Architecture Modeling of Stacked Point Bar Reservoir Based on Meander Loop Database and Automatic Pattern Reconstruction Algorithm: A Case Study in the M Oilfield, Bohai Bay Basin, East China

Abstract The spatial stacking structure and contact relationship between point bars in meandering river reservoir affect the reservoir continuity, interwell injection-production connectivity and distribution of remaining oil. But its characterization and modeling is hard to achieve in traditional modeling algorithm or deterministic microfacies drawing method, especially in offshore oil fields with large well spacing and low resolution seismic data. In order to solve this problem, we developed a new prediction and modeling method based on meander loop database and automatic reconstruction algorithm of sedimentary pattern. This algorithm honors the geometries and interrelationships between point bars while honoring the conditioning data. The general methodology consists of 2 steps. (1) Build a database storing the geometries of meander loops. (2) Select the meander loop from database sequentially and place it in the simulation space constrained by predefined rule and conditioning property. We applied this method to construct the stacked point bar model of M oilfield. Multiple equal probable architecture models were built and were compared with each other through history matching. The optimal predicted model reaches a history matching coincidence rate of 91.3% and was used to predict the abundance zones of remaining oil. Based on the prediction results, 13 adjustment horizontal wells were drilled in the study area to tap the remaining oil. The accurate predicted result implies that the pattern reproduction based on database are effective for characterizing the spatial structure of stacked point bar reservoir.