Accepted Manuscript A comprehensive study on measurement and prediction of viscosity of biodieseldiesel-alcohol ternary blends Mert Gülüm, Atilla Bilgin PII:
S0360-5442(18)30152-X
DOI:
10.1016/j.energy.2018.01.123
Reference:
EGY 12235
To appear in:
Energy
Received Date: 17 July 2017 Revised Date:
17 January 2018
Accepted Date: 25 January 2018
Please cite this article as: Gülüm M, Bilgin A, A comprehensive study on measurement and prediction of viscosity of biodiesel-diesel-alcohol ternary blends, Energy (2018), doi: 10.1016/j.energy.2018.01.123. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
11
ACCEPTED MANUSCRIPT 1
A comprehensive study on measurement and prediction of viscosity of biodiesel-diesel-
2
alcohol ternary blends
3
Mert Gülüm a,*, Atilla Bilgin a
4
a
5
ABSTRACT
6
Recently, biodiesel is receiving global attention as an alternative fuel for diesel engines because
7
of its many advantages compared to petroleum-based diesel fuel. Despite many advantages, it
8
has also some drawbacks such as: poor flow characteristics, lower energy content, higher price
9
and viscosity. Higher viscosity leading poor atomization with larger droplet size results in bad
10
mixing quality and incomplete combustion. Blending of biodiesel with diesel or alcohol is one of
11
the most appropriate techniques to overcome high viscosity problem. Although viscosity data of
12
biodiesel-diesel fuel binary blends at different temperatures are abundantly available in the
13
literature, these of ethyl ester-diesel-alcohol ternary blends including especially higher alcohols
14
are still insufficient. Therefore, in this study, waste cooking oil ethyl ester (biodiesel) was
15
produced by means of transesterification, and it was blended with diesel fuel and different
16
alcohols (methanol, ethanol, isopropanol, n-butanol and n-pentanol). Viscosities of the prepared
17
ternary blends were measured at different temperatures accordingly DIN 53015. Finally, a
18
rational model depending on blend temperature or alcohol fraction was proposed for estimating
19
viscosities of them, and it was tested with the well-known other models (Arrhenius, Kendall–
20
Monroe, Andrade and Vogel) for its predictive ability.
21
Keywords: Ethyl ester, Biodiesel, Viscosity, Ternary blend, Prediction, Models.
22
*Corresponding Author Tel.: +90 462 377 29 63, Fax: +90 462 377 33 36
23
E-mail address:
[email protected] (Mert GULUM)
24 25
AC C
EP
TE D
M AN U
SC
RI PT
Department of Mechanical Engineering, Karadeniz Technical University, Trabzon, Turkey
12
ACCEPTED MANUSCRIPT 26 27 28 29 30 31 32
Nomenclature and units :
Average absolute error (%)
ARE
:
Average relative error (%)
a, b, c
:
Regression constants
33
BF
:
Base fuel (i.e. 20% waste cooking oil biodiesel-80% diesel fuel)
34
Bu2, Bu4, Bu6, Bu8, Bu10, Bu15, Bu20 : Waste cooking oil biodiesel-diesel-butanol ternary blends
35
BD
: Waste cooking oil biodiesel (ethyl ester)
36
B20D75E5
:
20% rapeseed oil biodiesel-75% diesel-5% bioethanol
37
B15D75E10
:
15% rapeseed oil biodiesel-75% diesel-10% bioethanol
38
B10D75E15
:
10% rapeseed oil biodiesel-75% diesel-15% bioethanol
39
B5D75E20
:
5% rapeseed oil biodiesel-75% diesel-20% bioethanol
40
B25D70E5
:
25% rapeseed oil biodiesel-70% diesel-5% bioethanol
41
B20D70E10
:
20% rapeseed oil biodiesel-70% diesel-10% bioethanol
42
B15D70E15
:
15% rapeseed oil biodiesel-70% diesel-15% bioethanol
43
B10D70E20
:
10% rapeseed oil biodiesel-70% diesel-20% bioethanol
