In the Hangingstone Project (Athabasca, northern Alberta, Canada), oil sands destributes deeper than 250 m, they have been developed by employing the Steam-Assisted Gravity Drainage (SAGD) method. The depositional environment of the oil sands has been interpreted as a tidal-influenced meandering river, which involves extremely heterogeneous lithology with respect to the distribution of mudstones. Since this lithological heterogeneity can affect the growth of the SAGD steam chamber, reliable reconstruction of the lithological heterogeneity in the reservoir model is essential for forecasting SAGD performance via reservoir simulation.
In this study, we developed a geostatistical modeling workflow to reconstruct the lithological heterogeneity in consideration of the depositional concept. Frameworks of abandoned channel-fill and point-bar deposits, as well as the lateral accretion surfaces of point bars, were determined based on interpretation of 3D seismic data. Lithologies of abandoned channel-fill deposits were distributed based on two-point statistics in a horizontal grid system, in which grid cells were aligned horizontally. Lithologies of point-bar deposits were distributed based on multiple-point statistics in an inclined grid system where grid cells were aligned in parallel to the lateral accretion surfaces. Lithologies observed at well locations were used as hard conditioning data and the probability of the reservoir facies obtained from seismic attributes was used as soft conditioning data in the geostatistical lithology model. A training image reflecting the depositional concept of a point bar was constructed and used with the multiple-point geostatistics for the lithology modeling of point-bars.