I ran into troubles trying to use best fit command to determine centres of arcs from survey points. Points were shot on a cylindrycal structure lying in horizontal position in a local coordinate system with northing (Y axis) aligned with the axis of the cylinder. I have a few sets of points with constant northings for each arc, but easting and elevations vary. I tried changing UCS so that all the points for a particular set (with constant northing) lie in the XY plane and therefore have 0 elevation but it won't allow me to use the feature saying the "points are collinear". I even tried to draw lines between the points in that UCS (again all elevations 0) and snap on line ends when picking the points but the same issue. Anyone knows what might help ?

Thanks a lot for any help.