SYNGAS AND BIOGAS COMBUSTION AND EMISSIONS

Syngas can be produced using a variety of feedstocks and conversion processes, particularly using gasification, as discussed in the preceding section. On the other hand, biogas is generally produced by anaerobic digestion or fermentation of biodegradable materials in an oxygen-free environment (http://en. wikipedia. org/wiki/Biogas). There is significant potential for using syngas and biogas fuels for transportation and power generation. Both of these fuels represent a clean and renewable energy source, and offer great flexibility in their production and utilization. The next two sections provide an overview of the fundamental and applied research dealing with these fuels.

2.1.2 Syngas combustion and emissions

Syngas is a renewable energy source with wide flexibility in feedstock and conversion processes. Most of the harmful contaminants and pollutants can be removed in the post-gasification process prior to combustion. Moreover, technologies for its production and utilization are fairly developed, as several IGCC plants are currently operational around the world. There is also significant interest in using syngas as a transportation fuel. In addition, the use of syngas in fuel cells, such as solid oxide fuel cells, through the reforming of hydrocarbons and other routes is also being explored (Kee et al., 2005; 2008).

Considerable work has been reported on syngas combustion and emissions (Lieuwen, 2009; Cheng et al., 2009). Fundamental studies have focused on various aspects, including the devel­opment of thermo-transport and kinetic models, and examining the ignition and combustion characteristics. A major challenge identified in these studies is due to a substantial variation in its composition and heating value. This requires that the syngas combustion and emission behavior be analyzed for a wide range of composition. Thus, properties such as adiabatic flame tempera­ture, laminar burning velocity, flammability limits, flame stability, extinction, and blowout need to be determined for a wide range of syngas composition. This presents challenges while design­ing syngas combustors, requiring optimization for locally available fuels. As indicated in Table

2.1 (Kee et al., 2005), the main components in syngas are H2 and CO, with varying amounts of diluents, such as CO2, H2O, and N2, as well as CH4 in small amounts. Consequently, previous studies on syngas combustion have considered several representative compositions. Table 2.2 lists an average syngas composition, based on the values in Table 2.1.

Fundamental combustion properties can be analyzed by starting with the stoichiometric mass balance for a syngas-air mixture as:

xCO + (1 — x)H2 + a(O2 + 3.76N2) ^ xCO2 + (1 — x)H2O + dO2 + 3.76aN2

Table 2.1. Representative compositions (in terms of percentage of mole fractions) and related properties of syngas utilized in various IGCC plants; from Kee et al. (2005).

Syngas

PSI

Tampa

El

Dorado

Pernis

Sierra

Pacific

ILVA

Schwarze

Pumpe

Sarlux

Fife

Exxon

Singapore

Motiva

Delaware

PIEMSA

Tonghua

H2

24.8

37.2

35.4

34.4

14.5

8.6

61.9

22.7

34.4

44.5

32.00

42.30

10.3

CO

39.5

46.6

45.0

35.1

23.6

26.2

26.2

30.6

55.4

35.4

49.50

47.77

22.3

CH4

1.5

0.1

0.0

0.3

1.3

8.2

6.9

0.2

5.1

0.5

0.10

0.08

3.8

CO2

9.3

13.3

17.1

30.0

5.6

14.0

2.8

5.6

1.6

17.9

15.80

8.01

14.5

N2 + Ar

2.3

2.5

2.1

0.2

49.3

42.5

1.8

1.1

3.1

1.4

2.15

2.05

48.2

H2O

22.7

0.3

0.4

5.7

39.8

0.1

0.44

0.15

0.9

LHV [(Btu/ft3]

209

253

242

210

128

183

317

163

319

241

248

270.4

134.6

LHV [kJ/m3]

8224

9962

9528

8274

5024

7191

12492

6403

12568

9477

9768

10655

5304

Tfuel F/C

570/300

700/371

250/121

200/98

1000/538

400/204

100/38

392/200

100/38

350/177

570/299

338/170

H2/CO ratio

0.63

0.8

0.79

0.98

0.61

0.33

2.36

0.74

0.62

1.26

0.65

0.89

0.46

Diluent

Steam

n2

N2/Steam

Steam

Steam

Steam

Moisture

H2O

Steam

h2o/n

n2

n/a

Equivalent LHV [Btu/ft3]

150

118

113*

198

110

200

*

116

150

129

134.6

Equivalent LHV [kJ/m3]

5910

4649

4452

7801

4334

7880

4600

5910

5083

5304

*Always co-fired with 50% natural gas.

