Flow, Turbulence and Combustion (2025) 114:1315–1339

https://doi.org/10.1007/s10494-025-00653-8

RESEARCH

LES Modeling of the DLR Generic Single-Cup Spray

Combustor: Comparison of Exploratory Category C Jet Fuels

Arvid Åkerblom1 · Christer Fureby1

Received: 17 December 2024 / Accepted: 18 March 2025 / Published online: 3 April 2025
© The Author(s) 2025

Abstract
The combustion of conventional Jet A, alongside two alternative jet fuels, C1 and C5, is
simulated with Large Eddy Simulations (LES) in a generic single-cup spray combustor
during idle and cruise conditions. The spray is modeled using Lagrangian particle tracking
and the combustion chemistry of each fuel is modeled by skeletal reaction mechanisms.
The volatility and atomizability of each fuel directly affect the spray penetration depth,
with Jet A having the longest spray and C5 the shortest. All fuels have qualitatively similar
flames at idle conditions, but the Jet A flame is relatively lifted at cruise conditions. C1
and C5 have similar flames despite different spray lengths, likely due to the rapid breakup
of C1. The fuels produce different emission profiles, which is connected to their respec-
tive H/C ratios, equivalence ratios, and aromatics contents. NOx emissions are particularly
affected by the mixture fraction in the flame, resulting in high NOx emissions for the
compact C1 and C5 flames. Thermoacoustic oscillations are observed in all simulations
but are strongest for C1 and C5, which we hypothesize is a result of their high volatility.

Keywords Gas turbine combustor · Spray combustion · Large eddy simulation · Finite
rate chemistry · Thermoacoustics

1 Introduction and Background

Using Sustainable Aviation Fuels (SAF) is an effective measure for mitigating aviation car-
bon emissions. SAF is on the rise, with total production tripling to 600 million liters between
2022 and 2023, but this still only accounted for 0.2% of global consumption (ICAO, 2023b).
Economic challenges include limited production capacity and price competitiveness as well
as engine compatibility concerns resulting from the chemical differences between fossil

Arvid Åkerblom
arvid.akerblom@energy.lth.se
Christer Fureby
christer.fureby@energy.lth.se
1
Department of Energy Sciences, Division of Heat Transfer, Lund University, PO Box 118,
Lund SE 221-00, Sweden

13
1316 Flow, Turbulence and Combustion (2025) 114:1315–1339

fuels and SAF (Undavalli et al., 2023). Conventional jet fuel is defined by ASTM Standard
D1655 (ASTM International, 2024), but pure SAFs do not typically meet the specifications
of this standard and are hence used as drop-in fuels. As of July 2023, the International Civil
Aviation Organization (ICAO) lists 11 approved production pathways for SAF; examples
include Fischer-Tropsch (FT) fuels produced from gasified biomass, Hydroprocessed Esters
and Fatty Acids (HEFA) and Catalytic Hydrothermolysis (CHJ) fuels produced from oily
biomass, and Alcohol-To-Jet (ATJ) fuels produced from alcohols (ICAO, 2024). The SAFs
produced by these processes exhibit a wide range of properties and chemical compositions,
but they typically contain little to no aromatics unless these are deliberately added. A low
aromatics content reduces soot emissions, but at least 8% (by mass) is required to ensure
that elastomer seals in conventional aircraft fuel systems swell properly (Yang et al., 2019).
There is a need to thoroughly explore the combustion dynamics of SAF so that the rela-
tionship between chemical composition, engine performance, and emissions can be under-
stood. The research will gauge the impact of the variability of properties that SAFs exhibit
compared to fossil jet fuels, and provide knowledge to drive innovation in production and
certification. To facilitate this research, the US-based National Jet Fuels Combustion Pro-
gram (NJFCP) proposed a range of well-defined test fuels in three categories: A, B, and C
(Colket et al., 2017). Category A contains conventional petroleum-based jet fuels. Category
B contains fuels that were found to exhibit unacceptable combustion properties. Category C
fuels include SAFs, fossil-SAF blends, and synthetic fuels with borderline properties. Two
category C fuels, C1 and C5, are studied in the present work.
Large Eddy Simulations (LES) provide a powerful tool for studying the combustion
dynamics of jet fuels. With modern high-performance computing, LES can resolve partial
or full combustor geometries with enough detail to capture the large-scale turbulent motions
of the flame while chemically distinguishing between different fuels. To our knowledge,
all previous LES studies with category C fuels in full-scale combustors have targeted the
NJFCP Referee Rig of Stouffer et al. (2021), which is a swirl-stabilized spray combustor
designed to be sensitive to fuel effects. Esclapez et al. (2017) used LES to simulate the com-
bustion of Jet A, C1, and C5. They found that the short ignition delay time of C1 resulted in
a short flame lift, whereas the high volatility of C5 only had a marginal effect on the flame.
Panchal and Menon (2022a, b) simulated Jet A and C1. They found that C1 tended to burn
closer to the anchoring point, with a larger portion of premixed burning compared to Jet A.
C1 was predicted to undergo Lean Blowout (LBO) at a higher equivalence ratio than Jet A
(in accordance with experiments) but this trend was again determined to be mainly influ-
enced by chemistry rather than vaporization. Moreover, Jet A produced an acoustic response
with a single dominating frequency whereas C1 showed two excited frequencies and a more
irregular pressure oscillation. The authors hypothesized that the pressure oscillations were
primarily coupled to non-premixed burning along the spray, resulting in more prominent
oscillations for Jet A. Finally, Hasti et al. (2022) also used LES to simulate the combustion
of Jet A and C1 with adaptive mesh refinement and compact chemistry, capturing the same
LBO trend as Panchal and Menon (2022b).
Here we present LES of SAF combustion in a different model gas turbine combustor,
originally studied experimentally with fossil jet fuel by Behrendt et al. (1999) and Meier
et al. (2012). The same case has been studied numerically by teams at Imperial College
(Jones et al., 2014) and the University of Florence (Puggelli et al., 2016, 2018a, b; Mazzei
et al., 2019), but only with fossil jet fuel. Our initial results were included in Åkerblom et

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1317

al. (2022), which was followed by a thorough validation and sensitivity study (Åkerblom
and Fureby, 2024) with Jet A and JP-5. In our most recent study (Åkerblom et al., 2024) the
methodology was improved by adding a novel liquid thermodynamics model and tested on
four different fuels. The aim of the present work is to provide a broad and thorough com-
parison between the combustion of Jet A, C1, and C5. These three fuels represent a broad
spectrum of properties, particularly in aromatics content, combustion kinetics, and liquid
properties. Since there are no experimental data for alternative jet fuels, the simulations with
C1 and C5 presented here are entirely exploratory.
The study begins with an overview of the targeted fuels and their modeling. This is fol-
lowed by a description of the LES methodology, after which the results are presented. The
results include comparisons of temperature, velocity, heat release, mixture fraction, spray
statistics, chemical species distributions, thermoacoustic oscillations, as well as emissions
of CO2, H2O, CO, OH, and NOx. Trends among these parameters are identified and con-
nected to the base properties of each fuel. The simulations are carried out at stable operat-
ing conditions representing idling and cruising jet engines; LBO and re-ignition are not
considered.

2 Fuel Characterization and Modeling

This study targets three fuels: A2 (Jet A), C1, and C5. C1 is a commercial ATJ fuel and func-
tions as a SAF representative, whereas C5 is a test fuel with a high aromatics content and flat
boiling curve. The properties of these fuels are described in detail by Edwards (2017), and
surrogate compositions have been developed by Xu et al. (2015). A summary of the most
relevant properties for this work is given in Table 1. The composition spectra of the fuels,
i.e., the distribution of different classes of components across different carbon numbers, are
shown in Fig. 1(a) with measured compositions above and surrogate compositions below.
Jet A consists of a wide range of hydrocarbons and thus differs significantly from its sim-
plified surrogate. C1 primarily consists of two iso-paraffins; its surrogate, which neglects
minor fractions of other iso-paraffins, is a very accurate representative of the real fuel. C5
contains a small range of hydrocarbons, with the aromatic 1,3,5-trimethylbenzene account-
ing for 34.38% of all components. The fuels consequently have different molecular proper-
ties, e.g., Molecular Weight (MW), Hydrogen/Carbon ratio (H/C), and Lower Heating Value
(LHV), and their thermodynamic properties also differ. As shown in Fig. 1(b), Jet A has a

Table 1 Summary of ther- Jet A C1 C5

mochemical properties of the
n-paraffins [mol. %] 19.33 0 16.72
targeted jet fuels, obtained from
Edwards (2017). Surrogate com- iso-paraffins [mol. %] 26.09 100 48.90
positions and LHV are obtained cyclo-paraffins [mol. %] 31.16 0 0
from Xu et al. (2015) aromatics [mol. %] 23.42 0 34.38
MW [g/mol] 159 178 135
H/C (by mol) [-] 1.91 2.16 1.93
LHV [MJ/kg] 43.1 43.9 43.0
Viscosity @ 40°C [cSt] 1.31 1.53 0.83
Density @ 40°C [kg/m3] 785 742 749
Surface tension @ 40°C [dyn/cm] 23.6 21.0 22.2
CN [-] 48.3 17.1 39.6

