Effects of fuels on flash boiling spray from a GDI hollow cone piezoelectric injector

Fuel 257 (2019) 116080

Contents lists available at ScienceDirect

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

Full Length Article

Effects of fuels on flash boiling spray from a GDI hollow cone piezoelectric injector Libing Wang, Kaushik Nonavinakere Vinod, Tiegang Fang

T

⁎

Department of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh, NC 27695, United States

ARTICLE INFO

ABSTRACT

Keywords: Flash boiling Ambient-to-saturation pressure ratio Superheated degree Hollow cone spray Gasoline direct injection

Flash-boiling of fuel sprays can have a significant effect on spray formation and its characteristics due to bubble nucleation, growth, and phase change, producing explosive-like atomization and complex spray structures. In this work, experiments were conducted to study the spray of both pure substance fuels (pure isooctane, pure ethanol) and multicomponent fuels (50/50 mixture of isooctane and ethanol, commercial gasoline), under flash boiling conditions and non-flash boiling conditions. Under different temperature and ambient pressure, different superheated degrees can be achieved for the fuels. Pure substances have a single vapor pressure curve, while mixtures do not have a single boiling point at a given pressure, and a two-phase region exists for multicomponent fuel. Under the same conditions, ethanol has higher superheated degree compared to isooctane, but the heat of vaporization for ethanol is also much higher. This contributes to the fact that less boiling is observed in the ethanol spray with longer penetration in several cases. Mixture 50/50 shows a good average of isooctane and ethanol for spray penetration and spray front plume ratio analysis. Gasoline, due to its low initial boiling point and wide range of components, has the widest plume ratio distribution and smallest gradient, as well as complex peak penetration velocity distribution. The results also imply that adding low boiling point (preferably with low heat of vaporization as well) additive or component to high boiling point fuel can facilitate flash boiling, fuel vaporization and mixing.

1. Introduction When fuel is injected into an environment where the pressure is lower than its saturation pressure rapid boiling occurs causing the fuel to vaporize quickly, and this is known as the flash boiling phenomenon. This has been used as a method to atomize liquid fuels inside a combustion chamber and as a way to reduce emissions [1,2]. The same phenomenon was also observed by Vieira et al. to produce highly reactive and explosive two-phase clouds in laboratory experiments [3]. It was identified by Geroge et al. that flash boiling helps the liquid attain equilibrium [4]. The flash boiling mechanism was classified into three stages including bubble nucleation, bubble growth, and two-phase flow [5]. Once the bubble nucleation stage has developed the fluctuations in the pressure in the nucleation sites can cause the bubbles to either grow or collapse as discussed by Plasset et al. [6]. While the bubbles formed are still relatively small, the growth rate will be low and will be restricted by the fluid surface tension, this can be overcome if the degree of super heat is high enough that the growth rate will increase as the bubble size increases. The number of bubbles in a droplet is determined by the diameter of the droplet, surface tension, liquid viscosity and the

⁎

growth rate of bubble. Then droplets break up into smaller droplets and the momentum of the parent droplet is distributed to the new droplets. This leads to a two-phase flow once the flash boiling process is complete and creates a region with both liquid and vapor phases. Flash boiling can be considered as phase transition mechanism from liquid to vapor similar to cavitation, the difference being that the flash boiling effect is thermodynamically driven while the cavitation process is a mechanical phenomenon. Polanco et al. and Simões examined isooctane under flash boiling conditions and reported that flashing takes place at the surface of the liquid core through an evaporation wave which leads to a sudden discontinuous evaporation of liquid [7,8]. Studies have shown solid evidences relating flash boiling to liquid jet breakup and the importance of bubble growth that accompanies the boiling on the atomization process [9,10]. Steelant et al. conducted experiments on the characteristics and morphology of a flashing jet using high speed shadowgraph and found that the level of superheat had a strong influence on the inception of flash boiling within a depressurized jet [11]. One of the major factors affecting the combustion efficiency in a gasoline direction injection (GDI) engine is the mixture quality as shown by experiments conducted on mixture formation inside the

Corresponding author. E-mail address: [email protected] (T. Fang).

https://doi.org/10.1016/j.fuel.2019.116080 Received 6 June 2019; Received in revised form 19 August 2019; Accepted 21 August 2019 Available online 30 August 2019 0016-2361/ © 2019 Elsevier Ltd. All rights reserved.

Fuel 257 (2019) 116080

L. Wang, et al.

Table 1 All experimental settings and conditions and the vapor pressure value. Camera frame rate Resolution F/# Exposure time Injection pressure Injection duration

8113 fps 240 × 240 5.6 2 μs 8 MPa 1 ms

Fuel type

Isooctane

Ethanol

50/50 mixture

Commercial gasoline

Injected mass (mg) Boiling point under room pressure (°C)

34.13 99

33.32 78

34.07 78–99

31 35–200

Temperature (°C)

Ambient pressure (kPa)

25 50 75 90 100 125

1 1 1 1 1 1

3 3 3 3 3 3

5 5 5 5 5 5

7 7 7 7 7 7

10 10 10 10 10 10

30 30 30 30 30 30

100 100 100 100 100 100

Corresponding vapor pressure for isooctane (kPa)

Corresponding vapor pressure for ethanol (kPa)

6.578 19.533 48.252 77.417 103.525 198.896

7.833 29.368 88.682 158.534 225.978 498.165

Vapor pressure of ethanol (kPa)

Fig. 1. Schematic of the optical setup (1, point light source; 2, parabolic mirror; 3, constant volume chamber; 4, parabolic mirror; 5, knife edge; 6, high-speed camera; 7, injector). Table 3 Isooctane ambient-to-saturation pressure ratio (Pa/Ps) of all experimental conditions.

100

1 kPa 3 kPa 5 kPa 7 kPa 10 kPa 30 kPa 100 kPa

10

1

10

30

50

70

90

Temperature (°C)

110

130

Fig. 2. All experimental conditions and the vapor pressure curve of isooctane and ethanol.

