Now this is only my two cents, but...

Trackwork is difficult enough to do correctly on a flat surface. Even tiny mistakes can be drastic when measured in the scale you model in. Any design element you add to further complicate a design necessitates a higher level of precision in executing that design. This means the potential for mistakes not only increases but the margin of error decreases. Also, variables like temperature can wreak havoc on trackwork.

Superelevating a curve is a very real and practical solution for real railroads or highway departments. But honestly, I don't think it would serve any purpose on a model (unless it was a very large model operating at a high rate of speed). However, if you are dead set on modeling a superelevated curve, you should be acquainted with a few of it's characteristics.

Spiral (or easement) curves: these are the transition areas between the tangent (straight section) and the circular curve. This is a parabolic curve.

Runout: the change from flat to superelevated. Typically the runout occurs in the spiral portion of the curve, not in the circular curve. Just as the radius of the spiral curve is steadily changing, so should the cross-slope change throughout the runout.

Cross-Slope: the change in elevation from one railhead to the other measured perpendicular to the centerline of the track and divided by the distance between the railheads. This is usually expressed as a percentage (i.e., "x" vertical feet per 100 horizontal feet).

Although the design speed and radius of the curve dictate the runout length and cross slope, a general rule is the length of the spiral in and spiral out is usually about 1/3 the total delta of the curve. And a typical cross slope would be between 2-5%.

Whatever you do to construct this type of curve, be consistent. Aberrations will be obvious to your rolling stock but might be insignificant to the eye. Good luck if this sounds like something you'll try.