Streamline-Based Robust Rate Allocation Optimization for Polymer Flooding Considering Geological Uncertainty

Abstract

Abstract Optimal injection/production rate allocation is an integral part of effective reservoir management for enhanced oil recovery (EOR). Field-scale rate allocation optimization with hundreds of wells can be time consuming and complicated by the multitude of operational and well related constraints. The common practice is to either do the optimization manually or limit the simulation-based optimization to a single realization which ignores the geological uncertainty and causes suboptimal rate allocation. This paper proposes a streamline-based robust optimization algorithm that accounts for geologic uncertainty while allocating injection/production rates to maximize sweep efficiency and hydrocarbon recovery. The algorithm follows a sequence of time intervals. For each time interval, the well rates of each realization are tuned iteratively based on well pair rate multiplier that is consolidated for all realizations to account for geological uncertainty. The robust optimization is based on an expected value of the performance index over multiple realizations, and can be tuned to include appropriate risk tolerance. The optimization process continues to the next time interval, until the end of the field life. Power and utility of the proposed algorithm is applied to a large-scale polymer flooding field case (base model). Multiple geological realizations are generated by grid-connectivity transformation (GCT) based on the base model to account for geological uncertainty. Before optimization, multiple history-matched realizations are selected and divided into training realizations and one blind test realization. The training realizations and base model are used in their entirety to conduct robust optimization, but the base model is chosen to conduct nominal optimization for comparison purposes. The optimal rate schedules from both the robust optimization and the nominal optimization are applied to a blind test realization to examine the impact of geological uncertainty. The optimal rate schedule from the robust optimization consistently outperformed that from the nominal optimization in terms of sweep efficiency, hydrocarbon recovery, and polymer utilization.