Applied Energy 86 (2009) 630639

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Performance and exhaust emissions of a gasoline engine with ethanol blended

gasoline fuels using articial neural network
G. Naja a,*, B. Ghobadian a, T. Tavakoli a, D.R. Buttsworth b, T.F. Yusaf b, M. Faizollahnejad a
a
b

Tarbiat Modares University, Jalale-E-Aleahmad Highway, Tehran, P.O. Box: 14115-111, Iran
University of Southern Queensland, Toowoomba, 4350 QLD, Australia

a r t i c l e

i n f o

Article history:
Received 19 June 2008
Received in revised form 16 September
2008
Accepted 20 September 2008
Available online 4 November 2008
Keywords:
Articial neural network
SI engine
Engine performance
Exhaust emissions
Ethanolgasoline blends

a b s t r a c t
The purpose of this study is to experimentally analyse the performance and the pollutant emissions of a
four-stroke SI engine operating on ethanolgasoline blends of 0%, 5%, 10%, 15% and 20% with the aid of
articial neural network (ANN). The properties of bioethanol were measured based on American Society
for Testing and Materials (ASTM) standards. The experimental results revealed that using ethanolgasoline blended fuels increased the power and torque output of the engine marginally. For ethanol blends it
was found that the brake specic fuel consumption (bsfc) was decreased while the brake thermal efciency (gb.th.) and the volumetric efciency (gv) were increased. The concentration of CO and HC emissions in the exhaust pipe were measured and found to be decreased when ethanol blends were
introduced. This was due to the high oxygen percentage in the ethanol. In contrast, the concentration
of CO2 and NOx was found to be increased when ethanol is introduced. An ANN model was developed
to predict a correlation between brake power, torque, brake specic fuel consumption, brake thermal efciency, volumetric efciency and emission components using different gasolineethanol blends and
speeds as inputs data. About 70% of the total experimental data were used for training purposes, while
the 30% were used for testing. A standard Back-Propagation algorithm for the engine was used in this
model. A multi layer perception network (MLP) was used for nonlinear mapping between the input
and the output parameters. It was observed that the ANN model can predict engine performance and
exhaust emissions with correlation coefcient (R) in the range of 0.971. Mean relative errors (MRE) values were in the range of 0.465.57%, while root mean square errors (RMSE) were found to be very low.
This study demonstrates that ANN approach can be used to accurately predict the SI engine performance
and emissions.
2008 Elsevier Ltd. All rights reserved.

1. Introduction
Developing renewable energy has become an important part of
worldwide energy policy to reduce greenhouse gas emissions
caused by fossil fuel [1]. Alternative transport fuels such as hydrogen, natural gas and biofuels are seen as an option to help the
transport sector in decreasing its dependency on oil and reducing
its environmental impact [2].
According to [3], using ethanolgasoline blend fuel in a sparkignition (SI) engine caused a higher engine torque than that of
gasoline fuel. The maximum torque was obtained at 0.9 relative
airfuel ratio [3]. The effects of ethanolgasoline blends (E0, E10,
E20, E40 and E60) on engine exhaust emissions and performance
has been investigated by [4]. According to the results of the experiment, engine torque increased. It was also reported that blends
with ethanol allowed the compression ratio to increase without
* Corresponding author. Tel.: +98 261 2203289.
E-mail address: NAGAFY_14@yahoo.com (G. Naja).
0306-2619/$ - see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apenergy.2008.09.017

any knock [5]. Ref. [6] reported that the most suitable ethanolgasoline fuel blend in terms of performance and emissions was E50 in
a small gasoline engine with low efciency. Engine power increased by about 29% running with E50 fuel at high compression
ratio compared to running with E0 fuel. The specic fuel consumption was reduced by approximately 3% [6,7]. With increasing the
ethanol content in gasoline fuel, the heating value of the blended
fuels is decreased, while the octane number of the blended fuels
increases. Ref. [8] reported that blending unleaded gasoline with
ethanol increases the brake power, torque, volumetric and brake
thermal efciencies and fuel consumption, while it decreases the
brake specic fuel consumption (bsfc; the same conclusions can
be reached from Refs. [9,10].) The 20 vol.% ethanol in fuel blend
gave the best results for all measured parameters at all engine
speeds [8].
Using E40 and E60 blends led to a signicant reduction of CO
and HC emissions [4]. For E50, the CO, CO2, and HC emissions were
reduced by 53%, 10% and 12%, respectively [6,7]. NOx emission depends on the engine operating condition rather than the ethanol

