Ellen Meeks Fuels Research Review September 16, 2009
Transcript of Ellen Meeks Fuels Research Review September 16, 2009
![Page 1: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/1.jpg)
Fuels Research Review
September 16, 2009
Ellen Meeks
![Page 2: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/2.jpg)
Outline
Overview
– Participants
– Objectives of project
Results
– Surrogate blend for F-T and bio-derived jet fuels
– Experimental data obtained at USC
– Mechanism validation
– Mechanism reduction
Conclusions
– Comparisons of F-T and bio-derived jet fuels with
conventional fuels
– Outlook for detailed kinetics in CFD simulations
![Page 3: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/3.jpg)
Outline
Overview
– Objectives of project
– Tasks and participants
Results
– Surrogate blend for F-T and bio-derived jet fuels
– Experimental data obtained at USC
– Mechanism validation
– Mechanism reduction
Conclusions
– Comparisons of F-T and bio-derived jet fuels with
conventional fuels
– Outlook for detailed kinetics in CFD simulations
![Page 4: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/4.jpg)
Collaborators on this project
NASA Glenn (Project oversight)
– M. Rabinowitz, D. Bulzan
– Award # NNC07CB45C
– 05-2007 through 04-2009
University of Southern California (Co-PIs)
– Profs. F. N. Egolfopoulos and T. Tsotsis
– Y. L. Wang, P. Veloo, Q. Feng, A. Holley
Reaction Design
– E. Meeks (PI)
– C. V. Naik, K. V. Puduppakkam, A. Modak (Co-PIs)
– Consultant: C. K. Westbrook (Co-PI)
4
![Page 5: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/5.jpg)
Project Objectives
Obtain fundamental data on combustion
behavior of alternative jet fuels
– Fischer-Tropsch and Bio-derived
– Real fuels and associated model (surrogate) fuels
Assemble fuel-chemistry models for simulation
– Validate kinetics through comparison with experiment
– Recommend surrogate blends
– Provide accurate, reduced mechanisms
Identify differentiating characteristics of
molecular fuel components
5
![Page 6: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/6.jpg)
Tasks undertaken to meet objectives
Fuels survey and analysis
– F-T fuels
– Bio-derived fuels
Flame experiments for liquid / heavy fuels
– Laminar flame-speed and flame-extinction limits
– Augmentation of diagnostics to measure NOx and soot
Surrogate-model assembly and testing
– Build from state-of-the-art detailed mechanisms
– Refine NOx sub-model
– Flame modeling, including NOx and soot formation
Mechanism reduction for targeted conditions
6
![Page 7: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/7.jpg)
Ultimately, we want fuel-combustion
mechanisms suitable for CFD simulation
1. Identify appropriate surrogate components
for targeted alternative jet fuels
2. Assemble & test component mechanisms
against available experimental data
3. Determine appropriate component-blending
method to match real-fuel properties
4. Test surrogate blend against real fuel
behavior
5. Reduce surrogate mechanism for targeted
conditions
![Page 8: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/8.jpg)
Outline
Overview
– Objectives of project
– Tasks and participants
Results
– Surrogate blend for F-T and bio-derived jet fuels
– Experimental data obtained at USC
– Mechanism validation
– Mechanism reduction
Conclusions
– Comparisons of F-T and bio-derived jet fuels with
conventional fuels
– Outlook for detailed kinetics in CFD simulations
![Page 9: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/9.jpg)
Fuel data was collected to determine
appropriate fuel surrogates
2 F-T samples obtained from the Air Force
– Courtesy of Tim Edwards, AFRL
– GC/MS data provides class/size composition of fuels
– Cetane number (IQT data) and distillation points
Bio-derived jet fuel (R-8) also acquired from AF
– Same general characteristics as F-T samples
– Detailed chemical analysis not available
9
![Page 10: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/10.jpg)
F-T fuel analysis shows predominance of
iso-alkanes with very little branching
Summary results for S-8 (Syntroleum) sample: (based
on GCMS analysis courtesy of Tim Edwards of AFRL)
Most of the iso-paraffins consist of only one methyl
branch on a long and straight alkane chain
10
F-T fuel (S-8) analysis, mol%
iso-alkane
n-alkane
0
5
10
15
20
25
8 9 10 11 12 13 14 15
Mo
l %
Carbon #
Carbon number distribution in F-T fuel (S-8)
![Page 11: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/11.jpg)
F-T fuel from different sources are similar,
but have different C# distribution
Highlights about composition – Mostly 1 or 2 methyl side chain alkanes
Side chains responsible for lower cetane numbers of large alkanes
– C# varies significantly
C7-C17 for S-8 with C8–C13 consist majority of composition
C8-C12 for Shell
– Cetane number is ~60
Shell GTL Cetane
Numbers IQT
ASTM
D976
