Chemical Kinetics in Support of Syngas Turbine Combustion
Department of Energy Grant DE‐FG26‐05NT42544 Project Period 08/01/05‐07/31/06
Frederick L. Dryer Mechanical and Aerospace Engineering Princeton University Princeton NJ 08544-5263
University Coal Research Contractors Review Meeting Pittsburgh Marriot Center City Pittsburgh PA June 6, 7 2006
Research Motivation and Objectives
Motivation: Syngas turbine combustion conditions span a range of pressures, temperatures, and compositions atypical of those more generally encompassed by fundamental chemical kinetic validation and combustion experiments. More accurate computational design tools will require chemical kinetic models to be used either directly or as a skeletal model from which lower dimensional representations for CFD can be derived. Objectives Further investigate fundamental chemical kinetic characteristics of CO/H2/Air/diluent/contaminant mixtures at pressures, temperatures, and diluent concentrations of CO2 and H2O typical of syngas combustion in gas turbines. Further evaluate third body collisional properties of major diluents, carbon dioxide and water, in the H + O2 + M = HO2 + M reaction. Extend our own prior work in developing and validating kinetic models for the CO/hydrogen/oxygen/contaminant systems. Provide additional experimental data using a Variable Pressure Flow Reactor (VPFR) for mixture oxidation. Analyze the individual and interactive behavior of specific elementary and subsets of elementary reactions at the conditions of interest.
Mechanism Development
Glarborg et al. (1996) Mueller et al. (2000)
Sulfur Chemistry
(15 species; 67 rxns)
CO/H2/O2/NOX/SOX Mechanism
(60 species; 192 rxns)
Mueller et al. (1999b)
Allen et al. (1997)
Nitrogen Chemistry
(22 species; 97 rxns)
Li (2004) Li et al. (2004)
CO/H2/O2/NOX Mechanism
(45 species; 125 rxns)
H2/O2 Submechanism
(10 species; 19 rxns)
Mueller et al. (1999a)
CO/H2/O2 Mechanism
(13 species; 28 rxns)
Westbrook and Dryer (1977); Yetter et al. (1991); Kim et al., (1994)
Elementary rate constants, Thermochemistry, and Validation comparisons
C1 Mechanism Development
Marinov (1999) Li (2004) Li et al (2005)
Acetaldehyde/Ethanol Chemistry
Held and Dryer (1998)
Ethanol Mechanism
Li et al. (2004), (2006)
Methanol Chemistry
C1 Mechanism
Li et al. (2004)
H2/O2 Submechanism
Mueller et al. (1999a)
CO/H2/O2/NOx/SOx Mechanism
Mueller et al. (2000)
Available on the web at http://www.princeton.edu/~combust/
C1 Mechanism Used in This Work
The mechanism consists of 85 elementary reactions among 21 species, and is based on the CH3OH/O2 mechanism of Held and Dryer (1998) and new H2/O2 mechanism of Li et al. (2004). Revisions encompass recently published kinetic and thermochemical information, while continuing to predict both new experiments and the experimental targets investigated by the original mechanism. It is developed in a hierarchical manner: CH3OH CH2O CO H2/O2 At each level, the sub-mechanism is tested against a wide range of experimental data: • Species time histories in flow reactors • Ignition delay data from shock tubes, etc. • Laminar premixed flame speeds • Other data (stirred reactor, burner-stabilized flame, etc.) • New validations as they become available
Key C1Mechanism Refinements
Part I: CO + OH = CO2 + H
This reaction is the main pathway to convert CO to CO2 and is responsible for a major fraction of the energy release derived in hydrocarbon oxidation Recent theoretical calculations predict higher rates than experimental measurements at low to intermediate temperature range The temperature-dependent sensitivity analysis of Zhao et al. (2005) demonstrates that the laminar flame speed of H2/CO oxidation systems is most sensitive to this reaction at 300-1900 K, depending on H2/CO ratio The C1 mechanism uses a new, weighted least squares fit of all of the experimentally measured rate constants available in literature.
