Fuel 140 (2015) 116–125

Contents lists available at ScienceDirect

Fuel

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

Effect of di-n-butyl ether blending with soybean-biodiesel on spray andatomization characteristics in a common-rail fuel injection system

http://dx.doi.org/10.1016/j.fuel.2014.09.1040016-2361/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding authors. Fax: +86 29 82668789.E-mail addresses: chenglongtang@mail.xjtu.edu.cn (C. Tang),

zhhuang@mail.xjtu.edu.cn (Z. Huang).

Li Guan, Chenglong Tang ⇑, Ke Yang, Jun Mo, Zuohua Huang ⇑State Key Laboratory of Multiphase Flows in Power Engineering, Xi’an Jiaotong University, Xi’an 710049, People’s Republic of China

h i g h l i g h t s

�We examine the effect of DBE addition on the spray characteristic recover.� We investigate the macroscopic and microscopic spray characteristics of biodiesel, DBE/biodiesel.� DBE addition significantly improves the spray characteristics due to its lower viscosity and surface tension.

a r t i c l e i n f o

Article history:Received 15 June 2014Received in revised form 18 September 2014Accepted 25 September 2014Available online 7 October 2014

Keywords:BiodieselDi-n-butyl etherSpray characteristicsDroplets size distribution

a b s t r a c t

In this work, the spray and atomization characteristics of soybean biodiesel, di-n-butyl ether (DBE)/bio-diesel blends and 0# diesel were investigated by using a high pressure common-rail injection system. Themacroscopic spray characteristics such as the spray tip penetration (STP), the cone angle and the sprayprojected area were obtained from the spray images captured through high speed schlieren photography.The results show that as DBE is blended into biodiesel, the STP is decreased and the spray cone angle, pro-jected area are increased and when the DBE volume fraction in the DBE/biodiesel blends reaches 30%, theSTP, the spray cone angle and the projected area are comparable to that of diesel. For all the tested fuels,the microscopic spray characteristics such as the Sauter Mean Diameter (SMD) and statistical size distri-butions were measured by particle/droplet image analysis (PDIA) technique. The droplets number den-sity distribution shows that for all the tested fuels, the region near the central of spray has the largestdroplets number density and it decreases sharply as the position shifts from the center to the edge ofthe spray. As DBE is blended into biodiesel, smaller SMD is observed, which indicates that DBE additioncan promote the atomization of biodiesel.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction to those of fossil diesel, it can be directly used in the traditional

Due to its very high compression ratio, the diesel engines havehigher thermal efficiency when compared to regular petrol enginesand are expected to reduce fuel consumption. Additionally, dieselengines typically have higher torque output and thus have beenwidely used in heavy road vehicles such as trucks, buses and ships.Though fossil fuels are still available for 50–100 years, their limitedreserves are being depleted due to the increasing demands, thusresearch on alternative fuels have been attracting more and moreattention. Biodiesel is the mono-alkyl esters of long chain fattyacids and can be produced from renewable and biodegradableresources. It can be derived from triglycerides by transesterifica-tion with alcohols. Due to its similar physico-chemical properties

diesel engines or with minor engine modifications and it hasattracted great interest of researchers as an alternative fuel for die-sel engines [1–3]. Compared to fossil diesel, biodiesel is renewable,nontoxic and contains no sulfur. The effect of biodiesel fuels onengine emissions was reviewed by Lapuerta et al. [4] in whichbiodiesel was shown to be potentially capable for the reductionof carbon monoxide (CO), carbon oxides (CO2), unburned hydrocar-bons (HC), and particulate matter (PM) emissions. However,biodiesel produces more nitrogen oxide (NOx) because of its higheroxygen content and high bulk modulus, which leads to anincreased combustion temperature and NOx formation [5,6]. Prac-tically, the exhaust gas recirculation (EGR) technique is introducedto reduce the combustion temperature thus to reduce the NOxemissions of the biodiesel fueled engines [7,8].

It is well known that the characteristics of fuel spray and themixture formation are very influential on combustion processand final engine performance and emission characteristics.

L. Guan et al. / Fuel 140 (2015) 116–125 117

Recently, some investigations have been carried out on the biodie-sel spray characteristics. Lee et al. [9] studied the atomization char-acteristics of biodiesel-blended fuels and compared with theconventional diesel. Their results indicates that biodiesel-blendedfuels have similar spray tip penetration (STP) but larger SauterMean Diameter (SMD) compared to the conventional diesel. Wanget al. [10] studied the spray characteristics of two biodiesels anddiesel under high injection pressures. Their study shows that bio-diesel gives longer STP and smaller spray angle, projected areaand volume compared to diesel, and biodiesels give larger SMDthan diesel because of their higher viscosity and surface tension.Gao et al. [11] studied the spray characteristics of biodiesel basedon inedible oil and they found that spray is more concentratedand SMD of biodiesel-blended fuels is larger than that of dieselbecause of the higher viscosity and surface tension of biodiesel.These studies show that because biodiesel has higher surface ten-sion and viscosity, biodiesel spray is relatively more difficult forevaporation and atomization. Additionally, biodiesel has a highpour point and fuel flow problems become significant at low tem-peratures [12,13].

