Experimental Investigation of the Effects of Water

Addition on the Exhaust Emissions of a Naturally

Aspirated, Liquefied-Petroleum-Gas-Fueled Engine
Hakan O zcan and M. S. Soylemez*
Department of Mechanical Engineering, Faculty of Engineering, University of Gaziantep,
27310 Gaziantep, Turkey
Received June 25, 2004. Revised Manuscript Received March 8, 2005
The purpose of this study is to evaluate the effect of water addition on combustion in a
conventional SI engine. The manifold induction method is used for water addition in this study.
The exhaust emissions, ignition timing, and exhaust temperature values were measured for
different equivalence ratio values by using a naturally aspirated liquefied-petroleum-gas-fueled
spark ignition, four-cylinder engine. The water induction is accomplished over a wide range of
water to fuel mass ratios of 0.2-0.5. The results showed that water addition worked as a cooling
mechanism for the fuel-air charge and slowing the burning rates, yielding a reduction of the
peak combustion temperature, which in turn provides a 35% reduction in peak NO
x
emissions
without any significant change in CO and HC emissions. In addition, greater ignition advance is
obtained.
1. Introduction
The concept of water addition as a supplement to the
internal combustion engine has been around for over
50 years.
1
It is a well-known fact that water does not
burn but it is excellent at absorbing heat due to water
having a high specific heat capacity and latent heat of
evaporation. The latent heat of evaporation of water is
2256 kJ/kg, which is approximately 6 times greater than
that for gasoline under standard atmospheric pressure
and temperature.
Water addition, as a separate liquid or emulsion with
fuel for automobile engines, has been investigated and
reported in published papers extensively. These inves-
tigations are generally related to water effects on engine
performance, knock, and emissions.
Lestz et al.
2
showed that the NO
x
concentration
decreased while HC and CO emissions increased with
water injection for a diesel engine. Harrington
3
tested
a single-cylinder engine, and concluded that engines
could be calibrated to operate with small amounts of
water to gasoline and by doing this knock could be
suppressed, hydrocarbon emissions would slightly in-
crease, NO
x
emissions would decrease, CO would not
change significantly, and fuel and so energy consump-
tion would be increased. Peters and Stebar
4
investigated
the effect of water-gasoline fuels on spark ignition
engine emissions and performance. They showed that
40% of water addition by weight to the fuel produced
about a 40% drop in the peak nitric oxide emission level.
Conversely, direct manifold water injection in amounts
equal to the fuel flow caused about a 50% increase in
HC emissions. In addition, their results showed that the
effect of water addition on carbon monoxide emissions
was small. Nicholls et al.
5
reported that dramatic
reductions of about 50% occurred in nitric oxide emis-
sions by the effect of water addition at a water/fuel mass
ratio of unity. Several different methods of water
addition have been developed for diesel engines.
6,7
These
studies have shown that further reduction of harmful
emissions is still possible. For example, Kahketsu et al.
6
investigated the direct injection of water with a new
designed stratified fuel-water injection system. They
applied this new system to an automotive diesel engine.
Their results showed that NO
x
emissions were reduced
by 50%. Kegl and Pehan
7
discussed some aspects of
injecting or adding water either into the intake air or
to the fuel of a diesel engine to reduce harmful emis-
sions.
Several authors studied the performances of water-
emulsified diesel and water-emulsified gasoline engines.
Park et al.
8
showed that water-emulsified diesel reduced
NO
x
and smoke at the same brake-specific fuel con-
sumption at high speeds after their experiments. Tsu-
kahara and Yoshimoto
9
reported the reduction of NO
x
with emulsified diesel fuel for a low compression ratio
diesel engine. Abu-Zaid
10
investigated the effect of water
emulsification on the performance and exhaust gas
temperature of a single-cylinder diesel engine. His
results indicate that the exhaust gas temperature
decreases as the percentage of water in the emulsion
increases. The engine tests do not yield homogeneous
results, owing to various typologies of the engine and
different fuels used for the experiments.
LPG (liquefied petroleum gas) has been widely used
in automobiles recently as an alternative to gasoline due
to its comparably lower price. In addition, it is well-
known that experts set their hopes and expectations on
this fuel because of its sufficient reserves in the world.
When environmental effects of LPG are taken into
consideration compared with those of liquid fuels,
significant improvements in exhaust emissions can be
achieved.
11
LPG-fueled spark ignition engines produce
virtually zero emissions of particulate matter, very little
carbon monoxide, and moderate hydrocarbon emis-
sions.
