IMPROVING DIGITAL ROCK PHYSICS PREDICTIVE POTENTIAL FOR RELATIVE PERMEABILITIES FROM EQUIVALENT PORE NETWORKS

摘要

Both single-phase and multiphase transport properties computed using quasi-static pore network models show sensitivities to the pore network extraction methods. Bondino et al. [1] show that differences do exist between predicted single-phase petrophysical properties and relative permeabilities for the same sample as a result of differences in the extracted pore networks from digital rock images. Idowu et al. [2] also show that differences do exist in predicted single-phase properties from different pore networks generated from the same sample. Using the same geometrical characterization, they [2] however, show that predicted multiphase transport properties from different pore networks are consistent if the topology of the pore space is well preserved in the extracted skeletons. In addition to the differences in the predicted transport properties, comparison of water-wet and oil-wet relative permeability trends from Bondino et al. [1] and Zhao et al. [3] show counter-intuitive trends for some of the pore networks. Using the same sample as Bondino et al. [1] and a quasi-static pore network model, we examine the causes of the apparent differences in the predicted multiphase transport properties and counter-intuitive relative permeability trends. In real porous media, there is no demarcation of the pore space volume to throats and pores but this is required in pore network models. We show that this geometrical characterization/partitioning of pore network elements in general have significant impacts on predicted single-phase and multiphase transport properties. For the networks examined in this study in particular, differences in the predicted single-phase and multiphase properties are caused mainly by the estimated lengths for pores and throats. However, volumes associated with (or allocated to) pores and throats during the pore network extraction processes are responsible for the observed counter-intuitive relative permeability trends.