Simulation and experimental investigation of performance

Scientia Iranica B (2014) 21(4), 1424{1439

Sharif University of TechnologyScientia Iranica

Transactions B: Mechanical Engineeringwww.scientiairanica.com

Simulation and experimental investigation ofperformance and emissions of a turbocharged lean-burnnatural gas engine considering thermal boundary layer

Sh. Kharazmi�, A. Mozafari and A. Hajilouy-Benisi

Department of Mechanical Engineering, Sharif University of Technology, Tehran, P.O. Box 11155-9567, Iran.

Received 18 February 2013; received in revised form 5 September 2013; accepted 4 December 2013

KEYWORDSNatural gas;Turbocharged;Turbulence andcombustionsimulation;Thermal boundarylayer;NOx;CO;UHC;CO2.

Abstract. Emission characteristics of pure natural gas is fairly known, however moreexperimental investigation of natural gas combustion is necessary. In this research, emissionand performance of a turbocharged natural gas SI engine are investigated experimentallyat WOT and lean burn conditions. A computer code is also developed in MATLABenvironment for predicting engine performance and NO emission and the results arevalidated with experimental data. Thermal boundary layer, as a novel approach, isconsidered in the simulation and a better agreement of predicted mass fraction burnedwas achieved at end part of combustion process when compared with experimental results.Experimental results have revealed that turbocharger match at mid engine speed, waste-gate opening and increase of friction losses at high engine speed resulted in better torqueback-up. Boost pressure, controlled by compressor outlet pressure, remained nearlyconstant at speeds higher than 1450 rpm. NO emission reduced with engine speed increasedue to shorter time that burned gases remain at high temperatures, although they havehigher temperature at higher speeds. It however increased with the increase of excess airratio until it reached to its peak value at about 1.1 from stoichiometric mixture and thendecreased, while it decreased substantially with increase of spark timing retard. Brakespeci�c UHC and CO2 emissions were min at mid speed range and WOT.© 2014 Sharif University of Technology. All rights reserved.

1. Introduction

Fossil fuel consumption is steadily increasing as aresult of population growth and improvements of livingstandard demand. World total energy consumption hasgrown by about 36% over the past 15 years. Internalcombustion engine fuel resources depend on availabilityof fossil fuel supplies. Strict emission regulationsare another issue facing the current generation of ICengines [1].

*. Corresponding author. Tel: +98 21 66165667E-mail addresses: [email protected] (Sh.Kharazmi); [email protected] (A. Mozafari);[email protected] (A. Hajilouy-Benisi)

One of the favorite alternative fuels is naturalgas. It is composed of methane with molar H/C ratioof 3.8 comparing with ratios of 1.8-2.0 for gasolineand diesel fuels. This higher ratio reduces CO2emission and so its greenhouse e�ects. 1 g methaneproduces 0.056 g CO2/kJ while 1 g of gasoline anddiesel produces 0.068 g CO2/kJ and 0.071 g CO2/kJrespectively. CO2 emission of natural gas engine isabout 20% lower than that of gasoline engine at thesame power [2]. Natural gas produces lower CO andTHC emissions. Higher RON of natural gas allowshigher compression ratio thus producing higher thermale�ciencies. It also has relatively wider ammabilitylimits, with better combustion in lean-burn strategies.Natural gas gives lower engine volumetric e�ciency due

Sh. Kharazmi et al./Scientia Iranica, Transactions B: Mechanical Engineering 21 (2014) 1424{1439 1425

to its injection in the intake manifold (compared todiesel fuel) and its lower stoichiometric fuel/air ratio(compared to gasoline) thus provides less power output.CNG DI engines have good volumetric e�ciencies, butthey have problematic load control and di�culties ininjection control and technology and durability test.Turbocharger can o�set this power reduction. It usesexhaust gas energy to compress intake air; thereforemore fuel per displacement volume can be burned in acycle. Turbocharging of SI engines has been improvedless than diesel engines due to many di�culties, likeknock phenomena and turbocharger matching withthrottle valve [3].

Natural gas has signi�cantly larger proven re-serves compared with crude oil. The current knownReserves-to-Production (R/P) ratio for crude oil isabout 40 years, while for natural gas it is about 60years [1].

Emission standards are expected to becomestricter in future. The European Automotive Manu-facturers Association has determined 130 g/km CO2emission target for 2012 to 2015, with an additional10 g/km CO2 emission reduction coming from \comple-mentary measures" including a greater use of biofuels.65 percent of new cars will have to comply with theemission requirements in 2012, 75 percent in 2013, 80percent in 2014 and 100 percent in 2015 [4]. NOxemission has been the strictest Euro emission standard,with a limit of 0.4 g/(kWhr) in Euro VI. Table 1 showsstandard values of emissions in recent years.

Cho and He [2] reviewed SI natural gas engines.Engine operating map, fuel economy, emissions, cycle-to-cycle variations in IMEP and strategies to achievestable combustion of lean burn natural gas engineswere highlighted. Stoichiometric natural gas engineswere also brie y reviewed. They concluded that highactivity catalyst for methane oxidation and lean de-NOx system or three-way catalyst with precise air-fuelratio control strategies should be developed to meetstringent emission standards.

Rothlisberger and Favrat [6] investigated theoperation of a cogeneration natural gas engine withprechamber. They presented the limit of conventionallean burn operating mode with spark ignition (withoutprechamber) and discussed the prechamber geometrical

Table 1. The European emission standards, g/kWh [5].

Year Standard CO HC NOx PM

1992 Euro I 4.5 1.1 8.0 0.361996 Euro II 4 1.1 7 0.152000 Euro III 2.1 0.66 5 0.12005 Euro IV 1.5 0.46 3.5 0.022008 Euro V 1.5 0.46 2 0.022013 Euro VI 1.5 0.13 0.4 0.01

con�guration. They illustrated the in uence of pistongeometry and concluded that in comparison with thespark ignition (without prechamber), the prechamberignition strongly intensi�ed and accelerated the com-bustion process.

In another research [7], they studied the in u-ences of spark timing, turbocharger speci�cations andengine load on prechamber ignition operation. Theyalso presented a comparison between spark (withoutprechamber) and prechamber ignition. It was foundthat at rated power output, the prechamber ignitionoperation ful�lled the Swiss requirement for emissionand achieved fuel conversion e�ciency higher than36.5%.

We have also studied turbocharged aftercooledgasoline engines by simulation and experiment [8]. Weapplied quasi-dimensional simulation on the engineand validated it with experiment results. We alsofound that intercooler fan had considerable e�ects onthe turbocharged aftercooled engine performance atmedium and high engine speeds at WOT and couldbe o� at low loads and speeds.

Ghareghani et al. [9] investigated the thermalbalance of a turbocharged, spark ignition engine fu-elled by natural gas. Variation of output power,transferred energy to cooling uid and exhaust gasenergy at di�erent operating conditions were comparedexperimentally. The maximum thermal e�ciency ofthe turbocharged engine was 4% more than naturallyaspirated at 2500 rpm. Also, the maximum thermale�ciency of the gas engine at 2500 rpm was 4.5% morethan gasoline engine. Finally, based on experimentalresults, an empirical correlation was suggested to pre-dict the energy of exhaust gases by using exhaust gastemperature and air fuel ratio.

