PI-SGF: Physics-Informed Swin-GeoFormer for Unsupervised Pore Aspect Ratio Prediction in Reservoir Characterization (opens in new tab)

Pore aspect ratio ( $\alpha $ ) is a key microstructural parameter that controls elastic wave propagation in reservoir rocks, yet it cannot be directly measured from well logs, and existing estimation methods rely on computationally expensive iterative inversion or pseudo-labels. This article presents physics-informed Swin-GeoFormer (PI-SGF), a physics-informed framework that couples log-sequence learning with the Berryman–Gassmann rock-physics model as a differentiable constraint, enabling f...