G. Naja et al. / Applied Energy 86 (2009) 630639

631

Nomenclature
a
ANN
A/F
bsfc
MLP
MRE
n
p
R

actual output
articial neural network
air fuel ratio
brake specic fuel consumption (kg/k Wh)
multi layer perception
mean relative errors
number of the points in the data set
predicted output
correlation coefcient

content [11]. NOx concentrations are increased due to rising of the

cylinder temperature with increasing ethanol percentage in the
blends [9]. The effect of ethanol blended gasoline fuels on emissions was investigated in a spark-ignition engine with an electronic
fuel injection (EFI) system [10]. The addition of ethanol to gasoline
fuel enhances the octane number of the blended fuels and changes
distillation temperature. Ethanol is reported to be an important
contributor to decreased engine-out regulated emissions and decreased brake specic energy consumption [10]. Ref. [12] reported
that using E10 blended fuel at a relative airfuel ratio (k) slightly
greater than one can generally reduce pollutant emission components [12].
The effects of using ethanol-unleaded gasoline blends on cyclic
variability and emissions in a spark-ignited engine have been
investigated by Ref. [13]. It was reported that using ethanol-unleaded gasoline blends as a fuel decreased the coefcient of variation in indicated mean effective pressure. The 10 vol.% ethanol in
fuel blend gave the best results [13].
Articial neural networks (ANN) are data-processing systems
inspired by biological neural system and are used to solve a wide
variety of problems in science and engineering, particularly for
some areas where the conventional modelling methods fail. A
well-trained ANN can be used as a predictive model for a specic
application. The predictive ability of an ANN results from the training on experimental data and then validation by independent data.
An ANN has the ability to relearn to improve its performance if
new data are available [14]. Testing the engine under all possible
operating conditions and fuel cases is both time consuming and
expensive. As an alternative, the performance and exhaust emissions of an engine can be modelled using ANNs [15]. Refs. [16
24] investigated the performance of various thermal systems with
the aid of ANN. The ANN approach was used to predict the performance and exhaust emissions of internal combustion engines
[15,25,26] and the specic fuel consumption and fuel air equivalence ratio of a diesel engine [27]. Ref. [28] reported the effects
of valve-timing in a spark-ignition engine on the engine performance and fuel economy using ANN [28].
In this study, the use of ANN has been proposed to determine
the engine power, torque, brake specic fuel consumption, brake
thermal efciency, volumetric efciency and emission components
based on different gasolineethanol blends and speeds using results of experimental analysis.
2. Experimental work
2.1. Description of the experimental setup and ECU
In this study, the experiments were performed on a KIA 1.3
SOHC, four cylinder, four-stroke, and spark-ignition (SI) gasoline
engine. The engine specication is given in Table 1.
A 190 kW SCHENCK-WT190 eddycurrent dynamometer was
used in the experiments. Fuel consumption rate was measured in

RMSE
SI
SOHC

gb.th.
gi
gv
k
/

root mean square error

spark ignition
single over head camshaft
brake thermal efciency (%)
indicated efciency (%)
volumetric efciency (%)
relative airfuel ratio
equivalence fuelair ratio

the range of 0.445 kg/h by using laminar type ow meter,

Pierburg model. Air consumption was measured using an air ow
meter. The relative airfuel ratio, emission parameters and the
exhaust gas temperature from an online and accurately calibrated
exhaust gas analyser DIGAS 4000 were recorded. Five separate fuel
tanks were tted to the gasoline engine and these contained
gasoline and the bioethanolgasoline blends.
The engine control unit (ECU) that was used in this engine was a
Johnson Controls JCAE S2000. ECU function is to control the quantity of fuel, injection timing, ignition timing and engine speed by
receiving signals from seven sensors. These sensors are oxygen
sensor, knock sensor, manifold air pressure sensor, intake air temperature sensor, throttle position sensor, water temperature sensor
and engine speed sensor. A multi point fuel injection (MPFI) system with top-feed injectors is used to inject the fuel into the combustion chamber. The ignition system was semi-static distributor
less ignition (DLI). A schematic diagram of the experimental setup
is shown in Fig. 1.
2.2. Testing procedure
The performance and emission from the engine running on ethanol (derived from potato waste) and blended with gasoline (E5,
E10, E15 and E20) were evaluated and compared with gasoline
fuel. The properties of ethanol fuel are given in Table 2.
Above 20% ethanol, engine could not run smoothly, therefore,
only experimental results obtained up to this percentage of ethanol
will be presented. The fuel blends were prepared just before starting the experiment to ensure that the fuel mixture was homogenous and to avoid the reaction of ethanol with water. A series of
experiments were carried out using gasoline, and the various bioethanol blends. All the blends were tested under varying engine
speed conditions. The engine was started using gasoline fuel and
it was operated until it reached the steady state condition. The engine speed, fuel consumption, and load were measured, while the
brake power, brake specic fuel consumption (bsfc), brake thermal
efciency and volumetric efciency were computed. After the engine reached the stabilized working condition, emission parameters such as CO, CO2, HC, NOx and the exhaust gas temperature
from an online and accurately calibrated exhaust gas analyser were
Table 1
Main characteristics of the test engine
Engine type
Combustion order
Bore  stroke (mm)
Displacement volume (cc)
Compression ratio
Maximum torque (N m/rpm)
Maximum power (kW/rpm)
Maximum speed (rpm)
Cooling system

8 Valves 4 cylinder-inline-SOHC
1-3-4-2
71  83.6
1323
9.7
103/2750
47/5200
6200
Liquid, enclosed with forced
circulation of a cooling uid

632

G. Naja et al. / Applied Energy 86 (2009) 630639

Fig. 1. Schematic diagram of experimental setup.

