Work

Detailed CFD study of air–fuel, oxygen-enriched and partial oxy-fuel combustion with optimized fuel savings in an industrial reheating furnace

Year: 2026

Type: article

Abstract: Industrial reheating furnaces face increasing demands for higher efficiency and lower emissions, yet the impact of oxygen-enriched and oxy-fuel combustion on furnace performance is still insufficiently understood on an operational scale. Previous studies often applied fixed burner power distributions across oxidizer conditions, which neglected zone-specific heating requirements and led to deviations from the intended thermal heating behavior of the products. To address this gap, a thermodynamic redistribution method was developed to adjust the zonal fuel input to the oxidizers studied while maintaining the original furnace characteristics and heating curves of the products. A numerically inexpensive three-dimensional RANS model was then coupled with a transient bloom model, enabling consistent representation of both furnace flow and bloom heating and confirming the method. The procedure was validated against detailed furnace measurements, including bloom heating curves, furnace temperature, flue gas compositions, NO x emissions, and wall temperatures. Results showed that enrichment to 25 vol% O 2 reduced fuel demand and CO 2 emissions by 4.8%, but more than doubled NO formation, whereas partial substitution with oxy-fuel in one combustion zone offered comparable savings while keeping NO x emissions low. Radiation remained the dominant heat transfer mechanism, and overall temperature fields were largely unaffected. Comparison with CFD confirmed that the developed thermodynamic approach reproduces global power balances reliably, demonstrating its robustness and applicability to a wide range of oxidizer–fuel configurations in industrial furnaces. To incorporate the different combustion modes consistently, combustion-dependent coefficients were introduced for the gas absorption coefficient calculation, while burner-type-dependent values were applied for the turbulent Schmidt and Prandtl numbers.

Source: Applied Thermal Engineering

DOI: 10.1016/j.applthermaleng.2026.130276