DI-UMONS : Dépôt institutionnel de l’université de Mons

Recherche transversale
(titres de publication, de périodique et noms de colloque inclus)
2005-04-04 - Colloque/Article dans les actes sans comité de lecture - Anglais - 5 page(s)

Lupant Delphine , Pesenti Barbara , Evrard Patricia , Lybaert Paul , "Numerical and experimental characterization of a self-regenerative flameless oxidation burner operation in a pilot-scale furnace" in European Combustion Meeting, Louvain-la-Neuve, Belgique, 2005

  • Codes CREF : Recherche énergétique (DI2290), Combustion (DI2212)
  • Unités de recherche UMONS : Thermique et Combustion (F704)
  • Instituts UMONS : Institut de Recherche en Energétique (Energie)
Texte intégral :

Abstract(s) :

(Anglais) Diluted combustion technologies, also called “flameless oxidation”, have been introduced in the early nineties as a very effective way to abate NOx production in high temperature gas fired furnaces using preheated combustion air. Besides NOx production, these techniques also modify flame characteristics as well as heat transfer inside the furnace; their application to industrial processes still requires basic experimental data and developments in modeling techniques in order to be able to predict furnace performance. The present study compares the heat transfer, temperature and species profiles measured in a pilot-scale furnace fired with flameless oxidation with values computed with a commercial CFD code. Tests are carried out on a ~200 kW natural gas FLOX REGEMAT® burner manufactured by WS-Essen (Germany). It is a compact, high efficiency, self-regenerative burner through which gas and highly preheated air are injected at high momentum into the combustion chamber. The entrainment of flue gases at high rate by air and gas jets produces a high level of dilution of the reactants before mixing. The burner fires a rectangular fiber-lined pilot furnace of about 3 m3, where the temperature is controlled by sliding 10 water cooled tubes along the side walls. Burner characterization is performed by measuring temperature and flue gas composition profiles in the furnace axis and at the chimney with a suction thermocouple probe connected to a gas analysis unit. Profiles of radiative heat flux incident on the roof are measured with 4 elliptical radiometers and CCD imaging of OH self-emission in UV is performed to show the evolution of the combustion zone within the furnace for various furnace temperature conditions. The burner and the pilot scale furnace have been modeled using FLUENT 6 with a 340,000 hexahedral cells grid. Two different approaches have been tested for turbulent combustion modeling (solving of a transport equation for mixture fraction with a presumed probability density function for the turbulence/chemistry interaction or infinitely fast and irreversible single step reaction with EBU to compute the average reaction rate). Thermal and prompt NOx are predicted with FLUENT in a post-processing computation. Both combustion models with default parameters give relatively similar temperature fields and heat transfer profiles. Temperatures are over-estimated in the near burner zone and both models fail to predict the step-shaped temperature profile measured in the furnace axis for the lowest furnace temperature conditions. The measured NOx emission level is very low, ranging from 3 to about 40 ppm when the furnace temperature varies from 1090 to 1330°C with air preheating temperature as high as 1000°C. NOx levels are underestimated by the models but the trends are rather correctly reproduced. However, even if the CFD performance is quite encouraging, industrial extrapolation cannot be envisaged due to the prohibitive computation time needed to represent a full-scale furnace. Furthermore, combustion models have to be improved, by including chemical kinetic, to better fit with the measured values. Work is also done to find new models able to reproduce the correct order of magnitude of NOx emissions.