typinjection and atomiz

bus

. Vana

ing and

nt, Belg

Antonio

ce Engi

um

Pump-line-nozzle system

s. The results reported

t volume combustion

nder non-evaporating

medium speed diesel

type and temperature.

oils and animal fats.

igh speed camera was

enetration and spray

parameter to control

because it significantly affects the fuel properties. Both the injection timing and injection

roperties on the spray

ly deteriorated atomi-

td. All rights reserved.

These days, it is clear that the energy availability from non-

renewable sources is limited [1]. Furthermore, fossil fuels

have now been identified as one of the main culprit of climate

change and environmental pollution [2,3]. Therefore,

duction are required.

The automotive industry and the transport sector in general

have to meet future stringent emission limits and zero net

carbon dioxide emissions in the long-termview [4]. This aim is

also driven by the limit of the worldwide oil reserves and the

rising consumption and price of fossil fuels. As a first, short

* Corresponding author. Tel.: 32 9 264 34 53.

Available online at www.sciencedirect.com

.co

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8E-mail address: jonas.galle@ugent.be (J. Galle).Sprays

Medium speed diesel engines

duration are affected by the fuel properties. The influences of these p

development were less pronounced. At low temperatures, a strong

zation of oils and fats was observed.

2013 Elsevier L

1. Introduction solutions for carbon dioxide emission re22 June 2013

Accepted 5 July 2013

Available online 3 August 2013

Keywords:

Biofuels

Constant volume combustion

chamber

Physical fuel properties

affecting atomization, spray development and combustion processe

in this paper have been obtained by experimentation with a constan

chamber. The influences of physical fuel properties on injections u

conditions are studied, using a pump-line-nozzle system from a

engine with injection pressures up to 1200 bar, by changing the fuel

Experiments were conducted for diesel, biodiesel, straight vegetable

Injection pressure and needle lift measurements were analyzed. A h

used to visualize the spray, which enabled us to study the spray p

angle. Our results show that the fuel temperature is an importantReceived in revised form However, these fuels have different physical, chemical and thermodynamic propertiesin an optical com

J. Galle a,*, S. Defruyt a, CA. Verliefde c, S. VerhelstaGhent University, Faculty of Engineer

Sint-Pietersnieuwstraat 41, B-9000 GebTechnical University of Havana JosecGhent University, Faculty of Bioscien

Coupure links 653, B-9000 Gent, Belgi

a r t i c l e i n f o

Article history:

Received 30 August 20120961-9534/$ e see front matter 2013 Elsevhttp://dx.doi.org/10.1016/j.biombioe.2013.07.ation of liquid biofuelstion chamber

de Maele a, R. Piloto Rodriguez b, Q. Denon c,

Architecture, Department of Flow, Heat and Combustion Mechanics,

ium

Echeverria, Centro de Estudios de Energa Renovable, Havana, Cuba

neering, Department of Applied Analytical and Physical Chemistry,

a b s t r a c t

Due to the scarcity of fossil fuels and the future stringent emission limits, there is an

increasing interest for the use of renewable biofuels in compression ignition engines.influence of fuel e and properties on the

Experimental investigation concerning thehttp: / /www.elsevierier Ltd. All rights reserved004m/locate/biombioe.

Tf Fuel temperature

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8216term solution, considerable effort has been made to reduce

harmful emissions from diesel engines, to meet these stricter

emission standards [5]. However, as fossil fuels are limited, a

second, long term solution is essential. The use of renewable

fuels can be a solution for the above two problems simulta-

neously [6]. Biofuels can offer a significant contribution in

reducing harmful emissions. Besides, it is a biologically

derived product and thus one of the most efficient ways of

Abbreviations& Symbols

AF Animal Fats

ASOI After Start Of Injection

B Bulk modulus of the fuel

CI Compression Ignition

CVCC Constant Volume Combustion Chamber

D Diesel

DI Direct Injection

FLP Full Lift Period

fps Frames per Second

GUCCI Ghent University Combustion Chamber I

RSO Rapeseed Oil

LL Liquid Length

NL Needle Lift

NOP Needle Opening Pressure

PLN Pump-Line-Nozzle

PO Palm Oil

RME Rapeseed Methyl Ester

RSO Rapeseed Oil

SMD Sauter Mean Diameter

SVO Straight Vegetable Oilsreducing the carbon dioxide emissions [7].

It is generally accepted [8e11] that the fuel atomization

process is extremely important for the reduction of fuel con-

sumption and exhaust emissions of diesel engines. A good

atomization leads to an improvement of the airefuel mixing,

causing a more complete and more efficient combustion

process that results in a higher performance and less

pollutant emissions. So, in order to fulfill future emission

regulations, the atomization process has to be investigated

thoroughly. Although there are many literature reports

related to the review on performance and exhaust emissions

of compression ignition (CI) engines fueled with biofuels,

these studies mostly treat the engine as a black box, without

studying in depth how biofuels affect the injection process or

what are the repercussions of its use on the airefuel mixing

[6,12].

Biofuels have different properties, affecting atomization

and thus combustion. In addition, fuel properties are affected

by operation temperature and pressure. As an example, fuel

viscosity in the cold start phase can be doubled compared to

normal operating conditions. However, there is little knowl-

edge about the influences of the different fuel properties on

the atomization process [7,13]. This knowledge is essential in

order to adapt engines for the use of alternative fuels and to

ensure reliable calibration of the injection parameters at all

operating conditions and for all fuels [7]. Furthermore, this

knowledge is important facing the arrival of new alternativefuels and could be used to design fuels displaying better

characteristics [14].

We used a traditional pump-line-nozzle injection, as

implemented in current medium speed diesel engines system

to conduct single-shot injection experiments in an optically

accessible constant volume combustion chamber (CVCC)

under an inert, non-evaporating atmosphere. The experi-

mental conditions were varied and different fuels were used.

Tc Injector cooling temperatureV Volume of fuel between the plunger of the pump

and he injector needle

Greek symbols

a Cam angle position

Roman symbols

DP Pressure drop across the nozzle orifice

Cd Discharge coefficient

Fp Force of the fuel acting on the injector needle

Fs Force of the spring acting on the injector needle

Fv Viscous force on the injector needle

NLstart The cam position when the needle starts to lift

[CA]Pa Ambient pressure, pressure inside the

combustion chamber

rf Fuel density

ra Ambient density, gas density inside the

combustion chamber

Ta Ambient temperature, temperature inside the

combustion chamberWe studied the influences of engine parameters, physical fuel

properties and in-cylinder conditions on the injectionpressure

andneedle lift andon the spray development andatomization.