k = 2.23 ×105 T1.89 exp(+
583 ) T
Weighted Fit for CO + OH = CO2 + H
10 Yu (1996) Ravishankara et al. (1983) Westerberg et al. (1973) Wooldridge et al. (1994,1996) Vandooren et al. (1975) Lissanski et al. (1995) Yu (1996) Troe (1998) Senosiain et al. (2003) Yu (1991) Zhu et al. (2001) Joshi & Wang (2006) Present 583 5 1.89
(cm mol s ) k × 10
-1
8
-1
6
-11
3
4
2
k = 2.23 ×10 T
exp(+
T
)
0 0.4 0.8 1.2 1.6
-1
2
2.4
1000/T (K )
Key C1Mechanism Refinements
Part II: HCO + M = H + CO + M
This reaction is the main pathway generating CO during the high temperature combustion of hydrocarbons. The temperature-dependent sensitivity analysis of Zhao et al. (2005) demonstrates that the laminar flame speed of hydrocarbon combustion systems is most sensitive to this reaction at 13002000 K, which is above the temperature range of recent experimental studies of this reaction (Friedrichs et al., 2002). Extrapolation of Friedrichs et al. causes difficulties in reproducing flame speed and flow reactor results for numerous hydrocarbons In the present study, a new rate correlation was developed by a weighted least squares fitting of literature experimental data
k = 4.75 ×10 T
11
0.66
7485 exp(− ) T
Weighted Fit for HCO + M = CO + H + M
10
13
k (cm mol s )
10
10
10
3
7
10
4
10
1
Ahumada et al. (1972) Baldwin et al. (1972) Bowman (1970) Browne et al. (1969) Campbell et al. (1978) Cherian et al. (1981) Cribb et al. (1992) Friedrichs et al. (2002) Hidaka et al. (1993) Hochnadel et al. (1980) Krasnoperov et al. (1999) Pearson (1963) Schecker et al. (1969) Timonen et al. (1987) Wang et al. (1973) Westbrook et al. (1977) Timonen et al. (1987) Friedrichs et al. (2002) Present
-1
-1
0.4
0.8
1.2
1.6
2
-1
2.4
2.8
3.2
1000/T (K )
Full Set of Updated Kinetic Parameters
Reaction rate coefficients:
• H2/O2 sub-mechanism: Li et al. (Int. J. Chem. Kinet. 2004) • CO + OH = CO2 + H: this study – weighted least squares fitting of experimental results in literature • HCO + M = H + CO + M: this study – weighted least squares fitting of experimental results in literature • CH2O decomposition: Friedrichs et al. (Int. J. Chem. Kinet. 2004, 36, 157) • CH2O + H = HCO + H2: Irdam et al. (Int. J. Chem. Kinet. 1993, 25, 285) • CH2O + HO2 = HCO + H2O2: Eiteneer et al. (J. Phys. Chem. A 1998, 102, 5196) • CH3OH decomposition reactions: GRI-MECH 3.0 (1999)
Thermodynamic data:
• OH: Ruscic et al. (J. Phys. Chem. A 2002, 106, 2727) • HO2: Ruscic et al. (J. Phys. Chem. A 2006, In press) • CH3: Ruscic et al. (J. Phys. Chem. A 1999, 103, 8625) • CH2OH: Johnson and Hudgens (J. Phys. Chem. 1996, 100, 19874)
Literature CO Experiments Used for Validation
Method Shock Tube Laminar Premixed Flame Flow Reactor Source Gardiner et al. (1971) Dean et al. (1978) McLean et al. (1994) Huang et al. (2003) Yetter et al. (1991) Kim et al. (1994) Mueller et al. (1999) Mixture CO/H2/O2/Ar CO/H2/O2/Ar CO/H2/air CO/H2/N2/air CO/H2O/O2/N2 CO/H2O/O2/N2 CO/H2O /O2/N2 T (K) 1400– 2500 2000 – 2850 298 298 1033 960 – 1200 1038 P (atm) 0.15– 0.3 1.2 – 2.2 1 1 1 1.0 – 9.6 1.0 – 9.6 φ 0.40 1.6– 6.1 0.5 – 6.0 0.7– 1.4 0.4 – 1.4 0.3 – 2.1 1.0
