lable at ScienceDirect

Energy 77 (2014) 783e790

Contents lists avai

Energy

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

Evaluation of RME (rapeseed methyl ester) and mineral diesel fuelsbehaviour in quiescent vessel and EURO 5 engine

Luigi Allocca, Ezio Mancaruso*, Alessandro Montanaro, Luigi Sequino,Bianca Maria VagliecoIstituto Motori e CNR, Napoli, Italy

a r t i c l e i n f o

Article history:Received 26 May 2014Received in revised form15 September 2014Accepted 19 September 2014Available online 16 October 2014

Keywords:BiofuelInjection processDiesel enginePollutant emissionOptical diagnostic

* Corresponding author. Tel.: þ39 081 7177187; faxE-mail addresses: l.allocca@im.cnr.it (L. Allo

(E. Mancaruso), a.montanaro@im.cnr.it (A. Mon(L. Sequino), b.m.vaglieco@im.cnr.it (B.M. Vaglieco).

http://dx.doi.org/10.1016/j.energy.2014.09.0500360-5442/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

Alternative diesel fuels for internal combustion engines have grown significantly in interest in the lastdecade. This is due to the potential benefits in pollutant emissions and particulate matter reduction.Nevertheless at possible increase in nitrogen oxide (NOx), and almost certainly increase of fuel con-sumption have been observed.

In this paper, mineral diesel and RME (rapeseed methyl ester) fuels have been characterized in a non-evaporative spray chamber and in an optically-accessible single-cylinder engine using a Common Railinjector (8 holes, 148� cone opening angle and 480 cc/30s@10 MPa flow number) to measure the spatialfuel distribution, the temporal evolution and the vaporizationecombustion processes. The injectionprocess and mixture formation have been investigated at the Urban Driving Cycle ECE R15: 1500 rpm at0.2 MPa of break mean effective pressure.

Characteristic parameters of the spray like penetration length and liquid fuel distribution have beenanalysed and they have been correlated with the exhaust gaseous and particulate matter emissions.

In the spray-analysis in non-evaporative conditions, short events (pilot) are mostly affected byasymmetries in the fuel distributions with noticeable standard deviations at low injected quantities. Inthe engine tests, the jets reached immediately the stabilization. A comparative analysis on the liquidphase of the spray, in non-evaporative and evaporative conditions, has permitted to investigate betterthe mixture formation. Its effect on pollutant emissions has been analysed for both fuels.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

In the last two decades the diesel engine has met the increasingdemand of economy/performances of the powertrain conjugatedwith the legislative requirements of exhaust emission reduction,particularly focused on nitrogen oxides (NOx) and PM (particulatematter). These results have been achieved thanks to the substantialprogresses in the fuel injection equipments (Common Rail) andcombustionmanaging that make large use of the electronic control.First (FAME (fatty acid methyl ester)) and second (Fischer-Tropsch)generation of alternative diesel fuels comply this challenge withoutany modification with the powertrain, coupling these potentialemission reductions to the advantage of biodegradability/non-toxicity of the fuel and the global benefits on CO2 cycle due to

: þ39 081 2396097.cca), e.mancaruso@im.cnr.ittanaro), l.sequino@im.cnr.it

renewable fonts [1e3]. These fuels are generally referred as bio-diesel for diesel engine applications and regulation permits mix-tures up to 20% in volume to mineral diesel fuel.

However, some differences appear in the chemical-physicalcharacteristics of biodiesels, with respect to the mineral dieselfuel, that affect the air-fuel mixture preparation and the combus-tion in the engine. The spray characteristics have been widelyanalysed to study the influence of injection pressure and cylinderbackpressure on the fuel penetration [4e8]; moreover attempts indetermining the fuel-bulk and droplet fragmentation have beencarried out [9]. Some differences in the fuel injection rate have beenfound resulting biodiesel quantities lower than mineral fuels. Thishas been related to the different density and viscosity of the fluids[10]. Finally, effects of biodiesels on the injection apparatus havebeen widely investigated to associate the deposit formation in theinjector system to the quality and composition of the fuel [11,12].These sediments have been observed inside the injector body, onthe piston, on nozzle needle but, especially, in the spray-holesresulting in a reduced flow and dispersion modification of the

Table 1Physical-chemical properties of diesel and RME fuels.

Feature Method Diesel RME

Density @ 15 �C [kg/m3] EN ISO 12185 840.1 883Viscosity @ 40 �C [mm2/s] EN ISO 3104 3.141 4.254Oxidation stability [g/m3] EN ISO 12205 0.1 0.8Lubricity @ 60 �C [micron] EN ISO 12156-1 e 188Cetane number EN ISO 5165 51.8 52.3Low heating value [MJ/kg] ASTM D3338 43.1 37.34Distillation [�C] IBP 184.8 322[�C] 10% vol. 221.9 33.2[�C] 50% vol. 276.1 337[�C] 90% vol. 329.1 343.3[�C] 95% vol. 344.9 347[�C] FBP 358.3 360Carbon [%, m/m] ASTM D5291 86.5 78.5Hydrogen [%, mm] ASTM D5291 13.5 10.8Nitrogen [%, m/m] ASTM D5291 e 0.2Oxygen [%, m/m] ASTM D5291 e 10.5

L. Allocca et al. / Energy 77 (2014) 783e790784

injected fuel. These lead to loss of power and worst air-fuel mixturepreparation with strong consequences on the pollutant emission[13,14].

