Project
Flame stabilization in turbulent solid fuel combustion
Coal-fired power plants remain a major source of carbon dioxide emissions, yet they still provide a service the electricity grid depends on: dispatchable heat and power, available on demand regardless of whether the sun is shining or the wind is blowing. Substituting coal with cleaner solid fuels, without rebuilding existing plant infrastructure, could preserve this flexibility within a decarbonized energy system. This project investigated two candidate fuels for such a retrofit: biomass, fired under oxy-fuel conditions to enable carbon capture, and pulverized metallic iron, which releases no carbon dioxide at all and can be regenerated from its combustion product using surplus renewable electricity. Both fuels exhibit fundamentally different stabilization and conversion behavior than coal, and than each other, once burned inside a realistic, swirl-stabilized turbulent flame. Establishing how such a flame anchors itself, how hot individual reaction zones and particles become, and how completely the fuel converts is a prerequisite for safe and efficient operation at industrial scale. These quantities are not accessible everywhere within the flame through laboratory diagnostics alone, since the processes involve strongly turbulent, three-dimensional, particle-laden reacting flows with complex two-way coupling between gas and solid phases. Resolving this level of detail requires large-eddy simulations of a resolution and computational cost that only a high-performance computing system can deliver within a practicable turnaround time.
Project Details
Project term
July 1, 2025–June 30, 2026
Affiliations
TU Darmstadt
Institute
Institute for Simulation of reactive Thermo-Fluid Systems
Principal Investigator
Methods
The project employed large-eddy simulation coupled with Lagrangian particle tracking to compute turbulent solid-fuel flames, resolving the unsteady gas-phase flow field together with the trajectories, heating, ignition, and conversion of thousands of individual biomass or iron particles. Since fully resolved gas-phase chemistry, involving tens to hundreds of chemical species and reactions, is computationally prohibitive at this scale, combustion was represented through tabulated flamelet chemistry: one-dimensional flame structures were pre-computed for a wide range of conditions and stored in lookup manifolds accessed rapidly during the three-dimensional simulation. Each solid fuel required a dedicated conversion submodel, since the devolatilization and heterogeneous combustion pathways of biomass differ from the surface oxidation kinetics of iron. The simulated geometry matched a real laboratory-scale swirl burner, and simulation results were validated against dedicated velocity- and temperature-field measurements. Computational grids comprised several million cells, with each simulation distributed across hundreds of processor cores in parallel — feasible only on a shared high-performance computing system. To validate the tabulated-chemistry approach specifically for hybrid iron–methane flames, manifold-based simulations were additionally benchmarked against fully resolved detailed-chemistry simulations across a broad parameter range in a simplified flame configuration.
Results
In an earlier project phase, preceding the current reporting period, the first large-eddy simulation of a turbulent, laboratory-scale iron flame stabilized with a methane pilot flame was performed and validated against dedicated particle-tracking measurements. It demonstrated that the local availability of oxygen, itself governed by turbulent mixing, strongly controls the peak temperature reached by individual particles and their degree of conversion. This work was completed and published shortly before the present reporting period began.
During the current reporting period (July 2025 to June 2026), this foundation was extended along several lines. Large-eddy simulations of a biomass flame under air-firing and two oxygen-enriched, carbon-dioxide-diluted oxy-fuel atmospheres were completed and published, showing that flame topology and stabilization mode changed substantially with oxidizer composition, including whether a secondary reaction zone formed alongside the primary flame. Despite these topological differences, overall fuel conversion remained nearly invariant across conditions, since biomass devolatilizes readily even in oxygen-lean regions. The simulations agreed well with experimental diagnostics and identified an oxy-fuel composition that most closely reproduces air-firing behavior, relevant for plant retrofitting. Also within this period, a follow-up study on the iron flame showed that reducing the oxygen supply to part of the burner lowers peak particle temperatures but leaves a larger fraction of the iron incompletely oxidized — a trade-off between particle-temperature control and conversion efficiency that burner design must balance. A separate, more fundamental study, now submitted for publication, confirmed that the tabulated flamelet approach used in the large-scale simulations reproduces fully resolved detailed-chemistry results with high accuracy, corroborating the validity of all results above. Toward the end of this period, a new sub-project targeting a further metal fuel, aluminum, was initiated and remains at an early stage.
Discussion
Taken together, the results suggest that the overall large-eddy simulation framework, originally developed for pulverized coal, can be carried over to fundamentally different solid fuels. The fuel-specific building blocks within it, namely the particle conversion submodels and the tabulated-chemistry treatment, required further development. In particular, extending the tabulated flamelet chemistry to account for oxygen depletion by the particle phase was needed for the metal-fuel case. This extension was supported both by canonical flame configurations and by turbulent burner-scale simulations. That is encouraging for applying the framework to further fuels, though each will still require its own development and validation.
The two fuels studied so far appear to differ markedly in sensitivity. Biomass conversion seems comparatively robust to oxidizer composition, favoring flexible retrofitting. Iron combustion, in contrast, looks considerably more sensitive to local oxygen distribution, pointing to burner design and airflow control as important levers for constraining particle temperatures while sustaining conversion. This may be practically relevant, since high particle temperatures are associated with unwanted nanoparticle formation, though the link to actual nanoparticle yield still needs firmer establishment.
Looking ahead, the project intends to extend the framework to aluminum. It will also use additional computing time to resolve individual and small groups of particles in finer detail, refining the fuel-specific submodels underlying the large-scale simulations.
Additional Project Information
DFG classification: 404 Fluid Mechanics, Technical Thermodynamics and Thermal Energy Engineering
Software: OpenFOAM
Cluster: CLAIX
Publications
Pascal Steffens, Janik Hebel, Daniel Braig, Antje Vahl, Leon Loni Berkel, Sandra Schary, Hendrik Nicolai, Arne Scholtissek, Andreas Dreizler, Benjamin Böhm, Christian Hasse,
Exploring turbulent methane-assisted iron dust combustion: A combined experimental and numerical study of a 47 kW lab-scale combustor,
https://dx.doi.org/10.1016/j.fuel.2025.135205, April 2025
Antje Vahl, Pascal Steffens, Leon Loni Berkel, Christian Hasse, Hendrik Nicolai,
Large eddy simulations of swirl-stabilised gas-assisted oxy-fuel biomass flames under varying oxygen concentrations,
https://dx.doi.org/10.1016/j.fuel.2026.138602, August 2026
Thesis:
Emilija Petrosiute;
Characterization of Flame Behavior in Solid Fuel Particle-Gas Suspensions,
Bachelor Thesis, March 2025