Energy Conversion and Management 133 (2017) 178–195

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Application of response surface methodology in optimization of

performance and exhaust emissions of secondary butyl alcohol-gasoline
blends in SI engine
I.M. Yusri a,⇑, R. Mamat a,b,c, W.H. Azmi a,b, A.I. Omar a, M.A. Obed d, A.I.M. Shaiful e
a
Faculty of Mechanical Engineering, Universiti Malaysia Pahang, 26600 Pekan, Malaysia
b
Automotive Engineering Centre, Universiti Malaysia Pahang, 26600 Pekan, Malaysia
c
Karlsruhe University of Applied Sciences, 76131 Karlsruhe, Germany
d
Northern Technical University, Iraq
e
School of Manufacturing Engineering, Universiti Malaysia Perlis, 01000 Kangar, Perlis, Malaysia

a r t i c l e i n f o a b s t r a c t

Article history: Producing an optimal balance between engine performance and exhaust emissions has always been one
Received 21 August 2016 of the main challenges in automotive technology. This paper examines the use of RSM (response surface
Received in revised form 30 November 2016 methodology) to optimize the engine performance, and exhaust emissions of a spark-ignition (SI) engine
Accepted 1 December 2016
which operates with 2-butanol–gasoline blends of 5%, 10%, and 15% called GBu5, GBu10, and GBu15. In
the experiments, the engine ran at various speeds for each test fuel and 13 different conditions were con-
structed. The optimization of the independent variables was performed by means of a statistical tool
Keywords:
known as DoE (design of experiments). The desirability approach by RSM was employed with the aim
2-Butanol
Performance
of minimizing emissions and maximizing of performance parameters. Based on the RSM model, perfor-
Emissions mance characteristics revealed that increments of 2-butanol in the blended fuels lead to increasing trends
RSM of brake power, brake mean effective pressure and brake thermal efficiency. Nonetheless, marginal higher
brake specific fuel consumption was observed. Furthermore, the RSM model suggests that the presence of
2-butanol exhibits a decreasing trend of nitrogen oxides, carbon monoxides, and unburnt hydrocarbon,
however, a higher trend was observed for carbon dioxides exhaust emissions. It was established from
the study that the GBu15 blend with an engine speed of 3205 rpm was found to be optimal to provide
the best performance and emissions characteristics as compared to the other tested blends.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction aforementioned GHG, the research community as a whole are con-

Energy shortage of fossilized fuel, as well as its adverse environ-
mental impact are given due attention globally. Owing to the
unsustainable nature of fossilized fuel, its rapid depletion and
overdependence must be addressed immediately [1–3]. Moreover,
the utilization of these conventional energy resources, mainly in
transportation areas, has led to major environmental side effects
[4,5]. This trend of energy consumption is envisaged to continue
in the near future [6]. The emission of greenhouse gasses (GHG)
namely carbon dioxide (CO2), nitrogen oxides (NOx), carbon
monoxides (CO), and unburned hydrocarbon (HC) are of interest
as it affects the earth’s climate change [7]. As the utilization of
fossilized fuel is deemed to be the primary contributor of the
⇑ Corresponding author.
E-mail addresses: yusri890@yahoo.com, m.yusri890@gmail.com,
@ump.edu.my (I.M. Yusri).
tinuously investigating on the search for cleaner alternative fuels
such as alcohol, bio-diesel, and vegetable-oil [8–10]. These alterna-
tive energies are fundamentally environmental-friendly; however,
they are still required to be evaluated in terms of engine perfor-
mance and emission characteristics [11].
Transportation is one of the leading causes of environmental
problems in almost every part of the world [12,13]. Furthermore,
it is expected that the number of vehicles, especially cars, and light
trucks, are to increase to up to 1.3 billion by 2030 and to over 2 bil-
lion vehicles by 2050 [14]. In order to facilitate the effort for a bet-
ter environmental condition throughout the world, the European
Union (EU) have pledged that by the year 2020, 20% and 10%, of
its transportation fuels and energy supply, respectively must be
replaced by renewable resources [15]. For spark-ignition (SI) engi-
nes, alcohol is considered one of the feasible solutions for fuel sub-
rizalman
stitution [16]. This is because the presence of excess oxygen in
alcohol allows gasoline fuels to produce better engine combustion

http://dx.doi.org/10.1016/j.enconman.2016.12.001
0196-8904/Ó 2016 Elsevier Ltd. All rights reserved.
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195 179

Nomenclature

Abbreviations SI spark ignition

ANOVA analysis of variance SOHC single overhead camshaft
BMEP brake mean effective pressure WTO wide throttle open
BSFC brake specific fuel consumption n-butanol
BTE brake thermal efficiency primary butyl alcohol
CI compression ignition 2-butanol
CO carbon monoxide secondary butyl alcohol
CO2 carbon dioxide
DoE design of experiment Symbols
ECI-multi electronically controlled multi-point fuel injection A engine speed
EU European union Adj R2 adjusted R2
G100 gasoline AP adequate precision
GBu5 5% 2-butanol + 95 gasoline B fuel blend
GBu10 10% 2-butanol + 90% gasoline F – value value of F-statistic
GBu15 15% 2-butanol + 85% gasoline R2 coefficient of determination
GHG greenhouse gasses p-value probability value
HC unburned hydrocarbon SError standard error
LHV low heating value

X mean of data collections
NOx nitrogen oxides r standard deviation
ppm parts per million n number of data collections
RSM response surface method
RSE relative standard error
rpm revolutions per minute