2. Experimental method

2.1. Experimental setup and data acquisition

The experiments were conducted in the Ghent University

Combustion Chamber I (GUCCI). This constant volume com-

bustion chamber has an internal volume of 1603 mm3 and

optical access in 2 directions through quartz windows of

150 mm in diameter. The injector is placed along the space

diagonal (Fig. 1). The aim of this setup is to carry out and

visualize single shot injections with different types of fuels.

The injection system is an exact copy of the Pump-Line-

Nozzle (PLN) system, used by the engine manufacturer

(256 mm bore engines). The injection pump in the setup is

operated by a camshaft which is driven by an electric motor

with a power of 2.2 kW. Considering that the instantaneous

power from the electric motor is not enough to operate the

pump with a low coefficient of irregularity, the required en-

ergy is stored in a flywheel mounted on the camshaft.

A single shot injection is obtained by the pneumatic

operation of the control rod of the pump. At rest a pneumatic

cylinder retracts the control rod into zero load position.

co

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8 217A Bryce Woodward heavy-duty sac-hole nozzle with eight

440 mm orifices was used. During the experiments, the atmo-

sphere in the combustion chamber consisted of an inert gas, N2,

at a maximum pressure Pa of 80 bar and a maximum tempera-

ture Ta of 150 C. As a consequence, only the atomization andspraydevelopmentofnon-evaporatingsprayswas investigated.

All optical measurements are obtained with shadowgraph

imaging [15]. Continuous light from an Xe-arc lamp is focused

onto a pinhole of 1 mm to simulate a point source. The light is

collimated by a 1200 mm focus biconvex lens (f/8) to a parallel

light beam. Planar mirrors steer the beam through the com-

bustion chamber. A schematic overview of the setup is given

in Fig. 2.

The authors refer to [1] for a more detailed description of

Fig. 1 e Left: Mounting of the injector on the constant volume

the x-axis.the injection equipment, the optical measurement equip-

ment, the sensors and the data acquisition of the GUCCI setup.

2.2. Tested fuels

The fuels we tested are diesel (D), a biodiesel derived from

rapeseed oil (rapeseedmethyl esters, RME), animal fats (AF) and

twostraight vegetableoils (SVO): rapeseedoil (RSO) andpalmoil

(PO). Inthiswork,2differentbatchesof rapeseedoil, indicatedas

RSO1 and RSO2, are used of which the origin is unknown. Inter-

esting tomention is that theodor andcolorof the2batcheswere

different. When investigating some physical properties, rele-

vant for atomization, small but significant differences were

noticed (Figs. 3e5). A GCeMS analysis (cfr. Fig. 6) reveals a dif-

ference in chemical composition which will probably influence

the ignition and combustion process more than the atomiza-

tion. The combustion behavior is not studied in this work. No

significant differencewas between RSO1 and RSO2 in the optical

spray images. Nevertheless, for all results shown along this

study, theRSObatch is specified. Inorder to study the influences

of the physical fuel properties on the injection process, knowl-

edge of the exact values of these properties at different fuel

temperatures is needed. The dynamic viscosity was measured

using a Brookfield DV II viscometer, measurements of thesurface tension were based on theWilhelmy plate method and

for the density measurements an Anton Paar DMA5000 density

meter was used. All measurements were conducted at atmo-

sphericpressure.Theresults forviscosityanddensityareshown

inFigs. 3 and4.Thebulkmoduliof fuelswerenotmeasured.The

bulk modulus is a measure for the compressibility of a fluidum

and is defined as the ratio of the pressure increase to the

resulting relative decrease of the volume. Since the bulk

modulus isamainparameteraffecting the injectionpressure for

PLN-systems, data on the bulk moduli of fuels were gathered

from literature. These data are represented in Fig. 5.

2.3. Fuel temperature control

mbustion chamber. Right: Injector cooling and definition ofFuel properties strongly depend on the fuel temperature.

Therefore it is very important to know and control the fuel

temperature Tf, defined as the temperature of the injected fuel

at the moment of injection, for every experiment.

Between two injections, a certain amount of fuel resides

inside the injector. Since the internal volume of the injector

exceeds the injected volume (2000mm3} per injection), all fuel

that will be injected has been inside the injector since the end

of the previous experiment. The time between two single shot

experiments is sufficiently long to assume that the fuel inside

the injector adopts the equilibrium temperature of the

injector. Based on this assumption, we conducted tempera-

ture measurements inside the injector in order to determine

the regime temperature of the fuel.

The injector is both in contact with the chamber walls and

injector cooling (Fig. 1), resulting in a temperature gradient

inside the injector. The chamber walls are electrically heated

in order to control the gas temperature Ta inside the chamber.

The temperature Tc of the injector cooling oil is controlled as

well. Our goal was to set up a correlation, which expresses the

fuel temperature Tf as a function of the PID-controlled tem-

peratures Tc and Ta.

To set up this correlation, a set of experiments was con-

ducted. An injector was provided with a drilling along its

tan

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8218central axis, making it possible to use a K-type thermocouple

tomeasure the temperature inside the injector. This was done

along the central axis, using a stepsize of 1 cm and starting at

the sac volume (x 0 cm) up to the place where the fuel entersthe injector (x 18 cm). The location x along the axis is shownin Fig. 1. This was done for different combinations of Tc and Ta(Fig. 7). Note that the conditions inwhich Tc exceeds Ta are not

considered, since this is not relevant for engine research. To

make the correlationmore robust, artificial pointswere added;

these points correspond with the situation in which the tem-

perature of the injector cooling and chamber cooling are equal.

In these cases, the fuel temperature should be equal to the

cooling and chamber temperature as well. Afterward, the in-

ternal fuel volume of the injector per unit of length along the x-

axis was used to calculate the fuel temperature as the mass

Fig. 2 e Schematic overview of the consweighted average of the temperatures measured along the x-

axis. This resulted in a Tf-value for every measurement point

of Fig. 7. Finally, these datawere used to fit the coefficients of a

correlation by the use of the LevenbergeMarquardt algorithm

Fig. 3 e Viscosity as a function of temperature for the fuels

used in our experiments.[16], which expresses the fuel temperature Tf as a function of

Tc and Ta. The fitting of the coefficients resulted in differences

between the predicted and experimental fuel temperatures

which are below 1 C.

2.4. Measurements and standard deviation

2.4.1. Injection pressure profile and needle displacementFor PLN injection systems, the injection pressure, which is the

pressure of the fuel inside the injector and the injection pipe, is

not constant during an injection. Furthermore, this pressure

actsuponthe injectorneedle,affectingtheneedledisplacement.

