Anda di halaman 1dari 10

Fuel 89 (2010) 15731582

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Combustion of n-butanol in a spark-ignition IC engine


S. Szwaja a,b,*, J.D. Naber a
a
b

Michigan Technological University, 1400 Townsend Dr., Houghton, MI 49931, USA


Czestochowa University of Technology, ul. Dabrowskiego 69, 42-200 Czestochowa, Poland

a r t i c l e

i n f o

Article history:
Received 13 April 2009
Received in revised form 26 June 2009
Accepted 25 August 2009
Available online 9 September 2009
Keywords:
Butanol
Gasoline
Combustion
Internal combustion engine

a b s t r a c t
Alcohols, because of their potential to be produced from renewable sources and because of their high
quality characteristics for spark-ignition (SI) engines, are considered quality fuels which can be blended
with fossil-based gasoline for use in internal combustion engines. They enable the transformation of our
energy basis in transportation to reduce dependence on fossil fuels as an energy source for vehicles. The
research presented in this work is focused on applying n-butanol as a blending agent additive to gasoline
to reduce the fossil part in the fuel mixture and in this way to reduce life cycle CO2 emissions. The impact
on combustion processes in a spark-ignited internal combustion engine is also detailed. Blends of n-butanol to gasoline with ratios of 0%, 20%, and 60% in addition to near n-butanol have been studied in a single
cylinder cooperative fuels research engine (CFR) SI engine with variable compression ratio manufactured
by Waukesha Engine Company. The engine is modied to provide air control and port fuel injection.
Engine control and monitoring was performed using a target-based rapid-prototyping system with electronic sensors and actuators installed on the engine [1]. A real-time combustion analysis system was
applied for data acquisition and online analysis of combustion quantities. Tests were performed under
stoichiometric air-to-fuel ratios, xed engine torque, and compression ratios of 8:1 and 10:1 with spark
timing sweeps from 18 to 4 before top dead center (BTDC). On the basis of the experimental data, combustion characteristics for these fuels have been determined as follows: mass fraction burned (MFB) prole, rate of MFB, combustion duration and location of 50% MFB. Analysis of these data gives conclusions
about combustion phasing for optimal spark timing for maximum break torque (MBT) and normalized
rate for heat release. Additionally, susceptibility of 20% and 60% butanolgasoline blends on combustion
knock was investigated. Simultaneously, comparison between these fuels and pure gasoline in the above
areas was investigated. Finally, on the basis of these conclusions, characteristic of these fuel blends as
substitutes of gasoline for a series production engine were discussed.
2009 Elsevier Ltd. All rights reserved.

1. Introduction
Butanol or butyl alcohol can be demonstrated to work in the IC
engine designed for use with gasoline without modication. It can
be produced from biomass (biobutanol) as well as fossil fuels (petrobutanol). Both biobutanol and petrobutanol have the same
chemical properties. Butanol is less corrosive than ethanol and

Abbreviations: A/F, air-to-fuel ratio; BTDC, before top dead center; CA, h, crank
angle; CFR, cooperative fuels research; COV, coefcent of variance; CR, compression ratio; ER, equivalence ratio; IC, internal combustion; IMEP, indicated mean
effective pressure; LHV, lower heating value; MBT, maximum break torque; MFB,
mass fraction burned; MON, motored octane number; NIMEP, net indicated mean
effective pressure; PON, pump octane number; RON, research octane number; SI,
spark-ignited; WOT, wide open throttle.
* Corresponding author. Address: Czestochowa University of Technology,
Mechanical Engineering and Computer Science Department, ul. Dabrowskiego 69,
42-200 Czestochowa, Poland. Tel./fax: +48 343250555.
E-mail addresses: szwaja@imtits.pcz.czest.pl, sszwaja@mtu.edu (S. Szwaja).
0016-2361/$ - see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2009.08.043

has a higher energy content than ethanol, similar energy content


to gasoline. In comparison to ethanol, butanol is less prone to
water contamination. As a result it could be distributed using the
same infrastructure used to transport gasoline. It can be used alone
in an internal combustion engine, or it can be mixed with gasoline.
Four butyl alcohols can be distinguished. They all have the same
chemical composition consisting of four carbon atoms, 10 hydrogens and single oxygen and examined by identical chemical pattern C4H10O, but they differ each from others with respect to
their structure. They are as follows:





1-butanol: (n-butanol) CH3CH2CH2CH2OH,


sec-butanol: CH3CH(OH)CH2CH3,
tert-butanol: (CH3)3COH,
iso-butanol: CH3(CH2)3OH.