[17]. There are three types of alcohol that have recently attracted
the attention of automotive researchers for their potential future
use, viz. methanol, ethanol, and butanol [18–20]. A considerable
amount of literature investigated on the use of methanol and etha-
nol as sustainable alternative fuels, however, limited attention is
paid to the use of butanol. Although ethanol is presently the dom-
inant biofuel [21,22], butanol is seen as a viable fuel supplement to
gasoline fuels, and it has been proposed as one of the next-
generation of biofuels [23].
Butanol, as compared to ethanol and methanol, has a higher
heating value and lower volatility, stoichiometric air–fuel ratio,
octane number, and auto ignition temperature, thus making it
more readily extendable to be blended with gasoline fuels
[24,25]. Moreover, according to Fortman et al. [26], butanol has
been proposed as a viable alternative not only for gasoline and die-
sel fuel but also for ethanol as well. This is primarily due to buta-
nol’s higher energy content and lower solubility in water that
allows it to be easily transported through existing fuel pipelines.
The mixture of butanol-gasoline blends also accommodates high
compression ratios without inducing knocking [27]. Nonetheless,
the major drawback that prevents butanol’s use in internal com-
bustion engines is its much higher production cost compared to
gasoline, which has been the subject of other research studies
[28]. However, the recent enhancement of genetically modified
bacteria will increase butanol yield at a lower cost margin and
make it suitable on a life-cycle basis as an imminent fuel supple-
ment [29].
The successful use of butanol utilization under SI engine perfor-
mance and emissions characteristics has been demonstrated previ-
ously by studies that used them as one of the blending
components. One of the recent experimental studies by Elfasa-
khany [30] investigated the effects of adding dual butanol isomers
in an unleaded gasoline fuel towards its performance and emis-
sions characteristics of a single-cylinder, four-stroke, port fuel
injection SI engine. The fuels had been prepared according to these
percentages of volume: 3% (1.5% volume of iso-butanol and 1.5%
volume of n-butanol), 7% (3.5% volume of iso-butanol and 3.5%
volume of n-butanol), and 10% (5% volume of iso-butanol and 5%
volume of n-butanol) of butanol in gasoline blends. Following the
mixture of dual butanol isomers and gasoline blends, a slight
decrease in the torque and brake power of about 3.6%, 4%, and
2.1%, and of 5.9%, 7.2% and 4.6%, respectively were noted, for mix-
tures of 3%, 7%, and 10% blended fuels respectively, compared to
the unleaded gasoline. As for exhaust emissions, Elfasakhany dis-
covered that all emissions decreased by 2.9%, 4.3%, and 5.7%
(CO), 8.2%, 12%, and 15% (HC), and 42%, 41%, and 39% (CO2) for
3%, 7%, and 10% blended fuels in comparison to unleaded gasoline.
The author opined that the better engine performance offered by
the mixture of n-butanol and iso-butanol is presumably due to
the different laminar burning velocities of n-butanol and iso-
butanol, that in effect shortens the combustion duration in the
engine.
In another major study done by Feng et al. [31] 30% and 35%
n-butanol–gasoline blends were used in a single-cylinder, four-
stroke, high-speed motorcycle spark ignition engine at different
speeds and full load condition. A slight decrease in brake power
and brake torque was observed upon the utilization of the afore-
mentioned blend. Nonetheless, it is worth to note that as the igni-
tion timing is changed to a more advanced degree of crank angle,
higher brake power and torque were achieved by both blended
fuels as compared to neat gasoline. Conversely, the drawback
reported by altering the crank angle is the higher production of
NOx and CO2. According to Chen et al. [32], researchers have not
treated butanol–gasoline blend in a gasoline direct injection engine
in much detail. In their study, 15%, 30%, and 50% n-butanol–
gasoline blend fuels were tested in a gasoline direct injection
engine at a single engine speed kept constant at 2000 rpm and
three different engine brake mean effective pressures (BMEP), i.e.
0.2 MPa, 1.0 MPa, and 1.8 MPa. It was reported that the brake
specific fuel consumption increases with the increase in butanol
fraction. However, the brake thermal efficiency (BTE) improves
especially for the 50% fuel blends. Further results revealed that
increasing the n-butanol fraction reduces the exhaust temperature
and NOx emissions but increases the HC and CO emissions. Wallner
et al.’s [33] study of using an n-butanol–gasoline blend in a
gasoline direct injection engine found that there were few
180 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
discrepancies between the exhaust emissions of NOx, CO, and HC in
comparison to pure gasoline fuels. It is noteworthy to mention that
there are certain drawbacks associated with the direct injection
spark ignition engine namely the lack of fuel atomization of the
blended fuel, oil dilution, and cold start [34]. Alasfour [35–38]
run 30% of iso-butanol–gasoline blends over a wide range of fuel/
air equivalence ratios to investigate its performance, emissions
and availability analysis. Following the whole range of fuel/air
equivalence ratios, there was a reduction in BTE, specifically 4.5%
in equivalence ratios equal to 1. On emission characteristics of
NOx and HC, his investigation revealed that the maximum NOx
concentration for blended fuels was lower than the corresponding
maximum NOx emissions caused by pure gasoline, by approxi-
mately 9%, and both of the emissions increases with the inlet air
temperature increasing.
Response surface methodology (RSM) is a widely used tech-
nique in solving many industrial problems [39]. It is one of the
most effective and economical solutions for evaluating single and
combined factors of experiment variables that lead to output
responses [40]. The results obtained by RSM analysis will provide
the best system performance for optimized datasets [41]. The main
advantage of this technique is that fewer tests are required and is
less time consuming in contrast to real experimental study [42].
This approach is widely used and has been applied in many inves-
tigative studies especially in the optimization of engine output in
both SI and compression ignition (CI) engines [43].
Studies regarding the investigation of various blend ratios for
butanol blend optimization have been reported by researchers
[44,45]. Yüksel and Yüksel [46] emphasize that one of the central
problems in the successful application of gasoline and alcohol
blends as an alternative fuel is the realization of a stable homoge-
neous liquid phase. To achieve this main goal, several optimization
techniques under RSM such as central composite design, Taguchi
method, and factorial design, may be utilized. A growing body of
literature has been investigated on the application of RSM to inter-
nal combustion engines. An extensive study on engine perfor-
mance and emissions optimization using RSM was conducted by
Najafi et al. [47]. Two parameters, viz., engine speed and fuel
blends, were varied and the responses obtained were brake power,
torque, brake specific fuel consumption, CO, CO2, HC, and NOx,
respectively. The main objective of their study was to gauge the
finest conditions for superior performance and fewer exhaust
emissions. According to their results, the optimal values were
found at an engine speed of 3000 rpm and a blend of 10% bioetha-
nol with 90% of gasoline. The output responses indicated under
those particular conditions for brake power, torque, brake specific
fuel consumption, and CO, CO2, HC, and NOx were found to be
35.26 kW, 103.66 Nm, 0.25 kg/kWh, 3.5 vol.%, 12.8 vol.%,
136.6 ppm, and 1300 ppm, respectively. Atmanlı et al. [45] per-
formed a similar series of optimization using blend ratios of diesel
fuel, n-butanol and cotton oil in a ternary blend in a diesel engine.
Khoobbakht et al. [48] investigated the potential for energy recov-
ery in a four-cylinder, four-stroke, direct injection diesel engine
using blended ethanol and biodiesel in diesel fuel. The experiment
was designed to optimize the exergy efficiency of the blended fuel
varied with engine load and speed using a central composite rotat-
able design of RSM. Their results showed that the highest exergy
efficiency, of up to 36.92%, occurred for the fuel blend of 0.08 L
of ethanol and 0.17 L of biodiesel added to 1 L of diesel. Ganapathy
et al. [49] performed a two-zone Weibe’s heat release function to
simulate Jatropha biodiesel engine performance by using Taguchi’s
optimization method. Bharadwaz et al. [50] carried out an experi-
ment to study the effects of biodiesel-methanol blending, com-
pression ratio, and load to improve the engine performance as
well as to reduce the exhaust emissions in a variable compression
ratio engine. The results revealed that according to the desirability
approach of RSM, the condition that the engine demonstrates the
maximum and minimum emissions is for a fuel blend that com-
prises of 5% methanol in 95% biodiesel by volume with a compres-
sion ratio of 18, and a load of 9.03 kg. Under these operating
conditions, the responses indicated that brake specific fuel con-
sumption, brake thermal efficiency, NOx, CO, HC, and emitted
smoke were 0.37 kg/kWh, 31.95%, 531.23 ppm, 0.036%, 5 ppm,
and 15.35%, respectively. Poompipatpong [51] draws our attention
to the distinction between the use of average value and RSM in
decision support methodology accuracy with regards to the torque
comparison in an engine fuelled by waste plastic pyrolysis oil. The
experimental results implied that the decision support methodol-
ogy using RSM was better than conventional average method.
Optimization via RSM conducted by Pandian et al. [52] showed
that significant interactions between the input factors of injection
pressure, injection timing and nozzle tip protrusion on the engine
performance and emissions characteristics of twin cylinder water
cooled naturally aspirated engine operating with biodiesel, derived
from pongamia seeds. Ileri et al. [53] performed a similar investiga-
tion but with a different type of fuel i.e. canola oil methyl ester.
Hirkude and Padalkar [54] worked on the optimization of the
direct diesel engine with respect to engine performance and engine
emissions of waste fried oil methyl esters particularly on smoke
emissions through RSM.
A considerable amount of literature has been published on opti-
mizing parameters using RSM in the field of internal combustion
engine; see, e.g., [44–54]. However, to the best of the authors’ knowl-
edge, the optimization of a lower blend ratio of butanol-gasoline
blends in an SI engine on high performance, and low exhaust emis-
sions has not been reported yet in the literature. The main advantage
of optimizing the lower fuel blend ratios are the following: no major
and minor modifications are required to be made to the engine for a
low ratio of fuel blend with butanol-gasoline blends [55]; butanol
production is still less productive and has a much higher cost com-
pared to the gasoline fuels [23], which means that using a lower ratio
of 2-butanol is cost-effective in comparison to a higher ratio. More-
over, most studies of butanol-gasoline blends in an SI engine have
also not specifically dealt with the use of butanol from the secondary
butyl alcohol family.
This study aims to contribute to this growing area of butyl fam-
ily usage in SI engines by investigating the influences of fuel blend
and engine speeds towards engine performance and emissions
characteristics using RSM. In view of all that has been mentioned
in the literature review, it is believed that the combined effects
of factors from engine speeds and fuel blends of 2-butanol–
gasoline blends on engine performance and emissions characteris-
tics have yet been explored in great detail. Therefore, the novelty of
this study lies on the combined effects of factors from engine
speeds and fuel blends of 2-butanol–gasoline mixtures for eight
engine performance and emissions responses; brake power, BMEP,
BSFC, BTE, NOx, CO, CO2 and HC. This research article also intends
to optimize the combination of the aforementioned factors to
achieve simultaneous improvement in engine performance and
reduction of exhaust emissions in an SI engine fuelled by buta-
nol–gasoline blends. The optimize fuel blends results are then
compared with the baseline fuels of gasoline fuels (G100). A low
ratio of butanol from the secondary butyl alcohol family was used
in the fuel blends in volumes of 5%, 10%, and 15% in gasoline fuel
operated at 50% of wide throttle open (WTO) varied from 2000 to
4000 rpm with intervals of 500 rpm. It worthwhile to mentioned
that it is a widely held view that RSM is capable of predicting data
that is not measured in experimental analysis [55,56]. Interactive
graph trends for 7.5% and 12.5% volume of 2-butanol in gasoline
blends will be discussed in terms of the effects of different fuel
blends and engine speeds based on the measured trends of the
tested fuels.
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195 181

(a)

(b)
Fig. 1. Schematic diagram of the experimental setup (a) schematic and (b) actual engine.
182 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195

