Overview of Combustion in Power Plants

Combustion in power plants is the process of burning fossil fuels—such as coal, natural gas, or oil—to release thermal energy that is converted into electricity. Modern power plants must achieve high thermal efficiency while minimizing emissions of CO₂, NOₓ, SOₓ, and particulate matter. The combustion environment involves turbulent reacting flows, complex heat transfer (convection, conduction, and radiation), and rapid chemical kinetics across broad temperature and pressure ranges. Accurate modeling of these coupled phenomena is essential for burner design, fuel flexibility studies, and compliance with increasingly stringent environmental regulations.

Types of Combustion Systems

Power plant combustors vary widely: pulverized coal boilers, natural gas-fired burners, oil-fired furnaces, and fluidized bed reactors. Each system presents unique challenges. For example, pulverized coal combustion requires modeling of devolatilization, char oxidation, and ash deposition, while gas turbines firing natural gas demand fidelity in flame stability and lean blow-off predictions. Advanced concepts such as oxy-fuel combustion and chemical looping also rely on detailed CFD to assess feasibility and performance.

Why ANSYS Fluent for Combustion Modeling

ANSYS Fluent is a leading computational fluid dynamics (CFD) platform widely adopted in the power generation industry. Its solvers incorporate comprehensive physics models for turbulent reacting flows, making it well suited for simulating industrial-scale burners and furnaces. Fluent's ability to handle complex geometries, polyhedral meshes, and parallel computation allows engineers to resolve detailed flame structures without excessive computational cost.

Key advantages include:

  • Integrated reaction mechanisms for gas-phase, surface, and solid-fuel combustion
  • Multiple turbulence-chemistry interaction models (Eddy Dissipation, Finite-Rate/Eddy Dissipation, Laminar Flamelet, and Composition PDF Transport) (ANSYS Fluent Official Site)
  • Advanced radiation models (Discrete Ordinates, P1, and Surface-to-Surface) essential for high-temperature furnaces
  • Multiphase capabilities including Discrete Phase Model (DPM) for coal particles and Lagrangian sprays for liquid fuels
  • Pollutant formation modules for thermal and prompt NOₓ, N₂O, and soot

Rigorous validation against experimental data—from laboratory-scale flames to full-scale power plant measurements—has established Fluent as a trusted tool for combustion engineers.

Modeling Chemical Reactions in Fluent

Chemical Kinetics and Reaction Mechanisms

Accurate representation of chemical kinetics is central to combustion modeling. Fluent allows users to import detailed reaction mechanisms in CHEMKIN format. For natural gas combustion, mechanisms such as GRI-Mech 3.0 (with 53 species and 325 reactions) or DRM-22 offer good accuracy for predicting flame speed and NO formation. For coal, a combination of devolatilization models (single rate, two competing rates, or Chemical Percolation Devolatilization) and char combustion kinetics (intrinsic, effectiveness factor, or unreacted core) must be coupled. Users can also employ global reduced mechanisms when computational speed is prioritized, though with a trade-off in accuracy for pollutant predictions.

Turbulence-Chemistry Interaction (TCI)

In turbulent flames, the mixing rates of fuel and oxidizer strongly influence reaction rates. Fluent provides several TCI models:

  • Eddy Dissipation Model (EDM) – Assumes reactions are mixing-limited; suitable for fast chemistry in diffusion flames.
  • Finite-Rate/Eddy Dissipation Model (FR/ED) – Computes both Arrhenius and mixing rates, taking the minimum; appropriate for moderate Damköhler numbers.
  • Laminar Flamelet Model – Precomputes flame structure in mixture fraction space and accounts for flame stretch and quenching. Excellent for non-premixed and partially premixed gas flames.
  • Composition PDF Transport Model – Solves the joint probability density function of species and temperature using stochastic particles. Most accurate but computationally intensive; recommended for highly autoignitive regimes such as gas turbine combustors.

The choice of model depends on the fuel type, flame regime, and available computational resources. A typical approach for coal-fired boilers uses the EDM for gas-phase combustion coupled with DPM for coal particles, while natural gas burners often benefit from the flamelet model.

Radiation Heat Transfer

Radiation often dominates heat transfer in boilers, accounting for up to 90% of the thermal flux. Fluent's Discrete Ordinates (DO) model with spectral band options resolves radiative transport in participating media (CO₂, H₂O, soot). The Weighted Sum of Gray Gases (WSGG) model provides an efficient treatment of gas absorption coefficients. For pulverized coal flames, inclusion of particle radiation from char and fly ash is critical for predicting wall heat flux and flame temperature.

Setting Up a Combustion Simulation

Geometry and Mesh

A representative geometry of the combustion chamber—including burners, over-fire air ports, and flue gas outlets—is constructed in ANSYS DesignModeler or SpaceClaim. Mesh generation in ANSYS Meshing or Fluent Meshing focuses on capturing near-wall gradients and flame zones. For gas burners, a polyhedral mesh with local refinement in the flame region yields good compromise between accuracy and cell count. For coal boilers, the mesh must resolve the coal jet penetration and the recirculation zones that stabilize the flame.