We propose that the addition of less viscous fuels with smallersurface tension into biodiesel is potentially capable to recover thedeteriorated spray characteristics, compared to diesel. Previously,Yoon et al. [14] added up to 20% (by volume) of ethanol into biodie-sel (BE20), and the spray characteristics of fuel blends were studiedwith spray visualization and droplet size analyzing techniques.Their results show that the spray penetrations of BE10, BE20 anddiesel are similar, but the SMD of diesel is larger than those ofBE10 and BE20. Shi et al. [15] studied the effects of biodiesel/eth-anol addition on the emission characteristics of a heavy duty dieselengine and their results show that both PM and total hydrocarbons(THC) emissions are decreased but NOx emissions are slightlyincreased. The amount of unregulated emissions such as acetone,aldehyde and ethanol are also found to be increased. Ethers likedimethyl ether (DME) and diethyl ether (DEE) have shown to bepromising in terms of soot reduction and combustion enhance-ment [16,17]. Another ester of interest is di-n-butyl ether (DBE),which can be produced from lignocellulosic biomass with promis-ing material efficiency [18] through a series of bio-chemical pro-cesses. Additionally, DBE is the so called second generationbiofuel which does not threaten the food supply and biodiversity.Compared to biodiesel, it has lower density, surface tension andviscosity, but has higher cetane number (up to 100) [19–21]. Verylimited studies have been conducted on the use of DBE as alterna-tive fuel. Nabi et al. [22] investigated the effect of DBE blendinginto diesel on emissions and engine noise characteristics in a direct

Common Rail

HighPressure

Pump

Filter

ConvexLens

GratingLightSource

LensGroup

Fuel Tank

ConstaVolume V

Fig. 1a. Experimental setup for STP

injection diesel engine. Their results show that significant reduc-tions in smoke, CO, NOx, and THC are simultaneously achievedwith the increase of DBE blending ratio, and engine noise is alsoreduced because of better ignitibility of DBE. Beeckmann et al.[23] studied the effect of DBE blending (up to 20% by volume) onthe spray characteristics of CEC (Coordinating European Council)reference diesel at the ambient pressure of 50 bar and temperatureof 800 K. They reported that addition of DBE is potentially capableto improve the atomization process and ultimately the engine per-formance. Additionally, the ignition delay times are significantlyreduced as DBE is added into the CEC diesel because of its highcetane number and they stated that blending DBE into diesel mightenhance the premixed charge compression ignition (PCCI) combus-tion process which would result in a significant reduction in bothsoot and NOx.

Since compared to diesel, the higher viscosity of biodiesel dete-riorates the spray characteristics, we attempt to adopt DBE as anadditive into biodiesel and to see if the spray characteristics of thebiodiesel/DBE blends (with certain DBE blending ratio) are compa-rable to that of diesel. As far as it is known, there has been no studyon the spray atomization and fuel–air mixing of DBE/biodieselblends. In the following, the experimental setup and procedures willbe specified in detail. The macroscopic spray characteristics of bio-diesel/DBE blends at different injection pressures and ambient pres-sures in terms of spray tip penetration, spray angle and projectedarea will be examined from the spray images captured through highspeed schlieren photography. Additionally, microscopic and statis-tical spray characteristic parameters such as SMD and the dropletsnumber density distribution of different biodiesel/DBE blends willbe compared.

2. Experimental setup and procedure

2.1. Apparatus and procedure

Fig. 1a shows the sketch of the spray visualization apparatus. Ahigh-pressure, constant volume chamber is designed for the sprayvisualization and is filled with nitrogen gas. The ambient pressureis adjusted by a manual intake and exhaust valves that connects onthe chamber and the maximum pressure is 50 bar. Two quartzwindows with the diameter of 100 mm on two sides provide anoptical access for the spray visualization.

The fuel-injection facility consists of the fuel tank, the highpressure pump, the common rail, the injector and the electroniccontrol unit. A pressure transducer mounted on the common rail

Injector

High SpeedCamera

ECU

KnifeEdge

ConvexLens

PC

nt essel

and cone angle measurements.

Fig. 1b. The schematic of the injector nozzle.

118 L. Guan et al. / Fuel 140 (2015) 116–125

is used to monitor the injection pressure. The rail pressure andinjection rate characteristics including the injection time and pulsewidth, are controlled by an optical diagnostic controller OD2301with closed-loop control. This controller also controls the triggersignal of injector and camera, providing variable synchronizationof spray visualization. The injector used in the present study is aBOSCH second-generation common-rail injector with one single-hole and diameter of 0.25 mm, and details of the injector is por-trayed in Fig. 1b. The energizing time of the injector was set to1.5 ms in this study.

