مدیریت زمان واقعی فرایند جریان آب با استفاده از مدل مخزن پروکسی و نظریه همجوشی داده

Waterflooding is the use of water injection to enhance the oil recovery in mature oil reservoirs. In this paper an adaptive algorithm has been introduced for waterflooding management in oil reservoirs based on proxy modeling technique. The presented approach is capable to handle the time-varying nature and the inherent nonlinearity of the complex process. In addition, any variation either in market prices or in operational costs is compensated by the designed adaptive controller to fix the obtained profit (here, the net present value: npv) at a desired achievable value. The observed outcomes on 10th SPE-Model#2 benchmark case study have shown that by using this algorithm, any feasible desired trajectory for the expected benefit can be satisfied during the waterflooding-based production. Since the suggested controller has adaptive structure, it can be re-adjusted continuously in each time-step, using available operational data, to take into account the reservoir dynamical variations as well as the external disturbances to present an acceptable performance. By including a monitoring module in the algorithm structure based on data fusion technique, the updated profitability/productivity status of the reservoir is estimated. By using this information the npv setpoint induced to the closed-loop system can be automatically re-adjusted such that it always remains in an acceptable and reasonable range. In conclusion, the proposed methodology is an applicable solution for fairly profit-sharing in different kinds of contracts. In other words, the gained profit can be appropriately allocated to the shareholders according to the contractual obligations or a defined npv trajectory while considering the current condition of the reservoir. This strategy helps to prevent from ultra-production in a specific period of time by the clients or contractors which may lead to an unexpected reduction in the share of other parties in the reservoir life-cycle.