Conference PaperPDF Available

Shock-Tube Measurements of the Ignition Delay Times of Reference Fuels at Intermediate Temperatures and High Pressures

Authors:
Shock-Tube Measurements of the Ignition Delay Times of Gasoline Surrogates at
Intermediate Temperatures and High Pressures
M. Fikri
1,*
, J. Herzler
1,
, R. Starke , G. T. Kalghatgi , P. Roth , C. Schulz
1 2 1 1
1
IVG, Universität Duisburg-Essen, Germany
2
Shell Global Solutions
Abstract
Ignition times were determined in high-pressure shock-tube experiments for various stoichiometric mixtures of two
multi-component model fuels in air for the validation of ignition delay simulations based on chemical kinetic
models. The fuels were n-heptane (18%) / iso-octane (62%) / ethanol (20%) as well as n-heptane (20%) / toluene
(45%) / iso-octane (25%) / di-isobutylene (10%), by liquid volume. These fuels have octane numbers comparable to
a standard European gasoline of 95 RON and 85 MON. The experimental conditions cover temperatures from 690 to
1200 K and pressures at 10, 30, and 50 bar. The obtained ignition time data are scaled and compared to previous
results reported in literature. This study confirms that the autoignition of pure fuels differ from those of mixtures due
to the coupling of the chemical kinetics of the various components.
*
Corresponding author: mustapha.fikri@uni-due.de
Proceedings of the European Combustion Meeting 2007
Present affiliation: Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Verbrennungstechnik,
Pfaffenwaldring 38-40, 70569 Stuttgart, Germany.
Introduction
Among the major practical concerns of internal
combustion engines is to increase fuel efficiency while
reducing pollutant emissions such NO
x
and particulates.
Autoignition of the fuel/air mixture plays a major role
when optimizing the performance of modern automotive
engines.
In spark-ignition (SI) engines knocking limits the
allowable compression and thus efficiency. Knock,
however, arises from autoignition of the unburned part
of the fuel/air mixture while the spark-ignited flame
propagates through the combustion chamber. Thus,
fuels with long ignition delay times are needed for
increasing engine efficiency. There is also a great deal
of interest in Homogeneous Charge Compression
Ignition (HCCI) engines since they offer the prospect of
the high efficiency of the compression-ignition (CI)
engine but with low levels of nitrogen oxides and soot
which are very difficult to control in CI engines. One
shortcoming is the relatively high emission of unburned
hydrocarbons [1,2].
Practical fuels exhibit tremendous complexities due
to the quasi-continuous spectrum of hydrocarbon
constituents, as well as to the variability of their
compositions. They are affected by the oil source and its
refining process. In these cases it is beneficial to employ
a substitute fuel with a finite number of compounds and
a standard composition to nearly match practical fuels.
Theoretical studies of autoignition with complex fuel
mixtures have been not well established. The main
reason is that chemical reaction mechanisms applicable
to auto-ignition in IC engines have only been developed
for very few pure compounds – mostly paraffins of low
carbon numbers. It is known that the ignition behavior
of mixtures can differ significantly from that of the
single components due to the coupling of the chemical
kinetics of the various components [3]. A review
concerning the development of detailed chemical
models for hydrocarbon fuels is published in [4].
For characterizing fuels in respect to their resistance
in terms of ignition behavior, empirical measures such
as octane number are evaluated. The octane number of a
particular fuel blend is determined under specified
operating conditions. Two common standards are used
for the operating conditions, giving either the research
octane number (RON), or the motor octane number
(MON). However, the RON and MON scale are both
based on primary reference fuels (PRF) and the RON
and MON of any fuel containing non-paraffinic
components describes its auto-ignition behaviour only at
the respective test conditions. RON is higher than MON
for practical fuels but RON is the same as MON for
PRF. The resulting numbers are not always helpful for
characterizing fuels for modern IC engines.
Leppard et al. [5] suggested that under the MON test
condition the NTC range dominates the chemistry of
PRF slowing the reactions and making them more
resistant to knock compared to fuels containing olefins
and aromatics. While, in the RON test condition the
NTC chemistry becomes less important so that the PRF
fuels lose the resistance to autoignition compared to
non-PRF-fuels.
