Gasoline Surrogate Blends

Dec 15, 2016 - This paper presents an experimental and numerical investigation of the autoignition of ethanol blended with several primary reference f...
0 downloads 14 Views 7MB Size
Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

Modelling end-gas autoignition of ethanol/ gasoline surrogate blends in the CFR engine Tien Mun Foong, Michael J. Brear, Kai J Morganti, Gabriel da Silva, Yi Yang, and Frederick L. Dryer Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.6b02380 • Publication Date (Web): 15 Dec 2016 Downloaded from http://pubs.acs.org on December 21, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Modelling end-gas autoignition of ethanol/gasoline surrogate blends in the CFR engine

Tien Mun Foong a , Michael J. Brear

a,∗

, Kai J. Morganti b ,

Gabriel da Silva a , Yi Yang a , Frederick L. Dryer c a School b Saudi

of Engineering, University of Melbourne, 3010, Australia

Aramco Research and Development Centre, Dhahran 31311, Kingdom of Saudi Arabia

c Department

of Mechanical and Aerospace Engineering, Princeton University, NJ 08544, USA

Abstract This paper presents an experimental and numerical investigation of the autoignition of ethanol blended with several Primary Reference Fuels (PRFs) and Toluene Reference Fuels (TRFs) in a CFR engine. Autoigniting, in-cylinder pressure traces are acquired under standard Research Octane Number (RON) conditions. Equivalent, non-autoignitiing traces are obtained by adding a small amount of the tetraethyl lead (TEL) to each fuel in order to suppress knock, and these are used to calibrate a two-zone engine model of autoignition. The simulated autoignition timing of the PRFs without ethanol, TRFs without ethanol, ethanol/PRF and ethanol/TRF blends are then compared to those measured in the engine. These results suggest that the incorporation of residual NO

Preprint submitted to Elsevier Science

ACS Paragon Plus Environment

9 December 2016

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

significantly improves the agreement between simulations and experiment for all cases with no or low ethanol content. Ethanol also appears to suppress the low temperature activity of n-heptane. Both of these results are consistent with previous, more fundamental studies. However, close agreement between the simulated and measured autoignition timing across all fuel blends is not observed. Thus, the modelling does not comprehensively explain the significant synergism and antagonism observed in a recent study of these fuels’ octane numbers (Foong et al., Fuel, 115 (2014) 727-739), demonstrating that further research is required. These results also contribute to a growing body of evidence suggesting the importance of NO chemistry, which should be included in kinetic simulations that attempt to model autoignition and knock in real engines. Key words: Research Octane Number (RON), ethanol, gasoline, Primary Reference Fuels (PRFs), Toluene Reference Fuels (TRFs)

1

Introduction

Alcohol fuels have been used in internal combustion engines for many decades. Indeed, ethanol powered some of the first internal combustion engines designed by Otto in the 19th century [1], and both ethanol and methanol have often been used as fuels ever since. Whilst other alcohols, and particularly butanol, are of more contemporary research interest [2–4], ethanol remains the dominant alcohol fuel used today. A tremendous growth in ethanol use has been observed in recent years. World ethanol production nearly doubled from 2005 to 2010, the latter year in which ∗ Corresponding author. Email address: [email protected] (Michael J. Brear ).

2

ACS Paragon Plus Environment

Page 2 of 38

Page 3 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

roughly 100 billion litres was produced [5], and is projected to further increase to 180 billion litres in 2021 [5]. This is equivalent to roughly 14% of world gasoline consumption in 2010 [6]. The reasons for this growth are several, and include environmental, economic and social considerations. Determination of the best use of ethanol/gasoline blends is therefore a topic of growing practical importance. It is already well known that, relative to gasoline, ethanol can raise spark ignition engine efficiency while reducing specific emissions [7–16]. This is due to both its lower reactivity and a higher enthalpy of vaporisation relative to gasoline [11]. Still, the autoignition characteristics of ethanol when blended with gasoline in spark-ignition engines are not thoroughly understood. Also, whilst the fuel chemistry is of course one important effect, it is not the sole consideration. Others include charge cooling, engine heat transfer, charge homogeneity and residual gases. Importantly, there are key uncertainties in each of these effects and in their interactions. One of these uncertainties is the effect of nitric oxide (NO), which is present in the engine’s residual gases. Autoignition of hydrocarbons has been found to be sensitive to even very low concentrations of NO, although the effect of NO is complex [17,18]. Amano and Dryer [19] reported that NO had a significant promoting effect on the oxidation of methane at temperatures below 1200 K. Likewise, Faravelli et al. [17] showed that the addition of NO to C1 through C4 molecules enhanced their oxidation at low concentrations of NO, although the autoignition of other hydrocarbons may also be inhibited by the presence of NO. A similar observation was reported by Machrafi et al. [18], in which the presence of NO at 45 ppm would advance the ignition delay for an HCCI engine fuelled with PRF40, while autoignition was delayed at 3

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

higher concentrations of NO. More recently, it was reported that under HCCI conditions the autoignition of isooctane would monotonically be advanced with increasing NO content (up to 500 ppm) [20]. Similarly, our own recent works [21,22] have found that residual NO plays an important role in the autoignition of C3 and C4 paraffins, olefins and gasoline surrogates in spark ignition engines. This study therefore complements these other studies by investigating numerically and experimentally the autoignition of ethanol blended with different gasoline surrogates in a CFR engine. Autoigniting, in-cylinder pressure traces are acquired under standard Research Octane Number (RON) conditions. Equivalent, non-autoignitiing traces are obtained by adding a small amount of the dilute tetraethyl lead (TEL) to each fuel, and used to calibrate a two-zone engine model of autoignition. Varying levels of agreement between experiment and simulation are observed, and the source(s) of agreement and disparity are discussed.