In addition, each of the fuels has different thermodynamic properties and combustion characteristics. For the tests described in the

1574

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

paper n-butanol (1-butanol) was applied as a fuel. Characteristics


of n-butanol in comparison to gasoline and other alcohol fuels
are given in Table 1.
However, when taking into account the latent heat of vaporization of these fuels, butanol is less attractive than others. For port
fuel injection systems, when the fuel vaporizes in the inlet port it
affects a temperature decrease of the intake charge. Therefore,
fuels of higher latent heat of vaporization have larger decreases
in temperature of intake charge with complete vaporization in
the intake port. This increases the density of combustible mixture
and increases the charge mass. Furthermore, the cost of butanol
production is higher in comparison with ethanol [2]. However,
there are some promising circumstances for butanol production
from fermentation process of agricultural feedstock by cellulosic
enzymes [3] that have the potential to reduce its production cost.
Investigation of butanol usage as the engine fuel has been conducted by several research groups. It has been used both for blending gasoline and diesel fuel for IC engines [4]. Among others
Alasfour [5] has conducted extensive studies. He conducted research on n-butanol combusted in an SI engine with particular focus on NOx emission and availability analysis as functions of intake
temperature, air-to-fuel ratio and ignition timing. Yacoub et al. [6]
performed several studies on application of alcohols C1C5 (methanol to pentanol) as fuels blended with gasoline for the SI CFR
Waukesha engine. He determined optimal spark timings and upper
limits for compression ratio and knock resistance for these fuels.
His studies showed that n-butanol was more prone to generate
combustion knock than gasoline UTG-96. Additionally, he examined exhaust emission from these blends. Further research in this
eld was conducted by Gautam et al. and presented in [7,8]. He
worked with blends of 10% of the ve-alcohols mixture (from
methanol to pentanol) with gasoline UTG-96 with an 87 motored
octane rating and tested these fuels at the stoichiometric ratios
and steady state conditions. He determined that the maximum
compression ratio for these blends was 10:1 with respect to combustion knock. There have also been studies in the eld of butanol
transport properties with respect to the application for an IC engine. For instance, Aleiferis et al. [9] conducted research on sprays
of ethanol, butanol, and gasoline in a direct injected spark-ignited
engine. Dagaut and Togbe [10] examined theoretical and experimental analyses of combustion mechanism of n-butanolgasoline
mixtures, conducted several combustion tests of butanolgasoline
at ratio of 85%/15%, and studied oxidation mechanism using a jet
stirred reactor. They reported good agreement between experimental results and the computations of detailed chemical kinetic
scheme for the n-butanolgasoline blend. Furthermore, results of
various alcohols combustion were presented by Cooney et al.
[11]. They investigated the application of ethanol and butanol
blends with gasoline as fuels for a series engine working without
modications to an engine control unit. They conducted these tests
at partial loads, with torque 100 and 150 Nm. At lower loads, they
reported engine conversion efciencies were similar. As they presented in their work, at higher engine loads up to 150 Nm, maximum fuel conversion efciency changed by 4% from 0.36 to

0.377 at torque of 150 Nm with n-butanol 85 and ethanol 85,


respectively, but the difference in efciency between gasoline
and n-butanol 85 was small (approximately 0.002). In such insignicant changes, the error analysis for plots would have been helpful to draw proper conclusions. They also concluded that efciency
drop between n-butanol and gasoline was caused due to lower octane rating of n-butanol, even though knock combustion was not
observed. On the other hand, the ame speed of n-butanol was
the highest among the other applied fuels, so, theoretically faster
combustion should increase the efciency if combustion timing
was adjusted.
Although there are several valuable works concerning n-butanolgasoline-blended fuels combustion in SI engines, there is limited information of combustion characteristics over a range of
blends of n-butanol and gasoline including 100% n-butanol. Therefore, the combustion studies in this work including characteristics
of the mass fraction burned and heat release provide additional
important insights into the application of n-butanolgasoline
blends.
2. Experimental setup
The engine used for this research is a single cylinder CFR (cooperative fuels research) engine manufactured by the Waukesha Motor Company. The engine was chosen for its versatility and
robustness of construction which is important because of the
inclusive studies of combustion knock. A specialized attribute of
this engine is the ability to vary the compression ratio without disassembling the engine. Characteristics of the test bed and engine
are shown in Fig. 1 and listed in Table 2, respectively. Numerous
modications were made to the engine to meet the requirements
in these studies. This included changing the compression ratio of
the engine by modifying the piston. The modied piston increased
the range of compression ratio that could be studied from 410 to
5.418.5.
The experiments were conducted by sweeping the spark timing
and n-butanol concentration at constant engine speed of 900 rpm,
constant net indicated loads of 330 and 650 kPa mean effective
pressure, and stoichiometric equivalence ratio. Spark timing was
sweep from 18 to 4 before TDC. n-Butanol concentrations of 0%,
20%, 60%, and 100% n-butanol were studied, and compression ratios of 8:1 and 10:1 were examined. Real-time combustion analysis
in the test cell was available through a DSP ACAP combustion system. In addition, data for 300 engine cycles at resolution of 1024
samples/revolution was acquired and logged for later off-line analysis. Engine control and monitoring was performed using a targetbased rapid-prototyping system [1] with the electronic sensors and
actuators installed on the engine. Mototrons Mototune was used
as the calibration tool to record engine control parameters including intake manifold pressure, air ow-rate, fuel injection pressure,
injection duration, etc.
Four fuels of different n-butanol to gasoline fraction were tested
in this investigation. The rst fuel was pure gasoline with a PON
87. It was chosen as a reference fuel to make comparisons with

Table 1
Characteristics of several fuels [6,7].
Fuel

Chemical
formula

Specic
gravity
(kg/
dm3)

Lower heating
value (MJ/kg)

Stoichiometric air
fuel ratio (kgair/
kgfuel)

Energy density of a
stoichiometric airfuel
mixture (MJ/kg)

Latent heat of vaporization


(at boiling point) (kJ/kg)

Octane number
(RON + MON)/2

Methanol
Ethanol
n-Butanol
Gasoline regular
PON 87

CH3OH
C2H5OH
C4H9OH
CH1.87

0.7913
0.7894
0.8097
0.7430

20.08
26.83
32.01
42.9

6.43
8.94
11.12
14.51

2.750
2.699
2.641
2.769

1098
838
584
349

99
100
86
87

1575

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

Fig. 1. Diagram of the CFR engine test bench.