Park et al. [10] investigated the applicability ofa preexisting turbocharger designed for a CNG engineand correlated changes to the compression ratio withfuel economy, power, and torque levels of an HCNGengine. The potential bene�ts and the knockingcharacteristics of an HCNG engine with a higher com-pression ratio were assessed as considerations regardingits practical application in the �eld. Thermal e�-ciency was improved with the increase in compressionratio but adverse e�ects occurred and the excess airratio that helps realize maximum thermal e�ciencydecreased from 1.8 to 1.6 for an HCNG engine due to adecrease in exhaust gas energy available for intake-aircharging.

Use of stoichiometric air-fuel mixture with Ex-haust Gas Recirculation (EGR) technique in a SInatural gas engine was experimentally investigatedby Ibrahim and Bari [5]. Engine performance andNO emission were studied for both atmospheric andsupercharged inlet conditions. It was found that theuse of EGR had a signi�cant e�ect on NO emissions.

1426 Sh. Kharazmi et al./Scientia Iranica, Transactions B: Mechanical Engineering 21 (2014) 1424{1439

Ibrahim and Bari [11] also compared the e�ectsof both EGR and lean burn on natural gas SI engineperformance at similar operating conditions. It wasfound that EGR dilution strategy was capable ofproducing extremely lower NO emission than lean burntechnique. NO emission reduced by about 70% whenthe inlet charge was diluted at a rate of 20% using EGRinstead of excess air.

Kesgin [12] studied the e�ects of design andoperational parameters on the NOx emissions of aturbocharged natural gas engine. A gas engine seriesused in combined power plants was optimized regardingpower, e�ciency and emissions. Since evaluation of theemission behavior has become increasingly important,the nitrogen oxide emission and its dependence onengine operational and geometrical conditions were in-vestigated by using a zero-dimensional reaction kineticmodel. The results showed improvement potential fromemission point of view for developing new engines.They concluded that increase in engine speed, andalso increase in excess air ratio caused a signi�cantdecrease in NO emissions. He also showed that chargetemperature highly a�ected the level of NO emissions.In another paper [13], Kesgin predicted the e�ects ofdesign and operating parameters on the NOx emissionof the same engine series, by genetic algorithm andarti�cial neural network. Kesgin also [14] investigatedthe e�ects of turbocharging system on the performanceof the stationary engine. The e�ects of exhaustand turbocharging system parameters on the engineperformance were predicted. Furthermore, he [15]investigated the inlet and exhaust system design ofthe stationary engine. Proper sizing of the inlet andexhaust pipe systems were precisely predicted.

Pourkhesalian et al. [16] performed a detailedcomparison between some conventional and alternativefuels. Engine performance and exhaust emissionswere experimentally studied for gasoline, methaneand methanol in a wide range of engine operatingconditions. For propane, ethanol and hydrogen, athermodynamic model of SI engine was developed andvalidated. It was concluded that volumetric e�ciencyof the hydrogen engine was the lowest (28% less thatgasoline engine), gasoline produce more power thanthe all other alternative fuels, and BSFC of methanolwas 91% higher than that of gasoline while BSFC ofhydrogen was 63% less than that of gasoline.

Korakianitis [1] extensively reviewed use of natu-ral gas as a fuel in reciprocating CI and SI engines.Engine performance, combustion characteristics andemission levels were investigated.

There is more need for experimental and theo-retical studies of natural gas combustion in internalcombustion engines. For instance the turbulent amespeed, ame propagation characteristics, and emissionsgeneration characteristics of pure natural gas in these

engine operating conditions are not well known yet [1].Here, a turbocharged pure natural gas SI engine isinvestigated experimentally. A thermodynamic simu-lation of the turbocharged pure natural gas SI enginein Matlab environment has also been developed andvalidated. Laminar ame speed of Witt and Griebelfor CH4 quoted by Lammle [17] is applied. Mean valuemodeling of Bargende also quoted by Lammle [17] hasbeen used to calculate turbulence parameters duringcompression and combustion processes. Intake owe�ects on Turbulent Kinetic Energy (TKE) have beentaken as initial values. Compression, squish anddissipation e�ects on TKE have been considered instep by step calculations. Gulder equation [18] hasbeen used to calculate turbulent ame speed fromlaminar ame speed. In this research, ame frontarea was calculated considering hemispherical amefront and its intersections with Mexican hat top ofpiston and cylinder walls. Thermal boundary layerwas considered in combustion process to provide abetter agreement for predicted mass fraction burnedat the end part of combustion process when comparedwith experimental results. Thermodynamic propertiesof reactants and products including methane haveprecisely been obtained from correlations based onJANAF thermo-chemical tables [19]. Emission andperformance characteristics of the engine have beenpresented and discussed.

2. Engine speci�cations and experimentalsetup

2.1. Engine speci�cationsThe experimental engine was a Diesel-based tur-bocharged CNG engine. The main engine speci�cationsare summarized in Table 2. The compression ratio ofengine had been reduced from 16.82 to 10.5 throughmodi�cation of piston bowl geometry in order to useCNG instead of diesel fuel. The engine had beenequipped with new cylinder head, which enabled the�tting of spark plugs at the locations of diesel fuelinjectors. The engine had been turbocharged at SharifUniversity turbocharging lab within the frameworkof a previous research and development project incooperation with Sharif University and industry.

2.2. Experimental setupThe experimental investigation was carried out at theturbocharging lab of Sharif University of Technology(SUT); its schematic is shown in Figure 1. The testbed was equipped with an eddy current dynamometerwith an accuracy of �1 Nm and �1 RPM. Ambienttemperature, engine charge air temperature after inter-cooler, charge mixture temperature at inlet manifold,exhaust temperature before turbocharger turbine andinlet and outlet coolant temperatures were measured by


Figure 1. Schematic of turbocharging lab at Sharif University of Technology [20].

Table 2. The test engine main speci�cations.

No. and arrangementof cylinders

6 in line

Displacement volume 11.58 literStroke 128 mmBore 150 mmConnecting rod 280 mmCompression ratio 10.5:1Number of valves 2 inlets and 2 outlets

TurbochargerTwin-entry waste-gatedwater-cooled CNGSchweitzer S300G

Firing order 1-5-3-6-2-4Spark timing 25 �bTDC�mean 1.25Piston bowl vol. 184 cm3

Piston bowl uppersurface area

95 cm2

Combustion chamberin cylinder head vol.

0 cm3

J and K-type thermocouples calibrated within �1 K.Ambient pressure was measured by a strain-gaugeabsolute pressure transducer calibrated with accuracyof �136 Pa. Boost pressure was also measured withaccuracy of �136 Pa.

