sub-mechanism advancements for surrogate fuel modelling
TRANSCRIPT
Sub-mechanism Advancements for Surrogate Fuel Modelling
MULTI AGENCY COORDINATION COMMITTEE FOR COMBUSTION RESEARCH MULTI AGENCY COORDINATION COMMITTEE FOR COMBUSTION RESEARCH (MACCCR) FUELS RESEARCH REVIEW(MACCCR) FUELS RESEARCH REVIEW
Princeton, NJ 20Princeton, NJ 20--22 September 201022 September 2010
Dr. Henry CurranCombustion Chemistry Centre
NUI Galway, Irelandhttp://c3.nuigalway.ie/
Visiting Researcher, Princeton University
Motivation
Alkanes and iso-alkanes are a very large fraction of petroleum derived and alternative (Fischer Tropsch) gas
turbine fuels
Alkenes are primary products formed by the decomposition of higher
hydrocarbons
Alkenes also result from R + O2chemistry at higher temperatures
Limited studies are available for the oxidation of higher alkenes
Discuss C5 – C7 but higher alkenes still need further study
Gas Turbine Fuel Functional Class Distribution
(POSF-4658)
Cycloalkanes
Napthalenes
Aromatic(alkyl benenes)
iso-alkanes
25.53%
29.55%
3.32%
38.59%3.02%
n-alkanes
Flow reactor oxidation data for conditions of 12.5 atm, 0.3% carbon, φ= 1.0 and t =1.8s, for POSF-4658 (symbols), POSF-4658 surrogate (solid lines) and model simulation of POSF-4658 surrogate (dashed lines), Inset;DT. from Dooley et al. doi:10.1016/j.combustflame.2010.07.001
500 600 700 800 900 1000
POSF 4658 Surrogate ModelCO CO2 O2 H2O
Spe
cies
con
cent
ratio
n / p
pm x
103
Temperature / K
1
2
3
4
5
500 600 700 800 900 10000
10
20
30
40
50POSF 4658 Δ TSurrogateModel
Δ T
/ K
Temperature / K
•1st generation surrogate model predictions generally in line with actual real fuel and surrogate experimental behaviours!Modeling shows that n-decane is principally active in radical generation=> exclusively so at lowest temperatures.•Excellent emulation of POSF 4658 at low Ts indicates quantity of n-alkyl fragments are approximately. correctly prescribed in surrogate.•Aromatic/iso-alkanes mainly compete for radicals generated by n-decane oxidation, while having little or no radical production, slowing overall reaction.•Iso-octane/toluene start to contribute some radical generation above ~800 and ~900K respectively.•These and other data suggest that there may be deficiencies in n-decane model kinetic at conditions above the hot ignition, but where β-scission remains slow relative to reaction with molecular oxygen..
3
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.40
10
20
30
40
50
60
70
80
90
100
110
120
Spec
ies
conc
entra
tion
/ ppm
Time / seconds0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
0
20
40
60
80
100
120
140
160
ΔT /
K
Time / seconds
n‐Decane Oxidation Time‐History Princeton Variable Pressure Flow Reactor
4
Model predictions; Westbrook C.K., Pitz W.J., Herbinet O., Curran H.J, and Silke E.J., Combust. Flame 156 2009 181 ⎯199.
Data Copyright, Princeton University 9/20/2010
Data Copyright, Princeton University 9/20/2010
Model predictions show that 1-olefins are derived from QOOH+O2, R+O2, and R beta-scission channels at the present conditions.
Experimental data (symbols) versus simulation (lines) for n‐decane oxidation.
1-olefin species Profiles
Previous Studies 1
Experiments:• 1971 Tipper and Titehard: Effect of olefin addition to n-heptane cool
flames
• 1973 Hughes and Prodhan: Cool flames from 1- and 2-pentene
• 1980 Baldwin et al.: H2/O2 system plus 1- and 2-pentene (low T)
• 1996 Prabhu et al.: Flow reactor data for 1-pentene
• 1999 Minetti et al.: RCM data for 1-pentene
• 2005 Vanhove et al.: RCM data (low T) for the three isomers of hexene
• 2006 Yahyaoui et al.: Shock Tube and JSR study of 1-hexene
Previous Studies 2
Modeling Studies:• S. Touchard et al., Proc. Combust. Inst. 30 (1) (2005) 1073-1081
(1-pentene, 1-hexene low T)
• M. Mehl et al., 29th Meeting on Combustion, Pisa – Italy 2006 (1-pentene, high and low T)
• M. Yahyaoui et al., Combustion and Flame 147 (2006) 67–78 (1-hexene, intermediate and high T)
• M. Mehl et al., Comb. Flame 155 (2008) 756-772(1-, 2-, 3-hexene, high and low T)
• R. Bounaceur et al., Proc. Combust. Inst. 32 (1) (2009) 387-394(1-, 2-, 3-hexene, low T)
High and Low T Oxidation of Hexene Isomers
(M. Mehl et al., Comb. Flame 155 (2008) 756-772.)