44
B5D70E25
:
5% rapeseed oil biodiesel-70% diesel-25% bioethanol
45
CFPP
46
CN
47
D85B10E5
:
85% rape oil biodiesel-10% diesel-5% ethanol
48
D70B25E5
:
70% rape oil biodiesel-25% diesel-5% ethanol
49
D80B10E10
:
80% rape oil biodiesel-10% diesel-10% ethanol
50
DF
: Diesel fuel
51
E2, E4, E6, E8, E10, E15, E20
: Waste cooking oil biodiesel-diesel-ethanol ternary blends
52
HHV
: Higher heating value (kJ/kg)
53
M2, M4, M6, M8, M10, M15, M20 :
TE D
M AN U
SC
RI PT
AAE
AC C
EP
: Cold filter plugging point (℃) : Cetane number
Waste cooking oil biodiesel-diesel-methanol ternary blends
13
ACCEPTED MANUSCRIPT 54
IV
55
Pe2, Pe4, Pe6, Pe8, Pe10, Pe15, Pe20 : Waste cooking oil biodiesel-diesel-pentanol ternary blends
56
Pr2, Pr4, Pr6, Pr8, Pr10, Pr15, Pr20
57
R
: Correlation coefficient
58
SD
: Standard deviation
59
SN
: Saponification number
60
T
: Temperature (K)
x , x , x , … , x
63
X
64
Greek symbols
65
μ
69 70 71 72 73 74 75 76 77 78
ν
ρ
RI PT
: Dynamic viscosity (g/m ∙ s ≡ mPa ∙ s ≡ cP)
: Kinematic viscosity (mm /s ≡ cSt)
: Density (kg⁄m ), (g⁄cm )
TE D
68
: Alcohol fraction (v/v)
EP
67
: Independent variables
AC C
66
: Uncertainties of independent variables
SC
62
w , w , w , … , w
: Waste cooking oil biodiesel-diesel-isopropanol ternary blends
M AN U
61
: Iodine value
14
ACCEPTED MANUSCRIPT 79
1.
Introduction As a result of the industrialization of societies and population growing, the greater
81
consumption of energy, faster dwindling of world petroleum reserves, increase in fossil fuel
82
prices and enactment of more stringent environmental standards could be expected in the near
83
future [1,2]. These events have forced to find renewable clean alternatives to fossil fuels.
84
Biodiesel has become one of the most important alternative biofuels that can be used without or
85
little modification in diesel engines [3].
RI PT
80
Chemically speaking, biodiesel is defined as a mixture of mono-alkyl esters obtained from
87
(edible and non-edible) oils and animal fats [4,5]. It is usually produced by the transesterification
88
of triglycerides in oils with a mono-hydroxyl alcohol (methanol, ethanol, propanol, etc.) in the
89
presence of a catalyst [6]. The most important parameters affecting transesterification are
90
catalyst amount and type, reaction temperature, reaction time, mixing intensity, alcohol to oil
91
molar ratio and oil type [7,8]. Preference of the alcohol used in transesterification should be
92
made by considering its cost and performance [9]. Although methanol is worldwide preferred,
93
the use of ethanol also ensures several advantages such as reduction of toxicity and
94
environmental concerns due to the renewable character of ethanol [10]. Moreover, ethanol, as an
95
extraction solvent, has much higher dissolving power for oils than methanol [11]. For these
96
reasons, recently, ethanol is taking attention as a suitable alcohol for transesterification reaction
97
[11,12]. Ethyl esters have higher heating value and cetane number than methyl esters because of
98
the extra carbon atom supplied by the ethanol molecule. In spite of its benefits, the use of ethanol
99
also results in the formation of the more stable emulsion after transesterification making very
100
difficult phase separation of esters [13]. However, the addition of extra glycerine to the mixture
101
occurring after the transesterification reaction is found to be helpful in glycerine separation [14].