13
1318 Flow, Turbulence and Combustion (2025) 114:1315–1339

Fig. 1 Measured (top) and surrogate (bottom) composition spectra (a) alongside distillation curves for
all fuels (b) Measured spectra are obtained from Edwards (2017) and surrogate spectra from Xu et al.
(2015). Note the trace amounts of iso- and n-paraffins in the measured spectra of C1 and C5. Legend (a):
n-paraffins (cyan), iso-paraffins (orange), cyclo-paraffins (gray), aromatics (black). Legend (b): Jet A
(red), C1 (green), C5 (magenta)

mostly linear distillation curve and is completely distilled at 540 K. C1 has a bimodal curve
which increases sharply at high temperatures towards a similar final boiling temperature of
540 K. C5 is unique in that all its components boil at approximately the same low tempera-
ture of 440 K. The low boiling point of C5, alongside its relatively low density and notably
low viscosity, makes for a volatile fuel that is readily atomized and vaporized.
C1 primarily decomposes into i-C4H8 (iso-butene) at the initial stage of the combustion
process whereas Jet A and C5 primarily form C2H4 (ethylene). The Cetane Number (CN)
of C1 is 17.1, which is relatively low for jet fuels. CN is generally correlated with ignition
quality, which suggests that C1 has poor ignition properties. The reality is more nuanced:
Wang et al. (2018b) demonstrated that C1 ignites faster than Jet A for initial temperatures
between 1000 and 1250 K due to the relatively rapid decomposition of iso-paraffins. At
higher temperatures, however, the initial fuel decomposition step is not rate-limiting and
the relatively slow oxidation of iso-butene (compared to ethylene) causes C1 to ignite more
slowly. Zheng et al. (2022) demonstrated a clear relationship between CN (specifically
Derived CN, or DCN) and resistance to LBO, and found that C1 was particularly sensitive
to LBO.
The chemical kinetics of jet fuel combustion may be modeled via a chemical reaction
mechanism, i.e., a model consisting of a specific set of species participating in a specific set
of reactions with defined rate coefficients. Some comprehensive mechanisms, e.g., Ranzi
et al. (2014), model the fuel as a mixture of specific fuel components, but this level of
flexibility and detail comes at a computational cost that is infeasible for most CFD applica-
tions. Here we use the HyChem family of reaction mechanisms (Xu et al., 2018; Wang et
al., 2018a, b) instead, which includes mechanisms for all targeted fuels. These mechanisms
are reduced to ~50 species and ~250 reactions, and the fuel mixture is approximated as a
single “average” species which decomposes into smaller hydrocarbons (e.g, ethylene) in a
short series of lumped decomposition steps. The majority of included species and reactions
are thus used to describe the oxidation of small hydrocarbons with a relatively high level of
detail. Jet A, C1, and C5 are modeled as C11H22, C13H28, and C10H19, respectively. For com-

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1319

patibility with existing codes, the formulas are rounded from the experimentally measured
formulas C11.4H21.7, C12.5H27.1, and C9.7H18.7. This approximation has the disadvantage of
changing the H/C ratio trend among the fuels: the measured H/C of Jet A, C1, and C5 are
1.91, 2.16, and 1.93, whereas their modeled H/C are 2.0, 2.15, and 1.9.
To evaluate the performance of the chosen chemical reaction mechanisms, and to com-
pare the fuels on a fundamental level, laminar flame speeds (su ) and ignition delay times
(τign ) are computed using Ansys Chemkin (Ansys, 2021) and presented in Fig. 2 alongside
experimental measurements. Laminar flame speeds are obtained from 1D laminar atmo-
spheric flame simulations at 403 K and ignition delay times through 0D homogeneous stoi-
chiometric reactor simulations at 10 atm. The results from the HyChem mechanisms (which
are also used in the LES) are shown in Fig. 2(a) and (c). For comparison, results obtained
with the detailed 17790-step mechanism of Ranzi et al. (2014), are shown in Fig. 2(b) and
(d). The fuel mixtures are based on the surrogate compositions in Fig. 1.
The laminar flame speeds and ignition delay times predicted by the HyChem mecha-
nisms, Fig. 2(a) and (c), match the corresponding experimental data well. It should be noted
that these parameters may vary among different varieties of the same fuel; this is exempli-
fied by the difference between the laminar flame speeds of Wang et al. (2018a) and Kumar
et al. (2011), which are both measured for Jet A. C5 always ignites slower than Jet A as
it contains a larger portion of slowly igniting species (particularly aromatics). C1 has the
slowest ignition at high temperatures due to the slow oxidation of iso-butene, but the fastest
ignition at intermediate temperatures due to the rapid decomposition of the fuel species. At
low temperatures (1000/T > 1), C1 and C5 ignite very slowly as a result of the C1 and C5
mechanisms lacking a submodel for Negative Temperature Coefficient (NTC) behavior at

Fig. 2 Laminar flame speeds and

ignition delay times predicted
by the HyChem (a,c) and Ranzi
17790-step (b,d) chemical
reaction mechanisms, alongside
experimental measurements
(symbols). Legend: Jet A (red),
C1 (green), C5 (magenta),
squares (Xu et al., 2018; Wang
et al., 2018a, b; HyChem, 2020),
circles (Kumar et al., 2011)

13
1320 Flow, Turbulence and Combustion (2025) 114:1315–1339

present. The impact of this limitation should be low for the present LES study, where the
flame dynamics occur on a scale of 1 ms or less. The fuel trend predicted by the detailed
mechanism, Fig. 2(b) and (d), is different from that predicted by the HyChem mechanisms;
Jet A and C1 have identical laminar flame speeds whereas C5 is slower, and C1 has the fast-
est ignition across all temperatures. This difference in predictions demonstrates the uncer-
tainty associated with capturing the kinetics of complex mixtures of large hydrocarbons.
Since the HyChem mechanisms are used for the remainder of this work, their kinetics form
the theoretical basis in the discussion of the LES results.

3 LES Modeling and Numerical Methods

The LES model consists of implicitly filtered equations for continuity (Eq. 1a), momentum
transport (Eq. 1b), energy transport (Eq. 1c), and species mass fraction transport (Eq. 1d),
which are closed using constitutive equations for Fickian diffusion and Fourier heat con-
duction (Menon and Fureby, 2010) in a linear viscous ideal mixture. Overbars denote spa-
tial filtering and tildes denote density-weighted (Favre) spatial filtering. Pressure, density,
velocity, enthalpy, and species mass fractions are represented by p, ρ, ṽ, h̃, and Ỹi , respec-
tively. The stagnation enthalpy H̃K = h̃ + 12 |ṽ|2 is employed for a more compact energy
equation. The filtered viscous stress tensor is defined as τ = 2µD̃ − 23 µ(∇ · ṽ)I, where
D̃ = 12 (∇ṽ + ∇ṽT ) is the filtered strain rate tensor. The viscosity µ and thermal conduc-
tivity κ are computed using Sutherland’s law with coefficients for dry air. The thermal dif-
fusivity is defined as α = κ/cp , where cp is the specific heat capacity of the mixture, and a
unity Schmidt number is assumed for all species diffusivities Di . Because the targeted case
contains a highly turbulent flame, we expect laminar diffusion to be significantly weaker
than turbulent diffusion.
 ∂ ρ̄


 + ∇ · (ρ̄ṽ) = ρ̇l ,
 ∂t



 ∂ ρ̄ṽ + ∇ · (ρ̄ṽ ⊗ ṽ) = −∇p̄ + ∇ · τ + f − ∇ · B,

 l
∂t
(1)
 ∂ ρ̄H̃K ∂ p̄


 + ∇ · (ρ̄ H̃ K ṽ) = + ∇ · (τ ṽ) + ∇ · (ᾱ∇ h̃) + q̇ c + q̇ l − ∇ · b E ,

 ∂t ∂t



 ∂ ρ̄Ỹi + ∇ · (ρ̄Ỹ ṽ) = ∇ · (D ∇Ỹ ) + ω̇ + ṁ − ∇ · b .
i i i i il i
∂t

The effect of the unresolved turbulent scales is captured by the subgrid terms B bE , and bi ,
which (following Boussinesq’s hypothesis) are explicitly modeled as,
 1

B = 2ρ̄(−νsgs D̃D + kI),


 3
νsgs
bE = −ρ̄ ∇h̃, (2)

 PrT


bi = −ρ̄ νsgs ∇Ỹi ,
ScT

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1321

where D̃D is the deviatoric part of the strain rate tensor. The turbulent Prandtl and Schmidt
numbers, Prt and Sct , are set to 0.85 and 1, respectively. The Localized Dynamic K-equa-
tion Model (LDKM) (Kim and Menon, 1995), a dynamic extension of the One Equation
Eddy Viscosity Model (OEEVM) (Yoshizawa and Horiuti, 1985), is used to compute the
subgrid viscosity νsgs . LDKM involves solving a transport equation for the subgrid kinetic
energy k ,

