Integrated Modelling Approach Employing High Resolution Porosity Realizations Derived from Impedance-Porosity Joint Inversion in a UAE’s Complex Carbonate Reservoir

Abstract

Abstract This integrated study focused on developing an innovative workflow to populate vintage reservoir grids with seismically derived properties in a heterogeneous carbonate reservoir. The COHIBA workflow of time-depth model adjustment was the foundation for the joint porosity inversion and generation of high-resolution porosity realizations in the reservoir model. The multi-realization down-scaling workflow provided porosity realizations at the fine grid scale that fit the seismic information and can be further used in geomodelling workflows. The initial step of this integrated study workflow was a global update of the existing reservoir structure using the COHIBA methodology to model depth surfaces taking into consideration the uncertainties in the input parameters. The resulting depth surfaces were then used to build the reservoir grid. Deterministic geostatistical inversion was used to jointly invert for impedance and porosity. A downscaling workflow was added to obtain 100 porosity realizations, honoring well data, under control of low-resolution inverted porosities from seismic. The best ranked realizations were assessed against the porosity populated in the reservoir model using conventional gridding methods. After the depth horizons were adjusted by integrating all vertical and horizontal wells, the velocity model and depth prediction were successfully updated. Resulting depth surfaces gave an improved and more reliable structural model. The updated velocity model ensured the transfer of the depth prior models to the correct position in time. As a results from inversion and downscaled porosity realizations at the fine grid were consistent with the updated reservoir grid which helped realizations to be directly utilized in the reservoir grid. A ranking exercise was performed on the reservoir’s global pore volume to select a subset of 25 realizations out of 100 realizations. This allowed working with less realizations while still representing the global distribution and variability. The results were further assessed by generating alternative reservoir models populating the log porosities using conventional gridding algorithms and comparing those with the selected group of derived porosity realizations. Zone wise comparisons showed good correlations between porosity realizations & porosity derived from the conventional methods, specifically, in the areas of good seismic data quality. Additionally, results were checked using blind wells. Based on the outcomes of this analysis, the porosity trends were used to constrain the porosity population in reservoir modelling. This workflow provides a more reliable structural model and depth prediction. Joint inversion allows better integration of uncertainties on the computed porosity volumes. The multiple-realization down-scaling workflow provided porosity realizations at the fine grid scale that correlates with the seismic data.