We conducted an analysis on the influences of engine parame-

ters, fuel properties and in-cylinder conditions on the injection

pressure and the needle displacement. The injection pressure

t volume combustion chamber GUCCI.was measured using a Kistler 4067A2000 pressure sensor

mounted half-way the injection pipe. The standard deviations

on these data are below 2% during the injection event. To

Fig. 4 e Density as a function of temperature for the fuels

used in our experiments.

monitored by PID controllers.

article.)

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8 219measure the needle displacement, a small magnet was fixed at

the topof the injectorneedle. In theupper regionof the injectora

hall-effect sensorwasmounted, deliveringa signal proportional

to the distance of the magnet. Both the injection pressure and

the needle position weremeasured every 1/10 cam angle (ca).

2.4.2. Spray imagesA high speed CCD camera (Photron FastcamAPX RS) was used

to obtain two dimensional spray images with a pixel resolu-

tion of 512 512 at a frame rate of 10 kHz. Image processingwas able to extract the spray penetration length, spray edge

and spray angle. In literature, many different definitions of

these characteristics can be found. This might lead to

different results among institutes or evenmisinterpretation of

the users of the data. Our definitions are based on those re-

Fig. 5 e Bulk modulus as a function of temperature. Data

obtained from literature.ported and used by Siebers [17]. Details of the used image

processing are described by Galle et al. [1].

Similar standard deviations were obtained for all mea-

surement conditions. The standard deviation percentage is

not significantly reducedwith the amount of experiments, but

becomes more stable over the time after start of injection

(ASOI). This was shown previously by Galle et al. [1]. Galle et al.

obtained a stable value for the spray penetration of about 4%.

Fig. 6 e Fatty acid composition by GCeMSIn this work, smaller standard deviations were found with a

stable value being around 2%. Higher standard deviations for

the spray angle were noticed. However, once the spray is fully

developed the deviation on the spray angle varies between 1

and 3 for most experimental conditions.The reason for the lower standard deviations in thiswork is

meanly the changes to the setup for the capability of more

accurate setting of the boundary conditions: deviations on the

injection profile were reduced by the change of the fuel cir-

culation pump, the chamber pressure was controlled by a

more accurate pressure sensor, and temperatures wereFig. 7 e Visualization of the experimental measurement

points (red diamonds). The blue line represents Ta [ Tc.

The open circles represent artificial measurement points.

(For interpretation of the references to colour in this figure

legend, the reader is referred to the web version of this3. Results and discussion

3.1. Injection pressure profile and needle displacement

3.1.1. General view on the injection pressure profileBefore discussing the detailed results of the analysis of the

different phases of the fuel injection, we start with a general

for the investigated fuels in mass%.

view on the injection pressure profile. The main parameter

affecting these processes is the engine speed (dotted lines in

Fig. 8). The higher the engine speed, the higher the injection

pressure and the longer the injection duration expressed inca. On a time based scale however, the injection period de-creases with increasing engine speed. The period between the

start of pressure buildup and the end of injection for diesel is

approximately 18, 12 and 9ms for injections at 400, 700 and1000rpm respectively.

In this work, the main focus is the study of the influences

of fuel properties on the injection process. Fig. 9 and the solid

lines on Fig. 8 illustrate the significant effect of fuel properties

on the injection pressure profile. The height of the injection

pressure is mainly determined by the bulk modulus of the

fuel, while the injection duration strongly depends on the fuel

density. This will be discussed later on.

da V da

where a is the cam angle position. B is pressure and temper-

ature dependent and dV/da is determined by the geometry of

the cam only. The secondary parameters affecting the pres-

sure are the expansion of the fuel pipe when pressure rises

and the fuel leakages along the plunger. As a conclusion, the

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8220The pressure profile is quite wavy due to pressure waves

traveling along the injection pipe and the injector. However,

the exact same wave pattern is observed for different mea-

surements at the same conditions. Even under different con-

ditions, the wave pattern is very similar, as can be seen in

Fig. 9. The frequency of these waves is quite constant during

the injection. At 1000rpm for example, the distance betweentwo adjacent peaks is around 0.7ca. Small variations wereobserved between different fuels due to changes in the speed

of sound of the fuel. The amplitude of the waves is attenuated

due to the viscosity of the fuel [18]. For the more viscous SVO

and AF the curves are rather smooth at the top of the injection

pressure profile. This is not the case for the less viscous fuels

RME and diesel (Fig. 9). This effect is also illustrated by the

solid lines of Fig. 8: the lower the fuel temperature and thus

the higher the viscosity, the more smooth the injection pres-

sure profile. At the end of injection, after the closure of the

needle, the pressure signal is strongly oscillating. This is

caused by pressure waves due to the sudden closure of the

needle. In this region the waves are not completely repro-

ducible resulting in high standard deviations in this region.

However, the frequency seems to be quite constant for all

measurements. This oscillatory region after the end of injec-

tion will not be studied in detail in this text.

In the next paragraphs we will discuss some parts of the

injection process in more detail. We will focus on the mostFig. 8 e Injection pressure profiles of injections with rapeseed o

setpoints.important influences of fuel properties on the injection

process.

3.1.2. Pressure build-upFor a PLN injection, the movement of the injection pump

plunger is camshaft driven. Pressure is built up inside the

pump, causing the delivery valve of the pump to open. As the

camshaft is rotating, the plunger moves upward and the vol-

ume V of fuel between the plunger of the pump and the

injector needle is reduced. The increase of pressure related to

this decrease in volume is mainly determined by the bulk

modulus B of the compressed fuel:

dp BP;TdV (1)

Fig. 9 e Injection pressure profiles of injections at 1000 rpm

engine speed, using different fuels. Tc[ 50 C, Ta[ 140 C,Tf [ 56.5c and Pa [ 80 bar.il for different fuel temperatures and different engine speed

nt temperatures (Tf[ 26 and 45 C), 1000 rpm and Pa[ 80 bar.

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8 221bulk modulus is the most important parameter affecting the

pressure build-up.

A higher bulk modulus results in higher injection pres-

sures. This is illustrated by Fig. 9, which shows that the fuels

with the highest B have the highest injection pressure and the

solid lines of Fig. 8, which shows that the decrease in B for

higher fuel temperatures, results in a decrease of the injection

pressure.

3.1.3. Needle liftThe volume V is compressed until the injector needle lifts.

This moment is determined by the forces acting on the needle

and is defined as the point where the needle lift exceeds 4% of

the full lift. Same definition holds for the closing of the

injector. On the one hand there is the force of the spring (Fs),

who keeps the injector needle on its seat, on the other hand

there is the pressure of the fuel acting on the needle resulting

in the force Fp. Once the needle opening pressure (NOP) is

reached, F overcomes F and the needle starts to lift. The

Fig. 10 e Spray images of diesel and RSO1 injections at differe

Penetration lengths are around 180 mm.p s

dynamics of the needle displacement can be expressed as

follows:

mat Fpt Fst Fvt (2)where a(t) is the upward acceleration of the needle. Fs is pro-

portional to the needle displacement. The viscous force Fv is

proportional to theneedlesurfaceareaAandthe fuelviscositym:

Fv mAdvtdy (3)

dv(t)/dy is the velocity gradient in the fuel surrounding the

needle and depends on the needle velocity.