In a previous study, we investigated the ignition
delay time of mixtures containing 35% n-heptane and
65% toluene by liquid volume (fuel A) behind reflected
shock waves in the temperature range of 620
Τ
≤ 1180 Κ [6]. In a separate paper, the results of this
study have been used to validate a detailed chemical
THIRD EUROPEAN COMBUSTION MEETING ECM 2007
Back to Table of Contents
kinetics model proposed by Andrae et al. [7]. The main
objective was to test how well the shock-tube ignition
data as well as the data from two HCCI engines [8,9]
could be evaluated using this kinetic model. The
experimental results showed good agreement with the
predictions using detailed chemical kinetics. The
authors found that the autoignition of the multi-
component fuel mixture confirms that cross-acceleration
effects of hydroperoxy radicals produced during the n-
heptane oxidation increased the oxidation rate of
toluene compared to that of pure toluene.
Experimental facilities
Test section
MFC
MFC
High pressure section
Filter
PMT
Scope
He
Ar
Piezo pressure transducers
Aluminium diaphragm
Test section
MFC
MFC
High pressure section
Filter
PMT
Scope
He
Ar
Piezo pressure transducers
Aluminium diaphragm
Fig. 1: Experimental setup
The experiments were carried out in a heated high-
pressure shock tube with an internal diameter of 90 mm.
It is divided by an aluminium diaphragm into a driver
section of 6.1 m and a driven section of 6.4 m in length.
The driven section was pumped down to pressures
below 10
–2
mbar. Gas mixtures were prepared by
injection of a liquid mixture of the fuel components and
subsequent complete evaporation in a stainless-steel
mixing vessel. The total amount of fuel and air was
controlled manometrically
in order to ensure the desired
equivalence ratio. The shock tube and the mixing vessel
were heated to 130°C. The shock speed was measured
over two intervals using three piezo-electric pressure
gauges. The data were recorded with a time resolution
of 0.1 µs. The temperature and pressure behind the
reflected shock wave were computed from the measured
incident shock speed and the speed attenuation using a
one-dimensional shock-tube model (shock-tube code of
the CHEMKIN package [10]). The estimated
uncertainty in reflected shock temperature is less than
±25 K in the temperature and time range of our
measurements. The experiments were carried out in
synthetic air containting 79.5% N
2
and 20.5% O
2
. n-
heptane, toluene, isooctane, and diisobutylene (DIB)
which are liquid at normal conditions were mixed
according to the composition given in table 1
(percentages by volume) and injected as liquid into the
mixing vessel.
The ignition was observed by measuring pressure
profiles with a piezo-electric gauge (PCB HM 112 A03)
located 15 mm upstream of the end flange. Also, the
CH* emission at 431.5 nm was selected by a narrow
band pass filter and measured with a photomultiplier.
All ignition delay times shown in this work were
determined by extrapolating the steepest increase of the
emission signal to its zero level on the time axis.
The driver gas was mixed in-situ by using two high-
pressure mass-flow controllers (Bronkhorst Hi-Tec flow
meter F-136AI-FZD-55-V and F-123MI-FZD-55-V),
see Figure 1. Helium was used as the main component
and Argon was added to match the acoustic impedance
of the test gas. The required driver gas composition was
calculated by a spreadsheet analysis prior to the
experiments using equations by Oertel [11] and Palmer
and Knox [12]. Concentrations of 5 to 20% Ar in He
were required to generate tailored shock waves.
Scaling
The actual pressure at each experiment is accurately
measured but slightly scatters around the target
pressure. Therefore, for comparison of the ignition
delay times in the diagrams presented here, a pressure-
scaling factor is used for each fuel in order to normalize
all data to the actual target pressures. This is done by
performing a series of measurements for a limited
temperature range and fitting a power-law dependence
to the pressure according to the following equation:
τ =
f(T) p
x
x is assumed to be constant for each fuel. This
procedure oversimplifies the ignition behaviour. In fact
the temperature and pressure variations in ignition are
more complicated. Nevertheless, we used this method
similarly as in [6]. For the comparison with simulation
results, the full un-corrected data that is given in table 2
should be used.