A

B

C

Isooctane

194.64–298.44 297.51–373.28

3.94736 3.93679

1282.332 1257.840

−48.444 −52.415

Ethanol

216.15–353.15 350.15–516.15

8.20417 7.68117

1642.89 1332.04

−42.85 −73.95

50 °C

75 °C

90 °C

100 °C

125 °C

0.152 0.456 0.760 1.064 1.520 4.561 15.202

0.051 0.154 0.256 0.358 0.512 1.536 5.120

0.021 0.062 0.104 0.145 0.207 0.622 2.072

0.013 0.039 0.065 0.090 0.129 0.388 1.292

0.010 0.029 0.048 0.068 0.097 0.290 0.966

0.005 0.015 0.025 0.035 0.050 0.151 0.503

Table 4 Ethanol ambient-to-saturation pressure ratio (Pa/Ps) of all experimental conditions.

Table 2 Parameters for isooctane and ethanol in Antoine equation [45,46]. Temperature (K)

25 °C

1 kPa 3 kPa 5 kPa 7 kPa 10 kPa 30 kPa 100 kPa

2

25 °C

50 °C

75 °C

90 °C

100 °C

125 °C

0.126 0.379 0.631 0.884 1.263 3.788 12.628

0.033 0.098 0.164 0.230 0.328 0.984 3.279

0.011 0.032 0.053 0.074 0.106 0.319 1.062

0.006 0.018 0.030 0.041 0.059 0.177 0.591

0.004 0.012 0.021 0.029 0.041 0.124 0.413

0.002 0.006 0.009 0.013 0.019 0.056 0.187

Fuel 257 (2019) 116080

L. Wang, et al.

Isooctane

and differentiated it from the near-field collapse at an elevated ambient pressure condition. In Ref. [12], both the macroscopic and microscopic flash-boiling spray characteristics were studied in an optical GDI engine through PDPA and high-speed photography. Bornschlegel et al. [24] investigated flash-boiling spray from a single-hole injector in order to provide the basis for the study of jet-to-jet interaction in flash-boiling sprays from multi-hole injectors. Lacey et al. [25] developed a new framework to characterize the behavior of flash-boiling multi-hole GDI sprays and proposed a general criteria for both spray collapse and plume interaction due to flash boiling, which appeared to be applicable to any fuel injector. There has been quite a bit of work investigating the effects of ambient temperature, ambient pressure and fluid properties [26–28]. The results suggest that rapid evaporation of flash-boiling spray leads to a higher vapor concentration. In another study analyzing the dynamic interaction between the gas and liquid in and out of the nozzle, the dynamic feature of flash-boiling spray was closely connected with the dynamics of the in-nozzle flow [29]. Wu et al. [30] showed that nozzle length had contrary effects on spray breakup under subcooled and superheated conditions. Longer nozzle length was found to enhance atomization of flash-boiling sprays. Guo et al. [31] studied the radial expansion of a flashing jet from a single-hole GDI injector to show the extent of radial jet expansion under flashing conditions is determined by chemical potential of phase change and ambient resistance. Dong et al. [32] observed strong plume-plume and plume-air interaction under flash-boiling conditions during the intake stroke. They found that with the increase of superheated degree, the diffusion rate of blue flame becomes higher and yellow flame is less under flash-boiling, indicating that combustion is more complete. Yang et al. [33] studied the effect of flash-boiling sprays on the combustion characteristics of a GDI optical engine under cold start. They showed an improvement of indicated mean effective pressure (IMEP) and reduction of soot. Wu et al. [34] showed that spray processes of a multi-hole injector could not be well represented by the single-hole injector due to spray collapse. Strong plume interaction triggers severe spray collapse and results in longer spray penetration and smaller spray angles. Jiang et al. [35] analyzed the effect of GDI fuel injector hole geometry on flash-boiling sprays. It was found that a typically-sized hole diameter could produce collapsed sprays more easily than others. Several studies have been previously conducted to document the spray properties of alternative fuels in engines. One such study shows [36] that the main spray tip penetration decreases, and the spray angle increases with the increase in concentration of ethanol in the fuel. In another work done by Aleiferis et al. the spray plume was induced to collapse due to the low boiling point components in the mixture, even though the fuel was not under flash boiling conditions [37]. In some other experiments, the addition of ethanol made the evaporation poorer compared to straight gasoline [38]. Evaporation characteristics of E100 and E85 were compared under different conditions to determine that high boiling point components in gasoline are present mainly at middle to downstream of the spray after end of injection (EOI), which is caused by fast evaporation of the ethanol in the mixture [39]. In another study performed on E100, E85 and Gasoline in a GDI system the results showed that the increase in injection pressure caused an increase in axial spray tip penetration [40]. The study also suggested that E100 fuel has the largest droplet size among the tested E100, E85 and G100 with

10

Pa/Ps

1 0.1 0.01 0.001 10

30

50

70

90

110

130

90

110

130

Temperature (°C)

Ethanol 10

Pa/Ps

1 0.1 0.01 0.001 10

30

50

70

Temperature (°C)

Fig. 3. Ambient-to-saturation ratio (Pa/Ps) for isooctane and ethanol under all experimental conditions.

combustion chamber of a GDI engine and from studying the combustion processes in those same systems [12–14]. Chan et al. focused on studying the effects of injection pressure on the collapse of the flash boiling sprays in a GDI engine and found that for a fuel with a high superheated degree the spray plumes caused by flash boiling can be overcome to have the spray maintain its original direction with the increase in injection pressure thus causing the spray plumes to be reduced [15]. Further studies also suggested that flash boiling can enhance vaporization of the fuel and improve atomization if maintained within a limit above which it can increase the danger of vapor lock inside the fuel lines and alter the combustion performance [16]. Flash boiling can affect the fuel-air mixing process and spray by modifying the characteristics like cone angle, penetration, and droplet size and velocities. Many studies focused on flash boiling with the use of single and multi-hole injectors [17–21]. Montanaro et al. [19] identified parameters affecting flash-boiling, such as fuel temperature and corresponding boiling pressure point. Wood et al. [20] showed that under flash boiling conditions the large droplets within the spray break down into smaller droplets by flash boiling. Huang et al. [22] compared the spray and evaporation characteristics of ethanol and gasoline from nonevaporating to flash-boiling conditions. They suggest that for both fuels, modifying the fuel temperature significantly affects the macroscopic characteristics of flash-boiling sprays. Guo et al. [23] specifically investigated the far-field spray collapse under the flash boiling condition

3

Fuel 257 (2019) 116080

L. Wang, et al.

-52.7ºC

-52.7ºC

-52.7ºC

-52.7ºC

-4.31ºC

-4.31ºC

-4.31ºC

-4.31ºC

45.75ºC

45.75ºC

45.75ºC

45.75ºC

47.3ºC

47.3ºC

47.3ºC

47.3ºC

75ºC

75ºC

75ºC

75ºC

Fig. 4. Selected typical ethanol spray development under certain cases with different superheated degree (shown at the lower right corner for each image).

