Date of Award
Master of Science (MS)
Fluid-flow physics in porous media has been continually simplified by assuming isotropic homogeneous media with minimal rock-fluid interactions. Such simplification did not reflect reality and retained the ability to understand the flow behavior essence in unconventional reservoirs. The developed physics should be reevaluated on ideal porous media, indeed, which has minimal geometrical and interaction uncertainties. Therefore, image processing techniques were utilized to processs the CT scans of core samples to construct ideal 3D-printed replicas for coreflooding experiments and simulation models. The results from both were then compared for vcalidation and cross check. Methods, Procedures, Process: Grayscale CT-scan of a Berea core sample was digitally binarized to segment the grains cloud in the scan. That cloud was meshed and triangulated to form a 3D-printable object. The processed object was 3D printed with different 3D printing technologies and materials. The gypsum-replica, which had the closest petrophysical properties to the original Berea, was extensively investigated through a CO2 huff and puff experiment simultaneously with its original, geomechanical UCS (uniaxial compression strength) test, Nitrogen sorption, MICP (Mercury Injection Capillary Pressure), and contact angle wettability measurement. Based on the image processed CT scan, a finite difference model was created in which the petrophysical characteristics, porosity and permeability, were inferred from the CT scan. The model was used to simulate a transient permeability experiment on the Berea sample and the CO2 huff and puff experiment. The 3D printable volume was also used to create a finite element model to simulate the UCS test on the replica. Figure 1 shows the 3D printed replicas with their original Berea along with the coreflooding and geomechanical simulation models based on the reconstructed CT scan. Results, Observations, Conclusions: The 3D printed replica was able to represent their original sample with close storage and transport capacities. The used image processing workflow generated a precise static model for black oil (transient permeability) and compositional simulation models (CO2 huff and puff) of both samples. The CO2 effect on the core sample was pictured after breaking the replica to check its interior, and the simulation model was able to predict a similar saturation distribution. The simulation results accurately matched the replica’s measured oil recovery, pressure distribution during the transient permeability test. After including the UCS test schedule, the model succeeded in generating fatigue iso-surfaces, stress and strain contours, failure limits and modes, force reactions inside the core sample. Novel/Additive Information: The proposed image processing can produce the physical specimens for tests along with the needed models to simulate these tests. 3D-printed core replicas, which are created by reconstructing cores’ CT scans by image processing, are valuable for repetitive and destructive experiments and obey the criterion of ideality for laboratory research. The created coreflooding and geomechanical models are robust and precise for developing and understanding the physics of fluid flow in porous media.
Almetwally, Ahmed Galal Alqassaby, "Development Of Image Processing Techniques For Core-Scale Characterization And Synthetic 3D-Printed Core Replicas" (2021). Theses and Dissertations. 3906.