Lawrence Livermore National LaboratoryLawrence Livermore National Laboratory
Impact of Alternative Fuels on Gas Turbine and Diesel Engine Combustion
Mani Sarathy, William J. Pitz (PI), Charles K. Westbrook, Marco Mehl, LLNLHans-Heinrich Carstensen, Lam K. Huynh, Stephanie M. Villano, Anthony M.
Dean (PI), Colorado School of MinesFuel Summit, 20-23 September 2010, Princeton, NJ
LLNL-PRES-456812
Lawrence Livermore National Laboratory, P. O. Box 808, Livermore, CA 94551This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344
Acknowledgment: Sponsor
Office of Naval Research
Dr. Dave Shifler, Program Manager
2Lawrence Livermore National Laboratory2
Assess fuel impact on gas turbine engine performance
3Lawrence Livermore National Laboratory
From Med Colket, UTRC
Examine effects of alternative fuel (F-T and biofuels) on diesel engine combustion
• Navy has many diesel engines
•With Jim Cowart and Patrick Caton U S Naval Academy•With Jim Cowart and Patrick Caton, U.S. Naval Academy
N l A d C iNaval Academy Cooperative Fuels Research (CFR) diesel test engine
4Lawrence Livermore National Laboratory
Navy is interested in Fischer-Tropsch (FT) fuels: They can contain large amounts of singly-methylated iso-alkanes
FT analysis (NIST*)57% single
0.20cycloalkanes
other isoalkanes
Syntroleum S-8 synthetic jet fuel
• 57% single methyl branch alkanes
• 25% n-alkanes
0.16
single methyl branch
n-alkanes
• 16% multiple branched alkanes
• 2% cycloalkanes
0.12
e f
ract
ion
* Smith et al. Int. J. Propulsion (2008) 2% cycloalkanes
0.08Mo
le Propulsion (2008)
0.04
5Lawrence Livermore National Laboratory
0.00
C8 C9 C10 C11 C12 C13 C14 C15 C16
We are first focusing on 2-methyl alkanes have a single methyl branch and interesting ignition behavior
Species Cetane No.
nC7H16 54
C7H16-2 42
nC8H18 65
C8H18-2 49
nC9H20 75
C9H20-2 54
6Lawrence Livermore National Laboratory
Chemical Kinetic Mechanism for 2-methyl alkanesy
Includes all 2-methyl alkanes up to C20 which covers the entire distillation range for gasoline, jet and diesel fuels
Built with the same reaction rate rules as our successful iso-octane and iso-cetane mechanisms
7,900 species
mechanisms.
27,000 reactions
7Lawrence Livermore National Laboratory
Why do we need to get the distillation curve right for diesel combustion?
6- component diesel surrogate
diesel spray calculationVaporized fuel mass fraction
0.3 0.3
MW 165 0 MW 186 5
g
0.15
0.2
0.25
fract
ion
0.15
0.2
0.25
fract
ion
MW=165.0 MW=186.5
0
0.05
0.1mol
e
0
0.05
0.1mol
e
0c7h8 c10h22 c12h26 c14h30 c16h34 c18h38
component
0c7h8 c10h22 c12h26 c14h30 c16h34 c18h38
component
Light components come off first near the upstream portion of the diesel jet
8Lawrence Livermore National Laboratory
From : Ra Y, Reitz RD. A vaporization model for discrete multi-component fuel sprays. Int. J. Multiphase Flow 2009
Experimental Validation• Idealized chemically reacting flow systems with/without simplified
transport phenomenonCombustion Parameters
Jet Stirred Reactors Premixed Laminar FlamesCombustion Parameters
Temperature
Pressure
Mixture fraction (air-fuel ratio)
Shock tube
Non Premixed Flames
( )
Mixing of fuel and air
Engine
RCM
High pressure flow reactors
Engine CombustionElectric Resistance
