In the realm of energy and subsurface resource management, a team of researchers from the Earth Resources Laboratory at The University of Texas at Austin, the University of Tokyo, and Curtin University has developed a novel approach to enhance the accuracy of rock permeability predictions. Permeability, a measure of a rock’s ability to transmit fluids, is a critical factor in various energy-related applications, including carbon capture and storage, hydrocarbon extraction, and subsurface energy storage.
The researchers, led by Yaotian Guo and Fei Jiang, have proposed a convolutional neural network (CNN)-based upscaling method that integrates high-precision pore-scale permeability information into core-scale, low-resolution CT images. This approach aims to address the limitations of direct permeability estimation from low-resolution images, which often lack the detailed pore-scale structural features necessary for accurate predictions.
The team’s workflow involves partitioning a large core sample into smaller sub-core volumes. The permeability of these sub-core volumes is then predicted using CNNs, which are a type of artificial intelligence algorithm particularly well-suited to image analysis tasks. The upscaled permeability at the core scale is subsequently determined through a Darcy flow solver based on the predicted sub-core permeability map. This framework effectively incorporates small-scale heterogeneity, enabling accurate permeability upscaling from micrometer-scale pores to centimeter-scale cores.
One of the key aspects of this research is the examination of the optimal sub-core volume size. The team found that there is a balance to be struck between computational efficiency and prediction accuracy. Larger sub-core volumes may lead to more efficient computations but could compromise accuracy, while smaller sub-core volumes may improve accuracy but at the cost of increased computational resources.
The practical applications of this research for the energy sector are significant. Accurate permeability predictions can enhance the efficiency and effectiveness of carbon capture and storage projects, improve hydrocarbon extraction processes, and optimize subsurface energy storage solutions. By providing a rapid and reliable method for permeability assessment, this research can contribute to more informed decision-making and better resource management in the energy industry.
The research was published in the journal Fuel, a leading publication in the field of energy research. The study represents a significant advancement in the application of machine learning techniques to energy-related challenges, highlighting the potential of artificial intelligence to transform the energy sector.
This article is based on research available at arXiv.