These equations show that the inertia of the needle with

mass m and the viscous forces of the fuel, slow down the

needle displacement. This causes a significant delay between

the moment when the NOP is reached and the moment when

the needle starts to lift. During this period pressure buildup

continues. A higher bulk modulus of the fuel causes a faster

increase in Fp, resulting in an earlier start of needle lift.

Fig. 11 shows NLstart, which is the cam position at the

moment when the needle starts to lift, as a function of fueltemperature for different fuels. The most important fuel

properties affecting the moment of needle lift are the bulk

modulus and the viscosity. A higher bulk modulus causes a

higher injection pressure, a larger Fp and thus an earlier start

of needle lift. A higher viscosity increases the viscous forces

and delays the start of needle lift. The effect of viscosity

retarding the needle displacement was already observed for

common rail (CR) injections [19,20]. For CR injection systems,

the injection pressure is constant, so there is no effect of the

bulk modulus counteracting the effect of viscosity.

For diesel and RME the changes in viscosity are rather

small when temperature is varied (Fig. 3). The needle lift is

delayed when Tf increases, due to the decrease in B. For the

SVO and AF, the changes in viscous forces are more impor-

tant, especially in the low temperature range (Fig. 3). The ef-

fects of B and viscosity counteract each other resulting in an

NLstart which is less dependent on the fuel temperature. In the

high temperature range (70e90 C), where the viscosity israther constant, the effect of the bulk modulus becomesmore dominant: the needle lift is retarded with increasing

temperature.

Fig. 11 e NLstart [ca] as a function of fuel temperature fordifferent fuels. Average values of 9 experiments at engine

speed set point of 1000 rpm. Standard deviation

One can conclude that at elevated temperatures (above

60 C) the higher bulk moduli of SVO and AF lead to an earlierstart of needle lift compared to diesel injections.

The duration of the needle lift period did not seem to vary a

lot with temperature. The needle lift period is defined as the

time between opening and closing of the nozzle. The observed

increase of 0.1ca between 40 and 90 C is possibly the result ofa decreasing bulk modulus. The differences between different

fuels, however, are more clear. The needle lift period of diesel

injections (1e1.1ca) is longer than the one of RME injections(0.9ca) at 1000rpm engine speed. For injections at 700rpmno significant difference was observed (1.3e1.5ca) betweenthese two fuels. The needle lift period of SVO andAF injections

is shorter (0.8ca at 1000 rpm and 1e1.1ca at 700rpm). Thiscan be related to the higher bulk modulus of SVO and AF,

increasing Fp acting on the needle during needle lift. The

viscous forces do not seem to have a significant effect on the

needle lift period. Combinedwith the earlier start of needle lift

(at high temperatures), the higher bulk moduli of SVO and AF

result in 0.5ca earlier injections for these fuels compared todiesel (Fig. 12).

The earlier injection timing for SVO and AF will impact

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8222engine emissions and performance. Experiments of Bari et al.

[21] revealed that oils (waste cooking oil) and diesel respond

identically to injection timing changes. So, based on the

knowledge on diesel combustion, we can try to predict the

influences of the injection timing on the performances and

emissions of engines running on SVO or AF. The earlier start of

injection implies that the temperature and pressure inside the

cylinder will be lower at the start of injection. As a result, the

ignition delay will increase for the SVO and AF. The poorer

atomization and lower volatility of the bio-fuels can even

reinforce the increase of the ignition delay compared to diesel

fuel. The later start of ignition results in an increased amount

of heat release in the premixed combustion phase, increasing

the probability of diesel knock and causing a higher

Fig. 12 e Comparison of needle lift as a function of ca fordiesel, RME, animal fat and rapeseed oil injections.Tc [ 90 C, Ta [ 140 C, Tf [ 91.5 C, engine speed setpoint [ 1000 rpm, Pa [ 60 bar.temperature during the subsequent diffusive combustion

phase [22]. The period during which the mean temperature

inside the cylinder is above 1500 K, is a very important factor

affecting the formation of thermal NOx. As a result, the

increased ignition delay due to the earlier injection, leads to

higher NOx emissions [22]. However, the earlier injection

timing also has some advantages. First of all, more fuel is

burned in the premixed combustion phase. During a premixed

combustion less local over-rich spots are present compared to

the diffusive combustion phase. This results in a more com-

plete combustion and lower CO, HC and smoke emissions [22].

3.1.4. Pressure drop caused by needle liftWhen the needle rises, fuel starts to leave the nozzle through

the orifices and enters the chamber. This sudden decrease in

pressure at the nozzle tip results in a pressure wave traveling

upstream.When it arrives at the pressure sensor, a drop in the

pressure profile is observed. In Fig. 9 this drop can be seen

between 345 and 346ca. The distance between the sensor andthe needle tip is about 40 cm, the delay between the start of

the needle lift and the start of the drop is 0.8ca at 1000rpm(970rpm effectively) and 0.6ca at 700rpm engine speed. Bothcorrespond to 280 ms, which is the time needed to travel those

40 cm at approximately 1400 m/s, which is the velocity of

sound in the fuel. Although the differences are smaller than

the standard deviation on the delay, it was possible to detect

an increasing trend of the delay with increasing fuel temper-

ature, due to the decrease in speed of sound of the fuel.

3.1.5. Injection periodTo express the volumetric flow rate through the nozzle ori-

fices, the theoretical volume flow rate of the fuel is multiplied

with a discharge coefficient Cd:

_V Cd _Vth CdA2DPrf

s(4)

with rf the fuel density, A the orifice cross section and DP the

pressure drop across the orifice.

Different authors [7,14] stated that the discharge coeffi-

cient is independent of the fuel density for every DP. Although

CR injection systems were used for their experiments, we

expect their conclusions to be valid for PLN injections as well.

So _Vis directly proportional to the square root of the ratio of DP

to rf. For PLN systems, _V is not constant during an injection

because DP (and thus Cd) is not constant. These changes in DP

cause only very small changes in rf, so we can assume rf to be

approximately constant.