∂ρk ρk 3/2
+ ∇ · (ρkṽ) = −B · D̃ + ∇ · (ρ∇k) − Cε .(3)
∂t ∆
√
The subgrid viscosity is computed from k using the relation νsgs = Ck k∆, where ∆, the
LES filter width, is interpreted as the cubic root of the local cell volume. The model coef-
ficients are computed dynamically as described in Kim and Menon (1995).
∑N
The terms ω̇i and q̇c = i=1 hθi ω̇i (where hθi is formation enthalpy) represent heat
and mass transfer due to chemical reactions. Here, these terms are modeled using a finite-
rate chemistry approach with turbulence-chemistry interactions captured by the Partially
Stirred Reactor (PaSR) model (Sabelnikov and Fureby, 2013). The PaSR model assumes
that chemical reactions are localized in fine structures in a turbulent environment, with
the local volume fraction of fine structures represented by γ ∗ . Using the chemical time
scale τc and the turbulent mixing time scale τm , the filtered reaction rates are computed as
ω̇i ≈ γ ∗ ω̇i (ρ, Ỹk , T̃ ), where γ ∗ = τc /(τm + τc ). Here, τc is modeled as τc = δu /su , where
δu is the stoichiometric laminar flame thickness and su the stoichiometric laminar flame
speed. There are several alternative τc formulations (Péquin et al., 2022), but the one used
here is motivated by our previous observation that that the flame is locally premixed (Åkerb-
lom and Fureby, 2024) making the laminar premixed flame propagation rate a natural choice
√
for τc . The mixing time scale is modeled as τm = τK τ∆ , where τ∆ = ∆/v ′ is the time
scale of subgrid velocity stretch based on the magnitude of subgrid velocity fluctuations v ′
and τK is the Kolmogorov time scale. Both τ∆ and τK are computed dynamically based on
the local k .
The fuel spray is modeled separately from Eqs. 1a-1d, but interacts with the gaseous
phase via the source terms ρ̇l , fl , q̇l , and ṁil , which represent inter-phase transfer of mass,
momentum, energy, and species. The spray itself is modeled as a cloud of Lagrangian par-
ticles following a separate set of governing equations (Faeth, 1987), where each particle
represents a statistical group of fuel droplets (Dukowicz, 1980). A stochastic droplet colli-
sion model, based on O’Rourke (1981), captures inter-particle interactions near the injec-
tion point. Secondary droplet breakup is captured by the Reitz-Diwakar model (Reitz and
Diwakar, 1987), which accounts for the “stripping” and “bag” breakup modes. The droplets
are subjected to gravity and sphere drag, and CD is computed using the relation of Crowe
(1998). The vaporization rate is computed from droplet Reynolds and Schmidt numbers,
also following Crowe (1998). The inter-phase heat transfer coefficient is obtained using the
Ranz-Marshall model (Ranz and Marshall, 1952). The evaporation model is not affected
by subgrid fluctuations, which may lead to underpredicted spray dispersal. Since the spray
and flame are not expected to directly overlap, surface reactions and liquid-phase chemistry
are neglected. The thermodynamic properties of each fuel are provided by the correlations
proposed in Åkerblom et al. (2024).

13
1322 Flow, Turbulence and Combustion (2025) 114:1315–1339

Fig. 3 Cutaway of the targeted case. Solid sections of the burner are shown in red. Walls are colored by
temperature

Table 2 Experimental operating Idle Cruise

conditions, as reported in Meier
Air pressure [MPa] 0.4 1.0
et al. (2012)
Air preheat temperature [K] 550 650
Burner pressure loss [%] 3 3
Burner air mass flux [g/s] 60 140
Cooling air mass flux [g/s] 17 39
Liquid fuel mass flux [g/s] 3.0 6.8

4 Targeted Case and Simulation Setup

The targeted case is the high-pressure single-cup combustor (Behrendt et al., 1999; Meier et
al., 2012) at the DLR Institute of Propulsion Technology. A 3D model of the combustor is
presented in Fig. 3. It is 264 mm long and features a square cross-section of 102 × 102 mm2.
Compressed and preheated air is supplied to a plenum (the main air) and cooling slots by
the side walls. The main air passes through a burner, where two sets of swirlers induce co-
rotating flows. The burner uses prefilming airblast atomization; a thin film of fuel is injected
along the inner wall, then broken apart by the airflow at the edge of the prefilmer lip. Enter-
ing the combustion chamber, the fuel-carrying air expands into a cone of axial flow, or
Main Flow Cone (MFC), which encapsulates a Central Recirculation Zone (CRZ) and is
surrounded by an Outer Recirculation Zone (ORZ). The high-shear boundary between the
MFC and the CRZ is labeled the Inner Shear Layer (ISL), and there is likewise an Outer
Shear Layer (OSL) between the MFC and the ORZ. The ISL provides a region for the flame
to stabilize, but there is also a substantial amount of heat release inside the MFC. The fuel
is fully combusted within the first 100 mm of the combustion chamber, after which the hot
products are discharged through a converging sonic exit nozzle.
Two operating conditions, representing idle and cruise conditions in a jet engine, were
used in the experiments of Meier et al. (2012). The temperatures, pressures, and mass flows
associated with these conditions are given in Table 2, and both operating conditions are
included in the present study. Note that the mass flows in Table 2 are used for all fuels,
which results in slightly different global equivalence ratios ϕ. At idle conditions, the equiva-

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1323

lence ratios for Jet A, C1, and C5 are 0.735, 0.745, and 0.728 (disregarding cooling air). At
cruise conditions, they are 0.714, 0.724, and 0.707.
It should be noted that the targeted case lacks some common features of real aeroengines.
There is only a single fuel injection slot, with no separation between main and pilot fuel.
The air-to-fuel ratio in the burner is lean, and there are no dilution jets. This means that
combustion occurs fully under lean conditions, in contrast to conventional rich-burn com-
bustors. These differences heavily affect the emission profile of the rig, which should not be
expected to match measured emissions from any real engine. The case remains useful for
comparing fuels under conditions that are thermodynamically similar to real engines during
idle and cruise.
The simulation domain begins in the plenum and ends at the throat of the exit nozzle.
Fixed mass-flow inlets are used for the main air and cooling slots, and a wave-transmissive
condition reduces wave reflection at the outlet. All walls are treated as adiabatic no-slip
surfaces, and Spalding’s law of the wall (Spalding, 1961) is used to adjust subgrid viscosity
to account for unresolved near-wall effects. The fuel is introduced just downstream of the
prefilmer lip, fully atomized, at a rate of 109 Lagrangian particles per second. Initial droplet
sizes follow the Rosin-Rammler distribution suggested by Jones et al. (2014) for Jet A.
For C1 and C5, the distribution is shifted so that its relative Sauter Mean Diameter (SMD)
matches that predicted for each fuel by Lefebvre’s correlation (Lefebvre, 1980); this means
that the low-viscosity, low-surface tension fuel C5 has slightly smaller droplets than C1,
which in turn has slightly smaller droplets than Jet A. The injection velocity distribution was
found by iteration with experimental data from Meier et al. (2012) as a target.
In our previous work (Åkerblom and Fureby, 2024) the liquid was injected parallel to
the wall upstream of the prefilmer lip, forming a dense cloud of droplets similar to a liquid
film. That injection method resulted in very low droplet dispersion and an overly narrow
spray cone. Since the flow velocity was low along the lip, the mass of liquid fuel there could
oscillate in response to pressure oscillations in the combustion chamber, enabling a very
strong thermoacoustic response. Lo Schiavo et al. (2020) observed a similar effect, where
applying a liquid film model led to strong thermoacoustic fluctuations. The improved injec-
tion method used here was proposed in Åkerblom et al. (2024).
The assumption of adiabatic walls is not trivial; by neglecting heat losses to the walls the
combustion product temperature is overpredicted, which may affect the size of the CRZ as
the density gradient between the CRZ and ORZ is smaller. Since our focus is on comparing
fuels, we nevertheless prefer a simplified thermal boundary condition over more realistic
alternatives, which were found by Agostinelli et al. (2021) to have a strong influence on the
acoustic response in LES of a swirl-stabilized combustor.
The domain is discretized using a hexahedral mesh of 6.0 million cells, with characteristic
lengths of 0.33 to 0.5 mm in the combustion zone. The average ratio between resolved and
total turbulent kinetic energy is ~85%, with higher ratios near the flame. We have previously
demonstrated (Åkerblom and Fureby, 2024) that mesh refinement has an effect on flame lift,
but not more so than a change of turbulence-chemistry interaction model. Although a finer
mesh could be preferable, the 6.0 million cell mesh is adequate for comparing fuels.
The governing equations are solved using a finite volume code based on the OpenFOAM
C++ library (Weller et al., 1998). Convective fluxes are reconstructed by a monotinicity-
preserving scheme and diffusive fluxes are reconstructed using central differencing to
achieve second-order accuracy. The equations are temporally integrated using the single-

