Emissions Technology

Exhaust emissions reduction technology for Sulzer

marine diesel engines: General aspects

by Rudolf Holtbecker Markus Geist

Project Manager, Product Manager,
Emissions RTA-T Engines

Wärtsilä NSD Switzerland Ltd, Winterthur

July 1998

Contents Page

Introduction 3

Fundamentals of diesel exhaust components 5

Marine exhaust emissions legislation 9

Diesel engine developments towards low emissions 11

Primary measures (combustion control techniques) 13

Secondary measures (exhaust gas aftertreatment) 26

Conclusion 31

References 31

Summary
The paper reviews the various types of exhaust gas emissions of large diesel engines, particularly of low-
speed two-stroke types such as the Sulzer RTA-series engines, and the possible measures for controlling such
emissions in marine installations. It draws on the wealth of experience with exhaust gas emissions built up at
Wärtsilä NSD Switzerland Ltd with respect to both two-stroke and four-stroke diesel engines. The company
began studying exhaust gas emissions control techniques in the 1970s with research beginning in the early
1980s, and has since applied them in stationary applications. Emissions control is being extended to the
marine environment through the IMO limits to be introduced in the year 2000.
This paper is only intended as a general guide. Please consult the appropriate engine documentation re-
garding specific matters, such as engine adjustments for complying with emissions regulations, emissions
measurements, testing, emissions monitoring, etc., as they apply to specific Sulzer diesel engines.

1
Introduction
Everyone is more aware these days of environmen-
tal issues. Gradually efforts are being made globally
to curb pollution in all its forms. Even though the con-
tribution of large diesel engines to global pollution is
relatively small, attention has been given to their ex-
haust emissions since the 1970s. At first, the focus
was on stationary applications, and there are now
various countries which have environmental legisla-
tion requiring reductions of the exhaust emissions
from diesel power plants of varying degrees up to 100
per cent. For example, most stringent emissions
regulations apply for the city of Zurich, Switzerland,
where the NOX emissions of diesel engines must be
reduced by more than 99 per cent [1]. In January
2000, exhaust emissions control will be extended to
marine diesel engines world-wide through the regula-
tions of the International Maritime Organization (IMO).
We accepted the challenge of controlling the ex-
haust emissions from Sulzer diesel engines in the
1970s when the first exhaust emissions regulations
came into force in the USA (Fig. 1) through the EPA
(Environmental Protection Agency). From that time
on, intense efforts have been made to ensure that
future engines can be developed to comply with
forthcoming regulations. Emissions control has thus
always been one of the major topics in the develop-

Fig. 1
The Hoffmann-La Roche power plant in Belvidere NJ, USA, with a Sulzer
10RNF90M two-stroke engine. This was the first Sulzer diesel engine adapted
to give low exhaust emissions in compliance with EPA regulations
[0783-3316]

3
ment of new Sulzer diesel engines. Through these
many years of experience, we have accumulated
a thorough knowledge about emissions control, in-
cluding its influence on the engine, particularly with
respect to reliability, fuel consumption and other key
operating factors.
Owing to the high efficiency of diesel engines, the
emissions of carbon dioxide, carbon monoxide and
hydrocarbons are low in their exhaust gases but high
emissions of nitrogen oxides are also inherently char-
acteristic of the diesel cycle. The same high combus-
tion temperatures that give high thermal efficiency in
the diesel engine are also the most conducive to NOX
formation. By running on low-quality fuels with a low
fuel consumption, large diesel engines offer enor-
mous savings in fuel costs compared with those of
alternative prime movers. Yet the heavy fuel oil leads
to substantial emissions of oxides of sulphur and par-
ticulates. They are formed in the combustion process
out of the sulphur, ash and asphaltenes that are
always present in heavy fuel oil. However, as shown
in this paper, there are ways of reducing such emis-
sions, either by combustion control techniques or
by exhaust gas aftertreatment.
Summary of research activities to
reduce emissions of Sulzer engines
Research and development of low-emission tech-
nologies for Sulzer diesel engines were intensified
from the beginning of the 1980s. Tests were carried
out on many of the actual low-speed two-stroke
engines of that time starting with the Sulzer RND and
RNF series, going to the RLA and RLB, and later with
the RTA engine series. There has also been a contin-
ual two-way interchange of test results and experi-
ence between Sulzer four-stroke and two-stroke
engines. Naturally this course of R&D activities
involved many steps. Some of the principal ones,
however, may be summarised as follows:
1982–1986
The efficiency of fuel/water emulsions to reduce
NOX emissions was investigated. This included tests
on eight- and ten-cylinder Sulzer RNF90M engines
and 4RLB76 engines. As fuel, marine diesel oil as well
as heavy fuel oil were used.
During this period, tests were also made with
other primary measures on various other Sulzer
diesel engines.
1992–1995
On the Sulzer 6RTA38 research engine in Graven-
chon, an extensive test programme was carried out
jointly with Mobil Oil Corporation and Lloyd’s Regis-
4
ter of Shipping to investigate new selective catalytic
reduction (SCR) technologies and define the best
operating conditions for an SCR system.
1993
Fuel/water emulsion tests with marine diesel oil
were carried out on a 7RTA84T engine.
1993
An intensive research programme was carried out
on the 4RTX54 research engine in Winterthur to ana-
lyse different primary measures.
1994
At the Federal Institute of Technology (ETH) in
Zürich, Switzerland, a Sulzer 9S20 was installed with
an SCR system. A five-year research programme was
initiated focusing on primary NOX reducing technolo-
gies including a new concept for exhaust gas recircu-
lation, Miller supercharging and new fuel injection
systems.
1994
A research programme was started to investigate
possibilities for removing dust from exhaust gas by
the injection of different types of liquids in the exhaust
gas. A wet scrubber was installed after the 4RTX54
research engine in Winterthur.
1995
A direct water injection system was installed on the
4RTX54 research engine and showed the high poten-
tial of this technology for the reduction of NOX emis-
sions.
1995
The Diesel Technology Center was inaugurated.
The 4RTA58T and 8ZA40S engines were installed for
testing. Both engines are equipped with SCR units to
meet the local emission limits. On both engines ex-
tensive research programs were followed for the re-
duction of emissions by means of primary measures
covering the variation of all tuning parameters avail-
able on the engines.
1996
Dedicated exhaust gas recirculation tests were
carried out on the two-cylinder ZA40S high-pressure
gas engine.
1996
The 6ZA50S four-stroke engine was installed in the
Diesel Technology Center and adjusted for low emis-
sions and compliance with the IMO NOX regulation.
1998
A new type of cyclone for the reduction of particles
in the exhaust gas was tested in the Diesel Technol-
ogy Center with the 4RTA58T engine running on
heavy fuel oil.
Fundamentals of diesel engine exhaust components
The diesel engine process is well known and
therefore does not require discussion here. On the
other hand, it is appropriate to give a brief summary
of the various constituents of exhaust gases from
large diesel engines, particularly those that are re-
garded as pollutants, their effects on the environ-
ment in high concentrations, and their numerical
importance.
Exhaust gases emitted from large diesel engines
comprise nitrogen, oxygen, carbon dioxide and
water vapour with smaller quantities of carbon
monoxide, and oxides of sulphur and nitrogen, to-
gether with even lesser quantities of partially reacted
and unburned hydrocarbons and particulate mat-
erial (Figs. 2 and 3).
Molecular nitrogen and oxygen
Free nitrogen, N2, and oxygen, O2, comprise the
major constituents of both the air intake and the ex-
haust emissions from a diesel engine. Nitrogen, form-
ing 78 per cent by volume of the intake air, basically
does not react in the combustion process, although a
very small proportion will react chemically with oxy-
gen to form various oxides of nitrogen, jointly design-
ated NOX. Oxygen, originally forming 21 per cent by
volume of the intake air, will only be partially convert-
ed by the combustion process. Consequently the free
oxygen component of the exhaust will be a function of
the excess air ratio at which the engine is operated.
Air ➪ ➪
7.8 kg/kWh
Fuel ➪ N2
171 g/kWh
Lubricating
oil ➪ H2O 0.2 g/kWh
1.3 g/kWh
Fig. 2
Typical analysis of the
exhaust gases from
modern two-stroke
marine diesel engines
[98#126]
Carbon dioxide and water vapour
Carbon dioxide, CO2, and water vapour, H2O, are
formed in all combustion processes in which com-
plete, or nearly complete, combustion of a hydrocar-
bon fuel takes place. Their relative proportions are
determined primarily by the hydrocarbon composi-
tion of the fuel. Thus the production of both carbon
dioxide and water vapour is a function of the quantity
of fuel burned, which to a large extent is determined
by the plant efficiency and the elemental composition
of the fuel being burned.
Although traditionally not regarded as a pollutant,
carbon dioxide has become of increasing concern in
recent times through its importance as a ‘greenhouse
gas’ and the unknown consequences for the global
climate of the trend of rising carbon dioxide concen-
tration. Inputs of water vapour (also a principal green-
house gas) to the atmosphere are, as yet, of little con-
cern as the average global concentration is not
believed to be changing [2].
However, when applying conventional technol-
ogies to reduce NOX emissions, the fuel consumption
– and as a consequence the emissions of carbon
dioxide – tends to increase. The environmental im-
pact of such an increase in carbon dioxide is another
reason, apart from the financial consequences of the
increased fuel consumption, why it is absolutely im-
portant that the development of NOX reduction meas-
ures should not increase the fuel consumption.
Exhaust gas
8.0 kg/kWh
6.2 kg/kWh
O2 1.1 kg/kWh
CO2 0.5 kg/kWh
NOX 17.0 g/kWh
SOX 13.6 g/kWh
CO 0.4 g/kWh
HC 0.4 g/kWh
Dust 0.6 g/kWh
➪ Power kW
5