Since the engine was turbocharged, it was possible

Table 3. Technical parameters of AVL DiCom 4000 [23].

No. Gas Measuring range Measuringaccuracy

1 CO 0-10% by vol. 0.01% by vol.2 CO2 0-20% by vol. 0.1% by vol.3 UHC 0-20 000 ppm by vol. 1 ppm by vol.4 NOx 0-5000 ppm by vol. 1 ppm by vol.

5 O20-4% by vol. 0.01% by vol.4-22% by vol. 0.1% by vol.

to use a bellmouth for engine air consumption measure-ment [21], with throat diameter of 3 inch (76.2 mm).Di�erential static pressure of the bellmouth throat andambient was measured with an accuracy of �10 Pa.The ow calculation equation is:

_qac (kg/s) = 0:022p��P (mmH2O); (1)

where � is �=�s, i.e. ratio of air density at inlet ofbellmouth to that of standard conditions [22].

AVL exhaust gas analyzer of DiCom 4000 wasused to measure exhaust gas composition. Dry-basisvolumetric (or mole) fraction of CO2, CO, O2, UHCand NOx were measured in accordance with the UnitedNations Economic Commission for Europe (ECE) aswell as excess air ratio [23]. The measuring range andaccuracy is according to Table 3.

Spark timing was measured by AVL timing light.Engine cylinder pressure variation with crank angle wasalso measured by Kistler 6118B spark plug piezoelectric


pressure transducer, CA encoder 2614A and TDCsensor of 2629B. Experimental data were presentedaccording to DIN 70020 standard.

All experiments were performed in steady stateconditions at wide open throttle when the enginespeed became steady within �20 rpm, and exhausttemperature became to some extend constant.

3. Modeling

The engine simulation model is a quasi-dimensionaltwo-zone model for describing dynamic behavior duringthe compression, combustion and expansion strokes.In this simulation, the combustion chamber is dividedinto two zones by ame front. The �rst zone containsunburned and the second one contains burned mixture.The ame front is assumed to travel by turbulent ame speed which is a function of laminar ame speed.The engine model uses Hohenberg correlation [24]to estimate engine heat losses. It is assumed thatthe ame propagates in a spherical like shape. Thefrictional losses in an internal combustion engine can becategorized into three main components: The rubbingfriction, the pumping work and the accessory work.The simulation also includes Chen-Flynn friction modelto predict friction mean e�ective pressure [25]. Thecomposition of the reaction products is calculated fromthe chemical equilibrium of the 13 species at a givenpressure and temperature. Finally, using Newton-Raphson method, molar fraction of each species as wellas total mole fraction are calculated.

3.1. Turbulence simulationTurbulence and combustion have twofold e�ects. Tur-bulence wrinkles ame front and increases its areain the combustion chamber. On the other handspeci�c volume of products is about four times that ofreactants [26]. This rapid volume increase of the gases,passing through thin ame front in the limited volumeof the combustion chamber, produces turbulence. Infact, turbulence intensi�es combustion and also com-bustion increases the intensity of the turbulence.

Measurements at di�erent points of combustionzone show that turbulence intensity is relatively homo-geneous. In addition where there is no considerableswirl in intake ow, turbulence intensity is isotropicduring the steps near TDC [26].

The turbulence model has been built up usingmean values (averaged over the entire combustionchamber volume) and no di�erentiation between un-burned and burned zones has been applied. Intake owthrough intake port, which behaves like a jet ow, is thesource of turbulence in the cylinder. Turbulent KineticEnergy (TKE) at the time of Intake Valve Close (IVC)is assumed to be 20% of the intake ow mean velocitysquared [17]. There are 3 terms in k�" equation which

are related to TKE dissipation and TKE productiondue to compression and squish:

dkdt

=�dkcomp

dt

�+�dksquish

dt

��dkdiss

dt

�: (2)

Turbulent kinetic energy production due to compres-sion is calculated by assuming mass and angular mo-mentum conservation of eddy. Turbulent productiondue to squish is calculated from radial and axialcomponents of in-cylinder ow velocity near TDC.Dissipation term is based on assuming that an eddyis dissipated after a few turns [17].

Turbulent kinetic energy is calculated step by stepin each crank angle degree. Turbulence intensity iscalculated from TKE. Integral length scale is obtainedfrom mass conservation of eddy [17].

3.2. Laminar ame speed and EGR e�ectsThere are a lot of ame speed data for CNG andmethane, when mixed with other gases [27-34]. Thereare also some experiment-based relations for puremethane and CNG laminar ame speeds [17,26,35-38].

It is quoted by Lammle [17] that Witt and Griebelproposed methane laminar ame speed relation con-sidering special turbocharged combustion conditions,i.e. higher unburned mixture temperature and cylinderpressure. Witt and Griebel relation is valid for un-burned temperatures up to 850 K, while those of otherslike [35-37] are valid for up to 550 K and extended forup to 650 K at most [38]. Their laminar ame speedis in the following form for mixture of methane and airwithout EGR:

SL = c:p�dcyl : (3)

EGR e�ects on CNG laminar ame speed are consid-ered by experimental relation proposed by Liao [36].Therefore laminar ame speed for mixture of methaneand air containing EGR takes the form:

SL = c:p�dcyl(5:4825:rg2Vol � 4:1988:rgVol + 0:9952):

(4)

In this relation, pcyl is cylinder pressure, rgVol is totalEGR volumetric fraction and c and d are parametersdepending on Tu (unburned mixture temperature) and� (fuel air equivalence ratio) as:

For pcyl < 7 bar:

c =(�1:03� 10�2Tu + 3:645):�2

+ (�4:12� 10�6T 2u + 2:512� 10�2Tu � 7:68):�

+ (8:78� 10�6T 2u � 1:547� 10�2Tu + 4:19); (5)

d = 7:5� 10�1�2 � 1:6� + 1:337� 2� 10�4Tu; (6)


and for pcyl > 7 bar:

c =(�6:906�10�5T 2u+6:875�10�2Tu � 25:13):�3

+(1:155�10�4T 2u�1:1523�10�1Tu+46:47):�2

+(�4:185�10�5T 2u+4:922�10�2Tu�24:82):�

+(6:57�10�6T 2u � 9:55�10�3Tu+5:185); (7)

d = 0:45: (8)

3.3. Turbulent ame speedGulder [39] studied turbulent combustion models andproposed the ratio of premixed turbulent ame speedto laminar ame speed for 3 di�erent combustionregimes (wrinkled laminar- ame, amelets-in-eddiesregime, and distributed-reaction regime). He also usedexperimental ame speed results of methane by [38,39]for validation. Turbulent ame speed appropriate forengine combustion is in the wrinkled regime and isde�ned as:

STSL

= 1 + 0:62�u0SL

�1=2

Re1=4T : (9)

In Eq. (9), ST and SL are turbulent and laminar amespeeds, respectively. ReT is the turbulent Reynoldsnumber which is de�ned as:

ReT =�u0li�U�

�; (10)

where u0 is the turbulence intensity of the cylindercontents, �U and � are the unburned mixture densityand dynamic viscosity, respectively, and li is theintegral length scale.