13
1324 Flow, Turbulence and Combustion (2025) 114:1315–1339

point implicit Euler scheme. The gas phase is treated as weakly compressible and the gov-
erning equations are decoupled using the Pressure-Implicit Splitting of Operators (PISO)
(Bressloff, 2001) algorithm. Chemical source terms are computed separately with operator
splitting, and the chemical state is integrated in time using an extrapolation algorithm based
on the implicit Euler scheme (Hairer and Wanner, 1991). The Rosenbrock integration algo-
rithm used in our previous work (Åkerblom and Fureby, 2024) was later found to produce
mass conservation errors that likely resulted in overly strong thermoacoustic coupling. The
Courant number is kept below 0.95 to ensure stability. Each simulation is initialized from a
generic reacting state and allowed to stabilize over 30 ms, after which data is collected for
20 ms. The data collection period corresponds to ~3.2 and 3.5 flow-through times at idle and
cruise conditions, respectively.
Emission Indices (EI) are obtained directly from the simulations for all species except
NOx, which is not included by default in the HyChem chemical reaction mechanisms.
Instead, EINOx is computed by extracting a Probability Density Function (PDF) of heat
release at different mixture fractions z over the course of the simulation. This heat release
PDF is then multiplied with EINOx as a function of z , obtained from separate 1D laminar
flame simulations in Ansys Chemkin, Ansys (2021). Integrating the product of these func-
tions across all z produces an estimate for the total EINOx. Following Saggese et al. (2020),
the laminar flame simulations are carried out using the HyChem mechanisms coupled with
the NOx submodel of Glarborg et al. (2018). Estimating EINOx in this way assumes that
NOx production in laminar premixed flames is similar to that in spray flames, and that recir-
culated NOx does not greatly impact the combustion process or the production of new NOx.
Neither of these assumptions is trivial, and including NOx chemistry directly in the LES
would be preferable. However, Saggese et al. (2020) showed that both thermal and prompt
NOx formation pathways contribute significantly to the formation of NOx in Jet A flames,
and chemical reaction mechanisms that include both of these pathways are too computation-
ally expensive for this study.

5 Results

Figure 4 shows snapshots of the temperature and axial velocity in the simulations, obtained
from a central cut through the first 100 mm of the combustion chamber. The Heat Release
Rate (HRR) contour reveals a heavily turbulent flame stabilized at the ISL, with a significant
amount of heat also being released within the MFC and at the OSL. Peak temperatures are
found in small pockets inside the flame, where z is close to its stoichiometric value zst . The
hot combustion products are recirculated through the CRZ to stimulate continuous ignition.
The temperature in the ORZ is colder than in the downstream region due to mixing with
cold air from the window cooling film, and due to a lack of fuel at the base of the OSL.
Figure 5 shows the time-averaged temperature and axial velocity. Note that here, the
black contour demarcates ⟨HRR⟩ = max(⟨HRR⟩)/10 instead of a common fixed value.
Although the burner contains a central region of reversed flow, the hot combustion products
generally stay within the combustion chamber. At idle conditions, the fuels have qualita-
tively identical statistics. The flame lift (informally defined as the distance between the
burner and the HRR center) is shorter at cruise conditions, especially for C1 and C5.

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1325

Fig. 4 Instantaneous temperature (left) and axial velocity (right) for each fuel at idle (top) and cruise (bot-
tom) conditions. The HRR = 108 W/m3 contour is shown in black

The time-averaged temperature and axial velocity are also shown in Figs. 6 (idle) and 7
(cruise) as radial profiles extracted at four locations inside the flame and one in the down-
stream region, at x = 90 mm (with the origin at the end of the prefilmer lip). There is a
slight temperature difference between the fuels at idle conditions, which is correlated to
their slightly different ϕ and LHV. At cruise conditions, short flame lift is visible as a rapid
increase in temperature, but the fuels return to the ϕ- and LHV-based trend after combus-
tion is complete (x = 90 mm). The velocity profiles reveal a slight difference between the
fuels in the cone angle of the MFC, with C1 having the narrowest cone and C5 the widest.
We believe this to be the result of different product temperatures, as a high density gradient
between reactants and products may expand the CRZ.
The fuel spray is quantitatively analyzed in Fig. 8 using radial profiles of SMD, axial
velocity, and temperature at three different locations within the spray. The vaporization
progress, defined as the reduction in axial liquid fuel mass flux since injection, is given as
percentages. Results are only shown at locations where at least 100 Lagrangian particles
were observed during the simulation. The reliability of the data is low at x = 20 mm, as
only a small amount of liquid remains at that point. The SMD of Jet A is somewhat under-
predicted far from the centerline; Meier et al. (2012) noted that large droplets tend to be
transported further outward due to inertia, but this effect is not considered in the injection
model and does not manifest post-injection. The results could likely be improved by assign-
ing different velocity distributions to different droplet size classes.

13
1326 Flow, Turbulence and Combustion (2025) 114:1315–1339

Fig. 5 Time-averaged temperature (left) and axial velocity (right) for each fuel at idle (top) and cruise
(bottom) conditions. The ⟨HRR⟩ = max(⟨HRR⟩)/10 contour is shown in black

Fig. 6 Radial profiles of time-averaged axial velocity (a) and temperature (b) at idle conditions. Profiles
are presented at four locations inside the flame and one further downstream. Legend: Jet A (red), C1
(green), C5 (magenta)

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1327

Fig. 7 Radial profiles of time-averaged axial velocity (a) and temperature (b) at cruise conditions. Pro-
files are presented at four locations inside the flame and one further downstream. Legend: Jet A (red), C1
(green), C5 (magenta)

Fig. 8 Spray statistics at idle (a) and cruise (b) conditions, at three distances from the burner. Top to bot-
tom: SMD, axial velocity, temperature. Percentages refer to vaporization progress. Legend: Jet A (red),
C1 (green), C5 (magenta), experimental data obtained from Meier et al. (2012) for Jet A (black)

There are clear differences between the sprays of the three fuels. C5 has low viscosity,
surface tension, and density, which results in small droplets, whereas C1 has slightly larger
droplets due to its higher viscosity. This is partly an extrapolation of the injection condition,
but the Reitz-Diwakar model also ensures that C1 and particularly C5 form smaller drop-
lets via secondary breakup. The liquid properties of each fuel have a negligible impact on
the velocity profiles, but the vaporization progress demonstrates that Jet A has the longest
penetration depth, followed by C1, then C5. The liquid thermodynamics model (Åkerblom
et al. 2024) predicts the saturation temperatures of Jet A, C1, and C5 to be 573 K, 543 K,
and 495 K at 4 bar (idle), and 635 K, 601 K, and 547 K at 10 bar (cruise). The saturation
temperature represents the upper limit of possible droplet temperatures, which is reflected

13
1328 Flow, Turbulence and Combustion (2025) 114:1315–1339

in the average temperature profiles. The highest temperatures are found at the outer edge of
the spray, where the hot gases in the ORZ continuously heat it. The droplets are generally
somewhat colder at the inner edge of the spray, despite the considerably higher HRR there.
We attribute this to a greater variability in droplet temperatures near the flame, where bursts
of instant vaporization are followed by an influx of cold, unvaporized fuel. The temperature
in the ORZ may also be overpredicted in the simulations due to the assumption of adiabatic
walls, accelerating the vaporization of the outermost droplets.
The time-averaged spray density, HRR, and OH mass fraction distributions of the fuels
are qualitatively compared in Figs. 9 (idle) and 10 (cruise). The data is azimuthally inte-
grated, and there is a slight asymmetry since the integration is carried out separately for
two halves of the domain. Experimental data from Meier et al. (2012) are included for
comparison, obtained using Mie scattering and kerosene-LIF for the spray, Abel-inverted
OH chemiluminescence for HHR, and OH-PLIF for the OH concentration. These measure-
ments do not correspond directly to the data extracted from the simulations and should not
be expected to match up perfectly. The OH-PLIF images are actually temperature images
based on OH-PLIF signal intensity, but they still function as visualizations of the OH dis-
tribution. All simulation results are individually normalized and follow the same color map
as the experimental data.
Compared to the experiments, the flame appears overly lifted at idle conditions. As
shown by Fig. 8, the spray injection condition is quite accurate in terms of SMD and veloc-
ity, but the injection temperature may be too low. Describing the rig, Meier et al. (2012)
suggested that there should be some preheating of the fuel before atomization, due to its
proximity to the preheated air. A relatively cold spray also inhibits heat release along the
ISL, which, alongside the assumption of adiabatic walls, may contribute to the overly nar-
row angle of the MFC in the simulations. The spray, flame, and OH distribution for the Jet A
flame are nevertheless predicted adequately at idle conditions and well at cruise conditions.
The volatility of C5, and to a lesser extent C1, results in a short spray penetration depth and

Fig. 9 Time-averaged, azimuthally integrated liquid concentration (a), HRR (b), and OH mass fraction (c)
at idle conditions. Experimental kerosene-PLIF, OH chemiluminescence, and OH-PLIF measurements
obtained from Meier et al. (2012) are included in the leftmost column for comparison