2

2.1

Experimental methods

CFR engine and the test method

Experiments were conducted on a standard Waukesha CFR engine. An air dehumidifier was installed upstream of the carburettor to control the humidity of the intake air. Additional details of the engine setup have previously been described in [12]. Critical engine parameters were monitored and maintained within the accepted limits as defined in the ASTM D2699 standard [23] during the tests. The air-fuel ratio was measured by a Bosch LSU 4.9 wide-band 4

ACS Paragon Plus Environment

Page 4 of 38

Page 5 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

oxygen sensor that was installed on the exhaust system approximately 30 cm downstream of the exhaust port. Each fuel was examined at the airfuel ratio (lambda) that registered the maximum Knock Intensity (KI) value on the ASTM Knockmeter unit. The compression ratio was also set to the barometrically compensated value that corresponded to the measured RON of each fuel. This is referred to as the so-called Standard Knock Intensity (SKI) test condition [23].

2.2

Ethanol/gasoline-surrogate mixtures

It is of particular interest to study the combustion of gasoline surrogates with low-to-intermediate ethanol content, since a small amount of ethanol can change the octane number of a fuel significantly [12]. Nine mixtures of ethanol and gasoline surrogates were selected and studied in this work, with their properties presented in Table 1.

Isooctane (reference fuel grade) was obtained from Haltermann GmbH. Neat ethanol (anhydrous, not denatured) was at least 99.5% pure and obtained from Chem-Supply. The dilute tetraethyl lead (TEL), required for rating fuels with RONs above 100 and for suppressing autoignition (Section 2.3.2), was obtained from Innospec. All other chemicals used in the experiments were analytical grade and at least 99% pure. Blending of all fuels was done gravimetrically using a precision balance with a resolution of 0.001 g. 5

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 38

Table 1 Composition by liquid volume of the fuels used in this study

2.3

No.

Fuel

Isooctane

n-Heptane

Toluene

Ethanol

1

PRF100

100.0

0

0

0

2

PRF100-E10

90.0

0

0

10.0

3

PRF100-E20

80.0

0

0

20.0

4

PRF91

91.0

9.0

0

0

5

PRF91-E10

81.9

8.1

0

10.0

6

PRF91-E20

72.8

7.2

0

20.0

7

TRF91-30

53.2

17.0

29.8

0

8

TRF91-30-E10

47.9

15.3

26.8

10.0

9

TRF91-30-E20

42.6

13.6

23.8

20.0

In-cylinder pressure traces

A Kistler piezoelectric pressure transducer (model 6125C) was used to acquire the in-cylinder pressure at a resolution of 0.10 CAD. The diaphragm of the transducer was recessed by approximately 1 mm relative to the cylinder head, so that a thin layer of silicone rubber could be applied to the diaphragm in an attempt to reduce thermal stress and drift. Since the use of a piezoelectric pressure transducer precludes simultaneous knock reading by a detonation meter, standard conditions were assumed by setting the critical cylinder height and the non-dimensional air/fuel ratio (λ) for a given fuel to the critical values already measured in a previous study [12] (see Table 2). 6

ACS Paragon Plus Environment

Page 7 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Table 2 The RON and the corresponding lambda and critical compression ratio (CCR) for each fuel used in this study

a The

No.

Fuel

RON

Lambdaa

CCR

1

PRF100

100b

0.88

7.82

2

PRF100-E10

106.8c

0.93

9.43

3

PRF100-E20

109.4c

1

9.89

4

PRF91

91b

0.88

6.72

5

PRF91-E10

98.7c

0.92

7.58

6

PRF91-E20

103.8c

0.97

8.77

7

TRF91-30

91.3c

0.91

6.74

8

TRF91-30-E10

97.0c

0.93

7.31

9

TRF91-30-E20

101.4c

0.97

8.17

compression ratio and lambda were obtained under standard RON conditions. b The c The

RON of a PRF are by definition [23].

RONs of all other fuels were obtained from [12].

Two sets of in-cylinder pressure traces were obtained for each fuel at the same engine operating conditions, as detailed below.

2.3.1

In-cylinder pressure with autoignition

Raw pressure traces with autoignition were first obtained. This data set represents the autoigniting traces measured at the engine operating conditions established as per the ASTM procedure for measuring the RON [23]. Cycle7

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

by-cycle variability was accounted for by determining a ‘representative’ raw pressure trace using a cost function, as detailed in Appendix A.

