Multiscale Pore Network Modeling of Hierarchical Media with Applications to Improved Oil and Gas Recovery

Abstract

For complex geological materials such as carbonates and tight sandstones having pores at several scales, the conventional relationships are not adequate to quantify transport properties. Therefore, it becomes important to study these complex rocks at the pore scale and apply relevant physics for transport properties. However, with the current state of imaging technology it is not possible to obtain realistic images of the rock having pores at several orders of magnitude in a single image. Therefore, it becomes necessary to develop modeling tools that can study images with unresolved porosity. A pore network can be extracted such that the transport properties of the visible voids are calculated, while the interplay between micro- and macro-porosity can be studied by modeling the unresolved pores as effective continua. In this work, first we have attempted to generate three phase multiscale artificial images using PoreSpy and then devised a method of network extraction on these three phase images in a single step and thus created a multiscale pore network model using OpenPNM. Also, 3D and three-phase segmentation of real carbonate images were prepared where the developed algorithm was successfully tested. A cubic grid is applied to the micropores region which becomes the mesh for the continua simulation with each element endowed with effective properties. The macropores are then stitched together with the continua scale, thus creating a hybrid hierarchical pore network that possess information at several scales. The multiscale pore network algorithm prepared in this work is fast and robust and has been tested on several 2D and 3D artificial and real rock images. Porosity, permeability, and formation factor have been calculated on the resulting pore networks and validated with the real sandstone and carbonate images.