Heater
Evaporator
Fuel Inlet
Wall Heaters
9Lawrence Livermore National Laboratory
Slide Table
Oxidizer Injector
Optical Access Ports
Sample Probe
2-methylhexane comparison in the RCM
45
50
Stoichiometric/’air’ mixtures15 t
Entire compression stroke
30
35
40
me (m
s)
15 atmExperimental data:
Galway:
stroke simulated and heat loss simulated after end of
20
25
30
n Delay Tim Galway:
Silke et al. Proc. Comb. Inst. (2005)
compression
5
10
15
Ignition
SimulationsExperiments
RCM
0650 700 750 800 850 900 950
10Lawrence Livermore National Laboratory
Temperature at End of Compression (K)
2-methyl heptane model behaves well under high temperature shock tube conditionsp
100000 =3, P=21 atm, 1.355% fuel, O2/Ar
1000
10000
ay (μ
s)
Experiment Model
100
nitio
n de
l
10
Ign
experiment simulationExperiments: Dayton high pressure shock tubeKahandwala et al. Energy & Fuels (2008)
10.50 0.70 0.90 1.10 1.30 1.50
1000/T (1/K)
11Lawrence Livermore National Laboratory
1000/T (1/K)
Previously shown that simulated ignition behavior of n-alkanes showed little dependence with carbon length
1.0E-0113 bar
1.0E-02
[s]
Stoichiometric fuel/air mixtures
1.0E-03
elay
Tim
e
nC8H18nC9H20
1.0E-04
Igni
tion
De nC9H20
nC10H22nC11H24nC12H26nC13H28nc14H30
1.0E-050.7 0.9 1.1 1.3 1.5 1.7
I
nC15H31nC16H34Westbrook et al. Comb. Flame 2009
12Lawrence Livermore National Laboratory
1000/T [K]
2-methyl alkanes show effect of chain length at higher equivalence ratio and pressure (Dayton conditions):q p ( y )
10000 =3, P=21 atm, 1.355% fuel, O2/ArReactivity increases
(us)
increases with chain length
C10
1000
on d
elay
c10h22-2 - sarathy c12h26-2
C20
Igni
ti
y
c13h28-2 c14h30-2
c15h32-2 c16h34-2
Biggest fuel effect of chain length found in NTC region
1000 90 1 00 1 10 1 20 1 30 1 40 1 50 1 60
c18h38-2 c20h42-2region
13Lawrence Livermore National Laboratory
0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.601000/T (1/K)
2-methylalkanes show less NTC than n-alkanes
10000 (=3, P=21 atm, 1.355% fuel, O2/Ar)2-methylalkane
(us)
less reactive than n-alkane
1000
on d
elay
Less NTC behavior in 2 th l lk
Igni
ti
nc12h26 c12h26-2
2-methylalkane
1000 90 1 00 1 10 1 20 1 30 1 40 1 50 1 60
14Lawrence Livermore National Laboratory
0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.601000/T (1/K)
Comparison of reactivity for all the n-alkanes and 2-methyl alkanesy
10000 (=3, P=21 atm, 1.355% fuel, O2/Ar)Reactivity increases
2-methyl alkanes
(us)
increases with chain length
C10
1000
on d
elay
nc10h22 - westbrook c10h22-2 - sarathyn-alkanesC20
Igni
ti nc12h26 c12h26-2nc13h28 c13h28-2nc14h30 c14h30-2nc15h32 c15h32-2
Effect of chain length for n-alkanes now observed
1000 90 1 00 1 10 1 20 1 30 1 40 1 50 1 60
nc16h35 c16h34-2c18h38-2 c20h42-2
15Lawrence Livermore National Laboratory
0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.601000/T (1/K)
Opposed-flow Diffusion Flame (OPPDIF) comparisons for a 2-methyl heptane
Oxidizer
Fuel Mixture• The one dimensional flame structure is ideal
Port diameter = 25.4 mm
P t G 20
for modeling.
• The emissions and temperature profiles are dependent on chemical kinetics due to the non-turbulent flame
16Lawrence Livermore National Laboratory
Port Gap = 20 mmnon turbulent flame.
