Project
Modeling reacting flows in packed bed with the Lattice Boltzmann Method
Packed bed reactors are essential in chemical engineering for applications ranging from catalytic converters to industrial-scale chemical synthesis. Accurately modeling the complex interplay of fluid flow, heat transfer, and chemical reactions in these systems requires significant computational resources. Traditional computational fluid dynamics (CFD) approaches face challenges in capturing the multi-scale physics and geometry of porous media. The Lattice Boltzmann Method (LBM) offers an alternative kinetic-based approach that is particularly well-suited for complex geometries and parallel computing architectures. This project aims to develop and validate a high-fidelity LBM framework for simulating reacting flows in packed bed configurations, leveraging HPC resources to handle the millions of computational cells required for realistic domain sizes and chemical kinetics complexity.
Project Details
Project term
July 23, 2025–November 17, 2027
Affiliations
Otto-von-Guericke University Magdeburg
Institute
The Laboratory of Fluid Dynamics and Technical Flows
Principal Investigator
Methods
The simulation framework combines the Lattice Boltzmann Method for fluid dynamics with detailed chemical kinetics via Cantera integration. The D3Q19 lattice scheme is employed for three-dimensional flow simulations, with mass and energy transport equations solved on the same lattice structure. The chemical reaction mechanism is handled through direct coupling with the Cantera library, enabling accurate representation of multi-species reaction kinetics and thermodynamic properties. Special attention is given to boundary conditions for curved surfaces in the packed bed geometry, implemented using interpolation schemes to preserve accuracy at solid-fluid interfaces. The code is parallelized using domain decomposition strategies to efficiently utilize multiple compute nodes, with communication handled through optimized message passing protocols. Validation cases include canonical flow configurations and comparison with established benchmarks for reactive flows.
Results
A comprehensive reactive-flow simulation framework was developed within the ALBORZ framework by coupling the Lattice Boltzmann Method (LBM) with multispecies transport, thermodynamic modeling, and detailed chemical kinetics through Cantera. The framework supports detailed mechanisms, including GRI-Mech (53 species and 325 reactions), with consistent treatment of thermodynamic properties, transport coefficients, and chemical source terms.
The solver was validated progressively using one-, two-, and three-dimensional benchmark cases. The 1D premixed-flame simulations confirmed the correct coupling of flow, species transport, thermochemistry, and reaction kinetics, accurately reproducing flame structure, propagation speed, and species profiles for different chemical mechanisms. Three-dimensional Taylor–Green vortex simulations further verified the hydrodynamic accuracy of the framework, demonstrating appropriate energy-cascade behavior and viscous dissipation rates in agreement with analytical solutions.
Following these fundamental validations, the framework was applied to 2D and 3D packed-bed reactor configurations involving methane/air mixtures. Simulations were performed for different equivalence ratios and inflow conditions to investigate flame propagation, flame position, temperature distribution, heat-release patterns, and species transport within complex porous geometries. The results reproduced the characteristic interaction between the flame front and packed-bed structure and provide a basis for comparison with available experimental flame-position and thermochemical data. Ongoing simulations are focused on further assessing reactive-flow behavior in three-dimensional packed beds and establishing the predictive capability of the validated framework for complex porous-media combustion.
Discussion
The successful validation and initial simulations demonstrate that the LBM-Cantera framework is ready for application to realistic packed bed reactor configurations. The validated solver accurately captures both the fluid dynamics and detailed chemical kinetics, providing confidence in its predictions. The next major milestone involves extending the framework to three-dimensional packed bed geometries, which represents a significant computational challenge. Simulating flow through randomly packed spheres or structured particle arrangements with active chemistry requires resolving millions of lattice nodes while tracking dozens of chemical species at each point. This is precisely where HPC resources become essential, as the memory requirements and computational time scale rapidly with domain size and chemical complexity. The planned packed bed simulations will investigate how particle packing structure influences reaction efficiency, heat transfer characteristics, and pressure drop. These insights will be valuable for optimizing reactor designs in industrial applications. Beyond packed beds, the validated framework can be extended to other porous media systems such as catalytic converters, porous burners, and filtration systems where reacting flows occur in complex geometries.
Additional Project Information
DFG classification: 404-03 Fluid Mechanics
Software: Alborz
Cluster: CLAIX