This work presents a method to test the use of intelligent wells in probabilistic scenarios. The proposed algorithm assists the decision about the number, position and operation of the interval control valves (ICV). The procedure focuses on the evaluation of intelligent well controls when optimizing complex oil field production strategies considering geological uncertainties.
The method assists the ICV application in uncertain scenarios, seeking to enhance flexibility and to improve the expected monetary value (EMV) during the field development phase. This work presents a detailed analysis of a production strategy previously optimized for conventional wells. The algorithm divides the production period into smaller time sections and optimizes them individually in a time sequence. For each section, performance indicators evaluate wells and regions for the potential to install ICV. We simulate the different combinations only of promising regions, restricting search space, which results in fewer simulation runs.
We used the benchmark case UNISIM-I-D, which is based on the Namorado field (Campos Basin, Brazil). EMV of the representative models was the objective function and we optimized ICV controls for a long-term exploitation strategy (lifecycle optimization procedure). We compared the proposed proactive control with an optimized reactive control to assess the results.
The study considers two different platform capacities (one optimized for the conventional well strategy and one with reduced capacity). We also show that when a platform restricts water production, the ICV closure affects not only the nearest wells but also other wells with high water rates, so this also must be evaluated.
Most ICV studies use techniques that are computationally expensive in real field evaluations. Barreto and Schiozer (2015) optimizes ICV placement with a low number of simulations but they check the viability of a single ICV implementation in each cycle. We propose testing various ICV simultaneously, respecting the influence of cross parameters by partitioning the production period into a sequence of predetermined time sections and optimizing each time section individually, forward in time.
The proposed optimization process demands a feasible number of simulation runs, enabling use in practical applications and complex reservoirs.