Building the SCAL Dataset

From a Handful of Points to a Curve

Special core analysis (SCAL) returns only a few noisy relative-permeability points per sample, but a simulator needs a smooth curve in every cell. The modeler fits a model, usually the Corey form, through the measured points to bridge that gap.

Quality-Checking the Fit

A usable fit must reach zero at the right endpoints, stay monotonic, and pass through the trustworthy points. The Corey form enforces the endpoints and monotonicity by construction, so a good Corey fit is automatically a physically valid input rather than a curve that has to be policed cell by cell.

Least Squares and Judgment

An automatic least-squares fit picks the exponent that minimizes the misfit, but the modeler still decides which points to trust: a single bad measurement can drag the curve away from the true physics, so the best fit is part arithmetic and part judgment.