Enhanced oil recovery from fractured low permeability reservoir rock is challenging and while studied previously, significant challenges remain in improving the recovery factor. In many cases the reservoir has been pressure depleted and gas released from solution. In such cases, continuous gas pathways from fracture to the matrix may have been created aiding our ability to deliver gas to the matrix. Gas injection (i.e., CO2) then proceeds from immiscible to perhaps miscible conditions. The efficiency of gas transport to the low permeability matrix is limited by the presence of a gas phase in the matrix, injection pressure, and cocurrent or countercurrent flow conditions, among others. This work presents the results of a series of experiments using low permeability (0.02 - 1.3 mD), medium porosity (30 - 40 %) siliceous shale reservoir core samples initially saturated with either: live oil, depleted to a pressure of 200–300 psi; or dead oil brought to near miscible conditions; followed by CO2 injection at pressures proceeding from immiscible to CO2 near miscible conditions. Two gas injection modes were used: CO2 flow across one face of the core (countercurrent flow) which intends to account for the diffusive transfer mechanism; and CO2 flow along the length of the core (cocurrent injection). The experimental setup was monitored using X-ray computed tomography that helped to visualize phase flow and distribution during the processes. The incremental oil recovery caused by the immiscible CO2 injection ranged from 0–10% for countercurrent flow mode and from 18–25% for cocurrent flow mode, whereas for near miscible conditions the recovery was found to be 25% for the countercurrent flow mode and 10% for the cocurrent flow mode. Countercurrent recovery appears to be critically sensitive to the distribution of the gas phase near the fracture face. Oil recovery potential by CO2 injection into siliceous shale rock is challenged by low permeability, rock heterogeneities, distribution of oil within the rock matrix, but it is aided by the presence of continuous gas pathways that allow CO2 penetration into the matrix. Near miscible injection displays clear multiple contact miscibility; also countercurrent flow seems to exhibit a greater recovery under near miscible as compared to immiscible conditions.

This work presents the results of tests using a low permeability (1.3 mD), medium porosity (34%) siliceous shale reservoir core sample initially saturated with oil followed by CO2 injection at miscible conditions. The tests had 3 primary flow conditions that replicate aspects of expected field conditions: (i) countercurrent exchange of carbon dioxide from fracture to matrix with consequent production of oil (ii) cocurrent exchange of carbon dioxide from fracture to matrix with oil production at the downstream core face, and (iii) a flushing stage where carbon dioxide is injected at a pressure drop of 25 to 50 psi to drive the core to residual saturation. The countercurrent and cocurrent conditions are conducted at very low pressure difference across the fracture at the face of the core and along the length of the core, respectively. The miscible injection of CO2 recovered virtually all of the original oil in place. Results were simulated with a compositional reservoir simulator. The simulation model was constructed to be as similar as possible to the experimental conditions, including a heterogeneous permeability field. An exact match of the experimental results was difficult to obtain. Calculations agreed with the experiments on the total oil production but not the timing of recovery. Experimentally, more oil was recovered during countercurrent flow in comparison to simulation. Nevertheless, simulations did reveal CO2 penetration of the matrix via a countercurrent diffusion/dispersion mechanism. Incorporation of fracture-to-matrix CO2 transfer and matrix-to-fracture oil transfer were key elements of the simulation.

Serpentinization of olivine is often studied in the laboratory under batch conditions. Olivine conversion in situ with enhanced natural hydrogen production will likely be implemented via injection of aqueous solutions. Hence, transport is relevant to the extent of olivine reaction and, potentially, the morphology of precipitates formed. To test conditions for optimal H ₂ generation and outcomes, serpentinization was induced by injecting pH = 12.5 brine at 0.015 cm ³ /min (0.5 pore volumes per day) into an olivine sand pack (250 to <355 μm grain size) at 245°C generating, at minimum 76 and 89 mol% H ₂ at 35 and 57 d, respectively. Grain‐coating serpentine with radiating needles cemented the reacted sand grains. Importantly, pore space was maintained between the dissolving grains and the serpentine precipitates. Hence, reactivity continued as a result of fluid access to mineral surfaces, the large grain size, and the continuous injection of undersaturated alkaline fluids.

Image-based evaluation methods are a valuable tool for source rock characterization. The time and resources needed to obtain images has spurred development of machine-learning generative models to create synthetic images of pore structure and rock fabric from limited image data. While generative models have shown success, existing methods for generating 3D volumes from 2D training images are restricted to binary images and grayscale volume generation requires 3D training data. Shale characterization relies on 2D imaging techniques such as scanning electron microscopy (SEM), and grayscale values carry important information about porosity, kerogen content, and mineral composition of the shale. Here, we introduce RockFlow, a method based on generative flow models that creates grayscale volumes from 2D training data. We apply RockFlow to baseline binary micro-CT image volumes and compare performance to a previously proposed model. We also show the extension of our model to 2D grayscale data by generating grayscale image volumes from 2D SEM and dual modality nanoscale shale images. The results show that our method underestimates the porosity and surface area on the binary baseline datasets but is able to generate realistic grayscale image volumes for shales. With improved binary data preprocessing, we believe that our model is capable of generating synthetic porous media volumes for a very broad class of rocks from shale to carbonates to sandstone.