Detailed Heat Release From RME

7/31/2019 Detailed Heat Release From RME

1/19

Paper Number 07PFL-353

Detailed Heat Release Analyses With Regard To Combustion of RME and Oxygenated Fuels in an HSDI Diesel Engine

Uwe Horn, Rolf Egnell and Bengt JohanssonLund Institute of Technology

ivind AnderssonVolvo Car Corporation

Copyright 2007 SAE International

ABSTRACT

Experiments on a modern DI Diesel engine were carriedout: The engine was fuelled with standard Diesel fuel,RME and a mixture of 85% standard Diesel fuel, 5%RME and 10% higher alcohols under low load conditions(4 bar IMEP).

During these experiments, different external EGR levelswere applied while the injection timing was chosen in away to keep the location of 50% heat release constant.

Emission analysis results were in accordance withwidely known correlations: Increasing EGR rateslowered NOx emissions. This is explained by a decreaseof global air-fuel ratio entailing longer ignition delay.Local gas-fuel ratio increases during ignition delay andlocal combustion temperature is lowered. Exhaust gasanalysis indicated further a strong increase of CO, PMand unburned HC emissions at high EGR levels. Thisresulted in lower combustion efficiency. PM emissionshowever, decreased above 50% EGR which was also inaccordance with previously reported results.

Besides those similar trends, fuel dependent differencesin indicated thermal efficiency as well as CO, HC, NOxand especially PM emissions were observed.

These differences were evaluated by detailed heatrelease analysis and explanation models based uponfuel characteristics as fuel viscosity and fuel distillationcurve.

Fuel spray evaporation and heat release wereinfluenced by these fuel characteristics. Due to thesecharacteristics it was concluded that RME has a highertendency to form fuel rich zones at low load conditionsthan the other tested fuel types.

Moreover it was found that improved fuel sprayvaporisation is an option to improve exhaust emissionsat low load conditions.

INTRODUCTION

In the degree that costs and demand of crude oil rise,diminish the economical disadvantages for alternativeDiesel fuels, resulting in a variety of feasible substitutes.Both the Diesel itself engine and its fuel have beendeveloped concurrently since a long time and reached alevel of sophistication that makes it hard for alternativeDiesel fuels to compete.

Many fuel substitutes have been shown to havedeviating exhaust emissions from conventional fuel. Themethyl ester of rapeseed oil (known as RME/biodiesel)is receiving increasing attention as an alternative fuel forDiesel engines. RME is a non-toxic, biodegradable andrenewable fuel with the potential to reduce engineexhaust emissions [3]. Differences to petroleum basedDiesel fuel are a slightly increased cetane number, lowsulphur content, low amount of aromatics, lower volatilityand a short distillation temperature interval. This hasboth positive and negative effects on exhaust emissions:

The CO 2 neutrality of RME is one of the mostcommon arguments propagating the use RMEinstead of standard Diesel fuel. If the economicallymost worthwhile fuel (i.e. tax reduced standardDiesel fuel for the agriculture sector) is used forproduction of RME, CO 2 neutrality is allayedsomewhat.

During ECE tests it has previously been observedthat Diesel engines fuelled with RME emit a loweramount of unburned hydrocarbon (HC) and carbonmonoxide (CO) emissions decrease compared tostandard Euro Diesel fuel (EDF) whereas NOxemissions are slightly increased. At high engineloads particulate matter ( PM ) is reported to be onthe same level or slightly lower as for standardDiesel fuels and at a higher level under low loadconditions. [6, 12]


2/19

It has also been reported that engines fuelled withRME produce a higher fraction of soluble organicmaterial in exhaust PM compared to engine run withEDF. However, during ECE tests, PM is lowered ifthe engine is fuelled with RME. Non-regulatedemissions as aromatic hydrocarbons and specificalkenes are reported to be considerably lower withRME. Aldehydes and ketones are reported to be onthe same level as for standard Diesel fuels. [1, 6]

In addition, the mutagenicity of RME emissions islow compared to fossil fuels indicating a reducedhealth risk from cancer which is related to theextremely low sulphur content of RME [10]. Lowexhaust gas sulphur content reduces moreovercatalyst wear due to sulphate formation [1, 2]. Newconventional Diesel fuels (DF) provide also lowsulphur contents (Euro-DF


3/19

0 10 20 30 40 50 600

0.2

0.4

0.6

0.8

1

F S N [ - ]

EGR [%] Figure 1: Influence of intake EGR level on PM emissions

Ignition delay increases with increasing EGR level. If allfuel is injected well before SOC (see Figure 1 highestEGR level), no diffusive combustion occurs at all due toa sufficiently long ignition delay. Due to stronger mass

entrainment into the fuel spray, local temperatures arebelow 1500 K, which is the lower limit for soot formation.Hence, PM emissions are reduced and remainingresiduals after premixed combustion are not oxidiseddue to the lack of diffusive combustion. This entailsincreased HC and CO emissions (Figure 5 maximumEGR). As the length of ignition delay is highly dependenton fuel mixture generation, the timing of SOC becomeshard to control and results in increased variations inmaximum cylinder pressure as well as IMEP.

Even premixed reaction speed is reduced withincreasing EGR levels which is due to the lack of locallyavailable oxygen. This effect contributes to an improvedmixing of fuel mass with cylinder charge mass duringcombustion. Emerging diffusive hot spots can distributetheir heat to a larger amount of charge mass. In this waylocal temperatures are lowered and NOx formation isreduced, which was reflected in emission measurements(Figure 5).

As the equilibrium of the NOx reactions is dependent onthe local oxygen concentration, decreasing local oxygenconcentration shifts the reaction equilibrium in a way thatNOx output is lowered (Figure 5 - NOx).

Late stage premixed combustion shifts gradually toincreasing diffusive combustion with decreasing EGRlevel. Both combustion types overlay each other, whichmakes it hard to find an indicator for the start of diffusivecombustion.

The heat release rate (HRR) and the 2 nd heat releasederivate ( D2HR ) were used to characterise the ratio ofpremixed and diffusive combustion. Thischaracterisation is explained for the 40% EGR case inFigure 2:

-5 0 5 10 15-50

0

50

100

150

200

[CA] H R R [ J / C

A ] , D 2 H R [ J / C A 2 ]

HRRD2HR

d i f f u s i v e

p r e m

i x e d

-5 0 5 10 15-50

0

50

100

150

200

[CA] H R R [ J / C

A ] , D 2 H R [ J / C A 2 ]

HRRD2HR

d i f f u s i v e

p r e m

i x e d

-5 0 5 10 15-50

0

50

100

150

200

[CA] H R R [ J / C

A ] , D 2 H R [ J / C A 2 ]

HRRD2HR

d i f f u s i v e

p r e m

i x e d

d i f f u s i v e

p r e m

i x e d

Figure 2: Indicators for premixed and diffusive dieselcombustion (40% EGR)

The start of diffusive-only combustion is indicated by aslope change of HRR and D2HR. If the amount ofdiffusive HR shall be estimated from the HRR curve,

errors in estimation of the start of diffusive combustionare crucial due to the strong HRR during premixedcombustion: An estimation error of 0.2CA within thepremixed phase yields a difference in HR ratio of ~5%.

