Integrated Geomodel Accuracy Enhancement Based on Embedded MPS Geological Modeling for Thin Interbedded Reservoirs

Continental delta deposits are characterized by strong heterogeneity in the lateral direction; meanwhile, reservoir development is challenged by rapid changes in rock properties. Thus, it is critical to use proper methods for fine characterization to confirm the distributions of thin interbedded reservoirs. The aim of this study was to propose a novel workflow for integrated research on the 3D geomodeling of thin interbedded reservoirs, using the Triassic T 2 a 1 formation in the Tahe Oilfield B9 area of the Tarim Basin as a case study. The complicated representation of thin interbeds in a 3D geomodel was simulated using a multiscale joint controlling strategy, based on wells (Points), 2D geological cross-sections (Lines), and horizontal wells (Surfaces). The resistivity inversion results from the horizontal wells validated the proof of the plane distribution of the thin interbeds within the drilled area, and this quantitative statistic provided effective parameters and guidance for 3D interbed geomodeling. In this study, comprehensive 3D facies modeling was divided into 3D interbed geomodeling and 3D sedimentary facies modeling. An optimized interbed geomodel was picked out from multiple stochastic simulation realizations, and the drilled horizontal well data were used to constrain the simulation process, so the simulation results were more consistent with the real distribution of the thin interbed morphology. Classical two-point geostatistical methods, the multipoint simulation (MPS) geostatistical method, and the hierarchical mindset were integrated for the microfacies simulation. This procedure demonstrated a good ability to characterize thin interbed reservoirs in continental delta deposits. An MPS training image obtained from a high-resolution satellite photo was used to fix the issue of the relationships between the distributions and configurations of all microfacies within the spatial distribution. A 3D lithofacies interbed model was embedded into the 3D facies model. This comprehensive facies model served as a constraint condition in the property modeling process. A porosity model was simulated using separate stratigraphy and individual microfacies controls, as facies-controlled property modeling has been used as a prior foundation for field development planning in the Tahe Oilfield B9 case. The porosity model was then used as a basis for permeability modeling, and a water saturation model was created using the J function and all of the constraints from the other two property models. Finally, all the results were validated using dynamic production data from the Tahe Oilfield B9 wells, with good matching observed between the geological models. There was only a 0.92% difference in reservoir volume between the reservoir simulation results and the static geological model results using our solution.