4

Fuel 257 (2019) 116080

L. Wang, et al.

25°C

100 100kPa

Axial penetration(mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1

1.2

1.4

Time after trigger signal (ms)

50°C

100 100kPa

Axial penetration(mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

Time after trigger signal (ms) Fig. 5. Spray axial penetration development with time for ethanol.

increase in injection pressure due to E100 having a high kinematic viscosity and surface tension compared to other test fuels. Most of the previous studies on flash boiling spray in GDI applications, however, were conducted on a multi-hole or single-hole injector. Few works were done on the flash boiling of hollow cone sprays. In this study, experiments were carried out to study the flash boiling spray of a hollow cone GDI piezoelectric injector. The scope of this study is to characterize the spray of both pure fuels and multicomponent fuels under a wide range of conditions involving flash boiling under the effects of temperature and ambient pressure (i.e., different superheated

degrees). The effects of superheated degree and flash boiling on the spray shape and spray penetration development of different fuels are analyzed and compared. The differences between pure substance fuel and multicomponent fuel are also discussed. 2. Experimental setup All the experiments were conducted in a Constant Volume Chamber (CVC). The CVC is a steel chamber with an inner volume of 0.95 L with 6 access ports leading into the chamber volume. It has through optical

5

Fuel 257 (2019) 116080

L. Wang, et al.

75°C

100 100kPa

Axial penetration(mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1

1.2

1.4

Time after trigger signal (ms)

90°C

100

100kPa

90

30kPa

Axial penetration(mm)

80

10kPa 7kPa

70

5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

Time after trigger signal (ms) Fig. 5. (continued)

access on two opposing ports to allow visual data recording while other ports are closed with stainless steel plugs providing options for mounting instruments and lines for fuel and gases. The design and operation of the chamber can be studied in detail from the following publications [41,42].For the experiment, the chamber is vacuumed and then filled with the gases required for the specific experiment case. Fuel injection is handled by the fuel system consisting of a low pressure pump, a high pressure pump, a fuel rail and a piezoelectric outwardly opening hollow-cone fuel injector with all its control circuits. The whole system is controlled by LabVIEW on a PID control loop, further details on the system are available on following publications [43,44].

The injection pressure was set to 8 MPa with an injection duration of 1 ms maintained by an external pulse generator. The temperature of the chamber body, fuel injector body and gasoline common rail were all controlled at the set point with closed-loop controlled heaters to maintain a uniform temperature to eliminate any possibility of introducing uncertainties. Four different fuels were used in this study, including two pure substances: isooctane and ethanol, and two mixtures: commercial gasoline and 50/50 mixture of isooctane and ethanol by volume. The injected mass is 34.13 mg, 33.32 mg, 34.07 mg, 31.64 mg for isooctane, ethanol, 50/50 mixture and commercial gasoline, respectively. Table 1 lists the experimental setup and fuel

6

Fuel 257 (2019) 116080

L. Wang, et al.

100°C

Axial penetration (mm)

100

100kPa

90

30kPa

80

10kPa 7kPa

70

5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Time after trigger signa l(ms)

125°C

100 90

Axial penetration(mm)

80 70 60 50

100kPa

40

30kPa 10kPa

30

7kPa

20

5kPa 3kPa

10 0

1kPa 0

0.2

0.4

0.6

0.8

1

1.2

1.4

Time after trigger signal (ms) Fig. 5. (continued)

conditions. For these experiments a Schlieren imaging system was used to visualize the flash boiling of fuels. A simple ‘Z’-type Schlieren setup was used to capture data as shwon in Fig. 1. A Phantom V4.1 grayscale high framerate camera with a 50 mm lens was used as the image capture device. The lens settings were set to an F/# of 5.6 and the camera was set to have an exposure of 2 μs with a frame rate of 8113 fps at 240 × 240 pixel resolution. At this resolution 1 mm resolves into2.4 pixels. The camera receives a trigger simultaneously with the fuel injector and captures a total of 50 frames (4 pre-trigger and 46 post).

Later, Matlab was used to process the captured image data for frame by frame analysis. Digital filters and noise reduction were applied on the images and image quality was enhanced by changing contrast and setting backgrounds to a fixed value. Once the images have been modified an edge detection algorithm was used to resolve spray penetration and the averaged value of the spray penetration front was used to define the spray axial penetration length. Table 1 shows all the experimental settings and conditions, properties of the used fuels used, and the corresponding vapor pressure value of pure isooctane and ethanol. There are six different

7

Fuel 257 (2019) 116080

L. Wang, et al.

25°C

100 100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1

1.2

1.4

Time after trigger signal (ms)

50°C

100 100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

Time after trigger signal (ms) Fig. 6. Spray axial penetration development with time for mixture 50/50.

temperatures used: 25 °C, 50 °C, 75 °C, 90 °C, 100 °C and 125 °C. Seven different ambient pressures were used: 1 kPa, 3 kPa, 5 kPa, 7 kPa, 10 kPa, 30 kPa and 100 kPa. Vapor pressure curves of isooctane and ethanol under different temperature are also shown in Fig. 2. The vapor pressure curve of a pure substance denotes the saturation pressure (Ps) at the corresponding temperature, or the boiling point temperature (Tb) at the corresponding pressure. The equation used to calculate the vapor pressure of a pure substance is the Antoine equation:

P = 10 A

B C+T

or

logP = A

B C+T

(2)

The parameters for isooctane and ethanol in the Antoine equation are extracted from the NIST Chemistry WebBook, SRD69 [45] and Dortmund Data Bank [46]. The values used are shown in Table 2. With the vapor data, we can calculate both the superheated degree (SD) and the ambient-to-saturation pressure ratio. Superheated degree is defined as the difference between liquid temperature and boiling point (Tf-Tb)

(1)

8

Fuel 257 (2019) 116080

L. Wang, et al.

75°C

100 100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1

1.2

1.4

Time after trigger signal (ms)

90°C

100 100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

Time after trigger signal (ms) Fig. 6. (continued)

for a given pressure. It is an intuitive expression of superheating and widely used to indicate the superheat state of a liquid. However, this parameter is dimensional. Another way to indicate the superheated state would be to use the ambient-to-saturation pressure ratio (Pa/Ps), which is dimensionless. Tables 3 and 4 shows the Pa/Ps values of all experimental conditions for isooctane and ethanol, respectively. For a state above the boiling point (super-heated), Pa/Ps < 1. For a state below the boiling point (subcooled), Pa/Ps > 1. The non-dimensional ambient-to-saturation pressure ratio values can facilitate comparison of the experimental results under different pressure/temperature combinations, as shown in Fig. 3.