For the characteristics of partially premixed combustion,the following is summarised:

The amount of EGR has a major influence onignition delay and combustion duration.

Fuel/air mixture generation is considerablyinfluenced by ignition delay.

Exhaust emissions (i.e. CO, HC and PM) areconsiderably affected when diffusive combustionshifts to premixed-only combustion.

Local reaction temperature and oxygenconcentration are the major factors influencingNOx formation.

FUEL CHARACTERISTICS

OXYGENATES

The global reaction equilibrium is influenced bycombustion of oxygenated fuels in the following way:

2,2,

22

22,

22

22

2

)1(76.3

)76.3(

On N n

O H nCOn

N OnO H C

stoichO stoichO

O H CO

stoichO z y x

++++++

(1)

For oxygenated fuels, a part of the oxygen needed forchemical reactions is already bound in fuel molecules,which has two effects:


4/19

Due to a lower stoichiometric F A ratio, local cylindergas entrainment is lower which increases localcombustion temperatures.

The ratio of stoichiometric fuel mass and releasedfuel energy is higher, which is due to lower fuelenergy content. This entails lower local combustiontemperatures.

If vaporisation properties are not affected byoxygenates, increased local oxygen concentration mighthave a higher tendency towards NOx formation whereasPM is lowered due to improved oxidation properties.This effect, however, might be compensated by thehigher amount of fuel mass needed to release a certainamount of energy.

SAUTER MEAN DIAMETER

Spray vaporisation properties for RME deviateconsiderably from EDF. This is due to increased fuelviscosity and surface tension which entails larger

average droplet size. A measure for the average dropletsize is the Sauter Mean Diameter (SMD) [15]:

( ) 32

2

=

g g

f f

uC SMD

(2)

SMD is related to the fuel viscosity f [Pa s ], fuel

surface tension g [N/m ], fuel density f [kg/m3],

cylinder charge density g [kg/m3] and the velocity

difference between fuel and cylinder charge g u [m/s ].The empiric SMD relation is scaled by the correctionfactor C to fit to measurement results.

The velocity difference g u is highly dependent oninjection pressure; it is presumed that SMD isconsiderably influenced by increasing injection pressure(see Figure 19).

Calculation results for the SMD show how fuel propertiesinfluence the average droplet size at 500 bar injectionpressure. (Table 1):

Table 1: Influence of Fuel Properties on droplet sizeFuel Type SMD (at 500 bar)EDF 46 mAGRO15 46 mRME 84 m

Apart from injection pressure fuel viscosity has thestrongest influence on SMD. Fuel dependent differencesin viscosity (see Appendix Table 5) are hence reflectedin SMD.

DISTILLATION CURVE

The fuel distillation middle temperature (T50) for purefuel components like n-alkanes and aromates, is anindicator for the average molecule size. Hence, T50 is agood indicator for fuel CN. Aromates are reported tohave a somewhat less distinct relation between T50 andCN. A high amount of fuel additives changes overall fuelCN considerably and makes it hard to estimate standardfuel CN from T50 [5].

The most important fuel characteristics besides the fuelcetane number (CN) are described by the shape of thedistillation curve: The cetane number of the fuel fractionvaporising during the start of the distillation curve has aninfluence on exhaust emissions: A higher initial fuelcetane number (ICN) decreases ignition delay. Hence,both PM and NOx-emissions are lower with ICNimprovers even though the overall fuel cetane number isvirtually unaffected [17].

A flat distillation curve is advantageous due to a lowerfuel evaporation rate over the whole distillationtemperature interval. It has been reported that thermalengine efficiency increases slightly due to a bettercontrolled vaporized fuel supply during the wholecombustion process [5].

Even though the CN for RME is slightly higher than forEDF, combustion properties are reported to be poor atlow load conditions [1,5,12]. This is due to thecomposition of RME. RME consists of a few types ofacid ethyl esters with similar properties whereas EDFconsists of 200 different hydrocarbon types with varyingproperties. Hence, RME can be considered as a virtuallypure substance with a very high distillation curvegradient (cf. Figure 3):

200 250 300 3500

0.2

0.4

0.6

0.8

1

Distil lation Temperature [C]

D

i s t i l l a t i o n

F r a c t

i o n

[ - ]

EDFRME

Figure 3: Distillation curve for EDF and RME

The distillation middle temperature T50 as well as strongdistillation temperature gradient cause HC, CO and PMemissions. Especially under low load and cold startconditions HC emissions are produced by incompletecombustion due to poorly controlled fuel vaporisation as


5/19

a result of a high T50 and a short distillation temperatureinterval.

The initial vaporised fuel fraction is considered to have amajor influence on ignition delay. Thus, a high boilingpoint increases ignition delay as more time is needed tovaporise the initial fuel fraction.

In contrary to EDF, fuel compounds with PM precursors(i.e. polycyclic aromatic hydrocarbons PAH) are notincluded in RME. From this aspect, the affinity for RMEto form PM is lower due to the lack of PAH in themolecular structure but is most likely countervailed by itspoor evaporation characteristics.

For distillation curve characteristics, the following isconcluded:

A high distillation middle temperature (T50) isconsidered to increase ignition delay due todelayed fuel evaporation.

A high distillation curve gradient is assumed toincrease the affinity to form fuel rich zones asfuel is distributed under a shorter temperatureinterval.

For the differences in fuel characteristics between EDF,AGRO15 and RME, the following is summarised:

EDF (certification fuel) was chosen as thereference fuel. It contains ignition delayimprovers with comparatively low ICN andoxidation characteristics are improved by a lowvolatile fuel fraction with high FCN.

AGRO15 consists of 85% EDF and has thussimilar combustion characteristics as EDF. Thefuel fraction of 10% higher alcohols (i.e.oxygenates) might have a positive influence onoxidation characteristics.

For RME, fuel properties as viscosity, heatcapacity, boiling point, distillation curve, lowervolatility and chemical structure are moredifferent from EDF. This means that formationand growth of particulates during combustion aswell as oxidation characteristics might differconsiderably. The fuel oxygen content of RMEmight compensate increased emission formationdue to high fuel evaporation rates whenreaching the boiling point.

MEASUREMENT SETUP

The experiments discussed in this article were carriedout with a single cylinder research engine. Cylinderhead, engine piston, and common rail (CR) injectionsystem of a Volvo D5 HSDI Diesel engine with thefollowing specifications were used:

Table 2: Engine specificationsCylinder Head type Volvo D5Bore 81 mmStroke 92.3 mmDisplacement 480 cm 3 Compression Ratio r c,stat 16.15:1Injection Cone Angle 139Injector Type solenoid

Injection Nozzle Hole(Number x Diameter) 7x0.14 mm

The test rig was equipped with an adjustable EGRsystem and miscellaneous sensors. The most importantsensors for analysis and interpretation of measuringresults are depicted in the following figure:

exhaust

Dynamo-meter

EGRcooler

CR System

returnfuel

fuel

p

air

MEXA-7100DEGR:HC, CO, CO 2, O 2, NO x

AVL415S:FSN

CO 2 (MEXA-7100DEGR)

fuel balance( ) fuel m&

p,T

p,T

pRail

exhaust

Dynamo-meter

EGRcooler

CR System

returnfuel

fuel

pp

air

MEXA-7100DEGR:HC, CO, CO 2, O 2, NO x

AVL415S:FSN

CO 2 (MEXA-7100DEGR)



p,T

p,T

pRail

Figure 4: Engine setup and location of measuring points

Smoke emissions (in the unit filter smoke number FSN) were measured with a device ( AVL415S ) thatderived exhaust gas particle concentration from theblackening of filter paper that was exposed to theexhaust gas flow for a certain duration. Fuel effects onparticle properties as particle size distribution, solubilityor carcinogenic components can not be analyzedadequately with this measuring technique [6]. Thistechnique however is a commonly used measuringmethod and in accordance with todays emission teststhat focus on overall particle mass.

