Titles and Authors of ISF-3 Poster Abstracts
1. Mobility Size and Mass of Nascent Soot Particles in a Benchmark Premixed Ethylene Flame
(ISF-3 Premixed Flame 6)
J. Camacho, C. Liu, C. Gu, H. Lin, Z. Huang, Q. Tang, X. You, C. Saggese, Y. Li, H. Jung, L. Deng,
I. Wlokas, H. Wang
2. Thermal and soot fields measured in forced laminar non-premixed co-flowing jet flames
K.K. Foo, Z.W Sun, B. B. Dally, G.J. Nathan, Z. T. Alwahabi, P. R. Medwell
3. Mobility Size Distribution of Soot in Premixed Ethylene and Propene Flames
C. Gu, H. Lin, J. Camacho, B.Y. Lin, C. Shao, R.X. Li, H. Gu, B. Guan, Z. Huang, H. Wang
4. Dependence of Soot Surface Reactivity on its Internal Nanostructure and Chemical
Composition
M. R. Kholghy, M. J. Thomson
5. Preliminary experimental and numerical results of soot volume fraction for an ethylene coflow laminar diffusion flame (ISF-3 Smooke/Long burner)
M. Roussillo, B. Franzelli, A. Cuoci, P. Scouflaire, S. Candel
6. Soot formation characteristics of n-heptane/toluene mixtures in a burner-stabilized stagnation
premixed flame
Q. Tang, B. Ge, X. You
7. Soot formation in premixed rich oxygen-enhanced methane-flames
P. Vlavakis, M. Sentko, A. Loukou, B. Stelzner, D. Trimis
8. Formation and emission of large furans and oxygenated hydrocarbons from flames
K.O. Johannson, T. Dillstrom, M. Campbell, M. Monti, F. El-Gabaly, P. Elvati, D. Popolan-Vaida, N.
Richards-Henderson, P. Schrader, K. Wilson, A. Violi, H. Michelsen
9. Effects of pressure on soot morphology in an ethylene/air counterflow diffusion flame at
elevated pressures
H. M. F. Amin, W. L. Roberts
10. DNS Effects of hydrodynamics and mixing on soot formation and growth in laminar coflow
diffusion flames at elevated pressures “ISF-3 target flame 2”
A. Abdelgadir, I. A. Rakha, S.A. Steinmetz, A. Attili, F. Bisetti, W. L. Roberts
11. Monte Carlo simulation of nascent soot particles in a benchmark premixed flame
“ISF-3 Premixed Flame 6"
A. Abdelgadir, M. Lucchesi, A. Attili, F. Bisetti
12. A computational study of soot formation in opposed-flow diffusion flame
P. Selvaraj, P. G. Arias, H. G. Im
13. A sectional PAH model with reversible PAH chemistry for CFD soot simulations
C. Eberle, P. Gerlinger, M. Aigner
14. Predicting Soot in Flames formed with Aromatic Fuels
V.R. Katta, W.M. Roquemore
Titles and Authors of ISF-3 Poster Abstracts
15. Formation of Nascent Soot Clusters from Polycyclic Aromatic Hydrocarbons: A ReaxFF
Molecular Dynamics Study
Q. Mao, A.C.T. van Duin, K. H. Luo
16. Simulation of Laminar Sooting Flames using Sectional Method and Detailed Soot Surface
Mechanism
C. V. Naik, K. Puduppakkam, A. Modak, and E. Meeks
17. Molecular Dynamics Simulations for Structural Analysis of Combustion- Generated Particles
L. Pascazio, M. Sirignano, A. D’Anna
18. Influence of pentanol isomer additions on the sooting tendency of hydrocarbon blends
A. Matynia, P. Jacobs, W. Merhy, J. Bonnety, P. da Costa, G. Legros
19. Implementation of the detailed Naples soot model in the context of a quadrature-based
method of moment
S. Salenbauch, M. Sirignano, D.L. Marchisio, M. Pollack, A. D’Anna, C. Hasse
20. Trasient Flamelet Modeling of Laminar Non-Premixed Smoking and Non-Smoking
Ethylene/Air Flames
T.H. Kim, N.S. Kim, Y. Kim
21. Modeling for Turbulent C2H4/Air Non-premixed Sooting Flames
T.H. Kim, N.S. Kim, S.T. Jeon, Y. Kim
22. Effects of laser pulse shape and duration on particle size determination from Time Resolved LII
at elevated pressures
S.A. Steinmetz, E. Cenker, W.L. Roberts
23. An optical technique for the detection and characterisation of combustion formed particles in
laminar and turbulent flames
D. Bartos, M. Dunn, M. Sirignano, A. D’Anna, A. R. Masri
24. Experiments on Sooting Turbulent Non-Premixed Flames at Elevated Pressures
W.R. Boyette, S. Chowdhury, E. Cenker, W.L. Roberts
25. The Yale Coflow Burner as a Tool for Studying Soot Formation
D.D. Das, D. Giassi, N.J. Kempema, M.B. Long, C.S. McEnally, L.D. Pfefferle
26. Measurements in Turbulent Ethylene/Propylene Flames
K.N.G. Hoffmeister, T.W. Grasser, J.C. Hewson, S.P. Kearney
27. Measurements of soot concentrations, soot temperatures, and thermal radiation from a
turbulent, non-premixed, pre-vaporized aviation fuel jet flame
C.R. Shaddix, T.C. Williams, J. Zhang
28. Large Eddy Simulations of turbulent jet flames burning ethylene and a kerosene surrogate fuel
N. Burali, G. Blanquart
29. A multi-sectional / MMC-LES approach for the Adelaide/Delft flame
A. Majbouri, M.J. Cleary, J.H. Kent, M. Sirignano, A. D. Anna
Titles and Authors of ISF-3 Poster Abstracts
30. Large Eddy Simulation of a Turbulent Sooting Ethylene/Air Diffusion Flame Using a Detailed
Sectional Soot Model
P. Rodrigues, B. Franzelli, R. Vicquelin, O. Gicquel, N. Darabiha
31. Transported PDF Modelling of Soot in the Sandia Jet Flame
M.A. Schiener, R.P. Lindstedt
32. Large-Eddy Simulation of Soot Formation in a Model Aero Engine Combustor
A. Wick, F. Priesack, H. Pitsch
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Mobility Size and Mass of Nascent Soot Particles
in a Benchmark Premixed Ethylene Flame (ISF-3 Premixed Flame 6)
Joaquin Camacho1, Changran Liu1, Chen Gu2, He Lin2, Zhen Huang2, Quanxi Tang3, Xiaoqing
You3, Chiara Saggese4, Yang Li5, Heejung Jung5, Lei Deng6, Irenaeus Wlokas6, Hai Wang1
1
Mechanical Engineering Department, Stanford University, Stanford, California 94305-3032, USA
Key Laboratory for Power Machinery and Engineering of M.O.E, Shanghai Jiao Tong University,
Shanghai 200240, China
3
Center for Combustion Energy, and Key Laboratory for Thermal Science and Power Engineering
of the Ministry of Education, Tsinghua University, Beijing 100084, China
4
Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Politecnico di
Milano, Piazza Leonardo da Vinci 32, 20133 Milano, Italy
5
Department of Mechanical Engineering, and Center for Environmental Research and Technology
(CE-CERT), College of Engineering, University of California Riverside, Riverside, California
92521, USA
6
Institute for Combustion and Gasdynamics - Fluid Dynamics, University of Duisburg-Essen,
47057 Duisburg, Germany
2
The burner stabilized stagnation flame technique coupled with micro-orifice probe sampling and
mobility sizing has evolved into a useful tool for examining the evolution of the particle size
distribution of nascent soot in laminar premixed flames. Several key aspects of this technique are
examined through a multi-university collaborative study that involves both experimental
measurement and computational modeling. Key issues examined include (a) data reproducibility
and facility effects using four burners of different sizes and makers over three different facilities, (b)
the mobility diameter and particle mass relationship, and (c) the degree to which the finite orifice
flow rate affects the validity of the boundary condition in a pseudo one dimensional stagnation flow
flame formulation. The results indicate that different burners across facilities yield nearly identical
results after special attention is paid to a range of experimental details, including a proper selection
of the sample dilution ratio and quantification of the experimental flame boundary conditions. The
mobility size and mass relationship probed by tandem mass and mobility measurement shows that
nascent soot with mobility diameter as small as 15 nm can deviate drastically from the spherical
shape. Various non-spherical morphology models using a mass density value of 1.5 g/cm3 can
reconcile this discrepancy in nascent soot mass. Lastly, two-dimensional axisymmetric simulations
of the experimental flame with and without the sample orifice flow reveals several problems of the
pseudo one-dimensional stagnation flow flame approximation. The impact of the orifice flow on the
flame and soot sampled, although small, is not negligible. Specific suggestions are provided as to
how to treat the non-ideality of the experimental setup in experiment and model comparisons.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Thermal and soot fields measured in forced laminar non-premixed co-flowing
jet flames
K.K. Foo1, Z.W Sun1, B.B Dally1, G.J. Nathan1, Z.T. Alwahabi1, P.R. Medwell1
1
School of Mechanical Engineering, The University of Adelaide, Australia
kae.foo@adelaide.edu.au
Phase-resolved measurements of the flame temperature, soot volume fraction and primary soot particles
diameter in acoustically forced, laminar non-premixed co-flowing jet flames were conducted. These
measurements were used to study the effect of periodic oscillations and the accompanying vortex structures
on the thermal fields and the soot generated within the flames. Two oscillation frequencies were considered,
one closed to the natural flickering frequency of the flames (20 Hz) and its harmonics (40 Hz). To isolate the
effect of the frequency, the amplitude was maintained constant in both cases. Constant Temperature
Anemometry (CTA) was used to measure the axial velocity of the fuel exit at the centreline, for the nonreacting case. The amplitude of the forcing was then calibrated based on the axial velocity measured. A speaker
was used to modulate the fuel gas flow rate with an amplitude of 50% of the average gas flow. Two-line
Atomic Fluorescence (TLAF) was employed to measure the temperature fields within the flames. Timeresolved Laser Induced Incandescence (Ti-Re LII) was used to study the properties of the soot within the
flames. While the soot volume fraction was derived from the prompt signal of the Ti-Re LII, the time-resolved
signals were used to calculate the diameters of the primary soot particles generated within the flames. Measured
scalars and their dependency on the Strouhal number will be presented and discussed. In the future, three
burners with inner diameters of 4 mm, 5.6 mm, and 8 mm will be employed in the studies to assess the
dependence of burner geometry. The mean flow Reynolds number was kept constant at Re=200, which led to
the change in jet exit velocity as the diameter of the burners are different. In consequence, the Strouhal number
increased as the diameter of the burner increased. It aims to study the soot behaviour in a broader range of
local strain rate.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Mobility Size Distribution of Soot
in Premixed Ethylene and Propene Flames
C. Gu1, H. Lin1, J. Camacho2, B.Y. Lin1, C. Shao1, R.X. Li1, H. Gu1,
B. Guan1, Z. Huang1, H. Wang2
1
Key Laboratory for Power Machinery and Engineering of M.O.E, Shanghai Jiao Tong University,
Shanghai, PR China
2
Department of Mechanical Engineering, Stanford University, Stanford, CA, USA
linhe@sjtu.edu.cn
The evolution of mobility particle size distribution (PSDF) in premixed ethylene and propene flames
was systematically investigated in the burner stabilized stagnation (BSS) flame configuration.