Table 2
The properties of potatos wastes ethanol
Property

Method

Density (kg/m3)
Viscosity (cSt)
Caloric value (kJ/kg)
Research octane number
Pour point (C)
Flash point (C)
Ash content (mass%)

ASTM
ASTM
ASTM
ASTM
ASTM
ASTM
ASTM

D
D
D
D
D
D
D

Ethanol (E100)
4052
88
240
2699
97
93
482

785
1.1
27,000
108.6
<<50
14
0

recorded. All experiments have been carried out at full throttle setting. To adjust ignition timing, electronic ignition system was used.
Before obtaining data from the engine operated with a new
blended fuel, the engine was operated using the new fuel for sufcient time to clean out the remaining fuel from the previous blend.
Fuel properties were determined at the laboratories of Research
Institute of Petroleum Industry (RIPI) in Iran. In this paper, the quantity EX represents a blend consisting of X% bioethanol by volume,
e.g., E5 indicates a blend consisting of 5% ethanol in 95% gasoline.
Five test fuels were used in this study: 0% ethanol (E0); 5% ethanol
(E5); 10% ethanol (E10); 15% ethanol (E15); and 20% ethanol (E20).
The properties of the ve fuels have been summarized in Table 3.
3. Experimental results
3.1. Engine performance
3.1.1. Brake power
Fig. 2a shows the effect of various fuels on engine brake power.
When the ethanol content in the blended fuel is increased, the enTable 3
Properties of different ethanol/gasoline-blended fuels
Property item

Test fuel
E0

Vapour pressure (kPa),

reid@37.8 C
Octane number
(research)
Gross heat of
combustion (MJ/kg)
Distillation range (C)
@760 mmHg
IBP
10 vol%
50 vol%
90 vol%
End point

E5

E10

E15

E20

48.26

55.16

55.16

55.16

55.16

85.3

89.7

92.3

94.0

99.4

45.12

44.15

42.87

41.61

40.51

Test
method
ASTMD323
ASTMD2699
ASTMD340
ASTMD86

35.8
58.6
93.3
146.0
176.7

40.9
54.3
93.5
147.9
184.1

38.9
53.1
71.9
143.9
175.1

44.0
57.2
71.4
144.7
182.4

40.8
55.4
71.6
142.1
176.6

gine brake power slightly increased for all engine speeds. The gain
of the engine power can be attributed to the increase of the indicated mean effective pressure for higher ethanol content blends
[9]. The heat of evaporation of ethanol is higher than that gasoline,
this provides fuelair charge cooling and increases the density of
the charge, and thus higher power output is obtained [6]. With
the increase in ethanol percentage, the density of the mixture
and the engine volumetric efciency increases and this causes
the increase of power [8].
3.1.2. Torque output
Fig. 2b shows the inuence of different ethanolgasoline
blended fuels on engine torque. The increase of ethanol content increases the torque of the engine. Added ethanol produces lean mixtures that increase the relative airfuel ratio (k) to a higher value
and makes the burning more efcient [11]. The improved antiknock behaviour (due to the addition of ethanol, which raises the
octane number) allowed a more advanced timing that results in
higher combustion pressure and thus higher torque [7,29].
3.1.3. Brake thermal efciency
Fig. 2c presents the effect of using ethanolgasoline blends on
brake thermal efciency. As shown in this gure, the brake thermal
efciency increases as the ethanol percentage increases. The maximum brake thermal efciency (gb.th.) was approximately 35%
when 20% ethanol was in the fuel blend. As the E% increases in
the fuel blend, the indicated work increases (i.e., the indicated efciency gi increases). The relationship between the engine speeds
and the brake thermal efciency is given in Fig. 2c. As the engine
speed increases reaching 3000 rpm, the brake thermal efciency
increases reaching its maximum values.
3.1.4. Volumetric efciency
Fig. 2d shows the relationship between the volumetric efciency (gv) and the percentage of ethanol in the fuel blends. As
shown from Fig. 2d, the volumetric efciency increases as the ethanol percentage increases for all engine speeds.
3.1.5. Brake specic fuel consumption
The relationship between engine speed and brake specic fuel
consumption for different blends and gasoline is shown in
Fig. 2e. As shown in this gure, the bsfc decreases as the ethanol
percentage increases. As engine speed increases reaching
3500 rpm, the bsfc decreases reaching its minimum values. This
is due to the increase in brake thermal efciency.
3.1.6. Equivalence fuelair ratio (/)
Equivalence fuelair ratio (/) is one of the important parameters that effects engine performance parameters. Ethanol is an

633

G. Naja et al. / Applied Energy 86 (2009) 630639

50
45
40
35
30
25
20
15
10
5
0

110

E0

105

E0
E5
E10
E15
E20

Torque (Nm)