Overall good prediction of major and minor species of 2-methylheptane
8%
9%
N (%
) CO26000
7000
PM) C2H4
5%
6%
7%
NTR
ATIO
N CO
C8H18-24000
5000
RAT
ION
(P C2H2
CH4
2%
3%
4%
AR
CO
NC
EN
2000
3000
CO
NC
ENTR
C2H2
0%
1%
2%
0 2 4 6 8 10 12 14 16 18 20
MO
LA
0
1000M
OLA
R C
0 2 4 6 8 10 12 14 16 18 20
DISTANCE FROM FUEL PORT (mm)
Good prediction of
2 4 6 8 10DISTANCE FROM FUEL PORT (mm)
17Lawrence Livermore National Laboratory
Good prediction of major species profiles
Important minor species are well predicted
Further experimental validations planned,2-methylheptane
• Counterflow Flame Ignition/ExtinctionUC San Diego
y p
g
•Jet Stirred ReactorCNRS Orleans, FranceCNRS Orleans, France
• Shock TubeRensselaer Polytechnic Institute New YorkRensselaer Polytechnic Institute, New York
• Ignition Quality Tester (IQT)NREL ColoradoNREL, Colorado
•Flow ReactorPrinceton University New Jersey
Electric ResistanceHeater
Evaporator
18Lawrence Livermore National Laboratory
Princeton University, New Jersey Fuel Inlet
Slide Table
Oxidizer Injector
Optical Access Ports
Sample ProbeWall Heaters
Collaborating with the Naval Academy on fuel effects in diesel engine
10
100
ec]
Phi=4, 55 atm, fuel/air100% Hex
50%H
Mixtures of n-hexadecane and tolueneSimulations
0.1
1
10
nitio
n D
elay
[ms 50%Hex;
50%Tol30%Hex; 70%Tol20%Hex; 80%Tol10%Hex;
toluene(mixtures are by liquid volume)(Midshipman Aaron Carr)
Experiments on n hexadecane/toluene100% Hex
10% Hex; 90% Tol
770 K0.01
0.8 1 1.2 1.4
Ig
1000/T[K]
10%Hex; 90%Tol
Experiments on n-hexadecane/toluene mixtures in a Diesel CRF engine
(SAE 2010-01-2188: Mathes, Ries, Caton, Cowart, Prak, Hamilton, US Naval Academy)
Ignition delay is defined from start of injection
100% Hex 770 K
3
4
sec]
T=770K; Phi=4, 55 atm, fuel/air
AD)
Ignition delay is defined from start of injection to 10% mass burned
Simulations
Experiments
0
1
2
3
0 20 40 60 80 100gniti
on D
elay
[ms
Igni
tion
Del
ay (C
A
19Lawrence Livermore National Laboratory
Ig
%Toluene by Volume
Summary for LLNL work
A new chemical kinetic mechanism for 2-methyl alkanes has been developed:p• The mechanism covers the entire distillation curve of practical gasoline, jet and diesel fuels (C7-C20).• The model successfully reproduces the experimental data currently available for this class of fuels.•Further validations plannedFurther validations planned.• Key differences between 2-methylalkanes and n-alkanes are reproduced by the model.
20Lawrence Livermore National Laboratory
Detailed Modeling of Low-Temperature Alkane Oxidation:
High-Pressure Rate Rules for Alkyl+O2 Reactions
Hans-Heinrich Carstensen, Lam K. Huynh, Stephanie M. Villano, and Anthony M. Dean
Chemical Engineering DeptChemical Engineering Dept.
Colorado School of Mines
Third Fuels Summit
September 20 23 2010September 20-23, 2010
Development of large detailed chemical models facilitated by the use of rate estimation techniques
• Typical hydrocarbon gas phase reaction mechanisms contain — Hundreds of speciesHundreds of species— Thousands of reactions, many pressure-dependent
• Problem: How to assure consistency in rate constant assignments?— Experimental info only available for small fraction of reactions— Often literature data over limited range of conditions
High level calculations restricted to small molecules— High level calculations restricted to small molecules
• Solution: Develop accurate rate constant estimation method— Use high-level theory to calculate rate constants for smaller speciesg y p— Generalize results on a per-site basis and use these for larger species — Use rate rule based rate constants as input for pressure-dependence
l i
22Lawrence Livermore National Laboratory22
analysis— Approach applicable to either hand-generated or computer-generated
mechanisms
Model development: from electronic structure calculations to pressure-dependent rate constants
Geometry, Frequencies, Electronic Energy,
Gaussian Software®
Dipole Moment, Polarizability, …FANCY
Heat of Formation Entropy Heat Capacity
TSTdG
Heat of Formation, Entropy, Heat Capacity
3-Freq Representation, NASA Polynomials
CHEMDIS
High-pressure Rate Constants k(T) for
Elementary Reactions
k(T,P)
Pressure-dependent Rate Constants k(T,P) for Apparent Reactions
