Simulation of rock properties using 3D-printing and machine learning
Mohamed Soufiane Jouini, Ezdeen Raed Ibrahim
Characterizing porosity, absolute permeability and elastic properties in cored area from oilfield reservoirs is a crucial step to evaluate hydrocarbon reserves. This characterization is more challenging in carbonates rocks representing half of the world reservoirs and the main type of reservoirs in United Arab Emirates. Indeed, their high heterogeneity due to presence of pores at macro, micro and nano scales resulting from diagenesis makes rock properties characterization very complex. Measurements obtained in the laboratory are effective measures which do not reflect the variability and effect of each range of pore sizes (nano, micro and macro). In the last decades several developments in Digital Rock Analysis technology showed the ability to estimate reliably rock properties using images in sandstones but faced several limitations in carbonates due to their heterogeneity. The general strategy to deal heterogeneity is to extract physically small subsets from representative “homogeneous” zones to be scanned at a finer scale. Numerical simulations allow estimating numerically rock properties on digital models of subsets. However, due to the very small size of subsets it is not possible to validate these simulations through laboratory experiments. Also the multi scale strategy to upscale simulations from fine to coarse scales suffers from clear workflows in carbonates. In this study, we have two main objectives: the first consists on using 3D printing technology to print extracted subsets in larger scale to make laboratory experiments feasible. Producing these 3D print of rock subsets will allow validating our simulations experimentally at laboratory scale. The second objective of the study is to use texture analysis and machine learning to propose a reproducible workflow for multi scale analysis and to simulate rock properties in complex carbonates rocks. We will present application of our proposed approach on synthetic and real samples.
