International journal of automotive technology, tập 10, số 1, 2009

On comparison of the cases in Figure 3b where the equivalence ratio is 0.3, the cylinder pressure during combustion rises fast enough to cause obvious knock in the premix tests as charac

(Received 8 April 2008; Revised 10 June 2008)

ABSTRACT−Compression ignition of homogeneous charges in internal combustion (IC) engines is expected to offer highefficiency of DI diesel engines without high levels of NOx and particulate emissions This study is intended to find ways ofextending the rich limit of HCCI operation, one of the problems yet to be overcome Exhaust emissions characteristics are alsoexplored through analyses of the combustion products DME fuel, either mixed with air before induction or directly injectedinto the combustion chamber of a rapid compression and expansion machine, is compressed to ignite under various conditions

of compression ratio, equivalence ratio, and injection timing The characteristics of the resulting combustion and exhaustemissions are discussed in terms of the rate of heat release computed from the measured pressure, and the concentrations ofTHC, CO, and NOx are measured by FT-IR and CLD The experimental data to date show that operation without knock ispossible with mixtures of higher equivalence ratio when DME is directly injected rather than when it is inducted in the form

of a perfectly homogeneous fuel-air mixture Although fuel injected early in the compression stroke promotes homogeneity

of the DME-air mixture in the cylinder, it causes the mixture to ignite too early to secure good thermal efficiency and free operation at high loads Low temperature reactions occur at about 660K regardless of the fueling methods, fuel injectiontiming and equivalence ratio The main components of hydrocarbon emissions turned out to be unburned fuel (DME),formaldehyde and methane

knock-KEY WORDS : DME, HCCI, RCEM, Exhaust emissions, Perfectly homogeneous fuel-air mixture, Direct fuel injection

1 INTRODUCTION

New technologies that increase thermal efficiency and clean

exhaust gas of automobile engines are desperately being

sought to mitigate the problems of energy shortage &

en-vironmental pollution Homogeneous charge compression

ignition (HCCI) is one of those representative technologies

being developed in the field of engine combustion HCCI

engines are expected to have higher efficiency than SI

engines due to their high compression ratio and

dispens-ability of throttle valves Compared to diesel engines, they

emit less particulate matters and less NOx because only

lean premixed combustion without local fuel-rich zones is

present (Gray and Ryan, 1997; Thring, 1989; Chung et al.,

2008) Despite these advantages, commercial mass

produc-tion of HCCI engines have not yet been realized One of

the obstacles to the commercialization of HCCI engines is

objectionable knock occurring under heavy load conditions

Knock results from an excessive rate of pressure rise in the

combustion chamber (Gray and Ryan, 1997) Methods of

utilizing mixture stratification (Inagaki et al., 2006; Kumano

and Iida, 2004; Sjöberg and Dec, 2006) and thermal

strati-fication (Dec et al., 2006; Lim et al., 2006; Sjöberg et al.,

2005) have been suggested to resolve this difficulty byrelaxing almost simultaneous heat release of HCCI com-bustion In actual engines, various causes for non-homo-geneity exist including imperfect mixing of fuel, air, andresidual or EGR gas, differences in gas flow and boilingpoints of fuel blends, and varying levels of heat transferalong the cylinder walls

The objective of this study is to investigate the effects ofstratification caused by uneven fuel-air mixing on the HCCIcombustion and consequent emissions Different ways ofadding fuel to air are tried including premixing and directinjection at various timing A rapid compression and ex-pansion machine (RCEM) is used in this study in order tofocus on the effects of fuel-air mixing caused stratificationwith minimized contribution of other factors like residualgas and gas motion

DME (di-methyl ether) is the selected test fuel, since itscombustion-related properties are close to those of dieselfuel, and it burns cleanly without generating soot DMEalso exhibits two-stage heat release, which is one of the

distinctive characteristics of HCCI combustion (Ogawa et

al., 2003; Teng et al., 2004; Yao et al., 2003).