Brake power (kW)

E5
E10

100

E15
95

E20

90
85
80

1000

2000

3000

4000

5000

6000

1000

Engine speed (rpm)

3000

4000

5000

d
40

E0

35

E5
E10
E15
E20

30
25
20
0

1000

2000

3000

4000

5000

Volumetric efficiency (%)

Brake thermal efficiency (%)

6000

E0

105

E5
E10

100

E15

95

E20

90
85
80
75
0

6000

1000

Engine speed (rpm)

2000

3000

4000

5000

6000

Engine speed (rpm)

0.34
0.32
0.30
E0
0.28

E5

0.26

E10
E15

0.24

E20

0.22
0

1000

2000

3000

4000

5000

6000

Engine speed (rpm)

Equivalence fuel-air ratio ()

e
bsfc (kg/kW.hr)

2000

Engine speed (rpm)

1.3
1.2
E0

1.1

E5
E10

1.0

E15
E20

0.9
0

1000

2000

3000

4000

5000

6000

Engine speed (rpm)

Fig. 2. Experimental results of (a) brake power, (b) Torque, (c) gb.th., (d) gv, (e) bsfc and (f) equivalence fuelair ratio at different fuel blends and engine speeds.

oxygenated fuel, and for this reason, adding ethanol to gasoline

leads to leaner operation and improves combustion [9]. Experimental results show that equivalence fuelair ratio decreases as
the percentage of ethanol (in volume) in the blended fuel increases
(Fig. 2f). Engine performance parameters such as brake power,
torque, volumetric efciency and brake thermal efciency are increased when the ethanol amount in the blended fuel is increased.
This is due to the reduction in the / values. As previously mentioned, increasing ethanol blend ratio leads to lean and complete
combustion and therefore the bsfc was slightly decreased shown
in Fig. 2.

CO concentrations at 3000 rpm using E5, E10, E15 and E20 was
decreased by 13.7%, 24.31%, 27.93% and 45.42%, respectively in
comparison to gasoline. The reduction in CO concentration using
blended fuels is due to the fact that ethanol (C2H5OH) has less
carbon than gasoline (C8H18). Another signicant reason of this
reduction is that the oxygen content in the blended fuels increases
the oxygen-to-fuel ratio in the fuel-rich regions. The most
signicant parameter affecting CO concentration is the relative
airfuel ratio (k) [6,12]. Relative airfuel ratio (k) approaches 1 as
the ethanol content of the blended fuel increases, and consequently combustion becomes complete [9,11,12].

3.2. Engine emission studies

3.2.2. CO2 emission

Fig. 3b shows the relationship between the CO2 concentrations
and engine speeds for different blends percentage. Fig. 3b indicates
that CO2 concentration increases as the ethanol percentage increases. CO2 emission depends on relative airfuel ratio and CO
emission concentration [6,12]. The CO2 concentration in the exhaust gas emission at 3000 rpm for gasoline fuel was 12.4 (%V),
while the CO2 concentration of E5, E10, E15 and E20 at 3000 rpm
was 12.9, 13.2, 13.3 and 13.8 (%V), respectively. The CO2 concentrations at 3000 rpm using E5, E10, E15 and E20 was increased by

3.2.1. CO emission
Fig. 3a shows the concentrations of CO emission for different
engine speeds. It can be seen from this gure that when ethanol
percentage increases, the CO concentration decreases which means
the combustion is tuned to be completed. The CO concentration in
the exhaust gas emission at 3000 rpm for gasoline fuel was 4.69
(%V), while the CO concentration of E5, E10, E15 and E20 at
3000 rpm was 4.05, 3.55, 3.38 and 2.56 (%V), respectively. The

634

G. Naja et al. / Applied Energy 86 (2009) 630639

E0
E5

14
E0

E5
E10

CO2 (%V)

CO (%V)

15

E15

E10
13

E15
E20

12
11

E20
0

10
0

1000

2000

3000

4000

5000

6000

1000

Engine speed (rpm)

2000

3000

4000

5000

400

E0

2000

E0

E5
300

E5

E10
E15
E20

200

1500

NOx(ppm)

HC (ppm)

6000

Engine speed (rpm)

E10
E15

1000

E20

500

100

0
0

1000

2000

3000

4000

5000

6000

Engine speed (rpm)

1000

2000

3000

4000

5000

6000

Engine speed (rpm)

Fig. 3. Experimental results of (a) CO, (b) CO2, (c) HC and (d) NOx at different gasolineethanol blends and engine speeds.

