Agile Field Development Study Leveraging Fast-Tracked Reservoir Modeling and Data Analytics While Incorporating Reservoir Uncertainties

Abstract

Abstract A small onshore brownfield in south Oman has low oil recovery because of its heavy oil and high water production, which together with reservoir uncertainties poses development challenges. Petrogas implemented an innovative field development planning approach to quantitatively compare multiple field development scenarios and optimize the operational choices within each. The workflow started with a single history matched model for each of the two geological structures in the field. A set of 14 field development scenarios were defined, on injection rates, well locations, and injection fluids. Identification and quantification of subsurface uncertainties were performed. These uncertainties were included in the geomodel for each scenario, which generated an ensemble of realizations and corresponding production forecasts. Two sets of economic results were produced—a simple, discounted cashflow model and the fiscal terms of the operator's service contract. Each ensemble was run against these models to generate probabilistic performance indicators for each scenario. Using cloud-computing capability, the field development study was drastically accelerated without losing on the quality. Almost 800 simulations were run over 5 days, covering 32 development scenarios in total (for two structures), automatically integrated with the economics workflow, providing in-depth analyses. The scenarios were compared in a series of dashboards that presented the economic metrics and their corresponding cumulative distribution functions. The analysis yielded several important insights: longer wells did not provide enough additional production to offset the increased costs. Moreover, peripheral drive with horizontal wells was more effective than irregular vertical wells. The waterflood scenarios improved production, but the polymer-injection option with short horizontal wells and peripheral infill well pattern was the highest-performing scenario. The study also helped identify areas where more detailed optimization studies should be performed, e.g., to optimize polymer-injection scheduling and polymer design. Traditionally, subsurface uncertainties analysis was restricted to a small number of discrete model realizations. Results were quantified in terms of production ranges only. Here, production forecasts were based on an ensemble of models, capturing the full range of uncertainties. In addition, evaluation criteria included economics.