The result shows several similar sooting behaviors between premixed ethylene and propene flames as
follows. Qualitatively, bimodal PSDFs were observed in both ethylene and propene flames. The evolution of
the PSDFs with respect to flame stoichiometry, temperature and growth time is consistent with the
understanding of kinetic competition during soot formation. Finite rate kinetic limitations are observed at
lower temperatures and thermodynamic reversibility occurs at higher temperatures. The observed PSDF
features are highly sensitive to competition among the various processes of soot formation, from nucleation
to coagulation and gas-surface reactions. The evolution of the PSDF indicates a strong contribution to the
mass of coagulation-mode by the nucleation-mode particles.
The comparison of PSDFs between premixed ethylene and propene flames shows that soot nucleates
and grows in propene flames earlier and faster than in ethylene flames under comparable conditions. In
particular, the propene flames produce a substantially higher number of soot nuclei than in the ethylene
flames. The observations made for the ethylene and propene flames strongly indicate that parent fuel
structure has a notable effect on soot formation in premixed flames both in the global properties of soot
produced (volume fraction, number density etc) and detailed evolution of soot PSDFs and particle kinetics
including the onset of nucleation, persistency of nucleation during soot growth, and the surface growth.
Lastly, we note that the data reported here, including the temperature profiles and the soot PSDFs,
should be useful to soot model development and test.
References
[1] Camacho J, Liu C, Gu C, et al. Mobility size and mass of nascent soot particles in a benchmark premixed ethylene
flame[J]. Combustion and Flame, 2015, 162(10): 3810-3822.
[2] Gu C, Lin H, Camacho J, et al. Particle size distribution of nascent soot in lightly and heavily sooting premixed
ethylene flames[J]. Combustion and Flame, 2016, 165: 177-187.
[3] Gu C, Lin H, Camacho J, et al. Mobility size distribution of soot in premixed propene flames. Combustion and
Flame, 2016, (Accepted).
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Dependence of Soot Surface Reactivity on its Internal Nanostructure
and Chemical Composition
Mohammad Reza Kholghya1, Murray John Thomsona*
a
Department of Mechanical and Industrial Engineering, University of Toronto, 5 King’s College
Road, Toronto, Ontario, Canada, M5S 3G8
1
mkholghy@mie.utoronto.ca
A novel model called soot Surface Shell Formation (SSF) is used to simulate soot aging and
investigate the effects of soot internal nanostructure and chemical composition on its surface
reactivity in terms of the number density of the surface hydrogenated sites (𝜒𝑐𝑠𝑜𝑜𝑡−𝐻 ) in laminar
coflow ethylene flames. SSF predicts the average PAH molecular weight and the characteristic
shell-core structure of the soot primary particles based on the equilibrium internal nanoarrangement of Polycyclic Aromatic Hydrocarbons (PAHs) inside the primary particles. This
knowledge of the PAH arrangements is then employed to distinguish mature from nascent soot
particles based on the presence of the graphitic shell. It is shown that the shell formation for mature
particles changes the configuration of the surface PAHs from Edge On Surface (EOS) to Face On
Surface (FOS) which significantly and rapidly reduces the number of available surface
hydrogenated carbon (i.e. 𝐶𝑠𝑜𝑜𝑡 − 𝐻 ) sites for chemical growth reactions. The necessity of
considering the effects of soot internal nanostructure on its surface reactivity in terms of 𝜒𝑐𝑠𝑜𝑜𝑡−𝐻 is
illustrated by comparing the performances of four models which consider constant surface density
of available hydrogenated sites, i.e. 𝜒𝑐𝑠𝑜𝑜𝑡−𝐻 , as a function of soot chemical composition, 𝜒𝑐𝑠𝑜𝑜𝑡−𝐻 as
a function of soot internal nanostructure and 𝜒𝑐𝑠𝑜𝑜𝑡−𝐻 as a function of both soot chemical
composition and internal nanostructure. It is shown that the model which simultaneously relates
𝜒𝑐𝑠𝑜𝑜𝑡−𝐻 to its internal nanostructure and chemical composition has the best performance. For the
first time, SSF enables us to simultaneously relate soot chemical composition and its internal
nanostructure to 𝜒𝑐𝑠𝑜𝑜𝑡−𝐻 which is the dominant factor in determining soot surface reactivity.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Preliminary experimental and numerical results of soot volume fraction for an
ethylene co-flow laminar diffusion flame (ISF-3 Smooke/Long burner)
a
b
M. Roussilloa, B. Franzellia, A. Cuocib, P. Scouflairea, S. Candela
EM2C Laboratory, CNRS, CentraleSupélec, Université Paris-Saclay, 92295 Châtenay-Malabry, France
Departement of Chemistry, Materials and Chemical Engineering ’G. Natta’, Politecnico di Milano, 20133
Milano, Italy
Corresponding author: Benedetta Franzelli (benedetta.franzelli@cnrs.fr)
Soot production resulting from an incomplete combustion of hydrocarbons is a complex
process [1], associated with production processes that are not completely understood, in
particular, in a turbulent context. Investigations of laminar diffusion flames are essential for
the characterization of soot production because:
(1) Flames of this type provide a fundamental understanding of the physical processes
underlying soot formation in a well-controlled configurations,
(2) Such flames often serve to calibrate Laser Induced Incandescence (LII) techniques [2]
now extensively used to measure soot volume fraction in both laminar and turbulent
flames,
(3) These flames allow detailed reference calculations that can be validated by direct
comparisons with experiments before developing reduced models for turbulent
simulations.
In this general context, the reference ISF-3 Smooke/Long burner [2] is investigated
experimentally and subjected to detailed numerical simulations. The primary objective of
this work is to verify that the experimental and numerical tools available at EM2C and
Politecnico di Milano provide adequate experimental data and numerical results for the
investigation of soot and that these data and results retrieve those already available in the
literature. This work specifically reports:
• LII measurements of the reference ISF-3 Smooke/Long burner;
• Two-dimensional numerical simulations using a detailed kinetic scheme [3]
combining a detailed gas-phase mechanism and a detailed soot description.
Experimental observations indicate that the flame is quite sensitive to the operating
conditions and that it is subject to flickering, so that retrieving the results of the literature is
a challenging task. Improving the spatial stability of the flame requires slight modifications
of the operating conditions. These changes are described and their impacts on soot volume
fraction predictions are investigated numerically. Comparisons between the new
experimental results, results from the literature and the present numerical calculations are
finally discussed.
Bibliography
[1] H. Bockhorn, Soot Formation in Combustion (1994).
[2] M.D. Smooke, MD, M.B. Long, B.C. Connelly, M.B. Colket, R.J. Hall, Combust. Flame 143
(2005) 613–28.
[3] C. Saggese, S. Ferrario, J. Camacho, A. Cuoci, A. Frassoldati, E. Ranzi, H. Wang, T. Faravelli,
Combust. Flame 162 (2015) 3356–3369.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Soot formation characteristics of n-heptane/toluene mixtures in a burnerstabilized stagnation premixed flame
Quanxi Tang1,2, Boqing Ge1, Xiaoqing You1,2*
1
2
Center for Combustion Energy, Tsinghua University, Beijing 100084, China
Key Laboratory for Thermal Science and Power Engineering of the Ministry of Education,
Tsinghua University, Beijing, China
*xiaoqing.you@tsinghua.edu.cn
The soot formation characteristics of pure n-heptane and binary mixtures of toluene and n-heptane
with their liquid volume ratios ranging from 0.2-1 were studied. The C/O ratio and unburned gasmixture velocity were kept the same for all conditions. The soot particle size distribution functions
(PSDFs) at different burner-to-stagnation surface distances were measured by using the burnerstabilized stagnation (BSS) probe/ scanning mobility particle sizer (SMPS) technique. Meanwhile
the morphology of soot particles taken from the same probe was examined using the transmission
electron microscopy (TEM). With the amount of toluene addition increases, the experimental results
show that the particle nucleation is enhanced, while the primary particle sizes are smaller and the
soot volume fraction decreases. To explain the experimental observations, numerical simulations of
the gas-phase flame chemistry were carried out using measured temperature profile to obtain
species profiles along the distance above the burner surface. The results are consistent with our
PSDF and TEM observations, as in the toluene-added flames more benzene and pyrene but less
acetylene are produced, which may promote soot nucleation but slow down the soot surface growth
separately.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Soot formation in premixed rich oxygen-enhanced methane-flames
P. Vlavakis1, M. Sentko1, A. Loukou1, B. Stelzner1, D. Trimis1
1
Department of Chemical and Process Engineering, Karlsruhe Institute of Technology, Germany
petros.vlavakis@kit.edu
The present study deals with the issue of soot formation in premixed methane flames under conditions that
are representative for industrial high-temperature processes and in particular for gasification. In this context,
the focus of the experimental investigations is set at high equivalence ratios, ranging from 2.4 to 2.7 and at
high oxygen content in the oxidizer stream, from 50% to 80%, with the rest being Argon. The examined
flames are stabilized using a heat-flux burner [1] in order to ensure laminar conditions and one-dimensional
flame structure. Soot particle size distributions (SPSDs) are determined using online sampling and a
scanning mobility particle sizer. The applied sampling method extracts the soot-containing gas probe through
a pinhole in a pipe where it is rapidly diluted (dilution ratio > 10 4) with cold nitrogen [2]. Gas temperature
profiles along the burner axis are obtained from radiation corrected fine-wire thermocouple measurements.
In preliminary investigations, the laminar burning velocity was determined for the selected flames and
subsequently used as inlet velocity for the soot measurements. The influence of C/O-ratio was investigated
for a fixed oxidizer composition, 50% O2 - 50% Ar and equivalence ratios of 2.4, 2.5, 2.6 and 2.7. The
sooting limit was approximately at C/O of 0.6, which corresponds to an equilibrium temperature below
2000 K. In this case, a two order of magnitude increase in the soot volume fraction (from 10-4 to 10-2 ppm)
was observed with ϕ increasing from 2.4 (C/O=0.60) to 2.5 (C/O=0.63). The temperature influence on soot
formation was studied keeping ϕ constant at 2.6 and varying the Argon-dilution in the oxidizer stream from
50% down to 20% approaching oxyfuel conditions. The results show that with decreasing dilution the flame
temperature increases and consequently the soot volume fraction decreases for about one order of magnitude,
from 10-1 ppm for 50% dilution to 10-2 ppm for 20% dilution. Current investigations focus at spatial
resolution of SPSDs to study their evolution within the soot growth zone above the burner.
Reference(s)
[1]
[2]
L.P.H. De Goey et al., Combust. Sci. Technol. 92 (1993) 201–207.