N2
O2
H 2O
CO 2
CO
Pollutants
SO X
HC NO X
Fig. 3
The pollutants in the exhaust gases (% vol) of large diesel
engines are an extremely small proportion of the whole
[98#127]
Oxides of nitrogen, NOX
The formation of oxides of nitrogen, NOX, occurs
as a result of the oxidation of molecular nitrogen
in the combustion air and of organic nitrogen in the
fuel. In the latter case, it would be expected that the
bulk of the organic nitrogen will be oxidised during
the combustion process. As heavy fuel oil has a
greater content of organic nitrogen than marine diesel
oil or other distillate fuels, the NOX emissions of an
engine running on heavy fuel are thus greater.
The oxidation of atmospheric nitrogen is influ-
enced by local conditions in the combustion cham-
ber, such as for example the maximum cylinder pres-
sure, local peak temperatures and local air-to-fuel
ratios. Nitric oxide NO is a primary reaction product,
but around five per cent of it is converted to nitrogen
dioxide NO2 later in the combustion cycle, during
expansion and during the flow through the exhaust
system. At the same time, a very limited proportion of
nitrous oxide N2O is also formed. Further oxidation of
NO to NO2 subsequently continues at ambient
temperatures after the exhaust gases have passed
out to the atmosphere.
The environmental effects of NOX are diverse.
Nitrogen oxide is of particular concern because of its
detrimental effects on respiration and plant life, as
well as its significant contribution to acid rain. In ad-
dition, NOX, together with volatile organic com-
pounds (VOC), is also involved in a series of photo-
chemical reactions that lead to an increase in
troposphere ozone which, in turn, adversely affects
human health, crop yields and natural vegetation. It
should be mentioned that these problems are only
pronounced on land, and especially in urban areas.
6
Nitrogen oxides are one of the primary pollutants
which have been introduced in regulations, such as
those adopted by the IMO.
Sulphur oxides, SOX
The oxides of sulphur, SOX, derive directly from the
sulphur content of the fuels used. In the combustion
chamber, the sulphur is oxidised thereby forming
principally sulphur dioxide SO2 and, to a much lesser
extent, sulphur trioxide SO3. The alkaline lubricants
employed to protect the liner surface from acidic
corrosion convert a small proportion of the SOX
produced by the combustion process to calcium
sulphate. However, this is a relatively insignificant
proportion and the sulphur emissions from the engine
will essentially be proportional to the sulphur content
of the fuel and the fuel consumption.
Sulphur oxides are the major source of acid rain.
They thus have detrimental effects on plant life, veg-
etation, human respiration and buildings [2]. On the
other hand, it has to be stressed that the effect of acid
rain on the sea is in most cases negligible. This is be-
cause sea water is slightly alkaline. One litre of sea
water can, on average, neutralise some 300 mg of
sulphur. It also seems that SOX is only transported
over a relatively short distance (some ten or at most
100 km). It is not surprising, therefore, even in coast-
al regions bordering on the most heavily used ship-
ping waters in the world, such as the English Channel
or the Strait of Malacca, that the contribution of ships
to the total SO2 emission has been found to be
considerably less than five per cent of that from land-
based sources [3].
Nevertheless, limitation of the emission of sulphur
oxides has been introduced in the recent IMO regula-
tions.
Carbon monoxide, CO
Carbon monoxide, CO, is the result of incomplete
combustion of carbonaceous material. Its formation
in the diesel engine is thus principally a function of the
excess air ratio, the temperature of combustion and
the uniformity of the air/fuel mixture in the combus-
tion chamber. In general, carbon monoxide emissions
from large two-stroke diesel engines are low owing to
the high oxygen concentrations and the efficient
combustion process.
Broad environmental effects are not generally of
major concern from carbon monoxide, although the
gas may have some small influence on global climate
change [2]. Owing to the low level of carbon monox-
ide in diesel exhaust gases, it has not so far been
addressed in marine regulations.
Hydrocarbons, HC
The hydrocarbon fraction, HC, of the exhaust gas-
es will predominantly consist of unburned or partially-
burned fuel and lubricating oils. In reality this fraction
comprises a myriad of individual organic compounds
with almost every chemically allowable configuration
of carbon, hydrogen, nitrogen, oxygen and sulphur
being represented, albeit at extremely low concentra-
tions.
Individual components may be present in either
vapour or particulate phases, or may switch between
the two phases through evaporation, condensation
and polymerisation reactions leading to a constantly
changing distribution. Consequently, the diverse na-
ture of the hydrocarbon fraction components makes
for difficulties in both quantifying the emissions and in
identifying the specific health and environmental
problems.
In general, hydrocarbon emissions result from in-
complete combustion. The nature and level of hydro-
carbons in the exhaust will thus be largely dependent
upon the combustion characteristics and thermal
efficiency of the engine which, in turn, are influenced
by engine load, condition and maintenance. With
correctly-adjusted diesel engines, hydrocarbons are
emitted in very small quantities and thus present
only a minor problem with respect to emissions in the
marine environment.
Particulates
The particulate fraction of the exhaust emissions
represents a complex mixture of inorganic and organic
substances largely comprising elemental carbon, ash
minerals, heavy metals, condensed sulphur oxides,
water, wear debris and a variety of unburned or partial-
ly-burned hydrocarbon components of the fuel and
lubricating oils. An intermittent discharge of accumu-
lated deposits from the exhaust system may also be
encountered. With the exception of the latter, the ma-
jority of diesel particulates are likely to be less than one
micron in diameter, readily transportable by air currents
and with a low settling velocity. Potentially detrimental
effects may thus be encountered away from the imme-
diate vicinity of the exhaust gas plume [2].
Although the study of the particulate composition
of marine diesel exhaust is limited, extrapolation of
results in various studies of road vehicle applications
would suggest that particulates may potentially
cause some respiratory problems, as well as more
serious health effects.
To a large extent, the magnitude of particulate
emissions will be dependent upon the fuel oil quality
and the completeness of combustion with ‘smoke’
traditionally acting as a criterion of combustion qual-
ity. Yet smoke is simply a visual criterion and does not
necessarily imply that the engine is polluting. For
example, white smoke is usually condensing water
vapour, and some real pollutants are in fact invisible.
Quantification of the particulate emissions is diffi-
cult on account of their complex nature. Thus, a
variety of terms are employed to describe both the
nature and quantity of particulate matter. Many terms
are defined by sampling and quantification methods,
and include suspended particulate matter and total
suspended particulates.
Owing to the fact that the different measuring tech-
niques yield quite different results, it might be helpful
to review the different ways of measuring the particu-
late emissions [4]:
• Smoke spot methods: A certain volume of exhaust
gas is drawn through a filter paper and the reduc-
tion in reflectance of the used filter paper caused
by the soot captured on it is measured or evaluat-
ed. Examples are Bosch Smoke Number, SAE
Smoke Number, Bacharach (ASTM) Smoke Num-
ber, etc.
• Opacity methods: The percentage that light is re-
duced when it is passed through the exhaust gas
is determined (for example Hartridge).
• Hot filter method: A sample is taken from the ex-
haust gas, and drawn through a filter maintained at
a temperature similar to that of the exhaust gas.
The mass of the collected material is then deter-
mined gravimetrically and referred to as ‘dust’. An
example of a loaded filter mat is shown in figure 4.
• Dilution tunnel: A sample is taken from the exhaust
gases, diluted with clean air (to emulate the situa-
tion after the funnel), and then passed through a
filter maintained at a temperature of 52 °C maxi-
mum. The mass of the collected material is then
determined gravimetrically and referred to as
‘particulates’.
Other terms, such as ‘inhalable’ or ‘respirable’, re-
late to the site of deposition in the respiratory tract
whilst PM10 declares particulate matter with an aero-
dynamic diameter of 10 microns, a dimension that
influences the site of deposition as well as the
physiological effect. An inability, in general, to trans-
late the results of measurements using one method
into those using another further complicates the
issue. Quite obviously, smoke spot, smoke opacity,
dust and particulates are not the same and therefore
should not be confused. It should also be noted that
a high sulphur content makes measuring particulates
by using a dilution tunnel rather problematic. There-
7
 Fig. 4
An example of a loaded
filter mat from the hot filter
method of determining
the particulates content
of exhaust gases
[98#128]

fore, in ISO 8178, [5], it is stated that for engines run- Units
ning on fuels having a sulphur content above 0.8 per
While nothing else but mass emissions in terms of
cent by weight, only hot filter methods, as described
kilograms per hour or tonnes per year plays an impor-
for example in ISO 9096 [6], should be employed.
tant role for the environment, it is widely accepted
Contrary to NOX, particulates are not an inherent
that for industrial applications the amount of emis-
product of the diesel process itself. There has never
sions should be related to a certain produced output
been a shortage of hypotheses on soot and particu-
such as the engine power, resulting in the unit g/kWh.
late formation in diesel engines. Virtually all of these
Owing to the various measuring principles it is also
models however, only refer to middle-distillate fuels
common to relate the amount of emissions to the ex-
and, moreover, are so complex that they are impos-
haust gas volume flow. However, it has to be consid-
sible to use in practice. The principal known influen-
ered that, when using parts per million by volume
ces on the particulate emissions are the fuel injection
(ppmV) or mass of emissions per volume of exhaust
and the fuel used. Concerning the latter, several em-
gas (mg/Nm3), the quantity of air throughput plays an
pirical equations have been established for residual
important role.
fuels, driven by the fact that dust levels are roughly
Consequently, exhaust emissions concentrations
five times higher on heavy fuel oil than on marine
have to be related to a certain reference oxygen con-
diesel oil. For example, Dexter [7] found a linear
centration (for example 15, 13 or five per cent by vol-
dependence of the particulates upon the amount of
ume, which are written as 15%vol, 13%vol or 5%vol),
Conradson carbon residue in the fuel.
which does not normally have any relation to the real
Independent of the existing confusion, it has to be
exhaust oxygen concentration but is adopted to
stated that reduction of the particulate emissions of
assist in the numerical treatment of varying exhaust
diesel engines may represent the greatest challenge
figures*. The conversion from measured exhaust gas
for the future development of the diesel engine.
concentrations in ppmV to g/kWh is described in the
standard ISO 8178 part 1 [5].