Shell GTL 59.67 68
Syntroleum
S-8 59.08 57
![Page 12: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/12.jpg)
Reaction Design has developed a Surrogate
Blend Optimizer to match fuel properties
aromatics
olefins
c-paraffins
i-paraffins
n-paraffins
Set Target Characteristics
• Class composition
• Heating value
• Octane / Cetane #
• H/C ratio, O content
Other physical properties
Choose Fuel Palette
Select components or
size limitations
Surrogate Fuel
Composition
Match Properties
• Optimize Composition
45%
15% 3%
1%
15%
Chemical
Model for
Simulation
n-heptane
Iso-octane
1-pentene
mchexane
m-xylene
ethanol
19%
Fuel properties
to match
Mechanism
Reduction, as
necessary
![Page 13: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/13.jpg)
We determined surrogate blends based on
properties that are important for simulation
Iso-octane included to match Cetane Number
– Although highly branched, captures effect of low-branch components
– We had a well validated mechanism consistent with other mechanisms
Shell GTL Surrogate
Shell GTL Targeted
Properties (Measured)
S-8 Surrogate
S-8 Targeted
Properties (Measured)
Surrogate Blend Definition
iso-Octane (mol %) 28 32
n-Decane (mol %) 61 25
n-Dodecane (mol %) 11 42
Properties Comparison
Cetane Number 61 61 61 60
H/C molar ratio 2.21 2.17 2.20 2.17
Lower Heating Value (MJ/kg)
45 44 44 44
T50 boiling point (K) 404 445 447 474
![Page 14: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/14.jpg)
The F-T surrogate mechanism was
assembled based on previous work
n-decane and n-dodecane
– From Westbrook et al. (2008)
mechanism of n-alkanes
Removed species > C12
Removed low-temperature kinetics to focus
on flames
Added estimates of transport parameters
iso-octane mechanism merged in
– From Curran et al. (2002)
high-temperature reactions only
Enforced self-consistent rate rules
and thermodynamics
– SMILES strings identified for all species
![Page 15: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/15.jpg)
Improvements were made to address over-
prediction of laminar flame-speeds
Sensitivity analysis pointed to C0-C3 chemistry
as sources of systematic error
Changes made to C0-C3 core chemistry:
– H2 oxidation Updated based on recent studies by Curran et al. (2004) and Dryer et al.
(2007)
Updated Hf298K for OH and HO2
Significant effect on flame speeds
– C1 oxidation Updated based on Petersen et al. (2007)
– C2 and C3 oxidation Updated based on Naik and Dean (2006)
All reverse rate constants based on microscopic reversibility
15
![Page 16: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/16.jpg)
A NOx sub-mechanism was assembled,
including low- and mid-temperature pathways
NOx sub-model from GRI-mech 3.0
Dagaut, Glarborg et al. 2008
Rasmussen, Glarborg, et al. 2008 Low-T
Mid-T
High-T
Based on recent mechanisms reported in literature
– GRI 3.0 NOx sub-mechanism – High-T
– Dagaut, Glarborg, et al. 2008 mechanism – Mid-T
Complete and up-to-date HCN chemistry, as well as N2O and NNH
chemistry
– Rasmussen, Glarborg, et al. 2008 mechanism – Low-T
NOx-HC interactions
Final NOx sub-model includes fuel-NOx sensitization and
self-consistent set of thermodynamic properties
![Page 17: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/17.jpg)
Detailed mechanisms were systematically
reduced to high-temperature versions
High-temperature mechanisms extracted
based on chemistry logic
– Remove species deemed to be only important for low-
temperature chain-branching
Ketohydroperoxides and QOOH species
– Remove reactions associated with removed species
Method based on unique species identifiers
– SMILES strings tag each species in system
– Independent of any species naming convention
– Allows full automation for mechanism operations
Resulting mechanism:
– 681 species, 3934 reactions 17
![Page 18: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/18.jpg)
Improved mechanism provides reliable
flame-speed predictions for n-alkanes Tested with smaller alkane data from the literature first
Compares well with USC JetSurf mechanism
![Page 19: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/19.jpg)
Extensive study at USC resulted in high-
quality, reproducible data for liquid fuels
Effect of flow rate (strain rate)
on unburned gas temperature
– Temperature correction required
Effect of radial location of
measurement for velocities
– Consistent placement very close to
centerline required for reproducibility
19
![Page 20: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/20.jpg)
A wide range of data was collected at USC
for fuel comparison and model validation
Laminar flame speeds
Laminar flame extinction strain rates
NOx in premixed flames
Ignition temperature for premixed flames
Data was collected for real jet fuels and
surrogate components
![Page 21: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/21.jpg)