B. Zhao et al., Aerosol Sci. Technol. 37 (2003) 611-620.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Formation and emission of large furans and oxygenated hydrocarbons from flames
K. Olof Johansson1, Tyler Dillstrom2, Matthew Campbell1, Matteo Monti3, Farid El-Gabaly4, Paolo
Elvati2, Denisia Popolan-Vaida5,6, Nicole Richards-Henderson5, Paul Schrader1, Kevin Wilson5,
Angela Violi2,7,8,9, Hope Michelsen1
1
Combustion Research Facility, Sandia National Laboratories, Livermore, CA, USA
2
Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI, USA
3
Department of Materials Science & Engineering, Stanford University, Stanford, CA, USA
4
Materials Physics, Sandia National Laboratories, Livermore, CA, USA
5
Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA
6
Department of Chemistry, University of California, Berkeley, CA, USA
7
Department of Chemical Engineering, University of Michigan, Ann Arbor, MI, USA
8
Department of Macromolecular Science and Engineering, University of Michigan, Ann Arbor, MI,
USA
9
Biophysics Program, University of Michigan, Ann Arbor, MI, USA
okjohan@sandia.gov
Many oxygenated hydrocarbon species formed during combustion, such as furans, are highly toxic and
detrimental to human health and the environment [1-3]. Oxygen atoms bound to organic species on the
particle surface greatly affect soot hygroscopicity [4] and can thus affect the ability of soot particles to act as
cloud-condensation or ice nuclei. The effect of soot emissions on cloud-nucleation properties is a major
uncertainty in climate predictions [5-7]. However, large furans and associated oxygenated species have not
previously been observed in flames, and their formation mechanism and interplay with polycyclic aromatic
hydrocarbon (PAH) species are poorly understood. We present a synergistic computational and experimental
effort that elucidates the formation of oxygen-embedded compounds, such as furans and other oxygenated
hydrocarbon species, during the combustion of hydrocarbon fuels. We used ab initio and probabilistic
computational techniques to identify low-barrier reaction mechanisms for the formation of large furans and
other oxygenated hydrocarbons. We used vacuum-UV photoionization aerosol mass spectrometry and X-ray
photoelectron spectroscopy to confirm these predictions. We show that furans are produced in the hightemperature regions of hydrocarbon flames and become the main functional group of oxygenates that
incorporate into incipient soot. We discovered ∼100 oxygenated species previously unaccounted for. We
found that large alcohols and enols act as precursors to furans, leading to incorporation of oxygen into the
carbon skeletons of PAHs. Our results depart dramatically from the crude chemistry of carbon and oxygencontaining molecules previously considered in hydrocarbon formation and oxidation models and spearhead
the emerging understanding of the oxidation chemistry that is critical, for example, to control emissions of
toxic and carcinogenic combustion by-products, which also greatly affect global warming.
Reference(s)
[1]
[2]
[3]
[4]
[5]
[6]
[7]
L.A. Peterson, Drug. Metab. Rev. 38 (4) (2006) 615-626.
V. Ravindranath, L.T. Burka, M.R. Boyd, Science 224 (1984) 884-886.
B.H. Monien, K. Hermann, S. Florian, H. Glatt, Carcinogenesis (2011) bgr126.
M. Commodo, G.D. Falco, R. Larciprete, A. D'Anna, P. Minutolo, Experimental Thermal
and Fluid Sciences (2015).
R. Zhang, A.F. Khalizov, J. Pagels, D. Zhang, H. Xue, P.H. McMurry, Proceedings of the
National Academy of Sciences 105 (30) (2008) 10291-10296.
Y.J. Kaufman, I. Koren, L.A. Remer, D. Rosenfeld, Y. Rudich, Proceedings of the National
Academy of Sciences of the United States of America 102 (32) (2005) 11207-11212.
V. Ramanathan, C. Chung, D. Kim, T. Bettge, L. Buja, J. Kiehl, W. Washington, Q. Fu, D.
Sikka, M. Wild, Proceedings of the National Academy of Sciences of the United States of
America 102 (15) (2005) 5326-5333.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Effects of pressure on soot morphology in an ethylene/air counterflow diffusion
flame at elevated pressures
Hafiz M. F. Amin, William L. Roberts
Clean Combustion Research Center, King Abdullah University of Science and Technology
23955-6900 Thuwal, Saudi Arabia
hafiz.amin@kaust.edu.sa
High operating pressures are desirable in gas turbines and diesel engines to improve their thermodynamic
efficiency, but soot formation is increased with pressure. Soot emissions from these combustion devices have
serious effects on human health and the environment. Transport properties of soot and its interaction with
human respiratory system depend on its morphology. Therefore, it is necessary to understand soot
morphology at conditions relevant to practical combustion systems, in order to develop better emission
control strategies. In this work, effect of pressure on soot morphology is investigated using light scattering
and extinction technique. A nitrogen diluted ethylene/air counterflow diffusion flame has been stabilized, up
to 5 atm, in a pressure vessel that can provide optical access for multi-angle light scattering. An Ar/Kr ion
laser beam is passed through the flame and the light scattered by soot is measured using photomultiplier
tubes. These PMTs are mounted on a rotary stage that allows positioning them at desired angles. Multi-angle
light scattering and extinction measurements are performed in a counterflow diffusion flame up to 5 atm. At
all pressures, global strain rate (a) is maintained constant by adjusting the inlet mass flow rates. RayleighDebye-Gans polydisperse fractal aggregate (RDG-PFA) scattering interpretation is used to calculate soot
volume fraction, primary particle diameter, aggregate size distribution, mean radius of gyration of aggregates
and fractal dimension of soot. Figure 1 shows the soot volume fraction (fv) profiles measured along the
centreline of counterflow diffusion flame and a significant increase in the value of fv is observed with
pressure.
a = 30s-1 XF = 0.3
5 atm
4 atm
3 atm
2 atm
7
6
fv [ppm]
5
4
3
2
1
0
3.0
3.5
4.0
4.5
Distance from fuel nozzle [mm]
Fig. 1. fv measured along the axis of counterflow diffusion flame from 2 to 5 atm
Effects of hydrodynamics and mixing on soot formation and growth
in laminar coflow diffusion flames at elevated pressures
”ISF-3 target flame 2”
Ahmed Abdelgadir, Ihsan Allah Rakha, Scott A. Steinmetz, Antonio Attili,
Fabrizio Bisetti, William L. Roberts
A set of steady, laminar coflow diffusion flames are simulated to study the formation, growth, and oxidation
of soot in ethylene flames at pressures ranging from 1 to 8 atm. Our modeling approach combines detailed
finite rate kinetics mechanisms that model the formation of Polycyclic Aromatic Hydrocarbon (PAH) species
up to pyrene and a bivariate method of moments that describes soot particles and aggregates by their volume
and surface area. We assess the robustness of our conclusions with respect to the chemical kinetics model
by repeating simulations using two kinetics mechanisms of similar complexity. We also present detailed
comparisons with experimental data on PAH species and the soot volume fraction. We find that, since
fuel flow rates are kept constant as pressure increases, the scalar dissipation rate decreases, promoting the
formation of PAH species and soot. The scaling of the scalar dissipation rate is not straightforward due to
buoyancy effects, which result in narrowing of the flame and steepening of the mixture fraction gradients.
One of the two kinetics mechanisms predicts the spatial distribution of soot accurately, although the peak
value of the soot volume fraction is underestimated by more than a factor of three compared to measurements
based on light extinction. Soot growth was found to occur primarily through routes based on PAH species
such as nucleation and condensation, while HACA growth plays a small role. These trends are observed on
the centerline as well as on the flames wings. The effect of the radiation heat loss is explained in details and
related to the flame length and resident time.
ISF Workshop, 2016
1
Monte Carlo simulation of nascent soot particles in a benchmark
premixed flame ”ISF-3 Premixed Flame 6”
Ahmed Abdelgadir, Marco Lucchesi, Antonio Attili, Fabrizio Bisetti
A modeling frame-work based on Direct Simulation Monte Carlo (DSMC) is applied to simulate and
analyze the evolution of the soot particle size distribution in a benchmark ethylene-oxygen premixed flame.
The experimental measurements were taken in different facilities to insure data reproducibility [1]. The flame
is a burner-stabilized stagnation flame coupled with micro-orifice probe sampling and mobility sizing. The
DSMC approach is applied to the one-dimensional solution obtained using OPPDIF from the CHEMKINPRO package. The probe effect is accounted for by applying the correction to the sampling location as
proposed by Saggese et al. [2]. A good agreement between the 1D solution and the experimental data is
found for temperature. Comparisons are obtained for five different distances from the burner, at 0.4, 0.6, 0.8,
1, and 1.2 cm. The soot number density and soot volume fraction from DSMC are in a good agreement. At
large distances from the burner, the PSDF obtained via Monte Carlo is well reproduced. Both the span of
the tail of the second mode and the position of the trough between the modes are accurately captured. At
locations close to the burner, soot quantities are not reproduced as accurately. We relate these differences
to the inability of the one-dimensional solution to approximate the two-dimensional fields in the case of very
small distances between the sampling plate and the burner.
Referances:
[1] J. Camacho, C. Liu, C. Gu, H. Lin, Z. Huang, Q. Tang, X. You, C. Saggese, Y. Li, H. Jung, et al.,
Mobility size and mass of nascent soot particles in a benchmark premixed ethylene flame, Combust. Flame
162 (2015) 3810-3822.
[2] C. Saggese, A. Cuoci, A. Frassoldati, S. Ferrario, J. Camacho, H. Wang, T. Faravelli, Probe effects in
soot sampling from a burner-stabilized stagnation flame, Combust. Flame 167 (2016) 184-197.
ISF Workshop, 2016
1
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
A computational study of soot formation in opposed-flow diffusion flame
Prabhu Selvaraj1, Paul G. Arias2 and Hong G. Im1
1
Clean Combustion Research Center, King Abdullah University of Science and Technology
(KAUST), Thuwal 23955-6900, Saudi Arabia
2
Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI 48109, USA.
prabhu.selvaraj@kaust.edu.sa
Soot formation and growth in nonpremixed flames is an important process in combustion employing
hydrocarbon fuels. Soot emission causes adverse health and environment impacts. Thus there is a stronger
need to obtain the fundamental and detailed understanding of the soot formation in a non-premixed flames in
the wider range of thermo-physical conditions. The present study investigates the structures of PAH and
sooting zones of both SF (soot formation) and SFO (soot formation oxidation) flames in counterflow
diffusion flames of ethylene. The gas-phase chemical mechanism adopted the KAUST-Aramco PAH Mech
1.0, which utilized the AramcoMech 1.3 [1] for gas-phase reactions validated for up to C2 fuels. In addition,
PAH species up to coronene (C24H12 or A7) were included to describe the detailed formation pathways of
soot precursors. In this study, the detailed chemical mechanism was reduced from 397 to 99 species using
directed relation graph with expert knowledge (DRG-X) [2] and sensitivity analysis. The method of moments
with interpolative closure (MOMIC) was employed for the soot aerosol model. In order to validate the
reduced mechanism SF flames were simulated and the results are compared [3]. The analysis is also carried
out from low to high strain rate conditions in SF flame. Further studies on the interaction of PAHs with soot
during soot growth and oxidation is examined. In the soot growth region for SFO flames, the temperature is
high and PAH concentrations relatively low, whereas in SF flames, the temperature is low and PAH
concentrations are higher than that in the corresponding SFO flame. Especially in SFO flames the impact of
oxidation pathway plays an important role [4], thus the effect of OH and O2 pathway has been analysed. The
profiles of soot volume fractions are relatively symmetric for the SFO flames due to oxidation, but they
become asymmetric and skewed for the SF flames since there is no oxidation. SFO flame soot characteristics
are further examined by varying the strain rate conditions.