For the spray macroscopic characteristic investigation, theschlieren method was used. A Xenon lamp was used as the lightsource and two magnifying lens were used to produce parallel light

Fig. 1d. The sketch of

Fig. 1c. Spray edge ext

and accumulate the light that pass through the spray respectively.A high-speed camera (Phantom V611) was employed to record thespray images with the imaging speed at 20,000 frames/s and reso-lution of 352 � 640 pixels. The recorded spray images were pro-cessed before the measurement of spray characteristics. At first,the images were changed to gray images and subtracted from aprerecorded background image, in which no spray could be found.Secondly, for one injection, a fixed threshold value of gray was cho-sen from one frame image of the recorded high speed video. Allother images were processed by this threshold value. Thirdly,inverted binary images were processed and were used for thespray projected area measurement. Finally, the detected edge fromspray images were employed for the comparison of the spray evo-lution. Fig. 1c shows the typically processed spray image. It isfound that the spray edge can be extracted, and with these sprayedge images, the macroscopic spray characteristics can bemeasured.

For the microscopic characteristic investigation, the schlierenoptics is replaced by the particle/droplet image analysis (PDIA) sys-tem (as shown in Fig. 1d) through which the droplets size is mea-sured. Previously, Kashdan et al. [24] and Berg et al. [25] comparedthe PDIA technique with the well-established technique phasedoppler anemometry (PDA) and the results show that PDIA is areliable technique to measure the size of spherical and non-spher-ical droplets. An Nd:YAG laser with the wavelength of 532 nm,along with a diffusor attachment, was employed for the homoge-neous illumination. The diameter of the lens at the head of the dif-fusor is 120 mm, which is larger than the window of the chamber,to ensure the homogeneous illumination of the spray flow field. A

the PDIA system.

raction procedure.

Table 1Properties of the fuels.

Fuel type B100 DBE15 DBE30 Diesel DBE

Density (20 �C) (kg/m3) 880.5 863.6 846.7 831.5 772.5Viscosity (20 �C) (mPa s) 9.192 5.796 3.853 4.70 0.691Surface tension (20 �C) (mN/m) 31.2 29.9 25.7 29.3 23.0

L. Guan et al. / Fuel 140 (2015) 116–125 119

CCD camera (ImagerProSX 5M) connected with a long distancemicroscope (Queststar QM1) and a magnifying lens with an ampli-fication factor of 2 was used to capture the droplets. The cameraand microscope were mounted on an electric positioner MC600.The precise displacement control (1 lm) ensures the accuratescanning of the measurement position. The calibration was carriedout via a scaling plate with minimum scale of 25 lm. After calibra-tion, the field of view was 1.78 mm � 1.49 mm and one pixel rep-resents 1.44 lm in the focal plane. The commercial software DaVis8.0.0 of LaVision was used to process the shadow images.

2.2. Tested fuels

In this study, four fuels were tested: the soybean biodiesel(B100), fuel blends DBE15 (15% by volume DBE in DBE-biodieselblends), DBE 30 and 0# diesel. Physical properties of the fuels arelisted in Table 1. The density, viscosity, surface tension were mea-sured according to Chinese national standard GB/T 1884-1992, GB/T 265-1988 and GB/T 6541-1986, respectively. The surface tension,viscosity and density of biodiesel are larger than those of diesel andthese parameters decrease with the increase of the blending ratioof DBE in biodiesel.

3. Experimental results and discussion

In this section, the macro spray characteristics including spraytip penetration (STP), the spray cone angle and the projected areaof B100, DBE15, DBE30 and diesel will be presented. Injection pres-sures Pinj are 600, 800 and 1000 bar and ambient pressures Pamb are10 and 20 bar. The experiment was repeated three times undereach condition to ensure the reliability of the results. Dropletsnumber and size distribution and characteristic diameters wereanalyzed at injection pressure of 600 bar and ambient pressure of10 bar for the biodiesel and two DBE/biodiesel blends. The ambienttemperature is 293 K. The injection duration is 1.5 ms under allconditions.

3.1. Macroscopic spray parameters

Fig. 2 shows the definition of the macroscopic spray character-istics. STP is defined as the vertical distance from the nozzle tip to

Spra

y tip

pen

etra

tion

(STP

)

STP

/ 2

Spray cone angel

Fig. 2. Definitions of STP and the spray cone angle.

the bottom of the spray edge. Because the spray is not rigorouslycone-shaped, the spray angle at half of the STP axial location ismeasured and defined as the spray cone angle. The averaged sprayangle is the arithmetic angle during quasi-steady stage of eachinjection. The projected area is determined from the processed bin-ary image. The pixels represent the spray have a value of ‘1’ and theprojected area is measured by summing all pixels.

Fig. 3 presents the spray evolution process of B100, DBE15,DBE30 and diesel at injection pressure of 1000 bar and ambientpressure of 20 bar. B100 and DBE15 show similar STP but DBE15gives slightly larger spray cone angle compared with biodiesel atthe same elapsed timing. DBE30 and diesel gives the shortest STPand largest spray cone angle. A quantitative analysis of STP, coneangle and projected area were performed to further investigatethe effect of DBE blending ratio on the spray characteristics inthe following.

