What would be a good way to get an accurate collision model for my attached geometry? I want the inside to be complete, and I don't want holes anywhere except at the top where one should be.

I've tried Meta Balls, Ray Intersect, both with no Laser Scan, and Triangulated Surface, all with high divisions or LOD's, yet to no avail. It is usually okay, until you get about halfway up, and then holes and slits start appearing.

The attached geometry has been subdivided, and is available at a lower detail level if needed.

You could scatter points on the surface, and then use the particle fluid surfacer to generate a new surface. Read this new surface into DOPs to generate a usable collision geometry, then have your actual geometry read in the .sim data.

You're going to have a hard time getting a more accurate surface out of your model because it's one giant piece of incoherent intersecting geometry. If you need more accuracy that what the attached file gives you, I suspect you would need to re-model.

Okay. Thanks. I guess my lazy modeling has come back to bite me in the butt; it's not the first time.

I wasn't able to look at the attached file, but I have now, and good job. That looks like it'll work, only way to find out is to run the simulation. Thanks.