12
A major disadvantage of the LPG is the NO
x
emission is greater than that for liquid fuels.
13
In this work, an experimental investigation was
performed on a four-cylinder conventional spark ignition
engine using LPG fuel. The objective of the study is to
investigate the effect of water addition on the level of
exhaust emission and temperatures for a typical LPG-
fueled engine at different fuel/air equivalence ratios and
to investigate the occurrence of knock and misfiring
under different water/fuel mass ratios. The results
obtained in this study are compared with the general
results of exhaust gas recycling (EGR) in the literature.
2. Experimental Apparatus and Setup
Figure 1 shows a schematic diagram of the engine and test
setup that was used for the experiments. The principal
specifications of the water-cooled, four-cylinder SI engine of
the setup are listed in Table 1. LPG may be in the form of
propane (C3H8), butane (C4H10), or a mixture of both. An
ordinary LPG fuel (containing 30% propane and 70% butane)
was used as the test fuel.
The cylinder pressure was measured with a water-cooled
piezoelectric pressure transducer. The output signal of the
pressure transducer was amplified using a charge amplifier.
This amplified signal was transmitted to the data acquisition
systemand/or displayed on a DSO(digital storage oscilloscope).
Two magnetic proximity pickups were used to generate 0.25
(degrees of crank angle) incremental timing information,
faithfully track the crankshaft position, and trigger the
cylinder pressure recording. The cylinder pressures vs crank
angle were monitored on a DSO for observing any engine
instability at the beginning of each test. In addition, cylinder
pressures were monitored to detect the onset of knock and
measured as a function of the crank angle. The ignition timing
was varied over a wide range of water/fuel mass and equiva-
lence ratios. The ignition timing was monitored and measured
as a function of the crank angle by capacitive coupling pickups
with the high-voltage pulse associated with the spark event.
The airflow meter was used to determine the airflow rate
through the engine. Airflow was measured by using a precision
long-radius flow nozzle inserted into a pulse-damping drum.
All air which entered the engine was drawn through the
nozzle. The airflow rate was calculated directly by using a data
conversion formulation program and corrected for temperature
(measured with a thermometer in Kelvin) and pressure
(measured with a barometer in bars) in the test conditions to
obtain the actual airflow rate value during the experiments.
The discharge air temperature was measured by a T-type
thermocouple, and two precision pressure transducers were
used to measure the differential and absolute pressures. The
concentrations of NO
x, HC, and CO in the exhaust gas were
measured by the electrochemical sensors for the detection of
toxic gases at the ppm (parts per million) level except for
carbon monoxide, which was represented as a percentage of
volume (vol %). Overall, electrochemical sensors offered very
good performance for the routine monitoring of toxic gases and
(8) Park, J. W.; Huh, K. Y.; Lee, J. H. Reduction of NOX, Smoke,
BSFC with Optimal Injection Timing and Emulsion Ratio of Water-
emulsified Diesel. Prog. Inst. Mech. Eng. 2001, 215, Part D.
(9) Tsukahara, M.; Yoshimoto, Y. Reduction of NO
X, Smoke, BSFC
and Maximum Combustion Pressure by Low Compression Ratios in
Diesel Engine Fuelled by Emulsified Fuel; SAE Paper No. 920464;
Society of Automotive Engineers: Warrendale, PA, 1992.
(10) Abu-Zaid, M. Performance of Single Cylinder, Direct Injection
Diesel Engine Using Water Fuel Emulsions. Energy Convers. Manage.
2004, 45, 697-705.
(11) Bayraktar, H.; Durgun, O. Investigating the effects of LPG on
spark ignition engine combustion and performance. Energy Convers.
Manage., in press.
(12) Bass, E.; Bailey, B.; Jaeger, S. LPG conversion and HC
emissions speciation of a light-duty vehicle; SAE Paper No. 932745;
Society of Automotive Engineers: Warrendale, PA, 1993.
(13) Murillo, S.; Miguez, J. L.; Porteiro, J.; Gonzalez, L. M.; Granada,
E.; Moran, J. C. LPG: Pollutant emission and performance enhance-
ment for spark-ignition four strokes outboard engines. Appl. Therm.
Eng., in press.
Figure 1. Schematic of the engine test bed arrangement.
