This union fails in Rh7-8-9
FaultyBoolean.3dm (1.6 MB)
Rhino doesn’t find a complete intersection - if you run Intersect the curve is open in one spot. Edit - open in two spots actually… That (rail revolved?) surface is pretty funky…
I thought I could maybe make a better top surface by splitting the revolve at an isocurve and removing the top part, then using some extracted isocurves from the revolve and the new Patch tool, fill in the surface G2 at the edges. But I was unable to get a good result, however maybe I just don’t know how to set the parameters to use the tool correctly.
FaultyBoolean-PatchFail.3dm (890.1 KB)
@menno - is there a way to make this work with Patch?
Surface is from a degree 3 curves lofted loose ending with a point. Both are considered closed polysurfaces.
FaultyBooleanLoftSurface.3dm (848.1 KB)
Here are the curves 
… provides a very smooth surface.
Yeah, that’s not the issue I was trying to check, it was seeing if avoiding having a singularity at the top near the intersection could make the result work.
That was my concern but the distance seemed to me sufficient to avoid the singularity, I tried with different tolerance settings, both with more loose settings and more stringent settings, without any success.
There is a problem with one portion of the lower white surface between two knot curves. The red curve is the result of Intersect of the white and black polysurfaces.
But the problem may not be with that surface, or at least not with only that surface. I created a bounding box of the black surface and it intersects properly.
Mystery
That file does not contain any curves, only polysurfaces. Can you provide the curves?
FaultyBooleanLoftCurves.3dm (853.3 KB)
Curves for loft loose surface.
Thanks for the report. I opened RH-94079.
Patch (in Rhino WIP) can be used, but the guide curves make the result somewhat ‘cushion-like’, with slight protrusions in the four quadrants.
FaultyBoolean-PatchFail_MDvR.3dm (360.7 KB)