Fig. 4a–c illustrates the effect of DBE blending ratio on the STPat injection pressures of 600, 800 and 1000 bar and ambient pres-sures of 10 and 20 bar. For all the tested fuels and pressure condi-tions, the STP evolves in a very similar pattern: initially after thestart of injection, the STP increases linearly with time, and aftercertain critical time tbreak, the increasing rate of STP slows down.This is because for t < tbreak, the STP is mainly affected by the fueldensity and all test fuels show the similar density, leading to theSTP evolution insensitive to the DBE addition, however, fort > tbreak, DBE addition tends to decelerate the spray tip evolution.This is because blended fuels have lower viscosity and surface ten-sion and their droplets are easier to deform and breakup, resultingin a decrease of spray droplet size and loss in spray momentumwhich slows eventually down the penetration [26,27]. Further-more, for all the tested fuels at Pinj = 600 bar, tbreak is around3 ms. As the injection pressure increases, tbreak decreases: 2 and1.5 ms at 800 and 1000 bar, respectively. This is because the highinjection pressure promotes the breakup of the primary spraydue to the increased hydrodynamic instability. This is consistentwith the study of Hiroyasu et al. [26] and they proposed the empir-ical equation of tbreak that is inversely proportional to the differ-ence between the injection and ambient pressure. For given fuel,the STP at fixed time after start of injection is increased with higherPinj because of increased velocity of the spray tip. With the increaseof ambient pressure, the STP is minimally affected for t < tbreak,while for t > tbreak, it significantly reduces at the late stage sprayevolution due to the increased loss of momentum at higher ambi-ent pressure. The effect of ambient pressure is negligible becausein our study, the ambient pressures were 10–20 bar and the injec-tion pressure is 600–1000 bar, respectively. Previous work by Hiro-yasu and Arai [26], shows that tbreak is inversely proportional to thepressure difference between the injection and the ambientpressure.

Fig. 5 presents the spray cone angle of B100, DBE15, DBE30 anddiesel at injection pressure of 1000 bar and ambient pressure of20 bar. All the four fuels show stable spray cone angle after the ini-tial stage of the spray. DBE30 and diesel gives slightly larger spraycone angles than those of B100 and DBE15, and B100 and DBE15show similar spray cone angle. Fig. 6 gives the averaged spray coneangle for the test fuels at various injection pressures and ambientpressures. With the increase of DBE blending ratio, the averagedcone angle is slightly increased. This may be because of the low vis-cosity of DBE which decreases the overall viscosity of the blends,leading to the increase of the spray cone angle. This is consistentwith the study of Valentino et al. [28] that lower viscosity of fuelsbrought larger cone angles. It was also found that the injectionpressure does not strongly influence the averaged cone angle,while the increase in the ambient pressure results in an obviousincrease in spray angle. As discussed by Payri et al. [29], the injec-tion pressures as high as 800 bar had little effect on spray angle

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.80

20

40

60

80(a) Pinj=600bar

Spra

y Ti

p Pe

netr

atio

n / m

m

Time After Start Of Injection / ms

Pamb=10bar

Pamb=20bar

B100DBE15DBE30Diesel

tbreak

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60

20

40

60

80

tbreak

(b) Pinj=800bar

Spra

y Ti

p Pe

netr

atio

n / m

m

Time After Start Of Injection / ms

Pamb=10bar

Pamb=20bar

B100DBE15DBE30Diesel

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.40

20

40

60

80

tbreak

Pamb=20bar

Spra

y Ti

p Pe

netr

atio

n / m

m

Time After Start Of Injection / ms

(c) Pinj=1000bar

B100DBE15DBE30Diesel

Pamb=10bar

Fig. 4. STP evolution.

B100 DBE15 DBE30 Diesel0mm

20mm

40mm

60mm

80mm

0.2ms

0.5ms

0.8ms

1.1ms

Fig. 3. Typical spray evolution process for the four tested fuels.

600 800 10000

10

20

30

40Pamb=20barPamb=10bar

B100DBE15DBE30Diesel

Spra

y A

ngle

/ de

gree

Injection Pressure / bar

Fig. 6. The averaged spray angle for different fuels and at different conditions.

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60

10

20

30

40

50

B100 DBE15 DBE30 Diesel

Spra

y A

ngle

/ de

g

Time After Start Of Injection / ms

Pinj=1000bar, Pamb=20bar

Fig. 5. Spray cone angle evolution.

120 L. Guan et al. / Fuel 140 (2015) 116–125

because the flow regime was in fully developed turbulence condi-tions. While the increase of ambient pressure caused a significantincrease of spray angle because the high ambient gas density

inhibits the spray from axial evolution because of the strengthenedresistance and leads to an increased spray cone angle.

Fig. 7a–c give the projected spray area of B100, DBE15 andDBE30 and diesel versus STP at various injection and ambient pres-sures. The projected spray area quantifies the intensity of the fueland ambient gas mixing. The STP was used as x-axis to eliminatethe effect of injection timing. Results shows that DBE30 gives thelargest spray area under all conditions, indicating a better fuel–air mixing. The projected spray areas are increased with theincrease of ambient pressure and this is consistent with the changeof spray angle induced by ambient pressure. It is expected that lar-ger spray angle leads to larger projected spray area.