Literature CH2O Experiments Used for Validation
Method Source Dean et al. (1980) Buxton and Simpson (1986) Shock Tube Hidaka et al. (1993) Eiteneer et al. (1998) Friedrichs et al. (2002) BurnerVandooren et al. Stabilized (1986) Flame Hochgreb and Dryer (1992) Flow Reactor Scire (2002) Mixture CH2O/O2/Ar CH2O/Ar CH2O/O2/Ar CH2O/O2/Ar CH2O/Ar CH2O/O2 CH2O /O2/N2 CH2O /H2O/O2/N2 T (K) 1935 – 2150 1750 – 2100 1240 – 1950 1440 – 2120 955 – 975 300 945 – 1095 850 – 950 P (atm) 1.1 – 1.3 0.6 – 3.5 1.5 – 2.9 0.9 – 2.3 0.3 – 1.8 0.03 1 1.5 – 6.0 φ pyrolysis – 0.67 pyrolysis pyrolysis – 4.0 pyrolysis – 6.0 pyrolysis 0.22 0.013– 1.74 ~ 0.005
Representative Test Cases
Flame Speeds and 1 Atm Pressure Flow Reactor
200 Laminar Flame Speed (cm s ) CO/H = 50/50
2 2 -1
1
Mole Fraction (%)
150
CO/H = 95/5
0.8 0.6 0.4
CO2
100
50
CO 0.2 O2 0
0 0 1 2 3 4 5 6 7 Equivalence Ratio
0.02
0.04
0.06 Time (s)
0.08
0.1
Initial Conditions:
298 K, 1 atm, CO/H2/air mixture
Initial Conditions:
CO = 0.92%, H2O = 0.59%, O2 = 0.32% with balance N2 at 1034 K and 1 atm
Symbols: laminar flame speed data of McLean et al. (1994) _____ : Present model
Symbols: flow reactor data of Yetter et al. (1991) _____ : Present model
Representative Test Cases
Flame Speed Comparisons cont’d
250 200 Flame Speed (cm s )
-1
120 Flame Speed (cm s-1)
90
150 100 50 0 0 20 40 60 80
60
30
0
100
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
% CO in Fuel
Equivalence Ratio
Comparison with stoichiometric CO/H2/air mixture data at 1atm and 298 K (Mclean et al., 1994)
Comparison with reformer gas data (28% H2, 25% CO, and 47% N2) at 1atm and 298 K (Huang et al., 2004).
Representative Test Cases
Shock Tube Ignition and VPFR
1000
1.2 1 1 atm 2.4 atm 3.5 atm 9.6 atm
Ignition Delay Time ( s) µ
τ1 100 τ2
CO Mole Fraction (%)
0.8 0.6 0.4 0.2
10 0.35
0
0.4
-1 1000/T (K )
0.45
0.5
0
0.2
0.4
0.6 Time (s)
0.8
1
1.2
Initial Conditions:
H2 = 0.049%, O2 = 1.01%, CO = 3.28% with balance Ar τ1 and τ2 are defined as the time when [CO2] reaches 2.4×1016 and 8 ×1015 molecule/cm3, respectively. Symbols: shock tube data of Dean et al. (1978)
_____
Initial Conditions:
CO = 1.01%, H2O = 0.65%, O2 = 0.52% with balance N2 at 1038 K and 1.0 atm; CO = 1.01%, H2O = 0.65%, O2 = 0.50% with balance N2 at 1038 K and 2.4 atm; CO = 0.99%, H2O = 0.65%, O2 = 0.49% with balance N2 at 1038 K and 3.5 atm; CO = 0.99%, H2O = 0.65%, O2 = 0.49% with balance N2 at 1040 K and 9.6 atm Symbols: VPFR data of Mueller et al. (1999) _____ : Present model
: Present model
Computational Singular Perturbation (CSP)
(Lam, 1993) dz Chemical kinetic ODE system: = g (z ) dt
~ z = T y1 y 2 ... yn where
(
)
T
− state variable vector
g – vector of rates defined by reaction mechanism yi – species mass fractions (n total) ~ T - normalized temperature
At any given time, t, the rate vector can be “perturbed” with respect to time, i.e.