Boundary Conditions

Inlet boundaries specify fuel composition, temperature, velocity or mass flow, turbulence intensity, and length scale. For liquid or solid fuels, particle size distributions and injection positions are defined within the DPM. Wall boundaries set heat transfer coefficients (convective, radiative) or temperatures, often based on historical data or coupled to external water-wall models. Outlet boundaries are typically pressure outlets with backflow temperature and species.

Solution Strategy

Combustion simulations are often initialized with a cold flow solution (without reaction) to stabilize the flow field. Then the energy equation and reaction source terms are activated gradually to avoid divergence. Under-relaxation factors for species and energy may need adjustment. Steady-state solutions are common for burner design, while transient simulations are required for analyzing ignition, flame propagation, or load changes.

Pollutant Formation Modeling

Environmental constraints drive the need for accurate prediction of NOₓ and SOₓ. Fluent's NOₓ model calculates thermal NO (Zeldovich mechanism), prompt NO (Fenimore mechanism), and fuel NO (from coal-bound nitrogen). N₂O formation can also be included for low-NOₓ burners. Soot formation is modeled using the Moss-Brookes or the one-step model (semi-empirical). The models use post-processed temperature and species fields from the main combustion simulation, or can be coupled to the flow field for two-way coupling. (Combustion Modeling in ANSYS Fluent – Ansys White Paper)

Case Studies in Power Plant Optimization

Gas Turbine Combustor Optimization

A 500 MW gas-fired combined-cycle plant used Fluent to redesign the combustor liner for reduced NOₓ. The simulation employed the non-premixed flamelet model with GRI 3.0 kinetics. By adjusting fuel staging and dilution air injection locations, the design achieved a 20% reduction in NOₓ while maintaining flame stability and exit temperature profile. The validated model helped avoid costly full-scale tests.

Pulverized Coal Boiler Retrofit

An aging coal plant faced derating due to slagging and high unburned carbon. Fluent simulations with DPM for coal particles and EDM for gas-phase combustion identified uneven coal feed distribution as the root cause. Equalizing the fuel lines and adjusting the burner tilt angle reduced unburned carbon from 8% to 2.5% and decreased slagging by optimizing the near-wall flame temperature. Annual fuel savings exceeded $2 million.

Oxy-Fuel Combustion for Carbon Capture

For oxy-fuel combustion (burning fuel in O₂/CO₂ atmosphere), Fluent's ability to model real-gas thermodynamics and modified kinetics was key. Researchers at a European test rig replicated the boiler behavior and predicted CO₂ purity above 95% after condensation. The simulation helped size the flue gas recirculation system and identify optimal oxygen concentration for stable combustion. (CFD Modeling of Oxy-Fuel Combustion – ResearchGate)

Advanced Topics

Multiphysics Coupling

Simulating a coal-fired boiler often requires coupling Fluent to other physics modules: stress analysis for tube wall temperatures in ANSYS Mechanical, or system-level models in ANSYS Twin Builder for plant transient dynamics. The coupling ensures realistic boundary conditions (e.g., wall heat flux from boiler steam cycle).

Uncertainty Quantification and Machine Learning

Recent work integrates Fluent with uncertainty quantification tools to assess the impact of fuel variability (e.g., coal rank or moisture) on emissions. Machine learning models trained on Fluent data can then replace CFD for real-time optimization. (Data-driven surrogate models for coal combustion – Journal of Cleaner Production)

Best Practices for Accurate Modeling

  • Mesh independence: Perform a grid sensitivity study with at least three mesh sizes. Monitor key outputs like temperature, CO, and NO mass fractions.
  • Kinetic mechanism verification: Validate against laminar flame speed or ignition delay data before using in turbulent simulations.
  • Inlet turbulence specification: Use experimentally measured or correlated turbulence intensity and length scales. Incorrect profiles can shift flame location.
  • Run steady-state first: For most industrial burners, steady RANS with a realizable k-ε model provides adequate accuracy for mean flow and temperature. Use LES or SAS only for transient phenomena like combustion instability.
  • Post-processing: Calculate global parameters (adiabatic flame temperature, thermal NO index) to check sanity. Compare temperature profiles against thermocouple or pyrometer readings.

Limitations and Challenges

Despite its capabilities, Fluent combustion modeling has limitations. Detailed chemistry models can be slow for large meshes; adaptive mesh refinement (AMR) is not yet fully automated for reacting flows. Soot models remain semi-empirical and may fail in sooting flames at high pressure. Radiation property models like WSGG are calibrated for specific gas mixtures and may need validation for oxy-fuel or hydrogen-rich fuels. Users must also be aware of the assumptions behind each TCI model—using EDM for highly strained flames can underpredict NOₓ by ignoring finite-rate effects.

Conclusion

Modeling chemical reactions and combustion in power plants using ANSYS Fluent provides engineers with a robust framework to improve efficiency, reduce emissions, and ensure safe operation. By carefully selecting turbulence-chemistry interaction models, radiation schemes, and kinetic mechanisms, practitioners can achieve predictive accuracy that translates into tangible plant improvements. As computational resources continue to grow and new models emerge (for hydrogen combustion, ammonia co-firing, and carbon capture), Fluent will remain an indispensable tool in the transition toward cleaner power generation. (The Future of Combustion Modeling – ANSYS Blog)