3.4. Novel approach for the end part ofcombustion

When the ame approaches the cylinder walls, theturbulent length scales that can wrinkle the ame frontreduce and consequently mass burning rate decreases.Proximity of ame to the cylinder walls also leads tohigher heat losses and resulted in decrease of massburning rate. Finally the ame extinguishes due totermination of reactants, low temperature of reactantsand lack of chemical radicals at the proximity of thecylinder walls [40]. Keck [41] proposed an exponentialdecay for �nal mass burning rate. Therefore, it is clearthat turbulent burning velocity by itself is insu�cientto describe a turbulent ame in an engine [40]. Thus,in order to better simulate the end part of combustionin this research, a thermal boundary layer has beenconsidered during the combustion process. There arefew references about the cylinder thermal boundarylayer. Lyford-Pike and Heywood [42] measured the

thickness of thermal boundary layer in an IC engine.This measurement was done by Schlieren approach ondi�erent sections of combustion chamber, i.e. cylin-der liner, cylinder head, and piston during completeoperating cycle. Taking the idea of thermal boundarylayer from [42], it has been assumed in this researchthat thermal boundary layer thickness is equal toquench layer. The temperature of thermal boundarylayer is low due to heat losses to cylinder walls, andits thickness is calculated from mass fraction burnedcurves. This thickness should be equal to a valuethat if be subtracted from cylinder dimensions adjacentto its walls, total mass fraction burned at the endof combustion is equal to corresponding experimentalvalue.

3.5. Assumptions1. Cylinder content is considered to be homogeneous

during compression and expansion processes.2. Two-zone combustion model is used. Each zone is

assumed to be homogeneous.3. It is assumed that all gases are ideal.4. E�ects of blow-by and crevices are ignored.5. The ame front thickness is assumed to be negligi-

ble.6. Cylinder wall temperature is assumed to be con-

stant.7. Combustion chamber wall surfaces in contact with

burned and unburned zones are calculated by con-sidering a hemispherical ame progress with centerat spark plug.

8. Engine is running at steady state conditions.9. Cylinder pressure at the end of each calculating

step is assumed to be uniform through the wholecylinder.

10. Heat transfer between burned and unburned zonesis neglected; however each zone has heat losses withthe cylinder walls.

11. Thirteen species (CO2, CO, H2O, H2, O2, N2,O, H, OH, N, NO, NO2 and unburned fuel asCH4) are considered to be combustion productsand in chemical equilibrium during combustion andexpansion processes.

12. The e�ects of EGR are considered in theoreticalvalidation calculations.

13. Turbulence is assumed to be homogeneous andisotropic.

14. Intake manifold temperature and static pressure ofexperimental results are assumed for mixture at the�rst step of compression.

15. Every degree of crank revolution is considered asone step in theoretical calculations.


3.6. NO formation kinetic modelThe extended Zeldovich mechanism [26] is applied todetermine the rate of change of NO concentrationduring combustion and expansion processes as:

d[NO]dt

=2R1f1� ([NO]=[NO]e)2g

1 + ([NO]=[NO]e)R1=(R2 +R3); (11)

where:

R1 = k+1 [N2]e[O]e = k�1 [NO]e[N]e; (12)

R2 = k+2 [N]e[O2]e = k�2 [NO]e[O]e; (13)

R3 = k+3 [N]e[OH]e = k�3 [NO]e[H]e; (14)

for the principal reactions, respectively, given as:

N2 + Ok+

1$k�1

NO + N; (15)

N + O2k+

2$k�2

NO + O; (16)

N + OHk+

3$k�3

NO + H; (17)

where [ ] denotes the concentration, and the subscripte refers to the equilibrium value. The rate constantsk are in units of m3/kmol/s and have been calculatedfrom GRI-MECH 3.0 [43].

3.7. Theoretical calculationsCalculations are carried out step by step with everydegree of crank revolution as one step. Heat losses fromcylinder contents to its surrounding surfaces is calcu-lated by the Hohenberg semi-empirical relation [24],the most appropriate relation for CNG engines [44].Hence, the �rst law of thermodynamic is satis�ed foreach step. A hemispherical ame front propagatesfrom spark plug located at the center of combustionchamber cylinder head. When this hemisphere radiusis less than li, turbulence does not wrinkle the ame.As the ame radius increases, the e�ect of turbulencebecomes more important. It is assumed that when the ame radius reaches li, turbulent ame propagationis established and its speed is calculated accordingto Eq. (9). Two-zone combustion model is used tosimulate combustion process. Flame front propagationand intersection with Mexican-hat bowl of piston andcylinder liner are precisely calculated based on cylindergeometry. Initial pressure and temperature of mixtureis estimated from experimental results.

3.8. Thermal Boundary Layer (TBL) e�ectsIn order to show the e�ects of thermal boundaryLayer, two di�erent predicted mass fraction burned and

corresponding cylinder pressure for two di�erent op-erating conditions have been presented and comparedwith experimental results of [17] in Figures 2 to 5.Considering thermal boundary layer in combustionsimulation resulted in negligible deviation betweensimulation and experiment curves of the mass fractionburned at end part of combustion as seen in Figures 2and 4. Considering thermal boundary layer alsoresulted in well simulation of combustion duration inCAo. Cylinder pressure was also simulated well, shownin Figures 3 and 5.

Figure 2. Comparison of simulation, with and withoutconsideration of thermal boundary layer, with experimentfor mass fraction burned.

Figure 3. Comparison of simulation, with and withoutconsideration of thermal boundary layer, with experimentfor cylinder pressure.

Figure 4. Comparison of simulation, with and withoutconsideration of thermal boundary layer, with experimentfor mass fraction burned.


Figure 5. Comparison of simulation, with and withoutconsideration of thermal boundary layer, with experimentfor cylinder pressure.

Figure 6. Experimental and simulated cylinder pressurevariation with crank angle at 800 rpm, part load, sparktiming of 25� bTDC and �mean of 1.25.

4. Model validation, parametric study, testresults and discussions

It is necessary to study diesel-based NG engine perfor-mance and emissions by simulation and experimentalinvestigations. Emission generation characteristics ofpure natural gas is not especially well-known, henceexperimental investigation of NG combustion is neces-sary.

4.1. Model validationSimulation model was validated with experimentalresults for the pure CNG SI engine with main spec-i�cations summarized in Table 2. Spark timing was25 obTDC, �mean was 1.25 and mean total EGR massfraction was assumed to be 2%. Figure 6 illustrates theexperimental and simulated cylinder pressure variationwith crank angle. As seen, there is a good agreementbetween simulation and experiment.

Figure 7 shows the values of brake torque andBMEP versus engine speed. Simulation results havegood agreement with those of experiments withinless than 4% deviation. Trend lines of all data arethe same. Torque variation as a function of enginespeed re ects the variation of volumetric e�ciency withengine speed. Falling o� in volumetric, mechanical andthermal e�ciencies causes the decrease of the braketorque at high engine speed. Torque back-up as its bestdescription was 13% according to the de�nition [45]:

Figure 7. Variation of experimental and simulated valuesof torque and BMEP with engine speed at �mean = 1:25,spark timing of 25�bTDC and WOT.