* A given exhaust-gas concentration xref can be converted

from the reference oxygen concentration O2ref to another
oxygen concentration O2corr by using the following equa-
tion:
20.95% – O2corr
xcorr = Xref 20.95% – O
2ref

8
Marine exhaust emissions legislation
The international nature of shipping, of the ship-
building industry and of the marine equipment supply
infrastructure commends the international harmon-
isation of equipment standards and regulations. A
proliferation of national or regional regulations based
on different standards would be highly disruptive and
costly to the marine industry and would create need-
less barriers to international trade. The marine sector
is unique in having the benefit of an international reg-
ulatory body in the form of the IMO, the International
Maritime Organization.
In September 1997 the Conference of Parties to
the MARPOL 73/78 Convention adopted the new An-
nex VI. In Regulation 13 of this annex, limiting NOX
levels are defined for engines of more than 130 kW
output (Fig. 5). These NOX emissions limits will only
apply to engines installed in new ships built on or
after 1 January 2000, or to engines in existing ships
undergoing a major conversion. The basis of the reg-
ulation is that the average NOX emissions of new
ships must be some 30 per cent less than the emis-
sions from ships in 1990.
In Regulation 14 of Annex VI, the sulphur content
in fuels is limited to 4.5 per cent. To protect sensitive
areas from sulphur oxides emissions, the concept
of SOX special areas has also been introduced. The

Fig. 5
NOX emissions limit i) 17 g/kWh when n is less than 130
according to the IMO II) 45*n –0.2 g/kWh when n is 130 rpm or more, but less than 2000 rpm
regulation 13, Annex VI of iii) 9.84 g/kWh when n is 2000 rpm or more
MARPOL 73/78
where n = rated engine speed
[98#129]
NOx [g/kWh]
25

20

15

10

0
0 200 400 600 800 1000

Rated Engine Speed [rev/min]

The three test cycles D2, E2 and E3 apply to three different engine categories namely:
D2: Generating sets with intermittent load, including shipboard generating sets
E2: Constant-speed heavy-duty engines for ship propulsion, including engines with a
controllable-pitch propeller
E3: Variable-speed heavy-duty engines for ship propulsion equipped with a fixed-pitch
propeller
Table: Weighting factors (f) for the test cycles D2, E2 and E3 in the Technical Code of the
IMO regulation for NOX emissions

% rated speed 100 100 100 100 100 100 91 80 63

% rated load 100 75 50 25 10 100 75 50 25

cycle D2 0.05 0.25 0.30 0.30 0.10 – – – –

Weighting factor [f] cycle E2 0.20 0.50 0.15 0.15 – – – – –

cycle E3 – – – – – 0.20 0.50 0.15 0.15

g
NOxweighted = ∑f
i
i x NOxi
h ()
∑fi x Pi (kW)
i

9
Baltic Sea represents such an area and others may
follow in the future. In these areas, fuel with less than
1.5 per cent sulphur has to be burned or the exhaust
gas has to be treated to reduce the sulphur oxides
emissions to a level under 6 g/kWh.
In addition to regulation 13 of Annex VI, a Techni-
cal Code has been prepared which will be mandatory,
and describes the procedures and boundary condi-
tions for diesel engine testing, surveying and certifi-
cation for compliance with the NOX emissions limits.
The most important items are:
• NOX emissions measurements and calculations
have to be carried out in accordance with the Tech-
nical Code which, in this respect, is almost identi-
cal with ISO 8178. In particular, the actual NOX
value which has to be less than the limit value has
to be calculated by an averaging formula which
takes into account the emission values at different
load points of the engine operating range.
• Prior to installation on board, every engine shall be
adjusted to meet the applicable NOX emissions
limit and shall be provided with the requisite En-
gine International Air Pollution Prevention (EIAPP)
Certificate. In turn, the EIAPP Certificate qualifies
the ship for its requisite International Air Pollution
Prevention (IAPP) Certificate issued by the flag sta-
te. For those engines which, because of their size,
construction and delivery schedule, cannot be cer-
tified at the manufacturer’s works, the complete
NOX measurement procedure may be carried out
on board ship.
• Every engine should be provided with a technical
file specified in the Technical Code which identifies
those components and settings which influence
the NOX emissions and confirm the correct engine
specification to ensure compliance with the regu-
lations. The components and settings involved will
generally include parts of the fuel injection system,
valve cams, injection and valve timing, cylinder
cover, piston with piston rod and shims, connect-
ing rod and turbochargers. The file will need to
include adjustment settings, tolerances, and de-
tailed test measurements from the official shop
test, and will follow the engine throughout its life.
• For certification, the ‘engine family’ or the ‘engine
group’ concept may be applied. The ‘engine fami-
ly’ concept may be applied to series-produced en-
gines which, through their design, are proven to
have similar exhaust emissions characteristics,
and which are used as produced. The ‘engine
group’ concept may be applied to any engine type
having the same design features as specified but
subject to individual engine adjustment or modifi-
cation. Engine families or engine groups are repre-
10
sented by their parent engines. The certification
testing is only necessary for these parent engines.
Further engines can be certified by checking doc-
uments, components, settings, etc. which have to
show correspondence with those of the parent en-
gine. The parent engine of an engine family has to
incorporate those features which will most ad-
versely affect the NOX emissions level. During the
testing of the parent engine of an engine group, the
NOX influence of adjustments and modifications
which are usual for the group has to be demon-
strated.
• For confirmation after installation of a certified en-
gine, and for later repeat surveys on board, differ-
ent methods can be chosen by the ship operator,
namely checking components and parameters,
monitoring NOX, or making simplified measure-
ments. These methods are sufficient to deduce
the NOX performance by reference to the Technical
File.
US legislation from EPA and CARB
The EPA (Environment Protection Agency) of the
USA is also working on new regulations concerning
exhaust emissions. With a view to international har-
monisation, the EPA as well as CARB (Californian
Air Resources Board) have declared their intention to
recognise and accept the limits established by the
IMO for international shipping and not to introduce
lower limits. More stringent regulations are only
envisaged for smaller domestic vessels, such as for
example pleasure craft operating only in domestic
waters.
Marine legislation in Sweden
Swedish regulations, as complement to the IMO
regulations, aim for the goal of a 75 per cent reduc-
tion of NOX and SOX emissions by the start of next de-
cade. The regulations are implemented from the year
1998 and are based on a fee according to the actual
NOX emissions and the ship type. The fee is applied
to all ships, both new as well as existing, and all en-
gines except emergency units. It has to be paid when
entering Swedish waters. The full fee becomes due
for engines emitting more than 12 g/kWh NOX, for
emissions between 12 and 2 g/kWh the fee is propor-
tional to the NOX emissions, and below 2 g/kWh the
fee reaches its minimum. The procedure for the cer-
tification (to be renewed every two years) refers to 80
per cent load, and the measurements and conver-
sions are to be performed according to ISO 8178.
Diesel engine developments towards low emissions
As shown above, the most important gas compo-
nent which has to be reduced in diesel exhaust emis-
sions is NOX. The strongest motivation is given by the
forthcoming IMO regulations. Accordingly, the follow-
ing sections will predominantly deal with NOX emis-
sions and will give only basic indications of how other
emission components might be reduced in the future.
Analysis of the combustion process in the cylinder
and the reactions which are involved to form NO has
identified three main sources of NO [9] of which, as
mentioned above, some is converted to NO2 to give
the NOX mixture:
• Thermal NO: During combustion, high tempera-
tures are reached. Around 1500 K, and above,
enough thermal energy is available to dissociate
oxygen and nitrogen and also other molecules
formed during the combustion process itself. The
recombination of the elements leads to the forma-
tion of NO. The reaction processes are quite slow
so that most nitrogen oxides are formed during the
mixing of the stoichiometric combustion gases
with excess air in the cylinder.
• Fuel source: The main molecular components
of fuel are carbon and hydrogen. But marginal
amounts of nitrogen are also contained. During
combustion, 50 to 100 per cent of the available
nitrogen in the fuel is oxidised to NO owing to very
fast reactions.
• Prompt NO: Even if combustion occurs near stoi-
chiometric conditions, nitrogen originating from
the air is available near the combustion front and
can be combined to NO by the same reactions that
lead to NO from the fuel.
In low- and medium-speed diesel engines, by far
the most important part of NOX is generated in the
thermal NO process. However, fuel-derived NO be-
comes important when using heavy fuel oil because
such fuels contain more organic nitrogen than marine
diesel oil and other distillate fuels. Heavy fuel oil can
contain up to 0.5 per cent nitrogen which increases
the total NOX emission by as much as ten per cent.
Theoretical considerations
In addition to thorough experimental work, theo-
retical investigations have been undertaken at Wärt-
silä NSD Switzerland to determine the best solutions
for emission control.
A basic requirement for the calculation of NOX
is knowledge about the main reactions involved. It
was Zeldovich [10] who first described with sufficient
accuracy the chemical reactions leading to thermal
nitric oxide:
O2 + M ↔ 2O + M (1)
N2 + O ↔ N + NO (2)
N + O2 ↔ NO + O (3)
N + OH ↔ NO + H (4)
To obtain a first understanding of the main para-
meters influencing nitric oxide production, the Zeldo-
vich mechanism may be further simplified [11] to a
single equation showing the main dependence of the
NOX production rate on the temperature and the
nitrogen and oxygen concentrations.
Taking as an example combustion gas with three
per cent oxygen, the NOX formation rate can be

Fig. 6
Relative NO production
rate calculated with respect
to gas temperature in ∆NO/time [–]
combustion gas with three -4
per cent oxygen 10
[98#130]
10-5

10-6

10-7

10 -8

-9
10

-10
10
1500 1600 1700 1800 1900 2000 2100
Temperature [K]

11
shown to depend on the temperature (Fig. 6). As a
rule of thumb, it can be said that NOX formation at
temperatures above 1500 K increases by a factor of
ten for every 100 K rise.
In the cylinder, the molecular gas components are
not homogeneously distributed. In the axis of the fuel
oil spray, there is mainly fuel oil. At greater distances
from this axis, fuel is increasingly mixed with air until
only air is found. The uneven distribution influences
both the temperature and the availability of nitrogen
and oxygen. Wärtsilä NSD Switzerland has so far
followed two strategies for simulating this inhomo-
geneity:
• Process analysis with two-zone modelling;
• Computational reactive fluid dynamics (CRFD).
In process analysis of the combustion process, the
combustion chamber is divided into two zones [9].
The hot zone contains only fuel burned with air at
stoichiometric conditions with high temperatures
leading to high NOX production. The cold zone has
only air. During injection of fuel in the cylinder and the
ongoing combustion, the mass in the hot zone in-
creases and NOX production increases. However,
hot and cold zones mix together leading to a fall in
the temperature and accordingly diminishing of NOX
production in the hot zone (Fig. 7).
The two-zone model represents a raw simplification
of the processes in the combustion chamber. A far
more detailed analysis is possible with CRFD pro-
grams which solve the three-dimensional conserva-
tion equations for mass, momentum and energy of the
gas at each point in the combustion chamber. The
fuel spray is also calculated in detail with primary spray
and droplet break-up, coalescence and evaporation.
This latter form of analysis is still under develop-
ment and needs some further research before it is