Table 2
CFR engine characteristics.
Compression ratio
Bore (cm)
Stroke (cm)
Connecting rod length (cm)
Displacement (cm3)
IVO
IVC
EVO
EVC

5.418.5
8.26
11.43
25.40
611
10 ATDC
34 ABDC
40 BBDC
15 ATDC

other fuel blends. The second fuel was a blend of 20% n-butanol
with the 87 PON gasoline. Thus the LHV, A/F and energy density
for this 20% blend is close to that of gasoline. The third fuel applied
for the tests was a mixture of 60% n-butanol and 40% of the 87 PON.
Finally the fourth fuel was neat n-butanol.
On the basis of the data from Table 1, the characteristics of
these fuels were determined and are presented in Table 3.
During these tests environment parameters were as follows:
 Intake temperature: 30 C.

Table 3
Characteristics of blends applied for investigation.
Fuel no.

Fuel composition (volume


percentage)

Specic gravity
(kg/dm3)

Lower heating
value (MJ/kg)

Stoichiometric
airfuel ratio

Energy density of a stoichiometric


airfuel mixture (MJ/kg)

100% Gasoline
0% Butanol
80% Gasoline
20% Butanol
40% Gasoline
60% Butanol
0% Gasoline
100% Butanol

0.7430

42.90

14.5

2.77

0.7563

40.57

13.8

2.74

0.7830

36.14

12.4

2.70

0.8097

32.01

11.1

2.64

2
3
4

1576

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

Table 4
Tests parameters.
Test no.

Butanol percentage

Compression ratio (CR)

Spark timing (ST)


( BTDC)

Load NIMEP (kPa)

Number of work cycles


recorded/engine speed (rpm)

Air-to-fuel (per Table 3)

1
2
3
4

100
0, 20, 60, 100
100
100

10
10
8, 10
8

4, 8, 10, 14, 18
10
10
10

330
330
330
650 (WOT)

300/900
300/900
300/900
300/900

Stoichiometric
Stoichiometric
Stoichiometric
Stoichiometric

 Fuel temperature: 29 C.
 Temperature of the engine cooling water: 8590 C.
Table 4 summarizes the cases that were investigated in these
works.

3. Results and discussion


The study concentrated on combustion process of the fuel
blends. Thus, analysis of heat release expressed by normalized
mass fraction burnt (MFB) and net heat production rate examined
by rate of normalized MFB were particularly of the interest. The
analysis was focused on four issues, which are discussed separately
in the following sections:





Spark timing (ST) impact.


Butanol percentage in blends.
Compression ratio (CR) impact.
Load impact.

3.1. Spark timing impact


The analysis in this section concerns test no. 1 with 100% nbutanol with variable spark timing. Comparison of this test with
pure gasoline combustion is conducted in further section at constant spark timing. Fig. 2 shows the spark timing inuence on incylinder combustion pressure. The combustion pressure, presented
in Fig. 2, is the mean trace of 300 consecutive engine work cycles. It
is clearly seen that peak pressure increases with spark timing (ST)
change from 4 to 18 before top dead center (BTDC) of the piston.
It shows no misres with spark timing up to 4 BTDC for 20%,
60% n-butanol blends and pure butanol. The second limit for advanced spark timing was set by rst symptoms of combustion
knock, which were observed at spark timing (ST) higher 14 BTDC.

2500

2000

1500

Q ch

0.8

BTDC

Fuel: n-butanol
100%
CR=10
ST=18 BTDC

ST=10 BTDC
ST=4

ST=18 BTDC
ST=14 BTDC
BTDC


dV
1
dp

dh
V
c  1 dh c  1 dh

where c is the polytropic index, p is the in-cylinder pressure, V is the


cylinder volume, and h is the engine crank angle.
Assuming that primary notion is Qch(h), from which both MF and
MT are determined, it can be stated that the ratio MF upon MT is
considered as normalized MFB with limits from 0 to 1.
Fig. 4a depicts, considered as combustion duration, the 1090%
MFB, which was calculated from the normalized MFB traces from
Fig. 3 as a difference between 0.9 MFB and 0.1 MFB location (when
express these values as percentage they correspond to 90% MFB
and 10% MFB, respectively) measured by a crank position referring
to the TDC. Combustion duration of neat n-butanol and gasoline
got shorter with spark timing advancing due to higher mean temperature of combustion and smaller combustion chamber size at
50% MFB, which approaches to TDC. This can been seen in Fig. 3
for the case of 18 BTDC spark advance in comparison to the other
normalized MFB traces. The 18 spark advance case is centered
around TDC, with the other cases have the bulk burn durations
occurring at later timing. One should note, however, that the 18
and 14 spark advance cases were over-advanced with respect to
spark timing for optimum combustion phasing. Fig. 4b presents