Figure 8. Variation of experimental boost static pressureand its ratio with engine speed at �mean = 1:25, sparktiming of 25� bTDC and WOT.

Torque back-up

=maximum torque�torque at maximum speed

torque at maximum speed:(18)

Figure 7 also illustrates variation of experimental andsimulated values of BMEP of the engine versus enginespeed which is similar to torque variation.

The waste-gated turbocharger has in uences onbrake torque and its back-up by boost pressure asshown in Figure 8. This Figure shows variationof boost pressure and its ratio versus engine speed.Boost pressure increased linearly up to 1200 rpm, andthen its slope decreased. This was due to graduallyopening of waste-gate from 1200 to 1450 rpm. Boostpressure became nearly constant at speeds higher than1450 rpm, as waste-gate opening became nearly con-stant. By waste-gate opening, torque slope decreaseddue to lower boost pressure slope as shown in Figure 7.Opening range of waste gate had substantial e�ects onboost pressure. Figure 8 also shows boost pressureratio variation with engine speed. It became nearlyconstant at 1.2 at high speeds.

Figure 9 presents the experimental and simulatedvalues of brake power at di�erent engine speeds. Therewas a gradual increase in power at high speeds. This


Figure 9. Variation of experimental and simulated valuesof power with engine speed at �mean = 1:25, spark timingof 25� bTDC and WOT.

Figure 10. Variation of experimental and simulatedvalues of air consumption with engine speed at�mean = 1:25, spark timing of 25� bTDC and WOT.

trend can be conveniently justi�ed by the volumetric,thermal and mechanical e�ciency reductions as enginespeed increase. As it can be seen, the simulation resulthad good agreement with that of experiment.

Figure 10 compares variation of simulated air con-sumption values versus engine speed with experimentalresults. As seen, there is a good agreement betweensimulation and experiment.

Figure 11 shows variation of Brake Speci�c FuelConsumption (BSFC) values with engine speed. Thesimulation result has good agreement with that ofexperiment within at most 9% deviation. It shows thatBSFC �rst decreases with increase in engine speed upto 1000-1100 rpm and then increases as speed increasesfurther. Higher ratio of heat loss to inlet fuel energyat lower speeds and increase in friction losses includingpumping losses at higher speeds were the causes of suchvariation.

Figure 12 compares variation of simulated air con-sumption values versus engine speed with experimentalresults. As can be seen, there is a good agreementbetween simulation and experiment.

4.2. Parametric studiesThe validated code has been used to study the e�ects ofexcess air ratio (�) and spark timing on NO emissions

Figure 11. Variation of experimental and simulatedvalues of Brake Speci�c Fuel Consumption (BSFC) withengine speed at �mean = 1:25, spark timing of 25� bTDCand WOT.

Figure 12. Variation of experimental and simulatedvalues of NO emission with engine speed at �mean = 1:2,spark timing of 25� bTDC and WOT.

Figure 13. Variation of simulated NO emissions withexcess air ratio at 2030 rpm, spark timing of 25� bTDCand WOT.

and also the e�ects of excess air ratio (�), EGR andboost pressure on BSFC.

4.2.1. NO emissionsFigure 13 shows the e�ects of lean-burn on engine NOat 2030 rpm, WOT and spark timing of 25� bTDC.NO emission increased with the increase of excessair ratio until it reached to its peak value at about1.1 from stoichiometric mixture. The reason is thatNO formation mechanism depends on both temper-


Figure 14. Variation of simulated NO emissions withspark timing at 2030 rpm, � = 1:234 and WOT.

Figure 15. Variation of Brake Speci�c Fuel Consumption(BSFC) with engine speed at three di�erent excess airratios at spark timing of 25�bTDC and WOT.

ature and oxygen concentration. Although burnedgas temperature is higher for stoichiometric mixture,the availability of oxygen results in the maximum NOemission to be occurred at mixtures slightly leanerthan stoichiometric. The burned gas temperature andconsequently NO emission decreased when excess airratio was increased above 1.1.

Figure 14 indicates the e�ects of spark timing onengine NO at 2030 rpm, WOT and excess air ratio of1.234. NO emission decreased with increase of sparktiming retard. This is due to burned gas temperaturereduction as a result of combustion process shift toexpansion process. At temperatures lower than 1800 K,the extended Zeldovich mechanism is of low importancedue to high activation energy of (1) for detachment ofN atom from N2 molecule with the triple bond.

4.2.2. Excess air ratioFigure 15 gives e�ect of increase of excess air ratio onvariation of Brake Speci�c Fuel Consumption (BSFC)with engine speed. Both minimum and average ofBSFC increased at a rate of 1% for excess air ratios of1.38 and 3% for excess air ratios of 1.5 compared withexcess air ratios of 1.25. Although the temperature ofthe burned gas decreased with the increase of excess air

Figure 16. Variation of Brake Speci�c Fuel Consumption(BSFC) with engine speed at three di�erent EGR massfractions at �mean = 1:25, spark timing of 25�bTDC andWOT.

ratio, combustion duration increased substantially andled to a little increase in BSFC. At higher excess airratios, ame speed decreased and created drivabilityproblems and at most partial burning of the charge.Another e�ect of lean operation is the lower exhaustgas temperature which is of high importance for thee�ciency of the catalyst, if it exists. Engine durabilityand cost improve with the increase of excess air ratio.

4.2.3. Exhaust Gas Recirculation (EGR)Figure 16 indicates e�ects of three EGR mass fractionson variation of BSFC with engine speed. Increaseof EGR at a rate of 20% reduced the minimum andaverage of BSFC, 5% and 8%, respectively. EGRcontaining CO2, H2O and N2 caused higher overallheat capacity of the in-cylinder mixture in comparisonwith O2 and N2 which are normally a part of inletair. It resulted in lower burned gas temperature.CO2, H2O and N2 also dissociated at high ametemperatures and reduced burned gas temperaturemore. Lower burned gas temperature resulted in lowerheat loss importance to cylinder walls and improvedBSFC. The e�ect of lower burned gas temperature washigher than longer combustion duration due to lowerO2 concentration that shifts combustion to expansionprocess. Higher EGR also improves engine mechanicale�ciency; therefore BSFC decreases as EGR increases.

4.2.4. Boost pressure ratioFigure 17 illustrates BSFC variations at three curves ofincrease in the boost pressure ratio, which have similartrend with Figure 8. Here, by increasing the boostpressure ratio up to 0%, 10% and 20%, maximum boostratios of 1.24, 1.37 and 1.49 are obtained, respectively.The boost pressure ratio increase caused BSFC de-crease, because engine friction depends on speed andnot on boost pressure considerably. Therefore highercharge mixture at higher boost ratio produces higher


Figure 17. Variation of Brake Speci�c Fuel Consumption(BSFC) with engine speed at three di�erent boostpressure ratios at �mean = 1:25, spark timing of 25�bTDCand WOT.

