| OCR Text |
Show magnitudes higher when compared to reaction rate in ignition or post-flame zones. These "small" main-flame areas are the driving force behind the flame structure of type-2 swirl-stabilized flames. As this study will show the reason for the success of E B U combustion models is that they predict those areas of intense combustion with reasonable accuracy. New gas-firing technologies for achieving increased heat transfer and ultra-low N O x emissions are based on entraining flue gas into the fuel and/or oxidiser jet before combustion occurs (Matsumoto et al., 1995; Wunning, 1996). Entraining flue gases increases the thermal ballast and decreases the volumetric reactivity of a gas pocket. These effects lead to lower peak flame temperatures and, due to the strong mixing effects, result in a more homogeneous distribution of temperatures in the process chamber when compared to burner-stabilised flames. The decreased volumetric reactivity emphasises the need to focus improvements of combustion models towards the ignition zones. It should be noted, that, although fundamental research on mathematical modelling of turbulent combustion is about to close the gap between the broad spectrum of length and time scales of turbulent flow and combustion chemistry, a total direct numerical simulation (DNS) for hydrocarbon combustion in practical high Reynolds number turbulent Nomenclatura All physical properties are considered to represent stationary, time-mean values. a exponent T K combustion model (dimensionless ) AArrh Arrhenius law pre-exponential factor (consistent) Amix mixing rate constant ( dimensionless ) b exponent TK combustion model (dimensionless ) CEa constant for T K combustion model ( consistent) Cr correction factor (dimensionless ) Ct correction factor ( dimensionless ) Ea activation energy ( Jlkmol) k turbulent kinetic energy (m2 s2 ) R reaction rate (kg/m3s ) R gas constant ( Jlkmol • K) T temperature ( K ) W molecular weigth (kg; kmol) X species mol fraction ( kmol/kmol) Y species mass fraction ( kg; kg) flow is not feasable in the near future (Bray, 1996). Hence, improved predictive methods to aid the design of industrial scale equipment in the gas industry have to rely on simpler and faster turbulence/chemistry models. Objectives The objective of this paper is to analyse local turbulent reaction rates from C F D predictions of a type-2 swul-stabilised 2 M W t natural gas-fired turbulent diffusion flame. The analysis focuses on in-flame zones where ignition is supposed to occur and where high turbulence flows enter low to medium range temperature regions (700 - 1200K). The goal of the study is to identify ways to improve the E B U combustion model by incorporating quasi-global Arrhenius-type combustion models which account for the temperature dependency of the combustion process and as well for the local turbulence structure. Background In the framework of the multi-fuel burner (MFB-1) project at the International Flame Reseach Foundation (IFRF), Umuiden, The Netherlands, Dugue at al (1991) conducted detailed in-flame measurements of temperature, gas composition and velocity on flame number 258 in die IFRF furnace No. 1 (Fig. 1). This flame is an unstaged, high N O x swirling natural gas flame of 2 M W thermal input. Greek Symbols a Arrhenius law specie exponent (dimensionless ) £ turbulent dissipation rate (m2 s* ) A/ local stoichiometry (dimensionless ) v species mol fraction (mol) p mixture density ( kg/m1 ) Subscripts Arrh by means of finite rate Arrhenius law model C//4 methane concentration CO carbon monoxide concentration CXHV hydrocarbon concentration EBU by means of the EBU combustion model ff frozen-flow field calculation H20 water concentration 02 molecular oxygen concentration TK by means of the TK combustion model 2 |
