CONSEIL INTERNATIONAL
DES MACHINES A COMBUSTION
INTERNATIONAL COUNCIL
ON COMBUSTION ENGINES
PAPER NO.: 39
Modelling of the oxidation of fuel sulfur in low
speed two-stroke Diesel engines
Anders Andreasen, MAN Diesel, Denmark
Stefan Mayer, MAN Diesel, Denmark
Abstract: In large marine two stroke Diesel engines during combustion of sulfur containing fuel, the
sulfur is oxidised to SO2 , mainly, although substantial
amounts of SO3 and H2 SO4 will form as well. These
latter species may cause corrosional wear of the cylinder liner if not neutralised by lube oil additives. Potential attacks is due to either condensation of sulfuric
acid on the cylinder liner lube oil film or direct dissolution of oxidised sulfur species in the lube oil film in
which reaction with dissolved water may be the source
of acidic species. In order to evaluate and predict corrosional wear of the liner material, it is pivotal to have
realistic estimates of the distribution/concentration of
oxidised sulfur species as well as a reliable model of
c
CIMAC
Congress 2010, Bergen
formation, transport and destruction of acidic species
in the oil film. This paper addresses the former part
by invoking a detailed reaction mechanism in order to
simulate the oxidation of fuel bound sulfur and predicting the concentration of SO2 as well as the conversion
fraction into SO3 and H2 SO4 . The reaction mechanism
is coupled to a realistic model of the combustion process in which the air entrainment into the combustion
zone is accounted for. The results of the simulation
are evaluated with respect to previously applied models as well as existing data on the conversion fraction
of SO2 to SO3 and H2 SO4 . The conversion fraction is
found to be in a range of 2.6-6.7 %.
INTRODUCTION
During the combustion of sulfur containing fuel, the
sulfur will be oxidised mainly to SO2 but significant
fractions of SO3 and H2 SO4 will also form [1]. While
emissions of CO, HC and NOx to some extent can
be controlled through the combustion conditions
e.g. air excess ratio, combustion temperature etc.
the total emission of sulfur can only be controlled
through the sulfur content of the fuel [2] or by
appropriate exhaust gas after treatment e.g. by
scrubbing technologies [3]. From an environmental
point of view SOx emissions are undesired,
primarily due to the formation of atmospheric
H2SO4, which is the main source of acid rain.
From a technical/engineering point of view the
problems related to the sulfur content of the fuel are
mainly corrosional wear due to sulfuric acid
(produced from fuel sulfur, oxygen and moisture)
attacking various parts of the engine and auxiliaries
[4, 5, 6]. Especially the corrosional wear of the
cylinder liner has received special attention in the
past [7, 8, 9, 10, 11, 12, 13]. Sulfuric acid may
potentially cause corrosion either by condensation,
if the liner temperature drops below the dew point
of the combustion gas [7, 12] or by direct transport
of acidic gas phase species into the cylinder liner
lube oil film [13]. The acidic species may diffuse
through the cylinder lube oil and reach the metal
surface eventually leading to corrosional wear. This
can be avoided through neutralisation of sulfuric
acid with alkaline lube oil additives. Further it has
been proposed that SO2 is absorbed in the lube oil
film in which further reaction with water leads to
acid formation [11].
Standard textbooks on the topic of internal
combustion engines and combustion in general
only surface the topic of sulfur oxide formation
during combustion of sulfur containing fuels
[2, 14, 15]. To the best of our knowledge no results
on experimental measurements of SOx in the
exhaust gas from large two-stroke diesel engines
have been published in peer reviewed journals in
the past. Engel et al. reported measurements of
both SO2 and SO3 in the exhaust gas of a range of
heavy duty four stroke diesel engines [4]. They
found that approx. 3 % (volumetric) of the SO2 was
converted to SO3 on average. Previous modelling
studies devoted to the prediction of diesel engine
in-cylinder formation of sulfur oxides including
sulfuric acid [7] has been based either on the
assumption that i) SO2, SO3, and H2SO4 are in
equilibrium at temperatures above 1000∘C, viz. the
reactions responsible for their formation are much
faster than the time scales being considered. Below
© CIMAC Congress 2010, Bergen
1000∘C the reaction systems are considered frozen
i.e. the amount of the species already formed will
be constant, or ii) a fixed instantaneous conversion
ratio of SO2 to SO3 of e.g. 5% [12, 13]. In order to
provide a deeper understanding of corrosional wear
of the cylinder liner through modelling studies it is
pivotal to have reasonable concentrations of both
SO2, SO3 and H2SO4 as input.
In contrast to diesel engines the published
research on formation of SO x and H2SO4 from
aircraft turbines has been significantly more
extensive both experimentally as well as
theoretically [16, 17, 18, 19, 20] (and references
therein). The modelling studies reported are based
on large detailed reaction mechanisms describing
the oxidation of fuel bound sulfur formation
[18, 17, 19].