At low temperature reactivity is controlled by the saturated portion of the fuel. At high temperature predictions suggest that the relative reactivity of the three HCs is inverted (no experimental data).
Mechanism developed in the context of Milano’s kinetic mechanism(Ranzi’s group)
Igni
tion
Del
ay T
ime
[ms]
(G. Vanhove et al., Proc. Comb. Inst. 30 (2005) 1065-1072.)
Alkene Oxidation ProcessesMain reaction pathways involved in alkene autoignition
RH
R
ROO
QOOH
β‐decomposition
Initiation
OOQOOH
Ketohydroperoxide
HORH
HORHOO
Waddington
Temperature
HO2 activation
HO2 Elimination
600K 750K 900K 1050K 1200K 1350K 1500K
+ OH
Reactivity of Hexene Isomers at Low Temperatures
· · ·
OO
OO
OO
H H
1-hexene 2-hexene 3-hexene
H
H
· ·
· ·
·
· ·
OO H
OO
OOH
O
O
Radical Branching
0 500 1000 1500
Pres
sure
/ ar
b. u
nits
time / μs
2-Hexene diluted in AirT
5 = 1021 K, P
5 = 10.4 atm
0 1000 2000Pr
essu
re /
arb.
uni
tstime / μs
2-Pentene diluted in AirT
5 = 993 K, P
5 = 10.8 atm
High pressure and low pressure shock tube data
• High pressure shock tube (example)
Experimental (NUI Galway)
Hexene Isomers: High Temperature
The ignition delay time of the three isomer cross at
about 1200 K
2-hexene shows the highest reactivity
Decomposition rates are controlling the ignition
process determining the composition of the radical
pool
3-hexene
2-hexene
1-hexene
1.00.6 0.7 0.8 0.91000K/T
10
100
1000
10000
1.1
Igni
tion
Del
ay T
ime
[μs]
Pressure 10 atm, φ = 1 in air
1.0E+01
1.0E+02
1.0E+03
1.0E+04
1.0E+05
1.0E+06
0.6 0.8 1 1.2 1.41000K/T
1.6
Igni
tion
Del
ay T
ime
[μs]
3-hexene
1-hexene2-hexene
Pressure 10 atm, φ = 1 in air, Vanhove et al. 2005
Pentene Isomers: High and Low Temperature
2-pentene
1-pentene
Pressure 10 atm, φ = 1 in air, Minetti et al. (1999)
Good agreement with the low and high temperature regions
No low temperature data for 2-pentene (future work)
0.01
0.1
1
10
100
1000
0.6 0.8 1 1.2 1.41000K/T
1.6Ig
nitio
n D
elay
Tim
e [m
s]
Pentene Isomers: Low Pressure
The model performs satisfactorily at these conditions. This validation gives us confidence in the high pressure predictions.