3.87%, 6.06%, 6.76% and 10.14%, respectively in comparison to gasoline. As a result of the lean burning associated with increasing
ethanol percentages, the CO2 emission increased because of the
improved combustion [8,11,12].
3.2.3. HC emission
HC emissions for different speeds are illustrated in Fig. 3c. The
HC concentration in the exhaust gas emission at 3000 rpm for gasoline fuel was 183 ppm, while the HC concentration of E5, E10, E15
and E20 at 3000 rpm was 152, 139, 137 and 125 ppm, respectively.
The HC concentration at 3000 rpm using E5, E10, E15 and E20 was
decreased by 16.94%, 24.04%, 25.14% and 31.69% at 3000 rpm,
respectively in comparison to gasoline. This result indicates that
ethanol can signicantly reduce HC emissions. The concentration
of HC emission decreases with the increase of the relative airfuel
ratio, the reason for the decrease of HC concentration is similar to
that of CO concentration described above [11,12].

train the ANN so that it can precisely predict the system performance at other conditions. This technique has found application
in situations where the simulation of complex systems is required
but limited experimental data is available. ANN is a powerful, nonlinear tool and since many phenomena in industry have non-linear
characteristics, ANN has been applied widely. The performance of
the ANN-based predictions is evaluated by regression analysis of
the network outputs (predicted parameters) and the experimental
values [15]. The error identied during the learning process is
called the root-mean-squared-error (RMSE) and is dened as
follows:

3.2.4. NOx emission

Considering the NOx emission, Fig. 3d shows that the NOx concentration is higher when ethanol percentage increases. It shows
that as the percentage of ethanol in the blends increased, NOx
emission was increased. The NOx concentration in the exhaust
gas emission at 3000 rpm for gasoline fuel was 876 ppm, while
the NOx concentration of E5, E10, E15 and E20 at 3000 rpm was
1002, 1326, 1319 and 1609 ppm, respectively. The NOx concentrations at 3000 rpm using E5, E10, E15 and E20 was increased by
12.57%, 33.94%, 33.6% and 45.55%, respectively in comparison to
gasoline. When the combustion process is closer to stoichiometric,
ame temperature increases, therefore, the NOx emission is increased, particularly by the increase of thermal NO [9,11].
4. Articial neural networks
ANN is an analytical method for simulating system performance. The method relies on experimental data that is used to

Fig. 4. The structure of ANN for gasoline engine with gasolineethanol blended
fuels.

635

G. Naja et al. / Applied Energy 86 (2009) 630639

indicate a stronger positive linear relationship, while R values closer to 1 indicate a stronger negative relationship [15]. The mean
relative error, which shows the mean ratio between the error and
the experimental values, is dened as

Table 4
Summary of different networks evaluated to yield the criteria of network
performance
Activation
function

Training
rule

Neurons in
hidden layer

Training error

sig/lin
tan/lin
sig/lin
sig/lin
sig/lin
sig/lin
sig/lin
sig/lin
sig/lin
sig/lin
sig/lin

trainlm
trainlm
traingdx
trainscg
trainrp
trainlm
trainlm
trainlm
trainlm
trainlm
trainlm

20
20
20
20
20
19
18
17
21
22
23

9.99  106
4.61  104
0.0137
8.51  103
6  103
6.091  105
2.11  104
2.72  104
9.75  106
9.39  106
8.28  106

0.99997
0.99929

0.99106
0.99221
0.9993
0.9992
0.9995
0.99997
0.99995
0.99997

Pn
RMSE

 pi 2

i1 ai

MRE %

5. Modelling with ANN

An ANN model was developed to predict a correlation between
brake power, torque, brake specic fuel consumption, brake thermal efciency, volumetric efciency and emission components
using different gasolineethanol blends and speeds as inputs data.
Approximately 70% of the total experimental data (405 values) was
selected at random and was used for training purpose, while the
30% was reserved for testing. The experimental data set for every
output parameter includes 45 values, of which 30 values were used
for training the network and 15 values were selected randomly to
test the performance of the trained network. A standard Back-

40
30
20

R=0.999
RMSE=0.74 (kW)
MRE=2.32 (%)

10
0
0

110

Predicted Torque (Nm)

Predicted Brake power (kW)

The correlation coefcient (R) and mean relative error (MRE) are
used for characterising the network performance. The correlation
coefcient can vary between 1 and +1, but R values closer to +1

50

10

20

30

40

105
100
95
R=0.995
RMSE=0.49 (Nm)
MRE=0.46 (%)

90
85
85

50

90

Experimental Brake power (kW)

35
30

20
20

R =0.981
RMSE=0.59 (%)
MRE=1.28 (%)
25

30

35

Predicted bsfc (kg/kW.hr)

105

110

95
90
R=0.985
RMSE=0.48 (%)
MRE=0.48 (%)

85
80
80

40

Experimental Brake thermal

efficiency (%)
0.32

100

100

25

95

Experimental Torque (Nm)

Predicted Volumetric
efficiency (%)

Predicted Brake thermal

efficiency (%)

40

where n is the number of the points in the data set, and a and p are
actual output and predicted output sets, respectively [15].

!12


n 

1X
100  ai  pi :
n i1 
ai 

85

90

95

100

Experimental Volumetric
efficiency (%)

0.30
0.28
0.26
R=0.986
RMSE=0.003 (kg/kW.hr)
MRE=0.85 (%)

0.24
0.22
0.22

0.24

0.26

0.28

0.30

0.32

Experimental bsfc (kg/kW.hr)

Fig. 5. The ANN predictions for the (a) brake power, (b) Torque, (c) gb.th., (d) gv and (e) bsfc versus experimental values.