ST=18
bTDC
Fuel:
n-butanol
100%
data2
CR=10
data3

ST=8

Z 
h

Normalized MFB

In-cylinder Pressure kPa

3000

The tests were done at engine partial load expressed by net


indicated mean effective pressure (NIMEP) of 330 kPa. The main
aim of sweeping spark timing was to nd out the optimal spark
timing with respect to the highest indicated efciency with nonknocking conditions. Thus, normalized MFB plots, which can express heat release from combustion, were determined from the
each of the pressure traces and described in Fig. 3.
In this analysis the heat transfer to walls and fuel ow into crevices were neglected. Therefore, the chemical energy Qch released
by combustion can be easily determined on the basis of the rst
law of thermodynamics as a function of crank angle h following
[12].

0.6

ST=14 BTDC
ST=10 BTDC
ST=8 BTDC

0.4

ST=4 BTDC

1000

NIMEP=330kPa
500
-30

0.2

NIMEP=330kPa
-20

-10

0
10
Crank Angle deg

20

30

Fig. 2. In-cylinder pressure vs. crank angle with various spark timing (ST) for nbutanol and compression ratio (CR) 10.

0
-30

-20

-10
0
10
Crank Angle deg

20

Fig. 3. Normalized MFB vs. crank angle after top dead center.

30

1577

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

(a)

0.12

20

n-Butanol 100%
CR = 10
NIMEP = 330kPa

CR = 10
NIMEP = 330kPa
Gasoline PON 87

Max. Rate of MFB

CA deg

18
n-Butanol

10-90 % MFB

16

0.08

0.06

12

0.04
6

10

14

vs 50%MFB

0.1

14

vs Spark Timing

-2

18

10

14

18

Spark Timing deg BTDC, 50%MFB deg ATDC

Spark Timing deg BTDC

Fig. 6. Maximum rate of MFB vs. spark timing.

CR = 10
NIMEP = 330kPa

0-10 % MFB

CA deg

14

Gasoline PON 87

25
10%, 50%, 90% MFB and Max.Rate
MFB CA deg

(b)

15

13
12
11
n-Butanol

10
9
8
2

10

14

18

Spark Timing deg BTDC

n-butanol 100%, CR = 10
NIMEP = 330kPa
10% MFB
50% MFB
90% MFB
Max. Rate of MFB

20
15
10

5
0
2

10

14

18

22

-5
-10
Spark Timing deg BTDC

Fig. 4. (a) 1090% MFB as combustion duration vs. spark timing, (b) 010% MFB vs.
spark timing for n-butanol and gasoline PON 87.

0.1

Fuel: n-butanol 100%


CR=10

Fig. 7. A 10%, 50%, 90% MFB and maximum rate of MFB location as function of spark
timing.

NIMEP=330kPa

ST=18 BTDC

Rate of MFB

0.08

ST=14 BTDC
ST=10 BTDC

0.06

ST=8 BTDC
ST=4 BTDC

0.04

0.02

0
-30

-20

-10

0
10
Crank Angle deg

20

30

Fig. 5. Rate of MFB vs. spark timing.

Fig. 8. Indicated efciency vs. spark timing and location of 50% MFB.

rst stage of combustion 010% MFB. As plotted, the 010% MFB


changes insignicantly with spark timing for both n-butanol and
gasoline. However, longer times are observed with spark timing
more advanced because of lower temperature in pre-ignition
phase. The lower temperature of combustible mixture, then reactions rates are lower, so it suppresses combustion kernel to be
grown rapidly at this early phase. In the spark-ignited engine with

a at piston a squeezing effect by the piston is marginal therefore


laminar ow is mostly observed around a sparking plug. Thus,
front ame propagation, which can be expressed by laminar ame
speed of the combustible mixture there, strongly depends on temperature, and it is also lower.
In addition, in comparison the gasoline 1090% normalized MFB
is only longer by 3/4 to 1. Rate of normalized MFB presented in

1578

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

Fig. 5 and further in Figs. 14, 17 and 20 was calculated as crank angle based derivative of the normalized MFB. The rate of normalized
MFB provides a metric for the heat release rate during combustion.
Faster combustion, caused by the over-advanced ignition (spark
timing ST = 14 and higher), leads to higher rate of MFB as depicted
in Fig. 5. Fig. 6 shows maximal rate of MFB against spark timing
and additionally it shows this quantity as a function of 50% MFB.
The maximal rate of over 0.1 was reached at spark timing 18 BTDC
and it was located almost at TDC of the piston. These advanced
spark timings create conditions favorable for generating combustion knock. For these tests combustion knock was observed in tests
with spark timings of 18 and 14 BTDC. The knock that occurs at
these combustion phasings signicantly speeds up that combus-

(a)