Igni
tion
Del
ay T
ime
[μs]
10
100
1000
10000
0.5 0.6 0.7 0.8 0.9
2-penteneφ =1.0φ =0.5
φ =2.0
1000K/T0.5 0.6 0.7 0.8 0.9
1000K/T
10
100
1000
10000
1-penteneφ =1.0
φ =0.5
φ =2.0
0.50 0.55 0.60 0.65 0.70 0.75 0.8010
100
1000 φ = 2.0 φ = 1.0 φ = 0.5
Igni
tion
Del
ay T
ime
/ μs
1000 K / T
φ = 2.0 φ = 1.0 φ = 0.5
2000 1800 1600 1400
1-pentene2-pentene
T / K
Pressure 1 atm, 1% pentene in argon diluent
Alkenes: Conclusions• New set of experiments carried out to fill the gap in
the high temperature behavior of alkenes• High and low temperature mechanism for 1-, 2-
pentene and 1-, 2-, 3-hexene isomers incorporated in the kinetic mechanism
• Model validated successfully against ST data• Quantitative information provided to develop better
models for autoignition of more complex unsaturated fuels
Mechanism available:https://www-pls.llnl.gov/?url=science_and_technology-
chemistry-combustion-mechanisms
Small Hydrocarbon Kinetic Improvements
• Gallagher et al. C1-C4 kinetics combined with C0-C1 kinetics of Princeton as the base mechanism for the larger hydrocarbon kinetics (n-decane, n-dodecane, iso-octane, toluene, alkyl benzenes)
• Notes deficiencies in C2, C3 flame speed comparisons
• New work at Galway/Texas A&M on C1/C2 kinetics leads to significant improvements.
• Further revisions in Hydrogen/CO model for high pressure burning rates expected early this fall.
CH4/C2H6 Flame SpeedsBurning Velocities for Methane/Ethane
mixtures, P = 5 atm, Ti ~ 298 K
0.6 0.8 1.0 1.2 1.45
10
15
20
25
30 100 % C2H6
60/40 CH4 / C2H6
80/20 CH4 / C2H6
100 % CH4
Bur
ning
Vel
ocity
(cm
s-1)
φ
CH4/C2H6 Flame SpeedsBurning Velocities for Methane/Ethane
mixtures, P = 5 atm, Ti ~ 298 K
0.6 0.8 1.0 1.2 1.45
10
15
20
25
30
Dash - C3_49
100 % C2H6
60/40 CH4 / C2H6
80/20 CH4 / C2H6
100 % CH4
Bur
ning
Vel
ocity
(cm
s-1)
φ
CH4/C2H6 Flame SpeedsBurning Velocities for Methane/Ethane
mixtures, P = 5 atm, Ti ~ 298 K
0.6 0.8 1.0 1.2 1.45
10
15
20
25
30
Dash - C3_49Solid - C3_51.6
100 % C2H6
60/40 CH4 / C2H6
80/20 CH4 / C2H6
100 % CH4
Bur
ning
Vel
ocity
(cm
s-1)
φ
CH4/C2H6 Mixtures
010
2030
0
20
40
60
80
100
75
80
859095100
(100% C2H6)
Conditions of Current Study Sampling of Previous Studies
and below
Blen
d C
ompo
nent
CH
4
Pressure (atm) and above
Dilu
tion
(%)
• Studies Sampled Here:
• Burcat et al. (1971)
• Burcat et al. (1972)
• Cooke and Williams (1971)
• Cooke and Williams (1975)
• Crossley et. al (1972)
• de Vries and Petersen (2007)
• de Vries et al. (2007)
• Eubank et al. (1981)
• Herzler and Naumann (2009)
• Hidaka et al. (1999)
• Hidaka et al. (2000)
• Huang and Bushe (2006)
• Krishnan and Ravikumar (1983)
• Lamoureux and Paillard (2003)
• Lamoureux et al. (2002)
• Lifshitz et al. (1971)
• Petersen et al. (2007)
• Seery and Bowman (1970)
• Skinner and Ruehrwein (1957)
• Spadacinni and Colket (1994)
• Tranter et al. (2002)
• Tsuboi (1978)
• Zhukov et al. (2003)
CH4/C2H6 ST ignition Delay Times
6.5 7.0 7.5 8.0 8.5
100
1000
Mix 9 Mix 10 C4_51.6 San Diego JetSurf
Igni
tion
dela
y tim
e / μ
s
104 K / T
Mixture Blend Dilution Φ Pressure(CH4/C2H6) (%) (atm)
9 0 / 100 75 1 110 0 / 100 75 0.5 1
CH4/C2H6 ST ignition Delay Times
6.5 7.0 7.5 8.0 8.5
100
1000
Mix 11 Mix 12 C4_51.6 San Diego JetSurf
Igni
tion
dela
y tim
e / μ
s
104 K / T
Mixture Blend Dilution Φ Pressure(CH4/C2H6) (%) (atm)
11 75/25 95 1 1012 50/50 95 1 10
CH4/C2H6 ST ignition Delay Times
7.5 8.0 8.5 9.0 9.5
100
1000
Mix 21 Mix 22 C4_51.6 San Diego JetSurf
Igni
tion
dela
y tim
e / μ
s
104 K / T
Mixture Blend Dilution Φ Pressure(CH4/C2H6) (%) (atm)
21 0/100 85 1 1022 0/100 85 0.5 10
Changes to the sub-mechanismNew H2/O2 rates from recent Stanford University study
H + O2 <=> O + OHH2O2 (+M) <=> OH + OH (+M)H2O2 + OH <=> H2O + HO2
- new C2H4 + OH rate from recent Stanford paper- Alzueta's rate for HCCO + O2
- Klippensteins rate for CH2CHO decomposition (J. Phys. Chem. A 2006, 110, 5772-5781)
- Altered the branching ratio of C2H4 + O = CH2CHO + H (37%) / CH3 + HCO (63%) to 50:50
- San Diego Mech value for C2H5(+M) <=> C2H4+H (+M)- CH3 + CH3 <=> C2H5 + H chemically activated (Stewart et al. C&F 1989)- CH3 + CH3(+M) = C2H6 (+M) taken from Wang et al.
