InterPore2026

Impacts of permeability heterogeneities on foam flow in porous media: uncertainty quantification and sensitivity analysis

20 May 2026, 15:35

1h 30m

Poster Presentation (MS19) Uncertainty-Aware Decision Support in Porous Media Applications Poster

Speaker

Berilo de Oliveira Santos (Federal University of Juiz de Fora)

Description

Foam injection in porous media has been extensively studied for its ability to improve sweep efficiency and gas conformance by mitigating nonlinear phenomena, such as gravitational segregation and viscous fingering. However, modeling foam flow remains a significant challenge, particularly in geologically complex formations, due to difficulties in accurately characterizing the permeability field and foam behavior. These challenges are closely linked to reservoir heterogeneity, specifically the uncertainties inherent in absolute permeability fields, which remain underexplored in the literature. This work [1] addresses this gap by performing uncertainty propagation studies to investigate the influence of permeability heterogeneity on two-phase foam flow. The methodology couples the Karhunen-Loève Expansion (KLE), to generate Gaussian random permeability fields, with Polynomial Chaos Expansion (PCE), a machine learning method for computationally efficient uncertainty propagation. This approach evaluates the impact of permeability variations across three scenarios (strong foam, weak foam, and foamless) on key quantities of interest, including pressure drop, breakthrough time, and cumulative water production. Simulations involve water (with surfactant) and gas injection into a fully water-saturated medium using previously validated software. Results derived from Uncertainty Quantification (UQ) and Sensitivity Analysis (SA) reveal that foam behavior is highly sensitive to the spatial correlation structures of permeability, yielding critical insights for process optimization. The integration of KLE and PCE establishes the first systematic framework for uncertainty propagation in foam flow, unveiling previously unexplored correlations and behaviors. These findings highlight the necessity of incorporating permeability uncertainties into computational models to enhance the reliability of subsurface flow applications, including resource recovery and carbon sequestration.

The current work was conducted in association with the R&D project ANP 20715-9, “Modelagem matemática e computacional de injeção de espuma usada em recuperação avançada de petróleo” (UFJF/Shell Brazil/ANP). Shell Brazil funds them in accordance with ANP’s R&D regulations under the Research, Development, and Innovation Investment Commitment.