2. Materials and methods KANE Autoplus 5–2 gas analyser. The engine speed and engine
throttle position were controlled by the engine dynamometer con-
2.1. Engine test setup troller inside the engine room. The engine cooling water from the
external tank was arranged to flow and maintain a temperature
In this study, the experiments were conducted on a Mitsubishi of 80 °C. The sensitivity and measurement accuracy of the exhaust
4G93 SOHC (single overhead camshaft) four-cylinder, four-stroke, gas concentration are described in Table 2.
port fuel injection, naturally aspirated spark ignition gasoline
engine. The experimental schematic diagram consisted of an
2.2. Test fuels
engine dynamometer, fuel flow rate, fuel tanks, cooling water tank
for the engine and dynamometer, exhaust gas analyser, and air
The engine was fuelled by commercial gasoline fuel from a local
flow meter as shown in the engine testing schematic diagram of
petroleum station. The 2-butanol with a purity of 99.5% was pur-
Fig. 1(a). The actual engine layout is presented in Fig. 1(b). The
chased from the Merck distributor in Malaysia. Gasoline was set
main engine specifications are reported in Table 1. The engine
as a baseline fuel (G100) and blended fuels of 5%, 10%, and 15%
was coupled with the engine dynamometer by a coupling shaft.
2-butanol in volume in gasoline were denoted as GBu5, GBu10,
A 100 kW Dynalec Controls eddy current dynamometer was used
and GBu15, respectively. Both gasoline and 2-butanol were kept
to measure the brake torque and calculated brake power. The fuel
in the proper containers inside the engine chemical room. Table 3
flow rate was recorded using an AIC fuel flow meter with an accu-
shows the main properties of the tested fuels.
racy of 1%. The fuel line was adapted with a fuel return valve to cir-
culate the excess fuel back into the engine. Air flow rate was
recorded using a Benetech GM8903 hot wire anemometer with 2.3. Engine test procedure
an air speed resolution of 0.001 m/s. The emissions parameters
and relative air–fuel ratio were recorded using a precise calibrated Engine performance and exhaust emissions experimental inves-
tigations were performed at a constant 50% of wide throttle open
(WTO) with engine speeds from 2000 to 4000 rpm at intervals of
Table 1 500 rpm. During the engine tests, cooling water temperature was
The specifications of Mitsubishi 4G93 SOHC engines. adjusted at 80 °C from the outlet engine water. The engine was
Engine descriptions operated until the engine reached a steady state condition. After
the engine warmed up, the measurements for performance and
Bore
stroke 81.0 mm
89.0 mm
Piston displacement 1834 cc exhaust emissions characteristics were taken. Each set of experi-
Compression ratio 9.5:1 mental fuel tests was done on the same day in order to avoid
Fuel injection type ECI-multi (electronically
controlled multi-point fuel injection)
Maximum power 86 kW @ 5500 rpm
Maximum torque 161 Nm @ 4500 rpm Start

Introduction to case studies using RSM

Table 2
List of instruments used and its measurement accuracy.
Experimental preparation and engine testing
Instruments Measurements Measurement
domain accuracy
KANE Autoplus 5–2 gas NOx 0–1500 ppm ±5% or 25 ppm Mathematical modelling and RSM
analyser
CO 0–21% ±5% or 0.06% vol1
CO2 0–16% ±5% or 0.5% vol1 Anova test
HC 0–5000 ppm ±5% or 12 ppm vol1
Dynalec controls eddy Load cell ±0.25%
current dynamometer
Engine speed 0–9999 rpm ±1 rpm
Thermocouple probe 40 °C–1200 °C ±2.5 °C Good No
type K
Agreement?
AIC fuel flow meter 1–80 L/h ±1%

Yes
Model graph 3-D
Table 3
Properties of gasoline and 2-butanol [23,57–59].

Property Gasoline 2-Butanol

Optimization
Molar C/H ratio 0.44–0.50 –
Density (kg/m3) 736 806.3
Low heating value (kJ/kg) 44,300 33,000
Confirmation experimental analysis
Stoichiometric air/fuel ratio 14.6 11.2
RON/MON 95/85 101/92–97
Auto – ignition temperature (°C) 228–470 406.1 End
Boiling point (°C) 27–225 99.5
Heat of vaporization (kJ/kg) 349 551 Analysis of variance (ANOVA)
Flammable limits (vol.%) 1.4–7.6 1.7–9.8
Laminar flame speeds [60] 33 48
Fig. 2. Flow chart of engine optimization using RSM.
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195 183

exhaust emissions data deviations, because the experiments were where X ¼ ðX1 ; X2 . . . ; Xk Þ0 ; fðXÞ is a vector function of p elements
possibly affected by atmospheric humidity and temperature varia- that consists of powers of X1 ; X2 . . . ; Xk up to a certain degree
tions. To increase the reliability of the experimental test results, referred by d ð> 1Þ. For a first ðd ¼ 1Þ and second ðd ¼ 2Þ degree
the engine performance and exhaust emissions test was repeated polynomial, the equation can be written as in (4) and (5).
three times, and the measured values were averaged.
X
k
y ¼ b0 þ bi X i þ e ð4Þ
2.4. Measurement error i¼1

The measurement error of the data collections for both engine X

k XX X
k
performance and exhaust emissions was calculated based on per- y ¼ b0 þ bi X i þ bij X i X j þ bii X 2i þ e ð5Þ
cent relative standard error (RSE), as follows [61]: i¼1 i<j i¼1


SError where y is the predicted response, i is the linear coefficient, j is the
RSE ð%Þ ¼

100% ð1Þ
X quadratic coefficient, b is the regression coefficient, k is the number

is the mean of the data col- of factor and e is a random experimental errors assumed to have a
where SError is the standard error, and X
zero mean [62].
lections. The standard error was calculated based on the following
Analysis of variance (ANOVA) was evaluated from the data
formula:
obtained to prove the significant values between the input vari-
r ables and responses. The quality of the fit polynomial equation
SError ¼ pffiffiffi ð2Þ
n model was expressed by the coefficient of determination R2, and
Fisher’s F-test was used to check the value of significance. The
where r is the standard deviation, and n is the number of data col-
model was evaluated by the probability value (p-value) with a
lections. Data on the percentage of relative standard errors were
95% confidence level. Three-dimensional interactive plots and their
calculated for 3 repeatable readings during engine performance
respective contour plots were obtained based on effects of the two
and exhaust emission data collection. Based upon the percentage
factors (engine speeds and fuel blends) on the responses (brake
of relative standard error calculations for each type of performance
power, torque, brake specific fuel consumption, brake thermal effi-
and exhaust emission, these results featured an error of less than or
ciency, NOx, CO, CO2, and HC). The stepwise of the experimental
equal to 5%.
flow chart is presented in Fig. 2. Experimental design matrix with
input variables and responses was denoted in Table 4.
2.5. Experimental design

The version 7.0 of the Design Expert software was used to
develop the experimental strategy for RSM. The methodological
approach taken in this study is based on RSM. The main idea of
RSM is to use a sequence of designed experiments to obtain an
optimal response. In RSM, the input and response is denoted as
X1 ; X2 . . . ; Xk and ðyÞ, respectively. The relationship can be approx-
imated by a low-degree polynomial model as in Eq. (3).
Y ¼ f ðXÞb þ e
0
ð3Þ
3. Results and discussion
3.1. Analysis of the developed model
The evaluation of the model is analysed using analyses of vari-
ance (ANOVA) presented in Table 5. Thirteen different data points
had been constructed in this study. The output indicated that the
model was significant with p-value values less than 0.05. In ANOVA
analysis, a value of p that is less than 0.05 shows that the factors

Table 4
Experimental design matrix with input variables and responses.

Run order Input variables Responses

Engine speed (rpm) Fuel blend (vol.%) Brake power (kW) BMEP (MPa) BSFC (g/kWh) BTE (%) NOx (ppm) CO (%) CO2 (%) HC (ppm)
1 3000 10 29.3 0.64 287.0 29.0 900 1.9 8.5 50
2 3000 15 30.1 0.66 289.3 29.2 798 1.5 8.9 50
3 4000 10 36.7 0.60 291.0 28.6 1237 4.1 6.3 82
4 3000 10 28.9 0.63 285.1 29.2 905 1.9 8.8 47
5 2000 15 17.3 0.57 293.7 28.8 635 0.1 11.3 23
6 3000 10 30.3 0.66 286.2 29.1 890 1.8 8.7 53
7 3000 10 28.4 0.62 287.2 29.0 895 1.9 9.2 57
8 2000 10 16.7 0.55 291.7 28.6 680 0.3 10.5 32
9 3000 5 28.0 0.61 284.0 29.0 944 2.0 8.7 62
10 4000 5 35.7 0.58 288.0 28.6 1284 4.1 6.0 90
11 4000 15 38.8 0.64 293.7 28.8 1111 3.7 6.6 72
12 3000 10 27.5 0.60 288.1 28.9 910 1.9 9.0 59
13 2000 5 16.4 0.54 288.7 28.5 728 0.5 10.1 53

Table 5
ANOVA results for response parameters.