0 10 20 30 40 50 60 70 80 900

200

400

600

800

1000Pr

ojec

ted

Spra

y A

rea

/ mm

2

Spray Tip Penetration / mm

B100DBE15DBE30Diesel

Pamb=10bar Pamb=20bar (a) Pinj=600bar

0 10 20 30 40 50 60 70 800

200

400

600

800

1000B100DBE15DBE30Diesel

Proj

ecte

d Sp

ray

Are

a / m

m2

Spray Tip Penetration / mm

Pamb=20barPamb=10bar (b) Pinj=800bar

0 10 20 30 40 50 60 70 800

200

400

600

800

1000

Proj

ecte

d Sp

ray

Are

a / m

m2

Spray Tip Penetration / mm

B100DBE15DBE30Diesel

Pamb=20barPamb=10bar (c) Pinj=1000bar

Fig. 7. Spray projected area for different fuels and at different conditions.

L. Guan et al. / Fuel 140 (2015) 116–125 121

3.2. Droplets number density distribution

The droplets number and size distribution characteristics repre-sents the quality of spray and atomization and it affects the subse-quent combustion and emissions. Due to the high droplets densityat the central of the spray that attenuate the background illumina-tion light, the droplets size distribution of B100, DBE15, DBE30 anddiesel were measured at 40 mm below the nozzle tip and 6 mmfrom the spray axis at injection pressure of 600 bar and ambientpressure of 10 bar. Measurements are conducted at 1500 ls afterthe start of injection energizing when the spray shows quasi sta-tionary period. The recorded images were then processed by thesoftware DaVis 8.0.0 of LaVision. Fig. 8 shows the typical shadowimage of the droplets. Droplets on the focal plane appear sharpand others that are away from the focal plane appear blurred. Bysetting a threshold value of the pixel intensity (25% of the maxi-mum pixel intensity), all regions in the image above the thresholdvalue are considered and segmented. A pixel will belong to a cer-tain segment when its intensity is above this threshold value. Then,a high and low level as percentage of the minimum and maximum

intensities (70% and 30%, respectively) of each segment are definedand used to calculate the particle diameter. The low level diameteris clearly higher than the high level diameter, so a percentage(value used = 200%) that the low level diameter is allowed to beabove the high level diameter is selected to exclude the de-focusedparticles. Detailed algorithm of the droplet sizing can be found inthe product manual of DaVis 8.0. By this way, only the sharp drop-lets on the focal plane are detected as effective droplets and thesize of the droplets can be determined by applying the algorithmintegrated in the software. Depth of field correction was not per-formed in this study, as is discussed by Goldsworthy et al. [30]because the correction would distort the measured droplets sizedistribution towards small particles when the size of droplets isnot big enough. Moreover, it will not affect the results of compar-ative study on atomization characteristics of different fuels. In thisstudy, 1 pixel represents 1.44 lm, hence the droplets with a diam-eter under 8 lm are rejected because the minimum droplet diam-eter that can be accurately resolved is 5 pixel equivalent. The PDIAsystem was used to measure the size-certified polystyrene parti-cles (Thermo Fisher Scientific Inc., Fremont, CA) with diameter29.75 lm, and the measured SMD was 30.64 lm, thus the uncer-tainty of PDIA technique is within 3%.

A total number of 100 images were recorded and no less than4000 droplets were obtained for each fuel. When 100 more imagesof B100 spray were recorded and over 8000 droplets wereobtained, the calculated SMD is changed from 30.40 lm to30.26 lm with the variation less than 1%. Therefore, it is consid-ered that the number of droplets is sufficient to obtain the valuableresults.

Unlike the well-established PDA technique which only detectsthe spherical droplets and is a point-wise method, PDIA candirectly measure even non-spherical droplets on the focal plane.Therefore, the analysis of the droplets number density is of greatinterests. Fig. 9 shows the ensemble of all the detected dropletsof B100, DBE15 and DEB30 and diesel in the detected region. Theblack spots were employed to represent the droplets. As shownin Fig. 9, droplets tend to aggregate on the left half of the focalplane (closer to the spray center) for all fuels studied here. It isexpected that there is a larger number of droplets close to the cen-tral of the spray than the edge of it. A quantitative analysis of thenumber of the droplets is performed to further analyze the drop-lets distributions.

Fig. 10 presents the droplets number density distributions inboth radial and axial direction for all the tested fuels. In the radialdirection, the probability of droplets number density shows largevalue at the left position and it tends to decrease sharply whenposition shift to the right. While a relative even distribution wasfound in the axial direction. All fuels show similar droplets numberdensity distributions. The droplets number density distribution ofother regions of the spray is also important and will be performedin future as the main purpose of this work is to compare the spraycharacteristics of different fuels.