Although the two definitions (Pa/Ps and SD) are closely correlated, their relationship is not linear. More discussion can be found in an earlier paper [47]. In this paper all the Pa/Ps axes will be shown in a logarithmic scale in order to better illustrate the effects of different superheated conditions. 3. Results and discussions This section analyzes and presents the effect of different superheated conditions on the spray development and other macroscopic features over different temperature (25–125 °C) and ambient pressures

9

Fuel 257 (2019) 116080

L. Wang, et al.

100°C

100

100kPa

90

30kPa 10kPa

Axial penetration (mm)

80

7kPa

70

5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1

1.2

1.4

Time after trigger signal (ms)

125°C

100

100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

Time after trigger signal (ms) Fig. 6. (continued)

(1–100 kPa). The difference of the spray among different fuels (both pure substance and mixture) are also discussed.

condition (room condition), flash boiling does not occur and the spray develops as the originally designed hollow cone shape. Some minor evaporation may occur at the spray edges but has no obvious effect on the spray shape development. Under high superheated conditions, the flash boiling of the fuel causes the liquid sheet along the spray cone to expand with increased sheet thickness. The expansion of the sheet towards the cone center of an axisymmetric spray gradually converges together, forming a plume front which is being pushed out (row 3 and row 5 of spray development pictures in Fig. 4). This phenomenon of appearing central plume can be used as an indicator to identify flash

3.1. Spray images Fig. 4 shows the selected images of typical ethanol spray development under certain combinations of temperature and ambient pressure: 25 °C 100 kPa, 25 °C 10 kPa, 75 °C 10 kPa, 125 °C 100 kPa, and 125 °C 30 kPa. Under STP (25 °C, 100 kPa) the spray maintains its original hollow cone structure with time. In other words, under low superheat

10

Fuel 257 (2019) 116080

L. Wang, et al.

25°C

100 100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1

1.2

1.4

Time after trigger signal (ms)

50°C

100 100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

Time after trigger signal (ms) Fig. 7. Spray axial penetration development with time for commercial gasoline.

boiling from images. More discussions about this phenomenon can be found in Ref. [47]. With an increasing superheated degree, this flash boiling and plume front phenomenon becomes more prominent. When the superheated degree is further increased (Pa/Ps < 0.1), flare flash boiling happens (very strong inward and outward expansion and boiling, spray plume is nearly as wide as the original spray). The spray plume also develops much faster in this case. It is also observed that under higher flash boiling conditions, some liquid region in the spray does not appear as dark as that under lower superheated degree conditions due to quick fuel vaporization compared to a non-flash boiling spray. Some regions have quite high pixel value (8-bit grey scale pixel

value) comparing to the background gray scale pixel value (here the background pixel value is kept at 140 out of 0–255). 3.2. Spray penetration development The spray axial penetration with time was studied between different fuels to compare spry development. The spray axial penetration developments with time under different conditions are shown in Figs. 5–7 for ethanol, mixture 50/50 and commercial gasoline, respectively. The penetration development for pure isooctane has been discussed in an earlier publication [47]. All the cases with the same temperature and

11

Fuel 257 (2019) 116080

L. Wang, et al.

75°C

100 100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1

1.2

1.4

Time after trigger signal (ms)

90°C

100 100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

Time after trigger signal (ms) Fig. 7. (continued)

different ambient pressures are grouped in the same graph. For ethanol, it is seen that under 25 °C, all the penetration lengths are very close to each other, only the penetration of the 100 kPa cases gradually lag behind. As the temperature increases, the majority of penetration development under different pressure cases gradually become longer. For the 100 kPa cases, the axial penetration development does not change much with different temperatures. For 30 kPa cases, the early stage spray axial penetration development does not start increasing until the temperature reaches 100 °C, and under 125 °C the axial penetration development of the 30 kPa case becomes much faster than the 100 kPa case.

For mixture 50/50, the spray axial penetration development is similar, except that under 25 °C, the penetration of 1 kPa case is already faster than the other cases. We can also notice that the spray plume has already started to form under this case which indicates that flash boiling starts to occur for this condition. This shows a good averaged result of isooctane (flash boiling spray) and ethanol (non-flash boiling spray) at 25 °C and 1 kPa. On the other hand, the spray axial penetration development for commercial gasoline shows quite a different trend. It can be seen from Fig. 7 that even at 25 °C, flash boiling occurs for the majority of the cases under different ambient pressures. As the temperature increases, all the lines are moving upward except for

12

Fuel 257 (2019) 116080

L. Wang, et al.

100°C

Axial penetration (mm)

100

100kPa

90

30kPa

80

10kPa 7kPa

70

5kPa

60

3kPa 1kPa

50 40 30 20 10 0

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1

1.2

1.4

Time after trigger signal (ms)

125°C

100

100kPa

Axial penetration (mm)

90

30kPa

80

10kPa

70

7kPa 5kPa

60

3kPa

50

1kPa

40 30 20 10 0

0

0.2

0.4

0.6

0.8

Time after trigger signal (ms) Fig. 7. (continued)

100 kPa case. For the 30 kPa case, strong flash boiling has already occurred when the temperature is at 90 °C, which is much lower than the transition temperature for other fuels. This indicates that it is much easier for commercial gasoline to experience flash boiling.