Responses Brake power BMEP BSFC BTE NOx CO CO2 HC

P <0.0001 0.0028 0.0002 0.0006 <0.0001 <0.0001 <0.0001 <0.0001
F 212.41 11.61 27.98 18.85 690.98 750.42 143.88 33.08
R2 0.9935 0.8924 0.9523 0.9309 0.998 0.9981 0.9664 0.9594
Adj R2 0.9888 0.8155 0.9183 0.8815 0.9965 0.9968 0.9597 0.9304
AP 42.419 10.429 16.879 12.797 84.922 80.888 32.845 20.04
184 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195

Fig. 3. Normal probability plots of residual for (a) brake power, (b) brake mean effective pressure, (c) brake specific fuel consumption, (d) brake thermal efficiency, (e) NOx, (f)
CO, (g) CO2, and (h) HC.
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
endure significant effect at 95% confidence level [63]. A high R2
value, near to 1, is desirable, and a reasonable agreement with
adj R2 is necessary [60]. The lowest value of r was also necessary
to attain the fittest model types. Based on the value of the R2 and
the r suggested by DoE software, the quadratic model best fits
for all of the responses except fir CO2 whereby a linear model is
predicted. Adequate precision (AP) measures the signal-to-noise
ratio and a value greater than 4 is always desirable. It was also
shown that the value responses for an AP greater than 4 indicate
an adequate signal implying that these models can be used to nav-
igate the design space [64]. Based on the ANOVA statistical analysis
of the data in Table 5 revealed that the data shows positive results
of all the model responses. Fig. 3(a)–(h) presents the normal prob-
ability plots of residuals for each response. Taken together, these
results suggest that the data points form a straight line, indicating
that experimental data shows a good correlation, with a normality
assumption of the regression model equation.
3.2. Interactive effect of 2-butanol–gasoline blends and engine speed
3.2.1. Brake power
The first set of analyses examined the impact of engine speeds
and fuel blends on engine performance characteristics. The
response surface profile for a quadratic model of brake power is
Fig. 4. Interactive plot for engine brake power at different engine speeds and fuel blends with 50% of WTO (a) contour plot and (b) 3D plot.
185
shown in Fig. 4(a)’s contour plot and Fig. 4(b)’s three-
dimensional (3D) plot at 50% of WTO. Brake power may be defined
as a branched of engine performance characteristics commonly
referred to usable power delivered by an engine produced by the
product of multiplication of engine angular speed and torque
[65]. The relation between brake power with engine speed (A)
and 2-butanol percentage (B) mention in Eq. (6).
Brake power ¼ 28:87 þ 10:15A þ 1:02B þ 0:55AB  2:07A2 þ 0:23B2
ð6Þ
Fig. 4(b) reveals that there was a steady increase of brake power
with respect to engine speeds and fuel blends. GBu15 resulted in
the highest engine brake power across all engine speeds compared
to other fuels. Furthermore, according to the contour plot in Fig. 4
(a), GBu15 achieved a higher engine brake power of 35 kW at lower
engine speeds in comparison to other blended fuels at higher
engine speeds. This indicates that, although addition of 2-butanol
in gasoline fuels reduced the low heating value, it is believed that
the dominance effects of higher laminar flame speeds from 2-
butanol provided improvement in engine-indicated mean effective
pressure, and thus greater engine output in brake power was
obtained. This view can be supported by findings from Serras-
Pereira et al. [66] that butanol produces a large equivalent flame
radius that is almost similar to gasoline fuel during combustion
186 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
development. This seems to contradict findings previously
reported by Elfasakhany [67], who showed that engine brake
power reduces with the increase of fuel blends.
3.2.2. BMEP
Fig. 5(a)’s contour plot and Fig. 5(b)’s three-dimensional (3D)
plot at 50% of WTO show a quadratic model of BMEP for different
engine speeds and fuel blends. The BMEP is an important product
of the calculation of engine brake torque over engine displacement.
The BMEP produces two different series of trends. As can be seen
from Fig. 5(b), there is a clear trend of increasing BMEP with
increasing increments of 2-butanol percentage in the gasoline
fuels; meanwhile, for BMEP with respect to engine speeds there
is an improvement to a maximum point for all fuel blends, but then
it falls slightly at the end. The maximum BMEP is observed in the
red regions shown in Fig. 5(a) in the range of engine speed
3000–3500 rpm. A possible explanation for these results may be
attributed to the high latent heat of vaporization of 2-butanol indi-
cated in Table 2. When assessing the fuel blends which have a
much larger proportion of 2-butanol, there is a reduction of charge
cooling temperature owing to the increment of latent heat of
vaporization, which therefore gradually increases the BMEP. The
higher octane rating of 2-butanol (101) leads to the higher com-
bustion efficiency of the blended fuels, allowing for advance injec-
tion timing and thus superior BMEP. In a concurrent investigation,
researchers Masum et al. [68] attributed the reasons for the gain of
Fig. 5. Interactive plot for engine brake mean effective pressure at different engine speeds and fuel blends with 50% of WTO (a) contour plot and (b) 3D plot.
the BMEP to increases in combustion pressure due to the outcome
of the enhanced anti-knock behaviour of the alcohol fuels. How-
ever, similar trends cannot be expected at a much higher ratio of
butanol. Gu et al. [58] describe in their findings that 40% and
100% butanol yield a much lower engine brake torque. These draw-
backs are associated with a decrease of low heating value, which
reduces the BMEP. Finally, the relation between BMEP with engine
speed (A) and 2-butanol percentage (B) mention in Eq. (7).
BMEP ¼ 0:63 þ 0:02A þ 0:021B þ 0:0051AB  0:054A2 þ 0:0048B2
ð7Þ
3.2.3. BSFC
Fig. 6(a) and (b) depicts the contour and three-dimensional (3D)
plot at 50% of WTO for a quadratic model of BSFC for different
engine speeds and fuel blends. Apparently, a strong relationship
between BSFC and both fuel blends and engine speeds can be seen
in Fig. 6. Specific fuel consumption is the amount of fuel used with
regard to engine brake power [69]. Eq. (8) showed relation
between BSFC with engine speed (A) and 2-butanol percentage (B).
BSFC ¼ 286:76  0:29A þ 2:61B þ 0:1AB þ 4:47A2  0:23B2 ð8Þ
Fig. 6(b) shows that BSFC is observed to decrease with an
increase in engine speed for all tested fuels, and a further incre-
ment in engine speed results in increasing BSFC. On the other hand,
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
the BSFC is lower for the butanol-gasoline blends with a lower
butanol ratio at all engine speeds. In the contour plot in Fig. 6(a),
at 3000 rpm the trend of increasing BSFC by a range of 285.6 g/
kWh to 288.9 g/kWh for GBu5 to GBu15, respectively, is shown.
Turning now to the experimental possible reason on increase of
BSFC with fuel blend, are due to the low energy content of the les-
ser energy content of butanol, which enhances BSFC when it is
applied directly in the engines at much higher ratio. The findings
of the current study are consistent with those of Varol et al. [70].
In their comprehensive analysis of alcohols including butanol, they
stated that to deliver the same energy output of 1 L of gasoline,
approximately 1.23 L of butanol was needed. Another reason
why higher butanol percentage blends consume much more fuel
is that the butanol ascribed to the lower stoichiometric air – fuel
ratios. Therefore, to maintain the same BMEP, more fuel must be
injected in the combustion chamber for higher percentages of 2-
butanol to achieve the stoichiometric of gasoline fuels, hence
increasing the BSFC.
3.2.4. BTE
Fig. 7(a) and (b) shows the contour and three-dimensional (3D)
plot at 50% of WTO for a quadratic model of BTE for different
engine speeds and fuel blends. Thermal efficiency is a function of
the engine brake power to the heat input of BSFC and the low heat-
ing value [71]. The mathematical definition for BTE can be simpli-
fied based on the following equation:
Fig. 6. Interactive plot for engine brake specific fuel consumption at different engine speeds and fuel blends with 50% of WTO (a) contour plot and (b) 3D plot.
187
3600
BTE ¼
100 ð9Þ
BSFC
LHV
Compared to the BSFC graph trend in Fig. 6(b), BTE is roughly
inversely proportional to the BSFC. As can be seen in Fig. 7(b) the
BTE increased with an increase in engine speed, and then it
decreased at engine speeds of 3000–4000 rpm. From the graph in
Fig. 7(a) we can see the trend of improving BTE in the region of
2500–3500 rpm. What is interesting in this data is that the effi-
ciency for this range of engine speeds improves with successive
increases in percentage of the 2-butanol in the gasoline in the
range of 29–30%. This circumstance can be explained considering
that a faster burning rate resulted from higher laminar flame
speeds of the fuel blends contributed by the 2-butanol, leading to
a shorter combustion duration and hence less useful effective
power losses, thus increasing the BTE. Recent investigation
reported by Tornatore et al. was able to show that gasoline fuel
containing butanol generates faster burning velocity and similar
findings were found in research of Ref. [72]. Eq. (9) showed relation
between BTE with engine speed (A) and 2-butanol percentage (B).
BTE ¼ 29:07 þ 0:040A þ 0:13B  0:007128AB  0:45A2 þ 0:064B2
ð10Þ
3.2.5. Nox
Fig. 8(a) and (b) shows the contour and three-dimensional (3D)
plot at 50% of WTO for a quadratic model of engine exhaust
188 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
emissions of NOx for different engine speeds and fuel blends. The
engine exhaust emissions of NOx formed due to the combustion
temperature, residence time, impact of combustion area, and oxy-
gen concentration as defined in the Zeldovich mechanism [73]. The
NOx are a group of gases that are composed of nitric oxide (NO),
nitrogen dioxide (NO2), nitrous oxide (N2O), dinitrogen trioxide
(N2O3), dinitrogen tetroxide (N2O4), and dinitrogen pentoxide
(N2O5). But the primary group in NOx are NO and NO2 [74]. There
was a significant monotonically increasing trend in NOx exhaust
emissions corresponding to the engine speeds. However, as the
fuel blends increased from GBu5 to GBu15, the NOx emissions
tended to decrease, as shown in Fig. 8(b). From the contour plot
of Fig. 8(a), it can be seen that by far the GBu5 to GBu12.5 range
produced the highest NOx, approximately 1180.1 ppm, shown by
the red region ranging from 3700 to 4000 rpm. Strong evidence
of reduction in NOx exhaust emissions was found when the engine
was operating with GBu15, which indicates that within that range,
Gbu15 yielded approximately 1068.9 ppm of NOx, which was
lower than the other fuel blends by 9.4%. The reductions in NOx
