clean them up, and export them as SAT so I can model the Revit components (that needs to be on Revit 2018)

But now I need to model some caster wheels like the one shown in the screenshot attached. It has quite a complex organic geometry, so I believe I need to create it in the adaptive component environment.

My idea is to break the caster geometry into small parts like (‘tires’, wheels, brake pedal, and so on).

I saw a workflow where a guy recreates complex geometries on Revit, manually (!), by defining a set of ref planes, tracing intersecting splines on each ref plane, and then creating a form using those curves.

It provides a nice result and a low file size, but it’s extremely time-consuming. So I want to use Dynamo to speed-up things and have, at least, the splines drawn for me: selecting the geometry and the ref planes, setting an intersection, and then drawing the splines back on Revit.

How to draw these splines on Revit?

And other than that, do you know another workflow to get this ‘translation’ of imported SAT geometry done?

Thanks in advance!

if you wanted it to be “revit native geometry” you will need to recreate it yourself. But before you do that you have to ask yourself “do I really want it to be native revit geometry”? Could I just settle for it to be an SAT within a family"? If you still want it to be native revit geometry you will need to make the casters out of an adaptive component or Mass family. I would recommend you place reference planes in a radial patter then use intersect to get the those profiles then use loft along the profiles. You will not be able to use all of the profiles so you could just create half of the caster then the other half (0 to 180 then 180 to 360) These are just suggestions See image

Thanks Marcello, for taking the time to reply it. Yeah, I did this reasoning before and I had decided to go for using native geometry cause the sat file size was simply too huge that would impact the performance of anyone using that family.

I modeled it as an adaptive component but even so, I was getting more than 2,5MB(!) just for the ‘tires’: so I dropped the idea and simplified a bit the geometry to get a decent result without compromising the overall family size.