23Lawrence Livermore National Laboratory
Reaction classes analyzed
H and CH3 Abstraction from Alkanes, Cycloalkanes, AlkenesRate Constant Rules for the Automated Generation of Gas-Phase Reaction Mechanisms, J. Phys. Chem. A, 2009, 113, 367-380
• Intramolecular H-AbstractionThe Kinetics of Pressure-Dependent Reactions, in Comprehensive ChemicalKinetics (Elsevier), 2007, 42 105-187
H and CH3 Abstraction from Alcohols, Elimination of Water from AlcoholsDevelopment of Detailed Kinetic Models For The Thermal Conversion of Biomass Via First Principle Methods And Rate Estimation Rules, inComputational Modeling in Lignocellulosic Biofuel Production, ACS Symposium Series
(in press)
• RO2 Isomerizations• RO2 Concerted Elimination of HO2
QOOH C li Eth + OH• QOOH = Cyclic Ether + OH• O2QOOH Isomerization• O2QOOH Concerted Elimination of HO2
24Lawrence Livermore National Laboratory24
Alkyl isomerizations consistent with expectations
Primary to secondary H- atom shift reactions:
Lower activation energy for 5 and 6-
Alkyl radicals(primary/secondary)(Primary to secondary)
energy for 5 and 6member ring TS s with less ring strain
Lower A-factor as more CH2 rotors tied up in TS Additional calculations ongoing to confirm consistent behavior
• Results used as basis for rate rules
25Lawrence Livermore National Laboratory
N-propyl + O2 chemistry: Fast isomerization of CCCOO• adductleads to chain branching
CCCOO● adduct
This is the O2QOOH that leads to chain
• Only important pathways for n-propyl + O2 are isomerization, concerted elimination and
leads to chain branching
26Lawrence Livermore National Laboratory
y p p y p py 2redissociation
i-propyl+O2: Slower isomerization of CC(OO•)C adductleads to chain inhibition
CC(OO●)C adduct
• Concerted elimination dominant for isopropyl + O2b hi th i hibit d
27Lawrence Livermore National Laboratory
– branching pathway inhibited• Detailed calculations show other pathways can be neglected
RO2 isomerizations show same qualitative behavior as alkyl radicals
5-member TS 6-member TSOH
OHR
tertiaryprimary
RO O
primarytertiary
Activation energy depends on ring size and overall thermochemistry Again amenable to rule generation
28Lawrence Livermore National Laboratory
g g
Comparison of model predictions to propane ignition over the NTC region in a RCM
Expect all predictions to be faster than data • Predictions adiabatic
Experiments27 atm=0.5fuel/”air” mixtures
• Expect heat loss at longer times in data
All models capture NTC behaviorNTC i t hi h T ith CSM• NTC region at higher T with CSM model
• Reduced CSM model, with much smaller reaction subsets for both 37 atmRO2 and O2QOOH, very similar to full model
Analysis suggests substantial differences in Galway and CSMGalway
Experiments
=0.5fuel/”air” mixtures
differences in Galway and CSM models• Galway model* uses LLNL rate
rules
CSMReduced CSM
y
29Lawrence Livermore National Laboratory
Experimental data from Galway RCM:* Gallagher et al. Combustion and Flame 2008, 153, 316.
Significant differences in CSM vs. LLNL rate constants (e.g. RO2 isomerization)
CBS-QB3 results generally lower than LLNL values for 5-member TS CBS-QB3 results much higher than LLNL values for 6-member TS
• Mainly due to higher A-factors (much higher than alkyl i i ti )
30Lawrence Livermore National Laboratory
isomerizations) Differences lead to significantly different reaction pathways
Summary/Next Steps Mechanism constructed with rate constants based on unadjusted CBS-
QB3 potential energy surfaces captures NTC behavior in propane oxidation• Need to extend mechanism to larger systems where more data
available to reach firm conclusions on its success Relatively few reactions of RO2 found to be important for ignition
calculations• Significantly simplifies accounting for effects of pressure on
reaction rates and branching ratios Extension to larger systems facilitated by ability to generalize results
into reaction rate rules Extend approach to consider second addition of O2 (to QOOH) that
leads to low temperature chain branchingleads to low temperature chain branching
31Lawrence Livermore National Laboratory
Lawrence Livermore National LaboratoryLawrence Livermore National Laboratory
? QUESTIONS ?? QUESTIONS ?
Top Related