1.6
1.4

Gasoline PON 87

COV of NIMEP

1.2

gi

1
0.8

n-Butanol

0.6
0.4
0.2
0
0

10

12

14

16

18

20

Spark Timing deg BTDC

(b)
n-butanol 100%
CR=10
ST=10 deg BTDC

pdf

0.1

0.05

tion phase (as seen in Fig. 4 and causing higher rate of MFB) which
is located at the end phase of combustion as it is shown in Fig. 7.
The solid line in Fig. 7 expresses location of maximum rate of
MFB. With proper combustion, in the absence of combustion
knock, the maximum rate of MFB coincides with the location of
50% MFB. This is the case for the range of spark timings between
4 and 10 BTDC. For the spark timings of 14 BTDC and more advanced, the location of the maximum rate of MFB approaches to
the location of 90% MFB as a result of the rapid combustion that occurs during combustion knock in these cases. Similar characteristics of combustion knock were observed for gasoline combustion.
With respect to combustion knock under these conditions, n-butanol and gasoline show similar behavior.
These tests were conducted at a constant NIMEP of 330 kPa. In
addition, the fuel to air ratio was maintained at the fuel blends
stoichiometric condition, so, fuel mass ow-rate (MF in g/s) was
known and the indicated efciency gi for the engine work cycle
can be calculated following the equation.

Normal distribution
Mean=329.4
Sigma=2.983
R=0.93
NIMEP COV=0.91

0
300

310

320

330

340

350

rpm  NIMEP  V s
120  LHVfuel  MFfuel

where rpm is the engine crankshaft revolutions per minute, LHFfuel


is the lower heating value of fuel in MJ/kg, MFfuel is the mass owrate of fuel (g/s), NIMEP is the net indicated mean effective pressure
in kPa, and Vs is the engine displacement volume (m3).
The efciency was determined for both the n-butanol and gasoline fueled engine. As discussed further, in Section 3.2 entitled
n-butanol percentage in blends, n-butanol combusts faster than
gasoline at the same conditions, thus, through approaching the
combustion to the ideal constant volume (process) makes indicated efciency higher in this way. The maximum of the efciency
for the n-butanol fueled engine appears at spark timing of approximately 10 BTDC as plotted in Fig. 8. Thus, that spark timing was
taken for next tests.
Next, comparison of combustion stability, evaluated by applying coefcient of variance (COV) of net IMEP, was done for the tests
performed at spark timing sweep from 4 to 18 BTDC. COV is dened as standard deviation of net IMEP upon mean of net IMEP and
expressed in%. As presented in Fig. 9a, there is little difference between gasoline and n-butanol combustion with respect to COV value. Figs. 9b and c shows statistical results of this comparison,
which was based on probability density functions (pdf) of NIMEP
for n-butanol and gasoline combustion tests at spark timing of
10. It is noticed that n-butanol NIMEP distribution (Fig. 9b) is
slightly narrower than for gasoline (Fig. 9c), but the difference is

IMEP kPa

2500

(c)

pdf

0.1

0.05

In-cylinder Pressure kPa

Gasoline 87 100%
CR=10
ST=10 deg BTDC
Normal distribution
Mean=330.7
Sigma=4.099
R=0.98
NIMEP COV=1.24

CR=10
Spark Timing=10
deg BTDC

2000

NIMEP=330kPa
1500
n-butanol 100%
n-butanol 60%

1000

n-butanol 20%
0
300

Gasoline PON 87
310

320

330

340

350

IMEP kPa
Fig. 9. (a) COV of NIMEP for n-butanol and gasoline with PON 87 vs. spark timing,
(b) PDF of NIMEP distribution for n-butanol, (c) PDF of NIMEP for gasoline PON 87.

500
-10

10
Crank Angle deg

20

Fig. 10. In-cylinder pressure traces for n-butanolgasoline blends.

30

1579

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

2350

n-butanol 100%

2300

n-butanol 60%

2250
n-butanol 20%

2200
2150

CR = 10
NIMEP = 330kPa
ST = 10 deg BTDC

10.5

18
10-90% MFB

17.5

50% MFB

17
8.5
16.5

CR = 10
NIMEP = 330kPa
ST = 10 deg BTDC

16

n-butanol 0%
gasoline 100%

15.5

16

16. 5

17

17.5

18

18.5

20

Crank Angle Location deg

Normalized MFB

(b)
0-10 % MFB CA deg

Fig. 11. Location of in-cylinder peak pressure for n-butanolgasoline blends.

0.8

7.5

15.5

2100

CR=10
Spark Timing=10
deg BTDC
NIMEP=330kPa

6.5
100

40
60
80
Butanol Percentage %

11

0-10% MFB
10

CR = 10
NIMEP = 330kPa
ST = 10 deg BTDC

0.6
8

0.4

n-butanol 100%

20

40

60

80

100

Butanol Percentage %

n-butanol 60%

0.2

(c)

n-butanol 20%

18

n-butanol 0%
gasoline 100%

10
Crank Angle deg

20

30

Fig. 12. Normalized MFB proles for n-butanolgasoline blends.