In this preliminary study it will be investigated
how both chemical kinetics and thermodynamics
influence the distribution of oxidised sulfur species
during combustion of sulfur containing fuel in large
two-stroke diesel engines. The main motivation for
this study is the fact that the current knowledge
about the in-cylinder formation of SO3 and H2 SO4
in large two-stroke diesel engines is very limited.
Improving the understanding in this field will indeed
be helpful in evaluating potential in-cylinder
corrosional wear both for engines in service,
especially considering choice of lube oil type and
lube oil consumption, but certainly also for engines
in the design phase. Furthermore, exhaust gas
recirculation (EGR) is becoming an emerging
method for NOx reduction for large two-stroke
marine diesel engines running on heavy fuel oil
(sulfur content up to 4.5 w.w %) [21, 22]. A detailed
knowledge of the composition of sulfur species in
the EGR line will be helpful in the design and
control of the EGR scrubber in order to control the
level of sulfur species entering the cylinder through
the scavenge ports potentially increasing the risk of
cylinder liner corrosional wear if not neutralised.
The basis of this study is real measured engine
data from a large two-stroke diesel engine e.g.
directly measured in-cylinder pressure as a function
of degree crank angle as well as exhaust gas NO
concentraion, derived fuel mass burned fraction
and trapped air amount (air excess ratio). This is
used as input for simulations invoking a detailed
reaction mechanism describing the formation of
oxidised sulfur species from fuel injection start to
exhaust valve opening.
Paper No. 39
2
METHODS
Experimental
The engine used to acquire input data to the kinetic
model is the MAN Diesel custom built test engine
located in Copenhagen. A summary of engine
characteristics is given in Table 1. The engine is
equipped for full electronic-hydraulic control of the
fuel injection and the exhaust valve timing. A total
of four engine tests have been conducted at the
following loads: 100, 75, 50, and 25% (at 123, 112,
98 and 78 rpm, respectively) of the maximum
continuous rating (MCR). The engine speed
decrease as a function of load according to a
simulated propeller curve [23]. The air excess ratio
(trapped air only, scavenging air not included) for
the four tests was 2.27, 2.43, 2.36, 2.74 for 100, 75,
50, and 25% load, respectively. The used fuel was
diesel oil with the following atomic composition of C
and H atoms: 86.8 w/w % and 13.2 w/w %. The
sulfur content of the fuel was 0.05 w/w % and
heating value of the fuel is 42820 kJ/kg. All values
determined from an analysis performed by an
external laboratory. The exhaust gas concentration
of NO has been measured in the exhaust duct
between the turbine and the stack by the heated
chemiluminescence technique (HCLD) according to
ISO 8178 standards.
Figure 1 — Measured in-cylinder pressure
averaged over 50 revolutions as a function of
degree crank angle after top dead center (ATDC).
Table 1 Test engine specifications
Manufacturer
Type
Number of cylinders
Bore
Stroke
Connecting rod length
MCR speed
MCR power
MEP @ MCR
Turbocharger
MAN B&W Diesel
4T50ME-X
4
500 mm
2200 mm
2885 mm
123 RPM
7050 kW
20 bar
MAN TCA55-VTA
Figure 2 — Rate of heat release for the 4 load
cases
Model and calculational setup
The starting point in the kinetic model is measured
cylinder pressure as a function of crank angle.
From the measured in-cylinder pressure the fuel
mass burned fraction is derived from the integrated
heat release rate as obtained from a single zone
heat release analysis using an in-house code.
The measured cylinder pressure as a function of
degree crank angle for the four engine tests is
depicted in Figure 1. The rate of heat release as
determined from the in-cylinder measured pressure
traces shown in Figure 1 is shown in Figure 2.
© CIMAC Congress 2010, Bergen
In order to model the kinetics of the formation of
oxidised sulfur species i.e. SO2, SO3, and H2SO4, a
detailed reaction mechanism is called upon. For
this purpose the Glarborg sulfur mechanism is used
[24] which is a result of a continuous effort in
literature data review, experimental laboratory
reactor measurements as well as theoretical
predictions
[25, 26, 27, 28, 29].
Other
sulfur
oxidation mechanisms exists e.g. the Leeds sulfur
mechanism [30], however a substantial part of the
S subset is based on the work of Glarborg et al.
anyway. The S subset in the Glarborg mechanism
contains 96 elementary reactions. In addition to the
S subset the mechanism contains an O/H subset
describing the formation of O, OH, HO2, H2O2 . A
complete listing of all reactions along with rate
coefficient and thermodynamic parameters will not
be given. In order to describe the formation of
Paper No. 39
3
sulfuric acid a single reaction has been added in
which H2SO4 is formed from SO3 and H2O. The
kinetic parameters for this reaction have been taken
from [31].
In order to integrate the kinetic rate equations the
Cantera open source software [32, 33] is used. The
used mechanism include a thermodynamical
description of all involved species expressed
through 7 coefficient NASA polynomials [34]. In
order to simulate the oxidation of fuel bound sulfur
a multi-zone model is invoked and the details of the
model will summarised in following.