The effects on the combustion process of pure biodiesels andtheir blends with mineral diesel at different percentages have beeninvestigated both in quiescent vessels and in optically accessibleengines. High-pressure, high-temperature test rigs simulated thediesel engine conditions at the injection time and they have beenused to study the behaviour of the liquid and vapour phases and thefuels properties effects [15e17]. In particular, the continuouspenetration of the liquid phase has been related to the density,viscosity, and surface tension; while the droplet parts are mainlyinfluenced by the fuel volatility [15]. Investigations on the pene-tration of the liquid- and vapour-phases inside diesel engines havebeen carried out on pure biodiesel and blends at diverse percent-ages [16] as well as for single components fuels [17] where a rela-tionship between chemical-physical properties and spray evolutionhas been attempted.

In this paper, mineral diesel and RME (rapeseed methyl ester)fuels have been characterized both in a quiescent non-evaporativerig and in an optically-accessible single-cylinder engine. Apilotþmain strategy has been exploredwith a slight increase of themain duration for the RME condition to compensate its lowerheating value. The Urban Driving Cycle ECE R15: 1500 rpm at0.2 MPa of BMEP has been simulated. A comparison between theliquid spatial distributions of the 8 jets has been performed in orderto extract information about the air-fuel mixture and the pollutantemissions.

2. Test conditions and procedures

The injection process characterization has been carried out un-der both non-evaporative and evaporative environments to analysethe effects that the fluid density and viscosity of biodiesels-dieselfuels have on spray formation, mixing, and pollutants emissions.The delivering fuel rate and the spatial and temporal distribution ofthe liquid for non-evaporative conditions have been carried out in atest rig composed of aflow ratemeter and a quiescent high-pressureoptically accessible vessel, respectively. The evaporative phase hasbeen studied in an optically accessible single-cylinder diesel engineworking in real engine-like conditions.

2.1. Non-evaporative conditions

Diesel and RME fuels have been used to characterize the injec-tion process in non-evaporative quiescent vessel. Fuel injectionrates as well as spatial and temporal distribution of the fluid havebeen measured. The main physical/chemical characteristics of thetwo fuels are reported in Table 1.

The parameters mainly affecting the injection and spray for-mation are the density, viscosity and distillation curve. It isworthwhile to note that the RME has higher density and viscosityvalue than diesel fuel, parameters that influence the spray behav-iour. The tests have been performed on a Bosch second generationcommon rail solenoid-driven fuel injection system by using a Boschminisac-type injector (8� 480 cc/30sx148�). The nozzle diameter is0.136 mm and the length/diameter ratio is 5.29. The injectionstrategy is the same used in the single-cylinder engine and it hasbeen taken from the calibration of real engine. It consists of a trainof pilotþmain injections at the injection pressure of 62.0 MPa. Thisis a typical Euro 5 engine condition at 1500 rpm and 0.2 MPa ofBMEP (break mean effective pressure) and it is one of the mostfrequent point of the Urban Driving Cycle ECE R15 cycle. Theoptimization of this operating point on the real engine for biodieselhas produced an increase of the main pulse duration due to its

lower energy content with respect to the diesel [18]. To achieveequivalent power output, the main pulse of RME is 20 ms longerthan the diesel one. A PECU (programmable ECU) has managed theinjection apparatus enabling to set the strategies in terms of pulsenumber and timing. Fast electronic drivers in the PECU allowedsetting precise and stable injections for large as well as small fuelquantities like pilot ones. The high pressure pump supplying thefuel has been driven by a variable speed electric motor while a heatexchange system on the hydraulic circuit has been adopted to keepconstant the fuel temperature in the tank (40 ± 1 �C). Further de-tails as the sketch of the experimental apparatus, the injection flowrate bench and the high pressure vessel are reported in Ref. [19].

The global behaviour of the fuels has been studied by an AVLinjection meter for collecting the fuel injection rates and theanalysis of the spray images collected by a flash/CCD (chargecoupled device) system synchronized with the injection command.More details of the technique and procedures are reported inRef. [19].

2.1.1. Fuel injection rateFuel injection rates have been measured by an AVL meter

working on the “Bosch pipeline” principle [20]. The pressure in-crease, produced by the fuel injected through the nozzle in anadapted chamber, is registered by a GM12D e AVL piezoquartztransducer. The signal is proportional to the fuel rate through arelationship with geometrical parameters of the device andchemical-physical properties of the fluid. The time resolution of thefuel rate is less than 1 ms. The measured injected fuel has beencompared with the weight one collected at the Bosch tubedischarge and measured by a precision balance. The experimentalsetup is completed by electronic devices for managing the in-jections both in terms of starting and synchronization with theacquiring instruments. The data storage is a four channels oscillo-scope Tektronix TDS 684 B,1 GHz bandwidth. Further details on theinjection rate measures procedure have been reported in Ref. [19].