Results and discussion
The experiments were conducted in the temperature
range of 690
Τ
≤ 1200 Κ and at pressures of about 10,
30, 50 bar and
φ
= 1 . These conditions are relevant for
internal combustion engines.
An overview of the fuel composition and the RON and
MON of the used fuels is shown in Table 1. Fuels B and
C can be taken as gasoline surrogates.
Tab. 1: Octane number characteristics of the used
fuels.
fuel RON MON reference
A 84 73.2 [6]
B 95.1 89.5 this work
C 94.6 85 this work
n-heptane toluene isooctane ethanol DIB
A 35 % 65 %
B 18 % 62 % 20 %
C 20 % 45 % 25 % 10 %
2
THIRD EUROPEAN COMBUSTION MEETING ECM 2007
Back to Table of Contents
A typical measured pressure and CH* emission profile
is shown in Fig. 2. The pressure signal of the
experiment at T = 1062 K and 11.3 bar with an
equivalence ratio
φ
= 1.0 (upper part of Figure 2) shows
a two-step increase due to the incident and reflected
shock wave (time zero) followed by a constant pressure
for about 2 ms and then an increase to ~50 bar. The
CH* emission intensity (lower part of Fig. 2) remains at
zero for 1.9 ms, followed by a steep rise that indicates
ignition. The onset of the CH* emission was measured
through a side-wall window 15 mm from the end flange.
0 500 1000 1500 2000 2500
0
1
2
3
4
ignition delay time
CH*-Emission / a. u.
Time / µs
0.0000 0.0005 0.0010 0.0 015 0.0020 0.0025
0
10
20
30
40
50
60
Pressure / bar
Fig. 2: Typical pressure (upper part) and CH*-
emission (lower part) profiles indicating ignition
delay in a stoichiometric n-heptane/toluene (35% /
65% by liquid volume) / air mixture. Reaction
conditions: T
5
= 1062 K, p
5
= 11.3 bar.
Tab. 2: Measured ignition delay times for
stoichiometric mixtures of fuel B and fuel C in air.
T / K 1000 K /
T
τ
/ µs
p / bar
Fuel B 969 1.032 2870 10.0
1011 0.989 1170 8.8
1023 0.977 1130 10.8
1182 0.846 284 11.3
1222 0.818 157 12.2
1232 0.812 181 12.8
720 1.389 7800 31.7
762 1.312 6400 30.0
870 1.149 4420 30.4
976 1.024 1360 30.7
1057 0.946 467 31.3
1148 0.871 172 31.9
694 1.441 7690 47.7
751 1.331 3440 48.7
841 1.189 2200 47.0
953 1.049 1150 48.8
1037 0.964 431 49.8
1134 0.882 115 51.9
Fuel C 926 1.079 4600 10.7
979 1.021 2390 12.3
990 1.009 2300 11.0
990 1.009 2730 11.0
1024 0.976 2160 10.2
1076 0.929 704 10.6
1114 0.897 427 10.3
1148 0.871 275 12.3
717 1.395 8320 31.9
760 1.316 7490 31.0
832 1.202 5340 33.1
871 1.148 3870 32.0
932 1.072 2070 32.8
973 1.028 1250 31.6
999 1.000 896 29.9
1084 0.923 290 32.3
1123 0.890 191 31.4
691 1.447 10700 48.1
741 1.349 3540 48.2
786 1.272 2640 47.4
839 1.192 2680 48.2
901 1.110 1890 50.1
960 1.042 958 51.2
995 1.005 651 49.7
1030 0.971 381 47.7
1106 0.904 149 50.4
Experiments with fuel B were performed at
temperatures 690 – 1230 K, and at pressures near 10,
30, and 50 bar. The measured ignition-delay times are
shown in Tab. 2 and depicted in Fig. 3 as a function of
temperature. At lower temperatures no ignition was
observed during the maximum test time of 15 ms.
0,7 0,8 0,9 1,0 1,1 1,2 1,3 1,4 1,5
100
1000
10000
Ignition delay time
τ
/ µs
1000 K / T
10 bar
30 bar
50 bar
Fig. 3: Ignition delay times for a stoichiometric
mixture of fuel B in air at pressures of about 10
(diamonds), 30 (squares) and 50 bar (triangles). The
lines connect the data points to illustrate the trends.