Up to the time of fuel ignition the fresh charge of the
cylinder is considered as a single zone and the
working gas is compressed isentropically. By the
time of ignition and until the combustion has ended,
for each crank angle degree a new zone is created
by mixing the differential amount of fuel burned
during this particular time step (one crank angle
degree) stoichiometrically with fresh air and
subsequently equilibrating [35] at the corresponding
measured cylinder pressure. We will refer to such a
unit as a parcel. After its creation and subsequent
equilibration no more fuel is added to this particular
parcel. Only fresh air is allowed to be added at a
predefined mixing rate. The multi-zone approach
taken here is in close analogy with that in [36]. In
each parcel the rate equations describing O/H
chemistry and sulfur oxidation are integrated up to
the time of the exhaust valve opening.
in which Tad is the adiabatic flame temperature
calculated by equilibrating fuel and air at λ=1, cp is
the constant pressure heat capacity of the
combustion products and Crad is a heat radiation
constant with a value chosen to give a temperature
drop of approx 100 K for the parcels created early.
Throughout this paper we will refer to a so-called
conversion fraction or conversion efficiency of sulfur
into SO3 and H2SO4 which is denoted ε. The
definition of ε is
(2)
where ∑ [SOx] is the sum of the concentrations of
all sulfur containing species i.e. (SO2, SO3, H2SO4 ,
...).
RESULTS AND DISCUSSION
Examination of the equilibrium
An overall, yet simplified, description of the
oxidation of fuel bound sulfur is outlined in the
following reactions.
Some further assumptions of the model are:
Sulfur contained in the fuel is modeled as
elemental S in the fuel feed.
The fuel is mimicked by a mixture of noctane, benzene and ethanol. The exact
composition is tuned in order to match the
heating value of the real fuel as well as the
elemental composition.
The concentration of all gas phase species
is isotropic i.e. mixing is fast and complete.
For each time step it is assumed that
temperature and pressure is constant at the
starting value for that particular time step.
In order for the flame temperatures calculated not
being unrealistically high a temperature drop due to
radiation losses is included [36]
(1)
© CIMAC Congress 2010, Bergen
S + ½O2 = SO
(3)
SO + ½O2 = SO2
(4)
SO2 + ½O2 = SO3
(5)
SO3 + H2O = H2SO4
(6)
The enthalpy of formation, ΔH298o, of the individual
species are 6 kJ/mol (gas), -297 kJ/mol (gas), -396
kJ/mol (gas), and -814 kJ/mol (liquid) for SO, SO2,
SO3, and H2SO4, respectively [37].
To which degree the sulfur is oxidised or converted
into H2 SO4 depends on both thermodynamics and
reaction kinetics. The above reactions are listed in
the order of increasing thermodynamic stability of
the product. According to the principle of Le
Chatelier this implies that the higher the
temperature the less degree of oxidation i.e. SO is
favored vice versa at lower temperatures
SO3/H2SO4 is favored. Also pressure is important;
the higher the pressure, the higher degree of
oxidation.
To get a comprehension of the quantitative
influence of temperature and pressure on the
distribution of sulfur containing species the
O2/SO2/SO3/H2SO4 equilibrium composition is
Paper No. 39
4
calculated. The results of the equilibrium
calculations are shown in Figure 3. As seen from
the figure the main transition from SO3 to SO2 takes
place at around 1100 K at 10 bar and at approx.
1200 K at a pressure of 100 bar. Thus, in the early
phase of the combustion all sulfur will be in the form
of SO2. No significant amounts of SO were found
unless above 2200 K. At the lowest temperatures
sulfuric acid is favored. At a temperature of 1273 K
the fraction of SO3 relative to SO2 +SO3 equals 14
and 35 % at 10 and 100 bar, respectively. Upon
cooling from high temperatures, the amount of SO 3
and H2SO4 will increase from a thermodynamic
point of view. However, due to potential kinetic
limitations the SO3 and H2SO4 concentration may
be frozen by quenching the gas. The exact
concentration of sulfur containing species is thus a
complex function of temperature, pressure, initial
concentrations, cooling rate, and residence time.
The effect of kinetics on the formation of
SO3/H2SO4 is investigated in the next section.
Both limits are unrealistic for modelling real mixing
controlled diesel spray combustion and emission
formation. The former limit will result in too high
parcel temperatures and too low oxygen
concentration and for the latter limit the opposite is
the case. In reality the mixing rate is somewhere in
between the two. Thus the temperature (cooling by
fresh gas) and oxygen concentration in each parcel
is closely linked to the applied mixing rate.
In order to investigate the effect of applied mixing
rate on the calculated conversion fraction, ε, a
number of simulations have been performed in
which the mixing rate is varied. The results are
depicted in Figure 4.