Fig. 7
Temperatures calculated
for the hot zone (T1) where
Temperatures [K] ∆NO/time
there is only fuel burned
with air at stoichiometric
3000 140 conditions, and the cold
T1 zone (T2) where there is
2500 T2 120
only air, and the relative
Mean temperature NO production rate
∆NO/time 100
2000 [98#131]
80
1500
60
1000
40
500 20

0 0
-180 -90 0 90 180

Crank angle [degrees]

Fig. 8
Temperature field in the
combustion space of the
ZA40S medium-speed
four-stroke engine calculated
by three-dimensional
computational reactive fluid
dynamics. Vertical plane
through the fuel spray
at 20 degrees after start
of injection
[98#132]

12
able to replace experimental work. But it can already
give some detailed insights into the distribution of
physical parameters describing the gas state in the
cylinder. The graph in figure 8 shows the results for
temperature distribution obtained from calculations
for the Sulzer ZA40S four-stroke engine with an
extended version of the CRFD computer program
KIVA [12]. The original code has been extended with
improved models to describe primary spray break-up
and combustion [13]. The temperature distribution
Primary measures (combustion control techniques)
To reduce the exhaust emissions of today’s diesel
engines, both primary and secondary measures may
be applied. Primary measures focus on decreasing
the production of emission components during com-
bustion whereas secondary measures deal with the
abatement of the emissions in the exhaust gas.
The well-known drawbacks in employing catalytic
converters in ships, mainly the necessity of a reduc-
ing agent together with the additional space required
for the catalytic reactor, make them barely acceptable
to marine diesel engine users. Consequently, primary
reduction techniques are the first choice for to reduce
the formation of pollutants on board ships.
As shown above, the main factors influencing NOX
formation are the concentrations of oxygen and nitro-
gen and the local temperatures in the combustion
process. Therefore all primary measures (Figs. 9 and
Fig. 9
The various primary
measures for reducing NOX
emissions all focus on
the lowering of oxygen
concentration and peak
temperatures Temperature
[98#133]
Oxygen
concentration
shows distinct areas in the combustion chamber with
the highest temperatures being where high concen-
trations of NOX will be formed.
The above-mentioned simulation tools are con-
tinuously being improved and will more and more
replace experimental work and thereby lead to a
more efficient development process for Sulzer diesel
engines.
10), which aim to reduce NOX production, focus on
lowering the concentrations and peak temperatures.
On the other hand, methods for the reduction of
particulates and exhaust gas components, such as
hydrocarbons or carbon monoxide, deal with the
optimum mixing of fuel and air in the combustion
chamber to achieve even more complete combustion
of the injected fuel.
To investigate the effects of different primary
measures, extensive tests have been performed over
the past years on all types of Sulzer diesel engines.
Tests have involved four-stroke engines from the S20
to the ZA40S and ZA50S, as well as the two-stroke
engines available on the test beds in Winterthur, in
the Diesel Technology Center in Winterthur and in
production at the licensees, such as the most power-
ful engine available today, the Sulzer RTA96C.
Scavenge air temperature
Miller supercharging
Turbo cooling
Scavenge air pressure
Valve timing
Compression ratio
Delayed injection
Nozzle design
Plunger size
Exhaust gas recirculation
Water addition
13
NOX reduction by scavenge air cooling
and Miller supercharging
Both scavenge air cooling and Miller supercharg-
ing aim to reduce the maximum temperatures in the
cylinder by lowering the temperature before com-
pression. The straightforward method is the reduc-
tion of scavenge air temperature by improving the air
cooler efficiency. Tests showed that for every 3 °C
reduction there may be emitted around one per cent
less NOX. But the physical limit is given by the water
temperature entering the air cooler, which may be
as high as 32 °C in the tropics, so that further cooling
below the standard values becomes impossible.
Energy is required for further cooling. One solution
is the turbo-cooling concept with an extra turbo-
charger after the main air cooler. The principal idea is
to compress, cool down and re-expand the air in the
extra turbocharger to obtain lower air temperatures.
The disadvantages of this system are high investment
costs and loss of charging efficiency leading to a
higher fuel consumption. Therefore no tests have
been performed with this solution.

Fig. 10
Expected mean achievable
∆BSFC
NOX reductions for various
Compression ratio primary reduction measures
∆NOx max
Nozzle design and selective catalytic
reduction (SCR), and their
Delayed Injection corresponding fuel
consumption penalties
Water-fuel emulsion [98#134]
Valve timing

Plunger size

Scavenge air pressure

variation
Scavenge air cooling

Water injection

EGR

SCR

-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20

∆NOx max [%], ∆BSFC [g/kWh]

Fig. 11
Exhaust and inlet valve lifts
for a four-stroke diesel
Valve lift [–] engine with standard and
Miller supercharging, in
1.5 which the inlet valve is
Inlet valve standard closed before the piston
Inlet valve Miller reaches bottom dead centre
Exhaust valve
1
[98#135]

0.5

0
0 90 180 270 360 450 540 630 720

Crank angle [degrees]

14
 Fig. 12
Cycle pressures for an
Pressure [–]
RTA two-stroke engine with
the standard compression
ratio, with delayed Standard
injection, and with raised Delayed injection
compression ratio High compression ratio
[98#136] and delayed injection

-90 -60 -30 0 30 60 90

Crank angle [degrees]

A simpler system is the application of the Miller
supercharging strategy developed for four-stroke
engines. Using a high-pressure turbocharger and
closing the intake valves before the piston reaches
the bottom dead centre during the intake stroke
(Fig. 11), the same amount of air trapped in the cylin-
der can be guaranteed. But the expansion before
compression cools the air charge in the cylinder.
Tests on the Sulzer 9S20 engine in the ETH plant
showed that NOX emissions could be reduced by
15 to 20 per cent without any increase in fuel con-
sumption [14]. Miller supercharging is one of the very
few measures to reduce both NOX emissions and fuel
consumption. It was adopted in Sulzer ZA40S en-
gines more than ten years ago and is one reason for
their relative low NOX emissions.
NOX reduction by turbocharging and
valve timing
The excess air ratio in the combustion chamber
may be varied by changing the scavenge pressure or
valve timing. Tests have been performed on the
4RTX54 research engine and 4RTA58T prototype, as
well as on a 8ZA40S engine, and showed that by
decreasing excess air ratio NOX emissions can
be reduced. In two-stroke engine tests, the reduction
of the excess air from 2.2 to 1.9 was achieved by re-
tarding the exhaust valve closing and increasing the
compression ratio to keep the firing ratio (Pmax/
Pcomp) constant. A reduction of the NOX emissions
of about 15 per cent and a decrease in fuel consump-
tion by about 2 g/kWh have been measured.
Combining higher excess air ratios by increasing
the scavenge air pressure together with retarded fuel
injection may increase or decrease NOX emissions,
depending on the engine design.
NOX reduction by retardation of fuel
injection
An important factor in NOX production during com-
bustion is the after compression of burned gases.
When fuel and air have burned, high peak tempera-
tures are achieved. If this gas is further compressed,
even higher temperatures and pressures will be
reached leading to increased NOX emissions. The
problem may be overcome by later injection of the
fuel. This method may be the best known way to
reduce the NOX emissions. Basically the delayed
injection leads to lower peak pressures and therefore
to less compression after combustion. Lower peak
pressures lead to less engine efficiency.
In two-stroke engines, such as the 4RTA58T proto-
type, an injection delay of about two degrees crank
angle leads to around ten bar reduction of maximum
cylinder pressure, a ten per cent reduction in NOX
emissions and a fuel consumption rise of about
3 g/kWh. The maximum achievable reduction in NOX
by this measure is around 25 per cent. In view of the
increased fuel consumption, the possible reduction
may be limited because a higher exhaust gas energy
may lead to excessive turbocharger speed.

15
NOX reduction by increased
compression ratio
The decrease of efficiency by delayed fuel injection
as described above can be countered by increasing
the compression ratio. This can be accomplished by
either increasing the geometric compression ratio or
adjusting the valve timing. As valve timing would also
increase the excess air ratio, changing the geometric
compression ratio is preferred.
In figure 12, the pressure histories for an RTA two-
stroke engine with the standard compression ratio
and one with a higher ratio are shown. Figure 13
shows the corresponding production of NOX with a
lower amount for the case with higher compression.
The maximum NOX reduction that can theoretically
be achieved by this measure is approximately 25 per
cent with an increase in fuel consumption of about
one per cent.
NOX reduction by changing fuel injection
nozzles
The design of nozzles strongly influences the com-
bustion process itself. Orientation and size of the
nozzle holes define the depth of penetration and
location of the fuel spray, and as a consequence the
evaporation process, turbulence, mixing and com-
bustion.
Tests have been made with many different nozzle
configurations on the RTA58T, ZA40S and ZA50S en-
gines. Looking at the different geometries of two- and

Fig. 13
NOX production after top
dead centre for the same
NOx engine and conditions as
in figure 12
Standard
Delayed injection [98#137]
High compression ratio
and delayed injection

-5 0 5 10 15 20 25
Crank angle [degrees]

Fig. 14
Schematic spray patterns
in the combustion chambers
of a four-stroke engine
(far left) and a two-stroke
engine (left). The two-stroke
engine configuration is with
three nozzles
[98#138]

16
four-stroke engines, it becomes evident that the free-
dom in defining the parameters of the nozzles is
greater for two-stroke engines (Fig. 14). In four-stroke
engines, the central injection together with the flat
combustion chamber given by the relatively low
stroke-to-bore ratio demands an equal distribution of
the holes. With the side injection used in two-stroke
engines, the interference of the sprays coming from
the two or three nozzles in the combustion chamber
can be used to influence the combustion process.
Results of changes in fuel consumption and NOX
reduction measured for nozzle variations on the
ZA40S research engine and the 4RTA58T prototype
are shown in figure 15. It can be clearly seen that
there is a very close relation between these two
values. Whatever type of nozzle is used, the results
for fuel consumption and NOX emissions will lie close
to the curve shown.
For two-stroke engines, it has been possible to
develop nozzles which lead to reduced NOX emis-
sions at the same fuel consumption (Fig. 16). Compa-
red with today’s nozzles which are optimised in terms
of fuel consumption and low combustion chamber
surface temperatures, such low-NOx nozzles tend to
lead to higher surface temperatures which might
have a detrimental effect on engine reliability.