The variation of the ignition time with pressure for all
sets of data was determined for fuel B and was assumed
to vary as p
–0.76
. The scaling was performed using a
multiple linear regression analysis using ln(τ) as the
dependent variable and (1/T) and ln(p) as the
independent variables. The obtained normalized ignition
delay time are shown in Fig. 4. In this case, the ignition
delay times increase with decreasing temperature and
3
THIRD EUROPEAN COMBUSTION MEETING ECM 2007
Back to Table of Contents
after a certain temperature the ignition delay time
decrease slightly. The inserted lines indicate the trends.
0,8 0,9 1,0 1,1 1,2 1,3 1,4 1,5
100
1000
10000
10 bar
30 bar
50 bar
τ
30
=
τ
(30 bar / p)
-0.76
/ µs
1000 K / T
Fig. 4: Ignition delay times for stoichiometric
mixtures of fuel B with air at pressures of about 10
(diamonds), 30 (squares), and 50 bar (triangles)
normalized to 30 bar using (30 bar / p)
–0.76
.
The results of fuel B and fuel A ignition delay times
experiments, scaled to 30 bar with their respective
factors at stoichiometric conditions are shown in Fig. 5.
In our previous work the scaling factor -1.06 was used
for the fuel A [6]. It can be seen unequivocally from the
plot that the two fuels are effectively indistinguishable
in terms of ignition delay time except a slight difference
at high temperature.
0,8 0,9 1,0 1,1 1,2 1,3 1,4 1,5
100
1000
10000
τ
30
=
τ
(30 bar / p)
-n
/ µs
1000 K / T
Fig. 5: Comparison of the ignition-delay times for
stoichiometric mixtures of fuel A (open squares) [6]
and fuel B (diamonds) in air normalized to 30 bar
using the respective scaling factors.
In a similar way the ignition delay time of fuel C were
measured. The variation of the ignition-delay time with
pressure was determined by scaling the ignition times
values to 30 bar using the calculated exponent x =
–0.65. The results of the results for fuel C are shown in
Fig. 6. A slight NTC behavior is seen especially at high
pressure.
0,9 1,0 1,1 1,2 1,3 1,4 1,5
100
1000
10000
10 bar
30 bar
50 bar
τ
30
=
τ
(30 bar / p)
-0.65
/ µs
1000 K / T
Fig. 6: Ignition delay times for stoichiometric
mixtures of fuel C in air at pressures of about 10
(diamonds), 30 (squares), and 50 bar (triangles)
normalized to 30 bar using (30 bar / p)
–0.65
.
A comparison of the results for the different fuels
investigated in this work with data of the literature [13]
normalized with their respective scaling factors is
shown in Fig. 7. All data are under stoichiometric
conditions except the data of fuel A (black + sign) with
φ
= 0.3. The corresponding fuels from the work of
Gauthier et al. are gasoline (RD387) or ternary
surrogates containing isooctane, toluene and n-heptane
in the following proportions: surrogate A: 69/14/17%,
surrogate B: 63/20/17% by liquid volume. The power-
law pressure dependence calculated from the linear
regression was p
–0.83
, for surrogate A, p
–0.96
for surrogate
B, and p
–1.06
for gasoline – the pressure scaling factors
are those given in Ref. [13]. The pressure scaling of the
n-heptane mixture (circles) was determined to be p
–1.64
by Gauthier et al. [13]. Indeed the pressure scaling
exponent appears, in general, to be much larger for
paraffinic fuels compared to non-paraffinic fuels [17].
This experimental observation is consistent with the
observation that the difference between paraffinic and
non-paraffinic fuels in HCCI engines [18] and knocking
SI engines [19] could be explained only by assuming
such a difference in the pressure dependence of
autoignition delay times between such fuels. For
stoichiometric mixtures the difference in autoignition
delay between different gasoline fuel surrogate blends
and gasoline considered in Fig.7 is small for the high
temperature range – above about 1000 K. Fuel A (open
diamonds) seems to have a slightly longer ignition delay
at the same temperature compared to the other fuels.
This is in fact consistent with its observed behavior in
HCCI tests at high temperature where it was more
resistant to auto-ignition than expected from its octane
index [9].