Figure 4 — Conversion fraction, ε, resulting air
excess ratio in the burned zones at exhaust valve
opening, and calculated NO concentration as a
function of applied mixing rate. Results are
calculated for the 75 % load case
Figure 3 — Concentration of SO2, SO3, and H2SO4
as a function of temperature at (upper) a pressure
of 10 bar and at (lower) a pressure of 100 bar.
Initial concentration is xO2 = 0.15, xH2O = 0.05, xS =
0.01, no other components other than Argon
present.
Simulations of in cylinder formation of sulfur
species
Adjustment of mixing rate. The main tuneable
parameter in the model applied in this study is the
mixing rate at which fresh air is mixed into each
parcel. The limits of the mixing rate is zero, at the
low end, corresponding to fuel being burned at
stoichiometric conditions i.e. λ = 1 and cooling is
only due to the expansion of cylinder gas. The
higher limit corresponds to an infinite mixing rate i.e
air is mixed into each parcel increasing the local
lambda from 1 towards the global lambda very fast.
© CIMAC Congress 2010, Bergen
In addition to the calculated ε and the resulting
average air excess ratio in the burned zones, the
results of a NO model have been included as well.
The NO model used is the extended Zeldovich
mechanism [2, 14]
N2 + O = NO + N
(7)
N + O2 = NO + O
(8)
N + OH = NO + H
(9)
The kinetic parameters are taken from [14]. Since
NO has been measured for the tests investigated in
the present study comparison between measured
and calculated NO concentrations allows a tuning
of the mixing ratio in order to give satisfactory
agreement.
As seen from the figure the conversion fraction, ε,
increases due to increased mixing. The resulting
average air excess ratio in the burned zones
increase up to the global air excess ratio at a
mixing rate of 10. Both the lowest and highest
Paper No. 39
5
mixing rates applied may seem unrealistic,
nevertheless they provide a rough range for ε of
approx. 1-12%. The measured NO concentration is
1266 vol. ppm. In order for the calculated NO
concentration to equal this, a mixing rate of 3.29 is
required, as found by interpolation. The
corresponding ε is 4.43 %. Using the same
approach in order to match the experimental NO
concentration for the remaining load cases of 1119,
1236, and 1320 vol ppm for 100, 50, and 25% load,
respectively, the corresponding values of ε are
2.59, 4.25, and 6.72 %. Thus, for a wide range of
engine loads ε apparently lies in a relatively narrow
range of 2.5-6.7%. From these results as well as
from Figure 4 it is interesting to note the similar
effect of varying mixing rate on both NO and ε.
Engel et al. measured the conversion of SO2 to SO3
on a number of heavy duty diesel engines operated
at different load, different fuel quality, different
sulfur content etc. In their study they found a
conversion fraction ranging from 2–8%. Clearly our
results agree very well with these findings.
In the following the 75% load case with a mixing
rate of 3.29 will be investigated in more detail.
Crank angle resolved results Figure 5 shows
selected results from the 75% load case with an
applied mixing rate of 3.29. The calculated NO
concentration (B) is for the trapped cylinder gas
only. Diluting with additional air, used in the
scavenging process, results in a NO concentration
of 1266 ppm which is equal to the measured value
(after the exhaust gas turbine).
As seen from Figure 5 (A) the SO2 concentration
increase rapidly around TDC and the increase
follows the rate of combustion very closely. Both
SO3 and H2SO4 are at very low levels during the
combustion period. This is due to the fact that the
equilibrium conversion of SO2 (cf. Figure 5 (B)) is
low at the very high temperatures (cf. Figure 5 (C))
as well as the time required for a significant
conversion into SO3 and H2 SO4 has not elapsed.
The mean combustion temperature in the parcels in
Figure 5 is a result of both the expansion of the
cylinder content as well as the rate at which the
much colder fresh air is mixed into the parcels.
After the combustion has ended the temperature
has dropped sufficiently in order to favour a
significant conversion fraction of SO2 into SO3 and
H2SO4. As seen from Figure 5 (C) the reaction rates
controlling the oxidation of SO2 appear to be very
fast at least until up to 80 CAD ATDC. From this
point the temperature has dropped enough in order
to cause a slow-down in the rate of SO2 oxidation
and the result is a deviation from equilibrium. It
appears as though the oxidation of SO2 is not
frozen at the end of the simulation (corresponding
© CIMAC Congress 2010, Bergen
to the opening of the exhaust valve). However
during blow-down the hot combustion gases are
expanded rapidly causing a further temperature
drop, and once the scavenge ports are uncovered
cold intake is mixed with the combustion gases
causing a further cooling. Thus, it may seem
reasonable to assume that the further oxidation of
SO2 will be on a much smaller scale.
Comparison with equilibrium assumptions. In a
previous study, related to cylinder liner corrosive
wear due to sulfuric acid condensation, Teetz [7]
applied the assumption that reactions responsible
for SO2 oxidation are very fast i.e. equilibrated
o
above 1000 C, while below this temperature the
reactions become frozen i.e. no further reactions
take place. This assumption is valid according to [1]
and references within.