3.3. Droplets size distribution characteristics

Fig. 11 presents the droplets size distribution along with theradial and axial directions of the test fuels. B100 produces largerdroplets for both radial and axial directions. When DBE is added,the blended fuels give the smaller droplets and with the blendingratio of DBE increases, the diameter becomes smaller. This isbecause the addition of DBE leads to a more active breakup pro-cess. It is also found that the diameter decreases slightly as it shiftsfrom the central to the edge of the spray. While along the axialdirection, the droplet size is almost constant.

Fig. 12 shows the normalized droplet size distribution of B100,DBE15, DBE30 and diesel. Similar uni-modal droplets size

0 300 600 900 1200 15000

200

400

600

800

1000

1200

1400 y

Axi

al d

irect

ion

/ µm

Axi

al d

irect

ion

/ µm

Axi

al d

irect

ion

/ µm

Axi

al d

irect

ion

/ µm

Radial direction / µm Radial direction / µm

Radial direction / µm Radial direction / µm

B100

0 300 600 900 1200 15000

200

400

600

800

1000

1200

1400 DBE15

0 300 600 900 1200 15000

200

400

600

800

1000

1200

1400 DBE30

0 300 600 900 1200 15000

200

400

600

800

1000

1200

1400 Diesel

Fig. 9. The droplets assembly for the four tested fuels at the injection pressure of 600 bar and ambient pressure of 10 bar.

Fig. 8. Typical image illustrating the local droplets within the spray.

122 L. Guan et al. / Fuel 140 (2015) 116–125

distributions curves of different fuel spray are observed. For theB100 and diesel spray, the diameter with maximum frequency is25 lm and 20 lm. However, when DBE is blended into biodieselwith volume fraction of 15% and 30%, the diameter with maximumfrequency decreases to 21 lm and 16 lm, respectively. Moreover,the spray of DBE15 and DBE30 show a higher probability of relativelysmaller droplets compared with that of B100, and this indicates thataddition of DBE reduces the statistical average droplet size and willenhance the mixture formation. This can be explained by the factthat the higher viscosity and larger surface tension of biodieselinhibits its atomization process, and as DBE is added, the viscosityand the surface tension of the blended fuel is decreased, whichfavors the break of droplets due to reduced resistance to shear stressand thus better atomization performance is achieved. It also couldbe found that the atomization of DBE30 spray is even better thanthat of diesel.

Fig. 13 shows the cumulative volume distribution of B100,DBE15, DBE30 and diesel. The cumulative volume is significantlyinfluenced by the larger droplets as volume is proportional to thecubic of the average droplet diameter. It is observed that biodieselshows a flattened gradient of cumulative volume. With theincrease of DBE blending ratio, the curve shifts to the small drop-lets, indicating that addition of DBE significantly increases thenumber of the small droplets. For example, the cumulative volumefraction is 60% for biodiesel with a diameter smaller than 35 lm,but it increased to 80% for DBE30. DBE30 and Diesel show approx-imate cumulative volume distribution curves, while DBE15 locatesbetween biodiesel and diesel. It is concluded that increasing DBEblending ratio improves the atomization of biodiesel.

Table 2 gives the statistical results of the characteristic diame-ters calculated from all droplets of each fuel, including D10, D32,Dv10, Dv50 and Dv90. D10 represents the arithmetic diameter

0 300 600 900 1200 1500 18000.000

0.005

0.010

0.015

0.020

0.025

0.030B100

Dro

plet

num

ber p

roba

bilit

y de

nsity

Radial position / µm0 300 600 900 1200

0.0000.0020.0040.0060.0080.0100.0120.0140.0160.0180.020

B100

Dro

plet

num

ber p

roba

bilit

y de

nsity

Axial position / µm

Radial position / µm Axial position / µm

Radial position / µm Axial position / µm

Radial position / µm Axial position / µm

0 300 600 900 1200 15000.000

0.005

0.010

0.015

0.020

0.025

0.030DBE15

Dro

plet

num

ber p

roba

bilit

y de

nsity

0 300 600 900 12000.0000.0020.0040.0060.0080.0100.0120.0140.0160.0180.020

DBE15

Dro

plet

num

ber p

roba

bilit

y de

nsity

0 300 600 900 1200 15000.000

0.005

0.010

0.015

0.020

0.025

0.030DBE30

Dro

plet

num

ber p

roba

bilit

y de

nsity

0 300 600 900 12000.0000.0020.0040.0060.0080.0100.0120.0140.0160.0180.020

DBE30

Dro

plet

num

ber p

roba

bilit

y de

nsity

0 300 600 900 1200 15000.000

0.005

0.010

0.015

0.020

0.025

0.030

Dro

plet

num

ber p

roba

bilit

y de

nsity

Diesel

0 300 600 900 12000.000

0.005

0.010

0.015

0.020Diesel

Dro

plet

num

ber p

roba

bilit

y de

nsity

Fig. 10. Droplets number density distributions in both the spray radial and axial directions.