2.1.2. Spray evolution setupThe spatial and temporal evolution of the injected fuel has been

studied processing the images of the sprays captured at differenttime from the SOI (start of injection) in an optically-accessiblequiescent vessel at ambient temperature and filled with gas atdensities typical of the engine at the injection time. Taking intoaccount that the main controlling factor of the sprays developmentin non-evaporative conditions is the density of the gas [21], SF6(sulphur hexafluoride) (density 6.2 kg/m3) has been used for safetyreasons permitting to reach the desired densities at pressures lower

L. Allocca et al. / Energy 77 (2014) 783e790 785

than the air. A pressure of 0.30 MPa has been set to achieve thedensity of 18.6 kg/m3. The images of evolving jets have beencaptured by a CCD camera (1376 x 1040 pixels, 12 bit, 0.5 ms shuttertime) synchronized both with the injection system and a high in-tensity flash. A 50 mm focal lens has been used to collect the im-ages, realizing a spatial resolution of 10.10 pixel/mm. One image percycle has been caught illuminating the spray at different time fromthe SOI by a digital delay generator. Each image had a dedicatedpulsed flash with repetitive light intensity while the acquisitiontime has been controlled by the CCD shutter (0.5 ms). Five imageshave been acquired at each time-step and at the same conditions.The stability and uniformity of the evolving sprays made thisrepetition condition confident with the need of dispersion bars onthe measuring parameters. The fuel has been injected in thechamber in single-shot mode operation for preventing droplet fogand transparent windows dirtying.

Analysis of the liquid fuel spray images has been carried out byimage processing procedures: image acquisition, background sub-traction, filtering matrixes, edges determination, and tip penetra-tion. A low duty cycle of injections has been set to obtain a uniformbackground level for all the images. A computerized methodologyof image processing has been applied to the acquired images. Moredetails and specifications on the procedure have been reported inRef. [19].

2.2. Evaporative conditions

The analysis of the behaviour of the fuels in engine-like condi-tions has been investigated in a single cylinder research enginewith optical access to the combustion chamber through the pistonhead. Moreover, the combustion products in terms of gas andparticulate matter concentrations have been measured at theexhaust pipe.

2.2.1. Engine setupThe study of the spatial and temporal evolution of the fuel jets

has been carried out in an optically accessible diesel enginerepresentative of an evaporative system. It is a single-cylinder en-gine equipped with cylinder head and common rail injection sys-tem derived from a four-cylinder, 16 valves, 2.0 L and Euro 5production engine (Fig. 1).

The engine stroke and bore are 92 mm and 85 mm, respectively,and the compression ratio is 16.5:1. An external air compressor hasbeen used to supply pressurized intake air in order to obtain thesame inecylinder conditions of the real multi-cylinder engine andto compensate lower compression ratio typical of the optical

Fig. 1. Transparent engine.

engines. Moreover, the intake air, before reaching the intakemanifold, is filtered, dehumidified, and preheated. Thanks to thedehumidification, the measure of the air mass flow from the vol-ume flow, the pressure and the temperature is not affected by theuncertainties related to the presence of water. A reliable measure ofthe mass flow before the mixing with the recirculated gases allowscalculating the effective EGR (exhaust gas recirculation) systempercentage.

A VSA (variable swirl actuator) system has been employed inorder to manage the air swirl motion in the intake manifold. In theexhaust pipe, the presence of a pressure valves permits the recir-culation of the right amount of burned gases through the cooledEGR (exhaust gas recirculation system). Finally, the engine has beenequipped with a second-generation CR (common rail) system. Theinjector has been controlled by a fully flexible ECU (electroniccontrol unit) for the combustion optimization. In order to provide afull view of the combustion bowl a 46 mm diameter flat windowhas been fitted in the piston head and an appropriate 45� visiblefixed mirror has been set inside the extended piston. The windowhas been realized with UV-grade fused silica. In order to analyse theinjection signals, a Hall-effect sensor has been applied to the line ofthe solenoid current. Moreover, to acquire the inecylinder pressurein motored and fired condition, a piezoelectric pressure transducer(GH13P-AVL) has been set in the glow plug seat of the engine head.For each operating condition investigated, the cylinder pressureand the drive injector current have been digitalized and recorded at0.1 cad (crank angle degree) increments and ensemble-averagedover 150 consecutive combustion cycles. At motored condition,temperature and density at TDC (top dead centre) have been esti-mated assuming a polytropic coefficient of 1.36 [22]. Consideringthat the intake air temperature and pressure have been set to 317 Kand 0.11 MPa, respectively; the inecylinder temperature and den-sity have been 818 K and 18.6 kg/m3, respectively.