2.3.2

In-cylinder pressure with autoignition suppressed

Pressure traces obtained under standard RON conditions are accompanied by autoignition and (usually) significant pressure fluctuations, which complicates combustion analysis. In this study, a small amount of the dilute tetraethyl lead (TEL) [23] was therefore added to each fuel to suppress autoignition, while maintaining the same engine operating conditions as per the standard RON procedure. The goal is to estimate several key parameters, e.g. the heat transfer and the mass fraction burned (MFB) history, from non-autoigniting, in-cylinder pressure measurements for use in the modelling of autoignition.

Further analysis of in-cylinder pressure data (e.g. Figure 1) obtained at nonautoigniting conditions showed that TEL had little impact on flame propagation. This is in part due to the very low concentration of the dilute TEL added to the fuel, which did not exceed 1.5% by volume in any case. A similar observation was reported by Curry [24].

3

Numerical methods

The first author’s doctoral thesis [25] presents the two-zone model and associated method for examining in-cylinder combustion and autoignition developed for this study (Figure 2). 8

ACS Paragon Plus Environment

Page 8 of 38

Page 9 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Fig. 1. Effects of the dilute TEL on a) the in-cylinder pressure and b) the corresponding mass fraction burned profile for iso-octane at retarded spark (0 instead 13 CAD before TDC under RON conditions). The representative pressure trace was obtained using the method discussed in Appendix A.

3.1

Two-zone model of autoignition

The two-zone autoignition model makes several assumptions, particularly • the air-fuel mixture is spatially homogeneous in temperature and in composition within the unburned and burned zones, • chemical kinetics are modelled in both zones, 9

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 38

• after spark ignition, a flame of negligible volume exists between the unburned and the burned zones, • the mass burning rate is prescribed using the in-cylinder pressure traces with autoignition suppressed (Section 2.3.2 and 3.5), and is incorporated into the model via heat release in the flame, • the flame is always at chemical equilibrium, • mass and energy are transported from the unburned zone to the flame, and then instantaneously from the flame to the burned zone, • there is no heat transfer between zones. Compression starts at intake valve closure (IVC). Following spark ignition, the mass burning rate estimated from the non-autoigniting pressure traces (Section 3.5) is prescribed. Combustion ends as soon as either the end of combustion (as determined from the non-autoigniting pressure traces) is reached, or the fuel in the end gas is fully consumed by autoignition. Cantera (version 1.8) [26] was used to model the fuel chemistry, which was simulated from IVC in order to capture the important reactions that can occur prior to spark ignition.

3.2

Blowby

The rate of change of the total blowby is modelled as a simple function of the instantaneous in-cylinder pressure p(t) as well as the trapped mass m(t),

dml = Cb m(t)p(t), dt

(1)

where Cb is the blowby multiplier, which is merely a tuning parameter used 10

ACS Paragon Plus Environment

Page 11 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Fig. 2. Solution procedure for the two-zone model of [25]. The variables p, T, Y, m and V are the pressure, temperature, species mass fraction, mass and volume respectively. The subscripts u and b denote the unburned and the burned zones.

11

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 38

to attain a required blowby rate. In all cases the total blowby from IVC to the end of expansion was fixed at 1% of the total trapped mass at IVC, in keeping with measurements in the literature for a CFR engine [27].

3.3

Initial mixture temperature

Since it is difficult to accurately determine the mixture temperature at IVC experimentally, this temperature was estimated using the ideal gas law,

TIV C =

pIV C VIV C , mtotal R

(2)

where pIV C is the measured in-cylinder pressure at IVC, VIV C is the cylinder volume at IVC, and R is the specific gas constant. The term mtotal is the total mass trapped in the cylinder at IVC, which can be defined by,

mtotal = mf uel + mair + mresidual ,

(3)

where mf uel is the measured fuel mass per cycle, mair is the inferred air mass based on the measured λ and mresidual is the residual gas mass obtained from modelling (see Section 3.6).

3.4

In-cylinder heat transfer

The heat transfer coefficient (W/m2 K) is modelled using Woschni’s correlation [28], and is given by 12

ACS Paragon Plus Environment

Page 13 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

h = 3.26B −0.2 p0.8 T −0.55 w0.8 ,

(4)

where B is the bore, p the in-cylinder pressure, T the temperature and w the average cylinder gas velocity. The resulting heat flux in the model is then expressed as

q˙ = C0 h(T − Tw ),

(5)

where Tw is the wall temperature and C0 is a calibration constant (or heat transfer multiplier).

The heat transfer model was calibrated for each fuel using the representative in-cylinder pressure trace with autoignition suppressed. The optimum value of C0 was obtained when the sum of squared errors (SSE) between the measured and modelled pressure traces was minimised.

3.5

Prescribing the flame propagation

The simulated pressure traces without chemical kinetics and the optimum C0 were used to estimate the crank-angle resolved MFB profile for each fuel. This determined the amount of mass transported from the unburned zone to the burned zone at each time step. Details of this approach have previously been discussed in [29]. 13

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3.6

Page 14 of 38

Residual gas sub-model

A full-flow GT-Power [30] model of the standard CFR engine was used to calculate the residual gas fraction, taking into account the discharge coefficients of the intake and the exhaust valves [31] and their measured lift profiles. Depending on the fuel (and hence the critical compression ratio), the calculated residual gas fraction ranges from 5% to 7% (see Table 3) under standard RON conditions.