conditions using all four fuels. Fig. 8 shows the spray images at the same frame for isooctane. Cases on each row have the same temperature, and the temperature increases from top to bottom. Cases on each column have the same ambient pressure, and the ambient pressure increases from left to right. The superheated degree increases from right to left and from top to bottom. Thus the bottom left corner case is the most superheated case. It can be seen that with the increase of superheated degree, the spray plume front gradually appears and is being pushed out from the inside of the hollow cone. It is also noted that under 25 °C and 1 kPa condition, the spray plume has already appeared

3.3. Spray images at the same time frame In order to better compare the spray development under different conditions, spray images at the same frame with a similar time after the start of fuel injection are selected and compared for all the fuel injection

13

Fuel 257 (2019) 116080

L. Wang, et al.

Fig. 8. Isooctane spray images at the same time frame under all conditions.

and the spray axial penetration has been very long. Under high temperature and medium ambient pressure condition, the boiling and subsequent expansion of the spray cone on the leading edge of the spray cone can also be noticed. For ethanol, the spray development at the same time frame is shown in Fig. 9. The spray appears darker and the original cone shape is more visible for most cases than isooctane. There is also no spray plume observed under the 25 °C and 1 kPa condition, which is different from isooctane. On the other hand, the plume front generally appears denser and exhibits relatively lower vaporization caused by boiling along the spray cone than isooctane when comparing with Fig. 8. Fig. 10 shows the spray development at the same time frame for mixture 50/50 when

compared to Figs. 8 and 9 that show spray images for isooctane and ethanol. For mixture 50/50, spray plume appears under the 25 °C and 1 kPa condition, but is noticeably weaker than isooctane. For most of cases, the spray appears not as dark as ethanol. The boiling is not as strong as that of isooctane. Generally speaking, mixture 50/50 shows a good averaged result of isooctane and ethanol. Fig. 11 shows the spray development at the same time frame for commercial gasoline. It can be seen that even under 25 °C and 1 kPa, the flash boiling is already very strong, and spray plumes are seen for the 3 kPa, 5 kPa and 7 kPa cases. The spray development for commercial gasoline is much faster than the other three fuels under these conditions for the earlier spray penetration. The original cone shape is highly disturbed in most high

14

Fuel 257 (2019) 116080

L. Wang, et al.

Fig. 9. Ethanol spray images at the same time frame under all conditions.

superheated degree cases (cases near bottom left corner). Another thing to notice is that the plume width and the plume width to spray cone width ratio are obviously higher than the other fuels.

which is another important factor affecting flash boiling. Heat of vaporization is the amount of energy (enthalpy) that must be added to a liquid phase substance to transform a quantity of that substance into a gas phase. Fig. 12 shows the heat of vaporization for the four fuels under the different experiment temperatures. It can be seen that the heat of vaporization for ethanol is much higher than the other three fuels. The heat of vaporization value for isooctane and gasoline are close, and mixture 50/50 is taken as an average between isooctane and ethanol as suggested by previous works done on mixtures of ethanol and isooctane [48,49]. From their study involving testing multiple mixtures fractions under multiple conditions, they suggest that the HOV of an isooctane and ethanol mixture is relatively linear with increase in

3.4. Heat of vaporization discussion It is known that the boiling point of ethanol is lower than isooctane for most of the experimented conditions. However, it can be seen from the discussion above that the flash boiling phenomenon is stronger and easier for isooctane. Moreover, the ethanol spray appears darker and has less observed boiling than isooctane. In order to explain this, the effects of heat of vaporization for the tested fuels should be considered,

15

Fuel 257 (2019) 116080

L. Wang, et al.

Fig. 10. Mixture 50/50 spray images at the same time frame under all conditions.

ethanol concentration in the mixture. One of the main reasons stated in their work suggests that this phenomenon can be attributed to the similar densities of ethanol and isooctane. Imagine a single droplet experiencing the phase transition process. The temperature is kept at 125 °C, and the droplet goes from 8000 kPa to 1 kPa ambient pressure. As shown in Fig. 13, ideally the droplet moves along the same temperature line until it reaches the final state. Above the vapor pressure curve is the compressed liquid region, and below the vapor pressure curve is the superheated gas region. But in reality, when the droplet hits the vapor pressure curve, part of the droplet turns into vapor first and absorbs heat, thus lowering the temperature of the remaining liquid. The vapor/liquid mixture droplet

will move along the vapor pressure curve a little bit and then leave the vapor pressure curve when all liquid turns into vapor. During this fast process, because the ambient air density is very low, heat conduction to the droplet from the ambient environment can be neglected. In the end the vapor is gradually heated back to 125 °C. In this study, the initial flash boiling causes fuel vaporization and reduces the temperature of surrounding liquid. This has more effects on ethanol because of its much higher heat of vaporization than the other fuels. This explains why there is even less boiling observed in the spray images of ethanol, and why the ethanol spray appears darker and denser: when the liquid passes through the vapor pressure curve, it cools down the remaining liquid by a larger amount, thus the remaining

16

Fuel 257 (2019) 116080

L. Wang, et al.

Fig. 11. Commercial gasoline spray images at the same time frame under all conditions.

Heat of vaporization (kJ/kg)

Isooctane 1000 900 800 700 600 500 400 300 200 100 0

Ethanol

Mixture

liquid can continue moving without experiencing more boiling and phase change.

Gasoline

3.5. Comparison between different fuels

0

20

40

60

80

100

120

In order to compare among different fuels, some key parameters of spray development are used to facilitate the comparison: the peak penetration velocity, spray penetration length at the same time frame, and the plume ratio at the same time frame. To compare the spray plume front condition for different conditions, spray plume ratio is used in this study. The spray plume radius is not selected as the key parameter here, because plume radius value is usually dependent on the spray cone radius and comparing the absolute plume radius alone among different conditions cannot offer much insights. The plume radius to spray cone

140

Temperature (°C) Fig. 12. Heat of vaporization for the four fuels.