Digital imaging analysis has been performed by a CCD camera.The CCD camera (640 x 480 pixels) had a high sensitivity over awide visible range and it has been equippedwith a 55mmobjectiveF/3.5 realizing a spatial resolution of 10 pixel/mm. Images havebeen acquired with an exposure time of 0.055 ms corresponding to0.5 cad at 1500 rpm. Two external highly luminous CW (continuouswavelength) halogen lamps have been used to light the combustionchamber during the injection process. More details of the engineand the experimental set up are reported in previous paper [23].The images of the spray acquired in the combustion chamber havebeen automatically elaborated via software. The background hasbeen subtracted at each image of the jet recorded before the start ofthe main injection. The resulting image has been converted in bi-nary one, and the outliers in terms both of the dark and brightpixels have been removed. Finally, the image has been divided in 8sections, one for each hole, the edge of the jets detected and thespray parameters measured.

The engine operating condition has been extracted from the ECUmap of a commercial Euro 5 engine; it represents the condition at1500 rpm at 0.2 MPa of brake mean effective pressure (BMEP). Theimplemented injection strategy and the injector are the same usedin the previous non-evaporative tests. The engine runs in contin-uous mode. In Table 2, the engine operating conditions for thediesel and RME are reported.

Table 2Engine operating conditions.

Engine speed Fuel SOIPilot

ETPilot

DwellTime

SOIMain

ETMain

Pinj

[rpm] [cad] [ms] [ms] [cad] [ms] [MPa]1500 Diesel �15 274 1139 �4.5 532 62.01500 RME �15 274 1139 �4.5 552 62.0

Fig. 2. Spray distribution in the transparent engine (left) and in non-evaporative system (right).

L. Allocca et al. / Energy 77 (2014) 783e790786

2.2.2. Emission measurementsThe composition of the combustion products has been investi-

gated bymeans of exhaust gas analysers. The exhaust pipe has beeninstrumented with a series of probes for the sampling of the flow; adistance of 600 mm between the probes has been chosen in orderto prevent mutual influence. The most relevant emissions of adiesel engine have been measured. In particular, nitrogen oxides(NOx) have been measured by means of electrochemical sensors,carbon monoxide (CO) and dioxide (CO2), and hydrocarbon (HC)concentrations have been measured by non-dispersive infrareddetectors. An opacimeter has been used to measure the opacity ofthe particulate matter content (PM) in the exhaust gas. It consists ofa control volume where the light emitted by the lamps placed onone side is detected by a sensor on the opposite and attenuated bythe flowing gas. Then, the transmission coefficient is calculatedwith respect to the zero level. The opacity percentage obtained hasbeen converted in particulate mass concentration by an empiricalrelation [24].

It is important to note that, even if the maximum care forreproducing multi-cylinder operating conditions has beenemployed, absolute data cannot be compared between opticalsingle-cylinder and multi-cylinder applications, chiefly because ofthe effects induced by the special piston with optical access.However, the authors verified that in terms of relative trendsamong pollutant concentrations for the different tested fuels, aconsistent agreement can be found between the two engine ap-plications [18,25].

2.3. Spray analysis

Images of both pilot andmain injections have been acquired andsaved in 8-bit format, with 256 grey scale intensities representingthe several luminous levels. The jets shapes (Fig. 2) have been cutout and the chamber brightness would not affect the luminosityreflected and scattered by the jet. Images have been processed bydrawing a contour plot to delimitate the jet area. Moreover, themaximum penetration of each jet has beenmeasured and analysed.Five image repetitions at the same crank angle have been per-formed to reduce the uncertainty due to the engine cycle variation.The engine coefficient of variation has been measured and itresulted less than 1% with respect to the maximum pressure andless than 3% with respect to the indicated mean effective pressure.In Fig. 2 jets have been numbered in order to compare the spraythat emerges by the same nozzle hole.

Table 3Injection timing and injected fuel amount from pilot and main pulses for diesel andRME.

pinj[MPa]

tinj [ms]Pilot-dwell-Main

Pilot[mm3/str]

Main[mm3/str]

Total[mm3/str]

Diesel 62.0 274-865-532 1.1 8.14 9.24RME 62.0 274-865-552 0.93 8.82 9.75

3. Results and discussion

3.1. Non-evaporative spray characterization

The characterization in non-evaporative conditions of the fuelsprays has consisted in acquiring parameters by image analysis ofspatial/temporal evolution of the fuel at the investigated condi-tions. Preliminary measures of fuel injection rate have been made.In Table 3, the injection strategies and the fuel amount injected perstroke from pilot andmain pulses are reported for both investigatedfuels. The obtained quantity of injected fuel has been comparedwith the weighted ones at the Bosch tube discharge using a pre-cision balance. They have given a discrepancy less than 2%.