Exhaust gases were measured with an exhaust gasanalyser ( HORIBA MEXA-7100DEGR ) NOx and HCemissions were measured at a temperature of 191C.NOx emissions were measured with achemiluminescence analyser (CLA-755A) whereas HCemission measurements were done by means of a flameionization analyser from Horiba ( FIA-725A ). FIA resultscould be erroneous due to fuel dependent differences inthe molecular structure of unburned hydrocarbons.Intake gas, end exhaust O 2, CO and CO 2 sample gaseswere dehumidified before measurement. The EGR ratewas calculated from intake and exhaust gas CO 2concentration:


6/19

out CO

inCO EGR

,

,

2

2

= (3)

For Diesel combustion the major fuel dependentdifferences in exhaust emissions are observed at lowload conditions [5]. Hence, the following test conditionswere chosen:

Table 3: Engine test conditionsEngine speed 1200 rpmIMEP ~4 barInlet temperature 105 CInlet pressure 1.05 barRail pressure 500 barFuel Consumption (FC):FC EDF 0.119 g/sFC RME 0.129 g/sFC AGRO15 0.120 g/s

Fuelling rate was constant for each fuel. It was adjustedfor differences in density and heating value in order toinject the same energy amount for all fuels (Table 3).

Cylinder pressure was measured with a piezoelectricpressure sensor ( Kistler 6056 ) with sample every0.2CA. Both, fast and slow measurement data wereaveraged from 200 engine cycles.

MEASUREMENTS

Figure 5 shows how emissions changed during variationof EGR under constant fuelling rate. Fuel specificdifferences in indicated mean effective pressure of thecombustion phase (IMEP gross ) and exhaust emissionswere observed:

4

4.5

I M E P g

r o s s

[ b a r ]

0250500750

H C [ p p m C

3 ]

00.25

0.50.75

C O [ % ]

0250500750

N O

x [ p p m

]

0 10 20 30 40 50 600

1

2

F S N [ - ]

EGR [%]

EDFAGRO15RME

Figure 5: Exhaust emissions vs. EGR for three differentfuels

Compared to the reference fuel EDF AGRO15 showedslightly improved HC, CO and PM emissions at low EGRconditions whereas these emissions where slightlyincreased for the RME case (Figure 5). Exhaustemissions for all fuel types converged with increasingEGR level. Compared to EDF, NOx emissions wereslightly lower for both RME and AGRO15.

HC, CO, and PM emissions increased considerably withhigh EGR levels, which was due to incompletecombustion (see section Combustion Characteristics )NOx emissions were close to detection limit which wasdue to low temperature combustion.

To estimate the influence of injection pressure onmixture generation, a rail pressure variation at the pointof maximum PM emission (i.e. ~50% EGR) wasconducted (Figure 6). Changes in exhaust gasemissions were observed:

4

4.5

I M E P g

r o s s

[ b a r ]

200300400500

H C [ p p m C

]

0.2

0.3

C O [ % ]

10152025

N O

x [ p p m

]

500 600 700 800 900 10000

1

2

F S N [ - ]

pRail [bar]

EDFAGRO15RME

Figure 6: Influence of rail pressure on engine load exhaustemissions (50% EGR)

Higher local F A ratio during premixed combustioninfluenced the trade off between particle generation andoxidation in a way that PM output was lowered.Increasing injection pressure at 50% EGR, PM could bereduced for all fuels. HC and CO emissions were

improved for the RME case while exhaust emissions forAGRO15 and EDF were within measurement accuracy.

The observed emission characteristics led to thefollowing assumptions:

NOx and PM measurements indicate more fuel richcombustion for the RME case compared to the EDFcase.


7/19

The choice of higher alcohols in AGRO15 improvesoxidation characteristics which is reflected in theamount of unburned hydrocarbons (HC).

PM output is influenced by the trade off between PMgeneration and PM oxidation. Both mechanisms aredependent on local oxygen concentration, localtemperature and residence time. If injection pressureis increased, gas entrainment before ignition and

even during combustion is improved entailing higherlocal oxygen concentrations. Due to lower PMformation at higher local F A ratio and improvedparticle oxidation characteristics, PM emissions areimproved, especially for the RME case.

To be able to verify these assumptions, a detailedanalysis approach was chosen to analyse differences inemission and combustion characteristics due to fuelproperties.

ANALYSIS METHODOLOGY

For being able to investigate how calculated heatrelease (HR) was dependent on fuel propertydifferences, a HR analysis approach was chosen whichtook changes in cylinder gas composition into account.

The gas composition model based upon the apparentheat release rate HRR net which was derived from the 1

st law equation:

Vdp pdV HRRnet 11

1 +

=

(4)

is the adiabatic exponent, p the cylinder pressureand V the crank angle dependent cylinder volume. Forthe heat release analysis the following assumptionswere made:

The analysis is restricted to a control volume duringthe closed valve period (IVC-EVO). Blow-by is notconsidered.

The control volume is assumed to be a perfectlystirred reactor.

Cylinder pressure is assumed to be uniform in thewhole control volume.

The initial gas temperature at IVC was resolved fromthe inlet gas temperature. The cylinder mass at IVC was calculated from theaverage mole mass of cylinder gas composition andthe amount of moles at IVC, which was derived fromthe amount of injected fuel mass, the exhaust gasoxygen concentration, the EGR concentration andresidual gases.

The amount of residual gases is calculated by theideal gas law at TDC during gas exchange phase.

The level of external EGR is calculated from intakeand exhaust gas CO 2 measurement.

Convective heat transfer between gases and thecylinder wall was estimated by Woschniscorrelation.

The cylinder volume was divided into two regions:One with unburned fuel-air mixture and one withburned exhaust gas.

The crank angle dependent cylinder gascomposition was calculated from a normalised HRcurve and global chemical reaction.

Heat transfer between those two reaction zones wasneglected. The cylinder wall temperature was assumed to be

constant within the analysis interval.

EFFECTIVE COMPRESSION RATIO

The effective compression ratio ( eff cr , ) is different from

the static compression ratio ( stat cr , ) due to measuringerrors of the static compression volume and elasticstrain in engine components during engine run. Anestimation error in eff cr , shifts the phasing of calculatedHR, which induces analysis errors.

