If the vessel surface is fixed, then it might be much faster to compute a `distance field` of the min distance to the vessel from a set of points inside of the vessel volume, then spline that data (as a function of, say R,phi,Z), then use that splined data to determine the min distance instead of calculating it explicitly at each step. Our surfaces are continuous in space, so I would think the minimum distance maybe is likely continuous as a function of (R,phi,Z) space as well so the spline should be a good fit of it.
If the vessel surface is fixed, then it might be much faster to compute a `distance field` of the min distance to the vessel from a set of points inside of the vessel volume, then spline that data (as a function of, say R,phi,Z), then use that splined data to determine the min distance instead of calculating it explicitly at each step. Our surfaces are continuous in space, so I would think the minimum distance maybe is likely continuous as a function of (R,phi,Z) space as well so the spline should be a good fit of it.