The total amount of RME is greater than diesel fuel in agreementwith the longest duration of main energizing time. Fig. 3 shows thefuel injection rate profiles for the investigated injection strategies.The injected quantities are the integrals of each profile along thepulse durations. The signals are quite in good accordance for thepilot injections and the early main so the different physical-chemical properties of the fuels do not produce substantial differ-ences for both pulses. The results agree with those reported inRefs. [5,26,27]. The last part of main injections shows a slight dif-ference due to the diverse duration set. Moreover, a delay betweenthe start of solenoid energizing current and the first exit of the fuelfrom the nozzle of 0.37 ms is registered due to the inertias of theservo and main stages of the injector. This delay is constant andindependent from the fuel characteristics.

The spatial and temporal evolution of the sprays has been car-ried out in high-pressure optically accessible vessel both for pilotand main injection events. Fig. 4 reports, for the main event, the

Fig. 3. Fuel injection rate for the diesel and RME fuels at 62.0 MPa of injectionpressure.

Fig. 4. Spray sequence at different time from SOI for RME (top) and diesel (bottom).

L. Allocca et al. / Energy 77 (2014) 783e790 787

spray sequence at different time from the start of injection (SOI), forthe two fuels, at the injection pressure of 62.0 MPa. Due to the gasdensity slight variation in the evaporative environment, during theinjection duration, a constant value of 18.6 kg/m3 has beenconsidered in the non-evaporative condition. It corresponds to thestart of the main injection in the engine.

Each test point has been averaged on five images to correlateeach of them with a dispersion bar. At long time from the SOI thestructure of the spray depicts a uniform behaviour for each jetindicative of a good stability of the process. The sprays developuniformly in the pressure vessel reaching comparable penetrationlength at the same time. At early time from the SOI, and especiallyat 50 and 100 ms, smoky illuminated zones along the spray plumedirections, appear. They could be due to a ghosting residual illu-mination on the CCD camera such as to droplets present in thevessel due to the previous injection. In both cases they have not tobe referred to the interesting event and have not been considered inthe measurements. Finally, an analysis on the jet dispersion of thetip penetration reveals a greater stability of diesel fuel respect toRME. At injection time later than 200 ms, the <rms > ranged be-tween 9.1 and 1.5 for diesel while 14.0 to 5.6 for RME on the mainpulse.

The early stage injection processes needs particular mention.Here the sprays show a quite asymmetric distribution of the fuel upto 200 ms from the SOI, squared symbol in Fig. 3. The asymmetricspray behaviour is due to the not regular mechanical lift of theneedle in the first instants of the injection process. In this stage, thelesser amount of fuel injected by some holes is compensated by thegreater amount that comes out from the others. By looking at the

Fig. 5. Asymmetric distributions for RME (top) and diesel fuel (bottom) at 100 s afterthe start of the main pulse.

injection rate profiles (Fig. 3) during the rise time, the trends lookregular and no anomaly was pointed out. In fact, the pressuretransducer inside the Bosch tube is sensitive only to the totalamount of the injected fuel. This behaviour must be charged to thetransitory phase of the nozzle opening where the mechanical lift ofthe needle could be not regular for that device. Examples of thisbehaviour are reported in Fig. 5 for both the fuels.

In Fig. 5, three different images per each fuel at 100 ms from thestart of main injection are reported. The asymmetric fuel distri-bution is constant and repetitive for all the images. For the RMEfuel, some sprays on the right side show a shorter penetration thanthat on the left while it happens at the bottom for the diesel fuelwith respect to the top. The fuel distribution becomes quite sym-metric after 200 ms from the SOI for both fuels in agreement withthose reported in Ref. [28]. Moreover, the single shot images, ac-quired at the different times after the SOI, have depicted a quitenegligible cycle-to-cycle variation for the pilot and main injections.

3.2. Evaporative spray characterization

In Fig. 6, some images of main injection for diesel and RME fuel,respectively, are reported. They are in single-shot and have beenrecorded at 1500 rpm. For a fixed crank angle ASOI (after the start ofmain injection), images have been reported in order to evaluate thedifferences between themain injection of the two fuels. The start ofvisible injection occurs at about 2.5� crank angles (about 0.3 ms)after the start of the drive current to the injector and this delay isdue to electrical effects and the inertia of the needle. In Fig. 6 theeight jets can be clearly distinguished. The injection has durationaround 6� crank angles, in fact it can be consider finished at 3�

ATDC (after top dead centre) for both fuels.No impingement phenomena can be detected for both fuels. The

jets of diesel fuel show some asymmetry in the early stage of theinjection, they move in the bowl and have time to mix with airbefore the main autoignition occurs. The jets of RME fuel are morestable and did not have asymmetry in the spray behaviours. Duringthe main injection evolution, some luminous flames in the com-bustion chamber due to the combustion of the pilot injection can benoted. The diesel flames have greater luminous emissions thanRME. Due to these flames the temperature and pressure is changinglocally in the bowl. This affects the jets penetration and distribu-tion. Moreover, the RME fuel showed luminous flames of pilotcombustion just at 2� ATDC, earlier than diesel fuel. This is due tothe higher cetane number of the RME. Themain difference betweendiesel and RME regards the development of spray distribution intothe bowl. As it can be noted the RME jets are longer and wider thanthose of diesel. This could be due to the higher density, viscosity,and boiling point of RME fuel that affect the mixing of fuel with theair and the following pollutants formation.