insignicant. This statement is conrmed by the lower NIMEP


COV = 0.91% when compared with gasoline with pump octane rating (PON) 87 (COV = 1.24%). It can be also stated that n-butanol can
be considered as gasoline substitute as far as combustion stability
is concerned.
3.2. n-Butanol percentage in blends
These tests were performed to quantify the two n-butanolgasoline blends of 20% and 60% of n-butanol and two extra fuels: 100%
n-butanol and gasoline with PON 87. The test was done at the same
CR, spark timing, and for stoichiometric mixtures. The analysis was
based on in-cylinder pressure traces, normalized MFB prole and
rate of normalized MFB. Fig. 10 shows the in-cylinder pressure histories for these fuels during combustion.
Fig. 11 shows peak pressure of combustion and its location for
these fuels. As seen the highest peak pressure is for pure n-butanol
and it comes earlier when compared to other fuels. This is because
of faster n-butanol combustion than gasoline. This conclusion is
additionally conrmed by the normalized MFB proles for these
fuels, presented in Fig. 12.
From Fig. 12 quantities as 10% MFB, 50% MFB and 90% MFB can
be determined. As mentioned earlier, these values correspond to
0.1, 05 and 0.9 of normalized MFB, respectively. The combustion
duration presented in Fig. 13a is calculated as a difference between
0.9 MFB and 0.1 MFB and it is expressed in degrees of crankshaft

10-90 % MFB CA deg

Gasoline PON 87

0
-10

9.5

Location of 50% MFB CA deg

(a)
10-90 % MFB CA deg

In-Cylinder Peak Pressure kPa

2400

17

n-butanol 20%
n-butanol 60%

16
CR = 10
NIMEP = 330kPa
ST = 10 deg BTDC

n-butanol 100%

15
7

10

11

50% MFB CA deg


Fig. 13. Combustion parameters: (a) 1090% MFB vs. butanol percentage, (b) 010%
MFB vs. butanol percentage, (c) 1090% MFB vs. 50% MFB.

angle. As plotted in Fig. 13a, combustion duration of pure n-butanol (approximately 16) is shorter only by 10% in comparison to
pure gasoline (17.6). The more important issue is the location of
50% MFB, which shifts to an earlier position with increasing blends
of n-butanol. This is because of the shorter ignition delay for nbutanol as presented in Fig. 13b. The ignition delay was expressed
by 010% MFB duration. During this period of early combustion,
the combustion rate is impacted by the laminar ame speed of
the fuelair mixture. At later times which are in the fully developed bulk burn, the combustion is dominated by turbulent ame
propagation. The difference between 50% MFB for n-butanol and
gasoline is about 2. The 50% MFB represents the center of combustion, and it has been shown that the engine torque strongly depends on location of 50% MFB. Therefore to obtain the maximum

1580

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

0.1

0.08

CR=10
Spark Timing=10
deg BTDC

n-butanol 100%

n-butanol 100% CR=10

n-butanol 60%

n-butanol 100% CR=8

0.8

n-butanol 20%

Spark Timing=10
deg BTDC

0.06

Normalized MFB

Rate of MFB

Gasoline PON 87

NIMEP=330kPa

0.04

0.02

0
-10

0.6

0.4

0.2

10
Crank Angle deg

20

NIMEP=330kPa

0
-10

30

Fig. 14. Rate of MFB for several n-butanol share of butanolgasoline mixture.

10
Crank Angle deg

20

30

Fig. 16. MFB proles for two different compression ratios.

2500

0.1

Spark Timing =10 deg BTDC

n-butanol 100% CR=10

Rate of MFB

In-cylnder pressure kPa

n-butanol 100% CR=8

0.08

2000

1500

Spark Timing
=10 deg BTDC

NIMEP=330kPa

0.06

0.04

NIMEP=330kPa

1000
0.02

n-butanol 100% CR=10


n-butanol 100% CR=8
500
-10

10
Crank Angle deg

20

30

Fig. 15. In-cylinder pressure traces for two different compression ratios.

break torque when gasoline is replaced by n-butanol, the spark advanced should be adjusted.
As presented in Fig. 13c, the optimal 50% MFB goes from 10 to
approximately 8 after TDC with increase of butanol fraction in
the blends.
Although the combustion duration is shorter with increased nbutanol content, the relative change in the rate of combustion is
small. Fig. 14 shows rate of normalized MFB, which can be considered as a prole similar to heat release rate. As seen in the gure
the rate prole is advanced as the n-butanol fraction increases,
but the peak rate of normal MFB remains the same. The increase
in early combustion rate increases over the range of n-butanol fractions invested.
In summary, as shown under this condition, the normalized
MFB and MFB rate does not change signicantly with n-butanol
fraction. An advancement of the phasing of combustion relative
to the spark timing with increased n-butanol fraction does occur.
For optimal combustion phasing the spark timing should be adjusted or controlled as a function of the n-butanol fraction.
3.3. Compression ratio impact
In this section compression ratio impact on combustion of nbutanolgasoline blends is examined. A higher compression ratio

0
-10

10
Crank Angle deg

20

30

Fig. 17. Rate of MFB for different two compression ratios.

provides conditions for faster combustion due to higher temperature at the end of compression stroke and improved efciency.
As a result of the increased temperatures, an increase in compression ratio shortens ignition delay. However, these higher temperatures with increased compression ratio also increase the potential
for autoignition in the end gas and combustion knock. Combustion
pressure traces for two compression ratios of 8:1 and 10:1 are presented in Fig. 15. The test was carried out at the same NIMEP load
of 330 kPa, the same spark timing of 10 BTDC, for a stoichiometric
mixture, and pure n-butanol as a fuel.
The normalized MFB proles from the pressure traces were calculated and are presented in Fig. 16. The change in compression ratio from 8:1 to 10:1 does not signicantly shorten bulk burn
combustion duration but the start of combustion, which corresponds to the 10% MFB phase, is advanced by 3.3 for the higher
compression ratio case. In comparison the location of 50% MFB is
advanced by 4.3 for the 10:1 compression ratio relative to the
8:1 compression ratio.
Fig. 17 depicts rate of MFB determined from MFB proles presented in Fig. 16. Again, the advancement of combustion is evident
during the early portion of combustion; however, the peak rate of
combustion remains similar.
As presented in Fig. 17, both the maximal rate of MFB and the
MFB rate prole for compression ratio 10:1 is almost the same as