dg(z) dg( z ) and J = J • g(z ) , to define a local Jacobian matrix, J = , dt dz
can be decomposed ,
− J = VΛV 1
n +1 i =1
where V
∴ g (t + ∆t ) ≈ ∑ f i v i exp(λi ∆t )
mode amplitude (importance)
Λ - diagonal matrix containing eigenvalues
= (v1 v 2 ...v n +1 ) - matrix of eigenvectors
mode timescale may be complex (complex conjugate pairs)
CSP (Cont’d)
Mode classification: Re(λi) < 0 – stable (decaying) mode Re(λi) > 0 – unstable (explosive) mode – associated with ignition Im(λi) ≠ 0 – oscillatory mode; complex conjugate modes are transformed in the analysis into two real modes with the same damping coefficient Re(λi) and oscillatory frequency |Im(λi)| and a phase shift of π (Liu et al. 2001). Mode makeup (Participation Index):
g = ∑ rj S j
j =1
J
rj - jth reaction rate ∂g Sj = - vector of effective stoichiometric coefficients ∂rj
P j = rj V −1 • S j norm - Normalized vector of contributions of jth reaction for
individual modes Participation Index values present the importance of reactions promoting (positive) and inhibiting (negative) the thermal evolution (ignition) of the system.
CSP (Cont’d)
CSP advantages: More “direct” analysis of kinetic ODE system as opposed to typical sensitivity analysis Identifying and following the explosive modes allows determination of the factors controlling ignition unambiguously In conjunction with flux analysis, more complete interpretation can be derived Implementation: An in-house software package coupled with CHEMKIN library for postprocessing of SENKIN outputs Unlike most prior implementations, temperature is included in the variable vector (important for thermal feedback) Eigenvalue analysis is performed using LAPACK’s facilities
CSP – Different Experimental Systems
Rapid Compression Machine (RCM) Analyses
RCM’s can generate high pressure ignition information~ 20 to 40 atm
• As originally developed (Kazakov et al. 2006), the CSP analysis only treated constant volume systems (e.g. shock tubes) High pressure data obtained in Rapid Compression Machines (RCM) (e.g., Mittal and Sung, 2006) show heat losses need to be included in modeling these systems Sung and co-workers correct for heat losses by introducing a time-dependent volume expansion term after the piston reaches TDC This type of treatment effectively changes the rate vector g from its constant-volume counterpart A newly developed version of the CSP code can now implement these volumechanging systems
•
160 H2/CO/O2/N2/Ar - Pc = 50 bar; TC = 1,044 K 9.375/3.125/6.25/18.125/63.125 (%) tign = 1.35 ms Experiment Model - constant volume Model w/ volume expansion
•
120 Pressure (bar)
80
•
40
•
0 -5 -4 -3 -2 -1 0 1 Time (ms) 2 3 4 5
Results - Rapid Compression Machine Ignition
50
Mixture Composition:
Pressure (atm)
40 30
H2/CO/O2/N2/Ar = 6.25/6.25/6.25/18.125/63.125
20 10
0 0.02
0.024
0.028 Time (s)
0.032
0.036
