Speaker
Description
An acute understanding of multiphase flow in the subsurface and its interaction with different minerals is vital in solving challenging applications like CO2-sequestration, underground H2 storage, and enhanced oil recovery. Several dynamic pore-filling events occur at sub minute and sub second time resolution that fast dynamic scans of multiphase coreflooding experiments are required to study them. Advancements in synchrotron-based X-ray microcomputed tomography (micro-CT) have allowed direct in situ visualization of pore spaces and the fluids within it. In this work, the main objective is to study the presence and influence of hysteresis on nonwetting phase trapping in mixed-wet and water-wet Bentheimer sandstone samples through geometric (interfacial area), topological (Euler characteristics), and macroscopic (relative permeability) measurements. For this, fast multiphase flow scans of cyclic drainage and imbibition runs were acquired every 1s with 15-16s time lapse intervals using Australian synchrotron micro-CT beamline at 3.6µm resolution. A typical drainage cycle for both samples involved injection of decane at a low flow rate of 0.03cc/min during which fast dynamic batches of 50 scans were taken to observe the percolation of fluid in the pore spaces, and once the sample reached steady state, a slow scan with increased projections was captured. At the end of the drainage cycle, decane was swapped with brine (15% KI doped), and, in this manner, cyclic drainage-imbibition runs were achieved which aided to assess the influence of wettability in these cycles across the two samples. The acquired images were later processed and labelled using a customized deep learning, U-ResNet model. While Euler characteristics of these multiphase labelled images were measured using the imMinkowski package in MATLAB, the interfacial area between the nonwetting-wetting phase was measured using marching cubes algorithm in AvizoTM. Finally, effective permeability and subsequently relative permeability was estimated using a pore-finite volume (PFVS) solver which was further cross-analyzed using an artificial neural network (ANN) model. Our results indicate that both samples exhibit varied nonwetting phase trapping behaviors wherein, while the water sample showed residual oil saturation (Sor) of 0.52 at the end of primary imbibition, the mixed-wet sample recorded a Sor value of 0.15, suggesting the influence of wettability. Moreover, while hysteresis is observed between the primary drainage and primary imbibition cycles, there was little to no hysteresis present in secondary and tertiary cycles for both Euler characteristic and relative permeability measurements. However, in the case of interfacial area measurements, slightly more hysteresis was evident across the cycles. Broadly, these trends were unlike that observed and reported in literature previously for sandstones and even glass-bead systems, wherein reversibility and repeatability of fluid flow is visible along with prominent hysteresis across the cycles. These observations open new discussion dialogues especially in cases related to carbon dioxide (CO2) and hydrogen (H2) storage where phase connectivity and relative permeability hysteresis are the governing parameters that influence efficient trapping of the nonwetting phase without any leaks.
| Country | Australia |
|---|---|
| Student Awards | I would like to submit this presentation into the InterPore Journal Student Paper Award. |
| Acceptance of the Terms & Conditions | Click here to agree |
