Author : Beibei Wang
Publisher :
ISBN 13 :
Total Pages : pages
Book Rating : 4.:/5 (19 download)
Book Synopsis Study of Flow Mechanisms in Shale Using CT Imaging and Data Analytics by : Beibei Wang
Download or read book Study of Flow Mechanisms in Shale Using CT Imaging and Data Analytics written by Beibei Wang and published by . This book was released on 2019 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: Due to the decline of current conventional oil and gas reservoirs, the development of unconventional resources has received great attention in recent years (World Energy Outlook 2012). Shale, formations that are considered as both source rocks and reservoirs, play a significant role in the USA's hydrocarbon production (EIA 2019). Hence, understanding the effective and efficient development of unconventional resources is of crucial importance. Nevertheless, there are still numerous technical challenges related to fluid transport in shale. The nanoporous system of shale formations has relatively low porosity and ultralow permeability that has considerable influence on fluid transport by advection and diffusion (Javadpour et al., 2007). Moreover, cracks and natural fractures are also very common in shale and they play a very important role in production. Natural cracks and fractures contribute directly to storage and permeability, and they can interact with hydraulic fracturing treatments (Gale et al., 2010). The heterogeneous pore and network system together with the significant variation in mineral composition raise challenges for the understanding of fluid transport through shale. Mechanistic understanding of fluid transport in shale reservoirs is crucial for future production forecasts and for better field planning and development. This research work bridges the gap in understanding the storage and transport mechanisms of unconventional resources. Various experimental, simulation and data analysis techniques were applied, as follows. First, simulation of adsorption properties using statistical modelling based on Grand Canonical Monte Carlo (GCMC) techniques for CO2 adsorption in clay systems was performed. Significant CO2 is predicted to adsorb to clay. Results from simulation and experiment are compared to further investigate the adsorption properties of gas shale and to predict the adsorbed phase densities as a function of temperature, pressure, and pore size. It was observed that the simulated CO2 adsorption for the clay is smaller compared to organic matter. This result shows the same trend as the experimental measurement. At 60 bar and 80 °C, the CO2 adsorption in a 2 nm pore in clay is around 2 mmol/cm3; while in the 2 nm pore in the organic matter, the CO2 adsorption is around 13 mmol/cm3. Second, we carried out experiments to probe liquid behaviour in shale samples by X-ray CT imaging. CT scans were taken continuously after injecting water and water tracer into the core. From the change of CT signal of the shale core over time as the water flows through the porous medium, the water flow path is visualized. From CT image analysis, when injecting water into the dry core, a water front was observed to move along the core over time. The CT signal of the entire core increased substantially after breakthrough indicating that water preferably flowed through larger pore space and then transported into the matrix. Third, following on the success of imaging liquid movement in shale, experiments were carried out to visualize and study liquid diffusion in sandstone, carbonate, and two shale samples. The diffusion study is designed to be purely concentration driven with no pressure difference applied to the system. An effective diffusion coefficient was calculated by fitting the experimentally measured concentration profile data and analytical solutions from Fick's law. Then, sample tortuosity was analyzed based on the effective diffusion. The sandstone and carbonate had tortuosities of 1.34 and 1.36, respectively, in agreement with literature. The shale samples had tortuosity in excess of 10 indicating substantial geometrical complexity of shale porous networks. Finally, a data-driven deep learning approach was developed to infer the permeability distribution of shale samples. Through analyzing flow images of the shale sample from CT scans, a convolutional neural network model was trained to calculate the average and local permeability of the sample. Compared to traditional permeability measurement and calculation, this method presents a local 3-D permeability map of the shale and provides valuable information to understand the nature of shale and their production capabilities.