2 EXPERIMENTAL SYSTEM AND TEST SCOPE

As shown in Figure 1, the experimental system consists of

*Corresponding author e-mail: mtlim@chonnam.ac.kr

an RCEM and subsystems for fuel supply and injection, for

heating cylinder charge, and for acquisition of pressure and

temperature signals Basic structure of the RCEM

(specifi-cations shown in Table 1) is similar to that of a

recipro-cating piston engine The RCEM is a single-shot engine,

whose crankshaft is driven by the kinetic energy stored in a

large flywheel Before the combustion test, residual gas is

removed from the combustion chamber by a vacuum pump,

and fresh air is charged Design and specifications for this

RCEM are explained elsewhere in detail (Cho et al., 2004).

The initial temperature of the air or fuel-air mixture in the

combustion chamber of the RCEM is controlled by

operat-ing the electric heater wrapped around the cylinder head,

and also by circulating heated water through the jacket

around the cylinder

The DME fuel is supplied to the combustion chamber

either premixed with air or by direct injection Fuel and air

are mixed and stabilized for perfectly homogeneous charge

tests approximately a day before use Direct injection of

liquid DME is accomplished at 50 MPa by controlling the

timing and duration of the injector open time Table 2

provides the major specifications of the fuel injector in

detail An in-house designed computer program performs

the injection control based on the crank angle signal from a

rotary encoder (Autonics, E40S-360-3-5) Since viscosity

of DME is insufficient for self-lubrication of the nozzle

(0.33 cSt at 293 K and 50 MPa), 750 ppm of viscosity

improver (ETHYL, hitec-4140, 17.2 cSt) is added to DME

(Longbao et al., 1999)

Combustion tests were run with fuel introduced eitherpremixed or directly injected into the combustion chamberunder various conditions of equivalence ratio and injectiontiming as described in Table 3

Combustion characteristics are analyzed using the ssure in the combustion chamber taken by a piezoelectricsensor (Kistler, 6016B) and a charge amplifier (Kistler,5011A) Exhaust gas or combustion product is displacedout of the cylinder and fed to the analyzers after each run ofthe combustion tests Engine exhaust gas analyzers areinapplicable to RCEM tests because RCEM conducts onlyone cycle at a time, producing a fixed volume of the ex-haust gas (about 3.5 dm3 in the present case) which is toosmall to keep a continuous sample flow to the analyzers.Concentrations of unburned hydrocarbon (HC) and carbonmonoxide (CO) are therefore measured by an FT-IR(MIDAC, I2004) that requires about 0.3 dm3 of sample gas,and oxides of nitrogen (NO and NOx) are measured by aCLD (Thermo Environmental Instruments, 42C-HL) thatrequires about 25 cc/min of sample gas

pre-Figure 1 Schematic diagram of the experimental system

Figure 2 Definitions of LTR and HTR

Table 1 Specifications of RCEM

Premixed Direct injection

valence ratio of 0.2 The LTR and HTR are found to be

associated with formation of formaldehyde and formation

of CO and CO2 respectively (Hamada et al., 2005).

