Rock physics & Interpretation

[RP-07] Simulation and Validation of Porosity and Permeability of Synthetic and Real Rock Models Using 3D-Printing and Digital Rock Physics.

*Ezdeen Raed Ibrahim¹, Mohamed S Jouini¹, Fateh Bouchaala¹ (1. Khalifa University (United Arab Emirates))

Predicting rock properties of carbonate reservoir rocks is crucial for the oil industry. Nevertheless, standard models have many limitations in carbonate rocks due to the presence of complex pore structures at several length scales. Several studies showed that Digital Rock Physics has reliably characterized siliciclastic rock from 3D X-ray Micro tomography images, but this technique often fails in the case of carbonate rocks. Current scanning devices have not yet the capability to acquire images at various scales simultaneously due to physical limitations associated with detector sizes. This may be worked around by imaging the whole core plug sample at a coarse scale and acquire smaller subsets at a finer scale. Although rock properties of such subsets may be estimated through numerical simulations, their smaller size makes laboratory experiments not feasible and hence impossible to validate a simulated result. This is where lies the importance of 3D-printing and its potential to be a game-changer. 3D-printing said enlarged subsets to a diameter of 1.5 inch allows for poroperm (porosity & permeability) experiments which may validate the analytical and simulated petrophysical properties. With a vertical resolution of 25µm, the stereolithography 3D-printer is able to capture fine pore networks using a resin that has good mechanical properties when cured, sufficient for flush-cleaning and poroperm experiments. The input for the 3D-printer is a mesh file known as a STL (or STereoLithography) file, generated from processed raw data of micro-CT scans. Once the samples are printed, they are flush-cleaned under high pressure and temperature to remove any excess resin blocking the pore networks then experimented on. The validated petrophysical properties may then be mapped onto the texturally segmented core plug through the integration of textural analysis and machine learning.