No abstract

Field development optimization is a crucial and important problem considering economic and physical constraints. Advanced technologies such as smart wells and completions, and well performance optimization, allow for maximizing recovery without the need of additional drilling activities and all associated costs. This type of strategy applies to mature and challenging heterogeneous reservoirs with water and gas coning, and to low permeability and gas condensate deposits, and others. The presented method elegantly considers all field constraints and delivers predictive optimal well control settings taking into consideration reservoir uncertainties and offers optimized development strategies while minimizing on risk and cost. We will present a new field development technology to optimize well performance in both oil and gas fields. The methodology allows for simultaneous optimization under uncertainty of multiple wells using surface constraints and subsurface flow control devices. The method is very fast and uses a combination of gradient and stochastic methods. Starting from an ensemble of realizations of the reservoir model, the best well control settings are searched using a steepest ascent gradient search method, where the search direction can be approximated using the cross covariance of the ensemble of objectives predicted using cloned reservoir simulations for a user defined future time horizon. The workflow solution employs several levels of computational complexity reduction combining advanced hardware architecture and localization algorithms to speedup optimization procedures and cover wider range of spatial uncertainty. We present two examples that include control of surface liquid rates of producers and injection rates of injectors as well as controlling the flow control valves of producer and injector wells to demonstrate both efficiency and robustness of the optimization method. The method presented is very robust and offers with the highest confidence optimal field developments making best use of the most advanced software and hardware technology. The novel presented process results in significant cost reduction on drilling and other operational activities.

No abstract

Traditional reservoir management relies on irregular information gathering operations such as surface sampling and production logging followed by one or several treatment operations. The availability of both diagnosis and the prescribed remedial operations can cause severe delays in the reservoir management cycle, increasing unplanned down-time and impacting cash flow. These effects can be exacerbated in remote and offshore fields where well intervention is time-intensive. A new, innovative, all-electric, flow control valve (FCV) is now a reality for smart completions. This can support any well penetration scenario including multiple zones per lateral in maximum reservoir contact wells and multi-trip completion in extended reach wells. Each zone is equipped with a permanent intelligent flow control valve, allowing real-time reservoir management and providing high-resolution reservoir control. Valve actuation is semi-instantaneous and field data has shown that the frequency of updating such valves is at least 50 times compared to conventional valves, enabling near continuous closed-loop reservoir management. However, such a high frequency optimization demands computational efficiency as it challenges existing optimization applications, particularly when multiple realizations are considered to account for reservoir uncertainty. In this paper, we present a framework to support field-wide implementation of smart FCVs and hence maintaining a fast closed-loop reservoir management. The framework consists of history matching using Ensemble Kalman Filters (EnKF) where smart FCV data is assimilated to condition a suite of representative reservoir models at a relatively high frequency. Thereafter, a reactive optimizer utilizing a non-linear programming method is applied with the objectives of maximizing instantaneous revenue by determining the optimal positions of the downhole valves under user defined rate, pressure drop, drawdown and setting constraints. This optimization offers production control planning suggestions with the intent of immediate to short-term gain in oil production based upon the downhole measurement and the performance of the near wellbore model. At the same time, a proactive optimizer can be used to determine valve-control settings for longer term objectives such as delaying water/gas breakthrough. The objective of this optimization is equalization of the water/gas front arrival times based upon generation of streamlines and time-of-flight (TOF) analysis. Both modes of optimization are performed efficiently such that a single optimization run is sufficient per geological realization. We use the OLYMPUS reference model, a water flooding case, to demonstrate the workflow. The reactive optimization shows an increase of 25% in the net present value through minimizing water production and increasing injection efficiency, while proactive optimization delays water breakthrough time by 2-4 years. The paper showcases the effectiveness of complementary workflows where high frequency reactive and proactive optimizations support a near continuous closed-loop reservoir management.

No abstract