References
[1] W. K. Metcalfe et al., International Journal of Chemical Kinetics, 45, 638 (2013).
[2] T. Lu et al, Proc. Comb. Inst., 30, 1333 (2005).
[3] P. Selvaraj et al., Comb & Flame, 163, 427 (2016).
[4] J. Y. Hwang and S.H. Chung,S Comb & Flame, 125, 752 (2001).
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
A Sectional PAH Model with Reversible PAH Chemistry
for CFD soot simulations
C. Eberle1, P. Gerlinger2, M. Aigner1
1
German Aerospace Center (DLR), Stuttgart, Germany
2
University of Stuttgart, Germany
Soot modelling in technical combustion by means of computational fluid dynamics (CFD) is
challenging. On one hand, due to the high computational cost, a soot model for CFD applications
can consider only the most important processes of soot evolution using as few variables as possible.
On the other hand, technical combustion usually occurs at high pressures and high temperatures and
effects such as turbulence and partial premixing lead to locally and temporally varying combustion
conditions.
One of the least understood processes of soot evolution is nucleation. In simple models, soot is
formed from acetylene while more detailed models consider the slow chemistry of polycyclic
aromatic hydrocarbons (PAHs). Since detailed PAH kinetics is costly, models with lumped PAH
species have been developed by different authors to efficiently model PAH evolution.
In previous work, a sectional model for PAHs has been introduced and validated, where PAHs with
a molar weight in the range of 100 and 800 g/mol are discretized by three sections and nucleation is
modelled by PAH growth. This model had difficulties at particularly challenging conditions (strong
partial premixing and high temperatures) which were attributed to the irreversible treatment of PAH
chemistry. To improve the predictive capability under such conditions, an extended PAH model,
where a radical branch is assigned to each PAH section, is proposed in this work. The PAH radicals
enable a reversible formulation of PAH chemistry at no significant increase of computational cost.
The new PAH model was implemented in a framework where combustion is treated by finite-rate
chemistry and soot by a sectional approach.
The reversible PAH model leads to lower soot nucleation and a better agreement to measured soot
volume fractions in a series of laminar premixed flames. By the correct description of soot
nucleation, the sensitivity of soot morphological parameters including PSD functions with respect
to the equivalence ratio in a plug flow reactor is accurately resolved, while the irreversible PAH
model failed to predict PSD functions at equivalence ratios close to the sooting limit. It is found that
reversible PAH chemistry is required to correctly predict the temperature dependence of the soot
yield in ethylene pyrolysis.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Predicting Soot in Flames formed with Aromatic Fuels
V. R. Katta1, W. M. Roquemore2
1
Innovative Scientific Solutions, Inc., Dayton, OH, USA
2
Air Force Research Laboratory, Wright-Patterson Air Force Base, OH, USA
vrkatta@gmail.com
Experiments performed for steady laminar jet flames established with paraffin fuels such as n-dodecane, nheptane/n-dodecane and iso-octane/n-dodecane yielded subtle but unique differences in the soot structure.
While the addition of n-heptane to n-dodecane decreased the soot, addition of iso-octane increased. On the
other hand, addition of aromatic fuel (m-xylene) to n-dodecane caused a three-fold increase in soot [1].
Previous modeling efforts failed to predict such high amounts of soot emissions in laminar jet flames burning
m-xylene/n-dodecane surrogate mixtures or real JP-8 fuel. CFD simulations were underestimating soot in these
flames by more than 100% (Fig. 1a). It was believed that there is something fundamentally missing in the
chemical-kinetics models for these aromatic fuels for not yielding PAH species in enough concentrations for
allowing soot formation and growth. We have undertaken an extensive effort to identify this problem. Using
UNICORN code and incorporating five state-of-the-art full chemistry models we have computed jet flames
burning various alternative fuels. We have also used three soot-modeling approaches (2-equation, method of
moments and sectional). The vast number of calculations made with the combinations from gas phase
chemistries and soot models yielded more or less the same amounts of soot. This outcome was not that
surprising as all the current gas-phase chemistries and soot models were developed from the same fundamental
belief that production of soot in flames with aromatic fuels would be a lot more compared to that in flames
with paraffin fuels. After analysing the simulation results and the measured soot distributions critically, it has
been have identified that the soot in the flames formed with aromatic fuels is actually not getting oxidized as
quickly as that is happening in the flames with paraffin fuels. This lower oxidation rate left more soot in the
flames and made it look like as if the flame was producing more soot. Based on this hypothesis a new gas
phase chemical kinetics (SERDP-2015) has been constructed and the soot sectional method has been modified
for allowing decrease in oxidation rates as the soot particle grows in size. Using the new chemistry and the
soot models we were able to, for the first time, predict the soot in a jet flame formed with m-xylene/n-dodecane
blend (Fig. 1b). Later on we applied these models for the flames in the literature [2, 3] and able to obtain much
better predictions than those reported.
Figure 1. UNICORN predictions along centerline of a laminar jet flame burning a surrogate JP-8 fuel. (a)
with SERDP fuel chemistry and old soot model and (b) with the sectional-method soot model.
Reference(s)
[1] V. R. Katta, W. M. Roquemore, “Modeling soot in flames with complex fuels,” Proceedings of the U. S.
National Meeting of the Combustion Institute, 2015, University of Cincinnati, Cincinnati, OH (2015)
[2] M. Saffaripour, A. Veshkini, M. Kholghy, M. J. Thomson, Comb. Flame, 161, 848 (2014)
[3] V. R. Katta, K. Seshadri, J. Zelina, W. M. Roquemore, “Performance of JP-8 Surrogate Models in
Predicting Laboratory Jet Flames,” AIAA-2010-950, 48th AIAA Aerospace Sciences Meeting and Exposition,
Orlando, FL (2010).
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Formation of Nascent Soot Clusters from Polycyclic Aromatic Hydrocarbons:
A ReaxFF Molecular Dynamics Study
Qian Mao1, Adri C.T. van Duin2, K. H. Luo 1,3*
1
Center for Combustion Energy, Key Laboratory for Thermal Science and Power Engineering of
Ministry of Education, Department of Thermal Engineering, Tsinghua University, Beijing 100084,
China
2
Department of Mechanical and Nuclear Engineering, The Pennsylvania State University,
University Park, PA 16802, USA
3
Department of Mechanical Engineering, University College London, Torrington Place, London
WC1E 7JE, UK
*Email: K.Luo@ucl.ac.uk
The growth of polycyclic aromatic hydrocarbon (PAH) molecules to large soot particles involves
both chemical reaction and physical processes, which is still poorly understood.1 In the present
study, ReaxFF molecular dynamics simulation3 is implemented to study the growth mechanisms of
PAHs to soot particles comprehensively. Based on the mass spectrum from experiments, the mass
of soot monomer is detected ranging from 100 aum to 700 aum4. Therefore, PAHs constituted of 2-,
3-, 4-, 7-, 10-, and 19- numbered aromatic rings, named naphthalene, anthracene, pyrene, coronene,
ovalene and circumcoronene, respectively, are selected for simulations over a range of temperatures
from 400 K to 2500 K with and without Ni cluster and N2 molecules. In the absence of metal cluster
and N2 molecules, PAHs are more likely to grow into stacked clusters at low temperatures (e.g. 400
K) due to the binding effects of electrostatic and Van der Waals forces. Firstly, PAHs grow into
staking clusters, in parallel to each other. Then collisions between clusters lead to larger soot
particles. With the rise of temperatures (e.g. 1600 K), most of the PAHs form small clusters with 23 molecules, but they can not grow further into large particles. When the temperature reaches 2500
K, all of the above PAHs are chemically active. PAHs firstly dissociate into smaller molecules and
radicals. Then small radicals help to connect PAHs to form the stacked structure, which leads to
higher binding energy. Additionally, self-assembly formation of fullerene structure is also observed
in the simulations. After adding the Ni cluster to the gas phase PAHs system, the cluster acts as a
core nuclei to facilitate both physical and chemical growth of PAHs, by extending the physical
growth temperature range as well as lowering the dissociation temperature limit of PAHs and
consequently chemical growth temperature limit. Moreover, during the formation of soot particles,
the nonreactive collision between N2 molecules and PAHs dimer is non-negligible, which
influences the dimers’ life-time. Therefore, the influence of N2 on soot formation is carefully
scrutinized.
References
[1] Wang, H., 2011. Formation of nascent soot and other condensed-phase materials in flames. Proceedings of the
Combustion Institute, 33(1), pp.41-67.
[2] Totton, T.S., Chakrabarti, D., Misquitta, A.J., Sander, M., Wales, D.J. and Kraft, M., 2010. Modelling the
internal structure of nascent soot particles. Combustion and Flame, 157(5), pp.909-914.
[3] Van Duin, A.C., Dasgupta, S., Lorant, F. and Goddard, W.A., 2001. ReaxFF: a reactive force field for
hydrocarbons. The Journal of Physical Chemistry A,105(41), pp.9396-9409.
[4] Happold, J., Grotheer, H.H. and Aigner, M., 2009. Soot precursors consisting of stacked pericondensed PAHs.
In H. Bockhorn, A. D’Anna, A.F. Sarofim and H. Wang (eds.), Combustion Generated Fine Carbonaceous
Particles: Proceedings of an International Workshop Held in Villa Orlandi, Anacapri, May 13-16, 2007 (p. 277).
KIT Scientific Publishing.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Simulation of Laminar Sooting Flames using Sectional Method
and Detailed Soot Surface Mechanism
Chitralkumar V. Naik, Karthik Puduppakkam, Abhijit Modak, and Ellen Meeks
ANSYS Inc.
Three target laminar ethylene flames, Flame 2a, Flame 3a, and Flame 6 were considered, using a
detailed gas phase combustion mechanism coupled with a detailed surface mechanism for the sootparticle nucleation, growth and oxidation kinetics. Both the gas phase and soot surface mechanisms
are similar to those reported previously [1, 2]. The gas phase mechanism contains 223 species and
1435 reactions, and the soot surface mechanism contains 15 species and 43 reactions. A sectional
method for tracking particle size distributions, as implemented in the Chemkin-Pro premixed
burner-stabilized flame and in burner-stabilized stagnation flow (BSSF) flames, has been used to
model soot volume fraction and size distribution. The model includes particle coagulation, diffusion,
and thermophoresis, as well as particle aggregation. In addition, gas-radiation and a particleradiation model using a reformulated Mie solution have been included in the flame simulations. For
Flame 6, the BSSF flame model is used while solving for energy equation. Calculated temperature
profiles for various stagnation plane heights closely matched the measured temperatures. Despite
some discrepancies for the shorter stagnation-plane heights, calculated soot volume fractions are in
reasonable agreement with the measurements. Predicted trends in soot number densities also show
nucleation-dominated soot production at the shortest residence times, transitioning to a coagulationdominated regime at the largest residence times.