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1329

Fig. 10 Time-averaged, azimuthally integrated liquid concentration (a), HRR (b), and OH mass fraction
(c) at cruise conditions. Experimental kerosene-PLIF, OH chemiluminescence, and OH-PLIF measure-
ments obtained from Meier et al. (2012) are included in the leftmost column for comparison

more heat release along the ISL. At cruise conditions, the C5 flame also appears to retain
less OH in the downstream region compared to the other fuels.
The chemistry of the combustion process is visualized in Fig. 11, where the spray and
several important gaseous species are shown in a central cross-section. The “fuel” species
refers to the fuel surrogate species, i.e., C11H22, C13H28, and C10H19 for Jet A, C1, and C5,
respectively. The fuel spray (magenta) is heated by the surrounding air and transitions to the
gaseous phase (green). The gaseous fuel then decomposes into smaller hydrocarbons, nota-
bly ethylene and iso-butene, shown together in blue. The decomposition products ignite,
producing radicals such as OH (red) and major products such as CO2 (yellow). Regions
where OH and CO2 coexist are shown in orange. At idle conditions, the spray is typically
separated from the hot products by a layer of gaseous reactants ~10 mm thick. The evapora-
tion, mixing, and oxidation steps occur sequentially, giving the flame premixed characteris-
tics. The separation is much smaller at cruise conditions, where all steps of the combustion
process occur in very close proximity and are more coupled to each other. We argue that
this increased coupling is why the fuels are more distinct at cruise conditions; specifically,
the liquid thermodynamic properties of each fuel play a large role in determining the flame
shape.
To compare the combustion process of the fuels in greater detail, the time-averaged den-
sities of key species are radially averaged and plotted along the axial direction in Figs. 12
(idle) and 13 (cruise). Temperature is included as well. Note that each reaction mechanism
has its own list of species and pathways, which results in small density differences when
measuring specific species. Most of the combustion process occurs within x < 50 mm, and
the fuel decomposition products are notably different for each fuel. C5 produces ~40% more
ethylene than Jet A at idle conditions whereas C1 produces ~70% less. The ethylene densi-
ties in the Jet A and C5 flames are more similar at cruise conditions, but there the C5 flame
is also much more compact, which reduces the radially averaged density of all intermediate
species. Ethylene is considered a soot precursor, and the results therefore hint that C5 should

13
1330 Flow, Turbulence and Combustion (2025) 114:1315–1339

Fig. 11 Instantaneous cross sections of the spray (magenta, colored by temperature) and important gas-
eous species including the fuel species (green), decomposition products ethylene and iso-butene (blue),
OH (red), and CO2 (yellow)

Fig. 12 Radially averaged density of important gaseous species, as well as temperature, plotted along the
longitudinal direction at idle conditions. Density axes that do not start at zero are colored blue. Legend:
Jet A (red), C1 (green), C5 (magenta)

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1331

Fig. 13 Radially averaged density of important gaseous species, as well as temperature, plotted along the
longitudinal direction at cruise conditions. Density axes that do not start at zero are colored blue. Legend:
Jet A (red), C1 (green), C5 (magenta)

have the highest sooting propensity, followed by Jet A, then C1. The trend is corroborated
by Xue et al. (2017), who measured soot formation for Jet A and an ATJ fuel similar to C1,
and El Helou et al. (2023), who compared Jet A and C5. Although we expect the same soot-
ing trend here, we do not consider the densities of a small number of soot precursor species
to be sufficient information to make that conclusion; further simulations with a dedicated
soot formation model are required.
As shown in Sect. 2, C5 ignites more slowly than Jet A due to its high aromatics content.
C1 has the fastest ignition at intermediate temperatures due to its rapid initial decomposi-
tion, but the slowest ignition at high temperatures due to the relatively slow oxidation of
iso-butene. We may expect rapid ignition to result in short flame lift, but no such trend is
discernible in the LES results. Comparing the flame lift of different fuels in an unconfined
burner, Alsulami et al. (2021) observed that atomization and vaporization had a greater
impact than ignition chemistry. On the other hand, Esclapez et al. (2017) observed in
their simulations of the NJFCP Referee Rig that the rapid ignition of C1 reduced flame
lift whereas the volatility of C5 only had a marginal effect. We believe that both ignition
and vaporization can be determining factors, but the relative importance of each is case-
dependent. In the present work, where droplet diameters are similar to those in the study of
Alsulami et al. (2021), and smaller than the droplets in the study of Esclapez et al. (2017),
vaporization appears dominant. However, the rapid ignition of C1 at low temperatures may
also contribute to its short flame lift at cruise conditions, which is similar to that of C5
despite C1 being less volatile and having a longer spray penetration depth.
The densities of CO and OH are virtually identical between the fuels at idle conditions.
At cruise conditions, the high flame temperature results in a relatively high OH density in
the postflame zone, particularly for C1. The high OH density of C1 is a result of the differ-
ent chemical pathways taken by ethylene and iso-butene during combustion. The complete
combustion products CO2 and H2O follow the temperature curve, and the different H/C
ratios of the fuels visibly affect the relative densities of CO2 and H2O. As explained in
Sect. 2, however, the H/C ratios of the surrogate fuel species do not exactly match experi-

13
1332 Flow, Turbulence and Combustion (2025) 114:1315–1339

mental observations, particularly for Jet A and C5. The observed differences between the
CO2/H2O ratios of Jet A and C5 in the simulations are therefore likely exaggerated com-
pared to the real fuels.
The emission profiles of Jet A, C1, and C5 depend on the chemical composition of each
fuel but are also affected by flame dynamics. The local mixture fraction z in the flame deter-
mines the flame temperature and consequently impacts the rate of thermal NOx formation.
The EI of CO2, H2O, CO, OH, and NOx are given in Table 3. EI is computed by taking the
ratio between the selected species and the local combined mass of hydrogen and carbon
(which can only originate from the fuel) and radially averaging it at x = 200 mm, shortly
before the contraction at the end of the combustion chamber. The exception is EINOx, which
is computed indirectly following the procedure outlined in Sect. 4.
EICO2 and EIH2O are determined by the H/C ratio of each fuel and are only weakly
affected by operating conditions. This is reflected in Table 3, where EICO2 and EIH2O
change by ˂1% between idle and cruise conditions. As an inorganic radical, OH is unsur-
prisingly correlated with H/C ratio and temperature, with ~20% higher EIOH at cruise
conditions. The EICO results exhibit some unexpected trends that warrant further discus-
sion. EICO is ~1 at idle conditions, slightly lower than at cruise conditions. Meanwhile, the
average EICO across all engines in the ICAO Aircraft Engine Emissions Databank (ICAO,
2023a) is 26 at idle conditions, 3.9 during approach, and 0.59 at climb-out. Since cruise
conditions lie between approach and climb-out in terms of power, the predicted EICO is rea-
sonably representative at cruise conditions but very low at idle conditions. This discrepancy
is most likely due to the design of the targeted case. The single injector has a lean global
ϕ, meaning CO production is low overall. There is also no quenching from dilution jets,
which means that the CO is given time to form CO2. Although the same applies at cruise
conditions, the error is smaller since engines are designed for low EICO during high-power
operation. The predicted EICO is correlated with H/C despite the carbon content of CO, but
since H/C is also correlated with ϕ (for a constant fuel mass flux) and flame temperature, the
net effect is that EICO increases with H/C.
There are substantial differences between the NOx emissions of the three fuels; C5 has
relatively low NOx emissions at idle conditions, whereas Jet A has relatively low NOx emis-
sions at cruise conditions. The trend is heavily correlated with the local mixture fraction in
the flame, as illustrated by Fig. 14, where HRR and EINOx are presented as functions of
z/zst . The means µf uel and standard deviations σf uel of the HRR PDFs are also provided.
At idle conditions, all fuels burn at approximately the same z/zst ; this is logical consid-
ering how similarly the three fuels behave overall. The mixture fraction is correlated with
the global ϕ of each fuel, but Jet A also has a slight tendency to burn at near-stoichiometric

Table 3 Emission indices of CO, CO2, OH, H2O, and NOx, given in grams of emissions per kilogram of
burnt fuel
Idle Cruise
Jet A C1 C5 Jet A C1 C5
ϕ = 0.735 ϕ = 0.745 ϕ = 0.728 ϕ = 0.714 ϕ = 0.724 ϕ = 0.707
EICO2 3140 3110 3160 3150 3110 3160
EIH2O 1280 1370 1230 1280 1370 1230
EICO 1.1 1.2 0.9 1.3 1.7 1.0
EIOH 7.7 8.4 6.8 8.9 10.5 7.8
EINOx 2.4 2.3 1.4 15 27 24