Table 2.2. Average composition and standard deviation based on syngas mixtures listed in Table 2.1.

Syngas constituent

Average [% vol]

Standard deviation [% vol]

H2

31.0

14.9

CO

37.2

11.0

CH4

2.2

2.9

CO2

12.0

7.7

N2 + Ar

12.2

19.7

H2O

7.8

14.1

Table 2.3.

Heating values and adiabatic flame temperatures of various syngas mixtures.

H2 mole fraction

CO mole fraction

Mol. weight [kg/kmol]

Heating value [kJ/kg]

Heating value [kJ/kmol]

Adiabatic flame temp

(Ф = 10) [K]

0

1

28.0

10100.5

282814.0

2394.2

0.1

0.9

25.4

11145.3

283090.6

2385.1

0.2

0.8

22.8

12428.3

283365.2

2381.6

0.3

0.7

20.2

14041.3

283634.3

2379.3

0.4

0.6

17.6

16130.2

283891.5

2377.8

0.5

0.5

15.0

18942.6

284139.0

2376.9

0.6

0.4

12.4

22932.9

284368.0

2377.8

0.7

0.3

9.8

29036.9

284561.6

2379.3

0.8

0.2

7.2

39539.4

284683.7

2381.6

0.9

0.1

4.6

61871.7

284609.8

2385.1

1

0

2.0

141794.1

283588.2

2386.7

Here x defines syngas composition in terms of the mole fraction of CO, a, is related to the equiv­alence ratio Ф (Ф = AFstoichimetric/AFactual withAF = m Air/m Fuel = air to fuel ratio) through the relation a = 1/(2Ф), and d represents the excess O2 (for Ф < 1.0), given by d = (1 — Ф)/(2Ф). Air is assumed to contain 21% O2 and 79% N2 by volume. The above equation can easily be modified to include the presence of diluents in syngas. The syngas heating value can be deter­mined from the standard enthalpies of formation of reactant and product species (Turns, 2011). Table 2.3 lists the heating values of various syngas mixtures. For comparison, the heating val­ues of methane (representative of natural gas) on mass and volume basis are 55,500 kJ/kg and 888,000 kJ/kmol, respectively. Thus, the volumetric heat release rate from syngas combustion is low compared to those for methane. There are other such differences between the chemical and physical properties of syngas and natural gas. This presents challenges in replacing natural gas by syngas in existing combustion devices. Table 2.3 also lists the adiabatic flame temperatures (Tad) of various syngas-air mixtures at Ф = 1.0.

The variation of Tad with Ф for different syngas mixtures is plotted in Figure 2.5. The equi­librium temperature (Tad) was computed using the EQUILIBRIUM algorithm in CHEMKIN software (Chemkin, 2007). The algorithm is based on the application of the first and second laws of thermodynamics. As indicated in Figure 2.5, Tad is nearly independent of the CO fraction in syngas. However, diluents, such as CO2, H2O, and N2, can be used to modify its value.

Ignition of a fuel-air mixture is often characterized in terms of ignition delay time (tign), which has been measured using a variety of devices, including shock tube (Petersen et al, 2007), rapid compression machine (RCM) (Walton et al, 2007) and constant volume (or constant pressure) combustor.

The ignition delay also represents an important target for the development and validation of reaction mechanisms. The computations of tign are often performed using a homogeneous reactor

image007

Figure 2.5. Computed adiabatic flame temperature versus equivalence ratio (Ф) for three different syngas mixtures.

model (Aggarwal, 2011). Davis et al. (2005), Li et al. (2007) and others have reported such mechanisms for syngas oxidation. The GRI-3.0 mechanism (Smith web-link), which includes the oxidation chemistry of C1-C3 species, has also been employed. The homogeneous reactor model is based on the mass and energy conservation equations for a transient, spatially homogeneous system containing a gaseous reacting mixture. Figure 2.6 from Dryer (2008) summarizes the measured and predicted ignition delay data for different syngas mixtures reported by various researchers.

