. Non-premixed and partially premixed syngas flames

Combustion in many practical devices involves non-premixed (diffusion) and partially premixed flames (Bozzeli etal., 1995). While there exist numerous studies of such flames with hydrocarbon fuels, relatively few investigations have appeared with syngas fuels. Giles etal. (2006) numerically studied the effects of N2, CO2, H2O, and CH4 addition on the structure and NOX characteristics of syngas diffusion flames in a counter flow burner. A diffusion flame in this burner is established by having two opposing jets being issued from two coaxial nozzles that are placed one above the other. Fuel is supplied from the bottom jet and air from the top jet, and the flame is established near the stagnation plane formed by the two jets. Diluents can be introduced through either or both the

Подпись: Figure 2.10.

image022

Effect of adding N2, H2O, and CO2 in the airstream on the peak NO mole fraction and flame temperature for a syngas (50%H2/50%CO)-air diffusion flame (Giles et al., 2006).

jets. Simulations were performed using the OPPDIF algorithm and the GRI-3.0 mechanism. The algorithm computes the flow field and flame by solving the governing equations for temperature, species mass fraction, and velocity field. The 2-D axisymmetric flow field is transformed into a 1-D problem by employing a similarity transformation. Results indicated that syngas non­premixed flames are characterized by relatively high temperatures and NOX concentrations, and require diluents to control NOX emissions. Figure 2.10 from Giles etal. (2006) depicts the effects of three diluents (N2, H2O, and CO2) added to the airstream on the peak flame temperature and NO mole fraction for a 50%H2/50%CO syngas flame.

As the amount of dilution is increased, the flame temperature decreases with a corresponding reduction in the peak NO, indicating that NO formation in these flames is primarily be due to the thermal mechanism. CO2 and H2O are more effective than N2 in reducing NO, with CO2 being the most effective diluent on a mole basis. Giles et al. (2006) also observed that the presence of methane in syngas even in small amounts opens the prompt NO route, and decreases the diluent effectiveness in reducing NOX. Other studies on non-premixed syngas flames include those reported by Hui et al. (2007) and Park et al. (2004).

There have also been few investigations on syngas partially premixed flames (Hui etal., 2007). A partially premixed flame (PPF) in a counter-flow burner is established by introducing air from the top nozzle and a fuel rich mixture from the bottom jet. The important parameters characterizing a counter-flow PPF include the strain rate, equivalence ratio (Ф), and fuel composition. Som et al. (2010) investigated experimentally and numerically the influence of pressure and fuel composition on the combustion and NOX emissions in syngas PPFs. Figure 2.11 from this study depicts images oftwoPPFs established at Ф = 6 and 16, and strain rate as = 35 s-1. For Ф = 6, which is just above the rich flammability limit of 50%H2/50%CO syngas-air mixture, the flame exhibits a typical double flame structure with a weak rich premixed reaction zone (RPZ) established very close to the fuel nozzle and a non-premixed reaction zone (NPZ) on the oxidizer side near the stagnation plane.

As Ф is increased, the RPZ moves away from the fuel nozzle. Consequently, for Ф = 16, the RPZ and NPZ are much closer to each other. The computed structures of four syngas PPFs in terms of the profiles of temperature and heat release rate are shown in Figure 2.12. Two of these flames correspond to the same conditions as those for flames depicted in Figure 2.11. Again, for Ф = 6.0, the flame structure is characterized by two spatially separated reaction zones, namely the RPZ

image023

Figure 2.11. Images of syngas (50%H2/50%CO)-air partially premixed flames established at Ф = 6 (Flame a) and Ф = 16 (Flame b) in a counter flow burner. The strain rate is 35s-1(Som etal, 2010).

image024

Figure 2.12. Computed flame structure in terms of temperature and heat release rate profiles for four syngas (50%H2/50%CO)-air partially premixed flames. The two flames at strain rate as = 35 s-1 are the same as those depicted in Figure 2.13, while the other two flames are at as = 50 s-1 and Ф = 6 and 16.

and NPZ, which are easily located by the two heat release rate peaks. The RPZ is very close to the fuel nozzle, which is in agreement with the digital images presented in Figure 2.11. For Ф = 16, the temperature peaks indicate a nearly merged flame structure. However, the corresponding heat release rate profiles indicate two distinct peaks that are close to each other. This is again consistent with the digital images in Figure 2.11. At lower strain rates (as = 35 s-1), flame temperatures are slightly higher due to longer residence time.

Подпись: Figure 2.13.

image026

Peak NO mole fraction plotted versus CO fraction in syngas for partially premixed flames established at Ф = 6 and different pressures (Som et al., 2010).

Som et al. (2010) further observed that for the conditions investigated, the RPZ is characterized by H2 oxidation, while the NPZ is characterized by the oxidation of both H2 and CO. This is in contrast to hydrocarbon PPFs, in which the fuel is partially oxidized to produce H2 and CO in the RPZ, and the oxidation of H2 and CO occurs in the NPZ. However, similar to hydrocarbon PPFs, as the pressure is increased, the distance between the two reaction zones decreases, while the flame temperature increases. The reader is referred to Figure 8 in Som etal. (2010) for further discussion of the flame structure at different pressures and syngas compositions. With regards to NO emission, results indicated that as the pressure is increased, the amount of NO formed first increases rapidly with pressure, but then levels off at higher pressures. This can mainly be attributed to the increase in flame temperature with pressure, which increases the thermal NO. In addition, the peak NO exhibits a non-monotonic variation with the H2 fraction in syngas, as shown in Figure 2.13. As the H2 fraction is increased, the peak NO first decreases and then increases. This can be attributed to the combined effects of thermal and re-burn mechanisms, as the syngas composition is changed.

The re-burn mechanism consumes NO through reactions NO + H + M ^ HNO + M and NO + O + M ^ NO2 + M, which become important at higher pressures and as the H2 fraction in syngas increases. However, when the H2 fraction exceeds a certain value, the peak NO starts increasing with the increase in H2 fraction, which is due to the effect of higher flame temperature, which increases thermal NO. The contributions of various NO formation routes are depicted in Figure 2.14 from Som et al. (2010) which plots the NO emission index with respect to pressure for two different syngas mixtures. The emission index is defined as the ratio of NO production rate to fuel consumption rate. As indicated in the figure, the N2H, NNH, and re-burn mechanisms become important at high pressures.

Ouimette and Seers (2009) reported an experimental investigation on syngas partially premixed jet flames. The effects of Ф, CO2 dilution, and H2/CO ratio on the flame structure and NOX were reported. Figure 2.15 from this reference presents images of syngas jet flames at different Ф. As expected, the flame length is strongly influenced by the level of partial premixing.

As Ф decreases from the non-premixed to premixed regime, the flame length decreases mono — tonically. This has important consequences for the emissions of NOX, greenhouse gases, and

image027

Figure 2.14. Variation of emission indices of total, thermal, prompt, N2O, NNH and reburn NO mechanisms with pressure for syngas partially premixed flames (Som et al., 2010).

other pollutants, since the flame length directly influences the reacting volume and residence time. In addition, images at 2.0 and 1.6 indicate the existence of two reaction zones, with the NPZ enveloping the RZP Regarding NOX, results indicated that EINOX first increases as Ф is increased from 1.0 to 1.6, then remains nearly constant for 1.6 < Ф < 3.85, and subsequently decreases slowly as Ф is increased to the diffusion limit (Ф ^ X). In addition, results indicated that increasing CO2 dilution reduces EINOX in the entire range of Ф, consistent with previous studies while increasing the H2/CO ratio reduces EINOX for Ф < 2.0, and has negligible effect for richer mixtures.