The composition of the residual gases was assumed to be N2 , O2 , H2 , CO, CO2 , H2 O, N O and unburned hydrocarbons. The concentrations of N2 , O2 , H2 , CO, CO2 and H2 O were obtained from chemical equilibrium of these species based on the measured λ. The unburned hydrocarbon concentration in the exhaust for each fuel blend was measured using Cambustion’s HFR400 fast flame ionisation detector. The NO concentration in the exhaust was measured using Autodiagnostics’ ADS9000 gas analyser. This measurement was used to estimate the NO concentration in the trapped air/fuel mixture at IVC (Table 3),

[NO]IV C = xresidual [NO]exhaust ,

(6)

where xresidual is the residual gas fraction modelled by GT-Power and [NO]exhaust is the NO concentration measured in the exhaust. 14

ACS Paragon Plus Environment

Page 15 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Table 3 The residual gas fraction modelled by GT-Power for different fuel blends under standard RON conditions. The NO concentration in the trapped air/fuel mixture at IVC was estimated by Eq. 6.

No.

3.7

Fuel

Residual

Estimated NO

gas fraction

concentration

(% mass)

at IVC (ppm)

1

PRF100

6.3

47

2

PRF100-E10

5.3

73

3

PRF100-E20

5.1

205

4

PRF91

7.3

58

5

PRF91-E10

6.3

87

6

PRF91-E20

5.6

158

7

TRF91-30

7.3

82

8

TRF91-30-E10

6.7

98

9

TRF91-30-E20

6.1

165

Chemical kinetics model

Unless otherwise specified, the detailed ethanol/TRF/diisobutylene mechanism from Andrae [32] was used. The NO submechanism from Contino et al. [20] was incorporated into Andrae’s mechanism [32] in an attempt to capture the NO sensitisation on the oxidation of these ethanol/gasoline-surrogate 15

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

blends. The final kinetic mechanism, as a result of combining Andrae’s mechanism and the NO submechanism from Contino et al. [20], consists of 1150 species and 8525 reactions.

4

4.1

Results and discussion

Effect of NO in the residual gases

NO has been shown to advance autoignition onset in different studies at a very low concentration, e.g. [17,20]. A numerical study was thus conducted by modelling different concentrations of NO in the residual gases. The kinetic mechanism from Curran et al. [33,34] was used for the baseline case without NO, while those from Contino et al. [20] and Ranzi et al. [35] were used where NO was present.

Figure 3 shows that Curran et al.’s [33,34] and Contino et al.’s [20] models agree closely in the absence of NO, and that autoignition onset is always advanced with the addition of NO content in the models of Contino et al. [20] and Ranzi et al. [35]. Of note is that the NO concentration in the trapped mass at IVC for isooctane under RON conditions was estimated to be roughly 50 ppm based on exhaust measurement. This suggests that the presence of NO should significantly influence autoignition onset under RON conditions. 16

ACS Paragon Plus Environment

Page 16 of 38

Page 17 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Fig. 3. Effects of increasing NO concentration on a) in-cylinder pressure and b) autoignition onset of isooctane under RON conditions, as modelled by the two-zone combustion model. The NO concentration in the trapped mass at IVC was varied from 0 to 400 ppm (mole fraction). The reference case without NO used the mechanism from Curran et al. [33,34]. Mechanisms from Contino et al. [20] and Ranzi at al. (Version 1212) [35] were used where NO was present.

4.2

Autoignition of ethanol/isooctane blends

Combustion of ethanol/isooctane blends of up to 20% ethanol content by volume and under standard RON conditions was modelled with full chemical kinetics. Figure 4 shows the modelled and measured in-cylinder pressure for 17

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 38

these fuels, and Figure 5 shows the fuel oxidation profiles. Three different concentrations of NO were investigated for each fuel. For isooctane, agreement between the simulated and the measured pressure traces improves with increasing NO concentration. For isooctane-E10 and isooctane-E20, the modelled autoignition onset is more advanced than experiment. Clearly, the presence of NO increases the end-gas temperature at a given crank angle. Figure 4 shows that at MFB50, the end-gas temperature with NO present can differ from that without NO by approximately 50 K in the case of isooctane, and this difference is magnified with the addition of ethanol. For isooctane-E20, autoignition occurs prior to MFB50 with the presence of NO, and thus it is more relevant to compare the end-gas temperature attained when 50% of the fuel species has oxidised (Figure 4). For all fuels, the end-gas temperature at this instant is around 1000 - 1050 K, and remains relatively unchanged by different NO concentrations. Importantly, NO considerably shortens the time required to reach this temperature. Such results are very likely due to the ability of NO to form more OH radicals by reacting with HO2 [19,18] (a product of fuel decomposition) via NO + HO2 → NO2 + OH.

(7)

NO is further formed through the H atom assisted destruction of NO2 , NO2 + H → NO + OH,

(8)

which is typically a much faster process than reaction 7 [18]. Combining these two reactions yields HO2 + H → 2OH.

(9)

In the absence of NO and at low temperatures, HO2 reacts with H predomi18

ACS Paragon Plus Environment

Page 19 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Fig. 4. In-cylinder pressure and the end-gas temperature vs CAD for a) isooctane, b) isooctane-E10 and c) isooctane-E20 under RON conditions. The crosses denote the CAD at which 50% of the major fuel species has oxidised in the end gas.