Fig. 15
Measured trade-off between
NOX emissions and brake ∆NOx [%]
specific fuel consumption by
changing fuel injection 50
nozzle specification for two- 40 two-stroke
and four-stroke engines and four-stroke
30
by applying low-NOX nozzles two-stroke low NOx
for two-stroke engines. 20
These results were 10
measured on the 4RTA58T
and ZA40S research engines 0
[98#139] -10
-20
-30
-40
-50
-6 -4 -2 0 2 4 6 8

∆ BSFC [g/kWh]

Fig. 16
Two fuel injection nozzles
for Sulzer RTA two-stroke
engines with different
spray angles and increased
number of spray holes.
Right:standard nozzle.
Far Right: For low-NOX
emissions
[98#140]

17
NOX reduction by changing injection
pressure and duration
By changing diameter of the fuel pump plunger or
total fuel nozzle flow area, the maximum injection
pressure and duration of injection can be modified.
Both parameters influence the penetration of the
spray, the break-up process of the liquid core exiting
the fuel nozzle and the turbulence induced in the
combustion chamber. Experiments on the 8ZA40S
research engine showed that injection pressure alone
has only a marginal influence on the NOX emissions.
A positive effect could be detected by a longer injec-
tion period on the 8ZA40S, as well as on the 4RTX54
research engine. Starting from the standard configu-
ration, in the best case up to five per cent reduction
in NOX could be achieved with every ten per cent pro-
longation of injection, which can be explained by a
weaker combustion at the beginning.
Direct injection
Emulsion
Fumigation
18
NOX reduction by water addition
To achieve greater NOX reductions than those
achievable by the engine tuning measures described
above, techniques such as exhaust gas recirculation
(EGR), direct injection of ammonia, and addition of
water to the diesel process, may be applied. They can
result in NOX reductions in the order of, or even grea-
ter, than 50 per cent. However, some of these mea-
sures are either not compatible with the use of heavy
fuel oil, may lead to a significant increase in the capi-
tal cost or operational costs, may involve other,
undesirable emissions, or may even combine these
disadvantages.
It has long been well known that introduction of
water into the combustion chamber reduces NOX for-
mation. However, as diesel engineers take every con-
ceivable care to stop water being carried over
through the water separators after the scavenge air
cooler to the cylinders, “re-injecting” the water after
these separators is sometimes regarded as a dubious
Fig. 17
The three ways to add water
to the diesel engine combu-
stion process, by direct
injection in parallel with fuel
injection, by fumigation with
the scavenge air, and by
emulsion with the fuel oil
[98#141]
measure. Principally, these suspicions are justified,
as impingement of liquid water on the cylinder liner
surface leads to immediate disintegration of the lubri-
cation oil film. However, in this context, a clear sepa-
ration of two relevant aggregates of water is required.
Only liquid water destroys the oil film. Once the water
has evaporated, it becomes completely harmless for
affecting the oil film and for corrosion. Even at the
very high water/fuel ratio of 1:1, the dew point of sul-
phuric acid will rise by no more than 15° C [15]. In that
case, on one hand, evaporation of the injected water
is guaranteed and, on the other, condensation in the
combustion chamber is not possible at any time.
Still, this does not mean that all methods of water
addition are completely harmless (Fig. 17). Fumiga-
tion, meaning the addition of water to the scavenge
air, is the simplest way to apply water as an agent to
reduce NOX formation, and may reduce NOX emis-
sions by typically ten per cent for each 20 per cent
water addition to the fuel. However, as the water does
not completely evaporate in the cold scavenge air, it
will inevitably impinge on the cylinder liner, leading to
the above-described disintegration of the oil film. It is
therefore necessary to consider other ways to add
water, namely fuel/water emulsion and direct water
injection. To obtain the greatest NOX reduction, the
water has to be injected at the right time at the right
place, that is in the area where NOX are formed.
A simple approach may lead to the idea that only with
fuel/water emulsion can the water reach the flame
area [16]. However, a more detailed investigation
shows that both methods are at least equal with
regard to NOX reduction efficiency.
Fig. 18
Influence of percentage
water content in fuel/water NOx [g/kWh]
emulsions and fuel injection
timing on NOX emissions for 19
0% Water
a Sulzer 7RTA84T two-stroke
engine at 75 per cent load 17
[98#142]
15% Water
15
13
11
9
7
165
Whatever system of injecting water is employed,
consideration must be particularly given to the logis-
tics of providing sufficient fresh water on board ship.
Fresh water generators can be heated using the
engine cooling water, or using steam from an exhaust
gas economiser. In addition there must be sufficient
tank capacity for the water, with the necessary hand-
ling system.
Emulsion
Fuel/water emulsion is a well-known technique for
reducing NOX emissions. It was first tested on Sulzer
engines about 20 years ago and soon after that was
commercially applied on a Sulzer 10RNF90M two-
stroke engine in a power plant in the USA.
Running an engine on fuel/water emulsion makes it
theoretically possible to reduce NOX emissions by up
to 50 per cent with the required water quantity being
about one per cent for each percentage point reduc-
tion in NOX, as shown in figure 18 for 75 per cent
load. Those tests were carried out on a 7RTA84T
engine. The limiting factor for fuel/water emulsion is
the maximum delivery capacity of the fuel injection
pumps so that, in practice, the engine has either to be
derated or the maximum achievable reduction of NOX
limited to about 10 or 20 per cent. To obtain the
maximum NOX reduction also at full load, it may be
necessary to redesign not only the injection system
but also the camshaft with all the consequences in
the camshaft drive, etc. Another aspect of fuel/water
emulsion is that the injection nozzle design (hole dia-
meter, etc.) has to be adapted to the increased quan-
tity of liquid injected. With this modified nozzle
1.5° retardation
30% Water
2.3° retardation
40% Water
167 169 171 173
BSFC [g/kWh]
19
design, both fuel consumption and component tem-
peratures may be penalised when the engine is run-
ning without water. Additionally, it has to be consid-
ered that whereas heavy fuel oil and water can easily
be emulsified owing to the small difference in densi-
ties, emulsifying gas oil is only possible with the use
of an emulsifying agent, entailing additional costs.
Fuel nozzle
Water nozzle
Exhaust
3 Valve
2
1
Swirl
Fig. 19
Investigated water injector arrangements shown in a sche-
matic plan view of an RTA engine combustion chamber.
Locations 1 and 2 were tested on the 4RTX54 research
engine, while location 3 was investigated by computational
fluid dynamic (CFD) analysis
[98#143]
Direct injection
Owing to the above-mentioned disadvantages of
fuel/water emulsion, Wärtsilä NSD Switzerland start-
ed in 1993 the development of direct water injection
to achieve high NOX reduction rates. In this system,
the water is handled by a second, fully independent
injection system, preferably under electronic control.
This offers the possibilities firstly of injecting very
large amounts of water without the requirement to
derate the engine and secondly to inject the water
and fuel injection with different timing. The water can
be injected in parallel with the fuel and/or during the
compression stroke, so that optimising injection
timing, with respect to fuel and water consumptions,
NOX emissions and other emissions, such as hydro-
carbons and carbon monoxide, is possible without
influencing engine reliability. Independent injection
systems also allow water injection to be switched on
and off without influencing the fuel injection.
An early, detailed investigation, which included
computational fluid dynamic (CFD) analysis, showed
that emphasis has to be given to the location of the
water injector in the combustion chamber. On the test
bed, two solutions were investigated (Fig. 19). One
solution was with a tandem nozzle (Fig. 21) in which
the fuel injector and water injector are combined in
one nozzle body but the system for controlling injec-
tion timing of the two injectors is completely separate
(arrangement 1). The other solution was with a
completely separate system for injecting water with
separate fuel and water injectors (arrangement 2).
Arrangement 2 where the water injector is far from
the fuel injector (Fig. 19) clearly shows inferior beha-

Fig. 20
Measured NOX reductions
NOx [g/kWh] with respect to water
18 consumption, corresponding
to water injector arrange-
ments 1 and 2 in figure 19
[98#144]
15
Arrangement 2

12
Arrangement 1

6
0 30 60 90 120 150 180 210

Water consumption [g/kWh]

20
 viour in terms of the ratio of water quantity required
for the NOX reduction obtained (Fig. 20).
Based on the NOX level of the 4RTX54 research
engine tuned for low NOX emissions, reductions of
more than 60 per cent of the NOX emissions were
achieved with arrangement 1 by the combination of
water injection and retarded fuel injection when
injecting approximately 140 g/kWh of water (Fig. 22).
Theoretical investigations indicate that the entire
quantity of injected water evaporates if it is injected
after 60 degrees crank angle before top dead centre.
During the test period with such timing, no corrosion
on the liner surface occurred, even when operating
the engine on high-sulphur heavy fuel oil. This early
injection gives the advantage that fuel consumption
is only insignificantly increased, and can even be
improved.
The best congruence of fuel and water sprays is
achieved when the water is injected in parallel with
the fuel. However, injecting the water in this way does
not optimally correspond with the recognition that
most of the NOX is formed during the first quarter of
combustion [17]. For this reason, it is necessary to
have injected part of the water already at the begin-
ning of the combustion period. Injecting water at the
end or even after combustion is useless. Indeed the
results in figure 23 show that the timing of water
injection influences the increase in fuel consumption.
The later that the water is injected, the greater is the
increase in fuel consumption [23].
Fig. 21 After having performed many hours on the engine
Tandem nozzle for injecting fuel and water separately into the test bed to optimise the efficiency of the water injec-
cylinder of a two-stroke engine
[98#154] tion system, the next stage is to apply water injection
to a shipboard engine to confirm its long-term reliab-
ility.