Fig. 6. Spray sequence at different time from SOI for diesel (top) and RME (bottom).

L. Allocca et al. / Energy 77 (2014) 783e790788

3.3. Spray comparisons

The penetration of the sprays has been measured through theprocessing of the images. It is defined as the distance from thenozzle hole of the farthest fuel on the spray front. The spraypenetration is here presented as the result averaged on eight jets,which the spray is formed off, and on five repetitions at the samecondition in order to reduce the error of non uniform flow field andthe intrinsic irregularity. Same procedure has been applied in bothnon-evaporative system and transparent engine. The comparison ofthe liquid jet penetration under evaporative and non-evaporativeconditions is reported in Fig. 7 where the reported curves havebeen averaged on the eight jets of the complete spray and on thefive repetitions per each set point. Both pilot and main pulses forthe two used fuels are shown.

The plots prove that the evaporation process becomes thecontrolling parameter of the penetration for the liquid phase. Infact, the Fig. 7 shows that the penetration curves of the liquid startat the same time but develop in different way. Those evolving in theevaporative ambient start a reduction of the slope, 100 ms for dieseland 50 ms for RME with a collapses at 440 and 500 ms for diesel andRME, respectively. Under non-evaporative conditions the penetra-tion profiles don't show appreciable differences for the two fuelsconfirming that the different properties of the fuels induce lightvariations in the injection pattern. Moreover, spray patterns exhibitvery similar behaviour among the different fuels with deviationsover the standard dispersion that are noticeable at low injectedquantities. Finally, RME fuel showed longer tip penetration thandiesel fuel due to its high density, viscosity, and boiling point.

In order to understand how the inecylinder condition affectsthe evaporation of the spray, a comparison between the jets spatialdistributions in both systems has been made. Images have been

Fig. 7. Comparison of liquid jet penetration profiles under evaporative

calibrated in terms of pixel-mm ratio and the area of each jet, re-ported in Figs. 4 and 6, has been calculated for both fuels. Since thejets that spread in the engine are evaporating, while those in thespray vessel are in non-evaporative conditions, by means of thedifference between the jets areas it is possible to estimate theevaporated fraction on the jet. In Fig. 8 the calculated areas at twotimes after the start of the main injection are reported for dieseland RME fuel, respectively. At 100 ms ASOI (2� BTDC (before topdead centre)), bigger jet areas of diesel fuel has been noted in non-evaporative system, while the liquid jets spreading in the engineare greatly affected by the inecylinder condition that favours thejets evaporation. Similar behaviour has been noted for RME fuel.However, due to the jets distribution asymmetry noted in the non-evaporative system, some RME jets (#7 and #8 in Fig. 8) havesmaller areas than diesel. At 500 ms ASOI (1.5� ATDC), similarbehaviour has been observed. This means that a large amount ofthe liquid fuel injected is just evaporated in the bowl before themain combustion and the start of visible flames. Finally, the dieseljets areas resulted at both times ASOI lower than those of the RME.In particular, in the engine at 100 ms, a reduction of 61% for dieseland of 52% for RME of the jet areas with respect to those measuredin the non-evaporative system has been observed. The reduction isvery high also at 500 ms; they correspond to 64% for diesel and to60% for RME, respectively. This is due to the different properties ofthe RME fuel that resulted in lower evaporation efficiency.

3.4. Exhaust emissions

In Fig. 9, the gaseous and particulate matter (PM) exhaustmeasurements have been shown for the two investigated fuels.Even if the maximum care for reproducing multi-cylinder oper-ating conditions were employed, absolute data cannot be compared

and non-evaporative conditions for diesel (left) and RME (right).

Fig. 8. Comparison of liquid jet areas under evaporative and non-evaporative conditions at 100 s and 500 s ASOI for diesel (left) and RME (right).

L. Allocca et al. / Energy 77 (2014) 783e790 789

between an optical single-cylinder and a multi-cylinder applica-tions, chiefly because of the effects induced by the special pistonwith optical access. However, the authors verified that in terms ofrelative trends among pollutant concentrations for the differenttested fuels, a consistent agreement can be found between the twoengine applications [18,25]. Finally, it has to keep in considerationthat the gas measurements with the infrared detector are stronglyaffected by the exhaust temperature, which is low due to the shortoperation time of the optical engine. Regarding the exhaust emis-sions from the optical engine, it can be observed that RME produceshigher HC and CO emissions than diesel fuel. It is indicative of aworst mixing of the fuel with the air in the bowl [18,25]. The in-crease in HC emission may result from locally leaning out themixture during ignition delay of the pilot injection. Even if RME hashigher cetane number, it has lower LHV and higher boiling pointthan diesel and affects the evaporation process, as shown in theanalysis of Fig. 8. Moreover, because RME showed a longer pene-tration (Fig. 7), larger amount of RME droplets remains close to the

Fig. 9. Gaseous and particulate matter emissions for diesel and RME.

chamber walls during the combustion. Regarding the CO emission,its increase could be due to a worsening of the pilot of RME com-bustion because its CO oxidation was quenched by the heat evap-oration of the main injection (Fig. 6).

