Handling risk of uncertainty in model-based production optimization: a robust hierarchical approach

Model-based economic optimization of oil production suffers from high levels of uncertainty. The limited knowledge of reservoir model parameters and varying economic conditions are the main contributors of uncertainty. The negative impact of these uncertainties on production strategy increases and becomes profound with time. In this work, a multi-objective optimization problem is formulated which considers both economic and model uncertainties and aims to mitigate the negative effects i.e., risk of these uncertainties on the production strategy. The improved robustness is achieved without heavily compromising the primary objective of economic life-cycle performance. An ensemble of varying oil price scenarios and geological model realizations are used to characterize the economic and geological uncertainty space respectively. The primary objective is an average NPV over these ensembles. As the risk of uncertainty increases with time, the secondary objective is aimed at maximizing the speed of oil production to mitigate risk. This multi-objective optimization is implemented separately with both forms of uncertainty in a hierarchical or lexicographic way.