Fig. 22
Reduction in NOX emissions
measured on the 4RTX54 NOx, g/kWh ∆ BSFC [%]
research engine by tuning 24.1 105%
alone, adapted tuning with
water injection (arrangement BSFC 100%
20.5
1 in figure 19), and together
with reduced maximum
cylinder pressure, also 16.9 NOx 95%
showing in each case the
corresponding effect on fuel 13.3 90%
consumption
[98#145] 9.6 85%

6.0 80%

2.4 75%
Reference tuned to low NOx tuned to low NOx tuned to low NOx
+ 140 g/kWh water + 140 g/kWh water
Injection Injection
+ Pmax reduced

21
NOX reduction by exhaust gas The application of EGR to turbocharged two-
recirculation (EGR) stroke diesel engines is more complicated than in
automotive engines. In turbocharged two-stroke
Recirculating part of the exhaust gas is an efficient
engines, the exhaust gas pressure before turbine is
method for reducing NOX emissions. Yet currently it is
always lower than the scavenge air pressure after
only applied in some car and truck engines owing to
compressor. Thus it is not possible simply to recircul-
significant drawbacks associated with the cooling
ate exhaust gas on the high pressure side by a simple
and cleaning of the exhaust gas before it is circulated
pipe equipped with a control valve as normally done
back to the engine cylinders.
in automotive engines. Instead an additional exhaust
The tendency of exhaust gas recirculation to redu-
gas compressor has to be installed. Even if new
ce the formation of NOX is based on three factors:
research [18] shows that it is technically feasible to
• The reduced oxygen concentration obtained by
provide the additional compressor, we feel quite sure
EGR in the combustion air increases the quantity
that it will be rather impracticable when the engine is
of gas that has to be heated for combustion of the
run on heavy fuel oil.
fuel. The resulting lower adiabatic flame tempera-
On the other hand, other two-stroke engine de-
ture reduces the rate of NOX formation. As the
signers have, for test purposes, applied EGR on the
average temperature is not influenced when apply-
low-pressure side by recirculating exhaust gas from
ing cooled EGR, the efficiency of the engine does
after the turbine to before the compressor. However,
not deteriorate when the negative effects of the
when applying this method, the recirculated exhaust
lower oxygen concentration on the combustion
gas has to be led through the compressor and the
process are disregarded.
scavenge air cooler, both equipment being very sen-
• The reduced oxygen concentration also dimini-
sitive to fouling and acidic corrosion.
shes the reaction between nitrogen and oxygen to
The high concentration of sulphur oxides and par-
form NO, therefore reducing NOX emissions.
ticulates in the exhaust gas can cause corrosion and
• The inert compounds (such as H2O and CO2)
fouling problems with EGR unless taken fully into
recirculated to the engine cylinder have to be hea-
consideration from the beginning. Even with water
ted up during the combustion. At elevated tempe-
washing of the exhaust gas to be recirculated, the
ratures, the recirculated three-atomic compounds
deposition of particles, unburned hydrocarbons, sul-
H2O and CO2 have an approximately 25 per cent
phuric acid, etc., can never be avoided. In addition,
greater heat capacity than air which comprises
there is the problem of disposing of the acidic, dirty
two-atomic elements, mainly oxygen and nitrogen.
washings.
This leads to an increase of the overall heat capa-
In effect, conventional EGR can only be applied in
city of one to two per cent and therefore to a
combination with ‘clean’ fuels, for example natural
further reduction in local peak temperatures and
gas or gas oil. In this respect, some results of EGR
thus NOx emissions.

Fig. 23
Measured influence
Start of Water Injection [°CA] ∆ BSFC [g/kWh] of water injection timing
on the reduction of NOX
-80 -60 -40 -20 0 20 emissions and on the fuel
consumption
[98#146]

25%
water
50%
water
75%
water
100% -80 -60 -40 -20 0 20
water
∆ NOx [%]
Start of Water Injection [°CA]

22
 Fig. 24
Influence of the rate of
exhaust gas recirculation [g/kWh] NOx CO HC ∆ BSFC
(EGR) on various exhaust 10
gas emissions
and fuel consumption,
as measured on the ZA40S 8
high-pressure gas engine
[98#147]
6

0
0 13 23 33 43 0 13 23 33 43 0 13 23 33 43 0 13 23 33 43

Recirculation [%] Recirculation [%] Recirculation [%] Recirculation [%]

tests made with the ZA40S high-pressure gas engine
are shown in figure 24. The maximum NOX reduction
achieved was 85 per cent when recirculating 43 per
cent of the exhaust gas.
To avoid the above-mentioned fouling and corro-
sion problems, the most straightforward solution is
not to recirculate the exhaust gas outside the engine
cylinder but instead to recirculate it internally by
modifying the engine scavenging process so that the
purity of the gas inside the cylinder decreases. When
doing this in a simple way by reducing the size of the
turbocharger without any additional cooling inside
the combustion chamber, it is possible to gain half of
the potential of this measure. Tests made on the
4RTX54 research engine have shown that this mea-
sure has a considerable potential to reduce NOX
emissions. Lowering the purity by approximately six
per cent during the test period resulted in a reduction
of NOX emissions of 22 per cent (Fig. 25).
Owing to the increase in thermal load of the engine
with internal recirculation, we are working on a con-
cept to reduce NOX emissions by combining water
injection and internal exhaust gas recirculation [19].
By injecting certain quantities of water during the

Fig. 25
NOX reduction versus
deterioration in mass purity NOx [%]
as measured on the
4RTX54 research engine 100%
[98#148]
90%

80%

70%

60%

50%

0% -1% -2% -3% -4% -5% -6%

∆purity [%]

23
Fig. 26
Three fuel injection nozzles with different sac-hole volumes. Left is the ‘standard’ nozzle of the 4RTX54, centre is a nozzle with
reduced sac-hole volume, and right is another in which the sac-hole volume is reduced to virtually zero
[98#149]

compression stroke, the temperature level inside the
combustion chamber can be reduced down to the
original temperature level that would be reached
without internal EGR, so that firstly the thermal load
on the engine is diminished and secondly the NOX
emissions decrease [24].
It has to be noted that it is not meaningful to com-
bine internal EGR with fuel/water emulsion because
that combination will not effect an additional cooling
of the cylinder content during the compression stroke.

Fig. 27
Influence of the sac-hole
volume in the fuel injection
Dust [g/kWh] 100% load 50% load
nozzle on the dust
1.0 emissions, as measured on
the 4RTX54 research engine
running on heavy fuel oil
[98#150]
0.8

0.6

0.4

0.2
Standard Reduced Zero Standard Reduced Zero
Injector Sac-Hole Sac-Hole Injector Sac-Hole Sac-Hole

24
Hydrocarbon and particulate reduction
technologies
The optimisation of the diesel process for low fuel
consumption in past years has led to low hydrocar-
bon emissions, with incomplete combustion being
brought almost down to zero. In well-maintained mo-
dern diesel engines, a high proportion of the hydro-
carbon emissions are caused by losses of lubricating
oil and evaporation of fuel out of the sac hole in the
fuel injection nozzles. While the lubrication oil con-
sumption can hardly be further reduced, the sac-hole
volume can be minimised. For engines with off-centre
injectors, however, commercially-applicable nozzles
with zero sac-hole volume are still under develop-
ment. Several variants of nozzles with varying sac-
hole volumes have been tested on the 4RTX54 re-
search engine. Three examples are shown in figure 26
in which the first fuel injector shows the ‘standard’
fuel injector of the 4RTX54, the second shows a fuel
injector with a reduced sac-hole volume, and the third
shows a fuel injector with the sac-hole volume redu-
ced virtually down to zero. The measured dust emis-
sions of each nozzle are shown in figure 27.
The influence of the fuel injection parameters on
reducing exhaust emissions was investigated during
an extensive test programme with the 8ZA40S res-
earch engine that focused on reduced exhaust emis-
sions in general. By modifying fuel injection parame-
ters, such as injection duration, injection pressure,
etc., the dust and hydrocarbons emissions vary but it
could be shown that the standard fuel spray pattern
of the ZA40S engine type represents an optimum. All
variations showed increased hydrocarbon and parti-
culate emissions (Fig. 28). However, when reducing
the sac-hole volume down to zero (seat-hole nozzle),
similar to the tests on the 4RTX54, both hydrocarbon
and particulate emissions could be significantly
reduced.

Fig. 28
Influence of fuel injection
parameters on hydrocarbon HC [g/kWh]
(HC) emissions as measured
0.6
on the ZA40S engine running
on marine diesel oil
[98#151] 100% load 50% load
0.45