636

G. Naja et al. / Applied Energy 86 (2009) 630639

Propagation algorithm for the engine was used in this model. A

multi layer perception network (MLP) was used for nonlinear mapping between the input and the output parameters.
To get the best prediction by the network, several architectures
were evaluated and trained using the experimental data. The backpropagation algorithm was utilized in training of all ANN models.
This algorithm uses the supervised training technique where the
network weights and biases are initialized randomly at the beginning of the training phase. The error minimization process is
achieved using gradient descent rule.
There were two input and nine output parameters in the experimental tests. The two input variables are engine speed in rpm and
the percentage of bioethanol blending with the conventional gasoline fuel. The nine outputs for evaluating engine performance are
indicated in Fig. 4. There is 1 neural network structure with 2 inputs and 9 outputs; the input layer consisted of 2 neurons which
corresponded to engine speed and levels of biofuel blends and
the output layer had 9 neurons. The number of hidden layers and
neurons within each layer can be designed by the complexity of
the problem and data set. In this study, the network was decided
to consist of one hidden layer with 20 neurons. The activation
function for hidden layer was selected to be sigmoid function. A
linear function was best suited for the output layer. However,
many other networks with several functions and topologies were
examined.
Three criteria R, RMSE and MRE were selected to evaluate the
networks to nd the optimum solution. The complexity and size
of the network was also an important consideration, and therefore
smaller ANNs had to be selected. A regression analysis between the
network response and the corresponding targets was performed to
investigate the network response in more detail. Different training
algorithms were tested and LevenbergMarquardt (trainlm) was
selected. R values in Table 4 represent the correlation coefcient
between the outputs and targets. As seen in Table 4, R values did
not increase when neurons in the hidden layer is more than 20
such as 21, 22 and 23. Therefore, a network with one hidden layer

6. Results and discussion

Adding ethanol to gasoline will lead to a leaner better combustion. It was experimentally demonstrated that adding 20% ethanol
to the blends led to an increase in the engine brake power, torque
and brake thermal efciency, volumetric efciency and decreases
the brake specic fuel consumption. The lean combustion improves the completeness of combustion and therefore the CO emission was expected to be decreased. The experimental results
conrmed that by adding more ethanol, the CO was decreased.
The oxygen enrichment generated from ethanol increased the oxygen ratio in the charge and lead to lean combustion. CO2 emission
varies with the A/F ratio and CO concentration. As a result, the CO2
emission increased because of the improved combustion. Unburned HC is a product of incomplete combustion which is related
to A/F ratio. It is noted that adding ethanol to the blends reduces
the HC emission because of oxygen enhancement. When the combustion process is closer to stoichiometric, ame temperature increases, therefore, the NOx emission increased.
An articial neural network (ANN) was developed based on this
experimental work. The results showed that the training algorithm
of Back Propagation was sufcient for predicting engine torque,
brake power, brake thermal efciency, volumetric efciency, specic fuel consumption and exhaust gas components for different
engine speeds and different fuel blends ratios. The predicted versus
experimental values for the experimental parameters are indicated
in Fig. 5. The ANN predictions for the (a) brake power, (b) engine
torque, (c) brake thermal efciency, (d) volumetric efciency and
(e) brake specic fuel consumption yield a correlation coefcient

8
R=0.994
RMSE=0.18 (%V)
MRE=4.21 (%)

Predicted CO2 (%V)

Predicted CO (%V)

and 20 neurons was selected as the optimum ANN. Simulations

were performed using MATLAB. This ANN model is limited to the
research engine that was used in this project operating on at
wide open throttle conditions (the engine specications are given
in Table 1).

4
2

R=0.985
RMSE=0.24 (%V)
MRE=1.54 (%)

14
13
12
11
10
10

0
0

15

300
250

R=0.987
RMSE=5.41 (ppm)
MRE=2.23 (%)

Predicted NOx (ppm)

Predicted HC (ppm)

200
150
100
100

150

200

250

Experimental HC (ppm)

11

12

13

14

15

Experimental CO2 (%V)

Exxperimental CO (%V)

300

2000

R=0.973
RMSE=89.85 (ppm)
MRE=5.57 (%)

1500
1000
500
0
0

500

1000

1500

Experimental NOx (ppm)

Fig. 6. The ANN predictions for the (a) CO, (b) CO2, (c) HC and (d) NOx versus experimental values.

