Speaker
Description
Machine learning is increasingly being explored as a surrogate for pore-scale multiphase flow modelling in porous media, yet a clear understanding of how different model classes perform under realistic flow conditions remains limited. In particular, it is still unclear which modelling choices are most suitable for capturing interfacial dynamics, geometry–flow coupling, and temporal evolution at the pore scale.
In this presentation, we investigate representative machine learning models for pore-scale multiphase flow prediction using datasets generated from lattice Boltzmann method (LBM) simulations on micro-CT-based pore geometries, together with experimental observations from pore-scale flow imaging. The study focuses on physically meaningful cases that retain access to velocity fields, phase distributions, and geometric information, enabling controlled evaluation of model behaviour under well-defined physical settings.
We compare several classes of machine learning models, including convolutional neural networks operating on voxelised domains, graph-based approaches, and autoregressive temporal models. Performance is assessed in terms of prediction accuracy, stability under multi-step rollout, sensitivity to pore geometry, and generalisation across different settings. Beyond aggregate error metrics, we also examine qualitative failure modes, such as loss of interfacial sharpness, accumulation of long-term drift, and reduced robustness during rapid interface rearrangements.
By combining controlled simulation data with experimentally motivated examples, this work aims to provide a more practically grounded view of machine learning for pore-scale multiphase flow. The results highlight both the promise and the current limitations of these approaches, and help clarify how data-driven models can be used alongside conventional pore-scale simulation and experimental analysis.
| Country | UK |
|---|---|
| Student Awards | I would like to submit this presentation into the Earth Energy Science (EES) and Capillarity Student Poster Awards. |
| Acceptance of the Terms & Conditions | Click here to agree |
