Furnace: Hydrogen Combustion

Description

This project investigates non-premixed hydrogen–air combustion in a lab-to-semi-industrial-scale furnace, with a particular focus on how the accuracy of viscosity and thermal-conductivity modeling affects predicted flame behavior. Rather than relying on default property models, the simulation pairs Sutherland's law for viscosity with kinetic-theory-based thermal conductivity, since correctly capturing these fluid properties is central to how well turbulence and chemistry couple in a reacting flow. The furnace geometry has separate fuel and air inlets: air enters at 0.01 kg/s and hydrogen at 0.0003 kg/s, giving a stoichiometric mixture (equivalence ratio of 1), with the fuel inlet centered 0.25 m from the furnace floor and nested within the air inlet. The outlet is set to atmospheric pressure, and the side walls lose heat to the surroundings by convection and radiation through a 5 cm steel shell, with a heat transfer coefficient of 16 W/m²K, an ambient temperature of 300 K, and a sky temperature of 271.2 K.

Methodology

The domain is discretized with a structured mesh built in ANSYS Meshing, and the non-premixed air/hydrogen mixture is handled through the Species Transport model. Turbulence is closed with the standard k–ε model, while the combustion itself is resolved through the eddy dissipation approach, with energy source diffusion and turbulence–chemistry interaction both active; this method assumes species conversion is fast relative to turbulent mixing, which holds well for atmospheric-pressure reactions in an open furnace geometry. Pressure–velocity coupling uses the SIMPLE algorithm, gradients are discretized with Least Squares Cell-Based, pressure with Second Order, and all other variables with Second Order Upwind, with under-relaxation tuned to bring the solution to convergence.

Analysis

The centerline results at the fuel inlet show a peak velocity of 1063 m/s and a peak temperature of 1860 K, with H2 mass fraction dropping to zero by the facing surface and H2O mass fraction rising from zero to a steady 0.23 after about 0.3 m. The reaction rate peaks at 0.23 kmol/m³/s just 2 cm from the inlet and falls to zero by 0.4 m, marking where the flame effectively completes. Volumetric temperature contours show a furnace-wide maximum of 2102.56 K, with higher average temperatures concentrated in the lower furnace, while isotherm planes spaced 0.125 m apart confirm a peak of 2080 K. Velocity is presented both at full scale and capped at 100 m/s to expose lower-speed flow structures that would otherwise be masked, and mass fraction contours on the furnace's symmetry plane trace how combustion species distribute through the chamber. Eddy viscosity fields illustrate the turbulent kinetic energy cascading into internal energy, tying the flow's turbulent structure directly to the high-temperature zones identified elsewhere in the results. Taken together, the results indicate that the eddy dissipation method, combined with the Sutherland/kinetic-theory property treatment, reproduces the expected physics of a fast, mixing-limited combustion process with good fidelity across velocity, temperature, species, and reaction-rate fields.