Table 1. Test Engine Specifications
fuel LPG
swept volume 1297 cm
3
compression ratio 7.8/1
maximum torque 12.5 kg m (DIN) at 3000 rpm
maximum power 70 bHP (DIN) at 5500 rpm
Water Addition Effects on an LPG-Fueled Engine Energy & Fuels, Vol. 19, No. 4, 2005 1469
oxygen. Measured data for emissions were sent into a com-
puter by using the computers serial port with a processor. The
condensation trap that was installed in the measurement line
removed the water/moisture. It was in-line with the sampling
hose and placed before the filter. This trap collected the water
and needed to be emptied after each set of experiments. The
engine torque was measured by a dynamometer. The instan-
taneous engine speed was measured by an inductive pickup
clamp. The exhaust gas temperature was measured using a
K-type thermocouple located downstream of the exhaust port.
The temperatures at selected points on the engine block were
measured by a T-type thermocouple. These temperatures were
used to detect the steady-state conditions at the start of the
experiments. Two electronic scales, which were interconnected
to the load cells, were employed to weigh the fuel and water
consumption during the engine operation continuously. Liquid
water was introduced into the intake manifold upstream of
the intake valve using a capillary tube with a 0.35 mm bore
diameter. Control over the water flow rate to the engine was
obtained by varying the water supply pressure and/or by
needle valve adjustments. This valve was adjusted for different
engine loads by means of a stepper-motor-controlled mecha-
nism. A computer drove this stepper motor. The water addition
rate was controllable either automatically by the same com-
puter, via the help of proper software, or manually. In addition,
acceleration of the test engine was controlled with a linkage,
actuated by a stepper motor. This stepper motor was also
driven with the same computer through its parallel port and
via software similar to that of the water addition mechanism.
All these variables were monitored and measured by using
three personal computers. Analog voltages representing tem-
peratures, airflow, and engine torque were conducted through
an analog-to-digital converter. Proper software was used to
store the acquired data and for their analysis. A multiport
serial card was used with its software for continuous measure-
ments of fuel and water consumption rates, rpm (revolutions
per minute), and different gas compositions in the exhaust.
This software included the correction for exhaust emission
readings. A high-speed data acquisition system was used to
display and store the cylinder pressures, crankshaft positions,
and spark event data. Complex software was used for the
display and analysis of the acquired data.
The engine data were collected at 2000 rpm, which was
selected as a moderate engine speed for internal combustion
engines. The flow rate of fuel was adjusted for each different
fuel/air equivalence ratio. At each equivalence ratio and water
to fuel mass ratio, the ignition timing was set at maximum
brake torque (MBT). At each operating point, once every 5 s
during a 4 min period, about 48 values of speed, torque, water
and fuel consumption rate, temperature, and exhaust emission
were recorded continuously. These measurements were con-
ducted under steady-state conditions, that is, when the change
of temperatures over the surface of the engine block was
measured as close to zero.
3. Experimental Errors
Table 2 lists the accuracy and resolution of the
instrumentation used in detail. A calibration check of
the devices was made two times, before and after each
successive test. It should be noted that all data collected
by the data acquisition systems used in this experimen-
tal study are subject to small errors. These errors are
on the order of 0.02% for a 12-bit system and 0.001%
for a 16-bit system and are considered to be negligibly
small when compared to the other sources of error.
4. Results and Discussion
The effects of manifold water induction on an LPG-
fueled SI engine on knock and misfiring limits and
exhaust temperature and emissions were searched at
different fuel/air equivalence ratios. Results show that
the water induction affects the information gathered
such as knock, misfiring, and exhaust emission and
temperature during the tests. Production of NO
x
de-
pends on the fuel/air equivalence ratio, maximum cycle
temperature, and burning rate. Figure 2 shows the NO
x
emission as a function of the fuel/air equivalence ratio
for different water to fuel mass ratios. Results indicate
that the peak NO
x
emissions occur at slightly lean
conditions, where the combustion temperature is high
and there is excessive oxygen to react with the nitrogen
as a result of the tendency of dissociation. Figure 2 also
indicates that the water injection reduces the NO
x
emission in the lean region having a local maximum
between equivalence ratios of 0.9 and 1.0. Because the
combustion process is closer to a stoichiometric ratio and
produces a higher flame temperature, the NO
x
emission
is increased, particularly by the increase of thermal
nitrogen oxide. Experimental results also show that a
maximum 35% reduction in the NO
x
emission level was
achieved with water injection. The reduction of the NO
x
emission level is evident in the lean mixture. The drop
in temperature and reduction of the combustion rate
with water addition are the main reasons for the NO
x
reduction. A basic thermodynamic analysis was per-
formed for this purpose. The effect of 0.5 g of water
addition/g of fuel yields a lower adiabatic flame tem-
perature at a level of approximately 150 K for a
stoichiometric fuel/air mixture. It can be readily shown
that, for the Zeldovich mechanism of a stoichiometric
mixture, a 100 K drop in the combustion temperature
would cause a 33% reduction in the NO production rate.
