Coupled Lattice Boltzmann Modeling Framework for Pore-Scale Fluid Flow and Reactive Transport

Area of Science:

• Geochemistry
• Computational Fluid Dynamics
• Pore-Scale Modeling

Background:

• Accurate simulation of fluid flow and reactive transport in porous media is crucial for understanding subsurface processes.
• Existing models often face limitations in flexibility and efficiency when coupling fluid flow and geochemical reactions.
• Pore-scale modeling offers detailed insights into multiphase flow and reaction dynamics.

Purpose of the Study:

• To propose a novel modeling framework integrating Lattice Boltzmann Method (LBM) for fluid flow and PHREEQC for reactive transport.
• To enhance modeling flexibility and efficiency in simulating complex geochemical reactions within pore-scale geometries.
• To develop an AI-based optimization workflow for model parameter tuning.

Main Methods:

• Coupling of a parallel Lattice Boltzmann solver with the PHREEQC geochemical reaction solver.
• Development of multiple flow and reaction cell mapping schemes for seamless integration.
• Validation through benchmark numerical experiments including single-phase flow, diffusion, calcite dissolution, and surface complexation.
• Implementation of an AI-based optimization workflow for model parameterization.

Main Results:

• The coupled model demonstrates high flexibility, enabling complex reactions in desired cells with efficient data communication.
• Simulations accurately capture flow, diffusion, and reactions in complex pore-scale geometries.
• Validation results show good agreement with analytical solutions, experimental data, and other simulation codes.
• The AI-based optimization workflow provides fast and reliable results for surface complexation models.

Conclusions:

• The proposed coupled LBM-PHREEQC framework offers a powerful and flexible tool for pore-scale reactive transport studies.
• The developed mapping mechanism effectively handles complex geometries and coupled processes.
• AI-driven optimization significantly enhances the capacity and efficiency of the modeling framework.