0.3

0.15

0
Short injection

Short injection
Flat fuel spray

Flat fuel spray

High injection
High injection

High opening

High opening
Reference

Reference
Seat hole

Seat hole
pressure

pressure

pressure

pressure
duration

duration
pattern

pattern
nozzle

nozzle

25
Secondary measures (exhaust gas aftertreatment)
If primary countermeasures for exhaust emissions
are inadequate for reaching the limits set by the
legislation concerned, then it is necessary to apply
external treatment of exhaust gases after they have
left the engine cylinders. The devices available for this
purpose include catalytic converters (oxidising cata-
lysts for hydrocarbons and carbon monoxide, and
reducing catalysts for NOX), thermal reactors (for
hydrocarbons and carbon monoxide), and traps or
filters for particulates.
In contrast to spark-ignition engines operating
stoichiometrically, diesel engines always operate lean
as their load is controlled by the amount of fuel inject-
ed. Diesel exhaust gas therefore has a substantial
oxygen content (about 13 per cent) and is at a low
temperature (250° C to 400° C). Because of the high
oxygen content, the simultaneous removal of NOX,
hydrocarbons and carbon monoxide by employing
three-way catalysts is not possible. Other solutions
have to be applied.
Selective catalytic reduction
Theoretically there are various ways to decompose
NOX in diesel exhaust gas. However, research
during the past two decades has shown that selec-
tive catalytic reduction (SCR) is the only practicable
way to cut NOX emissions of diesel engines. The
overall SCR reactions can be written as*:
4NO + 4NH3 + O2 ➞ 4N2 + 6H2O (5)
2NO2 + 4NH3 + O2 ➞ 3N2 + 6H2O (6)
Theoretical work on SCR dates back to the 1960s
and the first commercial applications were made in
the 1970s, mainly in boiler plants. The introduction of
SCR to diesel engines came considerably later, pre-
sumably because no legislation was calling for such
technology. The situation changed in the 1980s,
making it necessary for all parties involved in the die-
sel engine industry to look into SCR. As the basic
SCR technology is already known (Fig. 29), some die-
sel engine manufacturers are developing their own
SCR designs whereas others are co-operating with
various independent SCR suppliers.
The application of SCR on large diesel engines,
however, still needs special care owing to the pecu-
liarities of these engines with regard to the fuel and
* If urea is used as the ammonium carrier, it decomposes at
temperatures above 160°C before entering the catalyst
according to the following process:
CO(NH2)2 + 6H2O ➞ CO2 + 2NH3
26
lubricants employed, type of operation, etc. Eventu-
ally, that led to a joint research programme by Wärt-
silä NSD Switzerland Ltd together with Mobil Oil Cor-
poration and Lloyd’s Register of Shipping [20]. In this
programme, SCR technology was tested on a Sulzer
6RTA38 two-stroke engine (Fig. 30). The engine ran
for up to 7500 hours a year on heavy fuel oil with an
average sulphur content of 3.5 per cent. It was locat-
ed at Mobil’s research centre in their refinery at
Notre-Dame-de-Gravenchon in France. The SCR
plant (including the catalyst) was designed by Sulzer
Chemtech Ltd and offered the possibility to vary the
1
space velocity between 0 and 25 000 h- , the tempe-
rature between 250 and 450°C, and the pressure bet-
ween 1.0 and 3.5 bar. A solution of urea in water was
used as the reducing agent. The research programme
was focused on fuel and lubricating oil related
aspects. The consortium was privileged to get sup-
port also from the National Energy Research Founda-
tion (NEFF) in Switzerland.
The first question to be answered when conside-
ring the installation of an SCR system on a diesel en-
gine is, of course, the positioning of the catalyst. The
correct location of the catalyst is determined by its
optimum working temperature, which is around
350° C for continuous operation. Hence, with a low-
speed two-stroke engine, the catalyst is preferably
installed before the turbocharger turbine (Fig. 31)
whereas, with a four-stroke engine, it is always
mounted downstream of the turbine. However, it is
also feasible to position the SCR unit after the turbine
Fig. 29
Honeycomb catalyst from Hug Engineering for use in
an SCR converter
[98#152]

Fig. 30
Sulzer 6RTA38 two-stroke engine at Mobil’s Gravenchon
refinery, France. The engine of 4080 kW at 200 rev/min is
fitted with a selective catalytic reduction (SCR) converter, and
is used for research into the aftertreatment of exhaust emissi-
ons and for lubricants
[7792-3003-1]
with a two-stroke engine, as installed in the Diesel
Technology Center in Winterthur (Fig. 32) but it then
requires an afterburner to heat the exhaust gases to
the appropriate working temperature for the catalyst.
The latter arrangement was selected for the DTC
because we first wanted to keep the 4RTA58T
research engine as close as possible to a standard
marine engine. Secondly, the cost for the fuel used in
the afterburner remains acceptable because of the
limited number of running hours compared with a sta-
tionary power plant. The higher gas pressure when
the SCR unit is arranged before the turbine, however
Fig. 31
Example of how catalyst
units can be fitted in a
marine power plant. This
is an ideal arrangement
with respect to gas flow.
Other arrangements are
possible
[98#153]
gives the great advantage of cutting the size of the
catalyst by some two thirds. Although the location
before the turbine results in a more complicated
design, the smaller size results in the overall invest-
ment costs tending to be no higher than for installa-
tion after the turbine.
To overcome the disadvantage of the large size for
SCR units installed after the turbine, Wärtsilä NSD
has developed the ‘Compact SCR’, which combines
an SCR unit and silencer, together with built-in soot
blowers.
The catalysts employed for SCR units are com-
monly vanadium pentoxide embedded in titanium
dioxide, and additionally are often dosed with tung-
sten trioxide and molybdenum trioxide to optimise
the catalytic properties. Such catalysts (Fig. 29) are
termed ‘full-contact catalysts’, in contrast to ‘coated
catalysts’ in which a porous carrier material is coated
with the catalytic material.
Essentially, many chemicals could be used as the
reducing agent. For various practical reasons, how-
ever, ammonia NH3, and urea CO(NH2)2, have turned
out to be the only commercially-applicable reducing
agents, despite the difficulties involved in their use.
Ammonia is the least desirable because of its delic-
ate handling requirements. When using urea, it first
has to decompose to ammonia before it enters the
catalyst. This requires an additional two to three me-
tres in the length of exhaust pipe before the catalyst.
At exhaust gas temperatures below some 300° C (the
exact value is very much dependent upon the partial
pressures of ammonia and sulphur trioxide in the ex-
haust gas, as well as the porosity of the catalyst sur-
27
face), ammonia combines with the sulphur trioxide to
form ammonium sulphate. Ammonium sulphate is an
adhesive and corrosive substance, and thus causes
serious fouling if the catalyst is operated at inappro-
priate temperatures. In addition, at temperatures
above 500° C, ammonia starts to burn, and thereby
loses its property as a reducing agent. These two
effects define the temperature window in which the
SCR plant can be operated to the region of 320–
480°C.
Control of an SCR plant is achieved by regulating
the urea dosage rate. This can basically be done in
two ways. The NOX level after the catalyst can be
measured on-line, and the urea dosage is then regu-
lated as a function of this measuring result (feedback
control). Alternatively, a typical NOX emission profile
(as a function of engine speed and power) is stored in
a computer and the urea dosage rate is adjusted as a
function of the NOX reduction being sought (feed-for-
ward control). While the latter approach is obviously
the more convenient, the first offers far more precise
Fig. 32
Exhaust stack on the Diesel Technology Center in Winterthur
showing the catalyst units for NOX reduction
[7795-3023-1]
28
control of the SCR plant. For example, the urea injec-
tion rate automatically adapts to fuel oil quality, which
is frequently not known, and the ambient conditions.
It is commonly accepted that, for NOX reductions
greater than 90 per cent and for operation dominated
by transient load changes, urea injection has to be
controlled by feedback whereas, for all other applica-
tions, a feed-forward control can be applied. For ex-
ample, the SCR plant supplied by Hug Engineering
fitted to the Sulzer 9S20 four-stroke engine in the
Swiss Federal Institute of Technology, ETH Zürich
(Fig. 33), is feedback controlled and achieves a NOX
reduction of 99.8 per cent.
Obviously, lower fuel oil quality (for example, in
terms of sulphur content) imposes a correspondingly
greater burden on the catalyst, especially with res-
pect to the concentrations of particulates and SOX in
the exhaust gases. These problems are usually
solved by choosing a catalyst geometry with a low
cell density to prevent clogging of the catalyst. Addi-
tionally, when using low-quality fuels, the catalyst
may suffer from hydrocarbon deposits, leading to a
physical clogging of the chemically-active surface.
However, this phenomenon only occurs at relatively
low temperatures below about 400° C and is revers-
ible. In four-stroke engines, this can be avoided by
employing a small burner to heat the exhaust gas up
to 420° C for one hour in every 24 hours of operation,
as first applied by Hug Engineering. In two-stroke
engines with the catalyst installed before turbine, this
high temperature can easily be obtained by loading
the engine so that no burner is needed. In a long-term
test, soot blowers have proved to be effective in pre-
venting the clogging of the catalyst by dust or parti-
culates.
After the research programme on the 6RTA38 at
Gravenchon, the following can be described as the
state of the art for SCR systems [21]:
• With an SCR system, NOX reductions greater than
90 per cent can be commercially achieved.
• Generally, the price for an SCR system is today
(mid 1998) around US$ 75 000 per MW engine
power (with great variations in price). The costs for
the catalytic material itself make up only about ten
per cent of the total price.
• The running costs (basically for the urea) are
around 3 US$ per MWh.
• The maintenance costs (approximately half of it for
replacement of the catalyst) amount to around
2 US$ per MWh.
• The catalyst housing (reactor) including insulation,
etc., has a volume of 2 to 5 m3 per MW engine
power (depending on the make of catalyst). The
 Fig. 33
Sulzer 9S20 diesel
generating set in a cogene-
ration plant at the ETH
Zürich is equipped with an
SCR unit that reduces NOX
emissions by 99.8 per cent.
Installed in 1994, the engine
develops 1440 kW at
1000 rev/min. The SCR is
the rectangular unit in the
background while the large
cylindrical unit overhead is
the exhaust silencer.
[7794-1001]

size of the catalyst housing is more or less inde- Exhaust gas desulphurisation
pendent of the input NOX concentration.
Although reducing the emissions of sulphur oxides
• The exhaust-gas back pressure imposed by an
(SOX) is technically straightforward and is used in
SCR plant is typically between 15 and 25 mbar.
thermal power stations, it has not yet been applied on
• For a two-stroke engine, the catalyst has to be in-
board ships because desulphurisation plants are
stalled either up- or downstream of the turbochar-
bulky and expensive. It is widely accepted that, for
ger turbine, depending on the operating character-
ships, it is easier and less expensive to reduce the
istic whereas, for a four-stroke engine, the catalyst
sulphur content of the fuel than reducing the sulphur
reactor ought to be placed somewhere between
oxides in the exhaust gas. Indeed, that is the basis for
the turbine and the exhaust-gas economiser.
the forthcoming IMO regulations summarised above.
• The temperature drop in an SCR plant (owing to
However, the desulphurisation technique or DeSOX
heat losses and the heat capacity of the injected
as it is colloquially known is outlined below as back-
urea solution) is generally small, of around 10° C.
ground information.
However, it has to be considered that the dew
As sulphur oxides do not originate in the constitu-
point of sulphuric acid is also increased by rough-
ents of the combustion air but from the fuel-borne
ly 10° C through the partial conversion of SO2 to
sulphur, it is not possible to reduce them by means of
SO3 in the catalyst.
a catalyst. They can only be washed out. This is nor-
• The catalyst reactor can be designed in a way that
mally done by lime milk or other alkaline solutions,
it serves as a silencer with achievable noise reduc-
such as sodium carbonate or caustic soda. As an ex-
tions in the order of more than 25 dBA.
ample, the overall chemical reactions when using
• Approximately 30 litres of 40 per cent urea solution
lime milk (solution of calcium carbonate in water) are:
(corresponding to about 15 kg of urea granulates)
are needed per MWh.
CaCO3 + SO3 ➞ CaSO4 + CO2 (7)
CaCO3 + SO2 + 1/2 O2 ➞ CaSO4 + CO2 (8)
• If the SCR plant is only for temporary use, a burner
In this case, lime milk is used as a washing agent
is absolutely necessary to heat the catalyst before
and reacts with the SO2/SO3 to yield gypsum as a
the engine is started. Otherwise the catalyst will be
final product. All other DeSOX processes also lead to
inevitably clogged by ammonium sulphates.
a waste product thereby ultimately entailing a dispo-
sal problem. Disposal of the gypsum itself generally
does not pose any problems, but the contamination
of the gypsum with particulates and heavy metals