Figure 18. Variation of experimental NOx and BSNOx

with engine speed at �mean = 1:25 and WOT.

brake power per 1 g fuel considering nearly constantengine friction with boost pressure change.

4.3. NOx emissionFigure 18 illustrates variation of NOx emission inexhaust gases with engine speed at �mean = 1:25 andWOT. NOx emission occurs in hot combustion prod-ucts. Three main factors which cause NOx emission,are oxygen availability, high temperatures (more than1800 K) and the time duration that combustion prod-ucts remain at high temperature [26]. Enough oxygenwas available in the combustion process at the presentlean running conditions. The spark timing was keptconstant over the speed range. Therefore, comparingwith MBT timing, spark timing was relatively retardat high engine speeds, causing a lower NOx emissiondue to shifting combustion process to expansion strokewith lower ame and combustion product tempera-tures. CNG fuel like other gaseous fuels does nothave the cooling e�ect of vaporization. Therefore, themixture temperature at IVC would be high. The timeduration of high combustion temperature decreases asthe engine speed increases. Also ow pressure dropof gas exchange processes was bigger at higher engine

Figure 19. Variation of experimental CO concentrationwith engine speed at �mean = 1:25 and WOT.

speeds during valve overlap, consequently, residual gasfraction increased and led to decrease in NOx emissionat higher engine speed.

Figure 18 also shows variation of brake speci�cNOx emission in exhaust gases with engine speed at�mean = 1:25 and WOT. Speci�c emission (g/kWh) isderived from emission concentration by:

Brake speci�c emission

= 3600MWemission

MWf

_mf

PbNmix;dry�emission; (19)

where _mf is fuel consumption mass ow rate in g/s,�emission emission volumetric concentration, MWf andMWemission are molecular weight of fuel and emission,respectively, in g/mole, Pb is brake power in kW andNmix;dry is the total mole number of products per moleof fuel CxHy for dry concentrations according to [46]:

Nmix;dry = 4:773�x+ (1� �O2;dry)y=4

1� 4:773�O2;dry

�� y=4;

(20)

in which �O2;dry is the concentration of dry oxygen inexhaust gases.

The BSNOx reduced with the increase of enginespeed due to spark retard (comparing with MBT tim-ing), smaller duration of high combustion temperaturesand lower residual gas (or internal EGR), as explainedfor NOx concentration.

4.4. CO emissionFigure 19 illustrates CO emission concentration in theengine exhaust gases. Poor mixing of air and fuel,incomplete combustion and rich regions produce higherconcentration of CO. It also substantially depends onexcess air ratio. At the lean mixture conditions, CO inexhaust gases was due to poor mixing and incompletecombustion. At constant spark timing during all speedrange, combustion shifted to expansion stroke as speedincreased and caused lower combustion temperatureand higher CO emission.


Figure 20. Variation of experimental UHC and BSUHCwith engine speed at �mean = 1:25 and WOT.

Figure 21. Variation of experimental CO2 and BSCO2

with engine speed at �mean = 1:25 and WOT.

4.5. Exhaust unburned hydrocarbonFigure 20 shows variation of unburned hydrocarbonand BSUHC in exhaust emission with engine speedat �mean = 1:25 and WOT. Flame quenching nearwall surfaces and ame extinguishment within crevices,as well as excessive cooling of the ame front, aresources of HC development in engines at lean andespecially rich conditions. The trend was compatiblewith expectations.

Figure 20 also indicates a minimum value of16 g/kWh for BSUHC at 1120 rpm. As seen fromFigure 11, the point of best BSFC was at speeds around1000-1100 rpm which corresponds to minimum BSFC.This region obviously had the optimum combustionand operation conditions and should provide the mini-mum BSHC.

4.6. CO2 emissionFigure 21 shows variation of exhaust gas brake speci�cCO2 and BSCO2 with engine speed at �mean = 1:25and WOT. BSCO2 had a minimum value of 852 g/kWhat 1030 rpm with an average value of 2.6 g CO2/g NGfor the test. The point of minimum BSCO2 corre-sponded to the point of minimum BSFC and optimumoperating conditions as seen in Figure 11.

Figure 21 also illustrates variation of CO2 in

exhaust gases with engine speed. CO2 reduced at highengine speed and attained a maximum in the mediumspeed range.

5. Conclusions

In this paper, a turbocharged natural gas SI enginewas experimentally investigated at lean operating con-ditions. A model was also developed and validated withexperimental performance results. Good agreementbetween simulated and experimental results showed theaccuracy of the developed model. E�ects of excessair ratio, EGR and boost pressure on BSFC as wellas e�ects of the excess air ratio and spark timing onNO emissions were investigated by parametric study.Experimental emission and performance characteristicsof the engine were studied and discussed. The mainresults are summarized as follows:

1. As a novel approach, it has been assumed thatthermal boundary layer thickness is equal to quenchlayer thickness. The thickness should be equal toa value if be subtracted from cylinder dimensions,total mass fraction burned is equal to correspondingexperimental value. This assumption provides goodagreement between simulated and experimentalmass fraction burned curves.

2. Considering a thermal boundary layer in simulationhas substantial importance in mass fraction burnedcurve, although its e�ect on cylinder pressure sim-ulation is low.

3. Waste-gated turbocharger a�ects the brake torque.By waste-gate opening, torque slope decreases dueto lower boost pressure slope. Torque decreaseswith further opening of waste-gate as engine speedincreases. This results in a better torque back-up.

4. The BSFC is min at medium engine speeds due toa lower heat and friction losses at WOT.

5. NO emission increases with the increase of excessair ratio until it reaches its peak value at about1.1 from stoichiometric mixture and then decreases.NO emission also decreases with increase of sparktiming retard substantially.

6. The use of higher excess air ratio at a rate of 20%increases minimum of BSFC 3% at constant designand operating parameters.

7. It was found that increase of EGR at a rate of 20%improves minimum of BSFC 5%.

8. Gradual opening of waste-gate decreases boostpressure gradient as speed increases at WOT. Thenboost pressure remains nearly constant at speedshigher than 1450 rpm.

9. The BSNOx reduces with engine speed increase. Itis due to the spark retard compared to MBT timing,


shorter time for high combustion temperatures andhigher residual gas mass fraction.

10. The trend of BSCO2 versus engine speed is similarto the BSFC, as expected.

Acknowledgements

The authors acknowledge research assistants for theirengineering support and also lab sta� for their coop-eration at the lab. Research vice-presidency of SharifUniversity of Technology is also acknowledged for their�nancial support.