19

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 5. Fuel component mass fraction and fuel mass fraction burned with respect to CAD for a) isooctane, b) isooctane-E10 and c) isooctane-E20 under RON conditions. The NO concentration used in modelling the mass fraction burned was estimated by Eq. 6.

20

ACS Paragon Plus Environment

Page 20 of 38

Page 21 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

nantly via the recombination process, HO2 + H → H2 O2 ,

(10)

whereas at high temperatures H2 O2 formed in reaction 10 would have sufficient energy to directly dissociate to two OH radicals, equivalent to reaction 9 [18]. However, consistent with the trends found in the in-cylinder pressure and temperature previously, the effects of NO diminish with further increases in its concentration. There are several reactions that may cause this effect, including those that diminish the ability of NO to oxidise hydrocarbons at higher temperatures [17,18]. These arguments cannot be confirmed on this engine without more experimental data and/or more fundamental study.

4.3

Ethanol/PRF91 blends

Figure 6 shows the modelled in-cylinder pressure and the end-gas temperature for ethanol/PRF91 blends under standard RON conditions. In all cases, autoignition is observed, although its onset again lags that observed in experiments. Similarly, NO again significantly advances autoignition onset, and the effects of NO diminish when its concentration is further increased. Likewise, NO also increases the rate of the end-gas temperature rise for these fuel blends. Overall, these results suggest that the autoignition kinetics of ethanol/PRF91 is similar to that of ethanol/isooctane. Regardless of NO content, two-stage ignition is modelled for PRF91 under RON conditions. While this is expected for n-heptane, it is interesting that both isooctane and n-heptane oxidise in two stages (Figure 7). This is contrary to the single-stage oxidation previously observed for isooctane in Figure 5, 21

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 38

demonstrating coupling between n-heptane’s and isooctane’s oxidation.

The increasing presence of ethanol appears to suppress such two-stage characteristics. A recent study by Haas et al. [36] reported that the oxidation of n-heptane at low temperatures generates radicals that will subsequently react with ethanol to form potentially more stable species. Such chemistry is initiated by

nC7 H16 + O2 → C7 H15 + HO2 ,

(11)

which is followed by the reaction between ethanol and HO2 to form α-hydroxyethyl,

C2 H5 OH + HO2 → CH3 CHOH + H2 O2 .

(12)

The species α-Hydroxyethyl then rapidly reacts with O2 to form acetaldehyde and thus regenerate HO2 ,

CH3 CHOH + O2 → CH3 CHO + HO2 .

(13)

Overall, reactions 12 and 13 when combined result in stable acetaldehyde and H2 O2 as products [36]. Haas et al. [36] suggested that these specific reactions are responsible for the lack of the negative-temperature-coefficient (NTC) or two-stage behaviour observed in blending ethanol with fuels like n-heptane. Specifically, the competition for HO2 and OH radicals between ethanol and n-heptane leads to slower regeneration of OH. This subsequently reduces the reactive radical pool that would otherwise be growing when n-heptane alone is oxidised, suppressing any low-temperature reactivity. 22

ACS Paragon Plus Environment

Page 23 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Fig. 6. The in-cylinder pressure and the end-gas temperature with respect to CAD for a) PRF91, b) PRF91-E10 and c) PRF91-E20 under RON conditions. The crosses denote the CAD at which 50% of the major fuel species has oxidised in the end gas.

23

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 7. Fuel component mass fraction and fuel mass fraction burned with respect to CAD for a) PRF91, b) PRF91-E10 and c) PRF91-E20 under RON conditions. The NO concentration used in modelling the mass fraction burned was estimated by Eq. 6.

24

ACS Paragon Plus Environment

Page 24 of 38

Page 25 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

4.4

Ethanol/TRF91-30 blends

Similar to that observed for ethanol/PRF91 blends, NO is predicted to significantly advance autoignition onset for ethanol/TRF91-30 blends under standard RON conditions (Figure 8). It is however interesting that the effect of NO on the in-cylinder pressure development is smaller in the case of TRF9130 compared with other fuels. It appears that the oxidation of toluene (when ethanol is not present) is less sensitive to NO, although such interaction is not validated experimentally and requires further study. Figure 9 shows the fuel oxidation and mass fraction burned profiles for different ethanol/TRF91-30 blends. Of all the fuels, toluene is consumed most slowly, followed by ethanol, isooctane and lastly n-heptane. In the case of TRF91-30, each fuel species appears to oxidise in two stages, regardless of NO content. This again suggests that although n-heptane is the only fuel with apparent two-stage ignition, the chemistry of isooctane and toluene is coupled to that of n-heptane. Adding ethanol again suppresses this two-stage oxidation. As previously discussed, this is likely a result of ethanol’s ability to suppress the low-temperature reactivity of n-heptane via reactions 12 and 13.

4.5

Model uncertainty and implications

Figure 10 compares the modelled and measured autoignition timing under standard RON conditions for all fuels. These results suggest that the incorporation of residual NO significantly improves the agreement between simulations and experiment for all cases with no or low ethanol content. The presence of NO is modelled to significantly advance autoignition onset, though such ef25

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 8. The simulated in-cylinder pressure and the end-gas temperature with respect to CAD for a) TRF91-30, b) TRF91-30-E10 and c) TRF91-30-E20 under RON conditions. The crosses denote the CAD at which 50% of the major fuel species has oxidised in the end gas.