1581

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

The test at full load was done only at compression ratio 8:1 as
the engine would be at heavy knock for this condition with compression ratio of 10:1. Fig. 18 shows in-cylinder pressure during
stoichiometric butanol combustion at the two loads of 330 kPa
and 650 kPa NIMEP. For the high load case, high frequency pressure oscillations are superimposed on the low frequency pressure
curve. Furthermore, the pressure rapidly increased at the end of
combustion. Both of these are symptoms typical for combustion
knock.
The combustion duration as determined from the normalized
MFB proles for these two loads is strongly affected. The initial
phase of combustion 010% MFB, the combustion phasing 50%
MFB and the end phase 90% MFB area all advanced for the high
load in comparison to with the partial load condition as depicted
in Fig. 19.
Furthermore, peak rate of MFB presented in Fig. 20 is much
higher during full load operation, and its maximum is located in
the end phase of combustion as result of rapid acceleration of combustion by the combustion knock effect.
3.5. Combustion knock analysis
When considering the MFB rate proles included in this work,
the threshold of MFB rate which does not accompany combustion
knock, can be estimated. After examination of Figs. 5, 14, 17 and 20
it can be noticed that combustion of n-butanol at rate of normalized MFB not higher than 0.06 does not contain combustion knock.
As presented in Fig. 20, high MFB rate (here over the 0.08) accompanies combustion knock, which strongly appears in combustion
initiated by over-advanced spark timing and high load. The correlation between spark timing and combustion knock at engine constant load for n-butanol and gasoline is shown in Fig. 21.

CR=8
Spark Timing=10
deg BTDC

0.8
Normalized MFB

3.4. Load impact

In this gure combustion knock intensity is expressed by the


peak pressure (PP) of the high frequency (oscillating) component
of the in-cylinder pressure. It was calculated from tests conducted
at the two loads, two compression ratios with n-butanol and gaso-

0.6

0.4
n-butanol 100%
NIMEP=330kPa
0.2

n-butanol 100%
NIMEP=650kPa

0
-10

10
Crank Angle deg

20

30

Fig. 19. MFB proles for different loads.

0.1

n-butanol 100% NIMEP=330kPa


n-butanol 100% NIMEP=650kPa

0.08

Rate of MFB

for compression ratio 8:1. Moreover, an increase in the MFB rate at


the end phase of combustion is not observed, indicating that combustion knock is not occurring with this increase in compression
ratio. It should be noted that is a part-load condition. The throttle
of the engine was adjusted such that the manifold pressure was below 50 kPa. The NIMEP at these terms reached 330 kPa, which is
approximately 50% of nominal load of this engine operating on
with n-butanol at wide open throttle (WOT). In the next section,
tests done at full load of the engine with the throttle maximally
open (WOT) are investigated.

0.06

0.04

CR=8
Spark
Timing=10
deg
BTDC

0.02

0
-10

10
Crank Angle deg

20

30

Fig. 20. Rate of MFB against crank angle.

4000
Peak Pressure of the Oscillating
Component kPa

3500
In-cylinder Pressure kPa

250

CR=8
Spark Timing
=10 deg BTDC

3000
2500
n-butanol 100% NIMEP=330kPa
n-butanol 100% NIMEP=650kPa

2000
1500

n-butanol 100% CR=8 NIMEP=650kPa


n-butanol 100% CR=10 NIMEP=330kPa
n-butanol 100% CR=8 NIMEP=330kPa
Gasoline 87 CR=10 NIMEP=330kPa

200

150

100

50

1000
0

500
-10

10
Crank Angle deg

20

Fig. 18. In-cylinder pressure traces for two loads.

30

10

12

14

16

18

Spark Timing deg BTDC


Fig. 21. Peak pressure of a high frequency component (oscillating component) of
combustion pressure vs. spark timing.

1582

S. Szwaja, J.D. Naber / Fuel 89 (2010) 15731582

Maximum Rate of MFB

0.11

CR=8
IMEP=650kPa
ST=10 deg bTDC

0.1
0.09

CR=8
IMEP=330kPa
ST=10 deg bTDC

0.08
0.07

CR=10
IMEP=330kPa
ST=var

0.06
CR=10
IMEP=330kPa
ST=10 deg bTDC
0,20,60,100% Butanol

0.05
0.04
0.03
2

10

12

14

16

18

20

Spark Timing deg BTDC


Fig. 22. Maximum rate of MFB vs. spark timing for 100% n-butanol (blue
rhombuses), for 0%, 20%, 60% blends (green circles) and for 100% n-butanol at two
loads (yellow triangles). (For interpretation of the references to colour in this gure
legend, the reader is referred to the web version of this article.)