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1333

Fig. 14 Probability distributions of HRR versus z/zst (solid lines), and EINOx as a function of z/zst
(dashed lines) at idle (a) and cruise (b) conditions. Means and standard deviations of HRR are denoted
µf uel and σf uel . Legend: Jet A (red), C1 (green), C5 (magenta)

z . These z occur where the flame comes close to the spray, which accelerates vaporization
and forms a cold, fuel-rich gaseous mixture. Because the Jet A reaction mechanism contains
an NTC submodel, it may allow the fuel to be ignited under these conditions where the C1
and C5 mechanisms would instead have the fuel undergo more heating and mixing first; we
consider this unlikely, however, as Fig. 2 shows that the time scales of the NTC effect are
on the order of 0.01 s, which is longer than the flame flow-through time. Although Jet A has
a slightly lower flame temperature than C1, its higher mean z/zst gives it a slightly higher
EINOx overall. C5 has a low mean z/zst and a low flame temperature, giving it the lowest
EINOx.
At cruise conditions, all fuels regularly burn under stoichiometric and fuel-rich condi-
tions, although the mean z/zst is still lean. The locations of the HRR peaks in Fig. 14
are determined by the global ϕ, but the amount of high-z/zst combustion is correlated
with the volatility of each fuel. The most volatile fuel, C5, has the most HRR close to the
spray, where z/zst is high. C1 is also quite volatile but experiences slightly less fuel-rich
combustion than C5. Jet A, with its comparatively low volatility and lifted flame, burns at
a considerably lower z/zst . The net trend is that C1 has the highest EINOx due to its high
flame temperature, closely followed by C5, then Jet A with a significantly lower EINOx.
Atomization and vaporization evidently play a large role in determining NOx emissions.
In our previous validation study with Jet A and JP-5 (Åkerblom and Fureby, 2024), we
observed thermoacoustic oscillations belonging to the first longitudinal (i.e., streamwise)
acoustic mode (~1.6 kHz) and the first azimuthal acoustic mode (~4.2 kHz) of the com-
bustor. The longitudinal and azimuthal oscillations both had amplitudes of 0.3% to 0.5%
(relative to the mean combusor pressure) at idle conditions. The longitudinal oscillations
were considerably stronger at cruise conditions, reaching 2.6% for Jet A and 6.1% for JP-5,
whereas the azimuthal oscillations remained relatively low at 0.3% and 1.2%. The longi-
tudinal oscillations significantly impacted the flame dynamics at cruise conditions, causing
visible resonant fluctuations in the flow velocity, HRR, and fuel concentration. The pressure
oscillated in phase with the HRR, satisfying the Rayleigh criterion for thermoacoustic insta-
bilities (Rayleigh, 1878). The same acoustic modes are excited in the present study, with
time-averaged amplitudes given in Table 4. Note that the amplitudes are significantly lower
in the present work, particularly at cruise conditions; this is the result of switching chemistry
integration algorithm and moving the spray injection point downstream of the prefilmer lip.

13
1334 Flow, Turbulence and Combustion (2025) 114:1315–1339

Table 4 Mean amplitudes of azimuthal and longitudinal pressure oscillations

Idle Cruise
Jet A C1 C5 Jet A C1 C5
Longitudinal [%] 0.22 0.24 0.25 0.18 0.97 1.31
Azimuthal [%] 0.049 0.059 0.049 0.09 0.12 0.09

Fig. 15 Normalized liquid mass fluctuation (black) and normalized pressure fluctuation (red) over time at
idle (a) and cruise (b) conditions. Top to bottom: Jet A, C1, C5

All fuels experience similar pressure oscillations at idle conditions, with longitudinal
amplitudes of ~0.2% and azimuthal amplitudes of ~0.05%. The azimuthal amplitudes are
also similar for all fuels at cruise conditions, but there is a clear trend among the longitudinal
amplitudes. Jet A has the same amplitude as at idle conditions. C1 has considerably stronger
oscillations at 0.97%, and C5 stronger still at 1.31%. Because the amplitude is subject to
low-frequency (˂100 Hz) changes over time due to flow fluctuations, the difference between
the time-averaged amplitudes for C1 and C5 is not statistically significant; a considerably
longer simulation would be required to conclusively establish a trend. Here, C5 exhibits
stronger longitudinal fluctuations than C1 over 85% of the data collection interval. The
substantial difference between Jet A and the other two fuels is significant, however. It is cor-
related with fuel volatility, suggesting that a fast vaporization rate allows the liquid fuel con-
centration to more easily couple to pressure oscillations. This is corroborated by Rajendram
Soundararajan et al. (2021), who observed stronger thermoacoustic coupling for n-heptane
compared to less volatile n-dodecane in an annular spray combustor.
The coupling between the spray and pressure oscillations is visualized in Fig. 15, where
the fluctuations of the total liquid mass and the near-spray pressure are presented over
10 ms. Both are given as percentages of their respective means. At idle conditions, there is
no visible coupling between the liquid fuel mass and the pressure. The same applies to Jet A
at cruise conditions, but not to C1 and C5. For the latter two fuels, the liquid oscillates at the

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1335

same frequency as the pressure but with a phase shift of π . There are two factors contribut-
ing to this phase shift. Firstly, the pressure peaks are accompanied by HRR peaks, which
increase the fuel consumption rate. Secondly, the pressure peaks create an adverse pressure
gradient between the prefilmer lip and the combustion chamber, which temporarily slows
the transport of liquid fuel and causes it to build up. Note that the liquid fuel injection rate
is constant.
In a recent experimental study on an afterburner rig, Bae et al. (2024) also observed
considerably stronger thermoacoustic oscillations for C1 and C5 compared to Jet A and
other fuels. This occurred despite the fact that the liquid jet was injected far upstream of the
flame holder. Among the fuels studied (which also included JP-8 and C9), a trend emerged:
higher CN resulted in lower ϕ for LBO and flashback. C1 and C5, which had the lowest CN,
experienced strong thermoacoustic oscillations at high ϕ which eventually led to blow-out
instead of flashback. The authors suggested that these relatively strong pressure oscillations
were connected to the low CN of C1 and C5. Paxton et al. (2020), studying a pre-vaporized
and premixed bluff-body flame, noted that fuels with high CN experience thermoacoustic
instabilities at lower ϕ compared to fuels with low CN, e.g., C1. At sufficiently high ϕ (the
limit cycle region), similar pressure amplitudes were observed for all fuels. Both Bae et al.
(2024) and Paxton et al. (2020) thus demonstrated that a high CN lowers the upper limit
of stable ϕ, where further increases in ϕ would would lead to thermoacoustic oscillations
and potentially flashback or blow-out. The same trend was not clearly demonstrated by
Rajendram Soundararajan et al. (2021), whose n-heptane spray flame experienced stronger
thermoacoustic oscillations than an n-dodecane flame despite the considerably higher CN
of the latter. However, the ϕ in that study were quite high (0.85 to 0.95) which means that
both fuels were likely above the instability threshold.
CN may be a dominant factor in the present study as well. The time scales associated
with the pressure oscillations are on the order of 10−4 s, and Fig. 2(c) shows that τign is
inversely correlated with CN at these time scales (according to the HyChem mechanisms).
However, the influence of CN cannot be determined without further simulation campaigns
at different ϕ and more fuels. If the ϕ used in the present work (~0.71 at cruise conditions)
is close to the limits of the stable range, CN may be an important indicator of the likelihood
and strength of thermoacoustic oscillations. If it is well within the stable range, then volatil-
ity (which is independent of ϕ) is likely the determining factor.

6 Conclusions

The combustion of conventional Jet A and two category C fuels, C1 and C5, is simulated
using LES in a single-cup swirl-stabilized spray combustor at idle and cruise conditions.
Our primary conclusions are summarized in the following points:

● All fuels exhibit qualitatively similar spray and flame structures at idle conditions. At
cruise conditions, Jet A produces a more lifted flame than C1 and C5. We attribute this
to the liquid thermodynamic properties of each fuel: C5 is most easily atomized and va-
porized, followed by C1, then Jet A. The rapid ignition of vaporized C1 may also reduce
flame lift at cruise conditions, resulting in similar flames for these fuels despite a shorter
spray penetration depth for C5.

13
1336 Flow, Turbulence and Combustion (2025) 114:1315–1339

● Species concentration profiles illustrate the significant chemical differences between

the fuels, particularly concerning initial decomposition products. C5 produces the most
ethylene, closely followed by Jet A, whereas C1 produces much less. Ethylene is con-
sidered a soot precursor, suggesting a sooting trend that matches the relative aromatics
content of the fuels. C1 primarily decomposes into iso-butene, leading to a different
chemical pathway and more OH in the postflame zone.
● At idle conditions, C5 produces slightly less NOx than the other fuels due to its low
equivalence ratio and heat of combustion. At cruise conditions, Jet A produces signifi-
cantly less NOx due to its large spray penetration depth and flame lift, which cause the
flame to burn at lower mixture fractions.
● Longitudinal and azimuthal pressure oscillations are observed in all simulations. These
are similar for all fuels at idle conditions, but at cruise conditions, the more compactly
burning C1 and C5 exhibit much stronger longitudinal oscillations than Jet A. We attrib-
ute this to the higher volatility of these fuels, but Cetane number may be an important
factor as well.

Acknowledgements The authors acknowledge Martin Passad and Elna Heimdal Nilsson for their insight on
the topic of chemical kinetics.

Author Contributions Conceptualization, A.Å. and C.F.; methodology, A.Å. and C.F.; investigation, A.Å.;
resources, C.F.; writing—original draft preparation, A.Å.; writing—review and editing, A.Å. and C.F.; visu-
alization, A.Å.; project administration, C.F.; funding acquisition, C.F.

Funding Open access funding provided by Lund University.

This work was supported by the European Union’s Horizon 2020 research and innovation program
MORE&LESS under grant agreement no. 769246, and by the competence center CESTAP funded by the
Swedish Energy Agency under grant agreement no. 52683-1. Computational resources were provided by the
National Academic Infrastructure for Supercomputing in Sweden (NAISS), partially funded by the Swedish
Research Council through grant agreement no. 2022-06725.

Data Availability The simulation results are available from the authors on reasonable request.

