چكيده به لاتين
Reservoir simulation is the most important tool in the petroleum industry for monitoring, controlling, developing and managing hydrocarbon reservoirs. The application of Simulation in reservoir engineering can provide us a precise prediction under different conditions and is able to make decisions for different scenarios. To simulate a reservoir, a set of nonlinear equations must be solved by the iteration method, which requires a lot of running time and memory. Besides, a reservoir has got lots of uncertainty parameters which can make the solution even more complicated and more difficult to simulate. So despite of the advances in computer processors, still reducing the simulation time is a problem for petroleum engineers. In recent years the reduced modeling has been widely used. These methods provide a good approximation of the reservoir behavior by reducing the order of dynamic equations and consequently can simulate model over a faster time. So by reducing the degree of freedom it is possible to reach an acceptable approximation of the final response. at the same time, we spend much less time to run a simulator. In this study, two full order models were used to simulate water injection for reservoir secondary recovery operation which can make a two-phase water-oil flow with high accuracy and run time. Then the proper orthogonal decomposition method was used to reduce the order of the full model to decrease the run-time. Since the reduction of the complexity in a reservoir model is associated with decrease in accuracy, the error of the model was 10% in comparison with the full model, while the run time was 80% lower. In the Next, to increase the accuracy of the simulation, an Adaptive Model was performed. This modeling simulates part of the simulation by a full order performance and then reduces it with an adaptive factor. Or even apply two completely different reduction models by different orders at specific times. adaptive factors were selected according to the values of water cut, pressure difference and mean saturation, which improved the accuracy of the oil and water production rate and water cut in comparison with the results of high or low-order reduction models.
Keywords: black oil simulation, water injection, model order reduction, proper orthogonal decomposition, adaptive simulation.