For a motored pressure curve, a compression interval closely after IVC is chosen, where cylinder wall

temperature is close to the global gas temperature. Byassuming a realistic for the considered gastemperature, the effective compression ratio iscalculated by adiabatic compression. The effectivecompression ratio eff cr , is varied by means of the

correlations )1( = cnt displaceme

r V

clearanceV and

nt displacemeclearance V V V +=in such a way that the HR for

the considered interval is minimised:

0)(

)(1

1)(

1)( 11

+

=

HR

pV V p HRR(5)

If eff cr , is estimated for a motored pressure curve,realistic assumption for the change of the adiabaticexponent due to cylinder gas composition can bedone.

CHANGE OF ADIABATIC EXPONENT

The adiabatic exponent (also called ratio of specific

heats) V P C C / = is used to describe adiabaticchanges of intensive gas state variables. Using the

relation uV P RC C += , one yields V uV C RC / )( += Hence, an increase of intake gas heat capacity lowers . The increase of exhaust gas heat capacity comparedto intake gas heat capacity is to be explained with itsdependency on the molecular degree of freedom (DOF):Gas molecules have different DOF (translation, rotation


8/19

and oscillation modes) which are dependent on theirmolecular structure. These different modes aresuccessively activated with higher temperatures. Atextremely high temperatures, even more heat can bestored due to effects such as dissociation and ionisation[14].

At standard conditions, the adiabatic coefficient of aircan be calculated from the molecular DOF to 4.1= which is very close to its literature value [14, pg. 514ff]. Ifintake air is mixed with warm exhaust gas, the amount ofcarbon dioxide and water molecules increases. Thosemolecules have a considerably higher heat capacity thannitrogen and oxygen molecules.

Assuming being constant and independent of cylindergas composition changes during combustion, a roughmean value for the whole combustion cycle between

4.1= (air at K T 300 ) and 3.1= (exhaust gas at1,650 = K T ) can be chosen.

The following figure shows the sensitivity of maximumHRR for different values of :

-10 0 10-500

50

100

150

200

[CA]

H e a t

R e

l e a s e

R a t e

[ J / C A ]

=1.4 =1.35 =1.3

decreasing

Figure 7: Influence of changes in constant adiabaticexponent on rate of heat release

A low adiabatic exponent increases the calculated HRRsince more energy can be stored in a control volume ata given temperature. Due to the higher molecular DOF,the specific heat capacity of the cylinder gas increasesconcurrently with the intake gas EGR level. Hence,underestimation of entails a higher overall HR. Sincethe HR calculation error is larger at high HRR, thecalculated HR is not only erroneous in amplitude butalso in phasing. Hence, it is important to make realisticassumptions for if the HR is used to characterisecombustion phenomena under use of excessive EGR.

A robust method to estimate the change in due tocylinder gas composition at IVC is to assume adiabaticcompression within a compression interval shortly afterIVC (Figure 8).

10 -5 10 -4 1010 4

10 5

10 6

10 7

ln(V)

l n ( p )

Volume-Pressure Traceprolonged "adiabatic" compression = 1.35"adiabatic" expansion = 1.33"realistic" adiabatic compression interval = 1.37

Figure 8: Influence of choice of interval on calculation

A realistic adiabatic compression interval is chosen forcylinder gas temperatures close to the cylinder walltemperature. In this compression interval, changes in due to heat losses can be neglected. The slope of the ln-p-V-diagram represents the adiabatic exponent and

can be calculated by:

( )( )12

21

/ ln / ln

V V p p

IVC = (6)

The subscripts 1 and 2 indicate start and end of theadiabatic compression interval. If the adiabaticcompression interval is prolonged until fuel injection, changes arbitrarily due to cylinder wall heat losses.

The change of cylinder gas composition due tocombustion can be taken into account by assumingadiabatic expansion during the expansion stroke afterthe end of combustion. Even here, heat losses to thecylinder walls are included in the adiabatic exponent.The difference in between compression andexpansion stroke (Figure 8 circle and square ) isinterpolated linearly or with a previously calculatednormalised HR.

The adiabatic exponent is highly temperature dependentdue to the change in DOF of the cylinder gas molecules.This temperature dependency was modelled with anexponential approach:

) / exp()( 210 T k k T = (7)

0 is a reference value for the intake gas, T is the actual

ambient gas temperature, 2,1k are model constants. Ithas previously been shown, that this temperature modelyields sufficiently exact results compared to results froma multi zone HR model [4]. However, deviations in due to cylinder gas compositions are not considered.Thus, the -temperature-model was scaled withrespect to the cylinder gas composition. The deviation in due to change of cylinder gas composition duringcombustion is depicted in Figure 9.


9/19

-20 -10 0 10 20

1.241.26

1.28

1.3

1.32

[CA]

a d i a b

a t i c

e x p o n e n

t

[ - ]

0%EGR50%EGR

0 = 1.38k1

= 0.2k

2= 900

Figure 9: Temperature and Gas Composition-Model for at different EGR levels

The response of the temperature-model (eq.7) onambient cylinder gas temperature is plotted as acontinuous line. The dashed lines show the -valuethat was fitted to the crank angle dependent cylinder gascomposition by calculating an averaged adiabaticexponent from the cylinder pressure trace until and aftercombustion (see Figure 8). Changes in cylinder gascomposition due to combustion were considered byinterpolating compression and expansion with a previouslycalculated normalized HR. The calculated adiabaticexponent ( dashed lines ) is lower than the result of thetemperature model because a fraction of cylinder wallheat losses are accounted as increased cylinder gasheat capacity. Thus, the model used during HR-analysis yields a higher amount of released heat thanthe temperature model only.

Ignition Delay and Injection Timing

The injection process is indicated by a local minimum inthe HRR before SOC and is due to heat losses duringfuel evaporation. Hence, the start of this local minimumwas defined as start of injection (SOI).

The injection duration DOI was derived from the 1 st lawequation. Assuming that the hydrodynamic pressureequals the pressure difference between injectionnozzle inj p and combustion chamber cyl p during injection,

the injection velocity injc can be estimated by Bernoullislaw:

2

2inj

fuel cyl inj

c p p = (8)

The injection mass flow rate is calculated from theinjection velocity, a hole dependent discharge coefficient

holesd C , and the nozzle hole area holes A :

holesholesd injinj AC cm = , & (9)

The overall hole area holes A is calculated as follows:

4)( 2holes D

holesholes Dcorr

N A

= (10)

with:

holes N being the number of holes [-]

holes D being the hole diameter [m]

Dcorr being the hole diameter correction factor [-].

The correction factor was chosen in a way that DOImatched to detailed injection simulation results for onereference point.

The discharge coefficient holesd C , is calculated from anempirical equation regarding flow characteristics due tonozzle hole geometry [4]:


10/19

Ambient Gas Temperature

The cylinder gas state is defined by three of the fourstate variables (p, V, n, T). The ideal gas law states thecorrelation between these state variables:

T nR pV u= (12)

The crank angle dependent cylinder gas temperature iscalculated from the gas composition at IVC:

.const RnT pV

nT pV

uCA IVC

==

=

(13)

Knowing the crank angle dependent pressure trace p CA,volume V CA and mole number n CA, the crank angledependent global cylinder gas temperature T CA iscalculated from (eq. 13):

( )

( ) CA IVC

IVC

CA IVC CA

n

n

pV

pV T T = (14)

This ambient cylinder gas temperature considerscylinder gas composition during combustion and istypically somewhat lower than ambient gas temperaturecalculated without influence of cylinder gas composition.This temperature difference influences heat transfercalculation.