exhaust emissions observed to occur with increasing butanol is
in accordance with the higher vaporization heat typical of alcohols
compounds, which yield lower air-fuel mixture temperatures dur-
ing intake and, as a result, lower peak temperatures in the combus-
tion chamber. This agrees in the result of other studies [75–77]
where smaller amounts of NOx exhaust emissions were observed
for blends of increasing butanol in the gasoline fuels. In contrast
to earlier findings, however, no evidence of major changes was
Fig. 7. Interactive plot for engine brake thermal efficiency at different engine speeds and fuel blends with 50% of WTO (a) contour plot and (b) 3D plot.
found with lower butanol-gasoline blends [58,78,79]. Finally, the
relation between NOx emissions with engine speed (A) and 2-
butanol percentage (B) mention in Eq. (11).
NOx ¼ 898:62 þ 264:83A  68:67B  20AB þ 63:33A2  24:17B2
ð11Þ
3.2.6. CO
Fig. 9(a) and (b) illustrates the contour and three-dimensional
(3D) plot at 50% of WTO for a quadratic model of engine exhaust
emissions of CO for different engine speeds and fuel blends. Gener-
ally, CO is tasteless, colourless, odourless, and highly toxic. The for-
mation of CO occurs from incomplete combustion when there is
insufficient oxygen to fully combust the entire carbon bond in
the fuel to CO2 [80]. It also represents the loss in chemical energy
[81]. Eq. (9) showed relation between CO emissions with engine
speed (A) and 2-butanol percentage (B).
CO ¼ 1:88 þ 1:85A  0:21B  0:028AB þ 0:35A2  0:14B2 ð12Þ
In Fig. 9(b), the quadratic function graph indicates that CO emis-
sions increased with increasing engine speeds, whereas in Fig. 9(a)
the CO emissions decreased with increasing 2-butanol–gasoline
blends. A possible cause for this could be due to the increasing
molecular mass corresponding to the increasing 2-butanol concen-
tration in the gasoline fuels. In addition, butanols are oxygenated
fuels; hence, they generate a leaning effect which reduces CO
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
Fig. 8. Interactive plot for engine exhaust emissions of nitrogen oxides at different engine speeds and fuel blends with 50% of WTO (a) contour plot and (b) 3D plot.
emissions. Balki et al. [82] held the view that decreases in CO emis-
sions are due to the carbon rate in the fuels as well as higher lam-
inar flame speeds which mean the engine has limited time to
complete the combustion cycle.
3.2.7. CO2
Fig. 10(a) and (b) depicts the contour and three-dimensional
(3D) plot at 50% of WTO for a quadratic model of engine exhaust
emissions of CO2 for different engine speeds and fuel blends. Com-
bustion of any carbon-containing material ideally produces water
(H2O) and CO2 [83]. Nevertheless, when a combustion is not
100% conserved, an engine also yields other remaining emissions
such as NOx, CO, and HC. Among all emissions, CO2 is regarded
as the most concerning exhaust emission [84]. This type of emis-
sion is significantly contributing to global warming, and it is men-
tioned in the Kyoto agreement as a major challenge to humankind
[85]. Mitigating the adverse effects of CO2 is a mandatory step for
harmonious stability between humans and the environment. A
much more systematic approach by RSM analysis would closely
identify how CO2 interacts with engine speeds and fuel blend vari-
ables that are believed to be linked to the lean air-fuel mixture. In
Fig. 10(b), the 3D interactive plot indicates that the CO2 emissions
in this engine produce an increasing trend 2–butanol additions.
The effects of the fuel blends are clearly visible in Fig. 10(a):
189
increasing the proportion of 2-butanol caused an increase in CO2
concentration. It is possible to hypothesise that these conditions
are likely to occur due to the increase of engine BTE as in Fig. 6
when it is operated with a higher 2-butanol percentage. Besides
that, another possible explanation for this result is that the higher
carbon-to-hydrogen ratio for higher percentages of 2-butanol pro-
duces a more complete combustion, hence yielding more CO2 and
enhancing the engine BTE performance. Elfasakhany [86] pre-
sented similar results by adding butanol in gasoline fuels. Those
results revealed that at a constant throttle position ranging from
engine speeds of 2600 rpm to 3400 rpm, CO2 emissions were the
highest specifically for 10% butanol in the gasoline fuels. The pre-
sent findings seem to be consistent with other research which have
also found that these increases in CO2 emissions can be attributed
to the extra oxygen in the alcohol fuel that allows partial reduction
of the CO by forming into CO2 [75]. Finally, the relation between
CO2 emissions with engine speed (A) and 2-butanol percentage
(B) mention in Eq. (13).
CO2 ¼ 8:66  2:17A þ 0:33B ð13Þ
3.2.8. HC
Fig. 11(a) and (b) depicts the contour and three-dimensional
(3D) plot at 50% of WTO for a quadratic model of engine exhaust
emissions of HC for different engine speeds and fuel blends.
190 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
Fig. 9. Interactive plot for engine exhaust emissions of carbon monoxides at different engine speeds and fuel blends with 50% of WTO (a) contour plot and (b) 3D plot.
HC emissions are primarily due to incomplete combustion of
hydrocarbon fuel. Exhaust emissions of HC are high for rich fuel
mixtures from a stoichiometric air-fuel ratio because of insuffi-
cient oxygen for the oxidation reaction [87]. Fig. 11(b) exhibits
the emissions of HC by all tested fuels, which increased with
engine speeds, while in Fig. 11(a) the trends indicate that emis-
sions of HC reduced with rising increments of 2-butanol in gaso-
line blends. These emissions were sharply lower, especially for
GBu12.5 to GBu15. This trend can be clearly seen at the range of
engine speeds from 3500 to 4000 rpm in the yellow regions. In
those particular regions the emissions of HC contributed by
GBu12.5 as well as GBu15 were significantly lower than GBu10,
GBu7.5 and GBu5 by approximately 13.9%. As the throttle position
does not change, the amount of intake alcohol fuel is about two
times higher when compared to the 2-butanol–gasoline blends.
Therefore, it seems that these results are due to oxygenated alco-
hol that homogenized the air-fuel mixture. This condition in turn
further improved combustion efficiency; hence we can see in
Fig. 10 that more CO2 was produced, whereas the amount of HC
emissions decreased due to better combustion in correspondence
to the higher 2-butanol concentrations in the gasoline fuels. It is
therefore likely that such connections exist between higher oxy-
gen content in the cylinder and exhaust caused by the oxygen con-
tent of the blended fuel. Finally, the relation between HC
emissions with engine speed (A) and 2-butanol percentage (B)
mention in Eq. (14).
HC ¼ 53:24 þ 22:67A  10:00B þ 3:00AB þ 3:66A2 þ 2:66B2 ð14Þ
3.3. Optimization
The optimization process is governed by the desirability param-
eter in which its objective function lies between 0.0 and 1.0 [88].
The overall goal is essentially to obtain a combined desirability
value for all of the factors and responses. The criteria for the opti-
mization process, namely the goal set for each response, lower and
upper limits used, weights used, and the importance of the factors,
are presented in Table 6. The importance of the responses can be
varied from 1 to 5 that indicates the least to the most significant
response, respectively. However, if the value of importance is the
same for all of the responses, the value of the objective function
will be reduced to the standard conditions for desirability. The
point prediction based on the criteria to determine optimum per-
formance and emissions characteristics is based on Table 7.
Through combined desirability-based approach, only one best
solution is obtained. A maximum desirability of 0.8 was found at
the following input: 15 vol.% of 2-butanol and 85 vol.% of gasoline
fuel (GBu15) and at 3205 rpm of engine speed. This suggests that
the best setting for all of the responses that does not conflict with
each other was achieved at the aforementioned conditions. As for
the optimal input parameters, the values of the brake power,
BMEP, BSFC, BTE, NOx, CO, CO2, and HC were found to be
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
Fig. 10. Interactive plot for engine exhaust emissions of carbon dioxides at different engine speeds and fuel blends with 50% of WTO (a) contour plot and (b) 3D plot.
32.2 kW, 0.66 MPa, 289.3 g/kWh, 29.2%, 858.7 ppm, 1.9%, 8.5% and
51.3 ppm, respectively. The results of the optimum condition for
GBu15 fuel was then compared with the similar condition of
G100 baseline fuel in Table 8. From this data, it is apparent that
the engine performance at 50% of WTO for GBu15 produces
approximately lower brake power, BMEP and higher BSFC but
improved in BTE by 5.3%, 4.3%, 2.2%, and 1.7%, respectively. Mean-
while, for emissions, the GBu15 provides lower NOx, CO, and HC
but higher CO2 by 17%, 26.9, 25.7% and 5.9%, respectively. In gen-
eral, GBu15 fuel produces almost similar engine performance to
the G100 fuel. It is interesting to note that, with GBu15, the engine
would benefit in terms of the emissions characteristics. This is pri-
marily due to the desirable fuel properties of butanol fuels namely
the higher vaporization properties as well as the oxygen content of
butanol that affects the emission characteristics of the NOx as well
as the CO and HC, respectively [31,75–77,89] The higher vaporiza-
tion allows the reduction of the combustion temperature which in
turn reduce the NOx emission. Meanwhile, the additional oxygen
content improves the combustion of GBu15 and thus significantly
reduces the emissions of CO and UHC.
3.4. Confirmation test results
In order to validate the optimized results, the experiments
were performed at the optimal condition of engine speed
191
3205 rpm and fuel blend of GBu15. There was a significant posi-
tive correlation between all of the output responses at the opti-
mum conditions, which indicates a combined desirability value
of 0.8. A maximum desirability value close to 1 represents an
ideal response [54]. In our cases, the value of 0.8 signifies that
there is a good trade-off regions existed that satisfies the model
criteria. The reason a confirmation test was done was to validate
the model’s accuracy. Kumar et al. [44] mentions that confirma-
tory trials are not compulsory to conduct if the RSM models pro-
duce a prediction error of less than 5%. Since our prediction error
was higher than 5%, the confirmation test was carried out. For the
actual responses, the mean of three measured results was calcu-
lated. Summary of the predicted values, the mean of experimen-
tal calculated values, and percentage of absolute error is
presented in Table 9. The formula for percentage of absolute error
is based on Eq. (15).
Percentage of Absoulete Error ð%Þ