4
THIRD EUROPEAN COMBUSTION MEETING ECM 2007
Back to Table of Contents
0,8 1,0 1,2 1,4
100
1000
10000
Normalized ignition delay
τ
30
/ µs
1000 K / T
fuel A, f = 0.3, n=0.88
fuel A, 84/73, n=1.06
fuel B, 95.1/89.5, n=0.76
fuel C, 94.1/84.4, n=0.64
n-heptane, Gauthier et al., n=1.64
Gasoline, Gauthier et al., n=1.06
surrogate A, Gauthier et al., n=0.83
surrogate B, Gauthier et al., n=0.96
Fig. 7: Comparison of the ignition-delay times for
the different fuels from this work with ignition delay
times from literature [13-16]. All data are
normalized to 30 bar using the respective scaling
factors.
Conclusions
Ignition delay times of stoichiometric mixtures of
two surrogate fuels consisting of n-heptane (18%) / iso-
octane (62%) / ethanol (20%), as well as n-heptane
(20%) / toluene (45%) / iso-octane (25%) / di-
isobutylene (10%) in air have been measured in a high-
pressure shock tube behind the reflected wave at
pressures of 10, 30, 50 bar and at temperatures of 690
Τ
≤ 1230. It is possible to scale the data at different
pressures for a given fuel by multiplying the
autoignition delay by p
x
where p is the pressure and x is
found from analysis of the data.
The data for the two gasoline surrogates are in
good
agreement with the data in the literature for other
gasoline surrogates, at high temperatures.
The experimental findings from this survey will be
used in the future work to validate an improved
kinetic model for the auto-ignition.
Acknowledgements
The authors thank N. Schlösser for her help in
conducting the experiments and Shell Global Solutions
for financial support.
[1] D. L. Flowers, S. M. Aceves, J. Martinez-Frias,
R. W. Dibble., Proc. Combust. Inst. 29 (2002)
687-694.
[2] M. Christensen, B. Johansson, A. Hultqvist,
SAE Paper 2001-01-1893 (2001).
[3] J. Andrae, D. Johansson, P. Björnbom, P.
Risberg, G.T. Kalghatgi, Combust. Flame 140
(2005) 267-268.
[4] J.M. Simmie, Prog. Energy Combust. Sci. 29
(2003) 599-634.
[5] W.R. Leppard, SAE Paper 902137 (1992).
[6] J. Herzler, M. Fikri, K. Hitzbleck, R. Starke, C.
Schulz, P. Roth, G.T. Kalghtagi, Combust.
Flame (2007),
doi:10.1016/j.combustflame.2006.12.015.
[7] J.C.G. Andrae, P. Björnbom , R.F. Cracknell ,
G.T. Kalghatgi Andrae, J.C.G.,. Combust.
Flame (2007),
doi:10.1016/j.combustflame.2006.12.014.
[8] G.T. Kalghatgi, P. Risberg, H.-E. Angström,
SAE Paper 2003-01-1816, (2003).
[9] G.T. Kalghatgi, R.A. Head, SAE Paper 2004-
01-1969 (2004).
[10] R. J. Kee, F.M. Rupley, J. A. Miller, M. E.
Coltrin, J. F. Grcar, E. Meeks, H. K. Moffat, A.
E. Lutz, G. Dixon-Lewis, M. D. Smooke, J.
Warnatz, G. H. Evans, R. S. Larson, R. E.
Mitchell, L. R. Petzold, W. C. Reynolds, M.
Carcotsios, W. E. Stewart, P. Glarborg, C.
Wang and O. Adigun, CHEMKIN Collection,
Release 3.6, Reaction Design, Inc., San Diego,
CA, 2000.
[11] H. Oertel, Stoßrohre, Springer Verlag,
Wien/New York, 1966.
[12] H.B. Palmer, B.E. Knox, ARS Journal 31
(1961) 826-828.
[13] B.M. Gauthier, D.F. Davidson, R.K. Hanson,
Combust. Flame 139 (2004) 300-311.
[14] H.K. Ciezki, G. Adomeit, Combust. Flame 94
(1993) 421-433.
[15] K Fieweger, R. Blumenthal, G. Adomeit, Proc.
Combust. Inst. 25 (1994) 1579-1585.