In the following we will apply the above assumption
from now on denoted the frozen equilibrium
assumption. The assumption is applied assuming a
single zone combustion i.e. combustion takes place
in the same zone containing the entire amount of
fresh air. This is in contrast to the previous multi1
zone approach . The result is shown in Figure 6 (A)
where a comparison with the equilibrium conversion
fraction as found in the burned zones using the
multi-zone approach is made. The temperature in
the single burned zone is depicted in Figure 6 (B).
The most important observation is the fact that the
mean temperature using the single-zone approach
is much lower than in the multi-zone approach,
coupled with the high pressure near TDC, this
results in a very high equilibrium conversion
fraction. Furthermore the resulting ε is approx. 18%
at the end of simulation, which is almost four times
larger than that found in Figure 5 with a mixing rate
of 3.3 and more than twice than assuming full
equilibrium in the parcels. Although we have found
evidence supporting the use of the frozen
2
equilibrium assumption for premixed combustion , it
is clearly insufficient for modelling emission
formation during the mixing controlled combustion
in diesel engines.
1
No parcels in the multi-zone approach are cooled below 1000o
C, thus the result would not differ from the equilibrium in Figure 5
2
Unpublished. A. Andreasen, S. Mayer
Paper No. 39
6
(A)
(B)
(C)
(D)
Figure 5 — Results from multi-zone approach for the 75% load case. (A) Calculated concentrations of SO 2,
SO3, and H2 SO4 (B) Calculated NO concentration (C) Conversion fraction of SO 2, equilibrium conversion
fraction of SO2, and oxygen concentration in the burned zones (D) Temperature of the fresh cylinder air and
the average temperature in the burned zones.
© CIMAC Congress 2010, Bergen
Paper No. 39
7
(A)
(B)
Figure 6 — (A) ε calculated with the frozen equilibrium assumption using a single zone approach compared
with the equilibrium ε from the multi-zone parcel model. (B) Temperature in the single zone model (dashed
line).
Effect of sulfur content In order to investigate
the effect of the fuel sulfur content on the resulting
conversion fraction, a number of simulations with
different sulfur content has been conducted. The
results from the 75% load case are depicted in
Figure 7. The simulations have been conducted
with all other parameters including the applied
mixing rate fixed for all sulfur contents.
As seen from Figure 7 the conversion fraction
decreases with increasing sulfur content. Taking
the relatively large range in sulfur content into
account the decrease in ε is modest.
Nevertheless, the result is in qualitative
agreement with the literature [1, 4].
Reactions controlling conversion fraction The
reaction mechanism applied in the present work
has provided valuable insight as already
discussed. Nevertheless the models investigated
are still coarse with many simplifying assumption.
In order to take the analysis to the next level it
would be desirable to use the applied reaction
mechanism coupled to a full CFD code. However
due to the high number of reaction steps this
approach will be expensive in terms of the time
required for simulating a single combustion cycle.
This cost can be significantly reduced if a reduced
mechanism, still capturing the most important
reaction paths, can be formulated. In fact it often
turns out that even large reaction mechanisms
contain a few steps which govern the overall rate
[38, 39]. It is not the purpose of the present paper
to propose a reduced mechanism. Nevertheless,
some hints and guidelines can be found in the
literature pin pointing the most important
reactions. These will be briefly summarized below.
Previous studies by Cerru et al. [40, 41] have
been devoted to the formulation of a reduced
mechanism for sulfur oxidation. Although a two
step mechanism is formulated for SO2 oxidation a
number of reaction rates (used in linear
combinations) are required.
Tremmel and Schumann [18] conducted a
sensitivity analysis study, and concluded that the
rate constant of the reaction
Figure 7 — Conversion fraction, ε, as a function of
fuel sulfur content. Results from simulations of the
75% load case
© CIMAC Congress 2010, Bergen
SO2 + OH(+M) = HOSO2(+M)
Paper No. 39
(10)
9
showed the highest impact on the conversion
fraction for conditions relevant for aviation turbine
conditions. At some conditions the rate constant
of the following reaction was found to also control
the conversion fraction
SO2 + O(+M) = SO3(+M)
and H2SO4 on a qualitative scale. Nevertheless, in
order to improve the predictive reliability of the
model on a more quantitative scale, a number of
recommended actions remain to be taken.
1. Real experimental evidence of the
conversion fraction in large two-stroke
diesel engines is severely lacking. In
order to improve the confidence in the
model the conversion fraction must be
measured and compared with model
predictions.
(11)
Reaction Eq. 11 is highlighted to be of major
importance by Glarborg [28], especially at high
temperatures. Reaction Eq. 10 in combination
with the reaction
HOSO2 + O2 = SO3 + HO2
2. In this study we have relied on the kinetic
parameters for conversion of SO3 to
H2SO4 as reported by Reiner and Arnold
[31]. Several other studies exists e.g.