HEAT TRANSFER

Due to the complexity of the flow patterns in thecombustion chamber, estimating the heat transfer to the

cylinder walls is very difficult. However, variouscorrelations have been derived from measurements topredict the heat transfer coefficient. During this workWoschnis correlation [10] has been used to estimatecylinder wall heat transfer:

[ ]53.08.08.02.0 = T w p BC h (15)The constant C was chosen according to Woschnisformulation, h is the heat transfer coefficient, B is acharacteristic length which was set equal to the borediameter in [m] , p is the cylinder pressure in [bar] , T isthe gas temperature in [K] and w is the characteristicgas speed in [m/s] given below:

IVC stroke

s IVC m p

p pV

V T C vC w 021

+= (16)

with:

28.21 =C

= )expansion(1024.3

)ncompressio(032

sK mC

Reference pressure and temperature were chosen atIVC. mv is the mean piston speed, T IVC is the referencetemperature, V s /V stroke is the ratio of swept volume and a

reference volume (here the stroke volume), IVC p

p p 0 is

the ratio of the pressure difference between actualpressure, motored pressure curve and the referencepressure.

High rates of EGR entail a considerable change of thecompression stroke pressure curve due to the change of . This change in , however is not reflected in amotored pressure.

Hence, the motored pressure curve p 0 was estimatedfrom the actual pressure curve until SOI by assumingisentropic compression until TDC . The expansion curvewas mirrored from the compression curve. Calculation ofthe pressure difference between compression andexpansion stroke of a real motored pressure curve (i.e.without EGR) indicated a maximum pressure loss of 1bar at 20CA after TDC due to cylinder wall heat losses.

These crank angle dependent pressure losses weresubtracted from the expansion curve to minimizeestimation error of an EGR dependent motored pressurecurve.

ACCURACY OF MEASUREMENTS AND ANALYSIS

COMBUSTION PHASING

The amount of EGR was varied while injection timingwas adjusted in such a way that the point of 50% HR(CA50) was kept constant. The fast HR analysisprogram that was used to estimate combustion phasingdiffered from the HR program used for detailed HRcalculation. In Figure 10, an EGR dependent differencein calculated CA50 is observed:


11/19

20 30 40 502

3

4

5

C A

5 0 , f a s t

[ C A ]

0 10 20 30 40 50 602

3

4

5

C A

5 0 , d e t a

i l e d

[ C A ]

EGR [%]

EDFAGRO15RME

Figure 10: Differences in combustion phasing due tochoice of evaluation model

At high EGR levels, combustion advanced somewhatcloser to TDC. The depicted difference in CA50 due toHR analysis method is related to the followingparameters:

Table 4: Differences in HR analysis parametersParameter Fast: HR net Detailed HR gross

Constant Estimated

fromisentropiccompression

Crank angledependent

Dependent oncylinder gascomposition

Temperaturedependent -Analysis

Interval [-11060]CACompression andexpansion cycle

CA50/SOCEstimation

Interval[HR0HR max ] [HRminHR max ]

Differences in indicated thermal efficiency due tocombustion phasing are negligible for the consideredvariation interval of CA50 (cf. A.3). To improve clarityand comparability, different HRR were synchronised atSOC (see Figure 13).

Determination of SOC however varies dependent onchoice of the analysis interval for calculation of massfraction burnt. If the reference point for determination ofSOC is set to the point where the HR exceeds 0 (Table4 HR 0), one obtains a later SOC as if the referencepoint is set to the minimum HR (Table 4 HR min). SOCshifts due to choice of analysis interval with ~3CA. AsHRmin is not influenced by the type of calculated HR(with/without heat losses), this criterion was used todetermine SOC during detailed HR calculations.

COMBUSTION EFFICIENCY

A strong increase of CO, PM and unburned HCemissions at EGR levels beyond 40% loweredcombustion efficiency ( C ) considerably:

0 20 40 600.95

0.96

0.97

0.98

0.99

1

EGR [%]

c

[ - ]

EDFAGRO15RME

Figure 11: EGR vs. combustion efficiency c

The energy content of CO and HC emissions was used

to calculate combustion efficiency. Fuel dependentdeviations in C were within an interval of 0.5%.

Moreover is C only slightly affected due to the low PMmass flow fraction and heating value. The FIDmeasurement result is dependent on molecularcharacteristics as molecule length and formation. HCemissions were measured as propane equivalentwithout any fuel dependent correction factor.

For the calculation of combustion efficiency it can beconcluded, that exhaust measurement results and theirinfluence on combustion efficiency have to beconsidered carefully. The effect of combustion efficiencyis investigated more precisely in the following twosections.

ENERGY CONVERSION EFFICIENCY

Energy conversion efficiency E was calculated as thequotient of the maximum HR (HR gross,max ) and thesupplied energy. The supplied fuel energy wascalculated as the product of measured fuel mass flow,LHV and the combustion efficiency C (see Appendixeq.2).

Dependent on EGR level, the maximum HR (HR maxvaried with ~5% . By comparing HR max with the amount ofreleased chemical energy, one gets an estimate on thequality of Woschnis empirical equation (Figure 12).


12/19

10 20 30 40 500.9

0.95

1

E

[ - ]

0 10 20 30 40 50 600.94

0.96

0.98

1

C [

- ]

EGR [%]

EDFAGRO15RMERMEifr

Figure 12: E and C dependening on EGR, fuel type andfuelling rate

For high EGR levels, HR max matched the amount ofsupplied fuel energy (Figure 12 E ). This was due to ahighly premixed combustion mode with low radiativeheat losses. The estimation of cylinder wall heat lossesby Woschnis convective formulation does not coverradiative heat losses. Hence it is concluded, that thelower E value at low EGR levels was caused byradiative heat losses due to increased localtemperatures during diffusive combustion.

Presumed that the Wochni formulation comprehendsconvective heat losses correctly, it can be concludedthat the balance between convective and radiative heatlosses changes with EGR level. At low EGR, the localflame temperature increases due to less time for chargeentrainment. Hence, less mass is involved with initialcombustion. This gives higher local temperatures, higherNOx (see Figure 5) and more radiative heat losses.

For the RME case, two sub cases are depicted in Figure12: For the standard RME case, fuelling rate was

adapted to the amount of energy supplied with EDFand AGRO15, according to their LHV.

For RME ifr, fuel rate was increased such that engineload matched the EDF and AGRO15 cases within anaccuracy of 0.2 bar IMEP.

For these two cases, a considerable difference in E was observed for the whole EGR sweep, which mightoriginate from a difference in C .

