Effects of Reaction Mechanisms and Differential Diffusion in Oxy-Fuel Combustion Including Liquid Water Dilution

摘要

The influence of chemistry and differential diffusion transport modeling on methane oxy-fuel combustion is analyzed considering different diluent characteristics. Analyses are conducted in terms of numerical simulations using a detailed description of the chemistry. Herein, different reaction mechanisms are employed to represent the combustion of methane. Simulations were performed with the computational fluid dynamics (CFD) code CHEM1D following different numerical setups, freely propagating flame, counter flow flame, and propagating flame in droplet mist reactors. The employed method is validated against experimental data and simulation results available in the literature. While the counter-flow flame reactor is exclusively used in the validation stage, different scenarios have been established for propagating flame simulations, as in single- or two-phase flow configuration. These comprehend variations in diluent compositions, reaction mechanisms, and different models to account for diffusion transport. Conducted investigations show that the choice for a specific reaction mechanism can interfere with computed flame speed values, which may agree or deviate from experimental observations. The achieved outcomes from these investigations indicate that the so-called GRI 3.0 mechanism is the best option for general application purposes, as a good balance is found between accuracy and computational efforts. However, in cases where more detailed information and accuracy are required, the CRECK C1-C3 mechanism demonstrated to be the best choice from the evaluated mechanisms. Additionally, the results clearly indicate that commonly applied simplifications to general flame modeling as the unitary Lewis number and mixture averaged approach strongly interfere with the computation of flame propagation speed values for single- and two-phase flows. While the application of unitary Lewis number approach is limited to certain conditions, the mixture averaged approach demonstrated a good agreement with the complex model for flame speed computations in the various tested scenarios. Such an outcome is not limited to oxy-fuel applications, but are straightly extensible to oxy-steam and air-blown combustion.