Quantifying geological data through stochastic modelling

Abstract

The genesis and properties of a depositional system can be understood by the use of facies models, al/owing the prediction of their spatial distribution and reservoir qualities. The qualitative approach, supplemented by numerical data, permits the quantification of flow units. The purpose oi the method presented herein is to translate geological data into numerical data byapplying geostatistical techniques ot stochastic modeling. Geological description of turbidite reservoirs was carried out in order to provide the basis for stochastic modeling. Four facies were initially recognized in well cores: channelized lobe, tobe, lobe fringe, and hemipelagic; then, these facies were correlated throughout the reservoir by using wel/ log data collected in 25 wells. In this geostatistical treatment, facies were considered as categorical variables, due to their close relationship with flow units. The variographic study and stochastic modeling were performed in stratigraphic coordinates. Indicator principal components were used in the modeling of facies. The results were compared to the geological model. Conditioned to the results of facies modeling, a sequential Gaussian simulation was used for porosity simulation. Absolute permeability simulation was conditioned to the facies modeling results. In this situation, the Monte Carlo methodology was applied.