Comparative study of different permeability upscaling techniques in porous media

Abstract

The development of the geological model, an intrinsic step in reservoir simulation, generates a large amount of data, including the petrophysical properties of the porous medium. This makes it impractical to run computational simulations directly on these models due to time and processing constraints, despite the ongoing technological advancements. To simplify this model, different property upscaling techniques can be applied. However, if the upscaling process is not performed properly, the results may be unreliable and inaccurate, ultimately compromising decision-making. This study aimed to evaluate the application of averaging techniques, specifically the harmonic, geometric, and arithmetic means, and the flow-based technique on the permeability data of a representative model with a ¼ five-spot geometry, with horizontal and vertical dimensions of 2,000 m and 40 m, respectively. The model described was discretized at three different mesh coarsening levels. By analyzing the oil recovery factor, its derivative, and the water cut, it was found that applying the flow-based technique to the intermediate mesh provided the best balance among the cases. It resulted in a shorter simulation time and a reasonable deviation of 2.44% compared to the finest mesh, with the water cut fixed at 0.8.