[42, 43, 44, 45]
signalling
both
the
complexity of the reaction as well as its
importance in relation to acid rain.
Unfortunately the studies are limited to
low pressure and close to ambient
temperature.
Thus
reliable
kinetic
parameters at elevated temperature and
pressure are needed.
(12)
is proposed to be important mainly at lower
temperatures i.e. when the burned gases are
cooled. In another study Glarborg and co-workers
[24] also point out the importance of the reactions
SO2 + OH = SO3 + H
(13)
and
SO2 + HO2 = SO3 + OH
(14)
3. Until now only the time until exhaust valve
opening has been considered. It should
also be investigated further what happens
after the exhaust valve opens (decrease
in temperature and pressure), during the
scavenging process and in the exhaust
receiver. It should be expected that
significantly more SO3 is converted to
H2SO4 during these processes.
through a rate of production and sensitivity
analysis.
Since the importance of the above reactions has
been proven at conditions not exactly matching
those in a large two-stroke diesel engines a future
sensitivity study should be devoted to the
formulation of a reduced mechanism.
4. It is well known that vanadium oxide
catalyse the oxidation of SO2 to SO3
[46, 47]. Since the fuel used for large twostroke engines may contain significant
levels of vanadium, the catalysed
oxidation may contribute significantly to
the overall conversion fraction. Future
analysis should take this possibility into
account.
CONCLUSIONS
In this paper the in-cylinder formation of SO3 and
H2SO4 of a large two-stroke diesel engine has
been studied. A detailed kinetic mechanism has
been used in combination with measured incylinder pressure in order to calculate the
concentration of sulfur containing species using a
multi zone approach with the rate of mixing of
fresh gas into the burned parcels as the main
adjustable parameter. By combining the multizone model with a NO formation model
(Zeldovich) the mixing rate has been adjusted in
order for the calculated NO to match the
measured value. Applying the fitted mixing rates
values of ε lying in a range from 2.6 to 6.7% is
found. The results are comparable with previous
measurements of the SO2 conversion fraction in
heavy duty diesel engines [4].
The presented model gives valuable information
about the process of in-cylinder formation of SO3
© CIMAC Congress 2010, Bergen
NOMENCLATURE
cp
Specific heat capacity
pressure, J/(mol K)
at
constant
Crad
Heat radiation constant, J/(mol K )
4
Conversion fraction of S to SO3 and
H2SO4
ΔH298∘ Standard formation enthalpy at 298 K
Paper No. 39
10
λ
Air excess ratio
MCR
Maximum continous rating
MEP
Mean effective pressure, bar
Tad
Adiabatic flame temperature, K
Trad
w/w
[7] Teetz, C., ―Beitrag zur Verminderung der
Naßkorrosion im Dieselmotor‖. VDI
Forschungsheft, Vol. 626, 1984, pp. 1–40.
[8] Behrens, R., and Groth, K., ―Problems
caused by burning heavy fuels in diesel
engines (paper 3)‖. Diesel Engine
Combustion Chamber Materials for Heavy
Fuel Operation, 1990, pp. 29–38.
Temperature drop due to radiation
[9] Behrens, R., ―Brennstoffbedingte korrosion im
dieselmotor‖. Motortechnische Zeitschrift,
Vol. 49, 1988, pp. 479–485.
Weight fraction
ACKNOWLEDGEMENTS
We would like to thank our colleagues Mrs.
Charlotte B. Røjgaard, Mr. Kjeld Aabo, Mr.
Michael F. Pedersen, and Mr. Svend Eskildsen for
invaluable discussions during the making of this
manuscript. Further we would like to express our
deepest gratitude to Prof. Peter Glarborg,
Department of Chemical and Biochemical
Engineering, CHEC Research Centre, Technical
University of Denmark, for many enlightening
discussions and, in particular, for commenting on
the manuscript.
[10] Behrens, R., ―Fuel-related wear
mechanisms in heavy oil-fuelled diesel
engines, as exemplified by first piston ring
(paper 8)‖. Maritime Systems Integrity, 1988,
pp. 83–91.
[11] Nagaki, H., and Korematsu, K., ―Relation
between diffusion process of sulphur oxides
in exhaust gas oil film and wear of cylinder
liner and piston rings in diesel engines‖. SAE
paper, 912400, 1991.
REFERENCES
[12] Schramm, J., Henningsen, S., and
Sorensen, S. C., ―Modelling of corrosion of
cylinder liner in diesel engines caused by
sulphur in the diesel fuel‖. SAE paper,
940818, 1994.
[1] Cullis, C. F., and Mulcahy, M. F. R., ―The
kinetics of combustion of gaseous sulphur
compounds‖. Combust. Flame, Vol. 18,
1972, pp. 225–292.