H2O2+OH=>HO2+H2O -- (R19) H2O2+H=>H2O+OH -- (R16) HO2+O=>O2+OH -- (R12) CO+O(+M)=>CO2(+M) -- (R20) H+CO+M=>HCO+M -- (-R24) CO+OH=>CO2+H -- (R23) HO2+H=>2OH -- (R11) HO2+H=>H2+O2 -- (R10) O+H2=>H+OH -- (R2) HO2+OH=>H2O+O2 -- (R13) H2+OH=>H2O+H -- (R3) CO+HO2=>CO2+OH -- (R22) 2HO2=>H2O2+O2 -- (R15) H2O2(+M)=>2OH(+M) -- (R15) H+O2(+M)=>HO2(+M) -- (R9) HO2+H2=>H2O2+H -- (-R17) H+O2=>O+OH -- (R1) -0.2 -0.1
31.509 ms 33.021 ms 34.506 ms 35.006 ms 35.422 ms Pc = 30 bar Tc = 1,011 K tign = 5.422 ms Mixture make-up (%): H2/CO/O2/N2 6.25/6.25/6.25/18.125 in balance Ar
Analysis of typical conditions (above) studied by Mittal et al. (2006) in an RCM Mittal et al. identify that CO+HO2=CO+OH (R22) is primarily responsible for disagreement of C1 model with experiment (R22) is only contributory during chemical induction phase (initial radical pool growth) Variation of ignition properties with CO/H2 ratio comes primarily from competition of (R22) and H2+HO2=H2O2+H (R17) The following reactions contribute to describing heat release: • HO2 + OH = H2O + O2 (R13) • HO2+ H = OH + OH (R11) • H + O2 + M = HO2 + M (R9) • H2 + OH = H2O + H (R3) • H2 + O = H + OH (R2) • H2O2 + M = OH + OH + M (R15) • CO + OH = CO2 + H (R23) (R22) is likely much slower that previously estimated.
0 0.1 0.2 Participation Index
0.3
Results – High Pressure Shock tube ignition
500 400 Mole Fraction (ppm) 300 200 100 0 1100 1200 1300 Temperature (K) 1400 1500 P = 300 bar, φ = 1 H2/CO/O2/Ar CO O2 CO2 296.6
(a)
Pressure (atm) 296.4
Po = 300 bar To = 1,355 K H2/CO/O2 160/450/320 ppm in Ar
(b)
296.2
296.0 0 0.0002 0.0004 Time (s) 0.0006 0.0008
H2O2(+M)=>2OH(+M) -- (R15) HO2+H=>2OH -- (R11) CO+OH=>CO2+H -- (R23) HO2+O=>O2+OH -- (R12) HO2+OH=>H2O+O2 -- (R13) 2HO2=>H2O2+O2 -- (R14) HO2+H=>H2+O2 -- (R10) CO+O(+M)=>CO2(+M) -- (R20) H2+OH=>H2O+H -- (R3) CO+O2=>CO2+O -- (R21) CO+HO2=>CO2+OH -- (R22) HO2+H2=>H2O2+H -- (-R17) O+H2=>H+OH -- (R2) H+O2=>O+OH -- (R1) * 1/3 H+O2(+M)=>HO2(+M) -- (R9) * 1/3 -0.2 -0.1 0 Participation Index 0.150 ms 0.200 ms 0.250 ms 0.300 ms 0.315 ms P = 300 bar T = 1,355 K
(c)
High pressure shock tube data of and analysis for Sivaramakrishnan et al.(2005) Present C1 model compares favorably with these high pressure shock tube experiments Dilute character of these studies emphasizes the importance of HO2 + OH = H2O + O2 (R13) Analysis at the above conditions and mixture composition shows no importance of (R22) Analysis for these P and T conditions with mole fractions of prior RCM experiments identifies same relative importance of reactions shown in RCM experiments at lower P and T. Under all conditions the importance of collision efficiencies in H + O2 + M = HO2 + M (R9) is important to both ignition and heat release events
0.1
Collision Efficiencies
Collision Rate (cm3 molecule-1 s-1)
Michael et al (2002) measured chemical rates and estimated collisional rates of various species for H+O2+M=HO2+M Water is a much more effective collider than other combustion species Are the collisional characteristics of strong and weak colliders interactive in mixtures?