If leakages are neglected, the total injected volume of fuel

is constant for PLN systems at constant pump position

(2000 mm3 for our experiments). Using the simplification that

DP is constant, equation (4) expresses that the injection

duration is directly proportional torf

p: This simplified ana-

lyses shows that the fuel density is amain parameter affecting

the injection duration. This was confirmed by our experi-

mental results. We observed a linear relationship between the

full lift period (FLP) andrf

p(Fig. 13). The FLP was defined as

the duration of the period of full needle opening. We canconclude that the higher density of SVO and AF results in

longer injections at lower volumetric flow rates, but higher

the height of the partial lift is between 25 and 40% of full lift,

compared to 8e20% for diesel and RME (Fig. 14). This results in

larger amounts of fuel injected during the partial needle lift for

SVO and AF. There is a slight effect of temperature on the

height of the partial lift (Fig. 15). The height decreases at

higher temperatures because of the temperature dependent

behavior of B. A more important parameter however, is the

chamber pressure. The chamber pressure acts up on the

needle tip, resulting in an extra upward force. At high engine

speeds, these forces were negligible compared to the fast and

high pressure buildup. At 400rpm, on the contrary, this af-fects the height of the partial lift significantly. For SVO and AF

an increase of 10% of full lift was observed when the chamber

pressure was increased from 40 to 80 bar (Fig. 15).

3.2. Spray development

All experiments were conducted in an (N2) atmosphere at

chamber pressures ranging from 40 bar to 80 bar, engine

speeds from 400rpm to 1000rpm. The chamber gas temper-

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8 223mass flow rates. This effect can be slightly reinforced by the

reduced leakages for SVO and AF, caused by their higher vis-

cosity and resulting in a slightly higher injected fuel volume

per injection [23].

Based on the longer injection duration of SVO and AF

compared to diesel, we would expect a longer combustion

duration as well. This would result in a decrease in thermal

efficiency. However, the earlier start of injection increases the

ignition delay, increasing the amount of fuel that is burned

abruptly in the premixed combustion phase. This effect

shortens the total combustion period because less fuel has to

be combusted during the diffusive combustion phase. It can

Fig. 13 e Full lift period [ca] as a function of square root ofrf for injections at 1000 rpm using different fuels. The solid

data have standard deviations below or equal to 0.13ca.The others have standard deviations above 0.22ca.partially counteract the effect of the longer injection period.

3.1.6. Partial needle lift at low engine speedsAt low engine speeds, the needle lift was observed to consist

of two phases. This can be explained as follows: on a degree

cam angle based scale, the pressure buildup before needle lift

is independent of the engine speed. Once the fuel pressure

exceeds the NOP, the dynamics of the needle can be expressed

using equation (2). If the needle starts to lift, a sudden pres-

sure drop is induced. At 400rpm, this pressure drop, causesthe pressure to fall below the NOP again. Thus, during a short

period of time, the resulting pressure on the needle is directed

downwards again. The needle drops again, but does not fully

fall back on its seat. During this partial needle lift, a small

amount of fuel is injected at low speed and is probably poorly

atomized. The influences of this partial needle lift on the spray

penetration are discussed in Section 3.2.1.

Themost important parameter affecting this partial needle

lift at low engine speeds is the bulk modulus. First of all, B

affects the pressure rise before and during the needle lift and

determines the pressure at start of needle lift. Furthermore,

the pressure rise together with the viscous forces will control

the moment, the velocity and the height of the first partial lift.

For SVO and AF, an earlier start of needle lift is observed.

Furthermore, the pressure at start of needle lift is higher andFig. 14 e Needle lift at 400 rpm engine speed as a functionature (Ta) was varied between 20 C and 150 C. Fuel temper-ature was changed by varying the injector cooling

temperature and the chamber gas temperature according to

the obtained correlation.

3.2.1. Spray length3.2.1.1. Influences of the engine parameters. The influences ofengine parameters and in-cylinder conditions on the spray

penetration can be investigated by equation (5). This equation

expresses the spray penetration as a function of time S(t) in

the fully developed zone for a CR injection [24]:

St

2DPrf

d0t

svuut (5)withDP the pressure drop across the nozzle, ra the chamber gas

density and d0 the orifice diameter. The influences of chamberof ca for different fuels. Tc [ 70 C, Ta [ 140 C,Tf [ 73.6 C and Pa [ 60 bar.

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8224gas density on spray penetration were investigated by varying

chamber pressure (Fig. 16). A higher chamber gas density racauses a higher aerodynamic drag [14,25]. More momentum is

transferred from the fuel to the ambient gas, resulting in lower

penetration rates [26]. Furthermore, the influence of chamber

gas density is non-linear [17,27]. This was also observed in our

experiments: the differences in spray penetration between 20

and 40 bar are larger than the differences between 40 and

60 bar, which are again larger than the differences between 60

and 80 bar. Further increase of the pressure and density will

have less and less impact on the liquid length but can increase

the efficiency and power output of the engine.

The influences of the engine speed on the spray penetration

Fig. 15 e Needle lift at 400 rpm engine speed as a function

of ca for rapeseed oil at different fuel temperatures andchamber pressures.can also be derived from equation (5). The higher the engine

speed, the higher the pressure difference across the nozzle and

thus the higher the velocity of the fuel at the nozzle exit,

resulting in fasterspraypenetrations.Thiswasalsoobserved for

our experiments aswas already reported inRef. [1]. Injectionsat

400rpmfordiesel, RMEandRSO1are shown inFig. 17. Thespray

Fig. 16 e Influence of ambient pressure on spray

penetration. Injections with Diesel at Tf [ 60 C. Pa [ 20,40, 60 and 80 bar at engine speed set point 1000 rpm.penetration is quite different at SOI for 400rpm injectionscompared to the injections at 700 and 1000rpm. In the earlystages, the spray tip penetration is very slow, while after some

time, the spray tip velocity increases. This can be explained by

the needle lift, which was found to consist of two phases at

400rpm (see the inset of Fig. 17). During the first phase, theneedle lifts partially and a small amount of fuel is injected. Af-

terward, the needle drops again and less liquid is injected. This

results in a low spray penetration velocity. When later on the

main needle lift starts, a sudden increase in spray penetration

velocity is observed. This is clearly reflected in the spray pene-

tration progress: for RME and RSO1, a significant decrease in

penetration speed is observed, which corresponds to the small

drop of the needle after the first phase of the needle lift. For

Fig. 17 e Spray penetration for diesel, RME and KKZO as a

function of time ASOI. The inset shows the corresponding

needle lift for RME and KKZO. Engine speed [ 400 rpm,

Pa [ 60 bar and Tc [ 70 C.diesel, the partial needle lift is rather small and probably no fuel

is injected before themain needle lift.