[13] van Helden, A. K., Valentijn, M. C., and van
Doorn, H. M. J., ―Corrosive wear in
crosshead diesel engines‖. Tribology
International, Vol. 22, 1989, pp. 189–193.
[2] Turns, S. R., ―An Introduction to Combustion‖.
McGraw-Hill, 1996.
[14] Heywood, J. B., ―Internal Combustion
Engine Fundamentals‖, McGraw-Hill, 1989.
[3] Andreasen, A., and Mayer, S., ―Use of
seawater scrubbing for SO2 removal from
marine engine exhaust gas‖. Energy Fuels,
Vol. 21(6), 1996, pp. 3274–3279.
[15] Stone, R., ―Introduction to Internal
Combustion Engines‖, 3rd ed. MacMillan
Press Ltd, 1999.
[4] Engel, P. K., Thompson, R. E., and
Silvestrini, R., ―Corrosion and fouling
potential in diesel exhausts‖. Trans. ASME
Journal of Engineering for Power, Vol. 101,
1979, pp. 598–607.
[5] McKinley, T. L., ―Modeling sulfuric acid
condensation in diesel engine EGR coolers‖.
Engine modelling, Vol. 1255, 1997, pp. 207–
218.
[6] Huijbregts, W. M. M., and Leferink, R. G. I.,
―Latest advances in the understanding of
acid dewpoint corrosion: corrosion and
stress corrosion cracking in combustion gas
condensates‖. Anti-Corrosion Methods and
Materials, Vol. 51, 2004, pp. 173–188.
© CIMAC Congress 2010, Bergen
[16] Sorokin, A., Katragkou, E., Arnold, F.,
Busen, R., and Schumann, U., ―Gaseous
SO3 and H2SO4 in the exhaust of an aircraft
gas turbine engine: measurements by cims
and implications for fuel sulfur conversion to
sulfur (VI) and conversion of SO3 to H2SO4‖.
Atmospheric Environment, Vol. 38, 2004,
pp. 449–456.
[17] Starik, A. M., Savel‘ev, A. M., Titova, N. S.,
and Schumann, U., ―Modeling of sulfur
gases and chemiions in aircraft engines‖.
Aerospace Science and Technology, Vol. 6,
2002, pp. 63–81.
[18] Tremmel, H. G., and Schumann, U., ―Model
simulations of fuel sulfur conversion
efficiencies in an aircraft engine:
Dependence on reaction rate constants and
Paper No. 39
11
initial species mixing ratios‖. Aerosp. Sci.
Technol., Vol. 3, 1999, pp. 417–430.
[19] Wilson, C. W., Petzold, A., Nyeki, S.,
Schumann, U., and Zellner, R.,
―Measurement and prediction of emissions
of aerosols and gaseous precursors from
gas turbine engines (PartEmis): an
overview‖. Aerosp. Sci. Technol., Vol. 8,
2004, pp. 131–143.
[20] Schumann, U., Arnold, F., Busen, R.,
Curtius, J., Kärcher, B., Kiendler, A.,
Petzold, A., Schlager, H., Schröder, F., and
Wohlfrom, K.-H., ―Influence of fuel sulfur on
the composition of aircraft exhaust plumes:
The experiments SULFUR-7‖. Journal of
Geophysical Research, Vol. 107, 2002,
pp. 2–1–22–7.
[21] ‗‗Exhaust Gas Emissions Control – Today
and Tomorrow‖‘ MAN Diesel technical paper,
Copenhagen No.: 5510-0060-00ppr,
http://www.mandiesel.com/files/news/filesof9
187/5510-0060-00ppr.pdf, Sep 2008
[22] Kaltoft, J. ―EGR application prototype test‖,
Deliverable: D7.2.b HERCULES Integrated
Project, Contract no. TIP3-CT-2003-506676,
Sixth Framework Program, 2007
[23] Basic priciples of ship propulsion. Tech.
rep., MAN Diesel A/S, Copenhagen SV,
Denmark.
[24] Hindiyarti, L., Glarborg, P., and Marshall, P.,
―Reactions of SO3 with the O/H radical pool
under combustion condistions‖. J. Phys.
Chem. A., Vol. 111, 2007, pp. 3984–3991.
[25] Glarborg, P., Kubel, D., Dam-Johansen, K.,
Chiang, H.-M., and Bozzelli, J., ―Impact of
SO2 and NO on CO oxidation under postflame conditions‖. Int. J. Chem. Kinet., Vol.
28, 1996, pp. 773–790.
[26] Alzueta, M. U., Bilbao, R., and Glarborg, P.,
―Inhibition and sensitization of fuel oxidation
by SO2‖. Combustion and Flame, 127, 2001,
pp. 2234–2251.
[27] Yilmaz, A., Hindiyarti, L., Jensen, A.,
Glarborg, P., and Marshall, P., ―Thermal
dissociation of SO3 at 1000–1400 K‖. J.
Phys. Chem. A., Vol. 110, 2006, pp. 6654–
6659.