References
1. Puduppakkam, K.V., et al., A Soot Chemistry Model that Captures Fuel Effects. Proc. ASME TurboExpo,
2014. GT2014-27123.
2. Naik, C.V., K.V. Puduppakkam, and E. Meeks, An Improved Core Reaction Mechanism for Un-saturated
C0-C4 Fuels and their Blends. Proc. ASME Turbo Expo, 2012. GT2012-68722.
International Sooting Flames Workshop
30 -31 July 2016 Nest Hotel, Incheon
Molecular Dynamics Simulations for Structural Analysis of CombustionGenerated Particles
Laura Pascazio, M. Sirignano, A. D’Anna
Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale - DICMaPI,
Università “Federico II”, Naples, Italy
laura.pascazio@unina.it
Introduction
Combustion-generated particle nucleation remains the least understood process of particle
formation. Stacking of polycyclic aromatic hydrocarbons (PAHs) has been recognized as key step of
particle inception, but a conclusive description of the process is not reached yet. Uncertainties remain
on the PAHs involved in the process and on their interaction mechanism. Recently, Molecular
dynamics (MD) has been proposed as a valid tool to explore the nucleation process in a physical
realistic way.
In this work, a study on the evolution of PAHs with different molecular masses and morphologies
and additional analysis on the structure of particles obtained has been made using a molecular
dynamics approach. Two different types of PAH molecules have been analysed in order to understand if they
exhibit a different coagulation efficiency: coronene (C24H12) representative of pericondensed aromatic
hydrocarbons (PCAHs), and dicoronene (C48H22) representative of aromatic aliphatic linked hydrocarbons
(AALHs). Simulations of homomolecular systems has been performed using a MD code
(GROMACS), at temperatures ranging from 500K to 1500K. The main objective is to provide a
correlation between the starting molecule morphologies and the clustering behaviours in order to gain
insights on the nucleation mechanism and soot structure.
A significative dependence of the coagulation efficiency has been found whether PCAHs or
AALHs are considered. Particle morphology has been systematically studied introducing the
distribution function of the distances between the centres of mass of the coagulated aromatics and a
structural parameter calculated from the inertia tensor of clusters, good indicator of cluster shapes. A
different size and morphology of the nascent particles has been found. Looking at the structural
parameter and at the internal disposition of molecules in the clusters, coronene clusters show mainly
an ordered arrangement of stacked molecules whereas an enhancement of disorder in the structure
and a more spherical shape has been observed for clusters of dicoronene molecules.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Influence of pentanol isomer additions on
the sooting tendency of hydrocarbon blends
A. Matynia*, P. Jacobs, W. Merhy, J. Bonnety, P. da Costa, G. Legros
Sorbonne Universités, UPMC Univ Paris 06, CNRS, Unité Mixte de Recherche UMR 7190, Institut
Jean Le Rond d’Alembert, F-75005 Paris, France
In the present study, the sooting tendency of laminar methane diffusion flames seeded with vapors
of isooctane/toluene blends potentially mixed with pentanol isomers is assessed. Isooctane and
toluene are two major components of gasoline surrogate. Pentanol constitutes a potential biofuel
naturally produced by the microbial fermentations from amino acid substrates. It presents chemical
(e.g. corrosive properties, octane number), physical (e.g. non-hygroscopic, vapor pressure), and
thermodynamical properties (e.g. lower heating value, vaporization enthalpy) compatible with a use
as additive to both gasoline and diesel engines. Since 2010, many work have been conducted on the
combustion processes of 1-pentanol, 2- and 3-methylbutan-1-ol because those isomers stand as the
principal components of pentanol mixture produced at industrial scale. However, only few
fundamental works exist on the other isomers. Moreover very few works have been carried out on
the pyrolysis and the sooting tendency of pentanol. Within this context, the Yield Sooting Indices
(YSIs) are measured on a Santoro burner by Laser Absorption method. The observed tendencies are
compared and discussed.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Implementation of the detailed Naples soot model in the context of a
quadrature-based method of moment
S. Salenbauch1, M. Sirignano2, D.L. Marchisio3, M. Pollack1, A. D’Anna2, C. Hasse1
1
TU Bergakademie Freiberg, Numerical Thermo-Fluid Dynamics, 09599 Freiberg, Germany
Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale, Università degli
Studi di Napoli Federico II, Piazzale Tecchio 80, 80125 Napoli, Italy
3
Department of Applied Science and Technology, Institute of Chemical Engineering, Politecnico di
Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy
2
steffen.salenbauch@iec.tu-freiberg.de
Understanding soot particle nucleation is one of the major scientific challenges in the field of
soot research [1-3]. As recent experimental studies have offered a detailed insight into the relevant
steps leading to soot formation [1-4], soot models have been further developed recently offering the
possibility to describe the pathways of the transition of large, polycyclic aromatic hydrocarbons
(PAHs) to solid particles in detail.
One of these detailed models has been formulated by D'Anna and coworkers at the University of
Naples [5]. It describes all important physical and chemical processes including soot nucleation by
dimerization of polycyclic aromatic hydrocarbons and molecular growth through acetylene and
aromatic addition. Furthermore, it accounts for different particle structures such as large molecules,
soot clusters and agglomerates in both radical and stable form. Dehydrogenation is also considered
describing the hydrogen/carbon-ratio (H/C-ratio) of each entity in the model. Associating the H/Cratio to the degree of pericondensation, the influence of both pericondensed and incompletely
condensed species is also accounted for by directly integrating the H/C-ratio in the formulation of
relevant source terms influencing HACA [6] reaction rates and the sticking probability of colliding
entities [7].
Tracking all these properties as independent coordinates yields a multivariate number density
function (NDF) with 4 internal coordinates describing the properties size (number of carbon atoms),
H/C-ratio, structure (molecule, cluster or agglomerate) and state (stable or radical). Different
numerical approaches are available to treat the evolution of population balance statements based on
multivariate NDFs, e.g. Monte Carlo approaches, sectional methods or moment-based methods.
Especially moment methods are known to be well-suited and established for soot simulation in
laminar and turbulent reactive systems. Among moment methods, quadrature-based moment
methods (QBMM) [8] have become a feasible choice.
It is within the scope of this study to formulate the extensive, quadvariate Naples soot model in
the context of a QBMM approach. The applied algorithm to close the moment equations is
presented and the derivation of the QBMM-based source term expressions involving the
molecule/particle properties described above is explained. The new QBMM-based implementation
of the soot model is finally applied to simulate the formation and growth of soot particles in a
laminar, burner-stabilized premixed sooting flame which is listed as a target flame of the ISF
workshop.
[1] A. D’Anna, Proc. Combust. Inst. 32 (1) (2009) 593–613.
[2] H. Wang, Proc. Combust. Inst. 33 (1) (2011) 41–67.
[3] P. Desgroux, X. Mercier, K. A. Thomson, Proc. Combust. Inst. 34 (1) (2013) 1713–1738.
[4] H. Bockhorn, A. D’Anna, A. Sarofim, H. Wang, Eds., Combustion Generated Fine Carbonaceous ‘ Particles. KIT
Scientific Publishing, 2009.
[5] A. D’Anna, M. Sirignano, J. Kent, Combust. Flame 157 (11) (2010) 2106 – 2115.
[6] M. Frenklach, H. Wang, Proc. Combust. Inst. 23 (1) (1991) 1559 – 1566.
[7] M. Sirignano, J. Kent, A. D’Anna, Combust. Flame 157 (6) (2010) 1211 – 1219.
[8] D. L. Marchisio, R. O. Fox, Computational Models for Polydisperse Particulate and Multiphase Systems,
Cambridge University Press, 2013.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
TRASIENT FLAMELET MODELING OF LAMINAR NON-PREMIXED
SMOKING AND NON-SMOKING ETHYLENE/AIR FLAMES
T.H. Kim1, N.S. Kim1, Y. Kim1*
1
Hanyang University, Korea
ymkim@hanyang.ac.kr
Recently, Kim et al. [1] has developed an Interactive Transient Flamelet (ITF) model combined with the
semi-empirical soot model. To realistically predict the laminar sooting flames with relatively long residence
time, the ITF model has developed a numerical procedure for strongly coupling the soot formation processes,
the heat release rate arising from chemical reaction, and the radiative heat transfer associated with gaseous
species and solid soot particles.
The present study has been mainly motivated to realistically predict the characteristics of the laminar nonpremixed smoking and non-smoking flames with the consistent formulation based on the ITF-MOMIC
approach and the detailed soot model. According to the previous literatures, only a few research groups [2, 3]
successfully simulated the smoking and non-smoking behavior with the consistent formulation. Liu et al. [2]
adopted the statistical particle tracking approach and two-equation model with GRI 3.0 mechanism which is
relatively simple model for soot formation. They proposed two temperature-dependent correction factors for
the O2 and OH oxidation, which allow flexibly capturing the difference of non-smoking and smoking flames.
On the other hand, Khosousi and Dworkin [3] employed the standard statistical particle tracking approach,
the detailed PAH chemistry, and a soot sectional model. In their approach, each section considers primary
particles size and aggregate structure. They suggested the new active surface sites available parameter for O2
oxidation as a function of thermal age including temperature and residence time. This new active parameter
is also able to distinguish non-smoking and smoking flames in terms of soot volume fraction and primary
particle size. Additionally, they [2, 3] both solved the all governing equations by full transport approach with
parallel processing and used the Discrete Ordinate Method together with a statistical narrow-band correlatedk-based model [4] for radiation.
In the present study, to realistically represent soot formation and radiation in the laminar non-premixed
flame with smoking and non-smoking behavior, the ITF-MOMIC approach with the consistent formulation
has been developed. In context with the ITF-MOMIC model, we have developed the strong coupling strategy
between physical space and mixture fraction space to deal with the slow processes including soot formation
and radiation and the non-unitary Lewis flamelet formulation is utilized to account for differential diffusion.
Moreover, the oxidation model of the zeroth moment in context with MOMIC is adopted to avoid unphysical
fractional reduced moments in the simulation of the non-premixed sooting flames and the active surface sites
available parameter for O2 and OH oxidation as well as C2H2 surface growth is evaluated as a function of
temperature and particle size [5]. Without tuning the model constants in the detailed soot model, the ITFMOMIC approach is investigated for consistently predicting the laminar non-premixed smoking and nonsmoking flames. To the best of author’s knowledge, it is the first effort to predict the smoking and nonsmoking behavior in laminar non-premixed flames by utilizing MOMIC based [5] model with the newly
proposed detailed PAH chemistry [6] and the transient interactive flamelet model [1].
Based on numerical results obtained in this study, the detailed discussion has been made for the capability
and limitations of the ITF-MOMIC approach to predict the precise flame structure and soot formation
characteristics in laminar non-premixed smoking and non-smoking ethylene/air jet flames with the consistent
model constants.
References
[1] T.H. Kim, Y.M. Kim, Combust. Flame 162 (2015) 1660-1678
[2] F. Liu, H. Guo, G.J. Smallwood, Ö.L Gülder, Combust. Theory. Model. 7 (2003) 301-315.
