Summary
Maximization of the yield of existing assets becomes more important than ever as the petroleum companies need to win in any business environment. In this context, model-based optimization technology plays an important role in managing efficiently the subsurface flow and can add significant values by maximizing the potential of reservoirs without a large capital investment. Yet, conventional optimization methods did not sufficiently respect expert knowledge and engineering requirements, which severely undermines their business impact in practice. This paper presents a novel interactive workflow that permits injection of expert knowledge into optimization process and ensures the final optimal solution executable. This workflow is unique because it allows to (i) interact with stakeholders, e.g., production engineers and operators, to capture engineering and economic requirements and constraints, (ii) interact with software to identify, screen, and maximize the opportunity, (iii) interact with reservoir to understand the physics for meaningful solutions, and (iv) interact with candidate solutions for the most rigorous one. Data analytics is used in this interactive workflow, boosting the optimization progress to reach the most trustworthy result. An offshore waterflooding example is used to illustrate the workflow proposed. Results show that the optimal solution generated significantly improves, compared to the existing strategy, the estimated short-term and long-term oil recovery (by more than 2% and 6%, respectively). Moreover, the water production volume is largely reduced. The proposed solution is feasible in engineering (meet engineering requirements and engineers’ judgements and expectations), meaningful in physics, optimal (convergence is guaranteed), robust (multiple uncertainties are considered), stable (immune to potential implementation errors), trustworthy (backed by data analytics), and thus executable in practice.