3.2.1.2. Influences of the fuel properties. The injector temper-ature Tcwas varied between50and 90 Cand experimentsweredoneat Pa 40, 60and80bar. Results of PO injections are shownin Fig. 18. The curves corresponding to the same Pa show a very

similar penetration behavior. So, the influence of the chamber

pressure on the spray penetration is more important than the

one of the fuel temperature and thus the fuel properties. This

wasobserved for all fuels thatwere tested. The sameconclusion

was drawn when studying the effect of the engine speed: vary-

ing the engine speed affects the spray penetration more signif-

icantly than varying the fuel temperature.

The influences of the fuel properties are further investigated

in Fig. 19: the penetration length for injectionswithDiesel, RME,

RSO1 and PO at 80 bar, 1000rpm and Tf 65 C are compared.During the early stages of spray formation, the penetration is

quite similar for all fuels. This observation differs from most

results in literature for CR systems [14,19,28], which report a

decreased penetration for bio-fuels at the early stages of the

injection. However, this can be attributed to the PLN system.

Injection pressures at the needle lift are higher for SVO due to

their higher bulk modulus, so that the outlet velocity and thus

coefficient of the droplet. First of all, this equation shows that

the higher density of SVO causes a decrease in droplet decel-

eration, causing the longer spray penetrations. Secondly, the

droplet sizes of SVO are accepted to be larger than those of

diesel and biodiesel [6]. Larger droplets experience less

deceleration and result in longer spray penetrations. The

larger droplet sizes are mostly attributed to their higher vis-

cosity [14], surface tension [13] or both [6]. These droplet size

studies were all conducted for CR systems. For PLN systems,

the higher injection pressures of the SVO will also decrease

droplet sizes [29]. This effect only partly counteracts the in-

fluences of viscosity and surface tension on droplet sizes.

We can conclude that during the early stages of the injec-

tion, the reduction in spray outlet velocity due to the high

viscosity and density of the SVO and AF is counteracted by the

increase in injection pressure at SOI due to the higher bulk

moduli of SVO and AF. In the fully developed zone, the higher

Fig. 18 e Influence of the fuel temperature and ambient

pressure on the spray penetration. Injections with PPALM

at Tc [ 50 C (red curves), 70 C (blue curves) and 90 C(black curves). Pa [ 40 (longest penetration length), 60

(curves in the middle) and 80 bar (shortest penetration

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8 225penetration increases. This counteracts the increased friction

due to the higher viscosity and the lower outlet velocity due to

the higher fuel density (Bernoullis law). Together this results in

similar spray penetrations at SOI for all fuels tested. In a later

stage, the spray tip penetrations of PO, RSO1 and less clearly AF,

are faster than those of Diesel and RME. A first explanation is

the higher injectionpressures for SVOandAFwhen using a PLN

system, due to their higher bulk modulus. However, the longer

spray penetrations in the fully developed zone for SVO

compared to diesel were also observed for CR systems

[14,19,28]. So there should be other explanations as well. These

can be derived from the following equation expressing the

aerodynamic drag on a liquid droplet:

length). (For interpretation of the references to colour in

this figure legend, the reader is referred to the web version

of this article.)a 83rairrfuel

Cw1rv2 (6)

where r is the radius of the droplet, v the speed of the droplet

relative to the surrounding gas, a the acceleration of the

droplet, rf the density of the fuel droplet and Cw the drag

Fig. 19 e Influences of fuel type on the spray penetration.

Injections with Diesel, RME, KKZO, PPALM and AF are

compared. Pa [ 60 bar, Tc [ 60 C.injection pressure and the larger SVO and AF droplets result in

a longer penetration length.

3.2.2. Spray angle3.2.2.1. Influence of the engine parameters. The spray angle isdetermined by the entrainment rate of ambient gas into the

spray. Because of the velocity difference between the fuel and

the gas, gas will entrain into the fuel spray while the fuel

penetrates along the spray axis. Due to the ambient gas, the

droplets will experience drag forces. This drag corresponds to

a momentum transfer from the droplets to the ambient gas,

decelerating the droplets [14]. Thus, a higher chamber gas

density causes an increased deceleration of the droplets and

as a result, faster droplets will push the slower ones outside

more rapidly. This stronger dispersion of the droplets results

in wider spray angles. This is expressed by different authors

[24,30] using the relationship tan(q)wrax, with x between 0.17

and 0.20 in most publications. In Fig. 20, measurements with

diesel, RME and RSO1 at Tc 60 C are shown for Pa 40 barand Pa 80 bar. In these figures, an increase of approximately5 was observed between 40 and 80 bar. We found x to bearound 0.19, which is in close agreement with the results of

Desantes et al. [30] (x 0.17). No significant difference wasfound among the different fuels.

Fig. 20 e Influence of ambient pressure on the spray angle.

Injections with diesel, RME and KKZO. Tc [ 60 C, Pa [ 40

and 80 bar, engine speed set point 1000 rpm. Moving

average over 500 ms, equal to 5 spray images.

probably results in a significant reduction in droplet sizes and

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8226Furthermore, when comparing the spray structures, dif-

ferences between the sprays at 40 and 80 bar were observed

(Fig. 10). The outer ranges of the sprays at low chamber

pressure appear to be less dense. On the two dimensional

spray images, the spray is surrounded by a more clear region.

This was observed most clearly for injections with RSO1 and

will be discussed at the end of the next section.

The influences of the engine speed on the spray angle were

investigated by comparing spray angles at engine speed set-

points of 400, 700 and 1000rpm. When varying the injectionspeed between 700 and 1000rpm, we did not observe anydifferences in the steady state spray angle. Similar steady

state spray angles for different injection pressures are also

reported in literature. Takahashi et al. [31] showed that once

the injection pressures is sufficiently high, the spray angle no

longer depends on the injection pressure. Next, Desantes et al.

[30] stated that at high injection pressures, a further increase

in injection pressure can affect the spray angle through cavi-

tation only. Although the steady state spray angle is similar,

slightly higher spray angles were sometimes observed at SOI

for higher engine speeds. Takahashi et al. [31] and Galle et al.

[1] mentioned that for relatively low injection pressures, the

spray angle increases with increasing injection pressures. So,

because the injection pressure shortly ASOI is higher for in-

jections at 1000rpm, this can be a first explanation. However,differences in cavitation regime can also influence the spray

angle ASOI.

For 400rpm injections, mostly a steady state spray anglewas reached similar to the 700 and 1000rpm injections.However, at high chamber pressures, sometimes the spray

angle increased as a function of time during the whole injec-

tion event. In these cases, the spray diverged strongly from

the theoretical conical spray shape.