Actual value  Predicted value
¼
100% ð15Þ
Actual value
It was observed that the percentage of absolute error in predic-
tion of developed models was in acceptable agreement. The valida-
tion of results indicated that the models developed were quite
accurate for performance characteristics with a percentage of
192 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195

Fig. 11. Interactive plot for engine exhaust emissions of unburned hydrocarbon at different engine speeds and fuel blends with 50% of WTO (a) contour plot and (b) 3D plot.

Table 6
Optimization criteria of responses for engine performance and exhaust emissions.

Parameter Limits Criterion Importance

Lower Upper
Fuel blend (vol.%) 5 15 In range 3
Engine speed (rpm) 2000 4000 In range 3
Brake power (kW) 16.4 38.8 Maximize 3
BMEP (MPa) 0.54 0.66 Maximize 5
BSFC (g/kWh) 284.0 293.7 Minimize 5
BTE (%) 28.5 29.2 Maximize 5
NOx (ppm) 635 1284 Minimize 5
CO (%) 0.1 4.1 Minimize 3
CO2 (%) 6 11.3 Minimize 5
HC (ppm) 23 90 Minimize 3

absolute error lower than 3%. However, for exhaust emissions charac-
teristics, the percentage of absolute error, particularly for CO and HC,
was a little higher: by 9.5% and 10%, respectively. This was perhaps
due to the lack of combined desirability value, which was contributed
by lack of experimental numbers. It is recommended for future
experimental analyses using RSM that additional repeatable experi-
ments must be done to improve the combined desirability values.
4. Conclusions
In this investigation, the aim was to assess the optimum engine
performance and exhaust emissions of a four-cylinder, four-stroke
spark ignition engine operating on 2-butanol–gasoline blended
fuels investigated experimentally. From the results of this study,
the following conclusions are drawn:
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195 193

Table 7
Point prediction of engine performance and exhaust emissions characteristics optimization.

Responses Combined
Factors Desirability
Performance Emissions

0.8

Table 8
Comparison at optimal conditions with gasoline baseline fuels.

Engine speed (rpm) Fuel type Brake power (kW) BMEP (MPa) BSFC (g/kWh) BTE (%) NOx (ppm) CO (%) CO2 (%) HC (ppm)
3205 GBu15 32.2 0.66 289.3 29.2 858.7 1.9 8.5 51.3
G100 34 ± 0.7% 0.69 ± 0.7% 282.6 ± 0.3% 28.7 ± 0.3% 1035 ± 0.3% 2.6 ± 4.5% 8 ± 4.2% 69 ± 4.8%

Table 9
Validation test results.

Experiment Engine Fuel Brake BMEP BSFC (g/ BTE (%) NOx CO (%) CO2 (%) HC
no. speed (rpm) type power (MPa) kWh) (ppm) (ppm)
(kW)
1 3205 GBu15 Predicted 32.2 0.66 289.3 29.2 858.7 1.9 8.5 51.3
Actual 33.3 ± 2.4% 0.68 ± 2.5% 292 ± 0.6% 28.9 ± 0.6% 874 ± 0.3% 2.1 ± 4.5% 8.25 ± 1.4% 57 ± 1.7%
Percentage of 3.3 2.9 0.9 1 1.75 9.5 3 10
absolute error (%)