Laminar flame speed or burning velocity represents another fundamental property of a fuel — air mixture. It is of critical importance with regards to flame spread, stabilization, flashback, and blowout in practical systems. In IGCC premixed burners, the problem of flashback and combustion instability represents a major challenge to the designer, especially due to the wide variation in fuel composition. Similarly, it is an important parameter for designing and optimizing the syngas-powered spark ignition (SI) engines, where backfire and inadequate mixing time due to rapid flame propagation represent important issues. The laminar flame speed and its response to stretch are also fundamental to the analysis of premixed turbulent flames. In this context, turbulent flame speed (ST) is another important property for the combustor design, as it has direct influence on important operational issues, such as flame blow-off, flashback, and combustion instability.

Numerous studies have been reported concerning laminar premixed syngas flames. The pri­mary objective of these studies is to determine the effects of various parameters, such as syngas composition, diluents, temperature, and pressure, on the laminar flame speed, flame stability, and emissions. Laminar burning velocities for H2-CO mixtures have been measured using different systems, including flat flame burner (Yan et al., 2011), bunsen burner (Natarajan et al., 2007), counter flow burner (Vagelopoulos et al., 1998), and spherically expanding flames (Prathap et al., 2008). Simulations have often been performed by considering a one-dimensional con­figuration and employing the PREMIX algorithm (Kee et al., 1993) in CHEMKIN software. Multi-dimensional flame simulations have also been performed using various algorithms (Briones et al., 2008). The computations are based on the solution of mass, momentum, species, and energy conservation equations, along with appropriate models for thermodynamic and transport proper­ties. Such properties include standard enthalpy of formation, viscosity, thermal conductivity, and diffusivity of each species. The number of species depends upon the particular kinetic mechanism employed to model the fuel oxidation chemistry. The above set of equations is closed by using

image008

Figure 2.6. Ignition delay times of various syngas and hydrogen mixtures under different pressure and temperature conditions. Filled and open circles correspond to strong and weak ignition events, respectively. All experimental data have been normalized to 20 atm assuming p-1 proportionality. Lines correspond to ignition delay predictions using the Li et al. mechanism at 20 atm; the solid line corresponds to the syngas mixture used in shock tube experiments (Li etal., 2007).

an appropriate equation of state. The numerical algorithms used for solving these equations have employed different approaches, such as finite-difference and finite-volume schemes. An adaptive grid refining of the computational mesh is often used, based on the first and second derivatives of the dependent variables. Further details can be found in the FLUENT user’s guide (2005). Important results from these studies are summarized below: •

Flame a; 50% со [1] 50% Fij

Подпись: 210
image010
image011
Подпись: Equivalence Ratio
Подпись: McLean et al. [1994]
Подпись: so
Подпись: 100
Подпись: Fiamo B! w5%CO • 5% H2

image017Vo ume Percent of CO in Fue

Figure 2.7. Measured and predicted laminar burning velocities for syngas-air mixtures. Variation of laminar flame speed with equivalence ratio Ф for Flames A and B (a), and with CO fraction in syngas at Ф = 2.0 (b) (Mclean etal., 1994).

extinction, turbulent flame propagation, flame stabilization, blowout, and transition to deto­nation. The classical approach yields the following relationship between the stretched flame speed and stretch rate (Mueller et al., 1999):

SL = SL — LaK

Here SL and S£ are the stretched and unstretched flame speeds, respectively, K the stretch rate, and La the Markstein length. Note that S°L corresponds to the burning velocity of a freely propagating planar flame discussed earlier. The flame stretch refers to the rate of change of

image018

Figure 2.8. Measured and predicted laminar flame speeds for various H2/CO spherically expanding premixed flames (Bouvet et al., 2011).

the flame surface area, which may be due to flame curvature, unsteadiness, and flow non­uniformity or hydrodynamic stretch (Bouvet et al., 2011). By determining SL as a function of K through measurements or computations, both SL and La can be obtained. Figures 2.7 and 2.8 from Bouvet et al. (2011) present such data for spherically expanding H2/CO flames. Results in Figure 2.7 for the unstretched flame speed are consistent with those presented earlier. The variation of Markstein length with Ф (Fig. 2.8) indicates that these flames are prone to thermo-diffusive instability under lean conditions, since La becomes negative for Ф < 1.

As discussed by Kishore et al. (2011), this instability is related to the non-unity Lewis number (Le) for stretched flames, with Le > 1 and Le < 1 corresponding to stable and unsta­ble situations, respectively. Similar behavior has been observed by Pratap et al. (2008) and Kishore et al. (2011). Further, previous studies have shown that the presence of H2 in syn­gas increases the flame propensity for instability, while that of CO has the opposite effect. However, the overall instability is predominantly determined by H2 rather than by CO.