Figure 3(a) shows a group of curves representing

respec-tively the motored cylinder pressure, and fired cylinder

pre-ssure along with the associated ROHR and the gas

temper-ature during a test run at an equivalence ratio of 0.1 This

figure gives an overall idea about what happens during the

entire period of compression and expansion strokes in the

combustion chamber during the test Figure 3(b) shows the

similar sets of data for the combustion tests run at

equi-valence ratios of 0.1 and 0.3, but the range of data is

restricted only to the late phase of compression and early

phase of expansion The blue curves in the latter figure

denote the results from the tests run with perfectly

homo-geneous fuel-air mixture supplied, while the red lines denote

those run with fuel directly injected at the start of

com-pression (i.e at bottom dead center) The combustion

pre-ssure in figure (a) gradually rises at first during the

com-pression, just tracing the motored cylinder pressure before

the start of reactions at about 19°BTDC, and then it rapidly

increases thereafter The in-cylinder gas completes

com-bustion, and its temperature reaches 930 K at TDC The

two peaks of LTR and HTR appear after ignition on ROHR

curve Although auto-ignitions are observed to occur at

about 20o~15o BTDC, in all the cases shown in Figure 3,

the early burning rate in the cases of the premixed

com-bustion is much larger, resulting in a shorter comcom-bustion

duration and higher peak pressure On comparison of the

cases in Figure 3(b) where the equivalence ratio is 0.3, the

cylinder pressure during combustion rises fast enough to

cause obvious knock in the premix tests as characterized by

the sharp peak and the fluctuations in the pressure curve,

but knock is not observed in the case of direct injection test

Figure 4 shows the maximum rate of combustion

pre-ssure measured in tests with various equivalence ratio and

fueling strategies In the cases of premixed mixture the

maximum rate of pressure rise exceeds 4.3 MPa/ms (Lim et

al., 2006), which is taken as knock borderline in this study,

at equivalence ratios over 0.25 However, knock is not seen

to take place at equivalence ratios of up to 0.4 when fuel is

directly injected at various injection timing This implies

that the rich operation limit of HCCI combustion can be

extended by direct injection of fuel The effect may be

attributed to the stratification of the mixture and latent heat

of vaporization The maximum rate of pressure rise seems

to gradually increase as either the equivalence ratio increases,

or the fuel injection is retarded at equivalence ratio over0.35

Figure 5 represents the timing and duration of LTR andHTR in the tests with various fuel injection timing andequivalence ratios The left, blue portions of the bars indi-cate time for LTR while the right, red portions indicate thisfor HTR Although the ignition delay is significantly short-ened, the start of LTR and HTR is delayed less than 5° CA

as the fuel injection is retarded from BDC to 30° BTDC.When compared with the premixed fuel-air mixtures, theignition starts later, and combustion lasts longer whenmixtures are formed through direct injection The ignitiontiming, taken as start of HTR, is slightly delayed when fuel

Figure 3 Cylinder pressure, temperature and rate of heatrelease as a function of crank angle for DME-air mixtures

of different equivalence ratios and mixing processes

injection is substantially delayed The delay is bigger as the

equivalence ratio increases, possibly due to the cooling

effect of the fuel vaporization and also due to the higher

degree of mixture stratification Mixtures of greater

equi-valence ratio also have shorter combustion durations

(especi-ally for HTR)

Figure 6 shows net heat release and combustion

effici-ency in tests of mixtures with various equivalence ratios,

where fuel is either premixed or directly injected at

differ-ent crank angles Hatched columns in the figure indicate

total amount of released heat, while green curves represent

the corresponding combustion efficiency The combustion

efficiency is rather low, between 40 and 70 percent, with

higher values associated with greater equivalence ratio

Further observation reveals that the combustion efficiency

does not depend on the fuel injection timing at equivalence

ratio of 0.2 or 0.3, but it gradually decreases at equivalence

ratio of 0.1 as the injection is retarded from BDC toward

30° BTDC

Figure 7 shows bulk combustion temperature of the cylinder mixtures during the combustion tests under vari-ous fueling strategies The blue bottom portion, the redupper portion, and the highest point of the bars, respec-tively, indicate the temperatures during LTR, during HTR,and at the final stage of combustion

in-The in-cylinder gas temperature before the ignition iscalculated from the equation (1) assuming an adiabaticcompression process because the fuel-air mixture in thecentral part of the combustion chamber, where the ignition

is most likely to occur first, gets least influenced by the

heat loss to the walls (Iida et al., 2004; Sato et al., 2003).