Declarations

Competing Interests The authors declare no competing interests.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as
you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons
licence, and indicate if changes were made. The images or other third party material in this article are
included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material.
If material is not included in the article’s Creative Commons licence and your intended use is not permitted
by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

References
Agostinelli, P.W., Laera, D., Boxx, I., et al.: Impact of wall heat transfer in large eddy simulation of flame
dynamics in a swirled combustion chamber. Combust. Flame. 234, 111728 (2021). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​
0​​1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​2​1​.​1​1​1​7​2​8

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1337

Åkerblom, A., Fureby, C.: LES modeling of the DLR generic single-cup spray combustor: validation and the
impact of combustion chemistry. Flow Turbul. Combust. 112. 557–585 (2024). ​h​t​t​p​s​:​/​/​d​o​i​.​o​r​g​/​1​0​.​1​0​0​7​
/​s​1​0​4​9​4​-​0​2​3​-​0​0​5​1​2​-​4​​​​
Åkerblom, A., Pignatelli, F., Fureby, C.: Numerical simulations of spray combustion in jet engines. Aero-
space. 9. 838 (2022). https://doi.org/10.3390/aerospace9120838
Åkerblom, A., Passad, M., Ercole, A., et al.: Numerical modeling of chemical kinetics, spray dynamics, and
turbulent combustion towards sustainable aviation. Aerospace. 11, 31 (2024). ​h​t​t​p​s​:​/​/​d​o​i​.​o​r​g​/​1​0​.​3​3​9​0​/​
a​e​r​o​s​p​a​c​e​1​1​0​1​0​0​3​1​​​​
Alsulami, R., Lucas, S., Windell, B., et al.: Experimental assessment of the impact of variation in jet fuel
properties on spray flame liftoff height. Appl. Energy Combust. Sci. 7, 100032 (2021). ​h​t​t​p​s​:​/​/​d​o​i​.​o​r​g​/​
1​0​.​1​0​1​6​/​j​.​j​a​e​c​s​.​2​0​2​1​.​1​0​0​0​3​2​​​​
Ansys: Ansys Chemkin-Pro, Ver. 2021 R2. (2021)
ASTM International: D1655-24 standard specification for aviation turbine fuels. (2024)
Bae, J., Chaudhuri, S., Canteenwalla, P., et al.: Combustion characteristics of sustainable aviation fuels in a
scaled-down afterburner test rig. Aeronaut. J. 128, 1429–1449 (2024). ​h​t​t​p​s​:​/​/​d​o​i​.​o​r​g​/​1​0​.​1​0​1​7​/​a​e​r​.​2​0​2​
4​.​4​9​​​​
Behrendt, T., Frodermann, M., Hassa, C., et al.: Optical measurements of spray combustion in a single sector
combustor from a practical fuel injector at higher pressures. In: Symposium on Gas Turbine Engine
Combustion, Emissions and Alternative Fuels, Lisboa, Portugal (1999)
Bressloff, N.W.: A parallel pressure implicit splitting of operators algorithm applied to flows at all speeds. Int.
J. Numer. Methods Fluids. 36, 497–518 (2001). https://doi.org/10.1002/fld.140
Colket, M., Heyne, J., Rumizen, M., et al.: Overview of the national jet fuels combustion program. AIAA J.
55, 1087–1104 (2017). https://doi.org/10.2514/1.J055361
Crowe, C.: Multiphase Flows with Droplets and Particles. CRC Press LLC, Boca Raton, FL, USA (1998)
Dukowicz, J.K.: A particle-fluid numerical model for liquid sprays. J. Comp. Phys. 35, 229–253 (1980).
https://doi.org/10.1016/0021-9991(80)90087-X
Edwards, T.: Reference jet fuels for combustion testing. In: Proceedings of the 55th AIAA Aerospace Sci-
ences Meeting. (2017). AIAA2017-0146. https://doi.org/10.2514/6.2017-0146
El Helou, I., Foale, J.M., Pathania, R.S., et al.: A comparison between fossil and synthetic kerosene flames
from the perspective of soot emissions in a swirl spray RQL burner. Fuel. 331, 125608 (2023). ​h​t​t​p​s​:​/​/​
d​o​i​.​o​r​g​/​1​0​.​1​0​1​6​/​j​.​f​u​e​l​.​2​0​2​2​.​1​2​5​6​0​8​​​​
Esclapez, L., Ma, P.C., Mayhew, E., et al.: Fuel effects on lean blow-out in a realistic gas turbine combustor.
Combust. Flame. 181, 82–99 (2017). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​1​7​.​0​2​.​0​3​5
Faeth, G.M.: Mixing, transport and combustion in sprays. Prog. Energy Combust. Sci. 13, 293–345 (1987).
https://doi.org/10.1016/0360-1285(87)90002-5
Glarborg, P., Miller, J.A., Ruscic, B., et al.: Modeling nitrogen chemistry in combustion. Prog. Energy Com-
bust. Sci. 67, 31–68 (2018). https://doi.org/10.1016/j.pecs.2018.01.002
Hairer, E., Wanner, G.: Solving ordinary differential equations, 1st edn. In: Chap II: Stiff and Differential-
Algebraic Problems. Springer, Berlin/Heidelberg, Germany (1991)
Hasti, V.R., Kundu, P., Som, S., et al.: Computation of conventional and alternative jet fuel sensitivity to lean
blowout. J. Energy Inst. 101, 19–31 (2022). https://doi.org/10.1016/j.joei.2021.12.006
HyChem: Performance. ​h​t​t​p​s​:​​/​/​w​e​b​​.​s​t​a​n​f​​o​r​d​.​​e​d​u​/​g​​r​o​u​p​/​​h​a​i​w​a​n​​g​l​a​b​​/​H​y​C​h​​e​m​/​p​a​​g​e​s​/​p​e​​r​f​o​r​​m​a​n​c​e​.​h​t​m​l
(2020). 26 November 2024
ICAO: ICAO aircraft engine emissions databank. ​h​t​t​p​s​:​​/​/​w​w​w​​.​e​a​s​a​.​​e​u​r​o​​p​a​.​e​u​​/​e​n​/​d​​o​m​a​i​n​s​​/​e​n​v​​i​r​o​n​m​​e​n​t​/​i​​c​a​
o​-​a​i​​r​c​r​a​​f​t​-​e​n​​g​i​n​e​-​​e​m​i​s​s​i​​o​n​s​-​​d​a​t​a​b​a​n​k (2023a). 29 May 2024
ICAO: SAF volumes growing but still missing opportunities. ​h​t​t​p​s​:​​/​/​w​w​w​​.​i​a​t​a​.​​o​r​g​/​​e​n​/​p​r​​e​s​s​r​o​​o​m​/​2​0​2​​3​-​r​e​​l​e​a​
s​e​s​/​2​0​2​3​-​1​2​-​0​6​-​0​2 (2023b). 7 Aug 2024
ICAO: Conversion processes. ​h​t​t​p​s​:​​/​/​w​w​w​​.​i​c​a​o​.​​i​n​t​/​​e​n​v​i​r​​o​n​m​e​n​​t​a​l​-​p​r​​o​t​e​c​​t​i​o​n​/​​G​F​A​A​F​​/​P​a​g​e​s​​/​C​o​n​​v​e​r​s​i​o​n​-​p​r​
o​c​e​s​s​e​s​.​a​s​p​x (2024). 7 Aug 2024
Jones, W.P., Marquis, A.J., Vogiatzaki, K.: Large-eddy simulation of spray combustion in a gas turbine com-
bustor. Combust. Flame. 161, 222–239 (2014). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​1​3​.​0​7​.​0​1​6
Kim, W.W., Menon, S.: A new dynamic one equation subgrid-scale model for large eddy simulations. In:
Proceedings of the 33rd Aerospace Sciences Meeting and Exhibit. (1995). AIAA 1995-356. ​h​t​t​p​s​:​/​/​d​o​i​
.​o​r​g​/​1​0​.​2​5​1​4​/​6​.​1​9​9​5​-​3​5​6​​​​
Kumar, K., Sung, C.J., Hui, X.: Laminar flame speeds and extinction limits of conventional and alternative
jet fuels. Fuel. 90, 1004–1011 (2011). https://doi.org/10.1016/j.fuel.2010.11.022
Lefebvre, A.H.: Airblast atomization. Prog. Energy Combust. Sci. 6, 233–261 (1980). ​h​t​t​p​s​:​/​/​d​o​i​.​o​r​g​/​1​0​.​1​0​1​
6​/​0​3​6​0​-​1​2​8​5​(​8​0​)​9​0​0​1​7​-​9​​​​
Lo Schiavo, E., Laera, D., Riber, E., et al.: Effects of liquid fuel/wall interaction on thermoacoustic instabili-
ties in swirling spray flames. Combust. Flame. 219, 86–101 (2020). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​c​o​m​b​u​​s​t​
f​l​a​m​​e​.​2​0​​2​0​.​0​4​.​0​1​5