Nomenclature

k Turbulent kinetic energyli Turbulence integral length scaleMW Molecular WeightN Engine speed (rpm)NOx Nitric oxidesNmix;dry Total mole number of products per

mole of fuelP Pressure (kPa)Pb Brake power (kW)pcyl Cylinder pressure (bar)_qac Engine air consumptionReT Turbulent Reynolds numberrgVol Residual gas volumetric fractionrpm Round per minuteSL Laminar ame speedST Turbulent ame speedTu Unburned mixture temperatureT Temperature (K)u0 Turbulence intensity�mean Mean excess air ratio� Unburned mixture dynamic viscosity� Air density�U Unburned mixture density�s Air density at standard conditions� Air density to its standard density

(= �=�s)� Fuel air equivalence ratio�emisson Emission volumetric (or molar)

emission�O2;dry O2 dry concentration in exhaust gases�P Pressure di�erence (mmHg and

mmH2O)

Acronyms

BMEP Brake Mean E�ective Pressure

BSCO2 Brake Speci�c CO2 emissionBSFC Brake Speci�c Fuel ConsumptionBSNOx Brake Speci�c NOx emissionBSUHC Brake speci�c unburned hydrocarbonbTDC Before Top Dead CenterCA Crank AngleCI Compression IgnitionCNG Compressed Natural GasECE Economic Commission for EuropeEGR Exhaust Gas RecirculationEXP ExperimentalGHG Green House GasIMEP Indicated Mean E�ective PressureIVC Intake Valve CloseMBT Minimum advance for Best TorqueRON Research Octane No.SI Spark IgnitionSIM SimulatedTBL Thermal Boundary LayerTDC Top Dead CenterTKE Turbulent Kinetic EnergyUHC Unburned hydrocarbonWOT Wide Open Throttle

Subscripts

b Burnedcyl. Cylinderf FuelL LaminarT Turbulentu Unburned

References

1. Korakianitis, T., Namasivayam, A.M. and Crookes,R.J. \Natural-gas fueled spark-ignition (SI) andcompression-ignition (CI) engine performance andemissions", Progress in Energy and Combustion Sci-ence, 37, pp. 89-112 (2011).

2. Cho, H.M. and He, B. \Spark ignition natural gasengines - A review", Energy Conversion and Manage-ment, 48, pp. 608-618 (2007).

3. Kharazmi, Sh., Mozafari, A. and Hajilouy-Benisi, A.\Simulating turbulence and combustion in CNG tur-bocharged engine considering cylinder thermal bound-


ary layer and EGR", Proceedings of the Seventh Inter-national Conference on Internal Combustion Engines,Olympic Hotel, Tehran, Iran (November 8-10, 2011).

4. Kandarp, Bhatt K. \Potential for meeting the EU newpassenger car CO2 emissions target", M.Sc. Thesis,MIT University (2010).

5. Ibrahim, A. and Bari, S. \An experimental investiga-tion on the use of EGR in a supercharged natural gasSI engine", Fuel, 89, pp. 1721-1730 (2009).

6. Rothlisberger, R.P. and Favrat, D. \Comparison be-tween direct and indirect (prechamber) spark ignitionin the case of a cogeneration natural gas engine, partI: engine geometrical parameters", Applied ThermalEngineering, 22, pp. 1217-1229 (2002).

7. Rothlisberger, R.P. and Favrat, D. \Comparison be-tween direct and indirect (prechamber) spark ignitionin the case of a cogeneration natural gas engine,part II: Engine operating parameters and turbochargercharacteristics", Applied Thermal Engineering, 22, pp.1231-1243 (2002).

8. Kharazmi, Sh., Hajilouy-Benisi, A. and Mozafari,A. \Computer simulation of turbocharged aftercooledgasoline engine", Proceedings of ESDA2006, 8th Bien-nial ASME Conference on Engineering Systems Designand Analysis, Torino, Italy (July 4-7, 2006).

9. Gharehghani, A., Koochak, M., Mirsalim, M. andYusaf, T. \Experimental investigation of thermal bal-ance of a turbocharged SI engine operating on naturalgas", Applied Thermal Engineering, 60(1-2), pp. 200-207 (2013).

10. Park, C., Kim, C. and Choi, Y. \Power output charac-teristics of hydrogen-natural gas blend fuel engine atdi�erent compression ratios", International Journal ofHydrogen Energy, 37, pp. 8681-8687 (2012).

11. Ibrahim, A. and Bari, S. \A comparison betweenEGR and lean-burn strategies employed in a naturalgas SI engine using a two-zone combustion model",Energy Conversion and Management, 50, pp. 3129-3139 (2010).

12. Kesgin, U. \Study on prediction of the e�ects of designand operating parameters on NOx emissions from alean burn natural gas engine", Energy Conversion andManagement, 44, pp. 907-921 (2003).

13. Kesgin, U. \Genetic algorithm and arti�cial neuralnetwork for engine optimization of e�ciency and NOx

emission", Fuel, 83, pp. 885-895 (2004).

14. Kesgin, U. \E�ect of turbocharging system on the per-formance of a natural gas engine", Energy Conversionand Management, 46, pp. 11-32 (2005).

15. Kesgin, U. \Study on design of inlet and exhaustsystem of a stationary internal combustion engine",Energy Conversion and Management, 46, pp. 2258-2287 (2005).

16. Pourkhesalian, A.M., Shamekhi, A.H. and Salimi,F. \Alternative fuel and gasoline in an SI engine:A comparative study of performance and emissionscharacteristics", Fuel, 89, pp. 1056-1063 (2010).

17. Lammle, C. \Numerical and experimental study of ame propagation and knock in a compressed naturalgas engine", PhD Thesis, Swiss Federal Institute ofTechnology ETH, Zurich, Switzerland (2005).

18. Gulder, O.L. \Turbulent premixed ame propagationmodels for di�erent combustion regimes", 23rd Sym-posium (International) on Combustion, pp. 743-750(1990).

19. NIST-JANAF Thermochemical Tables, published byAmerican Chemical Society and American Institute ofPhysics for National Institute of science and Technol-ogy, 4th Ed. (1998).

20. Kharazmi, Sh., Hajilouy-Benisi, A. and Mozafari, A.,\Experimental investigation of waste gate e�ects onCNG turbocharged SI engine and its turbocharger",Proceedings of the 20th Annual International IranianMechanical Engineering Conference, Shiraz University,Shiraz, Iran (in Persian, Presented Orally) (May 16-18,2012).

21. Martyr, A.J. and Plint, M.A., Engine Testing Theoryand Practice, Elsevier ltd, 3rd Edn. (2007).

22. Garret Airesearch Industrial Division (AID), FlowMeasurement with AID Flow Nozzles (1975).

23. AVL, Operational Manual, Engine Diagnostics, AVLDiCom 4000, AVL (2001).

24. Hohenberg, G.F. \Advanced approaches for heat trans-fer calculations", SAE International, S.A.E. Paper. No.790825, in SP- 449 (1979).

