Steam allocation and production optimization in SAGD reservoir under steam-to-oil ratio uncertainty

Abstract

Steam allocation is an important decision for bitumen thermo-recovery using Steam Assisted Gravity Drainage (SAGD) technique. This is due to the significant amount of steam requirement and often limited steam generation capacity. Steam-to-oil ratio (SOR) is an important parameter affecting the production performance. It is necessary to address uncertainty in SOR to prevent constraint violation in SAGD reservoir states such as subcool and also to maximize the overall steam utilization efficiency. In this work, we study the problem of steam allocation and oil production optimization in the SAGD process considering SOR uncertainty. A first principle model for the SAGD process is developed and further incorporated into the Nonlinear Model Predictive Control (NMPC) problem, which enforces the system to be within various constraints while optimizing an economic objective. The uncertainty is dealt with using three methods in this work: (i) open-loop worst-case optimization, (ii) scenario tree based closed-loop optimization and (iii) affine policy based closed-loop optimization. Performances of the above methods are compared through Monte Carlo simulations. Results demonstrate the superiority of affine policy based optimization method, which has around 50% improvement of economic performance over static robust and scenario based method in handling SOR uncertainty.