CR (cm3/mol/s) = A Tn From Michael et al.. (2002) Species He Ne H2 O2 H2O Ar N2 A 1.9437×10-10 1.3121×10-10 5.7633×10-10 1.9923×10-10 4.2451×10-9 1.6733×10
-10
2.5 10
-9
2 10
-9
1.5 10
-9
he Ne H2 O2 H2O Ar N2
1 10
-9
n 0.18036 0.17441 0.14711 0.16539 -0.11109 0.17565 0.11154
5 10
-10
0 0.5
1
1.5
1000/ T
2
2.5
3
3.5
3.3311×10-10
Assessment Method
Assess impact of possible interactions rather than develop a rigorous theory • Use an in house RRKM/Master code modified to accommodate mixtures of bath gases to estimate interactions • Apply a simple one-dimensional Master Equation • Assign loose transition state using properties of resulting dissociation (H & O2). • Evaluate the high pressure limit using the rate of recombination (R-1) and the equilibrium constant. • Use Michael et al. (2002) collision rate estimates (fitted with power curves) to input collision rates into the modified RRKM/Master code. Maximum energy considered in calculation was 130 kcal/mole above initial reactant (numerical equivalent of infinity) Determine all density and sum of state evaluations using an energy grain size of 1 cm-1 Solve the master equation on a grid with 5 cm-1 spacing.
Collision Energy Transfer
Use conventional “exponential down” model for collisional energy transfer.
<∆Edown> (cm )
1800
N2
1400
-1
H2O
Choose < ∆ Edown > for water and nitrogen to fit corresponding low-pressure limiting collision rate expressions for pure bath gases. Fit expressions for < ∆ Edown > as linear functions of temperature for final RRKM/master calculations for mixtures.
1000
600
200 500
1000
1500
2000
T (K)
Low Pressure limits for Mixtures at 1000 K
1.2 10
6
Linear scaling RRKM/Master
The low pressure limit of mixtures of water and nitrogen based upon linear interpolation vary less than 4% from the RRKM Master result at about 10% water in the mixture. For purposes of syngas combustion, the collisional efficiencies of mixtures even at high concentrations of water and/or carbon dioxide can be reasonably predicted based linearly on relative molar fractions.
k (cm mole s )
-1
8 10
5
3
-1
0
4 10
5
0
0
0.2
0.4
0.6
2
0.8
1
Mole fraction of H O
4
Deviation from llnear scaling (%)
3
2
1
0
0
0.2
0.4
0.6
2
0.8
1
Mole fraction of H O
H2/O2/NOX Kinetic Interactions
NOX promotes oxidation via: H+O2(+M) = HO2(+M)
pr essur e [atm] 10 Thermal/Chain Explosive Region
e ds co n im dl
Th ir d
it
lim it
1
NO+HO2 = NO2+OH NO2+H = NO+OH H2+OH = H2O+H and inhibits oxidation via: NO+X(+M) = XNO(+M) XNO+Y = XY+NO X+Y = XY where X,Y = H, O, or OH
Non Explosive
lim
E
n xt e
de
0.1
Se c
on
d
it
0.01
Fi r s t l i mit
Chain Branched Explosive Region
0.001 700 800 900 1000 temper atur e [K]
Overall Reaction Rate Variations with P
Conditions - H2/CO/O2/N2/Ar = 6.25/6.25/6.25/18.125/63.125
[H 2+CO]o / (d[H 2+CO]/dt)max (seconds) 10
2
0.5 atm 1 atm
XNO = 0 ppm XNO = 10 ppm XNO = 100 ppm XNO = 1000 ppm 2 atm 5 atm
[H 2+CO]o / (d[H 2+CO]/dt)max (seconds)
10
3
10
-1
10 atm
XNO = 0 ppm XNO = 10 ppm XNO = 100 ppm 20 atm XNO = 1000 ppm
10-2
101 100 10-1 10 10 10
-2 -3 -4
10
-3
40 atm
10
-4
10
-5
750
800
850 900 950 Temperature (K)
1000
1050
1000
1050
1100
1150 1200 1250 Temperature (K)
1300
1350
“Explosion limit” moves to significantly higher temperatures at increased operating pressures. Change in maximum reaction rate decreases dramatically at higher pressures. Small amounts of NO/NO2 can have dramatic effects on the maximum rate of reaction. Effect is even more dramatic at higher hydrogen content. Most conservative control of reaction rate is Temperature.