3.2.2.2. Influences of the fuel properties. The standard devia-tion of the spray angle (1e3) is mostly larger than differencesbetween experiments with different fuels or between experi-

ments at different fuel temperatures. So, the results that are

discussed here are trends that were discernible, but they have

to be verified by conducting more identical experiments for

every set of conditions. In Fig. 20, spray angles of different

fuels are very comparable. Varying the fuel temperature did

not alter the spray angle for diesel and RME. However, for PO

and AF, increasing the fuel temperature often resulted in an

increase in spray angle. The higher spray angles at higher fuel

temperatures observed for the SVO could be related to the

decrease in viscosity, density and surface tension. Larger (due

to high viscosity and/or high surface tension) and heavier (due

to high fuel density) droplets, have more momentum and

experience less deceleration (equation (6)). This reduces the

chance to get pushed outside, decreasing the spray angle.

However, this does not explain why spray angles between

diesel and SVO are similar. In general, the spray angle seems

to be altered only slightly by fuel properties for the PLN in-

jection system.

The spray angle is calculated by image processing soft-

ware. However, similar spray angles do not mean that there

are no structural differences betweendifferent fuel sprays. Forfuel temperatures above 45 C, no clearly visible structuraldifferences were observed between sprays of different fuels.explains why no clear region was observed around the spray

anymore. If the clear region around the spray was observed at

higher fuel temperatures, it was at 40 bar. Based on the pre-

vious explanation about the droplet sizes, this could be related

to a higher SMD at lower chamber pressures. However, the

influences of chamber pressure on SMD are not generally

accepted in literature [29], so this is only a hypothesis. Finally,

we observed that the clear region around the spray was more

explicit at lower engine speeds. Again, this can be related to

the higher SMD at lower engine speeds due to the lower in-

jection pressures [29]. Experimental investigations of SMD are

necessary to provide more insights into these phenomena.

4. Conclusions

In this paper, spray measurements were conducted in a con-

stant volume combustion chamber in inert non-evaporative

atmospheres at elevated pressures (up to 80 bar). The

research focuses on medium speed diesel engines and a

pump-line-nozzle injection system is used. Fuels investigated

were diesel, rapeseed biodiesel (RME), rapeseed oil (RSO), palm

oil (PO) and animal fats (AF). The influences of fuel properties,

engine speed and chamber pressure on the injection pressure

profile, needle lift, spray length and spray angle were inves-

tigated. The main conclusions of this paper are as follows:

Fuel properties are strongly temperature dependent. Even

small changes in fuel temperature result in a significant in-

fluence on the injection pressure of a pump-line-nozzle sys-

tem. Therefore, fuel temperature at the moment of injection

was determined and controlled for all experiments.

- The bulk modulus, a measure for the resistance to

compression, is the main parameter affecting the injection

pressure profile.Galle et al. [1] reported a strong structural difference for RSO in

comparison with diesel and RME. RSO droplets are larger than

those of diesel and thus for a same volume of fuel, they take

less space on the two dimensional spray images than more

and smaller droplets. Due to the large droplets, some light can

pass through the outer regions of the spray cone and there-

fore, on the two dimensional spray images, RSO is surrounded

by a more clear area. For diesel and RME sprays, the large

amount of small droplets do not allow light to pass through

the spray, resulting in very homogeneous two dimensional

spray images. It is important to note that the experiments of

Galle et al. [1] were conducted at Tc 20 C. We also observedthis more clear area for our RSO1 experiments at 27 C. How-ever, when fuel temperaturewas raised to 45 C, the clear areasurrounding the spray was much less observed for RSO1(Fig. 10) (except for some experiments at low chamber pres-

sure, 40 bar). Based on the reasoning of Galle et al. [1], this

means that increasing the fuel temperature from 25 to 45 Cresults in a significant improvement of the atomization of

rapeseed oil. This can be attributed to the decrease in vis-

cosity: in the range from 25 to 45 C, viscosity of rapeseed oildecreases from 60 to less than 30 mPas (cfr. Fig. 10). This- The higher bulk modulus of straight vegetable oil and ani-

mal fat results in an earlier and faster needle lift for

- The higher density of biofuels results in longer injection

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8 227durations for pump-line-nozzle injection systems.

- The spray length is strongly influenced by the engine speed

(or thus injection pressure) and the chamber gas density.

These parameters affect the spray penetration more

significantly than the fuel temperature or the fuel type.

- At start of injection, spray penetration is similar for all fuels

because of counteracting influences of bulk modulus, vis-

cosity and density. In the later stage of the injection, how-

ever, straight vegetable oil and animal fat show slightly

faster spray penetrations due to the higher injection pres-

sures and larger droplets of these fuels compared to diesel

and rapeseed biodiesel.

- The spray angle is not significantly influenced by engine

speed. Chamber gas density is the dominant parameter

affecting the spray angle. Similar spray angles were

observed for different fuels. However, an influence of fuel

temperature was often observed for the more viscous fuels.

- At low temperatures, despite similar spray angles, the spray

atomization for straight vegetable oil is clearly worse than

diesel and rapeseed biodiesel. Increasing fuel temperature

from 25 to 45 C already enhanced the spray atomizationsignificantly. This is most probably due to the strong

decrease in viscosity. Research on droplet sizes and droplet

distribution is necessary to quantify these influences of fuel

properties on atomization. Probably, the sauter main

diameter is the spray parameter that is most significantly

affected by the fuel properties.

Acknowledgments

The authors of this paper would like to acknowledge the sug-

gestions and technical assistance of Koen Chielens and Patrick

De Pue. The authors would like to thank Prof. Dr. ir. P. Van der

Meeren for the use of the lab facilities of the Department of

Applied Analytical and Physical Chemistry at Ghent University,

used in determining the fuel properties. J. Galle thanks the

Institute for the Promotion of Innovation through Science and

Technology in Flanders (IWT-Vlaanderen) for his Ph.D. grant

(SB-81139). The experimental equipment is financially sup-

ported by Anglo Belgian Corporation (ABC), the Ghent Univer-

sity Special Research Fund and the Institute for the Promotion

of Innovation through Science and Technology in Flanders

(IWT-Vlaanderen) within the R&D project Research and

development of a medium speed highly efficient internal

combustion engine with ultralow emissions, for diesel, dual

fuel, heavy fuel oil and alternative fuels (IWT110579).

r e f e r e n c e s

[1] Galle J, Demuynck J, Vancoillie J, Verhelst S. Spray parametercomparison between diesel and vegetable oils for non-experiments at high fuel temperatures (>60 C). In the lowertemperature range this effect is counteracted by the higher

viscosity of straight vegetable oil and animal fat: viscous

forces delay the needle lift.evaporating conditions. SAE (2012-01-0461) 2012.[2] Allocca L, Mancaruso E, Montanaro A, Sequino L, Vaglieco B.Bio- and mineral-fuel spray evolution in non-evaporatingand evaporating conditions by image processing techniques.Europe: ILASS; 2010.