[3] A. Khosousi, S.B. Dworkin, Proc. Combust. Inst. 35 (2015) 1903-1910.
[4] F. Liu, G.J. Smallwood, Ö.L Gülder, Int. J. Heat Mass Transfer 43 (2000) 3119-3135.
[5] J. Appel, H. Bockhorn, M. Frenklach, Combust. Flame 121 (2000) 122–136.
[6] V. Chernov, M.J. Thomson, S.B. Dworkin, N.A. Slavinskaya, U. Riedel, Combust. Flame 161 (2014) 592–601.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Modeling for Turbulent C2H4/Air Non-premixed Sooting Flames
T.H. Kim1, N.S. Kim1, S.T. Jeon1, Y. Kim1*
1
Hanyang University, Korea
ymkim@hanyang.ac.kr
The present ITF-MOMIC model has the basic advantages for extending to the simulation of turbulent nonpremixed flames. In this study, turbulence is represented by k-　model and the scalar dissipation rate should
be modeled. This methodology called the Transient Flamelet Model was widely used for turbulent nonpremixed jet flames which is capable of robustly treat the turbulence-chemistry interaction and the transport
effect for relatively slow process such as NOx and soot emission. To validate the capability of the present
ITF-MOMIC model to predict the structure and combustion process in turbulent non-premixed sooting
flames, three different atmospheric ethylene/air turbulent non-premixed flames were chosen as the
benchmark problems.These different flames have the quite simple inflow conditions and the fuel mixture is
injected from each un-cooled specific diameters vertical tube. The air is entrained from the annular ambient
region.For numerical convenience, the small axial velocity (0.5 m/s) is uniformly imposed. Computational
domain is extended to 1.0 m in axial direction and 0.5 m in radial direction. Computations are based on 201
(z)× 101(r) non-uniform mesh arrangement in axisymmetric computational domain. There are no
adjustments for k-　model constant. But a round jet correction as proposed by Pope[1] is employed. In the
approach, the oxidation model of the zeroth moment in context with MOMIC is adopted to avoid unphysical
fractional reduced moments in the simulation of the non-premixed sooting flames and the active surface sites
available parameter for O2 and OH oxidation as well as C2H2 surface growth is evaluated as a function of
temperature and particle size [2]. Without tuning the model constants in the detailed soot model, the ITFMOMIC approach is investigated for consistently analysing three turbulent non-premixed sooting flames.
This study adopts the newly proposed detailed PAH chemistry [6]. Schmidt numbers for scalars (such as
mean and variance mixture fraction) are used 0.7 for CASE 1 and 0.6 for CASE 2, 3. The each fuel nozzle
diameters and jet velocity conditions are listed in Table 1. The axial velocity profiles at fuel inlet is
prescribed using the 1/7 power law. In CASE 1 [4], emphasis is focused on the effects of soot radiation. For
CASE 2 [5] and CASE 3 [6], we investigate the turbulence-chemistry interaction and the transport effects in
the soot formation and oxidation.
Table 1 Boundary condition for each turbulent flame case. (Unit are m, s, K)
Case
Fuel (fully developed velocity profiles)
Ubulk
T
YC2H4
Dnozzle
Re
1 [Hu et al. 2003]
25.8
400
1.0
4.6E-3
13,500
2 [Young et al. 1991]
24.5
400
1.0
3.1E-3
8,600
3 [Kent et al. 1987]
52.0
400
1.0
3.0E-3
14,660
Co-flow air
(Uniform)
u
T
0.5
300
References
[1]Pope SB,(1978),AIAA J. 16(3)279–281
[2] J. Appel, H. Bockhorn, M. Frenklach, Combust. Flame 121 (2000) 122–136.
[3] V. Chernov, M.J. Thomson, S.B. Dworkin, N.A. Slavinskaya, U. Riedel, Combust. Flame 161 (2014) 592–601.
[4]Hu B, Yang B, Koylu UO,(2003), Combustion and Flame 134 93-106
[5]Young KJ, Stewart CD, Syed KJ, Moss JB, Proceeding on Tenth ISABE Meeting 1991
[6]Kent JH, Honnery D, (1987), Combustion Science and Technology 54 383-397
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Effects of laser pulse shape and duration on particle size determination from
Time Resolved LII at elevated pressures
S.A. Steinmetz, E. Cenker, W.L. Roberts
Clean Combustion Research Center, King Abdullah University of Science and Technology, Saudi
Arabia
scott.steinmetz@kaust.edu.sa
Time Resolved Laser Induced Incandescence (TiRe LII) is commonly used to measure soot volume fraction
and effective soot particle diameters. However, the efficacy of TiRe LII is known to depend on the temporal
and spatial characteristics of the laser pulse used [1,2]. This sensitivity can be exacerbated at elevated
pressures, where heat transfer timescale are greatly shortened. The result is a non-uniform distribution of
temperature, during and after the laser pulse, in poly-disperse soot samples [3,4]. In this work, the effect of
laser pulse temporal shape and duration on calculated particle size is investigated, with focus on elevated
pressure applications. Soot temperatures are measured in nitrogen-diluted ethylene flames at 1 and 15 bar
using two-color pyrometry. Tophat temporal profiles of 1.6 and 6 ns (FWHM) are produced with a custom
Nd:YAG laser, while typical Gaussian profiles of 4, 6, and 10 ns are produced with a conventional Nd:YAG
laser. Measurements are performed over a range of laser fluences from 0.05 to 0.3 J/cm2. Particle sizes are
calculated with an LII model via best-fit comparison of the temporal signal decay. The model accounts for
particle heat exchange via absorption, sublimation, conduction, thermionic emission, and radiation.
Simulations of LII signals for various particle size distributions are used to investigate potential errors related
to pulse shape.
Reference(s)
[1] M. Dansson et al., Applied Optics, 46, 8095 (2007).
[2] H. Michelsen, Applied Physics B, 83, 443 (2006).
[2] M Charwath et al., Applied Physics B, 104, 427 (2011).
[3] E. Cenker et al., Applied Physics B, 119, 745 (2015).
International Sooting Flames Workshop
30-31 July 2016, Nest Hotel Incheon
An optical technique for the detection and characterisation of combustion
formed particles in laminar and turbulent flames.
Daniel Bartos1,* Matthew Dunn1, Mariano Sirignano2, Andrea D’Anna2, Assaad R. Masri1
School of Aerospace, Mechanical and Mechatronic Engineering, the University of Sydney,
Sydney Australia.
1
2
Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale, Universita`degli
Studi di Napoli Federico II, Napoli, Italy.
*dbar4475@uni.sydney,edu.au
Abstract
In this work a novel optical technique on time-resolved laser induced fluorescence (LIF
and incandescence (LII) to simultaneously track soot particles and nanostructures in flames is
presented. An 80 picosecond pulsed laser @266nm is used to observe scattering and LIF,
while LII is observed under excitation from an 8ns pulsed laser @1064nm. For signal
collection, four fast photomultiplier tubes (PMTs) are placed at the spectral plane of a custom
built spectrometer. Pointwise, time resolved measurements are collected at four wavelength
bands to simultaneously to track Scattering, LIF and LII. The PMTs are placed within the
spectrum to capture scattering @266nm, UV LIF @350nm, Visible LIF @440nm and LII
@575nm. Data collected with 266 laser source are also compared with LIF measurements
taken with 8ns pulse and ICCD with good agreement. Data were firstly collected along the
axis of two rich premixed laminar flames as example of non-sooting (C/O=0.67) and sooting
(C/O=0.77) conditions. UV LIF and Visible LIF are attributed to nanoparticles or condensed
phase species made up largely of PAHs. In fact florescence decay times observed using this
technique are consistently above 2ns and thus cannot be attributed to gas phase fluorescence.
Scattering is very sensitive to particle size and is used track the largest particles in the flame
and LII is correlated with larger soot structures. This technique achieves a relatively high
signal to noise ratio on a single shot. On these premises, measurements were taken in a
turbulent conditions where the time-resolved measurements were found to be extremely
useful to provide insights on the evolution of combustion-generated particles in turbulent
sooting flames.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Experiments on Sooting Turbulent Non-Premixed Flames at Elevated Pressures
W.R. Boyette1, S. Chowdhury1, E. Cenker1, W.L. Roberts1
1
Clean Combustion Research Center, KAUST
wesley.boyette@kaust.edu.sa
Validation of various numerical models of turbulent non-premixed flames has been very
challenging due to the lack of experimental data. At the Clean Combustion Research Center at
KAUST, we have designed and built a replica burner based on the parameters of the ISF-2 turbulent
target flame 2 (Zhang et al 2011). Although the geometry is the same, the flame conditions are
slightly different. The burner includes a central fuel tube delivering ethylene with 65% nitrogen as
diluent to limit the soot loading. An annular fuel tube delivers premixed ethylene and air for the
pilot flame. The pilot mixture accounts for 6 % of the total heat release of the main jet. A
conditioned air co-flow jet is formed by a concentric tube with an outer diameter of 250 mm. At
atmospheric pressure, we have conducted soot particle measurements at different axial locations
along the centerline using a Scanning Mobility Particle Sizer (SMPS). The measurements were
made for two different Reynolds number of the jet: Re=10,000 and 20,000. A soot-sampling probe
is built with a tube having a 500-micron orifice diameter to pull in a small amount of gas sample
from the flame. A two-stage dilution system is used to “freeze” soot reactions along the sampling
line, allowing for accurate measurements of the soot particle size distribution function. The SMPS
measures the concentration of soot particles in the sample gas over a range of particle diameters
from 3 to 228 nm, using two different differential mobility analyzers. Measurements were taken at
multiple heights in each flame to provide an idea of how the soot evolves along the flame centerline.
A new high-pressure combustion duct (HPCD) has begun operation at KAUST that will allow for
experiments on turbulent non-premixed flames at pressures up to 40 atm. The duct is large enough
to accommodate a 250-mm burner with no flame impingement on the walls, making the flames
ideal for model validation. Six windows provide optical access to the flame. Early results, including
flame height measurements of the nitrogen-diluted flames at elevated pressures, will be presented.
Experiments planned for the near future include PIV/OH-PLIF for flow field and reaction zone
characterization and LII for quantification of soot volume fraction distribution. It is believed that
these flames offer the benefits of well-characterized boundary conditions and low soot loading,
making them excellent candidates for ISF pressurized target flames in the future.
Reference(s)
[1] J. Zhang et al., Rev. Sci. Instrum., 82, 074101 (2011)
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
The Yale Coflow Burner as a Tool for Studying Soot Formation
D.D. Das1, D. Giassi2, N.J. Kempema2, M.B. Long2, C.S. McEnally1, L.D. Pfefferle1
1
Department of Chemical Engineering, Yale University, New Haven CT USA
2
Department of Mechanical Engineering, Yale University, New Haven CT USA
dhrubajyoti.das@yale.edu
The Yale Coflow Burner is a burner with coannular air and fuel flow that is used for generating nonpremixed
laminar flames with a simple axisymmetric geometry. It was originally developed in the Long research
group for generating experimental data that could be compared with detailed flame simulations in a twodimensional configuration from the Smooke group [1, 2]. A key feature compared to other burners was that
the fuel tube was only 4 mm in diameter, which caused the flames to be lifted and the thermal boundary
conditions at the burner surface to be well-defined. It was subsequently extended to studies of soot
formation in collaboration with the Pfefferle group [3, 4]. In these studies a series of ethylene/air
nonpremixed flames were defined, where the air and fuel velocities are 35 cm/sec and the fuel is a mixture of
32 %, 40 %, 60 %, and 80 % ethylene in nitrogen. In the years since an extensive database of measurements
and computations has accumulated for these flames. They are included in the International Sooting Flame
Workshop (ISF) target cases as ISF Co-flow 3.
