A novel method to model water-flooding via transfer function approach

Water flooding is one of the most economical methods to increase oil recovery. To maximize oil recovery during water-flooding, it is essential to provide a forecast of reservoir performance. Hence, various methods are used to simulate reservoirs. Although grid-based simulation is the most common and accurate method, time-consuming computation and demand for large quantities of data restrict the use of this method. This study presents the development of a new method to predict the performance of water injection based on Transfer Functions (TF). This method is faster since it requires less data and the only requirements are injection and production rates. In this method, it is assumed that a reservoir consists of a combination of black boxes (TFs). The order and arrangement of the TFs are chosen based on the physical condition of the reservoir which is ascertained by checking several cases. The injection and production rates act as input and output signals to these black boxes, respectively. After analyzing input and output signals, unknown parameters of TFs are calculated. Then, it is possible to predict the reservoir performance. Different cases are employed to validate the derived model. The simulation results show a good agreement with those obtained from common grid-based simulators. In addition, we found out that the TF parameters depend on the characteristics and the pattern of different sections of the reservoir. This method is a rapid way to simulate water-flooding and could be a new window to the future of fast simulators. It enables prediction of the performance of water-flooding and optimization of oil production by testing different injection scenarios. The method also provides key parameters such as well connectivity.