17

Fuel 257 (2019) 116080

L. Wang, et al.

10000

Compressed liquid

Supercritical

Pressure (kPa)

1000

100

10

Superheated gas 1

0

50

100

150

200

250

300

350

400

Temperature (°C)

(a) 10000

Supercritical

Compressed liquid Pressure (kPa)

1000

100

10

Superheated gas 1

0

50

100

150

200

250

300

350

400

Temperature (°C)

(b) Fig. 13. State change of one single droplet during flash boiling: (a) ideal (b) reality.

radius ratio, as a non-dimensional value, can be a good indicator of the extent of flash boiling, and here we are using the plume ratio value of each experiment condition at the same time frame with the spray axial penetration discussion in the last part. A schematic of the spray plume front, as well as the plume radius and spray cone radius can be found in Ref. [47]. The comparisons of the key spray parameters are shown in Fig. 14–16. A color map plot is used in order to better show the results while displaying the experiment conditions, so that all the ambient pressure and temperature conditions can be shown on the same graph. The horizontal axis in each graph corresponds to the temperature. The

vertical axis in each graph is in log scale and corresponds to the ambient pressure. The color-map corresponds to the key parameters being compared. Fig. 14 shows the peak penetration velocity comparison among different fuels under all the experiment conditions. There are three corners to be highlighted: low temperature & low ambient pressure condition (bottom left corner); high temperature & high ambient pressure (top right corner); high temperature & low ambient pressure, which is the high superheated degree region (bottom right corner). It can be seen that generally ethanol has the highest peak penetration velocity under most conditions, while having the lowest value at low

18

Fuel 257 (2019) 116080

L. Wang, et al.

Fig. 14. Peak penetration velocity comparison among different fuels.

temperature and low ambient pressure condition (bottom left corner). Isooctane and gasoline have higher value at the bottom left corner. On the other hand, gasoline shows complex distribution of peak penetration velocity: a higher value region appears at mid-temperature, low pressure region. Fig. 15 shows the penetration at a certain time frame for all four fuels. The penetration result shows similar characteristics to the peak velocity. Again, ethanol has the longest penetration length, while having the lowest value at the low temperature and low ambient pressure condition (bottom left corner). This corresponds to the earlier discussion that ethanol does not experience flash boiling and spray plume does not appear at 25 °C and 1 kPa. However, due to higher heat of vaporization value for ethanol, under higher SD conditions, the initial flash boiling of ethanol spray causes fuel vaporization and cools down the remaining fluid by a larger extent compared to the other fuels. This reduces further boiling of the spray plume front, and enables the ethanol spray to have a faster and longer penetration than the other three fuels under flash boiling conditions. For isooctane, under the same flash boiling conditions, the spray plume front is still experiencing

boiling after it is being pushed out, and this continued boiling phenomenon slows down the spray plume penetration slightly. Overall isooctane has the smallest penetration value, while mixture 50/50 shows as a good average value between isooctane and ethanol. Gasoline has the largest penetration length at low temperature & low pressure region (bottom left corner), which corresponds to the earlier discussion that gasoline is already experiencing strong flash boiling and the original spray cone shape has already changed under the 25 °C and 1 kPa condition. Fig. 16 shows the spray plume ratio comparison at the same time frame for different fuels. Overall, for the plume ratio distribution, ethanol has the smallest region with a non-zero plume ratio, which means that ethanol has the fewest cases showing the plume. Ethanol also has the steepest gradient (changing from min to max) when the superheated degree increases (from top left to bottom right). On the contrary, gasoline has the widest distribution and smallest gradient. This indicates that it is easier for gasoline to experience flash boiling and generate a spray plume front under low superheated conditions. Another thing to note is the spray plume ratio value at low temperature

19

Fuel 257 (2019) 116080

L. Wang, et al.

Fig. 15. Spray axial penetration (at the same time frame) comparison among different fuels.

and low pressure region (bottom left corner): ethanol < mixture 50/ 50 < isooctane < gasoline. Again the mixture 50/50 shows a good average result between those of ethanol and isooctane.

multicomponent mixture fuels, instead of one single vapor pressure curve, there are two curves: a bubble point curve and a dew point curve. The bubble point curve separates the compressed liquid region and the region where bubbles starts to appear. The bubble point curve is mostly affected by the vapor pressure curve of the lowest boiling point component in the mixture. The dew point curve separates the region where liquid still exists and the superheated gas region. The dew point curve is mostly determined by the highest boiling point component in the mixture. Between the bubble point curve and the dew point curve is the two phase region. In the two phase region, the vapor of higher boiling point liquid fuel coexists with lower boiling point fuel vapor. In the two-phase region, illustrated in Fig. 17, the vapor of the lower boiling point component dominates, with the vapor of the higher boiling point component coexisting. The vapor of the higher boiling point component would not be present under the same conditions if it was the only pure substance used in the system, because this region is above the higher boiling point component’s saturated vapor pressure line. When using multicomponent fuels, in order to make initial flash

3.6. Pure substance and multicomponent fuels As discussed earlier, pure substances (isooctane and ethanol) show different flash boiling characteristics under the same ambient conditions, and mixture 50/50 shows good average results between those of isooctane and ethanol. Commercial gasoline, as a multicomponent mixture, experience flash boiling spray much earlier than other fuels. The low boiling point component in the gasoline mixture can play an important role here in this flash boiling spray. Fig. 17 shows the flash boiling process for both pure substances (a) and multicomponent mixtures (b). For the pure substances, above the vapor pressure curve is the compressed liquid region, and below the vapor pressure curve is the superheated gas region. There is no phase change until the liquid reaches the vapor pressure curve, where there are bubbles formed and flash boiling starts to take place. For the

20

Fuel 257 (2019) 116080

L. Wang, et al.

Fig. 16. Spray plume ratio (at the same time frame) comparison among different fuels.

boiling occur, the mixture liquid only needs to hit the bubble point curve. In most cases, the bubble point curve of a multicomponent mixture fuel is located higher (higher saturation pressure at the same temperature) than that of a pure substance in the plot. So, it is easier for multicomponent fuels to experience flash boiling, due to their low boiling point components. Adding lower boiling point additive/fuel into higher boiling point fuel is expected to facilitate flash boiling, enhancing the atomization and mixing. This will not only make the flash boiling occur easier, but also create some vapor of the higher boiling point component, which again could facilitate the mixing and combustion. This also implies that blending a low boiling component fuel with a high boiling component fuel may lead to an increase in fuel evaporation and hence multicomponent fuels, such as gasoline, are more susceptible to flash boiling than single component fuels.

