Prediction of Cementation Factor and Saturation Exponent Using Genetic Programming Algorithm

In this article, strong correlations for the cementation factor and saturation exponent were discovered by genetic programming (GP) algorithm. The cementation factord GP-base model was trained by input variables such as porosity, permeability, and grain density derived from 175 routine core analysis (RCAL) samples of 21 carbonated oil fields. Also, porosity, permeability, and wettability index were considered as input variables of saturation exponent model. The proposed correlations using GP improved greatly the average absolute error for the Archie’s parameters. The root mean square error and correlation coefficient of validation data for the new cementation factor correlation were 0.062 and 0.91, and for saturation exponent model, they were 0.051 and 0.96 respectively. The importance of theses correlations is in their dependence on simple measurable parameters, and except wettability index all of the independent parameters are simple routine core analysis parameters which can be measured easily and at no considerable expense.