Simulation- and experiment-driven development of an optimized numerical strategy to design low-NOx gas burners

Ortolani, Andrea and Campobasso, Sergio (2025) Simulation- and experiment-driven development of an optimized numerical strategy to design low-NOx gas burners. PhD thesis, Lancaster University.

Text (2025OrtolaniPhd)
2025OrtolaniPhd.pdf - Published Version
Available under License Creative Commons Attribution-NonCommercial-NoDerivs.
Download (6MB)

Abstract

This thesis presents a comprehensive study of fluid dynamics and pollutant formation in non-premixed industrial gas burners through advanced computational fluid dynamics (CFD) modeling and experimental validation. Two main objectives guide the research: developing accurate and computationally affordable models for nitrogen oxides (NOx) formation, and investigating the complex flow field within industrial burners. Validation studies were conducted using the Sandia Flame D experiment. Turbulent combustion was modeled using the Flamelet Generated Manifold (FGM) approach, with both Reynolds-averaged Navier-Stokes (RANS) and Large Eddy Simulation (LES) methods. While both showed good agreement with experimental data, LES better captured streamwise mixture fraction variations. RANS tended to overpredict mixing rates, causing premature peak predictions for temperature and CO/OH mass fractions. Three NO modeling approaches were evaluated: a baseline FGM scalar transport model (M1), a herein proposed non-adiabatic variant (M2), and a simplified ”sum of contributions” model (M3). M2 improved prediction accuracy over M1, while M3 provided the best match for thermal NO predictions. To address the challenges posed by the burner’s complex geometry, a cold flow analysis was performed using three RANS turbulence models. The Reynolds Stress Model (RSM) with the baseline omega equation offered the best balance between accuracy and efficiency. Hybrid grid refinement strategies proved beneficial in regions of flow separation. Subsequent unsteady RANS simulations of the industrial burner under varied thermal loads (41.7–87.5 kW) showed strong agreement with measured flue gas temperatures (within 0.65%). M3 predicted NO emissions most accurately (54 ppm vs. 55 ppm measured) at full power. Thermal NO dominated at high loads, while prompt and N2O-intermediate pathways were more relevant at low loads. M1 and M2 overpredicted emissions, with M2 performing slightly better. This work advances burner design by providing validated modeling strategies for flow and emissions, enabling optimized performance with lower environmental impact.