29
that are washed out from the exhaust gas leads to
discolouring of the gypsum.
Several organisations have been investigating the
use of sea water to scrub the exhaust gas and there-
by overcome this disposal problem in marine appli-
cations [22]. Although the removed oxides of sulphur
are neutralised by the calcium which is always pre-
sent in sea water, there is a big question mark as to
whether the direct dumping of SO2 into the sea would
be acceptable. Therefore, the most applicable me-
thod for reducing the emissions of sulphur oxides of
ships is still to reduce the sulphur content of the
fuel.
Particulate trapping
Trapping particulates in an exhaust gas filter is a
great challenge. It is not so far needed but it may
need to be done at some time in the future. Bearing in
mind that the particles are typically between 0.1 and
1.0 micron in size, it can easily be found impossible to
remove all particulates by centrifugal forces as, for
example, in a cyclone. On the other hand, although
monolithic ceramic particle traps have gained a cer-
tain popularity for small engines, they do not offer an
alternative for large diesel engines. Among other
deficiencies, ceramic particle traps cannot be rege-
nerated as the particle deposits (based on sulphur
and calcium) cannot be oxidised.
The above problems can be overcome by using
wet electrostatic precipitators. In these, the particles
are ionised thereby making them stick to the electro-
des. Eventually the dust can be washed out. Never-
theless this kind of installation requires a lot of space,
which is certainly not available onboard ship.
As a result of the above-mentioned disadvantages,
it was decided to investigate an integrated ‘electro-
cyclone’ which combines the advantages of both the
cyclone and the electrostatic precipitators. The ‘elec-
tro-cyclone’ was developed by Hosokawa Micropul
in Japan. A prototype of this equipment has been
installed with the 4RTA58T research engine in the
Diesel Technology Center. The first results with the
engine running on heavy fuel oil showed the potential
of the design with more than 60 per cent reduction in
particles being measured.
ter in an oxidation catalyst. Generally, the oxidation
catalyst is applied as an additional layer in the hous-
ing of the SCR unit. Noble metals such as platinum
and palladium are most suitable as catalytic material
because of their intrinsically high activity. The detai-
led composition of the catalytic material has to be ad-
justed to the substances to be oxidised, and to take
into consideration potential poisoning by lubricating
oil additives. For example, the specific activity of pal-
ladium is more suitable than that of platinum for the
oxidation of carbon monoxide, olefins and methane,
whereas platinum is more active than palladium for
the oxidation of paraffin hydrocarbons with molecular
sizes greater than C3 [9]. Since the catalytic beha-
viour is manifested exclusively by surface atoms, the
noble metals are dispersed as finely as possible on an
inert support such as g-Al2O3 (but which is not suita-
ble for exhaust gases carrying SO2) or ceramics
which prevent particle-to-particle metal contact,
otherwise they would sinter thereby inevitably leading
to a reduction of the active surface area.
The achievable efficiency of an oxidation catalyst
is typically some 50 to 80 per cent, depending on the
catalyst size and on the fuel oil quality.
In addition to the functions described above,
oxidation catalysts are often used to control the po-
tential ammonia slip in SCR plants. However, when
an oxidation catalyst should be applied it has to be
considered that it may lead to a significant conver-
sion from SO2 to SO3, potentially leading to addit-
ional corrosion in the exhaust gas economiser.

Post oxidation of exhaust gas (oxidation

catalysts)
Unburned or partially burned components of the
exhaust gas, such as carbon monoxide and hydro-
carbons, can be oxidised to carbon dioxide and wa-

30
Conclusion
Owing to the work begun in the early 1970s, it is to-
day possible to reduce the exhaust emissions of
Sulzer diesel engines to meet the forthcoming ex-
haust-gas emission regulations set by the IMO. In ad-
dition, development work and practical experience
has shown that there is sufficient scope to achieve
further reductions in emissions to comply also with
any further tightening of regulations.
• The first choice is the introduction of engine tuning
modifications that will achieve up to 30 per cent
reduction in NOX emission levels from those of
standard engines from 1990.
• For further reduction, separate direct water injec-
tion is considered to be the most appropriate solu-
tion. On the test bed, this technique has proved its
ability to reduce NOX levels by some 60 per cent
from those of today’s standard engines. Thus pro-
ven primary measures are available for considera-
ble reductions in NOX emission levels.
• To reduce NOX by 90 per cent or more, aftertreat-
ment by selective catalytic reduction (SCR) has
proved to be an effective solution despite the spe-
cial difficulties imposed by using high-sulphur fuel
oils.
• Lower emissions of sulphur oxides (SOX) are ex-
pected to be met by using fuel oils having reduced
sulphur content. Although the reduction of SOX
emissions by exhaust-gas aftertreatment is techni-
cally feasible, the desulphurisation process inevi-
tably imposes a disposal problem which would not
be acceptable for ships.
• Carbon monoxides and hydrocarbons can, if
necessary, be reduced by an oxidation catalyst
housed within an SCR reactor.
• Particulate reduction, when running on heavy fuel
oil, still poses a great challenge. Technical solu-
tions are available (for example, wet electrostatic
precipitators), but involve either great space requi-
rements or great expense. However, particulate
emissions are reduced by 50 to 90 per cent when
switching to distillate fuel oils.
References
1. ‘Luftreinhalte-Verordnung (LRV), 814.318.142.1’,
16 December 1985, status 1 January 1995.
2. Lloyd‘s Register, ‘Marine Exhaust Emissions Re-
search Programme’, 1995
3. OCIMF Submission to IMO, ‘CONCAWE study of
the contribution of SO2 emissions from ships to
overall deposition and air quality in the Channel
zone area’, BCH 24/7/13, 24 June 1994.
4. CIMAC: ‘Exhaust emissions measurement; re-
commendations for reciprocating engines and
gas turbines’, Number 12, 1991.
5. ISO 8178: ‘Reciprocating internal combustion
engines – Exhaust emissions measurement –
Part 1: Testbed measurement of gaseous and
particulate emissions.’
6. ISO 9096: ‘Stationary source emissions; deter-
mination of concentration and mass flow rate of
particulate material in gas-carrying ducts; ma-
nual gravimetric method.’
7. S. G. Dexter, ‘Residual fuels – combustion, emis-
sions, and particulates’, IMechE Seminar ‘Diesel
fuel injection systems’, Birmingham, October
1989.
8. F. Fleischer, on behalf of EUROMOT, ‘NOX Re-
duction – A technical challenge for marine diesel
engine manufacturers’, International Maritime
and Shipping (IMAS) Conference, London, 22–
24 October 1996.
9. J. B. Heywood, Internal combustion engine fun-
damentals, Massachusetts Institute of Technolo-
gy MIT, McGraw-Hill, 1988.
10. J. Zeldovich, ‘The oxidation of nitrogen in com-
bustion and explosions’, Acta psysicochim 21,
USSR, 1946.
11. H. Kremer, ‘NOX-Entstehung in Feuerungen’,
TB310, VGB Kraftwerkstechnik, Essen, 1984.
12. A.A. Amsden, P.J. Rourke and T.D. Butler, ‘KIVA
II: A Computer Program for Chemically Reactive
Flows with Sprays’, Technical Report LA-11560-
MS, Los Alamos National Laboratory, May 1989.
13. K. Boulouchos, B. Schneider, F.X. Tanner and
G. Weisser, ‘Experimentally validated atomiz-
ation and ignition models for diesel engine com-
bustion simulations’, ILASS-Europe 1997.
14. H. Stebler, G. Weisser, H.-U. Hörler and K.
Boulouchos, ‘Reduction of NOX emissions of D.I.
diesel engines by application of the Miller sys-
tem: An experimental and numerical investiga-
tion’, SAE Paper 960844.
15. J. Vollenweider, M. Geist and M. Schaub, ‘Resi-
dual fuels in emission-controlled diesel engines -
Background, developments and operational re-
sults’, CIMAC 1995, Interlaken.
31
16. A. Velji, W. Remmels, R.-M. Schmidt, ‘Water to
reduce NOX emissions in diesel engines, a basic
study’, CIMAC 1995, Interlaken.
17. S. Papadopoulos, ‘Reduktion der Stickoxid-
emissionen des direkteinspritzenden Dieselmo-
tors durch Dieselölwasseremulsionen bzw. Was-
sereinspritzung’, Dissertation ETH Zürich, 1987.
18. ABB, ‘Verfahren und Vorrichtung zur hochdruck-
seitigen Abgasrezirkulation an einer aufgela-
denen Brennkraftmaschine’, Patent pending,
1994.
19. Wärtsilä NSD Switzerland Ltd, ‘Verfahren zum
Vermindern der Stickoxydmenge im Abgas eines
Zweitakt-Grossdieselmotors und Motor zum
Durchführen des Verfahrens’, Patent, 1993.
20. J. Vollenweider, A Eyres (Mobil) and R.P. Hol-
brook (Lloyd‘s Register), ‘SCR research on a
two-stroke HFO engine’, IMarE, London, 1994.
21. J. Vollenweider, ‘SCR-Konzept für Grossdiesel-
motoren’, Abschlussbericht des NEFF-Projektes
Nr. 522, 1995.
22. J. Makansi, ‘Special Report: Controlling SO2
emissions’, Power, March 1993.
23. M. Geist, ‘The harmonisation of a broad spec-
trum engine program with prospective IMO regu-
lations’, MARIENV ’95, Tokyo, 1995.
24. M. Geist and R. Holtbecker, ‘Low and medium
speed diesel engines, compliance with prospec-
tive emission regulations’, ASME 1995, Techni-
cal Fall Conference, Milwaukee, 1995.

32
23.48.07.40-Printed in Switzerland