26

ACS Paragon Plus Environment

Page 26 of 38

Page 27 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Fig. 9. Fuel component mass fraction and fuel mass fraction burned with respect to CAD for a) TRF91-30, b) TRF91-30-E10 and c) TRF91-30-E20 under RON conditions. The NO concentration used in modelling the mass fraction burned was estimated by Eq. 6.

27

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 10. Modelled autoignition onset timing (CAD) for different fuels with and without NO. The measured autoignition onset is also included for reference.

fects diminish when its concentration increases. It is nonetheless concerning that the simulations for the hydrocarbon-only fuels agree worse with experiment than those with ethanol present. This is in contrast to earlier studies by the group, showing that similar simulations of C3 to C4 hydrocarbons in this same engine can agree reasonably well with experiment [21]. However, caution should be exercised when interpreting these results, as there is considerable uncertainty in several aspects of this problem. First, the NO submechanism used was originally developed only for the oxidation of isooctane [20], since it appears that none currently exists for the ethanol containing fuels of this study. The chemical interaction between NO and n-heptane, toluene and ethanol has also not yet been validated. Further, although the chemistry of neat ethanol is reasonably well understood, few studies in the literature have explored the oxidation of ethanol blended with different gasoline surrogates, even without NO present. Complicating the problem further, different published mechanisms predict different trends in the low-temperature reactivity of ethanol [37,38]. 28

ACS Paragon Plus Environment

Page 28 of 38

Page 29 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Based on these observations, implications for future kinetic model development include the following. (1) A validated kinetic model of the interaction between residual species, particularly NO, with common gasoline surrogates and ethanol at engine representative conditions is required. (2) The chemical interaction of ethanol with isooctane, n-heptane and toluene needs further modelling work. In particular, the modelled autoignition onset (Figure 10) of these fuel blends is not consistent with the observed synergism and antagonism in the octane rating experiments reported in a recent study by the authors [12]. (3) Whilst accurate kinetic models are challenging to obtain, so too is the modelling of several other phenomena that also affect autoignition significantly. There is little point in solving computationally more demanding, full chemical models if the flame propagation, the engine heat transfer or the mixture composition and temperature inhomogeneity at IVC cannot be measured or modelled with a similar degree of accuracy. Thus, the reliable modelling of autoignition and knock does not only involve work on the fuel chemistry, but also work on all aspects of the modelling of fluid motion through an engine.

5

Conclusions

This paper presented an experimental and numerical investigation of the autoignition of ethanol blended with several Primary Reference Fuels and Toluene Reference Fuels in a CFR engine operating under standard, Research Octane Number conditions. A systematic investigation of the effect of differ29

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ent kinetic and non-kinetic mechanisms on the autoignition of ethanol blended with these surrogates was conducted. Of particular interest was the sensitivity of these blends’ autoignition to NO, which was present in the engine’s residual gases.

In-cylinder pressure data was first acquired for each fuel for both autoigniting and non-autoigniting operation. While both sets of data were measured under the same, standard RON conditions, the non-autoigniting pressure traces were obtained by adding a small amount of dilute tetraethyl lead (TEL) to each fuel. These small amounts of TEL were shown to have an insignificant effect on turbulent flame propagation, and permitted the calibration of a two-zone model of the engine in the absence of autoignition.

The autoignition of isooctane at standard RON conditions was then examined using the kinetic models of Curran [33,34], Contino et al. [20] and Ranzi et al. [35]. The former two models gave very similar autoignition timing when there was no residual NO, but this timing lagged that measured in the engine. The NO submechanism in the model of Contino et al. [20] then enabled simulations with residual NO by its integration into the ethanol/TRF/diisobutylene model of Andrae [32], enabling kinetic simulations of ethanol blended with PRF and TRF surrogates with residual NO present. These simulations showed that the incorporation of NO chemistry improved the level of agreement between experiment and simulation for fuels with low ethanol content, and that ethanol suppressed the low temperature reactivity of n-heptane. This final result was in keeping with previous, more fundamental studies [36].

However, close agreement between the simulated and measured autoignition timing across all fuel blends was not observed. These kinetic models could 30

ACS Paragon Plus Environment

Page 30 of 38

Page 31 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

therefore not comprehensively explain the significant synergism and antagonism observed in a recent study by the group of these fuels’ octane number [12], demonstrating that further research is required. As part of this, these results also contribute to a growing body of evidence [19,17,18,21] suggesting that NO chemistry plays an important role in engine autoignition, and so should be included in kinetic models intended for this purpose.

6

Acknowledgements

This research was supported by the Advanced Centre for Automotive Research and Testing (www.acart.com.au), the Ford Motor Company and the Australian Research Council. We also take this opportunity to acknowledge Prof. Brian Haynes of the University of Sydney, whose recent 65th birthday is celebrated by this special volume. Prof. Haynes’ career continues as one that applies scientific rigour to the solution of important practical problems. We hope that this special volume in honour of this fine researcher plays a small part in continuing this tradition.