(JPC A 107:11414 (2003))
C2H4 Ignition
• Very important to alkane, cycloalkane, alcohol, ether and ester oxidation mechanisms among others:
• nC3H7 = C2H4 + CH3
• MCH oxidation → lots of C2H4
• C2H5OH = C2H4 + H2O• C2H5OC2H5 = C2H5OH + C2H4
• C2H5C(O)OC2H5 = C2H5C(O)OH + C2H4
C2H4 ignition delay times
Shock tube ignition delay times for C2H4with model predictions
7.5 8.0 8.5 9.0 9.5 10.0
100
1000
C2H4 in Air, 2 % C2H4 3 % O2, φ = 2.0, Paverage 10.1 atm
Igni
tion
Del
ay, μ
s
104 K / T
C4_50
7.8 8.0 8.2 8.4 8.6 8.8 9.0100
1000
C2H4 in Air, 1.0 % C2H4 3 % O2 φ = 1.0, Paverage 1.1 atm
C4_50
Igni
tion
Del
ay, μ
s
104 K / T
7.5 8.0 8.5 9.0 9.5 10.0
100
1000
C2H4 in Air, 2 % C2H4 3 % O2, φ = 2.0, paverage = 10.1 atm
Igni
tion
Del
ay, μ
s
104 K / T
C4_50
Improving model performance Identify a point where model performs poorly
Sensitivity analysisφ = 2.0, P = 8.8 atm, T = 1170 K
C2H3+O2<=>CH2CHO+OC2H4+CH3<=>C2H3+CH4
C2H4+HO2<=>C2H4O1-2+OHC2H4+OH<=>C2H3+H2O
H+O2<=>O+OHCH2CHO+O2=>CH2O+CO+OH
CH3+HO2<=>CH3O+OHH2O2(+M)<=>OH+OH(+M)C2H5+HO2<=>C2H5O+OH
C2H4+O<=>CH2CHO+HC3H8(+M)<=>CH3+C2H5(+M)
C2H4+O<=>CH3+HCOCH3+CH3(+M)<=>C2H6(+M)
CH2CHO<=>CH2CO+HC2H4+H(+M)<=>C2H5(+M)
CH3+HO2<=>CH4+O2
C2H5+O2<=>C2H4+HO2
C2H3+O2<=>C2H2+HO2
HO2+OH<=>H2O+O2
C2H3+O2<=>CH2O+HCO
-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6
C2H4 in Air, 2 % C2H4 3 % O2, φ = 2.0, p = 8.8 atm, T 1170 K
Sensitivity Coefficient
Sensitivity analysis
Inhibits reactivity
Promotes reactivity
7.8 8.0 8.2 8.4 8.6 8.8 9.0100
1000
C2H4 in Air, 1.0 % C2H4 3 % O2 φ = 1.0, paverage = 1.1 atm
C4_51.1 C4_50
Igni
tion
Del
ay, μ
s
104 K / T
Ethylene
7.5 8.0 8.5 9.0 9.5 10.0
100
1000
C2H4 in Air, 2 % C2H4 3 % O2, φ = 2.0, paverage = 10.1 atm
Igni
tion
Dela
y, μ
s
104 K / T
C4_51.1 C4_50