Modeling lifted jet flames in a heated coflow using an optimized Eddy dissipation concept model

Abstract

Moderate or intense low oxygen dilution (MILD) combustion has been established as a combustion regime with improved thermal efficiency and decreased pollutant emissions, including NOx and soot. MILD combustion has been the subject of numerous experimental studies, and presents a challenge for computational modeling due to the strong turbulence–chemistry coupling within the homogeneous reaction zone. Models of flames in the jet in hot coflow (JHC) burner have typically had limited success using the eddy dissipation concept (EDC) combustion model, which incorporates finite-rate kinetics at low computational expense. A modified EDC model is presented, which successfully simulates an ethylene-nitrogen flame in a 9% O2 coflow. It is found by means of a systematic study in which adjusting the parameters and from the default 0.4082 and 2.1377 to 3.0 and 1.0 gives significantly improved performance of the EDC model under these conditions. This modified EDC model has subsequently been applied to other ethylene- and methane-based fuel jets in a range of coflow oxidant stream conditions. The modified EDC offers results comparable to the more sophisticated, and computationally expensive, transport probability density function (PDF) approach. The optimized EDC models give better agreement with experimental measurements of temperature, hydroxyl (OH), and formaldehyde (CH2O) profiles. The visual boundary of a chosen flame is subsequently defined using a kinetic mechanism for OH* and CH*, showing good agreement with experimental observations. This model also appears more robust to variations in the fuel jet inlet temperature and turbulence intensity than the standard EDC model trialed in previous studies. The sensitivity of the newly modified model to the chemical composition of the heated coflow boundary also demonstrates robustness and qualitative agreement with previous works. The presented modified EDC model offers improved agreement with experimental data profiles than has been achieved previously, and offers a viable alternative to significantly more computationally expensive modeling methods for lifted flames in a heated and vitiated coflow. Finally, the visually lifted flame behavior observed experimentally in this configuration is replicated, a phenomenon that has not been successfully reproduced using the EDC model in the past.