Comparative study of flame structures and NOx emission characteristics in fuel injection recirculation and fuel gas recirculation combustion system

Abstract

SUMMARYA numerical study with momentum-balanced boundary conditions has been conducted to grasp the chemical effects of added CO 2 to fuel-and oxidizer-sides on flame structure and NO emission behaviour in H 2 -O 2 diffusion flames with varying flame location. A reaction mechanism is proposed to show better agreements with experimental results in CO 2 -added hydrogen flames.Oxidizer-side dilution results in significantly higher flame temperatures and NO emission. Flame location is dramatically changed due to high diffusivity of hydrogen according to variation of the composition of fuel-and oxidizer-sides. This affects flame structure and NO emission considerably especially the chemical effects of added CO 2 . The present work also displays separately thermal contribution and prompt NO emission due to the chemical effects caused by thermal dissociation of added CO 2 in NO emission behaviour. It is found that flame temperature and the flame location affect the contribution of thermal and prompt NO due to chemical effects considerably in NO emission behaviour.

Evaluation of chemical effects of added CO2 according flame location

“…Therefore, in Figure 4 EVALUATION OF CHEMICAL EFFECTS OF ADDED CO 2 ACCORDING TO FLAME LOCATION CO formation for oxidizer-side dilution is caused by the increase of added CO 2 mole fraction since maximum H mole fraction is located on the oxidizer-side. However, the decrease for oxidizer-side dilution implies that further CO 2 addition plays a role of a diluent due to an excessively added CO 2 (Park et al, 2003c). For fuel-side dilution with added CO 2 mole fractions of 0.1 and 0.3 the maximum H mole fractions are located on the oxidizer-side and thus the fuel-side dilution may not contribute to the CO formation reaction so much.…”

“…Especially for the description of CO 2 -added flames Li and Williams (1999) took a different approach, in which each rate parameter was evaluated independently but might provide reasonable agreement within references cited for other mechanisms. Park et al (2003c) took the basic skeleton of the Miller-Bowman mechanism and most of the rate parameters with the specific reaction rate constants were taken with those in the Li-Williams mechanism, since the Miller-Bowman mechanism was less sensitive (Kim et al, 2002), and showed good agreement with the experimental data of Rrveit et al (2002). The present study also adopted the reaction mechanism of Park et al (2003c), which consists of 49 chemical species and 239 elementary reaction steps.…”

“…Park et al (2003c) took the basic skeleton of the Miller-Bowman mechanism and most of the rate parameters with the specific reaction rate constants were taken with those in the Li-Williams mechanism, since the Miller-Bowman mechanism was less sensitive (Kim et al, 2002), and showed good agreement with the experimental data of Rrveit et al (2002). The present study also adopted the reaction mechanism of Park et al (2003c), which consists of 49 chemical species and 239 elementary reaction steps. The governing equations are solved using a CHEMKIN-based code (Kee et al, 1989) and a transport-based one (Kee et al, 1994).…”

“…For either fuel-or oxidizer-side dilution with CO 2 the boundary conditions of equal velocities sometimes fail to converge. The global strain rate is, therefore, obtained considering the balance of momentum flux (Park et al, 2003c):…”

“…The study of Feese and Turns might pass over the fact that the strain rate varied while they maintaining the fixed nozzle exit velocities or the fixed mass fluxes. The explanation with a characterized strain rate (Park et al, 2003b) might ignore the difference in fuel-and oxidizer-side boundary conditions; that with flame location (Park et al, 2003c) described NO emission behaviour well with the propriety of flow residence time according to the distance from the stagnation point to the flame location, even if it is not a basic answer. This may imply that even chemical reaction effects should also be affected by relative flame location to the stagnation point.…”