1. ANOVA analysis indicates that all of the models were found to
be statistically significant.
2. Performance characteristics indicated that adding a higher 2-
butanol proportion, of up to 15%, increased the brake power,
BMEP, and BTE; however, this also resulted adversely in a
higher BSFC.
3. Exhaust emissions characteristics showed that the presence of
2-butanol of up to 15% reduced the NOx, CO and HC but yielded
higher CO2.
4. The desirability approach of response surface methodology was
found to be an efficient optimization technique. A high desir-
ability of 0.8 was obtained for optimum engine performance
and exhaust emissions characteristics.
5. The solutions obtained using the desirability approach specified
that the most optimum condition was at engine speeds of
3205 rpm and fuel blends of GBu15.
6. A comparison of the two results between G100 and GBu15
revealed that the engine performance at 50% of WTO for
GBu15 produces approximately lower brake power, BMEP and
194 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
higher BSFC but improved in BTE by 5.3%, 4.3%, 2.2%, and 1.7%,
respectively. On the other hand, for emissions, the GBu15 pro-
vides lower NOx, CO, and HC but higher CO2 by 17%, 26.9,
25.7% and 5.9%, respectively.
7. Results validated by confirmatory experiments indicated that
the models developed using RSM for brake power, BMEP, BSFC,
BTE, NOx, and CO2 were adequate to describe the effects of
engine speed and fuel blend; and the error in prediction was
found to be less than 5%. However, for CO and HC the prediction
error was found to be a little higher, by less than 10%.
Overall, the RSM optimization technique can be used for any
combination of fuel blend of interest to find out the best blend
ratios for engine performance and exhaust emissions parameters.
Response surface methodology was very helpful in designing the
experiment, and the statistical analysis helped to identify the sig-
nificant parameters which most influence performance and emis-
sion characteristics. This experimental design considerably
reduced the time required by minimizing the number of experi-
ments to be performed and provided statistically proven models
for all the responses.
Acknowledgement
Appreciation and acknowledgment are due to the Ministry of
Higher Education (KPT) for providing the author with the scholar-
ship under My Brain 15 schemes and financial support under
University Malaysia Pahang Grand RDU 1403156. Sincere thanks
to Mr. Muhd Hafietz Bin Yusoff for bountiful assistance in terms
of technical support during the engine testing. Finally, the authors
thank the anonymous referees and the editor for carefully reading
this paper and making many helpful suggestions on improving the
original manuscript.
References
[1] Banos R, Manzano-Agugliaro F, Montoya F, Gil C, Alcayde A, Gómez J.
Optimization methods applied to renewable and sustainable energy: a
review. Renew Sustain Energy Rev 2011;15(4):1753–66.
[2] Xue J, Grift TE, Hansen AC. Effect of biodiesel on engine performances and
emissions. Renew Sustain Energy Rev 2011;15(2):1098–116.
[3] Demirbas MF, Balat M, Balat H. Biowastes-to-biofuels. Energy Convers Manage
2011;52(4):1815–28.
[4] Luo Y, Guda V, Wijayapala R, Steele PH. Upgrading of syngas hydrotreated
fractionated oxidized bio-oil to transportation grade hydrocarbons. Energy
Convers Manage 2016;115:159–66.
[5] Zarifi F, Mahlia T, Motasemi F, Shekarchian M, Moghavvemi M. Current and
future energy and exergy efficiencies in the Iran’s transportation sector. Energy
Convers Manage 2013;74:24–34.
[6] Ozan C, Haldenbilen S, Ceylan H. Estimating emissions on vehicular traffic
based on projected energy and transport demand on rural roads: policies for
reducing air pollutant emissions and energy consumption. Energy Pol 2011;39
(5):2542–9.
[7] Wan Ghazali WNM, Mamat R, Masjuki HH, Najafi G. Effects of biodiesel from
different feedstocks on engine performance and emissions: a review. Renew
Sustain Energy Rev 2015;51:585–602.
[8] Qu L, Wang Z, Zhang J. Influence of waste cooking oil biodiesel on oxidation
reactivity and nanostructure of particulate matter from diesel engine. Fuel
2016;181:389–95.
[9] Yusri IM, Mamat R, Azmi WH, Najafi G, Sidik NAC, Awad OI. Experimental
investigation of combustion, emissions and thermal balance of secondary
butyl alcohol-gasoline blends in a spark ignition engine. Energy Convers
Manage 2016;123:1–14.
[10] Sajjadi B, Raman AAA, Arandiyan H. A comprehensive review on properties of
edible and non-edible vegetable oil-based biodiesel: composition,
specifications and prediction models. Renew Sustain Energy Rev
2016;63:62–92.
[11] Agarwal AK. Biofuels (alcohols and biodiesel) applications as fuels for internal
combustion engines. Prog Energy Combust Sci 2007;33(3):233–71.
[12] Abbaszaadeh A, Ghobadian B, Omidkhah MR, Najafi G. Current biodiesel
production technologies: a comparative review. Energy Convers Manage
2012;63:138–48.
[13] Roberts A, Brooks R, Shipway P. Internal combustion engine cold-start
efficiency: a review of the problem, causes and potential solutions. Energy
Convers Manage 2014;82:327–50.
[14] Balat M. Production of bioethanol from lignocellulosic materials via the
biochemical pathway: a review. Energy Convers Manage 2011;52(2):858–75.
[15] Azadi H, Jongb Sd, Deruddera B, Maeyera PD, Witlox F. Bitter sweet how
sustainable is bio-ethanol production in Brazil. Renew Sustain Energy Rev
2012;16:3599–603.
[16] Surisetty VR, Dalai AK, Kozinski J. Alcohols as alternative fuels: an overview.
Appl Catal A 2011;404(1):1–11.
[17] Bayindir H, Yücesu H, Aydin H. The effects of k and e on engine performance
and exhaust emissions using ethanol–unleaded gasoline blends in an SI
engine. Energy Sourc, Part A: Recov, Util, Environ Eff 2010;33(1):49–56.
[18] Zhen X, Wang Y. An overview of methanol as an internal combustion engine
fuel. Renew Sustain Energy Rev 2015;52:477–93.
[19] Huang Y, Hong G. Investigation of the effect of heated ethanol fuel on
combustion and emissions of an ethanol direct injection plus gasoline port
injection (EDI+GPI) engine. Energy Convers Manage 2016;123:338–47.
[20] Uyumaz A. An experimental investigation into combustion and performance
characteristics of an HCCI gasoline engine fueled with n-heptane, isopropanol
and n-butanol fuel blends at different inlet air temperatures. Energy Convers
Manage 2015;98:199–207.
[21] Demirbas A. Political, economic and environmental impacts of biofuels: a
review. Appl Energy 2009;86(1):S108–17.
[22] Balat M, Balat H, Öz C. Progress in bioethanol processing. Prog Energy Combust
Sci 2008;34(5):551–73.
[23] Jin C, Yao M, Liu H, Lee C-FF, Ji J. Progress in the production and application of
n-butanol as a biofuel. Renew Sustain Energy Rev 2011;15(8):4080–106.
[24] No S-Y. Application of biobutanol in advanced CI engines – a review. Fuel
2016;183:641–58.
[25] Okoli CO, Adams TA, Brigljević B, Liu JJ. Design and economic analysis of a
macroalgae-to-butanol process via a thermochemical route. Energy Convers
Manage 2016;123:410–22.
[26] Fortman J, Chhabra S, Mukhopadhyay A, Chou H, Lee TS, Steen E, et al. Biofuel
alternatives to ethanol: pumping the microbial well. Trends Biotechnol
2008;26(7):375–81.
[27] Galloni E, Fontana G, Staccone S, Scala F. Performance analyses of a spark-
ignition engine firing with gasoline–butanol blends at partial load operation.
Energy Convers Manage 2016;110:319–26.
[28] Jang YS, Malaviya A, Cho C, Lee J, Lee SY. Butanol production from renewable
biomass by clostridia. Bioresour Technol 2012;123:653–63.
[29] Merola SS, Valentino G, Tornatore C, Marchitto L. In-cylinder spectroscopic
measurements of knocking combustion in a SI engine fuelled with butanol–
gasoline blend. Energy 2013;62:150–61.
[30] Elfasakhany A. Experimental study of dual n-butanol and iso-butanol additives
on spark-ignition engine performance and emissions. Fuel 2016;163:
166–74.
[31] Feng R, Fu J, Yang J, Wang Y, Li Y, Deng B, et al. Combustion and emissions
study on motorcycle engine fueled with butanol-gasoline blend. Renew Energy
2015;81:113–22.
[32] Chen Z, Yang F, Xue S, Wu Z, Liu J. Impact of higher n-butanol addition on
combustion and performance of GDI engine in stoichiometric combustion.
Energy Convers Manage 2015;106:385–92.
[33] Wallner T, Miers SA, McConnell S. A comparison of ethanol and butanol as
oxygenates using a spark ignition direct injection, spark ignition engine. J Eng
Gas Turbines Power 2009;131.
[34] Thewes M, Müther M, Brassat A, Pischinger S, Sehr A. Analysis of the effect of
bio-fuels on the combustion in a downsized DI SI Engine. SAE Int J Fuels
Lubricants 2012. 2011-01-1991.
[35] Alasfour FN. Butanol - a single-engine study: availability analysis. Appl Therm
Eng 1997;17(6):537–49.
[36] Alasfour F. The effect of using 30% iso-butanol-gasoline blend on hydrocarbon
emissions from a spark-ignition engine. Energy Sourc 1999;21(5):379–94.
[37] Alasfour F. NOx emission from a spark ignition engine using 30% iso-butanol–
gasoline blend: part 1—preheating inlet air. Appl Therm Eng 1998;18
(5):245–56.
[38] Alasfour F. NOx emission from a spark ignition engine using 30% iso-butanol–
gasoline blend: Part 2—ignition timing. Appl Therm Eng 1998;18(8):609–18.
[39] Namvar-Asl M, Soltanieh M, Rashidi A, Irandoukht A. Modeling and
preparation of activated carbon for methane storage I. Modeling of activated
carbon characteristics with neural networks and response surface method.
Energy Convers Manage 2008;49(9):2471–7.
[40] Ma L, Han Y, Sun K, Lu J, Ding J. Optimization of acidified oil esterification
catalyzed by sulfonated cation exchange resin using response surface
methodology. Energy Convers Manage 2015;98:46–53.
[41] Dharma S, Masjuki H, Ong HC, Sebayang A, Silitonga A, Kusumo F, et al.
Optimization of biodiesel production process for mixed Jatropha curcas–Ceiba