1. The spray plume generated from the mergence of inward expanding hollow cone can be a good indicator of flash boiling phenomenon, the ratio of the plume radius to the spray cone radius can be a good non-dimensional indicator to quantify flash boiling. 2. Pure substances have a single vapor pressure curve, while mixture does not have a single boiling point at a given pressure. Thus it is not able to use Pa/Ps and Tf-Tb to analyze mixture 50/50 and gasoline. When superheated degree is smaller, the spray penetration is mostly determined by the initial spray momentum, and the effect of flash boiling is minor. When superheated degree is larger, the peak penetration velocity increases with the increase of superheated degree, until it reaches around 200 m/s. In this range, the flash boiling and spray plume front becomes the dominant factor affecting the spray penetration velocity. 3. Under the same conditions, ethanol has higher superheated degree compared to isooctane, but the heat of vaporization for ethanol is much higher. This contributes to the fact that less boiling is observed in ethanol spray as seen in the video, while ethanol penetration is longer in serval cases. Mixture 50/50 shows a good average of isooctane and ethanol for both penetration and plume ratio.

4. Summary In this paper, the effect of superheated degree and flash boiling on the spray shape and penetration with different fuels and mixtures were studied with a GDI outwardly opening hollow-cone injector, the following is a summary of the findings:

21

Fuel 257 (2019) 116080

L. Wang, et al.

Fig. 17. Flash boiling process for different liquid: (a) pure substance, (b) multicomponent.

4. Gasoline, due to its low initial boiling point and wide range of components, has the widest plume ratio distribution and smallest gradient, as well as complex peak penetration velocity distribution. 5. Adding low boiling point (preferably low heat of vaporization as well) additive or component to high boiling point fuel could facilitate flash boiling, fuel vaporization and mixing.

iso-octane liquid jets. J Fluid Mech 2007;572:121–44. [4] George A, Tavlarides LL, Geana Dan. Phase transitions and thermal behavior of fueldiluent mixtures. Energy Fuels 2009;23(6):3068–77. [5] Mojtabi M, Chadwick N, Wigley G, Helie J, (2008) The effect of flash boiling on break up and atomization in gdi sprays. ICLASS. [6] Plesset MS, Prosperetti A. Bubble dynamics and cavitation. Annu Rev Fluid Mech 1997;9:145–85. [7] Polanco Geanette, Holdø Arne Erik, Munday George. General review of flashing jet studies. J Hazard Mater 2010;173(1):2–18. [8] Simões-Moreira José R. Oblique evaporation waves. Shock Waves 2000;10(4):229–34. [9] Brown Ralph, Louis York J. Sprays formed by flashing liquid jets. AIChE J 1962;8(2):149–53. [10] Lienhard John H, Day James B. The breakup of superheated liquid jets. J Basic Eng 1970;92(3):515–21. [11] Steelant J, High-Speed Shadowgraphy Investigations of Superheated Liquid Jet Atomisation H. Kamoun, G. Lamanna, B. Weigand Institute of Aerospace Thermodynamics Universität Stuttgart, 70,569 Stuttgart Germany. [12] Guo H, Ma X, Li Y, Liang S, Wang Z, Xu H, et al. Effect of flash boiling on microscopic and macroscopic spray characteristics in optical GDI engine. Fuel 2017;190:79–89. [13] Williams B, Ewart P, Wang X, Stone R, Ma H, Walmsley H, et al. Quantitative planar laser-induced fluorescence imaging of multi-component fuel/air mixing in a firing gasoline-direct-injection engine: effects of residual exhaust gas on quantitative PLIF. Combust Flame 2010;157(10):1866–78. [14] Zhang Z, Huang Z, Wang X, Zheng J, Miao H, Wang X. Combustion characteristics of methanol–air and methanol–air–diluent premixed mixtures at elevated temperatures and pressures. Appl Therm Eng 2009;29(13):2680–8.

Acknowledgements This research was supported in part by the Saudi Aramco Public R& D Center through the Clean Combustion Research Center of the King Abdullah University of Science and Technology. References [1] Kawano D, Ishii H, Suzuki H, Goto Y, Odaka M, Senda J, 2006. Numerical study on flash boiling spray of multicomponent fuel. Heat Transfer—Asian Research: Cosponsored by the Society of Chemical Engineers of Japan and the Heat Transfer Division of ASME, 35(5), pp. 369–385. [2] Cleary Vincent, Bowen Phil, Witlox Henk. Flashing liquid jets and two-phase droplet dispersion: I. Experiments for derivation of droplet atomisation correlations. J Hazard Mater 2007;142(3):786–96. [3] Vieira Marcelo Mendes, Simoes-Moreira JR. Low-pressure flashing mechanisms in

