Computational Modeling of Turbulent Ethanol Spray Flames in a Hot Diluted Coflow

Abstract

The phenomenon of turbulent spray combustion occurs in all industrial furnaces that consume liquid fuels. It is essential that a furnace is capable to have high efficiency and performance while the pollutant emissions meet the stricter national and international regulations. One of the recently proposed solutions to improve the aforementioned conditions is flameless spray combustion. In this process the combustion of fuel is done with oxidizer diluted with recirculated exhaust gases, which results in a lower peak temperature and more distributed reaction zone, so the NOx produced in conventional burners is highly reduced. Hence, it can be a promising approach in order to increase efficiency and decrease pollutant emission. Because flameless combustion is a novel concept, it should be fully investigated and optimized before being applied to large-scale cases. Due to the considerable costs of experimental tests, numerical simulation is used more and more to predict the performance of the flameless furnaces before utilizing them in a real case. The objective of this study is to develop and validate computational models for flameless spray combustion base on a validation study using the Delft Spray-in-Hot-Coflow (DSHC) flame. The burner has been designed to mimic the flameless oxidation of light oils. The properties of DSHC ethanol flames are computed by using a combination of CFD models for turbulence, chemistry and dispersed multiphase flow. The results are validated by comparison with the available experimental data for gas-phase velocity and temperature as well as droplets statistics. Furthermore, the experimental data are available for another case in which air is used instead of air diluted with exhaust gases. This case is also simulated to validate the models. This study employs a Reynolds-Averaged Navier Stokes (RANS) simulation approach. The combination of different turbulence and combustion models is investigated while the standard Lagrangian spray model is kept the same. The steady flamelet and the Flamelet Generated Manifold (FGM) models are two combustion models that are validated. It is shown that the FGM model can predict flame structure such as double-reaction region, lift-off, etc. better than the steady flamelet model. Moreover, it was observed that the Reynolds stress and the realizable k-epsilon models show similar results while the standard k-epsilon model performs differently. The effects of other models and parameters are also investigated. The results show that radiative heat transfer, secondary atomization and coalescence of droplets as well as buoyancy have negligible effects on the DSHC ethanol flame. Still, the effects of complete two-way coupling between two phases (liquid and gas) are demonstrated to be important.