pentandra biodiesel using response surface methodology. Energy Convers
Manage 2016;115:178–90.
[42] Silva V, Rouboa A. Combining a 2-D multiphase CFD model with a Response
Surface Methodology to optimize the gasification of Portuguese biomasses.
Energy Convers Manage 2015;99:28–40.
[43] Patel PD, Lakdawala A, Patel RN. Box-Behnken response surface methodology
for optimization of operational parameters of compression ignition engine
fuelled with a blend of diesel, biodiesel and diethyl ether. Biofuels 2016;7
(2):1–14.
[44] Kumar BR, Saravanan S, Rana D, Nagendran A. Combined effect of injection
timing and exhaust gas recirculation (EGR) on performance and emissions of a
DI diesel engine fuelled with next-generation advanced biofuel–diesel blends
 I.M. Yusri et al. / Energy Conversion and Management 133 (2017) 178–195
using response surface methodology. Energy Convers Manage
2016;123:470–86.
[45] _
Atmanlı A, Yüksel B, Ileri E, Karaoglan AD. Response surface methodology
based optimization of diesel–n-butanol–cotton oil ternary blend ratios to
improve engine performance and exhaust emission characteristics. Energy
Convers Manage 2015;90:383–94.
[46] Yüksel F, Yüksel B. The use of ethanol–gasoline blend as a fuel in an SI engine.
Renew Energy 2004;29(7):1181–91.
[47] Najafi G, Ghobadian B, Yusaf T, Ardebili SMS, Mamat R. Optimization of
performance and exhaust emission parameters of a SI (spark ignition) engine
with gasoline–ethanol blended fuels using response surface methodology.
Energy 2015;90:1815–29.
[48] Khoobbakht G, Akram A, Karimi M, Najafi G. Exergy and energy analysis of
combustion of blended levels of biodiesel, ethanol and diesel fuel in a DI diesel
engine. Appl Therm Eng 2016;99:720–9.
[49] Ganapathy T, Murugesan K, Gakkhar R. Performance optimization of Jatropha
biodiesel engine model using Taguchi approach. Appl Energy 2009;86
(11):2476–86.
[50] Bharadwaz YD, Rao BG, Rao VD, Anusha C. Improvement of biodiesel methanol
blends performance in a variable compression ratio engine using response
surface methodology. Alexandria Eng J 2016;55(2):1201–9.
[51] Poompipatpong C, Kengpol A. Design of a decision support methodology using
response surface for torque comparison: an empirical study on an engine
fueled with waste plastic pyrolysis oil. Energy 2015;82:850–6.
[52] Pandian M, Sivapirakasam S, Udayakumar M. Investigation on the effect of
injection system parameters on performance and emission characteristics of a
twin cylinder compression ignition direct injection engine fuelled with
pongamia biodiesel–diesel blend using response surface methodology. Appl
Energy 2011;88(8):2663–76.
[53] Ileri E, Karaoglan AD, Atmanli A. Response surface methodology based
prediction of engine performance and exhaust emissions of a diesel engine
fuelled with canola oil methyl ester. J Renew Sustain Energy 2013;5(3):1–19.
[54] Hirkude JB, Padalkar AS. Performance optimization of CI engine fuelled with
waste fried oil methyl ester-diesel blend using response surface methodology.
Fuel 2014;119:266–73.
[55] Bejaoui MA, Beltran G, Aguilera MP, Jimenez A. Continuous conditioning of
olive paste by high power ultrasounds: response surface methodology to
predict temperature and its effect on oil yield and virgin olive oil
characteristics. LWT-Food Sci Technol 2016;69:175–84.
[56] Hamid HA, Jenidi Y, Thielemans W, Somerfield C, Gomes RL. Predicting the
capability of carboxylated cellulose nanowhiskers for the remediation of
copper from water using response surface methodology (RSM) and artificial
neural network (ANN) models. Ind Crops Prod 2016;93:108–20.
[57] Balki MK, Sayin C. The effect of compression ratio on the performance,
emissions and combustion of an SI (spark ignition) engine fueled with pure
ethanol, methanol and unleaded gasoline. Energy 2014;71:194–201.
[58] Gu X, Huang Z, Cai J, Gong J, Wu X, Lee C-F. Emission characteristics of a spark-
ignition engine fuelled with gasoline-n-butanol blends in combination with
EGR. Fuel 2012;93:611–7.
[59] Christensen E, Yanowitz J, Ratcliff M, McCormick RL. Renewable oxygenate
blending effects on gasoline properties. Energy Fuels 2011;25(10):4723–33.
[60] Nithyanandan K, Wu H, Huo M, Lee C-F. A preliminary investigation of the
performance and emissions of a port-fuel injected SI engine fueled with
acetone-butanol-ethanol (ABE) and gasoline. SAE technical paper 2014; 2014-
01-1459.
[61] Peters CA. Statistics for analysis of experimental data. Environ Eng Process Lab
Manual 2001:1–25.
[62] Ghafari S, Aziz HA, Isa MH, Zinatizadeh AA. Application of response surface
methodology (RSM) to optimize coagulation–flocculation treatment of
leachate using poly-aluminum chloride (PAC) and alum. J Hazard Mater
2009;163(2):650–6.
[63] Rajmohan T, Palanikumar K. Application of the central composite design in
optimization of machining parameters in drilling hybrid metal matrix
composites. Measurement 2013;46(4):1470–81.
[64] Hubadillah SK, Othman MHD, Harun Z, Ismail A, Iwamoto Y, Honda S, et al.
Effect of fabrication parameters on physical properties of metakaolin-based
ceramic hollow fibre membrane (CHFM). Ceram Int 2016;42(14):15547–58.
[65] Abedin MJ, Masjuki HH, Kalam MA, Sanjid A, Rahman SMA, Fattah IMR.
Performance, emissions, and heat losses of palm and jatropha biodiesel blends
in a diesel engine. Ind Crops Prod 2014;59:96–104.
195
[66] Serras-Pereira J, Aleiferis P, Richardson D, Wallace S. Characteristics of ethanol,
butanol, iso-octane and gasoline sprays and combustion from a multi-hole
injector in a DISI engine. SAE Int J Fuels Lubricants 2008;1:893–909.
[67] Elfasakhany A. Experimental investigation on SI engine using gasoline and a
hybrid iso-butanol/gasoline fuel. Energy Convers Manage 2015;95:398–405.
[68] Masum B, Kalam M, Masjuki H, Palash S, Fattah IR. Performance and emission
analysis of a multi cylinder gasoline engine operating at different alcohol–
gasoline blends. RSC Adv 2014;4(53):27898–904.
[69] Mat Yasin MH, Yusaf T, Mamat R, Yusop A Fitri. Characterization of a diesel
engine operating with a small proportion of methanol as a fuel additive in
biodiesel blend. Appl Energy 2014;114:865–73.
[70] Varol Y, Öner C, Öztop H, Altun Sß. Comparison of methanol, ethanol, or n-
butanol blending with unleaded gasoline on exhaust emissions of an SI engine.
Energy Sourc, Part A: Recov, Util, Environ Effects 2014;36(9):938–48.
[71] Fattah IR, Masjuki H, Kalam M, Mofijur M, Abedin M. Effect of antioxidant on
the performance and emission characteristics of a diesel engine fueled with
palm biodiesel blends. Energy Convers Manage 2014;79:265–72.
[72] Szwaja S, Naber JD. Combustion of n-butanol in a spark-ignition IC engine. Fuel
2010;89(7):1573–82.
[73] Masumn BM, Masjuki HH, Kalam MA, Fattah IMR, Palash SM, Abedin MJ. Effect
of ethanol–gasoline blend on NOx emission in SI engine. Renew Sustain Energy
Rev 2013;24:209–22.
[74] Normann F, Andersson K, Leckner B, Johnsson F. Emission control of nitrogen
oxides in the oxy-fuel process. Prog Energy Combust Sci 2009;35(5):385–97.
[75] Wigg B, Coverdill R, Lee C-F, Kyritsis D. Emissions characteristics of neat
butanol fuel using a port fuel-injected, spark-ignition engine. SAE technical
paper 2011; 2011-01-0902.
[76] Singh SB, Dhar A, Agarwal AK. Technical feasibility study of butanol–gasoline
blends for powering medium-duty transportation spark ignition engine.
Renew Energy 2015;76:706–16.
[77] Puli D, Ravi Kumar P. Performance and emission characteristics of Tertiary
Butyl Alcohol gasoline blends on a spark ignition engine. Biofuels 2015;6(1-
2):71–8.
[78] Costagliola MA, De Simio L, Iannaccone S, Prati MV. Combustion efficiency and
engine out emissions of a S.I. engine fueled with alcohol/gasoline blends. Appl
Energy 2013;111:1162–71.
[79] Dernotte J, Mounaim-Rousselle C, Halter F, Seers P. Evaluation of butanol–
gasoline blends in a port fuel-injection, spark-ignition engine. Oil Gas Sci
Technol – Revue de l’Institut Français du Pétrole 2009;65(2):345–51.
[80] Daniel R, Tian G, Xu H, Wyszynski ML, Wu X, Huang Z. Effect of spark timing
and load on a DISI engine fuelled with 2,5-dimethylfuran. Fuel 2011;90
(2):449–58.
[81] Masum BM, Masjuki HH, Kalam MA, Palash SM, Wakil MA, Imtenan S.
Tailoring the key fuel properties using different alcohols (C2–C6) and their
evaluation in gasoline engine. Energy Convers Manage 2014;88:382–90.
[82] Balki MK, Sayin C, Canakci M. The effect of different alcohol fuels on the
performance, emission and combustion characteristics of a gasoline engine.
Fuel 2014;115:901–6.
[83] Ganesan V. Internal combustion engines. McGraw Hill Education (India) Pvt
Ltd; 2012.
[84] Demirbas A, Demirbas MF. Importance of algae oil as a source of biodiesel.
Energy Convers Manage 2011;52(1):163–70.
[85] Olah GA, Goeppert A, Prakash GS. Chemical recycling of carbon dioxide to
methanol and dimethyl ether: from greenhouse gas to renewable,
environmentally carbon neutral fuels and synthetic hydrocarbons. J Org
Chem 2008;74(2):487–98.
[86] Elfasakhany A. Experimental study on emissions and performance of an
internal combustion engine fueled with gasoline and gasoline/n-butanol
blends. Energy Convers Manage 2014;88:277–83.
[87] Paul A, Bose PK, Panua RS, Banerjee R. An experimental investigation of
performance-emission trade off of a CI engine fueled by diesel–compressed
natural gas (CNG) combination and diesel–ethanol blends with CNG
enrichment. Energy 2013;55:787–802.
[88] Hatami M, Cuijpers MCM, Boot MD. Experimental optimization of the vanes
geometry for a variable geometry turbocharger (VGT) using a Design of
Experiment (DoE) approach. Energy Convers Manage 2015;106:1057–70.
[89] He B-Q, Chen X, Lin C-L, Zhao H. Combustion characteristics of a gasoline
engine with independent intake port injection and direct injection systems for
n-butanol and gasoline. Energy Convers Manage 2016;124:556–65.