[16] K. Fieweger, R. Blumenthal, G. Adomeit,
Combust. Flame 109 (1997) 599-619.
[17] G.T Kalghatgi, SAE Paper 2005-01-0239
(2005)
[18] D. Bradley, C. Morley, H. Walmsley, SAE
2004-01-1970 (2004)
[19] G.T. Kalghatgi, K. Nakata, K. Mogi, SAE
2005-01-0244 (2005).
5
THIRD EUROPEAN COMBUSTION MEETING ECM 2007
Back to Table of Contents
... In this paper the model developed for TRFs in [7] is extended to include the olefin diisobutylene and the oxygenate ethanol. Model predictions are compared with shock tube ignition delay times for neat ethanol and diisobutylene as well as for gasoline surrogate mixtures measured in the shock tube at University Duisburg-Essen [16]. A critical test of the model is then conducted for qualitative prediction of synergistic and antagonistic non-linear MON blending. ...
... Ignition delay times have been determined in high-pressure shock-tube experiments for various stoichiometric mixtures of two multi-component model fuels [16]. The fuels were n-heptane (18% by liquid volume)/iso-octane (62%)/ethanol (20%) with RON 95 and MON 89 and n-heptane (20%)/toluene (45%)/iso-octane (25%)/diisobutylene (10%) with RON 94 and MON 85. ...
... In an engine, a useful approximation is that autoignition occurs when the integral of the reciprocal of s (T, p) with respect to time t attain a value of unity [25]. With non-PRF autoignition delay times indicated by s (T, p) and corresponding PRF delay times indicated by p (T, p) this implies for complete autoignition [26] [16]. Conditions behind reflected shock (p 5 , T 5 ) simulated as a closed homogeneous constant volume adiabatic reactor. ...
Article
A detailed chemical kinetic model to describe the autoignition of gasoline surrogate fuels is presented consisting of the fuels iso-octane, n-heptane, toluene, diisobutylene and ethanol. Model predictions have been compared with shock tube ignition delay time data for surrogates of gasoline over practical ranges of temperature and pressure, and the model has been found to be sensitive to both changes in temperature and pressure. Moreover, the model can qualitatively predict the observed synergistic and antagonistic non-linear blending behaviour in motor octane number (MON) for different combinations of primary reference fuels (PRFs) and non-PRFs by correlating calculated autoignition delay times from peak pressures and temperatures in the MON test to experimental MON values. The reasons for the blending behaviour are interpreted in terms autoignition chemistry.
Article
The objective of this work is to correlate the ignition delay times of reference diesel, convectional diesel, ethanol and biodiesel from soybean oil measured in a shock tube with a 5% additive increase in the cetane number. The results were correlated with the cetane number of the respective fuels and compared with the ignition delay times available in the studies by others authors. The shock tube is a metal tube in which gas at low pressure and at high pressure is separated by a diaphragm. When the diaphragm breaks in predetermined conditions (high pressure in this case) produces shock waves that move from the high-pressure chamber (known driver section) to the low-pressure chamber (known driven section). The tests were performed under the following initial conditions: reflected shock wave temperature from 903 to 1260 K, equivalence ratio of 1 and reflected shock wave pressures of 24 bar. For determining and recording the ignition delay time, pressure sensors with high acquisition rate and luminosity sensors were used. With information from pressure sensors and the luminosity sensor, ignition delay times of reference diesel, convectional diesel, additive ethanol and biodiesel from soybean oil were recorded. We concluded that the ignition delay time of additive ethanol was twice as large as the ignition delay time of reference diesel. The ignition delay time of biodiesel from soybean oil was approximately three times greater than the ignition delay time of reference diesel. The ignition delay time of the reference diesel was smaller than the convectional diesel. The contribution of this work is that it shows why additive ethanol and pure biodiesel should not be used as substitutes for diesel compression ignition engines without any major changes in the engines. The ignition delay times of these fuels are at least two to three times longer than the ignition delay time of reference diesel.
