Smart Coupled Flow-Geomechanical Upscaling Technique for Oil Sands

ABSTRACT: Interbedded shales play a critical role in the flow-geomechanics of reservoirs, influencing slope stability of open-pit mining, as well as steam chamber propagation and caprock integrity in steam-assisted gravity drainage (SAGD) operations used to extract bitumen from oil sands. However, conventional flow-geomechanics analyses often rely on oversimplified assumptions (homogeneity and isotropy) for interbedded shales in the Alberta oil sands. This is largely due to a limited understanding of the complex fluid flow and geomechanical interactions within these heterogeneous oil sand reservoirs. A robust and efficient geomechanical upscaling technique is essential for advancing the application of three-dimensional (3D) reservoir-scale geomechanical simulations that account for detailed geological heterogeneities. The objective of this study is to propose an innovative upscaling technique using a 3D convolutional neural network (3D CNN) that computes upscaled geomechanical properties directly from geological realizations. This approach significantly reduces the computational time of coupled fluid flow and geomechanical simulations while accurately reproducing the stress-strain behavior, volumetric strain, plastic strain and pore pressure changes of heterogeneous oil sands under partially drained condition during SAGD operation and open-pit mining of oil sands. 1. INTRODUCTION Interbedded shales, also termed as inclined heterolithic stratification (IHS), are widely distributed in the Alberta oil sands and create significant heterogeneities in both reservoir and caprock (Thomas et al. 1987). The interbedded shales act as flow barriers that hinder steam chamber development in SAGD operations. The long and continuous mud layers can form critical slip planes, leading to slope instability in open-pit mining of oil sands (Hein, 2017). Understanding and accurately modeling these geological heterogeneities are essential for efficient bitumen extraction from oil sands, necessitating advanced techniques to optimize production performance and ensure operational safety (Thomas et al., 1987). In the Wabiskaw Member and McMurray Formation of Alberta oil sands, coupled flow-geomechanical models help address challenges posed by geological heterogeneities like Inclined Heterolithic Stratification (IHS) (Darymple et al., 2007; Van den Berg et al., 2007; Martinius et al., 2011). Early upscaling techniques are focused on homogenization of geomechanical properties considering microscopic heterogeneities in a Representative Elementary Volume (REV) (Hill, 1963, Drugan et al., 1996). To further investigate the impact of heterogeneities across different scales, Multiscale Finite Element Methods (MsFEM) were developed to solve fine-scale problems within a coarse-scale framework without explicitly modeling the entire fine-scale system. (Hou et al, 1997, Efendiev et al., 2004). Fine-scale simulations, which perform detailed analyses on small-scale models and average the results (Durlofsky, 2005), have been further advanced by recent research employing machine learning and high-performance computing to address the challenge in multiscale simulations and to predict the constitutive geomechanical response of the REV at the microscopic scale (Ghaboussi et al., 1991, Hkdh, 1999, Koutsourelakis, 2007, Qu et al., 2021a, 2023, Wu et al., 2023).

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.