Concerning the NOx emissions, lower values have been noted forRME fuel [18,25]. Investigations on images of diesel and RMEcombustion made in previous papers [29,30], show that low localtemperatures have been detected for RME flames due to low LHV.Also Fang and Lee in Ref. [31] noted that a multiple injectionstrategy can greatly change the emission behaviour for differentfuels under similar load conditions. In particular, the multiple in-jection strategy is quite different from the single injection caseswith conventional combustion, where NOx emissions increase forbio-diesel fuel. Moreover, no defined trends have been obtained forthe effects of bio-diesel on the NOx emissions.

Lower values of PM have been measured at the exhaust due tothe higher oxygen content of RME that promotes the oxidation ofparticulate matter. Finally, higher values of the CO2 emission areexpected for RME because of the higher fuel injected quantityneeded to obtain the same BMEP as REF.

4. Conclusion

Mineral diesel and Rapeseed Methyl Ester fuels have beencharacterized by means of a 8-hole diesel injector in both a non-evaporative rig and an optically accessible single-cylinder engineat the engine operating condition of Urban Driving Cycle ECE R15:1500 rpm at 0.2 MPa of BMEP. The main results can be summarizedas:

- The different properties of the fuels do not affect the rates ofinjection except for the greatest amount of RME because of itslower net heat value;

- The early stages of the injection processes are characterized by astrong asymmetry in the penetration of the eight sprays for both

L. Allocca et al. / Energy 77 (2014) 783e790790

the fuels, nevertheless at time longer than 200 ms the penetra-tions become regular;

- The evaporation process controls the penetration of the liquidpart of the fuel: RME vaporizes faster than mineral resulting inreduced slope of the liquid penetration;

- In optical engine, RME pilot and main average tip penetration islower than diesel.

- Differences in the jet areas have been noted in the early stage ofthe main injection and when the jets are fully developed. Thedifference is due to the evaporation process occurring in theengine.

- Diesel fuel evaporates better than RME. This also affected thegaseous emissions, in fact, HC and CO for RME are higher thandiesel.

The main target of the work has been to extract the engineparameters, mostly the gaseous emissions, from the fuel disper-sions both in non-evaporating and evaporating conditions. Thedifferences between the two conditions have been qualitativelyrelated to the exhaust emissions.

Acknowledgements

The authors thank Mr. Carlo Rossi and Mr. Bruno Sgammato fortheir precious help.

Nomenclature

ASOI after start of injectionATDC after top dead centreBMEP brake mean effective pressureBTDC before top dead centrecad crank angle degreeCCD charge coupled deviceCR Common RailCW continuous wavelengthDI Direct InjectionECU electronic control unitEGR exhaust gas recirculationET energizing timingFAME fatty acid methyl esterPECU Programmable ECURME rapeseed methyl esterSF6 sulphur hexafluorideSOI start of injectionTDC top dead centreVSA variable swirl actuator

References

[1] Sidib�e SS, Blin J, Vaitilingom G, Azoumah Y. Use of crude filtered vegetable oilas a fuel in diesel engines state of the art: literature review. Renew Sust En-ergy Rev 2010;14(9):2748e59.

[2] Hossain AK, Davies PA. Plant oils as fuels for compression ignition engines: atechnical review and life-cycle analysis. Renew Energy 2010;35(1):1e13.

[3] Baquero G, Esteban B, Riba JR, Rius A, Puig R. An evaluation of the life cyclecost of rapeseed oil as a straight vegetable oil fuel to replace petroleum dieselin agriculture. Biomass Bioenergy 2011;35(8):3687e97.

[4] Zhao XW, Zan XK, He C, Tan JW. Experimental study on spray characteristicsof biodiesel oil. Chin Intern. Combust Engine Eng 2008;1:16e9.

[5] Senatore A, Cardone M, Allocca L, Vitolo S, Rocco V. Experimental character-ization of a common rail engine fuelled with different biodiesel. SAE TechPaper 2005-01-2207.

[6] Valentino G, Allocca L, Iannuzzi S, Montanaro A. Biodiesel/mineral diesel fuelmixtures: spray evolution and engine performance and emission character-ization. Energy 2011;36:3924e32.

[7] Atadashi IM, Aroua MK, Abdul Aziz A. High quality biodiesel and its dieselengine application: a review. Renew Sustain Energy Rev 2010;14:1999e2008.