L. Guan et al. / Fuel 140 (2015) 116–125 123

0 10 20 30 40 50 600.0

0.2

0.4

0.6

0.8

1.0

B100 DBE15 DBE30 Diesel

Diameter / µm

Cum

ulat

ive

Volu

me

Dv10 Dv50 Dv90

Fig. 13. The cumulative volume percentage as a function of droplet diameter for thefour tested fuels.

0 300 600 900 1200 150015

20

25

30

35

40 B100 DBE15 DBE30 Diesel

Dia

met

er /

µ m

Radial direction / µm

0 300 600 900 120015

20

25

30

35

40 B100 DBE15 DBE30 Diesel

Dia

met

er /

µm

Axial direction / µm

Fig. 11. Droplets size distribution along with the radial and axial directions.

10 15 20 25 30 35 40 45 50 55 600

2

4

6

8

10

12 B100 DBE15 DBE30 Diesel

Prob

abili

ty %

Diameter / µm

Fig. 12. Normalized droplet size distribution of the four tested fuels.

Table 2Statistics of droplet size and numbers in the target area of the spray.

(lm) B100 DBE15 DBE30 Diesel

D10 24.63 22.71 21.66 22.57D32 30.40 27.62 26.42 27.35Dv10 21.99 19.98 18.63 19.33Dv50 32.46 29.59 28.16 28.01Dv90 43.67 39.64 39.26 39.51Total number 4345 4481 4786 4436

124 L. Guan et al. / Fuel 140 (2015) 116–125

and D32 is the SMD, which is defined as a droplet having the samevolume to surface area fraction as the total droplets. Dv10, Dv50and Dv90 are extracted from the cumulated volume of all the drop-lets as is shown in Fig. 13. They represent that 10%, 50% and 90% ofthe total volume is made up droplets having a diameter smallerthan Dv10, Dv50 and Dv90 respectively. The results show that bio-diesel gives the largest characteristics diameter. With the increaseof DBE blending ratio, the characteristic diameter is decreased.DBE30 have even smaller characteristic diameters than diesel. Thisobservation can be explained by the differences in viscosity, sur-face tension and density of the test fuels. DBE blending decreasesthe viscosity and surface tension and consequently the resistanceof droplets to shear stress which favors the breakup of dropletsor liquid ligaments. It is concluded that the atomization perfor-mance of biodiesel is enhanced by the addition of DBE. It is furthernoted that the decreasing tendency of STP when DBE is blendedcan be explained by the decreased SMD. Specifically, the STP isexpected to increase with the increase of the inertia of spray andresistance of the ambient. When DBE is added, the droplets inthe spray have a statistically smaller radius, decreasing the inertiaof the spray and consequently resulting in the decreased STP.

4. Conclusions

In this paper, the spray and atomization characteristics of bio-diesel, biodiesel-DBE blended fuels and conventional diesel wereinvestigated at three injection pressures and two ambient pres-sures. Schlieren method was employed to measure macro spraycharacteristic and PDIA technique was used to measure dropletssize distribution and characteristic diameters. Addition of DBEhas an obvious influence on spray evolution process and STP whenvolumetric ratio of DBE increases to 30%, while similar trend wasfound for biodiesel and 15% DBE blended fuel. Biodiesel gives smal-ler spray cone angle and projected area compared with the blendedfuels and spray cone angle and projected area are increased withincrease of blending ratio. Droplets number distribution indicatesthat droplets tend to aggregate to the central of spray and numberdensity probability decrease sharply when the position shifts fromleft to the right of the field of view. Similar uni-modal curves arepresented in the droplets size distributions for the biodiesel andblended fuels. However, a discrepancy was found in which thespray of biodiesel consists of a higher probability of relatively lar-ger droplets while the spray of DBE30 consists of a higher probabil-ity of relatively smaller droplets. The characteristic diameters ofbiodiesel and blended fuels show that all characteristic diametersare decreased with the increase of DBE addition. The results revealthat, compared to the diesel, the deteriorated spray performance ofbiodiesel can be recovered by the addition of DBE, which signifi-cantly improves the spray and atomization process of biodiesel.

Acknowledgement

This work is supported by the National Natural Science Founda-tion of China (51206131, 51121092).

L. Guan et al. / Fuel 140 (2015) 116–125 125

References

[1] Kalam M, Masjuki H. Biodiesel from palmoil-an analysis of its properties andpotential. Biomass Bioenergy 2002;23(6):471–9.

[2] Wang W, Lyons D, Clark N, Gautam M, Norton P. Emissions from nine heavytrucks fueled by diesel and biodiesel blend without engine modification.Environ Sci Technol 2000;34(6):933–9.

[3] Canakci M. Combustion characteristics of a turbocharged DI compressionignition engine fueled with petroleum diesel fuels and biodiesel. BioresourceTechnol 2007;98(6):1167–75.

[4] Lapuerta M, Armas O, Rodriguez-Fernandez J. Effect of biodiesel fuels on dieselengine emissions. Prog Energy Combust Sci 2008;34(2):198–223.

[5] Hoekman SK, Robbins C. Review of the effects of biodiesel on NOx emissions.Fuel Process Technol 2012;96:237–49.