13
1338 Flow, Turbulence and Combustion (2025) 114:1315–1339

Mazzei, L., Puggelli, S., Bertini, D., et al.: Numerical and experimental investigation on an effusion-cooled
lean burn aeronautical combustor: aerothermal field and emissions. J. Eng. Gas Turbines Power. 141,
041006 (2019). https://doi.org/10.1115/1.4041676
Meier, U., Heinze, J., Freitag, S., et al.: Spray and flame structure of a generic injector at aero-engine condi-
tions. J. Eng. Gas Turbines Power. 134, 031503 (2012). https://doi.org/10.1115/1.4004262
Menon, S., Fureby, C.: Computational Combustion. In: Blockley, R., Shyy, W. (eds.) Encyclopedia of Aero-
space Engineering. John Wiley & Sons, Hoboken, NJ, USA (2010). https://doi.org/10.2514/4.754405
O’Rourke, P.J.: Collective drop effects on vaporizing liquid sprays. PhD thesis, Los Alamos National Labo-
ratories, Los Alamos, NM, USA (1981)
Panchal, A., Menon, S.: Large eddy simulation of fuel sensitivity in a realistic spray combustor I. Near
blowout analysis. Combust. Flame. 240, 112162 (2022a). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​2​
2​.​1​1​2​1​6​2
Panchal, A., Menon, S.: Large eddy simulation of fuel sensitivity in a realistic spray combustor II. Lean
blowout analysis. Combust. Flame. 240, 112161 (2022b). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​2​
2​.​1​1​2​1​6​1
Paxton, B.T., Fugger, C.A., Tomlin, A.S., et al.: (2020) Experimental investigation of fuel chemistry on
combustion instabilities in a premixed bluff-body combustor. In: Proceedings of the AIAA Scitech 2020
Forum. https://doi.org/10.2514/6.2020-0174, AIAA 2020-0174
Puggelli, S., Bertini, D., Mazzei, L., et al.: Assessment of scale-resolved computational fluid dynamics meth-
ods for the investigation of lean burn spray flames. J. Eng. Gas Turbines Power. 139, 021501 (2016).
https://doi.org/10.1115/1.4034194
Puggelli, S., Bertini, D., Mazzei, L., et al.: Modeling strategies for large eddy simulation of lean burn spray
flames. J. Eng. Gas Turbines Power. 140, 051501 (2018a). https://doi.org/10.1115/1.4038127
Puggelli, S., Paccati, S., Bertini, D., et al.: Multi-coupled numerical simulations of the DLR generic single
sector combustor. Combust. Sci. Technol. 190, 1409–1425 (2018b). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​8​0​/​0​0​​1​0​2​2​0​​2​
.​2​0​1​8​​.​1​4​5​​2​1​2​4
Péquin, A., Iavarone, S., Malpica Galassi, R., et al.: The partially stirred reactor model for combustion clo-
sure in large eddy simulations: physical principles, sub-models for the cell reacting fraction, and open
challenges. Phys. Fluids. 34, 055122 (2022). https://doi.org/10.1063/5.0090970
Rajendram Soundararajan, P., Vignat, G., Durox, D., et al.: Effect of different fuels on combustion instabili-
ties in an annular combustor. J. Eng. Gas Turbines Power. 143, 031007 (2021). ​h​t​t​p​s​:​/​/​d​o​i​.​o​r​g​/​1​0​.​1​1​1​
5​/​1​.​4​0​4​9​7​0​2​​​​
Ranz, W.E., Marshall, W.R.: Evaporation from drops. Chem. Eng. Prog. 48, 141–146 (1952)
Ranzi, E., Frassoldati, A., Stagni, A., et al.: Reduced kinetic schemes of complex reaction systems: fossil and
biomass-derived transportation fuels. Int. J. Chem. Kinet. 46, 512–542 (2014). ​h​t​t​p​s​:​/​/​d​o​i​.​o​r​g​/​1​0​.​1​0​0​2​
/​k​i​n​.​2​0​8​6​7​​​​
Rayleigh, J.W.S.: The explanation of certain acoustical phenomena. Nature. 18, 319–321 (1878). ​h​t​t​p​s​:​/​/​d​o​i​
.​o​r​g​/​1​0​.​1​0​3​8​/​0​1​8​3​1​9​a​0​​​​
Reitz, R.D., Diwakar, R.: Structure of high-pressure fuel sprays. SAE Technical Paper 870598. (1987).
https://doi.org/10.4271/870598
Sabelnikov, V., Fureby, C.: LES combustion modeling for high Re flames using a multi-phase analogy. Com-
bust. Flame. 160. 83–96 (2013). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​1​2​.​0​9​.​0​0​8
Saggese, C., Wan, K., Xu, R., et al.: A physics-based approach to modeling real-fuel combustion chemistry
– V. NOx formation from a typical Jet A. Combust. Flame. 212, 270–278 (2020). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​
1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​1​9​.​1​0​.​0​3​8
Spalding, D.B.: A single formula for the “Law of the wall”. J. Appl. Mech. 28. 455–458 (1961). ​h​t​t​p​s​:​/​/​d​o​i​.​
o​r​g​/​1​0​.​1​1​1​5​/​1​.​3​6​4​1​7​2​8​​​​
Stouffer, S., Hendershott, T., Boehm, R., et al.: Referee Rig. In: Colket, M., Heyne, J. (eds.) Fuel Effects on
Operability of Aircraft Gas Turbine Combustors. American Institute of Aeronautics and Astronautics,
Inc, Reston, VA, USA (2021). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​2​5​​1​4​/​5​.​​9​7​8​1​6​​2​4​1​0​6​0​​4​0​.​0​​1​1​5​.​0​1​4​2
Undavalli, V., Bilikis, O., Gbadamosi-Olatunde, O.B., et al.: Recent advancements in sustainable aviation
fuels. Prog. Aerosp. Sci. 136, 100876 (2023). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​p​a​e​r​o​​s​c​i​.​2​0​​2​2​.​1​​0​0​8​7​6
Wang, H., Xu, R., Wang, K., et al.: A physics-based approach to modeling real-fuel combustion chemistry - I.
Evidence from experiments, and thermodynamic, chemical kinetic and statistical considerations. Com-
bust. Flame. 193, 502–519 (2018a). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​1​8​.​0​3​.​0​1​9
Wang, K., Xu, R., Parise, T., et al.: A physics-based approach to modeling real-fuel combustion chemistry
– IV. HyChem modeling of combustion kinetics of a bio-derived jet fuel and its blends with a conven-
tional Jet A. Combust. Flame. 198, 477–489 (2018b). ​h​t​t​p​s​:​​/​/​d​o​i​​.​o​r​g​/​1​​0​.​1​0​​1​6​/​j​.​​c​o​m​b​u​​s​t​f​l​a​m​​e​.​2​0​​1​8​.​0​7​
.​0​1​2
Weller, H.G., Tabor, G., Jasak, H., et al.: A tensorial approach to computational continuum mechanics using
object-oriented techniques. Comp. Phys. 12, 620–631 (1998). https://doi.org/10.1063/1.168744

13
Flow, Turbulence and Combustion (2025) 114:1315–1339 1339

Xu, R., Wang, H., Colket, M., et al.: Thermochemical properties of jet fuels. ​h​t​t​p​s​:​​/​/​w​e​b​​.​s​t​a​n​f​​o​r​d​.​​e​d​u​/​g​​r​o​u​
p​/​​h​a​i​w​a​n​​g​l​a​b​​/​H​y​C​h​​e​m​/​a​p​​p​r​o​a​c​h​​/​R​e​p​​o​r​t​_​J​​e​t​_​F​u​​e​l​_​T​h​e​​r​m​o​c​​h​e​m​i​c​a​l​_​P​r​o​p​e​r​t​i​e​s​_​v​6​.​p​d​f (2015). 7 Aug
2024
Xu, R., Wang, K., Banerjee, S., et al.: A physics-based approach to modeling real-fuel combustion chemistry
– II. Reaction kinetic models of jet and rocket fuels. Combust. Flame. 193, 520–537 (2018). ​h​t​t​p​s​:​​​/​​/​d​o​​i​
.​o​r​​g​/​​1​0​.​​1​0​​1​​6​​/​j​.​c​o​m​​b​u​s​t​f​​l​​a​m​e​​.​​2​0​1​​8​.​0​3​.​0​2​1
Xue, X., Hui, X., Singh, P., et al.: Soot formation in non-premixed counterflow flames of conventional and
alternative jet fuels. Fuel. 210, 343–351 (2017). https://doi.org/10.1016/j.fuel.2017.08.079
Yang, J., Xin, Z., He, Q., et al.: An overview on performance characteristics of bio-jet fuels. Fuel. 237,
916–936 (2019). https://doi.org/10.1016/j.fuel.2018.10.079
Yoshizawa, A., Horiuti, K.: A statistically-derived subgrid-scale kinetic energy model for the large eddy simu-
lation of turbulent flows. J. Phys. Soc. Jpn. 54, 2834–2839 (1985). https://doi.org/10.1143/JPSJ.54.2834
Zheng, L., Boylu, R., Cronly, J., et al.: Experimental study on the impact of alternative jet fuel properties and
derived cetane number on lean blowout limit. Aeronaut. J. 126, 1997–2016 (2022). ​h​t​t​p​s​:​/​/​d​o​i​.​o​r​g​/​1​0​.​
1​0​1​7​/​a​e​r​.​2​0​2​2​.​3​3​​​​

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

13