The point of maximum E is situated at 50% EGR andis concurrent with maximum PM emissions. This showsthat soot oxidation reactions abate with increasing EGRlevel due to decreasing local temperatures. Combustionefficiency is included in calculation of E . The decreaseof E at the highest EGR level is either due to a more

effective combustion or due to an underestimation ofcombustion efficiency (Figure 12 C ). Anunderestimation is due to emission measurement errorat instable combustion conditions.

The analysis results presented in this section lead to thefollowing preliminary conclusions regardingmeasurement accuracy: HC and CO emissions arelowered with decreasing EGR. Hence, exhaust

measurement errors have the strongest effect at highEGR conditions and are moreover dependent on engineload. The estimation error for C is dependent on

exhaust gas emissions and influences E . As E isoverly lowered for the low load RME case, it is plausiblethat exhaust gas emission measurement results areunderestimated for this case.

A more exact investigation of the dependency between

C and indicated thermal efficiency (ITE) is given in thesection Thermal Efficiency in the appendix.

ANALYSIS RESULTS

In this section, previously described explanation modelsare applied to investigate differences in combustioncharacteristics due to EGR, fuel type and injectionpressure.

Influence of EGR and Fuel Type

Experimental results showed fuel dependent differencesfor engine load and exhaust emissions due to fuel typeat low EGR levels (Figure 5). These differences areexplained by comparison of HRR between 0 and 50%EGR. For clearness reasons, HR plots depicted inFigure 13 were synchronized at SOC .

As shown in Figure 13 both premixed and diffusive HRRis lower for the RME case. At low load conditions, RMEis considered as a poorly ignitable diesel fuel [5 pg.125ff] which is due to a high distillation middletemperature (see Figure 3): If it is assumed that theoverall liquid fuel spray has to be heated until T50 tobecome vaporised, and if it is moreover assumed thatheat capacities for both fuel types are alike, the heatneeded for warming up the liquid spray phase tovaporisation conditions is higher. Due to the highdistillation temperature gradient, more time is needed forvaporisation of the initial fuel fraction, which increasesignition delay somewhat.


13/19

0.4 0.6 0.8

0

200

4000...20 %EGR

H R [ J ]

0.4 0.6 0.8

30...50 %EGR

0.4 0.6 0.8

0

100

200

H R R [ J / C A ]

0.4 0.6 0.8

Premixed

Diffusive

0.4 0.6 0.8-50

0

50

D 2 H R [ J / C A 2 ]

time after SOC [ms]0.4 0.6 0.8

time after SOC [ms]

EDFRME

increasing EGR

Figure 13: Differences in HR, HRR and D2HR betweenEDF and RME due to EGR synchronized at SOC

Due to RMEs mono component like short distillationtemperature interval (cf. Figure 3) a large fuel sprayfraction vaporises concurrently within a smalltemperature and spatial interval if enough ambient heatis available. Together with a larger average droplet sizedue to higher fuel viscosity (Table 1, Table 5) these fuelcharacteristics increase the affinity of RME to form ahigher amount of fuel rich reaction zones at SOC. This isreflected in a lower overall HR due to poor premixed anddiffusive combustion (Figure 13 HR).

Both overall and premixed HR is delayed for the RMEcase between 0...20% EGR which is in accordance witha decrease in maximum HRR (Figure 16). If EGR level isincreased above 30% (Figure 13) the point of maximumHR is more delayed for the EDF case which is incontrast to the 020% EGR cases. This behaviour isrelated to the ignition delay which increases with EGRlevel. The time delay until auto ignition is a function ofmixture quality; the ignition delay for the RME case atEGR levels above 30% is an indicator for improvedmixture generation.

The ignition delay (Figure 14 SOC-SOI) of RME wassimilar to EDF. For EDF, the ICN is lower than theoverall CN to increase ignition delay. For RME on theother hand, the ICN is equal to the overall CN. The ICNis considered to have the strongest influence on ignitiondelay. Hence, the similar ignition delay between EDFand RME is related to a longer evaporation process forthe RME case; more time is needed to form an ignitablefuel-air mixture.

Due to the lower energy content of RME compared toEDF, more fuel mass had to be injected to reach thesame fuel energy content. Hence, the injection control

signal was lengthened from ~9.2 to ~10.0CA, entailinga larger fuel fraction being injected after SOC (Figure 14

SOC-EOI):

10 20 30 40 50

9

9.5

10

D O I [ C A ]

5

10

15

S O C - S

O I [ C A ]

EDFRMEAGRO15

-0.6-0.300.30.6

[ m s ]

0 10 20 30 40 50 60-5

0

5

S O C - E

O I [ C A ]

EGR [%]

Figure 14: Ignition delay and injection duration vs. EGR

For RME, a slow-down in HRR could be observed atEGR levels between 0 and ~30% (Figure 13 - circle ). Asthe EOI is situated in this region (Figure 14), it is mostlikely that heat losses due to fuel vaporisation areobserved in the HRR. As the fuel injection mass washigher for the RME case, more heat is needed for fuelspray vaporisation.

This HRR slow-down due to EOI was coincident withincreased PM level. Hence, it is most likely that earlyigniting fuel rich diffusive combustion zones evolvingfrom injection during concurrent premixed combustionare the reason for increased exhaust PM.

The injection process is split into two injection fractionswith help of SOC-EOI (Figure 14) as discussed before.The injected fuel mass was split into a mass fractioninjected before SOC and a mass fraction injected duringcombustion:

10 20 30 40 500.65

0.7

0.75

0.8

0.85

0.9

0.95

1

b e f o r e

S O C

[ - ]

0 20 40 60

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0

a f t e r

S O C [ - ]

EGR [%]

EDFAGRO15RME

Figure 15: Injection mass ratios vs. EGR


14/19

The injection mass ratio (Figure 15) is related to fueldependent differences in injection mass which wasadapted to keep the amount of fuel energy constant. AtEGR levels beyond 45%, ignition delay is considerablyincreased compared to combustion without EGR. Thus,all fuel was injected well before SOC.

For the AGRO15 and the RME case a larger amount offuel mass was injected after SOC. For AGRO15 this wasdue to a shorter ignition delay (Figure 14 SOC-SOI).For the RME case, a larger fuel mass was injectedentailing longer DOI and thus more fuel mass beinginjected after SOC.

Emission results (Figure 5) however, did not show anydifferences for AGRO15 in comparison to the referencefuel (EDF) whereas for the RME case differences in PMemissions were observed at low EGR conditions.

For the AGRO15 and EDF cases, concurrent fuelinjection and premixed combustion was not observed inHRR. This indicates improved fuel spray evaporationdue to the following properties:

Due to low viscosity, the average spray droplet sizeis lower (Table 1), which improves evaporationcharacteristics.

Due to a wide distillation temperature interval and alow initial distillation temperature the injected fuelspray starts to evaporate early (Figure 3).

Both factors influence the fuel spray evaporationprocess in a way that less spray is evaporated at a laterinjection stage.

