Filletproblem (both FilletEdge and FilletSurf)

I’m trying to fillet a airfoilshaped geometry, but it gives some weird errors. FilletEdge gives the following result:

FilletSurf gives back: “FilletSrf failed to create fillets.”
I also tried some stuff to determine what might go wrong and found out that OffsetCrvOnSrf doesn’t work as well.

The problem might be the curve I used to create the airfoil. However I tried some similar curves and they give the same problems.

Thanks in advance :slight_smile:

Fillet_Problem.3dm (197.5 KB)


Remodel your shape in a way that it reduces control points. It might be a good idea to use more than one surface. If your shape is cleaner, filleting will work much better

1 Like

The surface domain is kinds screwy, Extract the surface and run the : Reparameterize command, Use the automatic option,

After that you should be able to get a fillet up to about 25 mm to work.

Hello Mariska
The reason your fillet is failing here is because it’s crossing itself at the tapered end. A radius of 1.8 would be about the limit

Yes, if you don’t want to consume the trailing edge you need to keep it under 2.35 there, but that doesn’t mean the other edges can’t be filleted bigger

Fillet_Problemx.3dm (581.7 KB)

Thanks you all!
@TomTom I used fair on the original curve, but since I will use this shape for CFD simulations, it’s quite important to keep the original shape as much as possible. I checked the curvature as well, that seems to be okay.
@Tone it’s true that it fails at the end as well, but I wanted to trim that part of the fillet away and fix the surface with BlendCrv and NetworkSrf. This approach has worked in the past
@Jim I tried the reparametrize command, but it didn’t work for me. Thanks for your file! The fillet at the trailing edge looks beautiful. Could you explain how you did it? And I noticed the surface at the leading edge looks weird now, what happened there?

I used RebuildCrvNonUniform with RequestedTolerance=0.001 and MaxNumberPoints=201 to rebuild the extracted edge curve within 0.001 of the original. RebuildCrvNonuniform puts control points closer in areas of high curvature.
Fillet_Problem_DC01.3dm (217.4 KB)

The render mesh is not very good. The actual shape of the surface is the same as the original. if you split the main surface at an isocurve on the leading edge that helps with the render mesh.

Reparameterize allows commands like OffsetSrf and FilletSrf to work. Both those commands fail in Rhino5 which appears to be the Rhino version you are using.

You have to make the fillets manually with filletsrf and then do the trimming and joining.
The large fillet needs to be split also to create the 2.35 fillet between the ends. Rhino won’t let you make a fillet that bridges the same surface