Optimization of Waterflooding Performance in a Layered Reservoir Using a Combination of Capacitance-Resistive Model and Genetic Algorithm Method

Managing oil production from reservoirs to maximize the future economic return of the asset is an important issue in petroleum engineering. In many applications in reservoir modeling and management, there is a need for rapid estimation of large-scale reservoirs. The capacitance-resistive model (CRM), regarded as a promising rapid evaluator of reservoir performance, has recently been used for simulation of single-layer reservoirs. Injection and production rates are considered as input and output signals in this model. Connections between the wells and the effects of injection rates on production rates are calculated based on these signals to develop a simple model for the reservoir. In this study, CRM is improved to model a multilayer reservoir and is applied to estimate and optimize waterflooding performance in an Iranian layered reservoir. In this regard, CRM is coupled with production logging tools (PLT) data to study the effects of layers. A fractional-flow model is also coupled with the developed CRM to estimate oil production. Genetic algorithm (GA) method is used to minimize the error objective function for the total production history and oil production history to evaluate model parameters. GA is then used to maximize oil production by reallocating the injected water volumes, which is the main purpose of this research. The results show that our fast method is able to model liquid and oil production history and is in good agreement with available field data. Taking into account the reservoir constraints, the optimal injection schemes have been obtained. For the proposed injection profile, the field hydrocarbon production will increase by up to 1.8% until 2016. Also, the wells will reach the water-cut constraint 2 yr later than the current situation, which increases the production period of the field.