In Figure 16, the maximum HRR of all fuel types arecompared with each other. It is observed that themaximum rate of premixed combustion is lower for bothAGRO15 and RME:

0 10 20 30 40 50 6050

100

150

200

250

300

H R R

m a x

[ J / C A ]

EGR [%]

EDFAGRO15RME

Figure 16: Maximum HRR vs. EGR level

The fuel dependent difference in HRR max is related to thevaporisation characteristics of RME. Due to anincreased SMD (Table 1) less zones of air-fuel mixtureclose to stoichiometry are created during ignition delaywhich lowers premixed HRR. Fuel rich zones need more

time to mix with cylinder charge and combust at a laterstage if sufficient local heat and oxygen is available forself ignition. Hence, the HR of the RME case is bothdelayed and decelerated due to inferior fuel sprayevaporation characteristics.

If HR is separated into a diffusive and premixed HRfraction (cf. Figure 2), one obtains a different picturefrom what injection mass ratios (Figure 15) imply:

10 20 30 40150

175

200

225

P r e m

i x e

d H R

[ J ]

10 20 30 40225

250

275

300

325

D i f f u s

i v e H R

[ J ]

EGR [%]

EDFAGRO15RME

Figure 17: HR combustion ratios in dependence of EGR

The overall amount of heat released (Figure 17:HRDiffusive +HR Premixed ) is lower for the RME case eventhough the injection mass was chosen in a way to keepthe amount of supplied fuel energy constant for all fuels.

For all cases, the premixed HR ratio is lower than theenergy equivalent of the mass ratio injected before SOC.Maximum premixed HR is lowered due to the mixturegeneration process which takes some time and is notregarded for the estimation of injection mass ratio beforeand after SOC (Figure 15).

For the AGRO15 and EDF case, the diffusive HR ratio(Figure 17) is basically constant until 40% EGR. For theRME case, diffusive HR was considerably lower than isanticipated from injection mass ratio (Figure 15). Thisindicates poor diffusive combustion which is related to alow FCN. Moreover, a slight wear-off in diffusive HR wasobserved with increasing EGR levels.

For the influence of EGR on different fuel types, thefollowing is concluded:

For all fuels, ignition delay (Figure 14) andcombustion duration is increases with EGR level. Alonger ignition delay promotes mixture generation.This however was not reflected in the amount ofpremixed HR.

For the RME case, overall HR as well as themaximum HRR is lower. For EGR levels above 30%,combustion phasing was slightly earlier than for theEDF case which was in contrary to low EGR levels.


15/19

This is an indication for an improvement in mixturegeneration for the RME case with higher EGRlevels.

The ignition delay was shorter for AGRO15 than forthe reference fuel. HR characteristics however, weresimilar for both fuel types: Increased fuel massinjection after SOC was not reflected in HR, whichwas an indication for improved fuel sprayevaporation characteristics.

Influence of Rail Pressure

The variation of rail pressure led to a further indicationfor the assumption of locally fuel rich combustion ofRME as the source for PM. In Figure 18, fuel specificdifferences at diffusive-only HR are obvious. Observeddifferences in HR for EDF and RME were considerable:

0

200

400

H R [ J ]

0

50

100

150

H R R [ J / C A ]

0.4 0.6 0.8 1 1.2-20

0

20

D 2 H R [ J / C A 2 ]

time after SOC [ms]

EDFRME

increasing p Rail :500 ... 1000 bar

Figure 18: Influence of injection pressure on HR (50%EGR)

In Figure 18, a decrease in ignition delay andcombustion duration is observed, while HR max andHRR max increased concurrently with higher injectionpressure. Increased HR max is an indicator for improvedmixture generation.

This indicator is strengthened by the calculation ofaverage droplet size. In Figure 19 it is depicted, thatSMD is considerably influenced by injection pressure.

500 600 700 800 900 100020

3040

506070

8090

S M D [ m

]

p Rail [bar]

EDFAGRO15RME

Figure 19: SMD size reduction due to increasing injectionpressure

The injection velocity is the strongest factor influencingSMD (eq.2). Increasing injection pressure increasesrelative velocity between injected fuel und cylindercharge. Hence, average droplet size is reduced whichimproves vaporisation characteristics for all fuels. RME,however, which has the largest SMD due to higher fuelviscosity benefits strongest from increasing injectionpressure which is reflected in decreasing emissions(Figure 6) and increasing HR max (Figure 18).

Ignition delay and combustion duration were shorter withincreasing injection pressures. The shortening of ignitiondelay at high rail pressures was compensated by theshortening of injection duration. No considerablevariation in premixed and diffusive HR was observed.

For the influence of increasing injection pressure on HR,the following is concluded:

Average droplet size decreases with injectionpressure, which improves fuel spray evaporationcharacteristics.

Fuel mixture generation is improved which isreflected in improved HR max , shorter ignition delayand shorter combustion duration.

The fuel type RME benefits the most from increasedinjection pressure, which is related to a strongerreduction in SMD compared with EDF.

CONCLUSIONS

Measurement results from an HSDI Diesel engine werestudied by detailed heat release analysis. For all fueltypes, increasing EGR levels increase ignition delay(Figure 14), which improves mixture generation andpremixed combustion. Increasing EGR level entails alonger premixed combustion duration (Figure 13), whichlowers maximum HRR. In this way the amount of fuelrich, locally hot zones is minimised. Hence, PMemissions originating from fuel rich combustion wereimproved.


16/19

RME and AGRO15 showed different combustioncharacteristics from EDF at low load conditions whichdepends on the following factors:

Injection duration was longer for the RME case dueto lower fuel energy content. This resulted in lessavailable time for fuel mixture generation; a largerfuel fraction was injected close to or after SOC.

For RME, the high ICN resulted in early ignition offuel rich zones entailing suboptimal premixed HRwhich was indicated by lower maximum HR(Figure 13).

Due to higher fuel viscosity, SMD is increased forRME (Figure 19). Increased average droplet sizeincreases the affinity to form locally fuel rich zones.Fuel dependent SMD differences can becompensated by increased injection pressure.

RME combustion characteristics can be consideredas poor at low load conditions. Due to a high

distillation middle temperature, short distillationtemperature interval (Figure 3) and increased SMD,fuel vaporisation is delayed. Premixed HR isunsteady due to concurrent fuel spray evaporationand premixed combustion (Figure 13).

The ratio of diffusive and premixed HR is lower forRME than for the reference fuel (Figure 17). This isrelated to the lack of FCN improvers. The lowerdiffusive HR entails poor emission oxidationcharacteristics.

Supplementary RME measurements with slightlyincreased engine load resulted in considerablyimproved E (Figure 12) which is an indicator for astrong load dependency of E at low loadconditions.

The AGRO15 and EDF case showed that shortenedignition delay, concurrent fuel mass injection andpremixed combustion have minor influence on PMemissions due to good fuel spray vaporisationcharacteristics.

For AGRO15, HC and CO emissions weredecreased compared to RME and EDF at lowload/low EGR conditions even though diffusive HRwas lower (Figure 17). This indicates improvedoxidation characteristics during diffusive combustion.

If alternative diesel fuels with varying properties shall beincluded in future emission legislations, technicalpossibilities exist to meet these legislations: TodaysDiesel engines are equipped with advanced combustioncontrol devices like EGR, Common Rail and TurboCharging. In addition, fuel and/or cylinder pressuresensors open possibilities for a multi fuel productionengine. If engine control strategy is adapted to a specific

fuel type during operation, emission demands can bemet for a variety of fuels.

