Speaker: Enrico Scarpa

Date: Thursday 24th of November 2022, 1:15 pm.

Channelized turbidite systems are often gathered into complexes and display various stacking patterns. Their internal architectures represent one of the fundamental properties of a reservoir because they control the connectivity of high-permeability and low-permeability of sedimentary bodies. Some works have analyzed the static connectivity of various stacking patterns; however, few have quantitatively evaluated the dynamic implications of different stacking patterns on fluid flow circulation. In this work, we investigate the impact of different classes of geostatistical modeling methods on static and dynamic connectivity using several metrics. The stacking patterns are generated with an object-based method using Lindenmayer systems. 300 stochastic realizations are grouped into three categories: disorganized stacking channels, disorganized stacking conditioned to a vertical sand proportion map, and organized stacking reproducing channel vertical and lateral migration. To study the hydrodynamic responses, we set a two-phase system containing oil and water. We perform reservoir simulations in all stochastic scenarios and compute the connectivity on the simulations grid to determine the statistical relationship between metrics. This approach facilitates the comparison among flow simulations. It highlights a delay of water breakthrough time in disorganized stacking patterns and a less optimistic recovery efficiency in organized stacking patterns. Our study confirms the positive forecasting bias observed in customary geostatistical modeling that overestimates the actual reservoir connectivities.