At the second Workshop it was suggested that this burner be widely adopted by the community to simplify
comparing measurements and computations from different research groups. In response we developed CAD
drawings and a detailed parts list, and made them available online [5]. We also initiation a campaign – now
closed – to purchase the parts in bulk and make kits available to ISF participants at a reduced price.
Ultimately burners were distributed to 12 research groups on five continents.
Since the second Workshop our experimental work has focussed on further characterizing the standard
ethylene flames and on using fuel-doped methane flames to study soot formation from liquid transportation
fuels and their components. In the ethylene flames we have made TEM and 2-D multi-angle light scattering
measurements of soot aggregate properties [6]. We have also developed a method to directly measure the
thermal boundary conditions in these flames [7]. In the methane flames we have measured 2-D soot volume
fractions in a series of flames doped with reference jet fuels and diesel fuels, surrogate jet fuels and diesel
fuels, and individual components of the surrogates [8].
References
[1] M.D. Smooke, P. Lin, J.K. Lam, M.B. Long, Proc. Combust. Inst. 23, 575-582 (1990).
[2] M.D. Smooke, Y. Xu, R.M. Zurn, P. Lin, J.H. Frank, M.B. Long, Proc. Combust. Inst. 24, 813-821 (1992).
[3] C.S. McEnally, A.M. Schaffer, M.B. Long, L.D. Pfefferle, M.D. Smooke, M.B. Colket, R.J. Hall, Combust. Inst. 27,
1497-1505 (1998).
[4] M.D. Smooke, R.J. Hall, M.B. Colket, J. Fielding, M.B. Long, C.S. McEnally, L.D. Pfefferle, Combust. Theory
Modelling 8, 593-606 (2004).
[5] http://guilford.eng.yale.edu/yalecoflowflames/steady_burner.html
[6] N.J Kempema, M.B. Long, Combust. Flame 164, 373-385 (2016).
[7] N.J. Kempema, M.B. Long, unpublished results.
[8] D.D. Das, W.J. Cannella, C.S. McEnally, C.J. Mueller, L.D. Pfefferle, Proc Combust. Inst. (in press).
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Measurements in Turbulent Ethylene/Propylene Flames
K.N.G. Hoffmeister, T.W. Grasser, J.C. Hewson, S.P. Kearney
Sandia National Laboratories*, Albuquerque, NM, USA
kngabet@sandia.gov
We present a detailed set of measurements in the ISF-Target Flame 2 – Sandia Flame [1,2], a piloted,
turbulent ethylene/air flame with Re=20,000. Hybrid femtosecond/picosecond coherent anti-Stokes Raman
scattering (CARS) is used to monitor temperature and oxygen, while laser-induced incandescence (LII) is
applied to image the soot volume fraction. Soot measurements were also taken during a parametric study
investigating the effects fuel composition on soot formation in flames analogous to the target flame with fuel
compositions of 75%/25% and 50%/50% ethylene/propylene by volume. Soot volume fraction
measurements at varying downstream positions in all three flames are shown in Fig. 1. As expected, soot
volume fractions increase with increased propylene. Because our measurements were taken at an altitude of
~1,630 m above sea level, we compare LII measurements in the pure ethylene target flame to those taken by
Shaddix et al. at Sandia Livermore (~150 m above sea level) [3].
*Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed-Martin Company, for the
United States Department of Energy’s National Nuclear Security Administration under Contract DE-AC0494AL85000.
Fig. 1 Soot volume fraction versus downstream position
Reference(s)
[1] J. Zhang, C.R. Shaddix, and R.W. Schefer, Rev. Sci. Instr., 82, 074101, (2011)
[2] Turbulent Flames, International Sooting Flame (ISF) Workshop, http://www.adelaide.edu.au/cet/isfworkshop/datasets/turbulent/ (2016)
[3] Shaddix et al., SANDIA Report, SAND2010-7178, (2010)
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Measurements of soot concentrations, soot temperatures, and thermal radiation
from a turbulent, non-premixed, pre-vaporized aviation fuel jet flame
Christopher R. Shaddix, Timothy C. Williams, and Jiayao Zhang
Combustion Research Facility
Sandia National Labs
Livermore, CA 94550 USA
crshadd@sandia.gov
We have performed measurements in a piloted, non-premixed jet flame burning pre-vaporized aviation fuel
surrogate burning in lightly coflowing air with a fuel tube exit Reynolds number of 20,000. The burner
design and operation mirror that which currently forms ISF-3 Turbulent Flame Target 2 for burning ethylene
in air. Thus, the aviation fuel flame forms a natural target for interrogation of predictions of soot formation in
turbulent flames when using higher hydrocarbon fuels. The aviation fuel surrogate is composed of 77 liquid
vol-% n-dodecane and 23 liquid vol-% m-xylene, and was chosen to match the typical soot formation
tendency of Jet A as well as most of its other functional fuel qualities.
Planar, IR-excited laser-induced incandescence (LII) was used to measure soot concentrations across the
flame, and a 3-line measurement technique was used to measure soot temperatures across 10 mm long
pathlengths in the flame. Spatially resolved thermal emission was measured using a thermopile with an
attached small-diameter anodized tube to limit the spatial sampling. These same measurements were also
applied to the corresponding ethylene flame.
Fig. 1 shows the measured instantaneous, mean, and rms soot concentrations in the flame and Fig. 2 shows
the measured thermal radiation emitted along the flame axis, in comparison to that emitted from the
ethylene-fueled jet flame. The soot persists to greater heights in the aviation fuel flame, particularly when
plotted as a function of the dimensionless height (the fuel tube ID is 2.5 mm for the aviation fuel flame and
3.2 mm for the ethylene flame). We believe this aviation fuel jet flame could make a suitable simulation
target for future ISF workshops.
Fig. 1. Instantaneous (left), mean (middle) and
RMS fv fields measured in the aviation fuel flame.
Fig. 2. Mean thermal radiation measured from a
narrow-angle pyrometer along the jet axis for the
aviation fuel and ethylene jet flames.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Large Eddy Simulations of turbulent jet flames burning ethylene and a
kerosene surrogate fuel
Nicholas Burali and Guillaume Blanquart
Department of Mechanical Engineering, California Institute of Technology, Pasadena (USA)
In the present work, results from Large Eddy Simulations (LES) of turbulent jet flames burning
ethylene and a kerosene surrogate fuel are shown. The spatially-filtered continuity, momentum, and
scalar transport equations are solved using a low-Mach formulation. The subfilter quantities are
modeled using the dynamic Smagorinsky model with Lagrangian averaging. Local thermo-chemical
quantities are tabulated using a flamelet/progress variable approach. Soot particles are described
using a bivariate model, based on particle volume and surface area. The statistical evolution of soot
particles is described by solving transport equations for moments of the soot Number Density
Function (NDF), using the Direct Quadrature Method of Moments. In the present work a bimodal
distribution for the NDF is used. The LES results are compared against the experimental data of
Shaddix and Zhang [8th U.S. National Combustion Meeting, 2013]. Detailed comparisons include
soot volume fraction and soot temperature statistics.
The CaltechMech [G. Blanquart, CaltechMech, version 2.3, http://www.theforce.caltech.edu]
chemical model is used to generate the chemistry tables used in both simulations. This model has
been shown to lead to good agreement with experiments in laminar flames when burning both fuels
considered in this work. However, the present results reveal that, while soot is in fairly good
agreement in the JP-8 LES, it is significantly underpredicted in the ethylene jet. This suggests that
the chemical pathways leading to soot formation may need to be revised.
The comparison between the two simulations is complemented by a discussion of the inlet
conditions in LES of piloted jet flames. More specifically, an immersed boundary simulation of the
base of the jet provides insight into the effect of the use of perforated plates in pilot streams on the
inflow conditions.
Finally, a novel treatment of radiation effects based on the correction of tabulated quantities in
adiabatic chemistry tables is also discussed.
International Sooting Flames Workshop
30 -31 July 2016 Nest Hotel, Incheon
A multi-sectional / MMC-LES approach for the Adelaide/Delft flame
A. Majbouri1, M.J. Cleary1, J.H. Kent1, M. Sirignano2, A. D’Anna2
1
School of Aerospace, Mechanical and Mechatronic Engineering, University of Sydney, Australia
2
Dipartimento di Ingegneria Chimica, Università “Federico II”, Naples, Italy
m.cleary@sydney.edu.au
Introduction
The multi-sectional approach to model soot evolution in combustion makes use of detailed
kinetic pathways for soot nucleation, growth, coagulation and oxidation [1]. This approach is based
on sections (or bins) representing particulate species with similar properties. (e.g. molecular weight,
H/C ratio, morphology). The multi-sectional method is fully coupled with, and is a natural
extension of, the reacting gas phase. Soot sections and their kinetic expressions are treated similarly
to gas phase species. Consequently, implementation of the multi-sectional model into existing
combustion codes is straightforward. To date, reasonably accurate predictions of soot have been
obtained for laminar diffusion and premixed flames. The present research applies the multisectional method to turbulent sooting diffusion flames.
Method
Gas phase turbulent mixing and reaction is being modelled the multiple mapping conditioning
turbulent combustion model coupled to a large eddy simulation. The method is known as MMCLES and it has been implemented in a new OpenFOAM compatible code – mmcFoam – that is a
collaborative endeavour involving the universities of Sydney, Queensland, Stuttgart and New South
Wales.
We use a reduced sectional scheme containing 40 sections (20 Carbon number bins by two H/C
ratio bins) and 4270 steps. Although reduced, the kinetics scheme is large by the standards of
conventional gas-phase turbulent combustion modelling. Therefore, there is a significant
computational benefit provided by the sparse Lagrangian emulation of the filtered density function
(FDF), made possible by the MMC closure, involving significantly fewer stochastic particles than
there are LES grid cells.
Preliminary results
We report work-in-progress in simulating the Adelaide/Delft piloted methane jet flame.
Experimental data for temperature, major gas phase species, mixture fraction and soot volume
fraction are available [2]. Close to the burner exit the model predicts gas phase temperature, major
species and velocity quite well although the rate of decay of mixture fraction along the centreline is
underpredicted. The soot formation zone is located well downstream in this flame and it is strongly
influenced by the characteristics of the mixing zone close to the burner exit. The model shows good
agreement with experimental data for the axial distance corresponding to maximum soot volume
fraction and subsequent burn out, although the mean soot volume fraction is over predicted. Initial
tests suggest that the model can reproduce the experimental soot intermittency quite well.
Acknowledgements
This work is financially supported by the Australian Research Council grant DP160105023.
