Reshaping geomechanical characterization of unconventional reservoirs with digital rock physics and artificial intelligence

Unconventional reservoirs are increasingly important for oil and gas production, climate mitigation and subsurface engineering, demanding more refined geomechanical characterization. Current approaches are limited by poorly resolved engineering-scale parameters and weak integration of heterogeneous multi-source data into iterative model updating. Digital rock physics and artificial intelligence offer two routes to address these limitations. Digital rock physics builds digital representations of cross-scale rock heterogeneity and uses experimentally constrained simulations to upscale stiffness, strength, damage evolution and stress-dependent deformation into parameters for wellbore-, fracture- and reservoir-scale models. Artificial intelligence extracts geomechanical constraints from multi-source data, linking static characterization with dynamic reservoir responses to support iterative model refinement. These routes move unconventional reservoir characterization toward cross-scale parameterization, physics-constrained modelling and observation-updated prediction, strengthening the basis for safer resource development and complex subsurface engineering. Document Type: Perspective Cited as: Yan, D., Zhao, L., Zhang, F. Reshaping geomechanical characterization of unconventional reservoirs with digital rock physics and artificial intelligence. Advances in Geo-Energy Research, 2026, 21(2): 105-108. https://doi.org/10.46690/ager.2026.08.02