This result agreed with the measured trends in NO
x
.
Table 2. Accuracy of Instrumentation
no. measured value accuracy resolution
1 fuel flow rate 1% 0.02 g/s
2 air flow rate 0.75% 0.01 g/s
3 water flow rate 1% 0.02 g/s
4 temp 0.75%, (0.1 C
a
0.1 C
5 exhaust gas temp 0.75%, (2.2 C
a
1 C
6 torque 0.25%, 0.1 N m
a
0.01 N m
7 speed (1 rpm 1 rpm
8 ignition timing 0.5% 1
9 [NOx] 1%, 5 ppm
a
1 ppm
10 [CO] 1% 0.01%
11 [HC] (12 ppm 1 ppm
a
Whichever is greater.
Figure 2. Effect of water addition on NOx emissions as a
function of the fuel/air (F/A) equivalence ratio.
Several experimental results in the literature
14-16
showed that EGR lowers the NO
x
concentration in the
exhaust gas since the recirculated exhaust gas is mixed
with the fresh fuel-air mixture to dilute the concentra-
tion of fresh fuel-air charge.
The new mixture has a reduced oxygen concentration
due to a higher mean specific heat and higher temper-
ature compared to the air alone. This is like the effect
of water induction. Water is added into the intake port
and will produce a reduction in the peak flame temper-
ature similar to that of EGR, lowering the thermal NO
x
content. With EGR, a large reduction in NO
x
emissions
was obtained with a significant increase in HC emis-
sions. Substantial reductions in NO
x
concentrations are
achieved with 10-25% EGR.
14
This result agrees with
that of Woo et al.,
15
who worked on a modified com-
mercial heavy-duty diesel engine using LPG as a fuel.
Figure 3 shows the HC emission as a function of the
fuel/air equivalence ratio and water to fuel mass ratio.
HC emissions are not strongly affected near the stoi-
chiometric mixture with an increased water injection
rate. The HC emission slightly increases in the lean and
rich regions as the water to fuel mass ratio increases.
The higher heat of vaporization of water reduced the
temperature of the mixture. In addition, water addition
decreases the burning rate of the fuel-air mixture. The
lower mixture temperature, combustion chamber de-
posits, and longer burning period could have contributed
to higher HC emissions from the engine. EGR increases
the HC emissions. The increase of HC emissions may
reach over 60% at higher rates of EGR.
14
Woo et al.
15
also concluded in their experimental work that the NO
x
decrease caused HC to increase and the increase rate
in HC emissions was more moderate than the rate of
NO
x
reduction.
Results in Figure 4 explain the variation of the CO
concentrations as a function of the fuel/air equivalence
ratio and water to fuel mass ratio. For lean mixtures,
CO concentrations vary little with the equivalence ratio
and increase rapidly as the fuel-air mixture becomes
richer than the stoichiometric level. CO concentrations
in the exhaust increase as the fuel/air equivalence ratio
for fuel-rich mixtures increases in the amount of excess
fuel naturally. A reduction in the CO level of about 1%
is observed in the rich mixture region since a higher
CO concentration exists and water in combustion helps
to complete the combustion process by improving the
oxidation of CO. The combustion of CO occurs slowly
and yields a late burning in the combustion of the fuel.
The presence of water vapor in the combustion process
fundamentally changes the combustion chemistry. Specif-
ically, it does help the combustion of CO. EGR has little
effect on CO emissions.
14
The exhaust temperature significantly decreases for
the lean mixture as presented in Figure 5. This drop in
the exhaust temperature is relatively small for the rich
mixture. This is because the addition of water has
increased the total mass in the cylinder and water
absorbs a great deal of heat due to its high specific heat
similar to a supercharging and intercooling effect. Woo
et al.
15
show that EGR reduced the exhaust gas tem-
perature. This observation agreed with that of Caton.
16
He concluded that, as the homogeneous EGR rate
increased, the exhaust temperature decreased due to a
slower burning rate.
Figure 6 shows the MBT ignition timing for various
fuel/air equivalence ratios at different water addition
levels. The MBT ignition timing for an LPG-fueled
engine increases as the water addition increases. Water
addition to the reactant mixture decreases the burning
rate of the fuel-air mixture during combustion in a
(14) Mozafari, A. Exhaust Gas Recirculation in Spark Ignition
Engine. Adv. Heat Transfer ASME 1994, PD64 (1), 197-202.