[8] He C, Ge YS, Tan J, Han XK. Spray properties of alternative fuels: a comparativeanalysis of biodiesel and diesel. Int J Energy Res 2008;32:1329e38.

[9] Lee CS, Park SW, Kwon SI. An experimental study on the atomization andcombustion characteristics of biodiesel-blended fuels. Energy Fuels 2005;19:2201e8.

[10] Seyken XLJ, Somers LMT, Baert RSG. Modeling of common rail fuel injectionsystem and influence of fluid properties on injection process. In: Proceedingsof VAFSEP conference; 2004. Dublin.

[11] Caprotti R, Bhatti N, Nobuyiki I. Protecting diesel fuel injection system. SAETech Paper 2011-01-1927.

[12] Birgel A, Ladommatos N, Aleiferis P, Milovanovic N, Lacey P, Richerds P.Investigation on deposit formation in the holes of diesel injector nozzles. SAETech Paper 2011-01-1924.

[13] Ullman J, Geduldig M, Stutzenberger H, Caprotti R, Balfour G. Effect of fuelimpurities and additive interactions on the formation of internal dieselinjector deposits. TAE Essl Symposium; 2009.

[14] Barker J, Richards P, Snape C, Meredith W. Diesel injector deposits e an issuethat has evolved with engine technology. SAE Tech Paper 2011-01-1923.

[15] Ochoterena R, Larsson M, Andersson S, Denbratt I. Optical studies of spraydevelopment and combustion characterization of oxygenated and Fischere-Tropsch fuels. SAE Tech Paper 2008-01-1393.

[16] Pastor JV, Pastor JM, Gimeno J, Nerva J. The effect of biodiesel fuel blend rateon the liquid-phase fuel penetration in diesel engine conditions. SAE TechPaper no. 2009-24-0051.

[17] Fisher BT, Mueller CJ. Liquid penetration length of heptamethylnonane andtrimethylpentane under unsteady in-cylinder conditions. Fuel 2010;89:2673e96.

[18] Beatrice C, Di Iorio S, Guido C, Mancaruso E, Vaglieco BM, Vassallo A. Alter-native diesel fuels effects on combustion and emissions of an Euro4 Auto-motive diesel engine. SAE Tech Paper 2009-24-0088.

[19] Allocca L, Mancaruso E, Montanaro A, Vaglieco BM, Vassallo A. Renewablebiodiesel/reference diesel fuel mixtures distribution in non-evaporating andevaporating conditions for diesel engines SAE Tech Pap 2009-24-0054.

[20] Bosch W. The fuel rate indicator: a new measuring instrument for display ofthe characteristics of individual injection. SAE Tech Paper 6607496.

[21] Alfuso S, Allocca L, Corcione FE, Di Stasio S. Image diagnostics of common raildiesel sprays evolving in nitrogen ambient at different densities, ICE'99 in-ternal combustion engines: Experiments and Modeling, Capri 12-16September, 1999.

[22] Heywood JB. Internal combustion engine fundamentals. NewYork: Mc Graw-Hill; 1988.

[23] Mancaruso E, Sequino L, Vaglieco BM. First and second generation biodieselsspray characterization in a diesel engine. Fuel 2011;90:2870e83.

[24] AVL Model 415S Smoke Meter Product Literature, Updated June, 2005.[25] Mancaruso E, Vaglieco BM, Ciaravino C. Combustion analysis in an optical

diesel engine operating with low compression ratio and biodiesel fuels. SAEPaper 2010-01-0865. http://dx.doi.org/10.4271/2010-01-0865.

[26] Desantes JM, Arregle J, Ruiz S, Delage A. Characterisation of the injection-combustion process in a D.I. diesel engine running with rape oil methylester. SAE Tech Paper 1999-01-1497. http://dx.doi.org/10.4271/1999-01-1497.

[27] Allocca L, Mancaruso E, Montanaro A, Vaglieco BM. Use of Mineral diesel, 1stand 2nd generation biodiesels in modern common rail injection system undernon evaporative and evaporative conditions. Atomization Sprays 2012;22-n(2):1044e5110. http://dx.doi.org/10.1615/AtomizSpr. 2012004362.

[28] Allocca L, Montanaro A. Characterization of RME, RME aged and mineral dieselfuel injected by a common rail apparatus for EURO5 diesel engines. SAE TechPaper 2011-01-1938. http://dx.doi.org/10.4271/2011-01-1938.

[29] Mancaruso E, Vaglieco BM. Premixed combustion of GTL and RME fuels in asingle cylinder research engine. Appl Energy 2012;91(1):385e94.

[30] Mancaruso E, Vaglieco BM. An experimental comparison of n-heptane, RME,and diesel fuel on combustion characteristics in a compression ignition en-gine. Fuel Process Tech 2013;107:47e9.

[31] Fang T, Lee CF. Spray and combustion visualization in an optical HSDI dieselengine fuelled with biodiesel and diesel using multiple injection strategy,ILASS Americas 21st Annual Conference on liquid atomization and spraysystems. 2008. Orlando, FL.