line as for reference. It is noticed that the PP curve is strictly inline


with the maximum rate of MFB as for instance presented in Fig. 6
for n-butanol, CR = 10:1 and NIMEP = 330 kPa. Fig. 22 shows maximum rate of MFB vs. spark timing. It can be noticed that both the
PP metrics for knock intensity and maximum rate of MFB are correlated with each other. Moreover a strong dependence between
the knock and the engine load was noticed (points marked as yellow triangles in Fig. 22).1 Additionally, as expected the compression ratio also strongly affects combustion knock for the nbutanol fueled engine working at full load.
4. Conclusions
The following conclusions can be drawn from the analysis presented above:
 At the 330 kPa NIMEP condition examined in this study, the
location of 50% MFB for n-butanol is located approximately 2
earlier when compared with 50% MFB for gasoline at the same
combustion conditions.
 The bulk burn combustion durations for pure n-butanol and nbutanol blends are comparable to those for gasoline. The shift
in the 50% MFB location for n-butanol shifts towards TDC as a
result of the shorter early burn duration. This causes non-optimum combustion phasing at constant spark timing. Thus, optimal spark timing (to obtain MBT) for pure n-butanolair
stoichiometric mixture should be retarded with respect to the
spark timing adjusted for gasoline by an engine electronic control unit.
 Combustion stability at the part-load 330 NIMEP condition as
measured by the COV of NIMEP was slightly lower for n-butanol
in comparison to gasoline.
 In this engine at a compression ratio of 8:1, n-butanol combustion with spark timing of 14 BTDC or further advanced, results
in combustion knock, even at the part-load condition of 330 kPa
NIMEP. Combustion knock was also observed for the test with
spark timing 10 BTDC at 650 kPa NIMEP, the full load condition
for this engine.

1
For interpretation of colour in Fig. 22, the reader is referred to the web version of
this article.

 At the 330 kPa NIMEP partial load condition invested, the optimal spark timing of approximately 10 BTDC for pure n-butanol
provides maximum efciency without combustion knock. Similar results were observed for gasoline, but spark timing should
be slightly (by nearly 1) advanced.
 The behavior of neat n-butanol with respect to combustion
knock is similar to the characteristics to that of PON 87 gasoline.
This includes the sensitivity to spark timing and compression
ratio.
 At the part-load condition examined in this work, the combustion duration and rate of MFB are comparable for n-butanol
blended gasoline at ratios of 20% and 60% so it could be extended
to full blending range from 0% to 100%. From this point of view,
n-butanol can directly substitute for gasoline either as a neat
fuel or blended from a combustion and energy density perspective for fuel for a SI engine because of the similar thermo-physical properties (Table 3).
Acknowledgements
The authors acknowledge the support of Michigan Technological University and specically the Mechanical Engineering and
Engineering Mechanics Department and the Advanced Power Systems Research Center for support and facilities utilized in this
work. This work is conducted under the Universities Wood-toWheels research enterprise as a subgroup of the Sustainable Futures Institute.
References
[1] Naber JD, Bradley EK, Szpytman JE. Target-based rapid prototyping control
system for engine research. SAE Trans J Eng 2006;115(3):22343 [originally
appeared as SAE Technical Paper No. SAE 2006-01-0400; 2006].
[2] Tsao GT. A novel 4A process ready for commercial production of ethanol,
butanol and hydrogen from cellulosics. In: Biofuels symposium 2008, May 19
20, Stewart Center, Purdue University, West Lafayette, Indiana; 2008.
[3] Busche R, Allen B. Technoeconomics of butanol extractive fermentation in a
multi-phase uidized bed bioreactor. Appl Biochem Biotechnol 1989;20/
21:35774.
[4] Laza T, Kecskes R, Bereczky A, Penninger A. Examination of burning processes
of regenerative liquid fuel and alcohol mixtures in diesel engine. Periodica
Polytech Mech Eng 2006;50(1):1129.
[5] Alasfour FN. Butanol a single cylinder engine study: availability analysis.
Appl Therm Eng 1997;17(6):53749.
[6] Yacoub Y, Bara R, Gautam M. The performance and emission characteristics of
C1C5 alcoholgasoline blends with matched oxygen content in a singlecylinder spark ignition engine. Proc Inst Mech Eng A Power Energy
1998;212:36379.
[7] Gautam M, Martin DW. Combustion characteristics of higher-alcohol/gasoline
blends. Proc Inst Mech Eng 2000;214A:497511.
[8] Gautam M, Martin DW. Emission characteristics of higher-alcohol/gasoline
blends. Proc Inst Mech Eng 2000;214A:16582.
[9] Aleiferis P, Pereira JS, 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 2008-01-1591.
[10] Dagaut P, Togbe C. Oxidation kinetics of butanolgasoline surrogate mixtures
in a jet-stirred reactor: experimental and modeling study. Fuel
2008;87:331321.
[11] Cooney C, Wallner T, McConnell S, Gillen JC, Abell C, Miers SA, et al. Effects of
blending gasoline with ethanol and butanol on engine efciency and
emissions. In: Proceedings of the ASME internal combustion engine division
2009 spring technical conference ICES2009-76155, Milwaukee, Wisconsin,
USA; 2006.
[12] Heywood JB. Internal combustion engines fundamentals. McGraw Hill Inc.;
1988.

Anda mungkin juga menyukai