Article
Full-text available
The ignition delay times of mixtures containing 35% n-heptane and 65% toluene by liquid volume at room temperature (i.e., 28% n-heptane/72% toluene by mole fraction) were determined in a high-pressure shock tube in the temperature range 620⩽T⩽1180 K at pressures of about 10, 30, and 50 bar and equivalence ratios, ϕ, of 0.3 and 1.0. The equationτ/μs=9.8×10−3exp(15,680 K/T)(p/bar)−0.883 represents the data for ϕ=0.3 in the temperature range between 980 and 1200 K. At lower temperatures no ignition was found at 10 bar within the maximum test time of 15 ms, whereas for 50 bar, a reduced activation energy was observed. A pressure coefficient of −1.06 was found for the data with ϕ=1.0. No common equation for the data at ϕ=1.0 could be found analogous to that for ϕ=0.3 because the ignition delay times show no Arrhenius-like behavior. A comparison with ignition delay times of n-heptane/air and toluene/air for ϕ=1.0 and 30 bar shows that the values of the mixture of the two components are between the values of the single substances. Furthermore, the results confirm the negative temperature coefficient behavior found for the mixtures at 30 and 50 bar, similar to n-heptane/air. A comparison for the other pressure and equivalence ratio values of this study was not possible because of the lack of data for pure toluene. These experimental data have been used in the development of a chemical kinetics model for toluene/n-heptane mixtures as described in a companion paper.
Article
Full-text available
A detailed chemical kinetic model for the autoignition of toluene reference fuels (TRF) is presented. The toluene submechanism added to the Lawrence Livermore Primary Reference Fuel (PRF) mechanism was developed using recent shock tube autoignition delay time data under conditions relevant to HCCI combustion. For two-component fuels the model was validated against recent high-pressure shock tube autoignition delay time data for a mixture consisting of 35% n-heptane and 65% toluene by liquid volume. Important features of the autoignition of the mixture proved to be cross-acceleration effects, where hydroperoxy radicals produced during n-heptane oxidation dramatically increased the oxidation rate of toluene compared to the case when toluene alone was oxidized. Rate constants for the reaction of benzyl and hydroperoxyl radicals previously used in the modeling of the oxidation of toluene alone were untenably high for modeling of the mixture. To model both systems it was found necessary to use a lower rate and introduce an additional branching route in the reaction between benzyl radicals and O2. Good agreement between experiments and predictions was found when the model was validated against shock tube autoignition delay data for gasoline surrogate fuels consisting of mixtures of 63–69% isooctane, 14–20% toluene, and 17% n-heptane by liquid volume. Cross reactions such as hydrogen abstractions between toluene and alkyl and alkylperoxy radicals and between the PRF were introduced for completion of chemical description. They were only of small importance for modeling autoignition delays from shock tube experiments, even at low temperatures. A single-zone engine model was used to evaluate how well the validated mechanism could capture autoignition behavior of toluene reference fuels in a homogeneous charge compression ignition (HCCI) engine. The model could qualitatively predict the experiments, except in the case with boosted intake pressure, where the initial temperature had to be increased significantly in order to predict the point of autoignition.
Article
The self-ignition of several spark-ignition (SI) engine fuels (iso-octane, methanol, methyl tert-butyl ether and three different mixtures of iso-octane and n-heptane), mixed with air, was investigated experimentally under relevant engine conditions by the shock tube technique. Typical modes of the self-ignition process were registered cinematographically. For temperatures relevant to piston engine combustion, the self-ignition process always starts as an inhomogeneous, deflagrative mild ignition. This instant is defined by the ignition delay time, τdefl. The deflagration process in most cases is followed by a secondary explosion (DDT). This transition defines a second ignition delay time, τDDT, which is a suitable approximation for the chemical ignition delay time, if the change of the thermodynamic conditions of the unburned test gas due to deflagration is taken into account. For iso-octane at p = 40 bar, a NTC (negative temperature coefficient), behaviour connected with a two step (cool flame) self-ignition at low temperatures was observed. This process was very pronounced for rich and less pronounced for stoichiometric mixtures. The results of the τDDT delays of the stoichiometric mixtures were shortened by the primary deflagration process in the temperature range between 800 and 1000 K. Various mixtures of iso-octane and n-heptane were investigated. The results show a strong influence of the n-heptane fraction in the mixture, both on the ignition delay time and on the mode of self-ignition. The self-ignition of methanol and MTBE (methyl tert-butyl ether) is characterized by a very pronounced initial deflagration. For temperatures below 900 K (methanol: 800 K), no secondary explosion occurs. Taking into account the pressure increase due to deflagration, the measured delays τDDT of the secondary explosion are shortened by up to one order of magnitude.