Results show strong similarity between fuels
for flame propagation and extinction JP-7, S-8, R-8 and n-dodecane give same flame speeds
– JP-8 is slightly lower
JP-7, S-8, R-8 show the same extinction strain rates
– n-dodecane is slightly higher
![Page 22: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/22.jpg)
We performed a large number of simulations
to validate our chemistry model
Laminar flame-speeds for three surrogate components:
iso-octane
n-decane
n-dodecane
![Page 23: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/23.jpg)
We compared the model to literature data, as
well as to USC flame data
Ignition-delay time and species profiles in flames
for surrogate-fuel components
iso-octane/O2/Ar at 1.5 atm
Ignition data of Davidson et al. 2002
n-decane/air at 13 atm,
phi=1
Ignition data of Zhukov et al. 2008
Burner-stabilized flame species-profile data of
Doute et al. 1997
n decane/O2/N2 at 1 atm, fuel-rich
![Page 24: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/24.jpg)
We also looked at very recent ignition
temperature measurements
Data from Bieleveld et al. 2009
![Page 25: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/25.jpg)
Comparisons of the surrogate model with
F-T fuel data show good agreement
Measured species in the well-stirred reactor exhaust
by Stouffer et al. 2007
Extinction strain rate data from USC
Laminar flame-speed data from USC
Flame-speed, flame-extinction, and species data for
F-T fuel sample, S-8
![Page 26: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/26.jpg)
The NOx sub-mechanism has also been
tested against USC flame data
Agreement not as good for larger
hydrocarbons
![Page 27: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/27.jpg)
Uncertainty and reaction path analysis
suggest discrepancy may be due to data
Reaction path analysis show same dominant
reactions under both conditions
Uncertainty analysis suggests small perturbations in
velocity or temperature measurements could account
for difference
![Page 28: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/28.jpg)
Two automated mechanism-reduction
methods in CHEMKIN have been tested
Directed Relation Graph (DRG) method
– Produces skeletal mechanism
– Maintains original rates and species identity
Computational Singular Perturbation (CSP) method
– More severe reduction, based on quasi-steady assumptions
– Global, lumped reactions solved for active species
– Algebraic set of equations solved for non-active species
– Requires skeletalization (DRG) as preliminary step in reduction
Both methods are fully integrated into a (pre-release)
version of CHEMKIN-PRO
![Page 29: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/29.jpg)
Results show that accuracy can be
maintained with about 85% reduction
Mechanism # of species after DRG
reduction # of species after subsequent CSP
reduction
Master mechanism contains 549 species
Reduced mech 1 174 148
Reduced mech 2 95 95 (no CSP applied)
Reduced mech 3 64 56
S-8 Surrogate Model
![Page 30: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/30.jpg)
Outline
Overview
– Objectives of project
– Tasks and participants
Results
– Surrogate blend for F-T and bio-derived jet fuels
– Experimental data obtained at USC
– Mechanism validation
– Mechanism reduction
Conclusions
– Comparisons of F-T and bio-derived jet fuels with
conventional fuels
– Outlook for detailed kinetics in CFD simulations
![Page 31: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/31.jpg)
Conclusions (1 of 2)
Comparing F-T fuels to bio-derived jet fuels, we find no
difference in behavior
Despite differences in the C# distribution for two F-T
fuels, the flame speed & extinction were the same
– Long-chain normal alkanes > C6 have similar flame behavior
– Still may be important to distinguish for NOx emissions
Comparing F-T fuels with JP-7 and JP-8, we found that
the F-T fuels have same laminar flame-speed as JP-7
n-dodecane also shows similar flame behavior as F-T
– This is a reasonable 1-component surrogate for flame-speed and
extinction behavior only
– Need more complex surrogate for other fuel properties and emissions
![Page 32: Ellen Meeks Fuels Research Review September 16, 2009](https://reader031.fdocuments.net/reader031/viewer/2022012409/616a4cf011a7b741a350ffb0/html5/thumbnails/32.jpg)
Conclusions (2 of 2)
Our chemistry model underwent much improvement
during the course of the project
– Flame-speed, flame-extinction, ignition and NOx predictions are within experimental uncertainty
A 3-component fuel surrogate for the F-T and bio-
derived jet fuels matches data well
– n-dodecane, n-decane, iso-octane
Automated mechanism reduction provides a practical
model for use with CFD
The chemistry models are available and will be
published with NASA’s approval
A new CHEMKIN-based extinction model is currently
in beta testing