22

Fuel 257 (2019) 116080

L. Wang, et al. [15] Chan QN, Bao Y, Kook S. Effects of injection pressure on the structural transformation of flash-boiling sprays of gasoline and ethanol in a spark-ignition directinjection (SIDI) engine. Fuel 2014;130:228–40. [16] Lee J, Madabhushi R, Fotache C, Gopalakrishnan S, Schmidt D. Flashing flow of superheated jet fuel. Proc Combust Inst 2009;32(2):3215–22. [17] Weber D, Leick P, 2014. Structure and velocity field of individual plumes of flashing gasoline direct injection sprays. In 26th Annual Conference on Liquid Atomization and Spray Systems (ILASS), Bremen, Germany, Sept (pp. 8–10). [18] Allocca L, Montanaro A, Di Gioia R, Bonandrini G, 2014. Spray characterization of a single-hole gasoline injector under flash boiling conditions (No. 2014-32-0041). SAE Technical Paper. [19] Montanaro A, Allocca L, 2015. Flash boiling evidences of a multi-hole GDI spray under engine conditions by Mie-scattering measurements (No. 2015-01-1945). SAE Technical Paper. [20] Wood A, Wigley G, Helie J, “Flash Boiling Sprays produced by a 6-hole GDI Injector”. In: 17th International Symposium on Applications of Laser Techniques to Fluid Mechanics, July 2014. [21] Wu S, Xu M, Hung DL, Li T, Pan H. Near-nozzle spray and spray collapse characteristics of spark-ignition direct-injection fuel injectors under sub-cooled and superheated conditions. Fuel 2016;183:322–34. [22] Huang Y, Huang S, Huang R, Hong G. Spray and evaporation characteristics of ethanol and gasoline direct injection in non-evaporating, transition and flashboiling conditions. Energy Convers Manage 2016;108:68–77. [23] Guo H, Ding H, Li Y, Ma X, Wang Z, Xu H, et al. Comparison of spray collapses at elevated ambient pressure and flash boiling conditions using multi-hole gasoline direct injector. Fuel 2017;199:125–34. [24] Bornschlegel S, Conrad C, Eichhorn L, Wensing M, 2017. Flashboiling atomization in nozzles for GDI engines. In Ilass Europe. 28th European conference on Liquid Atomization and Spray Systems (pp. 321–328). Editorial Universität Politècnica de València. [25] Lacey J, Poursadegh F, Brear MJ, Gordon R, Petersen P, Lakey C, et al. Generalizing the behavior of flash-boiling, plume interaction and spray collapse for multi-hole, direct injection. Fuel 2017;200:345–56. [26] Montanaro A, Allocca L, Lazzaro M, 2017. Iso-octane spray from a GDI multi-hole injector under non-and flash boiling conditions (No. 2017-01-2319). SAE Technical Paper. [27] Li T, Dong X, Hung DL, Li X, Xu M. Analysis of evaporation characteristics and heat transfer for flash-boiling sprays. Int J Heat Mass Transf 2018;127:244–54. [28] Li Y, Guo H, Ma X, Qi Y, Wang Z, Xu H, et al. Morphology analysis on multi-jet flashboiling sprays under wide ambient pressures. Fuel 2018;211:38–47. [29] Yang S, Li X, Hung DL, Xu M. Characteristics and correlation of nozzle internal flow and jet breakup under flash boiling conditions. Int J Heat Mass Transf 2018;127:959–69. [30] Wu S, Xu M, Yang S, Yin P, 2018. Contrary Effects of Nozzle Length on Spray Primary Breakup under Subcooled and Superheated Conditions (No. 2018-010302). SAE Technical Paper. [31] Guo H, Li Y, Ma X, Xu H, Wang Z, 2018 Experiment Study on Radial Expansion of Flashing Jet from Single-Hole GDI Injector. ICLASS 2018, 14th Triennial International Conference on Liquid Atomization and Spray Systems, Chicago, IL, USA, July 22–26, 2018. [32] Dong X, Yang J, Hung DL, Li X, Xu M. Effects of flash boiling injection on in-cylinder spray, mixing and combustion of a spark-ignition direct-injection engine. Proc

Combust Inst 2019;37(4):4921–8. [33] Yang J, Dong X, Wu Q, Xu M. Influence of flash boiling spray on the combustion characteristics of a spark-ignition direct-injection optical engine under cold start. Combust Flame 2018;188:66–76. [34] Wu S, Yang S, Wooldridge M, Xu M. Experimental study of the spray collapse process of multi-hole gasoline fuel injection at flash boiling conditions. Fuel 2019;242:109–23. [35] Jiang C, Parker MC, Helie J, Spencer A, Garner CP, Wigley G. Impact of gasoline direct injection fuel injector hole geometry on spray characteristics under flash boiling and ambient conditions. Fuel 2019;241:71–82. [36] Gao J, Jiang D, Huang Z. Spray properties of alternative fuels: a comparative analysis of ethanol-gasoline blends and gasoline. Fuel 2007;86(10–11):1645–50. [37] Aleiferis PG, Malcolm JS, Todd AR, Cairns A, Hoffmann H, 2008. An optical study of spray development and combustion of ethanol, iso-octane and gasoline blends in a DISI engine (No. 2008-01-0073). SAE Technical Paper. [38] Andersson M, Warnberg J, Hemdal S, Dahlander P, Denbratt I, 2011. Evaporation of Gasoline-Like and Ethanol-Based Fuels in Hollow-Cone Sprays Investigated by Planar Laser-Induced Fluorescence and Mie Scattering (No. 2011-01-1889). SAE Technical Paper. [39] Chen R, Nishida K. Spray evaporation of ethanol–gasoline-like blend and combustion of ethanol-gasoline blend injected by hole-type nozzle for direct-injection spark ignition engines. Fuel 2014;134:263–73. [40] Park SH, Kim HJ, Suh HK, Lee CS. Atomization and spray characteristics of bioethanol and bioethanol blended gasoline fuel injected through a direct injection gasoline injector. Int J Heat Fluid Flow 2009;30(6):1183–92. [41] Jing W, Wu Z, Roberts WL, Fang T. Spray combustion of biomass-based renewable diesel fuel using multiple injection strategy in a constant volume combustion chamber. Fuel 2016;181:718–28. [42] Zhang J, Jing W, Roberts WL, Fang T. Soot measurements for diesel and biodiesel spray combustion under high temperature highly diluted ambient conditions. Fuel 2014;135:340–51. [43] Wang L, Badra J, Roberts W, Fang T. Characteristics of spray from a GDI fuel injector for naphtha and surrogate fuels. Fuel 2017;190:113–28. [44] Wu Z, Wang L, Badra J, Roberts W, Fang T. GDI fuel sprays of light naphtha, PRF95 and gasoline using a piezoelectric injector under different ambient pressures. Fuel 2018;223:294–311. [45] Huber KP, Herzberg G, 1997. NIST Chemistry WebBook, NIST Standard Reference Database Number 69, Eds. PJ Linstrom and WG Mallard, National Institute of Standards and Technology, Gaithersburg MD, 20899. https://webbook.nist.gov/ cgi/cbook.cgi?ID=C540841&Mask=FFF. [46] Dortmund Data Bank. http://www.ddbst.com/ddb.html, http://ddbonline.ddbst. com/AntoineCalculation/AntoineCalculationCGI.exe?component=Ethanol& tunit=%C2%B0C&punit=mmHg&TemperaturesEdit=&calculate=Calculate. [47] Wang L, Wang F, Fang T. Flash boiling hollow cone spray from a GDI Injector under different conditions. Int J Multiph Flow 2019;118:50–63. [48] Chupka GM, Christensen E, Fouts L, Alleman TL, Ratcliff MA, McCormick RL. Heat of vaporization measurements for ethanol blends up to 50 vol% in several hydrocarbon blendstocks and implications for knock in SI engines. SAE Int J Fuels Lubr 2015;8(2):251–63. [49] Chen L, Stone R. Measurement of enthalpies of vaporization of isooctane and ethanol blends and their effects on PM emissions from a GDI engine. Energy Fuels 2011;25(3):1254–9.

23