102
In addition, ethyl esters have poorer low-temperature properties than methyl esters [15].
AC C
EP
TE D
M AN U
SC
86
15
ACCEPTED MANUSCRIPT
As a fuel, biodiesel offers many benefits over petroleum diesel fuel as: (i) it is renewable,
104
non-toxic, non-aromatic and biodegradable fuel, (ii) it has higher cetane number and flash point,
105
and (iii) it significantly reduces pollutant emissions (CO, PM, HC, etc.) [16,17]. Moreover, the
106
use of biodiesel affects the CO2 balance by limiting its accumulation in the biosphere, which
107
may be considered as one of the principal causes of the greenhouse effect [18].
RI PT
103
Although these properties make biodiesel an ideal fuel in a way, biodiesel has also some
109
shortcomings such as higher cloud and pour points because of its higher viscosity, lower heating
110
value and volatility, and generally higher NOx emissions [19-22]. In addition, the high cost of
111
biodiesel as another shortcoming is the main impediment to its commercialization. It costs about
112
1.5 times higher than diesel fuel depending on sources of feedstock oils [23]. Although the
113
higher viscosity of biodiesel may eventually result in higher spray penetration into the
114
combustion chamber [24], there are more adverse effects including poor atomization, larger
115
droplet size, bad mixing quality and low combustion efficiency, causing to decrease engine
116
performance and increase exhaust emissions [25,26]. Even though viscosity measurement of
117
biodiesel is not very difficult, simple and reliable models are greatly necessary for researchers
118
[27] because they ensure both a rapid estimation of viscosity and helps in simulation and
119
optimisation of process equipment such as heat exchangers, reactors, mixing vessels and process
120
piping, as well as development of better combustion models [27,28]. In fact, in literature, many
121
studies have been performed to measure and predict viscosities of ternary blends (biodiesel-
122
diesel fuel-alcohol or biodiesel-vegetable oil-alcohol) at different temperatures [29-35].
123
However, in these studies, (1) ethanol was generally preferred rather than higher alcohols
124
(propanol, n-butanol and n-pentanol), thereby there is still lack of reliable measured viscosity
125
data for ternary blends including especially ethyl ester and higher alcohols at different
126
temperatures, (2) the well-known models (Arrhenius, Andrade, Kay, Vogel, Yuan, Ceriani, etc.)
127
were generally evaluated to predict viscosities and there is a scarcity of new reliable models in
AC C
EP
TE D
M AN U
SC
108
16
ACCEPTED MANUSCRIPT
the existing literature depending on temperature or alcohol fraction, and finally (3) in the existing
129
literature, generally the upper limit of alcohol fraction in a ternary blend is high and the effect of
130
alcohol fraction on viscosity is examined with a relatively large step size. However, as well-
131
known, the cetane number of ternary mixture decreases with increasing alcohol content due to
132
the lower cetane number of alcohol, resulting in extending the duration of ignition delay and
133
diesel knocking. Therefore, alcohol fraction in ternary blend should be less than about 20% to
134
avoid decreasing of the cetane number to undesirable levels, and the effect of alcohol fraction on
135
viscosity of ternary blends should be examined in relatively smaller step size. As an example
136
approving the above mentioned claims, Barabás and Todorut [29] measured viscosities of
137
rapeseed oil biodiesel (methyl ester)-diesel fuel-bioethanol ternary blends at the temperature
138
range of 273.15-343.15 K, and they evaluated the well-known models (Vogel-Fulcher-Tammann
139
equation, Kay’s rule, Refutas equation, etc.) for estimating the viscosity of them. In their study
140
[29], they changed bioethanol fraction from 5% to 25% in the blend with a step size of 5%.