Article
Autoignition characteristics of n-heptane/air, gasoline/air, and ternary surrogate/air mixtures were studied behind reflected shock waves in a high-pressure, low-temperature regime similar to that found in homogeneous charge compression ignition (HCCI) engine cycles. The range of experiments covered combustion of fuel in air for lean, stoichiometric, and rich mixtures ( Φ = 0.5 , 1.0, 2.0), two pressure ranges (15–25 and 45–60 atm), temperatures from 850 to 1280 K, and exhaust gas recirculation (EGR) loadings of (0, 20, and 30%). The ignition delay time measurements in n-heptane are in good agreement with the shock tube study of Fieweger et al. (Proc. Combust. Inst. 25 (1994) 1579–1585) and support the observation of a pronounced, low-temperature, NTC region. Strong agreement was seen between ignition delay time measurements for RD387 gasoline and surrogate (63% iso-octane/20% toluene/17% n-heptane by liquid volume) over the full range of experimental conditions studied. Ignition delay time measurements under fuel-lean ( Φ = 0.5 ) mixture conditions were longer than with Φ = 1.0 mixtures at both the low- (15–25 atm) and high- (45–60 atm) pressure conditions. Ignition delay times in fuel-rich ( Φ = 2.0 ) mixtures for both gasoline and surrogate were indistinguishable in the low-pressure (15–25 atm) range, but were clearly shorter at high-pressures (45–60 atm). EGR loading affected the ignition delay times similarly for both gasoline and surrogate, with clear trends indicating an increase in ignition delay time with increased EGR loading. This data set should provide benchmark targets for detailed mechanism validation and refinement under HCCI conditions.
  • D L Flowers
  • S M Aceves
  • J Martinez-Frias
  • R W Dibble
D. L. Flowers, S. M. Aceves, J. Martinez-Frias, R. W. Dibble., Proc. Combust. Inst. 29 (2002) 687-694. [2] M. Christensen, B. Johansson, A. Hultqvist, SAE Paper 2001-01-1893 (2001).
  • J Andrae
  • D Johansson
  • P Björnbom
  • P Risberg
  • G T Kalghatgi
J. Andrae, D. Johansson, P. Björnbom, P. Risberg, G.T. Kalghatgi, Combust. Flame 140 (2005) 267-268. [4] J.M. Simmie, Prog. Energy Combust. Sci. 29 (2003) 599-634. [5] W.R. Leppard, SAE Paper 902137 (1992).
  • J Herzler
  • M Fikri
  • K Hitzbleck
  • R Starke
  • C Schulz
  • P Roth
  • G T Kalghtagi
J. Herzler, M. Fikri, K. Hitzbleck, R. Starke, C. Schulz, P. Roth, G.T. Kalghtagi, Combust. Flame (2007), doi:10.1016/j.combustflame.2006.12.015. [7]
  • J C G Andrae
  • P Björnbom
  • R F Cracknell
  • G T Kalghatgi Andrae
  • G T Kalghatgi
  • P Risberg
  • H.-E Angström
  • Paper
J.C.G. Andrae, P. Björnbom, R.F. Cracknell, G.T. Kalghatgi Andrae, J.C.G.,. Combust. Flame (2007), doi:10.1016/j.combustflame.2006.12.014. [8] G.T. Kalghatgi, P. Risberg, H.-E. Angström, SAE Paper 2003-01-1816, (2003). [9] G.T. Kalghatgi, R.A. Head, SAE Paper 2004- 01-1969 (2004).
  • D L Flowers
  • S M Aceves
  • J Martinez-Frias
  • R W Dibble
D. L. Flowers, S. M. Aceves, J. Martinez-Frias, R. W. Dibble., Proc. Combust. Inst. 29 (2002) 687-694.
  • J Andrae
  • D Johansson
  • P Björnbom
  • P Risberg
  • G T Kalghatgi
J. Andrae, D. Johansson, P. Björnbom, P. Risberg, G.T. Kalghatgi, Combust. Flame 140 (2005) 267-268.