(15) Woo, Y.; Yeom, K.; Bae, C.; Oh, S.; Kang, K. Effects of Stratified
EGR on the Performance of a Liquid-Phase LPG Injection Engine; SAE
Paper No. 2004-01-0982; Society of Automotive Engineers: Warren-
dale, PA, 2004.
(16) Caton, J. A. Effects of Burn Rate Parameters on Nitric Oxide
Emissions for a Spark Ignition Engine: Results from a Three Zone,
Thermodynamic Simulation; SAE Paper No. 2003-01-0720; Society of
Automotive Engineers: Warrendale, PA, 2003.
Figure 3. Effect of water addition on HC emissions as a
function of the equivalence ratio.
Figure 4. Effect of water addition on the CO concentration.
Figure 5. Effect of water addition on the exhaust tempera-
ture.
spark ignition engine. As a result of this, a minimum
MBT is advanced. The thermodynamic heat release
analysis showed that the burning period was decreased
in the case of stratified EGR. The combustion speed was
increased as the EGR rate was decreased. As the EGR
amount was increased, the fuel-air mixture was diluted
by the inert gas and the total heat capacity was
increased.
15
Therefore, the MBT spark advance require-
ment increases with the rate of EGR directly.
The experimental results in Figure 7 indicate that,
as the water to fuel mass ratio decreases, the knock and
misfire limits occur at less advanced ignition timing
since water induction leads to slowing of the burning
rate and increases the ignition delay period and com-
bustion period. In addition, the cooling effect due to
evaporation of water has a significant effect on the
reduction of the charge temperature in the cylinder. As
a result of these, water addition decreases the knock
tendency, which yields a desirable result for IC engines.
The range of ignition timing between knock and misfire
limits increases from 15 to 22 crank angle degrees when
the water to mass fuel ratio increases from 0 to 0.5.
For this study, it was observed that water input
caused a slight but sustained increase (0.2-2%) in the
thermal efficiency as compared to the no water addition
case. Figure 8 shows the effect of water addition on the
thermal efficiency of the engine for a fuel/air equivalence
ratio of 0.9. The results show that, as the water addition
level increases, the thermal efficiency increases. This
occurs due to formation of steam as the compression
stroke results in a pressure drop that causes a decrease
in the compression work. At the same time, an increase
in the thermal efficiency occurs as a result of the
lowered tendency of dissociation reactions as a result
of a decrease in the average cylinder temperature during
combustion. EGR also lowered the specific fuel con-
sumption and decreased the heat transfer from the
cylinder contents to the surrounding surface.
14
The
engine thermal efficiency increased due to the decrease
in the brake-specific fuel consumption with EGR.
5. Conclusions
The effect of water addition on exhaust emissions of
an LPG-fueled SI engine is investigated experimentally.
In the experimental measurements, water is added into
the fuel-air charge at the intake manifold with the
induction method, which is presumably the simplest and
most effective method. The following conclusions are
drawn from the experimental measurements to inves-
tigate the effect of water addition on exhaust emissions
with LPG fuels: (1) The CO concentration is slightly
affected by different water addition rates. This yields a
small drop in the rich fuel-air mixture region. In
addition, the CO concentration slightly increases with
water addition in the lean fuel-air mixture region. (2)
HC emissions increase with all water addition levels for
all equivalence ratios. (3) The experimentally observed
reduction in the NO
x
levels by water injection for all
cases in the rich mixture is more significant than that
for the lean mixture. (4) The knock and misfiring occur
at more advanced ignition timing by water injection, but
an excessive amount of water addition relative to the
fuel-air mixture mass caused additional misfiring. (5)
The exhaust gas temperature is reduced for all equiva-
lence ratios by increasing the amount of water addition.
This reduction is more significant for the lean mixture
case as compared to the rich mixture case. (6) Water
addition to the intake port seems to be a simpler method
of EGR since injected water certainly dilutes the charge,
exactly as EGR does, but due to the higher latent heat
of evaporation of water, the magnitude of the effects on
combustion and emissions is relatively different.
Acknowledgment. This study was financially sup-
ported by the Research Fund of the University of
Gaziantep in terms of a research project coded MF
01-06.
EF049850G
Figure 6. Effect of water addition on MBT timing.
Figure 7. Effect of water addition on knocking and misfire.
Figure 8. Effect of water addition on the engine thermal
efficiency.