M AN U
SC
RI PT
128
To overcome the mentioned scarcities in the literature above, in this study, waste cooking
142
oil biodiesel was produced by means of ethanolysis reaction. Then, it was blended with diesel
143
fuel at a volume ratio of 20% because European strategy for the security of energy supply sets
144
20% substitution of conventional fuels by biodiesel in the road transport sector by 2020
145
according to European Directive 2003/30/EC published in 2003 [36]. The blend of 20% ethyl
146
ester + 80% diesel fuel was taken in this study as a base fuel (BF), and then the base fuel was
147
mixed with 2, 4, 6, 8, 10, 15 and 20% volume ratios of methanol, ethanol, isopropanol, n-butanol
148
and n-pentanol to prepare ternary fuel blend. The resulting blends were named to reflect their
149
composition. For example, the name Pr2 indicates a blend consisting of 2% isopropanol + 98%
150
base fuel. For this blend, the real percentages of biodiesel and diesel fuel become 19.6% and
151
78.4%, respectively. Similarly, the ternary blends including 2% of methanol, ethanol, butanol
152
and pentanol were denoted as M2, E2, Bu2, Pe2. Similar naming and fractions are also valid for
AC C
EP
TE D
141
17
ACCEPTED MANUSCRIPT
the other ternary blends. Kinematic viscosities of the prepared ternary blends were measured at
154
273.15, 278.15, 283.15, 288.15, 293.15, 303.15, 313.15, 323.15, 333.15, 343.15 K (i.e. 0, 5, 10,
155
15, 20, 30, 40, 50, 60 and 70 ℃) in accordance with the international DIN standards. In other
156
words, totally 422 viscosity points were determined at different temperatures for all test fuels
157
including ternary blends, base fuel (BF), diesel fuel (DF), waste cooking oil biodiesel (BD) and
158
pure alcohols. Finally, the rational model as a function of alcohol fraction (X) or temperature (T)
159
was proposed to predict viscosities of them. The model was also tested against the experimental
160
data reported in the other studies [29,35] and it was compared to well-known models previously
161
suggested in the literature.
162
2.
163 164 165
2.1. Biodiesel production
166
over methyl ester as explained in ‘Introduction’ section. Waste cooking oil was obtained from a
167
canteen in Karadeniz Technical University. Ethanol, sodium hydroxide and anhydrous sodium
168
sulphate were of analytical grades. Transesterification reaction parameters were determined
169
according to the detailed parametric investigation in the authors’ previous study as: 1.25%
170
catalyst concentration, 70℃ reaction temperature, 120 minutes reaction time and 12:1 alcohol/oil
171
molar ratio [37]. More details regarding biodiesel production can be found in [19,22].
172
2.2. Density and viscosity measurements
M AN U
Materials and methods
SC
RI PT
153
AC C
EP
TE D
In this study, waste cooking oil ethyl ester was produced because of its many advantages
173
The densities of all fuels were measured at different temperatures by means of pycnometer
174
accordingly ISO 4787 using a top loading balance with an accuracy of ±0.01 g. More details of
175
the measurements were given in [19,22,38].
176
Dynamic viscosities of them were also measured using universal Haake Falling Ball Viscometer,
177
Haake Water Bath, a stopwatch (±0.001 s) and a thermometer (±0.5 ℃) according to DIN 53015.
18
ACCEPTED MANUSCRIPT 178
Details about the viscosity measurement were also given in [19,22,39]. All measurements were
179
repeated three times, and then average of them was taken to minimize the measurement errors.
180
The kinematic viscosity was computed by dividing dynamic viscosity to density at the same
181
temperature. If μ"#$%#&'&( and ρ"#$%#&'&( are in the units of (mPa. s) and (kg/L), respectively, then ν"#$%#&'&( is obtained in unit of (mm /s).
183
Saponification number (SN), iodine value (IV) and cetane number (CN) of BD were determined
184
using Eqs. (1-3) [40-42]. /01∙2
SN = ∑# . 45 3 6
186
IV = ∑# .
187
CN =