References

[1] C. L. Cummins, Internal fire, Society of Automotive Engineers, Warrendale, PA, 1989. [2] T. Wallner, S. A. Miers, S. McConnell, A comparison of ethanol and butanol as oxygenates using a direct-injection, spark-ignition engine, Journal of

31

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Engineering for Gas Turbines and Power 131 (2009) 032802. [3] C. Jin, M. Yao, H. Liu, C. F. Lee, J. Ji, Progress in the production and application of n-butanol as a biofuel, Renewable and Sustainable Energy Reviews 15 (8) (2011) 4080–4106. [4] B. He, M. Liu, J. Yuan, H. Zhao, Combustion and emission characteristics of a HCCI engine fuelled with n-butanol/gasoline blends, Fuel 108 (2013) 668–674. [5] OECD-FAO Agricultural Outlook 2012, Tech. rep., OECD-Food and Agriculture Organization of the United Nations (2012). [6] U. S. Energy Information Administration website (2014). [7] W. D. Hsieh, R. H. Chen, T. L. Wu, T. H. Lin, Engine performance and pollutant emission of an SI engine using ethanol-gasoline blended fuels, Atmospheric Environment 36 (3) (2002) 403–410. [8] F. Y¨ uksel, B. Y¨ uksel, The use of ethanol-gasoline blend as a fuel in an SI engine, Renewable Energy 29 (7) (2004) 1181 – 1191. [9] M. Al-Baghdadi, Measurement and prediction study of the effect of ethanol blending on the performance and pollutants emission of a four-stroke spark ignition engine, Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering 222 (5) (2008) 859–873. [10] R. A. Stein, J. E. Anderson, T. J. Wallington, An overview of the effects of ethanol-gasoline blends on SI engine performance, fuel efficiency, and emissions, SAE Int. J. Engines 6 (2013) 470–487. [11] T. M. Foong, K. J. Morganti, M. J. Brear, G. da Silva, Y. Yang, F. L. Dryer, The effect of charge cooling on the RON of ethanol/gasoline blends, SAE Int. J. Fuels Lubr. 6 (2013) 34–43.

32

ACS Paragon Plus Environment

Page 32 of 38

Page 33 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

[12] T. M. Foong, K. J. Morganti, M. J. Brear, G. da Silva, Y. Yang, F. L. Dryer, The octane numbers of ethanol blended with gasoline and its surrogates, Fuel 115 (2014) 727–739. [13] S. Mani Sarathy, P. Oswald, N. Hansen, K. Kohse-Hinghaus, Alcohol combustion chemistry, Progress in Energy and Combustion Science 44 (2014) 40–102. [14] D. Splitter, J. Szybist, Experimental investigation of spark-ignited combustion with high-octane biofuels and egr. 1. engine load range and downsize downspeed opportunity, Energy and Fuels 28 (2014) 14181431. [15] D. Splitter, J. Szybist, Experimental investigation of spark-ignited combustion with high-octane biofuels and egr. 2. fuel and egr effects on knock-limited load and speed, Energy and Fuels 28 (2014) 14321445. [16] J. Bergthorson, M. Thomson, A review of the combustion and emissions properties of advanced transportation biofuels and their impact on existing and future engines, Renewable and Sustainable Energy Reviews 42 (2015) 13931417. [17] A. Frassoldati, T. Faravelli, E. Ranzi, Kinetic modeling of the interactions between NO and hydrocarbons at high temperature, Combustion and Flame 135 (1-2) (2003) 97–112. [18] H. Machrafi, S. Cavadias, P. Guibert, An experimental and numerical investigation on the influence of external gas recirculation on the HCCI autoignition process in an engine: Thermal, diluting, and chemical effects, Combustion and Flame 155 (3) (2008) 476–489. [19] T. Amano, F. L. Dryer, Effect of dimethyl ether, NOx, and ethane on ch4 oxidation: High pressure, intermediate-temperature experiments and modeling, Symposium (International) on Combustion 27 (1) (1998) 397 – 404.

33

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[20] F. Contino, F. Foucher, P. Dagaut, T. Lucchini, G. D’Errico, C. Mouna¨ımRousselle, Experimental and numerical analysis of nitric oxide effect on the ignition of iso-octane in a single cylinder HCCI engine, Combustion and Flame 160 (2013) 1476–1483. [21] K. J. Morganti, T. M. Foong, M. J. Brear, G. da Silva, Y. Yang, F. L. Dryer, The autoignition of liquidied petroleum gas (lpg) in spark-ignition engines, Proceedings of the Combustion Institute 35 (2014) 2933–2940. [22] H. Yuan, T. Foong, Z. Chen, Y. Yang, M. Brear, T. Leone, J. Anderson, Modeling of trace knock in a modern si engine fuelled by ethanol/gasoline blends, SAE Technical Paper 2015-01-1242. [23] ASTM International, Standard test method for research octane number of spark-ignition engine fuel, ASTM D2699-11 (2011). [24] S. Curry, Effect of antiknocks on flame propagation in a spark ignition engine, Symposium (International) on Combustion 9 (1) (1963) 1056 – 1068. [25] T. Foong, On the autoignition of ethanol/gasoline blends in spark-ignition engines, Ph.D. thesis, The University of Melbourne (2013). [26] D. G. Goodwin, An open source, extensible software suite for CVD process simulation, Tech. rep., Division of Engineering and Applied Science, California Institute of Technology (2003). [27] J. B. Heywood, Internal combustion engine fundamentals, McGraw-Hill, New York, 1988. [28] G. Woschni, A universally applicable equation for the instantaneous heat transfer coefficient in the internal combustion engine, SAE Technical Paper 670931. [29] K. J. Morganti, T. M. Foong, M. J. Brear, G. da Silva, Y. Yang, F. L. Dryer,