[6] Szybist JP, Song J, Alam M, Boehman AL. Biodiesel combustion, emissions andemission control. Fuel Process Technol 2007;88(7):679–91.

[7] Muncrief RL, Rooks CW, Cruz M, Harold MP. Combining biodiesel and exhaustgas recirculation for reduction in NOx and particulate emissions. Energy Fuel2008;22(2):1285–96.

[8] Shudo T, Nakajima T, Hiraga K. Simultaneous reduction in cloud point, smoke,and NOx emissions by blending bioethanol into biodiesel fuels and exhaust gasrecirculation. Int J Engine Res 2009;10(1):15–26.

[9] Lee CS, Park SW, Kwon SI. An experimental study on the atomization andcombustion characteristics of biodiesel-blended fuels. Energy Fuel2005;19(5):2201–8.

[10] Wang X, Huang Z, Kuti OA, Zhang W, Nishida K. Experimental and analyticalstudy on biodiesel and diesel spray characteristics under ultra-high injectionpressure. Int J Heat Fluid Flow 2010;31(4):659–66.

[11] Gao Y, Deng J, Li C, Dang F, Liao Z, Wu Z, et al. Experimental study of the spraycharacteristics of biodiesel based on inedible oil. Biotechnol Adv2009;27(5):616–24.

[12] Bhale PV, Deshpande NV, Thombre SB. Improving the low temperatureproperties of biodiesel fuel. Renew Energy 2009;34(3):794–800.

[13] Boudy F, Seers P. Impact of physical properties of biodiesel on the injectionprocess in a common-rail direct injection system. Energy Convers Manage2009;50(12):2905–12.

[14] Yoon SH, Park SH, Suh HK, Lee CS. Effect of biodiesel–ethanol blended fuelspray characteristics on the reduction of exhaust emissions in a common-raildiesel engine. J Energy Resour ASME 2010;132(4):042201.

[15] Shi X, Pang X, Mu Y, He H, Shuai S, Wang J, et al. Emission reduction potentialof using ethanol–biodiesel–diesel fuel blend on a heavy-duty diesel engine.Atmos Environ 2006;40(14):2567–74.

[16] Arcoumanis C, Bae C, Crookes R, Kinoshita E. The potential of di-methyl ether(DME) as an alternative fuel for compression–ignition engines: a review. Fuel2008;87(7):1014–30.

[17] Cinar C, Can Ö, Sahin F, Yucesu HS. Effects of premixed diethyl ether (DEE) oncombustion and exhaust emissions in a HCCI-DI diesel engine. Appl Therm Eng2010;30(4):360–5.

[18] Melin K, Hurme M. Evaluation of lignocellulosic biomass upgrading routes tofuels and chemicals. Cell Chem Technol 2010;44(4):117.

[19] Bi S, Zhao G, Wu J. Surface tension of diethyl ether, diisopropyl ether, anddibutyl ether. J Chem Eng Data 2009;55(4):1523–6.

[20] Meng X, Wu J, Liu Z. Viscosity and density measurements of diisopropyl etherand dibutyl ether at different temperatures and pressures�. J Chem Eng Data2008;54(9):2353–8.

[21] Harvey BG, Meylemans HA. The role of butanol in the development ofsustainable fuel technologies. J Chem Technol Biotechnol 2011;86(1):2–9.

[22] Nabi MN, Ogawa H, Miyamoto N. Approaches to extremely low emissions andefficient diesel combustion with oxygenated fuels. Int J Engine Res2000;1(1):71–85.

[23] Beeckmann J, Aye M, Gehmlich R, Peters N. Experimental investigation of thespray characteristics of di-n-butyl ether (DNBE) as an oxygenated compoundin diesel fuel. Stroke 2015:06–22.

[24] Kashdan JT, Shrimpton JS, Whybrew A. A digital image analysis technique forquantitative characterisation of high-speed sprays. Opt Laser Eng2007;45(1):106–15.

[25] Berg T, Deppe J, Michaelis D, Voges H, Wissel S. Comparison of particle size andvelocity investigations in sprays carried out by means of differentmeasurement techniques. ICLASS’06, Kyoto, Japan, 2006.

[26] Hiroyasu H, Arai M. Structures of fuel sprays in diesel engines. SAE Tech Paper900475; 1990.

[27] Reitz RD, Diwakar R. Effect of drop breakup on fuel sprays. SAE Tech Paper860469; 1986.

[28] Valentino G, Allocca L, Iannuzzi S, Montanaro A. Biodiesel/mineral diesel fuelmixtures: spray evolution and engine performance and emissionscharacterization. Energy 2011;36(6):3924–32.

[29] Payri R, Salvador FJ, Gimeno J, Novella R. Flow regime effects on non-cavitatinginjection nozzles over spray behavior. Int J Heat Fluid Flow2011;32(1):273–84.

[30] Goldsworthy L, Bong C, Brandner P. Measurements of diesel spray dynamicsand the influence of fuel viscosity using PIV and shadowgraphy. AtomizationSpray 2011;21(2).