2000

637

G. Naja et al. / Applied Energy 86 (2009) 630639

50

Brake power (kW)

(R) of 0.999, 0.995, 0.981, 0.985 and 0.986, respectively. It was

found the root mean square error (RMSE) values were 0.74 kW,
0.49 N m, 0.59%, 0.48% and 0.003 kg/kW h for the (a) brake power,
(b) engine torque, (c) brake thermal efciency, (d) volumetric efciency and (e) brake specic fuel consumption, respectively.
The ANN predictions for the brake power yield a mean relative
error (MRE) of 2.32% (Fig. 5a). Fig. 5b indicates the predicted versus
experimental values for the engine torque with mean relative error
(MRE) of 0.46%. Mean relative error (MRE) of brake thermal efciency was 1.28% (Fig. 5c). Fig. 5d indicates the predicted versus
experimental values for the volumetric efciency. The ANN predictions for the volumetric efciency yield a mean relative error
(MRE) of 0.48%. The ANN predictions for the brake specic fuel
consumption (bsfc) yield mean relative error (MRE) of 0.85%
(Fig. 5e).
The ANN predictions for the performance of the tested SI engine yielded a good statistical performance. Comparisons of the
experimental results and the ANN predictions demonstrate that
SI engines using ethanolgasoline blended fuels can be accurately simulated using ANN. The predicted versus experimental
values for the CO emission are indicated in Fig. 6a. The ANN pre-

40

dictions for the CO yield a correlation coefcient (R) of 0.994,

root mean square error (RMSE) of 0.18%V and mean relative error (MRE) of 4.21%. Fig. 6b indicates the predicted versus experimental values for the CO2 emission with correlation coefcient
(R) of 0.985, root mean square error (RMSE) of 0.24%V and mean
relative error (MRE) of 1.54%. The predicted versus experimental
values for the HC emission are indicated in Fig. 6c. The ANN predictions for the HC yield a correlation coefcient (R) of 0.987,
root mean square error (RMSE) of 5.41 ppm and mean relative
error (MRE) of 2.23%. Fig. 6d indicates the predicted versus
experimental values for the NOx emission. The ANN predictions
for the NOx yield a correlation coefcient (R) of 0.973, root mean
square error (RMSE) of 89.85 pm and mean relative error (MRE)
of 5.57%.
A comparative presentation of the error during testing by using
ANN and experimental results is shown in Figs. 7 and 8. It is seen
that the test pattern consist of the results of 15 tests. It was observed that the ANN model can predict engine performance and
exhaust emissions with a correlation coefcient (R) in the range
of 0.971. Mean relative errors (MRE) values were in the range of
0.465.57%, while root mean square errors (RMSE) were found to

110

Experimental
Predicted

Torque (Nm)

105

30
20
Experimental

10

100
95
90

Predicted

85

0
0

10

15

Test pattern

10

15

Test pattern

d
40

Experimental

Volumetric efficiency (%)

Brake thermal efficiency (%)

Predicted
35
30
25
20
0

10

15

100

Experimental
Predicted

95
90
85
80
0

Test pattern

bsfc (kg/kW.hr)

10

15

Test pattern

0.32

Experimental
Predicted

0.30
0.28
0.26
0.24
0.22
0

10

15

Test pattern
Fig. 7. Comparisons of experimental results and the ANN predictions for the (a) brake power, (b) Torque, (c) gb.th., (d) gv and (e) bsfc for various test patterns.

638

G. Naja et al. / Applied Energy 86 (2009) 630639

Experimental

Predicted

CO2 (%V)

CO (%V)

Experimental
Predicted

14

6
4

15

13
12
11

10

10

15

Test pattern

300

Experimental

2000

15

Experimental
Predicted

Predicted
1500

NOx (ppm)

250

HC (ppm)

10

Test pattern

200

1000
500

150

100
0

10

15

Test pattern

10

15

Test pattern

Fig. 8. Comparisons of experimental results and the ANN predictions for the (a) CO, (b) CO2, (c) HC and (d) NOx for various test patterns.

be very low. There is a good correlation between the simulations

from ANN and the measured data. Therefore, ANN appears to be
a useful method for simulating engine parameters.
7. Conclusions
The present work demonstrates that the use of ethanolgasoline blended fuels will marginally increase the brake power and decrease the brake specic fuel consumption. It was also found that
the brake thermal efciency and volumetric efciency increase
when ethanolgasoline blends are used. The CO2 and NOx concentrations were increased while the concentrations of CO and HC
were decreased when ethanolgasoline blends are used. The ANN
results are very good, R values in this model are very close to
one, while root mean square errors (RMSE) were found to be very
low. Analysis of the experimental data by the ANN revealed that
there is a good correlation between the ANN-predicted results
and the experimental data. Therefore ANN proved to be a useful
tool for correlation and simulation of engine parameters. ANN provided an accurate and simple approach in the analysis of this complex, multivariate problem, the analysis of the SI engine
performance and emissions.
Acknowledgements
The authors wish to thank the Iranian Fuel Conservation Organization (IFCO) of NIOC for the research grant provided to complete
this project and Mega Motor Company for providing of laboratory
facilities.
References
[1] Huang Y-H, Wu J-H. Analysis of biodiesel promotion in Taiwan. Renew Sust
Energy Rev 2008;12:117686.
[2] Wiesenthal T, Leduc G, Christidis P, Schade B, Pelkmans L, Govaerts L et al.
Biofuel support policies in Europe: lessons learnt for the long way ahead.
Renew Sust Energy Rev, in press. doi:10.1016/j.rser.2008.01.011.