[3] Jeong K, Park S, Lee C. Injection spray and emissionscharacteristics in a diesel engine using diesel-bioethanolblended fuel. Europe: ILASS; 2011.

[4] Heilig A, Kaiser M, Dinkelacker F. Spray analysis andcomparison of diesel and biodiesel-methanol blends. Europe:ILASS; 2011.

[5] Caresana F. Impact of biodiesel bulk modulus on injectionpressure and injection timing. The effect of residualpressure. Fuel 2010;90(2):477e85.

[6] No S-Y. Spray characteristics in CI engines fuelled withvegetable oils and its derivatives. In: 23rd Annual conferenceon liquid atomization and spray systems. Europe: ILASS; 2010.

[7] Battistoni M, Grimaldi C. Analysis of transient cavitatingflows in diesel injectors using diesel and biodiesel fuels. SAEInt (2010-01-2245).

[8] Arcoumanis C, Flora H, Gavaises M, Badami M. Cavitation inreal-size multi648 hole diesel injector nozzles. SAE Int (2000-01-1249) 2000.

[9] Han J-S, Lu P-H, Xie X-B, Lai M-C, Henein NA. Investigation ofdiesel spray primary break-up and development for differentnozzle geometries. SAE Int (2002-01-2775) 2002.

[10] Martynov S. Numerical simulation of the cavitation processin diesel fuel injectors. Ph.D. thesis; 2005. University ofBrighton.

[11] Wahlin F. Experimental investigation of impinging dieselsprays for HCCI combustion. Ph.D. thesis. Stockholm (: RoyalInstitute of Technology; 2007.

[12] Salvador FJ, Martinez-Lopez J, Romero JV, Rosello MD.Influence of biofuels on the internal flow in diesel injectornozzles. Math Comput Model 2011;54(7e8):1699e705.

[13] Reddemann MA, Mathieu F, Martin D, Kneer R. The influenceof fuel properties on spray propagation, atomization andevaporation. In: 23rd Annual conference on liquidatomization and spray systems. Europe: ILASS; 2010.

[14] Dernotte J, Hespel C, Foucher F, Houille S, Mounaim-Rousselle C. Experimental study of the influence of fuelphysical properties on the diesel injection process in non-vaporizing conditions. In: 24th European Conference onLiquid Atomization and Spray Systems, Estoril, Portugal.Europe: ILASS; 2011.

[15] Settles G. Schlieren and shadowgraph techniques 2006.[16] More J. The LevenbergeMarquardt algorithm:

implementation and theory. In: Proceedings of the 1977Dundee conference on numerical analysis 1970. p. 630e68.

[17] Siebers DL. ALiquid-phase fuel penetration in diesel sprays.SAE Int (980809) 1998.

[18] Steven S-J. Power transformer partial discharge (PD) acousticsignal detection using fiber sensors and wavelet analysis,modeling, and simulation. Masters thesis; 2006. Virginia Tech.

[19] Payri R, Salvador FJ, Gimeno J, Bracho G. E_ect of fuelproperties on diesel spray development in extreme coldconditions. Proc Inst Mech Eng Part D J Automob Eng2008;222(D9):1743e53.

[20] Deshmukh D, Mohan AM, Anand TNC, Ravikrishna RV. Highpressure spray characterization of vegetable oils. In: 23rdAnnual conference on liquid atomization and spray systems,Brno, Czech Republic. Europe: ILASS; 2010.

[21] Bari S, Yu CW, Lim TH. E_ect of fuel injection timing withwaste cooking oil as a fuel in a direct injection diesel engine.J Automob Eng 2004;218:93e104.

[22] Ma Z, Huang Z, Li C, Wang X, Miao H. Effects of fuel injectiontiming on combustion and emission characteristics of adiesel engine fueled with diesel propane blends. Energy &

Fuels 2007;21:1504e10.

[23] Hassaneen AE. Injection characteristics of rapeseed methylester versus diesel fuel in pump-line-nozzle injectionsystem. SAE Int (2008-01-1590) 2008.

[24] Naber JD, Siebers DL. Effects of gas density and vaporizationon penetration and dispersion of diesel sprays. SAE Int(960034) 1996.

[25] Morgan R, Wray J. The influence of injector parameterson the formation and break-up of a diesel spray. SAE Int(2001-01-0529) 2001.

[26] Siebers DL. Scaling liquid-phase fuel penetration in dieselsprays based on mixing-limited vaporization. SAE Int (1999-01-0528) 1999.

[27] Espey C, Dec JE. The e_ect of TDC temperature and density onthe liquid phase fuel penetration in a D.I. diesel engine.Trans. SAE 104 1995;Sect. 4:1400e14.

[28] Gostic I, Senghaas C, Willmann M, Wensing M. Howbiodiesel/diesel fuel blends affect spray breakup andemissions. In: 23rd Annual conference on liquid atomizationand spray systems. Europe: ILASS; 2010.

[29] Chi Y, Kim E. Measurement of droplet size distribution oftransient diesel spray. SAE Int (931949) 1993.

[30] Desantes JM, Payri R, Salvador FJ, De la Morena J.Influence of cavitation phenomenon on primary break-upand spray behavior at stationary conditions. Fuel2010;89(10):3033e41. http://dx.doi.org/10.1016/j.fuel.2010.06.004.

[31] Takahashi H, Tomaru K, Shiga S, Karasawa T, Kurabayashi T.Characteristics of diesel spray with unsteady and higherinjection pressure using a rapid compression machine. SAEInt (910226) 1991.

b i om a s s a n d b i o e n e r g y 5 7 ( 2 0 1 3 ) 2 1 5e2 2 8228

Experimental investigation concerning the influence of fuel type and properties on the injection and atomization of liquid ...1 Introduction2 Experimental method2.1 Experimental setup and data acquisition2.2 Tested fuels2.3 Fuel temperature control2.4 Measurements and standard deviation2.4.1 Injection pressure profile and needle displacement2.4.2 Spray images

3 Results and discussion3.1 Injection pressure profile and needle displacement3.1.1 General view on the injection pressure profile3.1.2 Pressure build-up3.1.3 Needle lift3.1.4 Pressure drop caused by needle lift3.1.5 Injection period3.1.6 Partial needle lift at low engine speeds

3.2 Spray development3.2.1 Spray length3.2.1.1 Influences of the engine parameters3.2.1.2 Influences of the fuel properties

3.2.2 Spray angle3.2.2.1 Influence of the engine parameters3.2.2.2 Influences of the fuel properties

4 ConclusionsAcknowledgmentsReferences