RCM Ignition Delay Variations with NO Present
(prediction)
80
0 ppm NO - tign = 5.422 ms 10 ppm NO - tign = 4.734 ms 100 ppm NO - tign = 2.006 ms 250 ppm NO - tign = 0.317 ms
60 Pressure (bar)
40
Pc = 30 bar Tc = 1,011 K Mixture composition (%): H2/CO/O2/N2 6.25/6.25/6.25/18.125 in balance Ar
20
0 0 0.01 0.02 Time (s) 0.03 0.04
Small amounts of NO/NO2 can have dramatic effects on the ignition delay at inlet temperatures relevant to gas turbine combustion. What other contaminants may be present in Syngas compositions?
Determining Collisional Efficiencies
The determination of rate data for H+ O2 (+M) = HO2 (+M) at temperatures of interest for combustion chemistry is complicated by competition with H + O2 = OH + O. At pressures where HO2 formation dominates the branching reaction, H2/O2 kinetics are influenced by HO2-H2O2 reactions which are not well-characterized. The addition of NO simplifies the high-pressure kinetics by providing an alternate consumption route for HO2 via HO2+NO=NO2+OH. NO2 reacts with H atoms via NO2 + H = NO + OH to form a catalytic cycle which results in steady-state NO2 concentrations.
[ NO2]ss =
kH + O 2( + M ) = HO 2( + M )[O 2] kNO 2 + H = NO + OH
Determining Collisional Efficiencies
High flow rates and dilute mixtures (typically 1% fuel) minimize spatial gradients within the VPFR and permit neglect of the diffusion terms in the governing equations. The experiment can thus be model as a zero-dimensional system using SENKIN with isobaric and adiabatic assumptions. Experimental data for H2/NO/O2/N2 seeded with amounts of water vapor or CO2 are shifted in time to agree with model predictions at point of 50% fuel consumption. To determine k2,0, reaction profiles are modeled with the baseline C1/NOx reaction mechanism in which k2,0 is treated as an adjustable parameter. The Troe formulation with k2,∞=4.52x1013(T/300)0.6 cm3 mole-1 sec-1 and FcN2(Ar)=0.5 (0.45) is used to model fall-off behavior.
Variable-Pressure Flow Reactor
Design Capabilities: P: 0.3-20 atm T: 300-1200 K τres: 0.001-3 sec Wall Heaters Slide Table Electric Resistance Heaters Oxygen Inlet
10-cm Diameter Quartz Test Section Sample Probe
Fuel Injector Mixer/Diffuser Type of Tests: Optical Access • Reaction profiles as function of time with Windows To Exhaust
fixed initial conditions • Reactivity traces with fixed residence time and changing initial conditions and/or initial reactant mole fractions
Analytical Equipment
Sample Probe
Silica-Coated Type R Thermocouple
out
Exhaust
Cooling water in
Diaphragm Pump
Needle valve
Calibration Gases
Flow Controller H2O ± 10% NO ± 5% NO2 ± 5%
H2
± 5% ± 2%
O2
FTIR cell
Flow Meter
Exhaust
Reaction Profiles (M=N2)
2 .4
P = 12.5 atm; Tin = 808 K; ϕ = 0.25; xNO = 107 ppm
O2
Mole Fraction [%]
2 .0 1 .6 1 .2 0 .8 0 .4 0 .0 140 120 100 80 60 40 20 0 0 .0 0 0 .2 5
H2 H 2O
Mole Fraction [ppm]
NO2
NO 0 .5 0 0 .7 5 1 .0 0 1 .2 5 1 .5 0 1 .7 5 2 .0 0
tim e [se c o n d s]
Symbols represent experimental data; lines are model predictions with k2,0N2 = 4.03x1015 cm3 mole-1 sec-1.
