Hierarchical controls on fracture sustainability in deep coal: Integrating stress-dependent permeability, microscale mechanics and ANN analysis
Author Identifier (ORCID)
Faisal Ur Rahman Awan: https://orcid.org/0000-0003-2394-0735
Abstract
Deep coal seams are increasingly considered for coalbed methane recovery and CO2 storage; however, injectivity and fracture sustainability at depth remain uncertain due to strong stress sensitivity and pronounced mechanical heterogeneity. In this study, an integrated framework combining stress-dependent permeability, microscale fracture-mechanics evaluation and artificial neural network (ANN)–based nonlinear integration is used to constrain fracture viability under deep in-situ conditions. Triaxial permeability measurements (5–20 MPa confining stress) reveal a near two-order-of-magnitude permeability reduction, indicating that bulk coal permeability alone is insufficient to sustain injectivity at depth. It should be treated as a first-order screening criterion rather than a performance indicator. Mineralogical and microstructural analyses demonstrate substantial microscale heterogeneity, which translates into geomechanically distinct zones with contrasting deformation and fracture responses. Nanoindentation-derived fracture mechanics show that fracture sustainability is spatially selective, with brittle mineral rich zones exhibiting higher resistance to irreversible deformation and greater fracture toughness, while clay-rich and organic-dominated zones deform readily and promote fracture closure or self-sealing. Correlation analyses reduce fracture behavior to a minimum controlling parameter set, showing that fracture toughness is primarily governed by hardness and contact depth, whereas elastic modulus and short-term stiffness play secondary roles. An ANN is subsequently applied to formalize the nonlinear coupling between these physically constrained parameters, providing a stable representation of fracture-relevant mechanical variability within the studied coal seam. Overall, the results demonstrate that fracture sustainability in deep coal is controlled by a limited subset of geomechanically favorable micro-zones that persist following stress-induced permeability collapse. This outcome cannot be resolved from permeability, fracture mechanics or data-driven modeling alone, but emerges from their combined application. The proposed framework offers a physics-guided basis for selective fracture design in deep coal reservoirs, with site-specific calibration required for broader implementation.
Keywords
artificial neural network, CO2 sequestration, deep coal seams, energy-based fracture toughness, microscale mechanical heterogeneity, stress-dependent permeability
Document Type
Journal Article
Date of Publication
8-1-2026
Volume
152
Publication Title
Gas Science and Engineering
Publisher
Elsevier
School
School of Engineering
Funding Information
The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China (Nos. 51874206, 52274222), the National Joint Fund for Regional Innovation and Development (No. U22A20167), the Natural Science Foundation of Shanxi Province (No. 202403021211211) and the China Scholarship Council (CSC) for providing the doctoral scholarship for the first author.
Copyright
subscription content
Comments
Butt, I. A., Liang, W., Rene, N. N., Ali, A., Yang, H., Awan, F. U. R., Shi, Q., & Liu, X. (2026). Hierarchical controls on fracture sustainability in deep coal: Integrating stress-dependent permeability, microscale mechanics and ANN analysis. Gas Science and Engineering, 152, 205947. https://doi.org/10.1016/j.jgsce.2026.205947