• Since syngas typically contains significant amounts of CO2 and H2O, and N2, it is important to examine the effects of these diluents on syngas combustion and emissions. Moreover, dilution is often used to lower the flame temperature and thereby limit NOX emissions. The effects of various diluents on laminar flame speed and stability have been reported by several researchers (Das et al., 2011; Sun et al., 2007; Law, 2006; Burke et al., 2006; Pratap et al., 2008; Kishore et al., 2011). A general observation is that the addition of these diluents decreases the laminar burning velocity due to the increase in heat capacity and the decrease in heat release rate. For a given amount of dilution, the effect is more pronounced with CO2 and H2O dilution compared to that with N2 dilution, mainly due to different heat capacities. The addition of a diluent also shifts the location of peak laminar burning velocity to leaner mixtures (Kishore et al., 2011). Some studies have also observed that the CO2 and H2O addition can affect the combustion chemistry and modify the syngas combustion characteristics (Das et al., 2011). For example, Das et al. (2011) observed that the laminar flame speed varies non-monotonically with H2O addition for CO rich mixtures, but decreases monotonically with H2O for H2-rich mixtures.

• Laminar burning velocity and cellular stability of flames burning other biomass-derived gaseous (BDG) fuels have also been investigated (Burbano et al., 2011). Such studies have con­sidered BDG fuels consisting of varying amounts of H2, CO, CH4, CO2 andN2.Yanetal. (2011) determined unstretched laminar burning velocities for four different BDG mixtures using a per­forated flat flame burner. Vu et al. (2011) reported laminar burning velocities and Markstein lengths for spherically expanding flames for three different BDG mixtures. The PREMIX 1- D algorithm was used for computing the corresponding burning velocities in these studies.

image019

Figure 2.9a. Measured Markstein Lb length versus equivalence ratio Ф for 50/50% H2/CO spherically expanding premixed flames (Bouvet et al., 2011).

A representative result from Vu et al. (2011) is depicted in Figure 2.9a and Figure 2.9b, which plots the unstretched burning velocity versus Ф for three BDG-air mixtures.

As indicated, the revised GRI-3.0 mechanism provides much closer agreement with mea­surements, especially under rich conditions. In the revised mechanism, rate constants of key reactions were modified based on the data in Davis et al. (2005) and Li et al. (2007). The Markstein lengths extracted from measurements for the three BDG-air flames were found to be negative, indicating a propensity for cellular instability. In addition, it was observed that the propensity increases and decreases with H2 and CH4 addition, respectively, and remains essentially unchanged with CO addition.

• There have been relatively few investigations on emissions from premixed syngas flames, although extensive data have been reported for the hydrocarbon flames. While it is impor­tant to consider both soot and NOX emissions from hydrocarbon flames, only NOX formation is relevant in syngas flames. NOX formation in hydrocarbon flames is essentially due to four mech­anisms, namely the thermal (Zeldovich), the prompt (Fenimore), N2O, and NNH mechanisms (Das et al., 2011; Vu et al., 2011; Briones et al., 2007). Thermal NO involves the following reactions: O + N2 ^ N + NO, andN + O2 + NO + O, andN + OH ^ NO + H. Here the first reaction is the rate limiting step, and becomes significant at high temperatures due to its high activation energy. Prompt NO formation is initiated through the reaction CH + N2 ^ NCN (or HCN) + H (or N). Thus the prompt mechanism is absent in syngas flames, since it is directly linked to hydrocarbon combustion chemistry, which produces a CH radical from acetylene. The prompt NO, however, may be important for syngas mixtures containing CH4. The N2O — intermediate mechanism involves N2 + O + M ^ N2O + M as the initiating reaction, with subsequent NO formation occurring through reactions such as N2O + H ^ NO + NH and N2O + O ^ NO + NO. This route is found to become important for lean mixtures and high pressures. Finally, the NO formation through NNH route involves reactions: N2 + H ^ NNH and NNH + O ^ NO + NH (Guo et al, 2007). Ding et al. (2011) investigated the extinc­tion and emission behavior of lean premixed syngas flames in a counter-flow configuration. Numerical simulations were performed using the OPPDIF algorithm in CHEMKIN and the Davis Mechanism (Davis, 2005). It was observed that the NO in these flames was formed predominantly through the NNH and N2O intermediate routes. The contribution of thermal NO was small due to the low flame temperatures. In addition, increasing the CO fraction in syngas was found to increase the amount of NO formed.