34

ACS Paragon Plus Environment

Page 34 of 38

Page 35 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Design and analysis of a modified CFR engine for the octane rating of liquefied petroleum gases (LPG), SAE Int. J. Fuels Lubr. 7 (2014) 283–300. [30] Gamma Technologies, GT-Power user’s manual, version 7.3 (2012). [31] J. Vancoillie, L. Sileghem, Personal communication (Nov. 2012). [32] J. C. G. Andrae, Development of a detailed kinetic model for gasoline surrogate fuels, Fuel 87 (10-11) (2008) 2013–2022. [33] H. J. Curran, P. Gaffuri, W. J. Pitz, C. K. Westbrook, A comprehensive modeling study of n-heptane oxidation, Combustion and Flame 114 (1-2) (1998) 149–177. [34] H. J. Curran, P. Gaffuri, W. J. Pitz, C. K. Westbrook, A comprehensive modeling study of iso-octane oxidation, Combustion and Flame 129 (3) (2002) 253–280. [35] E. Ranzi, A. Frassoldati, R. Grana, A. Cuoci, T. Faravelli, A. P. Kelley, C. K. Law, Hierarchical and comparative kinetic modeling of laminar flame speeds of hydrocarbon and oxygenated fuels, Progress in Energy and Combustion Science 38 (4) (2012) 468–501. [36] F. M. Haas, M. Chaos, F. L. Dryer, Low and intermediate temperature oxidation of ethanol and ethanol-PRF blends: An experimental and modeling study, Combustion and Flame 156 (12) (2009) 2346 – 2350. [37] N. M. Marinov, A detailed chemical kinetic model for high temperature ethanol oxidation, International Journal of Chemical Kinetics 31 (3) (1999) 183–220. [38] L. R. Cancino, M. Fikri, A. A. M. Oliveira, C. Schulz, Ignition delay times of ethanol-containing multi-component gasoline surrogates: Shock-tube experiments and detailed modeling, Fuel 90 (3) (2011) 1238–1244.

35

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 38

[39] A. Swarts, A. Yates, C. Viljoen, R. Coetzer, Standard knock intensity revisited: Atypical burn rate characteristics identified in the CFR octane rating engine, SAE Technical Paper 2004-01-1850.

A

A.1

Determining the representative in-cylinder pressure trace

Non-autoigniting traces

To study the effect of TEL on flame propagation, non-autoigniting pressure traces were acquired for isooctane with and without TEL added but at a much retarded spark timing. A spark timing at TDC, as opposed to 13 CAD before TDC in the standard RON tests, was chosen to ensure that autoignition did not occur even when no TEL was added, while all other operating parameters were set to those at standard knock intensity. For each case, one representative raw pressure trace which minimises the cost function was selected out of 300 cycles. This cost function is given by



IMEPi − IMEPavg Cost =  IMEPavg

!2

PPi − PPavg + PPavg

!2  ,

(A.1)

where IMEP denotes the indicated mean effective pressure, PP denotes the peak in-cylinder pressure, and the subscripts i and avg denote the ith cycle and the average quantity (evaluated from all cycles) respectively. 36

ACS Paragon Plus Environment

Page 37 of 38

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

A.2

Autoigniting traces

A different cost function is used for traces with autoignition, due to significant variance in the peak pressure as well as its location, and more difficult estimate of the IMEP using common methods. Any cycle which minimises the cost function is deemed the representative cycle, with the cost function now defined as

h

i

Cost = (MFB50i − MFB50avg )2 + (KPi − KPavg )2 ,

(A.2)

where M F B50 is the crank angle at which 50% mass fraction burned is reached, and KP is the ‘knock point’. Again, the subscripts i and avg denote the ith cycle and the average quantity (evaluated from all cycles) respectively. The ‘knock point’ is the crank angle at which there is a distinctive change in the slope of the in-cylinder pressure development under knocking conditions in a CFR engine [39]. This is the crank angle at which the pressure rise due to autoignition dominates that due to flame propagation, although significant end-gas reactions may have already occurred prior to this point. To estimate the knock point, the second derivative of the in-cylinder pressure with respect to CAD was evaluated. The knock point was then assumed to occur at the instant where the absolute value of the second derivative is greater than a threshold (2000 kPa/deg2 ). Figure A.1 shows selected in-cylinder pressure traces with the corresponding MFB50 and knock points labelled.

37

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. A.1. In-cylinder pressure traces measured for isooctane at the standard RON conditions, along with the associated MFB50 points (diamonds) and the knock points (circles). Traces shown are the representative trace (dashed line), and those with the most advanced and retarded knock points out of 900 cycles.

38

ACS Paragon Plus Environment

Page 38 of 38