Integrating Design of Experiments, Proxy Modeling, and Monte-Carlo Simulation for Combined Uncertainty Quantifications of Geological and Production Data in the Cyclic GAGD Process

Abstract

Abstract Uncertainty assessment in reservoir simulation studies is a crucial issue to quantify future planning risks through the field development scenarios. The uncertainty comes from the lack and sparseness of geological and reservoir data that negatively affect capturing the most realistic geological environment for the reservoir flow simulation. Therefore, the uncertainty should be quantified for more plausible and convincible procedure to make a trustful decision regarding future production plans. In this paper, the uncertainty was quantified with respect to the geological and production parameters through the cyclic implementation of the Gas Assisted Gravity Drainage (GAGD) process in heterogeneous sandstone oil reservoir. The geological uncertainty assessment includes generating multiple stochastic reservoir models for horizontal permeability and anisotropy ratio. In a successive step, the production uncertainty was assessed with regards to the decision parameters of the cyclic GAGD process simulation. The cyclic production parameters are durations of injection, soaking, and production in addition to the minimum bottom hole pressure in production wells. To quantify the geological uncertainties, a large number of stochastic reservoir models (realizations) were created honoring geological constraints. Since it is impractical to simulate all these realizations, ranking was applied to select the nine quartiles of P10, P20, and P90 of permeability and anisotropy ration models that represent the overall reservoir uncertainty. The selected realizations were evaluated in the compositional reservoir model to calculate the field cumulative oil production. The full factorial design was implemented to produce 81 reservoir models to quantify the geological uncertainty. After that, the most likely reservoir model was adopted for the production data uncertainty assessment to determine the least uncertain space of the cyclic parameters that leads to true optimal solution. In the production uncertainty assessment, the proxy-based Box-Behnken Design and Monte-Carlo simulation approach was adopted to create and evaluate the training and validation jobs through the compositional reservoir simulation. The presented successive uncertainty quantification workflow has led to obtain higher field flow response than the base case of default settings of the decision parameters. That was achieved after minimizing the uncertain space of the geological and production data that highly impact the cyclic GAGD process performance.