A geostatistical analysis approach for facies and porosity modelling of a heterogeneous sandstone reservoir to compare four practical stochastic methods

چکیده مقاله

The complexity of a hydrocarbon reservoir is one of the most challenging subjects to choose a proper method for static reservoir modelling. Geostatistical methods are among the widely used techniques to characterise the hydrocarbon reservoirs. Geostatistical analysis deals with estimation of statistical parameters that specify the spatial distribution of variables as well as uncertainty distribution. In the literature, there have been provided variety of geostatistical methods for property and facies modelling. In this article, the main industry methods are employed and compared to propose the most appropriate method. At the first step, a static model was constructed in a sandstone reservoir using fifteen wells. To do so, the structure was generated on the first step. For facies and property modelling: Sequential Indicator Simulation (SIS) and Truncated Gaussian Simulation (TGS) methods for facies modelling and Sequential Gaussian simulation (SGS) and Gaussian Random Function simulation (GRFS) methods for porosity modelling were employed and compared. The validity of outputs from SIS and TGS were investigated by using proportions, histograms and variograms. In this study, SIS showed a better performance in reproducing the proportions as well as indicator variograms in 25 realisations, compared with those of the TGS. Furthermore, for the simulation of the porosity, a better variogram reproduction was observed for GRFS in 25 different realisations with faster runtime speed (almost five times faster) compared with those of SGS. Finally, cross-sectional models in SIS method show the continuity for the facies correlation better than those of TGS. In addition, GRFS reproduced the porosity (from less than 1 to 31%) with a normal distribution better than SGS.