Sensitivity coefficients (Sij = ∂lnxi/ ∂lnAj) for the H2 and NO profiles in previous figure.
4 3 2 1 0 -1 -2 -3 -4 3 2 1 0 -1 -2 -3 -4 0.0 0.5
H+O2 (+N2 ) = HO2 (+N2 ) H+O2 = OH+O NO+OH(+N2 ) = HONO(+N2 ) OH+H2 = H2O+H NO2 +H = NO+OH
Sensitivity Analysis Results (M=N2)
H2 Sensitivities ( /10 )
-3 H+O2 (+N2 ) = HO2 (+N2 )
H+O2 = O+OH
NO2 +H = NO+OH
OH+H2 = H2O+H
NO Sensitivities ( /10
-4
)
1.0
1.5
2.0
2.5
3.0
3.5
4.0
time [seconds]
Summary
Further analyses of a C1 reaction mechanism we have developed show the mechanism can be utilized to predict phenomena over a wide range of parameter space that encompasses chemical system behavior under syngas gas turbine combustion conditions. The reactions CO+O+M = CO2+M (R22) and CO + HO2 = CO2 + OH (R20) primarily affect chemical induction time in pure carbonmonoxide-hydrogen-oxygen mixtures diluted by nitrogen, water, and/or carbon dioxide, and are not important to predicting energy release rate. CSP based analyses of recent RCM experiments show variation of ignition properties with CO/H2 ratio comes primarily from competition of (R22) and H2+HO2=H2O2+H (R17) during chemical induction (initial radical pool growth). Similar analyses of recent high pressure shock tube data are disparate with this result because of mole fraction differences in reactants and not the pressure/temperature range studied. The collisional efficiencies of mixtures even at high concentrations of water and/or carbon dioxide can be reasonably predicted based linearly on relative molar fractions.
Summary (cont’d)
Similar analyses show that the following reactions are important to heat release at high pressure conditions: HO2 + OH = H2O + O2 (R13) HO2+ H = OH + H (R11) H + O2 + M = HO2 + M (R9) H2 + OH = H2O + H (R3) H2 + O = H + OH (R2) H2O2 + M = OH + OH + M (R15) CO + OH = CO2 + H (R23) The change in overall reaction rate before and after crossing the extended second explosion decreases and temperature remains the principal parameter affecting overall rate as reaction pressure increases from 0.5 to 40 atm. The presence of small amounts of NO has the potential to essentially remove the importance of the above HO2-H2O2 reactions altogether. Contaminants in syngas are important to defining initial reaction behavior. Experimental assessment of collisional efficiency issues for water and carbon dioxide in (R15) at high dilutions remain to be fully investigated experimentally and are the principal issues in the remaining work under this effort.
Acknowledgements
Other contributors: • Dr. Andrei Kazakov (Research Staff), now at NIST, Boulder CO • Dr. Marcos Chaos (Research Associate) • Mr. Ken Kroenlein (Graduate Student) Thanks to Dr. Chich-Jen Sung for an advance copy of: • G. Mittal, C.J. Sung, R. A. Yetter (2006) “Autoignition of H2/CO at Elevated Pressures in a Rapid Compression Machine”, Int. J. Chem. Kin. (In Press). • Initial conditions and heat loss volume corrections for experiments • Other technical discussions