[3] Ycesu HS, Sozen A, Topgl T, Arcaklioglu E. Comparative study of

mathematical and experimental analysis of spark ignition engine
performance used ethanolgasoline blend fuel. Appl Therm Eng 2007;27:
35868.
[4] Ycesu HS, Topgl T, inar C, Okur M. Effect of ethanolgasoline blends on
engine performance and exhaust emissions in different compression ratios.
Appl Therm Eng 2006;26(1718):22728.
[5] Topgl T, Ycesu HS, inar C, Koca A. The effects of ethanolunleaded gasoline
blends and ignition timing on performance and exhaust emissions. Renew
Energy 2006;31(15):253442.
[6] Celik MB. Experimental determination of suitable ethanolgasoline blend rate
at high compression ratio for gasoline engine. Appl Therm Eng
2008;28:396404.
[7] Agarwal AK. Biofuels (alcohols and biodiesel) applications as fuels for internal
combustion engines. Prog Energy Combust Sci 2007;33:23371.
[8] Al-Hasan M. Effect of ethanol-unleaded gasoline blends on engine performance
and exhaust emissions. Energy Conv Manage 2003(44):154761.
[9] Bayraktar H. Experimental and theoretical investigation of using gasoline
ethanol blends in spark-ignition engines. Renew Energy 2005;30:
173347.
[10] He BQ, Wang JX, Hao JM, Yan XG, Xiao JH. A study on emission characteristics
of an EFI engine with ethanol blended gasoline fuels. Atmos Environ
2003;37(7):94957.
[11] Hsieh WD, Chen RH, Wu TL, Lin TH. Engine performance and pollutant
emission of an SI engine using ethanolgasoline blended fuels. Atmos Environ
2002;36:40310.
[12] Wu CW, Chen RH, Pu JY, Lin TH. The inuence of airfuel ratio on engine
performance and pollutant emission of an SI engine using ethanolgasolineblended fuels. Atmos Environ 2004;38:7093100.
[13] Ceviz MA, Yksel F. Effects of ethanolunleaded gasoline blends on cyclic
variability and emissions in an SI engine. Appl Therm Eng 2005;25:
91725.
[14] Hertz J, Krogh A, Palmer RG. Introduction to the theory of neural
computation. Redwood City, NJ: Addison-Wesley Publishing Company;
1991.
[15] Sayin C, Ertunc HM, Hosoz M, Kilicaslan I, Canakci M. Performance and exhaust
emissions of a gasoline engine using articial neural network. Appl Therm Eng
2007;27:4654.
[16] Kalogirou SA. Application of articial neural-networks for energy systems.
Appl Energy 2000;67:1735.
[17] Pacheco-Vega A, Sen M, Yang KT, McClain RL. Neural network analysis of ntube refrigerating heat exchanger with limited experimental data. Int J Heat
Mass Transfer 2001;44:76370.
[18] Bechtler H, Browne MW, Bansal PK, Kecman V. Neural networks a new
approach to model vapour-compression heat pumps. Int J Energy Res
2001;25:5919.
[19] Prieto MM, Montanes E, Menendez O. Power plant condenser performance
forecasting using a non-fully connected ANN. Energy 2001;26:6579.

G. Naja et al. / Applied Energy 86 (2009) 630639

[20] Chouai A, Laugeier S, Richon D. Modelling of thermodynamic properties using
neural networks application to refrigerants. Fluid Phase Equilibr
2002;199:5362.
[21] Sozen A, Arcaklioglu E, Ozalp M. A new approach to thermodynamic analysis of
ejector-absorption cycle: articial neural networks. Appl Therm Eng
2003;23:93752.
[22] Arcaklioglu E. Performance comparison of CFCs with their substitutes using
articial neural network. Int J Energy Res 2004;28:111325.
[23] Ertunc HM, Hosoz M. Articial neural network analysis of a refrigeration
system with an evaporative condenser. Appl Therm Eng 2006;26:62735.
[24] Hosoz M, Ertunc HM. Articial neural network analysis of an automobile air
conditioning system. Energy Conv Manage 2006;47:157487.

639

[25] Canakci M, Erdil A, Arcaklioglu E. Performance and exhaust emissions of a

biodiesel engine. Appl Energy 2006;83:594605.
[26] Arcaklioglu E, Celikten I. A diesel engines performance and exhaust emissions.
Appl Energy 2005;80:1122.
[27] Celik V, Arcaklioglu E. Performance maps of a diesel engine. Appl Energy
2005;81:24759.
[28] Golcu M, Sekmen Y, Erduranli P, Salman S. Articial neural network based
modelling of variable valve-timing in a spark ignition engine. Appl Energy
2005;81:18797.
[29] Mouloungui Z, Vaitilingom G, Berge JC, Caro PS. Interest of combining an
additive with diesel ethanol blends for use in diesel engines. Fuel
2001;80(4):56574.