What’s ‘the desired result’? I think @pascal is simply implying that fewer control curves will liberate the ability for Rhino to make smoother surface geometry, and therefore the ‘raised parts’ you’re having trouble with, should become much smoother with less curve density constraining them up.
You can see it’s possible to make it smoother, but the shapes that are derived from the curves are only going to be constrained in the manner the designer intends them to be.
A bystander such as myself can only look on and observe the overall design and guess onto infinity as to what the designer wants the shapes to be.
Locking down some curve shapes for a network srf process might be interesting in GH.
I basically played around with some curve networks, and smoothed out control points on UV a few times; did some rebuilding a few times, and used subD conversion a few times.
But I’m not sure how to feel about some of the deviations from the original, without knowing more about the design intent. I’m guessing there’s areas where I deviated too much.
Here again I’d use more of @pascal 's approach next time, and isolate a few curves from the original to prevent so much deviation. I guess I was being hasty and wasn’t careful enough in a few areas.