[28] Rasmussen, C. L., Glarborg, P., and
Marshall, P., ―Mechanisms of radical
removal by SO‖. Proceedings of the
Combustion Institute, Vol. 31, 2006,
pp. 339–347.
© CIMAC Congress 2010, Bergen
[29] Glarborg, P., ―Hidden interactions - trace
species governing combustion and
emissions‖. Proceedings of the Combustion
Institute, Vol. 31, 2007, pp. 77–98.
[30] ―Sulphur mechanism extension v. 5.2‖.
http://www.chem.leeds.ac.uk/Combustion/so
x.htm.
[31] Reiner, T., and Arnold, F., ―Laboratory
investigations of gasseous sulfuric acid
formation via SO3+H2O+M → H2SO4+M:
Measurement of the rate constant and
product identification‖. J. Chem. Phys., Vol.
101, 1994, pp. 7399–7407.
[32] Goodwin, D. G., ―Cantera C++ user‘s guide‖,
Tech. rep., California Institute of Technology,
October, 2002.
[33] Goodwin, D. G., ―Defining phases and
interphases – cantera 1.5‖, Tech. rep.,
California Institute of Technology, August,
2003
[34] McBride, B. J., Zehe, M. J., and Gordon, S.,
―NASA Glenn coefficients for calculating
thermodynamic properties of individual
species‖, Tech. Rep. NASA TP-2002211556, National Aeronautics and Space
Administration, September, 2002.
[35] Bishnu, P. S., Hamiroune, D., and
Metghalchi, M., ―Development of constrained
equilibrium codes and their applications in
nonequilibrium thermodynamics‖. Trans.
ASME Journal of Energy Ressources and
Technology, Vol. 123, 2001, pp. 214–220.
[36] Andersson, M., Johansson, B., Hultquist, A.,
and Nöhre, C., ―A real time NOx model for
convetional and partially premixed
combustion‖. SAE paper, 2006-01-0195,
2006.
[37] Lide, D. R., ed., “Handbook of Chemistry
and Physics‖, 78th ed. CRC Press LLC,
1997.
[38] Campbell, C. T., ―Towards tomorrow‘s
catalysts‖. Nature, Vol. 432, 2004, pp. 282–
283.
[39] Stegelmann, C., Andreasen, A., and
Campbell, C. T., ―Degree of rate control:
How much the energies of intermediates and
transition states control rates‖. Journal of the
American Chemical Society, Vol. 131(23),
2009, pp 8077–8082.
[40] Cerru, F., Kronenburg, A., and Lindstedt, R.,
―A systematically reduced reaction
mechanism for sulphur oxidation‖. Proc.
Paper No. 39
12
Combust. Inst., Vol. 30, 2005, pp. 1227–
1235.
[41] Cerru, F., Kronenburg, A., and Lindstedt, R.,
―Systematically reduced chemical
mechanisms for sulfur oxidation and
pyrolysis‖. Combust. Flame, Vol. 146, 2006,
pp. 437–455.
[42] Wang, X., Jin, Y. G., Suto, M., and Lee,
L. C., ―Rate constants of the gas phase
reaction of SO3 with H2O‖. J. Chem. Phys.,
Vol. 89, 1988, pp. 4853–4860.
[43] Hofmann, M., and von Ragué Schleyer, P.,
―Acid rain: Ab Initio investigation of the
H2O⋅SO3 complex and its conversion into
H2SO4‖. J. Am. Chem. Soc., Vol. 116, 1994,
pp. 4947–4952.
[44] Lovejoy, E. R., Hanson, D. R., and
Gregory Huey, L., ―Kinetics and production
of the gas-phase reaction of SO3 with water‖.
J. Phys. Chem., Vol. 100, 1996, pp. 19911–
19916.
[45] Jayne, J. T., Pöschl, U., Chen, Y.-M., Dai,
D., Molina, L. T., Worsnop, D. R., Kolb,
C. E., and Molina, M. T., ―Pressure and
temperature dependency of the gas phase
reaction of SO3 with H2O and the
heterogeneous reaction of SO3 with
H2O/H2SO4 surfaces‖. J. Phys. Chem. A.,
Vol. 101, 1997, pp. 10000–10011.
[46] Livbjerg, H., and Villadsen, J., ―Kinetics and
effectiveness factor for SO2 oxidation on an
industrial vanadium catalyst‖. Chemical
Engineering Science, Vol. 27(1), 1972,
pp. 21 – 38.
[47] Froment, G. F., and Bischoff, K. B.,
nd
Chemical Reactor Analysis and Design, 2
ed. John Wiley & Sons, Inc., 1990.
Anders Andreasen and Stefan Mayer
Basic Research, Process Development
Reasearch & Development, Marine Low-speed
MAN Diesel
Teglholmsgade 41
DK-2450 Copenhagen SV Denmark
© CIMAC Congress 2010, Bergen
Paper No. 39
13