Ali I. Mese

Rock Property Determination Using Digital Rock Physics

The Digital Rock Physics (DRP) technology is based on a rigorous numerical simulation of physical experiments in a realistic pore space, at the pore-scale level. DRP complements lab measurements and, at the same time, enormously enhances the geoscientist’s capabilities because digital experiments can be conducted in real time and on small fragments of rock, such as drill cuttings. We report a feasibility study on DRP technology applied to drill cutting samples to obtain porosity, permeability, and the Pand S-wave velocity. Introduction Porosity and permeability are key petrophysical properties in petroleum industry and environmental applications. Currently, detailed distribution of permeability cannot be remotely obtained in-situ. Decades of analysis of numerous laboratory data points failed to produce universal and robust transforms between permeability and other rock properties such as porosity, lithology, and texture, due to extreme variability of the pore space topology in rocks. This variability is caused by two principal factors: (a) variations in deposition and (b) variations in diagenesis. The most reliable, and, essentially, the only way of measuring permeability is in the laboratory on physical core plugs. In this study, we adopted the concept of virtual (or numerical) experimentation. Specifically, we simulated viscous fluid flow through a realistic pore space numerically represented by zeros (pores) and ones (mineral phase). One advantage of virtual experimentation over physical experimentation is that the former is non-destructive, i.e., a 3D pore space structure that is reconstructed from very small rock drill cuttings can be reused many times in a number of numerical experiments. Also, a 3D pore space can be reconstructed from sidewall plugs that often cannot be used for physical permeability measurements because of the damage during plug recovery. Finally, once a 3D numerical representation of a pore space is constructed, it can be numerically altered to reflect variations in diagenesis and sorting. By so doing, the virtual experimentalist can tremendously expand the database without using additional cores/cuttings when the DRP technique is implemented. Direct estimation of permeability by simulating fluid flow in a pore space has been studied in the past. The complex pore geometry often makes modeling and simulation of transport properties in porous media very difficult. As a result, simplified geometry is often used where pores are replaced with pipes (the network models). Advanced results of simulating fluid flow for a realistic pore space configuration are by Bryant et al. (1993) and Cade et al. (1994). Their approach is to approximate a complex pore space with a set of idealized geometrical figures. Essentially, by using network models, one can arrive at any desired result and match any experimental data. Such a virtual experimentation tool is not predictive. A robust and simple computational tool that can directly handle a complex 3D pore space without adjusting any free parameters is needed for virtual experimentation. One of such tools is the LatticeBoltzmann method (LBM) that is based on statistical description of fluid flow phenomenon. LBM describes fluid flow as collisions of imaginary particles that are much bigger than water molecules. The collision rule preserves mass and momentum and is implemented on a 3D lattice superimposed onto a realistic pore space. This numerical technique for one-phase flow has been perfected by Bosl (e.g., Bosl et al., 1998). Current speed of calculations allows one to calculate the permeability of a million-cell digital sample within seconds on a modern PC. Bosl’s numerical implementation is used in the below examples. 3D pore space can be reconstructed from 2D microscopic images of rock. Consider a 2D slice of a porous material, such as a thin section (Figure 1, left). A 2D digitized image of a thin section can be numerically converted into a binary 2D image represented by pixels with “1” assigned to the pixels that fall into the pore space and “0” assigned to the pixels that fall into the solid phase (Figure 1, middle). Keehm et al. (2001) show that the 2D image statistics (porosity and the variogram) can be used to statistically reconstruct a 3D pore space from its 2D slice via statistical indicator simulation (Deutsch and Journel, 1998).

An Investigation for Disposal of Drill Cuttings into Unconsolidated Sandstones and Clayey Sands

This project include experimental data and a set of models for relating elastic moduli/porosity/texture and static-to-dynamic moduli to strength and failure relationships for unconsolidated sands and clayey sands. The results of the project should provide the industry with a basis for wider use of oil base drilling fluids in water sensitive formations by implementing drill cutting injection into existing wells at abandoned formations and controlling fracture geometry to prevent ground water contamination.