25. Watson, N. and Janota, M.S., Turbocharging the Inter-nal Combustion Engine, MACMILLAN PRESS LTD(1982).

26. Heywood, J.B., Internal Combustion Engine Fun-damentals, New York, McGraw-Hill Book Company(1988).

27. Shy, S.S., Chen, Y.C., Yang, C.H., Liu, C.C. andHuang, C.M. \E�ects of H2 or CO2 addition, equiva-lence ratio, and turbulent straining on turbulent burn-ing velocities for lean premixed methane combustion",Combustion and Flame, 153, pp. 510-524 (2008).

28. Yoshida, A., Naito, H. and Mishra, D.P. \Turbulentcombustion of preheated natural gas-air mixtures",Fuel, 87, pp. 605-611 (2008).

29. Bromberg, L. \In-cylinder laminar ame propagationspeed: e�ects of hydrogen and hydrogen rich gasaddition", MIT Plasma Science and Fusion Center,(August 25, 2005).

30. \Laminar ame speed of stoichiometric methane/airpremixed ame", Application note of CHEMKIN-PRO, Pro-APP-AUTO-3 (v1.0) (May 14, 2008).


31. Ji, M., Miao, H., Jiao, Q., Huang, Q. and Huang, Z.\Flame propagation speed of CO2 diluted hydrogen-enriched natural gas and air mixtures", Energy Fuels,23, pp. 4957-4965.

32. Rozenchan, G., Zhu, D.L., Law, C.K. and TSE,S.D. \Outward propagation, burning velocities andchemical e�ects of methane ame up to 60 ATM",Proceeding of the Combustion Institute, 29, pp. 1461-1469 (2002).

33. Yu, G., Law, C.K. and Wu, C.K. \Laminar amespeeds of hydrogen + air mixtures with hydrogenaddition", Combustion and Flame, 63, pp. 339-347(1986).

34. Bell, J.B., Day, M.S., Almgren, S., Cheng, R.K. andShepherd, I.G. \Numerical simulation of premixedturbulent methane", Lawrence Berkeley National Lab-oratory.

35. Gu, X.J., Haq, M.Z., Lawes M. and Woolley, R.\Laminar burning velocity and Markstein lengths ofmethane-air mixtures", Combustion and Flame, 121,pp. 41-58 (2000).

36. Liao, S., Jiang, D. and Cheng, Q. \Determination oflaminar burning velocities for natural gas", Fuel, 83,pp. 1247-1250 (2004).

37. Elia, M., Ulinski, M. and Metghalchi, M. \Lami-nar burning velocity of methane-air diluent mixtures",Transactions of ASME, Journal of Engineering forGas Turbines and Power, 123, pp. 190-196 (2001).

38. Rahim, F., Elia, M., Ulinski, M. and Metghalchi, M.\Burning velocity measurement of methane-oxygen-argon mixtures and an application to extendedmethane-air burning velocity measurements", Inter-national Journal of Engine Research, 3(2), pp. 81-92(2002).

39. Gulder, O.L. \Turbulent premixed ame propagationmodels for di�erent combustion regimes", 23rd Sym-posium (International) on Combustion, pp. 743-750(1990).

40. Verhelst, S., Sheppard, C.G.W. \Multi-zone modelingof spark-ignition engine simulation - An overview",Energy Conversion and Management, 50, pp. 1326-1335 (2009).

41. Keck, J.C. \Turbulent ame structure and speedin spark-ignition engines", 19th Symposium (Interna-tional) on Combustion/ the Combustion Institude, pp.1451-1466 (1982).

42. Lyford-Pike, E.J. and Heywood, J.B. \Thermal bound-ary layer thickness in the cylinder of a spark ignitionengine", International Journal of Heat Mass Transfer,27(10), pp. 1873-1878 (1984).

43. Smith, G.P., Golden, D.M., Frenklach, M., Moriarty,N.W., Boris Eiteneer, B., Goldenberg, M., Bowman,C.T., Hanson, R.K., Song, S., Gardiner, W.C. Jr.,Lissianski, V.V. and Qin, Zh., Gas Research Institute,http://www.me.berkeley.edu/gri mech/

44. Lounici, M.S., Loubar, Kh., Balistrou, M. and Tazer-out, M. \Investigation on heat transfer evaluation for amore e�cient two-zone combustion model in the case ofnatural gas SI engines", Applied Thermal Engineering,31, pp. 319-328 (2011).

45. Atkins, R.D. \An introduction to engine testing anddevelopment", SAE International, SAE No. R-344(2009).

46. Turns, S.R., An Introduction to Combustion: Conceptsand Applications, 2nd En., McGraw Hill (2000).

Biographies

Shahaboddin Kharazmi received his BSc in Me-chanical Engineering in 2003 and MSc in Automo-tive Engineering in 2005 from Sharif University ofTechnology, Tehran, Iran. He then followed theMechanical Engineering PhD program at Sharif Uni-versity of Technology in 2006 as a talented student,without need to entrance exam. Throughout hisuniversity and industry career, he has worked withautomotive power train system especially internal com-bustion engine. He is author or co-author of over10 national and international conference and jour-nal papers. He has worked as a MSc and PhDstudent in engine research lab and turbocharger &turbocharging lab and developed engine combustionanalysis system as well as eddy current dynamometersystem at Sharif University of Technology. He hastaught Automechanic Workshop and also Heat Trans-fer, Thermodynamics and Fluid Mechanics Labs aswell as being Teacher Assistant in Fluid MechanicsII and Thermodynamics I all in Mechanical Engi-neering Department. of Sharif University of Technol-ogy.

Aliasghar Mozafari received his PhD degree fromUniversity of London (Q.M.C.), UK in 1988. He hasbeen a faculty member of Mechanical Engineering De-partment at Sharif University of Technology (SUT) for38 years. He has been head of Mechanical EngineeringDepartment for six years and Educational Deputy ofMechanical Engineering Deptartment for four years.He has been member of Center of Excellence in EnergyConversion as well as Director of International Stu-dents' O�ce at SUT. He is author or co-author of over90 national and international conference and journalpapers and has supervised over 40 graduate theses. Hehas also published a book in 1993.

Ali Hajilouy Benisi received his PhD degree fromthe Mechanical Engineering Department of ImperialCollege at the University of London in 1993. He hasbeen faculty member of the Institute of Water andEnergy and then the School of Mechanical Engineer-


ing at Sharif University of Technology (SUT) since1978. He served as Director of the Fluid MechanicsLab. from 1993-1995, Founder and Director of theTurbocharger Lab. from 1993, Founder and Directorof the Turbocharging Lab. from 2000, Founder andDirector of the Gas Turbine Lab. from 2008, at theSchool of Mechanical Engineering of SUT. He has alsoserved as Research Director of the SUT from 1993-1995.His research interests are experimental and theoretical

investigations of turbochargers, turbocharging, and gasturbines.

Dr Hjilouy Benisi is author or co-author of over90 conferences and journal papers, and coordinator forcompiling and publishing four volumes research reportsof SUT in 1991-1995. He and one of his students aretranslators of an English text book to Farsi (Persian).He has also supervised over 40 graduate theses. He hasconducted more than seven research funded projects.

Documents

Simulation and experimental investigation of performance