Reference(s)
[1] M. Sirignano, J.H. Kent, A. D’Anna, “Modeling formation and oxidation of soot in nonpremixed flames”, Energy
Fuel 27, 2303-2315, 2013.
[2] N. H. Qamar, Z. T. Alwahabi, Q. N. Chan, G. J. Nathan, D. Roekaerts, K. D. King, “Soot volume fraction in a
piloted turbulent jet non-premixed flame of natural gas”, Combust. Flame 156, 1339–1347, 2009.
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Large Eddy Simulation of a Turbulent Sooting Ethylene/Air
Diffusion Flame Using a Detailed Sectional Soot Model
P. Rodrigues, B. Franzelli, R. Vicquelin, O. Gicquel, N. Darabiha
Laboratoire EM2C, CNRS, CentraleSupélec, Université Paris-Saclay, Grande Voie des Vignes,
92295 Châtenay-Malabry Cedex, France
Corresponding author: Pedro Rodrigues (pedro.rodrigues@centralesupelec.fr)
I.
Introduction
Soot modeling is today recognized as an extremely challenging problem in turbulent combustion. Indeed,
very strong interactions between turbulence, chemistry and particles formation and dynamics coexist in this
kind of flames. Soot particles also play an important role in radiative energy transfer. The final objective of
this study is to investigate these coupled phenomena using a LES soot model based on a detailed sectional
method.
II.
Numerical setup
The Sandia Flame of the ISF workshop is studied in this work following the experimental setup presented in
[1]. The mesh is presented in Fig. 1 and details on numerical and modeling setup are given in Table 1.
Following the FPV approach [2] and its non-adiabatic extension [3], adiabatic laminar strained diffusion
counterflow flames are first computed using the REGATH package and with the KM2 chemical scheme [4],
in order to obtain the so-called S-curve as a function of the strain rate. Heat losses are then taken into account
in the flamelet manifold by i) adding an optically-thin radiative source term to the
energy equation of the corresponding flames and ii) computing unsteady flames with the
initial adiabatic flame profile. This flame database is then used in order to generate a 4D
table parameterized by (Z,Sz,C,H). For the subgrid-scale combustion model, a β-PDF is
used for the Z (mixture fraction) direction and a δ-dirac PDF for the C (progress variable)
and H (enthalpy) directions.
Number of cells/nodes
Solver
LES model
Numerical Scheme
Combustion Model
Discretization of the table
10M/1.7M
AVBP(unstructuredcompressiblesolver)
WALE
rd
TTGC(3 orderinspace/time)
RFPV(β-PDFonZ,δ-diraconCandH)
(Z,Sz,C,H):100x20x100x20
Fig 1. Computational grid
Table 1. Numerical and modeling setup characteristics
A soot sectional model (presented in [5]) is used for the solid phase in order to describe the evolution of soot
particles. PAHs unsteady response behavior treatment and soot subgrid model are adapted to the soot
sectional model according to the corresponding models developed by Mueller and Pitsch [6].
0.15
0.1
1000
0.05
500
-60
-40
-20
0
20
40
0
Radial Position [mm]
0.25
Fig 2. Mean
temperature field
0.2
400
0.15
300
0.1
200
100
-100
XO2/XN2 ratio RMS
500
[1] J. Zhang et al., Review of Scientific Instruments, 82, 074101 (2011) 1-10.
[2] C.D. Pierce and P. Moin, J. Fluid Mech., 504, (2004) 73-97.
[3] M. Ihme and H. Pitsch, Phys. Fluids., 20, 5 (2008) 1-20.
[4] Y. Wang et al., Comb. & Flame, 160, 9 (2013), 1667-1676.
[5] P. Rodrigues et al., Proc. Combust. Inst., under revision.
[6] M. Mueller and H. Pitsch, Comb. & Flame, 159, 6 (2012) 2166-2180.
[7] S.P. Kearney et al., Sandia Report, SAND2015-7968, October 2015.
Acknowledgements
0.2
1500
Mean XO2/XN2 ratio
Mean Temperature [K]
2000
Temperature RMS [K]
III.
Preliminary results
Only results for the non-sooting adiabatic case are
presented in this abstract. Figure 2 presents the
corresponding mean temperature field. Comparison
of mean and rms temperature, XO2/XN2 ratio radial
profiles are presented for x/D=134 and compared
with experimental results presented in [7]. Nonadiabatic sooting results will be presented at the third
ISF workshop and are expected to provide a better
agreement in mean temperature profiles.
References
0.05
-50
0
50
100
Radial Position [mm]
Fig 3. Comparison between numerical
results (lines) and experiments (symbols) of
mean and rms radial profiles of temperature
and XO2/XN2 ratio
This study is supported by the Air Liquide, CentraleSupelec and CNRS Chair on oxycombustion and heat transfer for energy and environment
and by the OXYTEC project, grant ANR-12-CHIN-0001 of the French Agence Nationale de la Recherche.
This work was performed using HPC resources from GENCI-CINES (Grant 2016-020164).
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Transported PDF Modelling of Soot in the Sandia Jet Flame
M.A. Schiener1, R.P. Lindstedt1
1
Department Mechanical Engineering, Imperial College London, United Kingdom
marcus.schiener09@imperial.ac.uk
The present work focuses on the prediction of soot in the Sandia turbulent ethylene diffusion flame at a
Reynolds number of 20,000 (ISF-3 Target Flame 2). A transported probability density function (PDF)
approach closed at joint-scalar level, allowing the inclusion of turbulence-chemistry interactions without
modelling, in conjunction with a parabolic solution of the velocity field is coupled to a method of moments
approach with interpolative closure [1] for modelling of soot coagulation and agglomeration. Closure of the
turbulence field is obtained via a second moment model and radiation (including the impact of turbulent
fluctuations) from soot and gas phase species is modelled assuming an optically thin medium with enthalpy
included as a solved scalar. The gas phase chemistry is represented by a systematically reduced C1-C2
mechanism featuring 144 reactions, 15 solved and 14 steady-state species. The sensitivity to soot oxidation
via reactions with O, OH and O2 is taken into account and the approximate inclusion of soot surface
reactions via a second ring PAH analogy [2] is investigated. In absence of the latter, the axial peak location
of soot and the radial distribution are matching experimental data reasonably well. However, maximum
values are over-predicted by a factor of about two and there is a tendency to under-predict soot near the
burner and far downstream. Overall agreement with experimental data is significantly improved when taking
into account the impact of the local gas phase composition by using a soot surface based reaction model for
mass growth. A parametric study regarding the model parameter αs, expressing the fraction of available
reaction sites on the soot surface, shows that absolute soot levels are arguably well-matched for a value in the
range 0.75 to 1.0 with the lower value consistent with previous results for diffusion flames [2].
0.06
1
α S = 1.00
α S = 0.75
Exp.
x/d = 39
x/d = 133
α s = 1.00
α s = 0.75
Exp.
0.04
0.8
0.02
0.4
0.4
0.75
x/d = 70
0.8
x/d = 164
0.3
0.6
0.2
0.4
0.1
0.2
f!v [ppm]
f!v [ppm]
0.5
1.2
0.8
x/d = 102
0
0.4
x/d = 195
0.6
0.3
0.4
0.2
0.2
0.1
0.25
0
0
75
150
Axial coordinate x/d [–]
225
0
0
5
10
15 0
5
10
15
0
Radial coordinate y/d [–]
Fig. 1 Axial (left side) and radial (right side) profiles of mean soot volume fraction from predictions
modelling soot growth via a PAH analogy soot surface chemistry model for αs = 0.75 and 1.00.
References
[1] M. Frenklach, Chem. Eng. Sci., 57, 2229-2239 (2002)
[2] R.P Lindstedt, S.A. Louloudi, Proc. Combust. Inst., 30, 775-783 (2005)
International Sooting Flames Workshop
30-31 July 2016 Nest Hotel Incheon, Seoul
Large-Eddy Simulation of Soot Formation in a Model Aero Engine Combustor
Achim Wick1, Frederic Priesack1, Heinz Pitsch1
1
Institute for Combustion Technology, RWTH Aachen University, Germany
a.wick@itv.rwth-aachen.de
Large-Eddy Simulations (LES) of an aero engine model combustor experimentally investigated by Geigle et
al. [1] at the German Aerospace Center (DLR) are performed using flamelet-based tabulated chemistry and a
detailed soot model.
The experimental setup mimics relevant flow features and flame conditions of a real aero engine combustor.
Primary air is supplied through swirlers, generating a recirculation with a highly turbulent rich primary
combustion zone, and secondary air jets lead to a Rich-Burn/Quick-Quench/Lean-Burn (RQL)-type soot
oxidation region. This recently collected dataset [1] is well suited for model validation as it provides an
extensive amount of accurate data under engine-relevant conditions.
Both cold flow and reactive simulations are performed using
a fully unstructured LES code. Combustion is modeled with
a flamelet/progress variable model based on diffusion
flamelets, which are computed using the FlameMaster code
and the detailed chemical kinetic mechanism by
Narayanaswamy et al. [2]. Heat losses due to radiation are
accounted for using the model by Ihme et al. [3]. Detailed
modeling of soot evolution is done using the Hybrid Method
of Moments [4]. The computational domain includes the
swirler and air piping (more than what is visible in Fig. 3),
and the numerical mesh consists of 8.9 million nodes, with
local refinement in regions of strong shear and turbulence.
Fig. 1 Axial velocity in non-reacting flow
Two cases are simulated: the reference case defined in [1]
without (left) and with oxidation air
(p=3bar, φ=1.2 based on the primary air, oxidation air
(right).
accounting for 30% of the total air) and an equivalent case without oxidation air. Figure 1 shows the axial
velocity for both cases at two axial positions located close to the swirler and 15mm upstream of the secondary
air inlets. Both the swirling and recirculating flow in the primary zone and the fraction of the secondary air
that is transported upstream are well captured.
The reacting flow of the reference case is analyzed in terms of temperature and soot volume fraction. As seen
in Fig. 2, good agreement for the temperature is obtained. Figure 3 shows the instantaneous temperature
distribution in the symmetry plane along with isolines of soot volume fraction. As in the experiments, the soot
volume fraction peaks on the centerline close to the swirler. Soot pockets are transported into the wings close
to the walls, while the center of the combustor is soot-free. Although the soot pockets are oxidized slightly
faster than in the experiments, model results and measurements are qualitatively in very good agreement.
Reasonable to good quantitative agreement
is also found for both instantaneous and
time-averaged soot volume fractions.
Fig. 2 Mean centerline temperature
profile (ref. case with oxid. air).
Fig. 3 Instantaneous temperature and soot isolines at 1, 50,
150, 250 ppb (from light to dark blue) (ref. case with oxid. air).
References
[1] K.P. Geigle et al., Proc. Combust. Inst., 35, 3373-3380 (2015)
[2] K. Narayanaswamy et al., Combust. Flame, 162, 1193-1213 (2015)
[3] M. Ihme & H. Pitsch, Phys. Fluids, 20, 055110 (2008)
[4] M.E. Mueller & H. Pitsch, Combust. Flame, 159, 2166-2180 (2012)