ACKNOWLEDGMENTS

The author would like to thank the Volvo CarCorporation for supplying the measurement data. Thiswork has been financed by the Swedish CompetenceCentre for Combustion Processes (KCFP).

REFERENCES

1. Chang, D., Van Gerpen, J.: Determination ofParticulate and Unburned Hydrocarbon Emissions fromDiesel Engines Fueled with Biodiesel, SAE 982527

2. May, H., Hattingen, U., Theobald, J., Weidmann, K.,Knig, A.: Untersuchung des Betriebs- undAbgasemissionsverhaltens eines Dieselmotors mitOxidationskatalysator - Verwendung von Rapsl-Methyl-Ester (RME), MTZ Ausgabe Nr.: 1998-02

3. Hopp, M.: Untersuchung des Einspritzverhaltens unddes thermischen Motorprozesses bei Verwendung vonRapsl und Rapsmethylester in einem Common-Rail-Dieselmotor, Dissertation Universitt Rostock 2005

4. Egnell, R.: Combustion Diagnostics by Means ofMultizone Heat Release Analysis and NO Calculation,SAE 981424

5. Garbe, T.: Senkung der Emissionen eines PKW mitdirekteinspritzenden Dieselmotor durch Verwendung vonKraftstoffen mit abgestimmtem Siede- undZndverhalten, Dissertation Universitt Hannover 2002

6. Munack, A., Schrder, O., Stein, H., Krahl, J., Bnger,J.: Systematische Untersuchungen der Emissionen ausder motorischen Verbrennung von RME, MK1 und DK,Landbauforschung Vlkenrode FAL AgriculturalResearch 2003

7. Krahl, J., Munack, A., Schrder, O., Stein, H., Bnger,J.: Influence of Biodiesel and Different Designed DieselFuels on the Exhaust Gas Emissions and HealthEffects, SAE 2003-01-3199

8. Taylor, J., McCormick, R., Clark, W.: Report onrelationship between molecular structure andcompression ignition fuels, both conventional andHCCI, National Renewable Energy Laboratory 2004

9. Murphy, M., Taylor, J., McCormick, R.: Compendiumof Experimental Cetane Number Data, NationalRenewable Energy Laboratory 2004

10. Woschni, G.: A Universally Applicable Equation forthe Instantaneous Heat Transfer Coefficient in theInternal Combustion Engine, SAE 670931


17/19

11. Dec, J.E.: A Conceptual Model of Di DieselCombustion Based on Laser-Sheet Imaging, SAE970873

12. Mayer, A.C.R., Czerwinski, J., Wyser, M.: Impact ofRME/Diesel Blends on Particle Formation, ParticleFiltration and PAH Emissions SAE 2005-01-1728

13. Munack, A., Krahl, J.: Erkennung des RME-Betriebes mittels eines Biodiesel-KraftstoffsensorsLandbauforschung Vlkenrode FAL AgriculturalResearch 2003

14. Young, H.D., Freedman, A.R., et al.: UniversityPhysics 9 th edition Addison Wesley PublishingCompany, Inc.

15. Valdsoo, T.: Studier av atomiseringsfenomen underbrnsleluftpreparering STU dnr: 86-4560

16. R. L. McCormick, C. J. Tennant, R. R. Hayes, S.Black, J. Ireland, T. McDaniel, A. Williams, M. Frailey:Regulated Emissions from Biodiesel Tested in Heavy-Duty Engines Meeting 2004 Emission Standards, SAE2005-01-2200

17. Musculus, Mark. P. B.: On the Correlation betweenNOx Emissions and the Diesel Premixed Burn, SAE2004-01-1401


18/19

CONTACT

Uwe Horn can be contacted via the following emailaddress: [email protected]

DEFINITIONS, ACRONYMS, ABBREVIATIONS

AGRO15 Mixture of 85%EDF, 5%RME and 10%higher alcohols

COV Coefficient of variationDOF Degree of FreedomDOI Duration of InjectionD2HR 2nd Heat Release derivateECE Energy Consumption EfficiencyEDF Euro Diesel FuelEVO Exhaust Valve OpeningFCE Fuel Consumption EfficiencyFCN Final Cetane NumberFSN Filter Smoke NumberHR Heat ReleaseHRR Rate of Heat ReleaseHRnet Net Heat ReleaseHRht Heat Release due to heat transfer

HRgross Gross Heat Release

ht net gross HR HR HR += IVC Intake Valve ClosingICN Initial Cetane Number

ITEgross/net

Indicated Thermal Efficiency:

LHV m

W ITE

f C

i

=

gross/net,

IFCEgross/net

Indicated Fuel Conversion Efficiency:

LHV mW IFCE

f

i

= gross/net,

RME Rape Oil Methyl Ester / Biodieselr c,eff Effective compression ratior c,stat Static compression ratioSMD Sauter Mean DiameterSOC Start of CombustionSOI Start of Injection E Energy Conversion Efficiency

C Combustion Efficiency

ITE

IFCE C =

APPENDIX

COMBUSTION CHARACTERISTIC VALUES

Emission analysis data can be used to calculate thecombustion efficiency C from the global air-fuel-ratio

F A , the emission gas mole fraction i , the ratio of

exhaust gas products mole weights p

i

M M and the ratio of

the lower heating values from emission gas components

and fuel f

i

LHV LHV :

( ) +=i

i LHV LHV

M M

F A

C f i

p

i 11 (A.1)

C determines how effectively fuel energy is convertedto heat energy. For a given load, higher combustionefficiency implies in general lower fuel consumption.

As heat losses are solely modelled by Woschnisapproach, energy conversion efficiency ( E ) is used as ameasure for the quality of HR:

LHV m HR

c fuel E

=

&

max (A.2)

E makes it possible to determine how combustionmode dependent thermodynamic properties as HRR-

shape, -duration and -location influence heat losses.

FUEL PROPERTIES

During engine measurements, the engine was operatedunder low load conditions (4 bar IMEP) with threedifferent fuel types: Euro Diesel fuel (EDF), RME, and amixture of 10% higher alcohols, 5% RME and 85% EDF(AGRO15)

Table 5: Fuel Property DataParameter EDF RME AGRO15flashpoint [C] 64 >110 2 water content [ppm]


19/19

ITE CALCULATIONS

If highly premixed diesel combustion is simplified as aconstant volume heat release at CA50 with a constantadiabatic exponent, the difference in thermal efficiencydue to combustion phasing can be approximated as:

1,

11 =

eff cT

r (A.3)

with

==

=CA4.5CAfor23.15

CA2.5CAfor51.15

max50,

min50,,eff cr (A.4)

and

37.1= (A.5)

one obtains

==

=CA4.5CAfor0.6349

CA2.5CAfor0.6374

max50,

min50,T (A.6)

Differences in HR due to combustion phasing can not belarger than ( ) %4.0

Date post:	05-Apr-2018
Category:	Documents
Upload:	mohd-suardi-suhaimi
View:	217 times
Download:	0 times

Detailed Heat Release From RME

Documents