(1)

Since the adiabatic assumption included in the isentropicprocess is not valid once ignition has occurred, the gastemperature is calculated using the equation (2), the idealgas equation of state

T ( )=Tθi (θi 1– ) P( )θi

P(θi 1– ) -

⎛ ⎞γ 1 -–γ

⋅

Figure 4 Maximum rate of pressure rise and knock as a

function of crank angle for DME-air mixtures of different

equivalence ratios, formed by premixing or direct injection

Figure 5 Combustion timing as function of fuel injection

timing and equivalence ratio

Figure 6 Released heat and combustion efficiency as afunction of equivalence ratio and injection timing

Figure 7 Combustion temperature as a function of valence ratio and injection timing

(2)

T [K], P [Pa], V [m3], n [mol], γ, and θ[deg.] in the above

equations represent the temperature, pressure, volume,

specific heat ratio of the gas in the cylinder, and the engine

crank angle, respectively A specific heat ratio of 1.3 was

obtained from the logarithmic plot of the measured

pre-ssure and volume pertaining to the engine operating

condi-tions

It is clear from the figure that LTR begins at about the

same temperature of 660 K ± 10 K regardless of either the

equivalence ratio or the fuel injection timing

Figure 8 shows indicated mean effective pressure (imep)

obtained under various fueling conditions As the fuel

injec-tion is retarded from BDC to 30° BTDC, imep steadily

increases a little (up to about 0.05 MPa) except at the leanest

condition The general trend of increasing imep with

delay-ed ignition, causdelay-ed by retarddelay-ed fuel injection, is attributdelay-ed

to the decreasing compression work due to the delay in

pressure rise during the compression stroke

fuel injection The trend can be easily explained since fuelinjected in the center of the combustion chamber at anearlier timing will have more chance of getting trapped inthe top ring crevice, which is considered to significantlycontribute to DME emission in HCCI operation Amongthe cases where fuel is supplied in the same manner, THCseems to decrease as the equivalence ratio is raised from0.1 to 0.2, and then slightly increases (premix fueling), orstays about at the same level (fuel injected at BDC) forfurther increases of equivalence ratio from 0.2 to 0.3 Themajor decrease of THC in the first interval of equivalenceratio results from a similar drop in DME Since equival-ence ratio 0.1 is close to the lean limit of auto ignition,which will tend to help more DME near the walls or in thecrevices avoid combustion, the highest concentration ofDME and the low combustion efficiency is measured atequivalence 0.1 When the equivalence ratio is increased to0.2 or above, more DME will burn to reduce THC in theexhaust gas, while more DME becomes trapped in thecrevices to increase THC The latter effect will play a moreimportant role when the cylinder charge is more homo-geneous, as explained above, and also become more prono-unced at conditions of greater equivalence ratio This may

be the factor to explain the significant, in the case of mixed fuel, or marginal, in the case of fuel injected at BDC,increase of DME and also THC at the equivalence ratio of0.3

pre-Summarizing the above discussions, one may cautiouslymodel HC formation in these cases so that the prevailingmode in the very lean range of the equivalence ratio below

Figure 9 Concentration of unburned hydrocarbons in

combustion products of DME-air mixtures formed by

premixing or direct injection

Figure 10 NOx and CO concentrations in combustionproducts of DME-air mixtures formed by premixing ordirect injection

0.1 is bulk quenching of the extra-lean mixture due to

incomplete ignition or cold walls in the neighborhood, while

the crevice volume effect is the dominant one in the richer

range

Figure 10 shows NOx and CO concentrations in the

combustion products of mixtures formed through either

premixing or direct injection Smaller amounts of CO are

generated at higher equivalence ratios regardless of the

fueling strategy, most likely due to higher combustion

temper-ature associated with the greater equivalence ratios

Pre-mixed mixtures look to generate a little more CO at every

equivalence ratio than do mixtures prepared by direct

injec-tion This is considered quite natural because more fuel will

exist in the close vicinity of the cold walls if fuel is

pre-mixed before induction than if it is injected later near the

center of the cylinder NOx concentration appears to be

roughly proportional to equivalence ratio, which is logical

because of the expected variation of the combustion

temperature Nevertheless the absolute levels of NOx

concen-tration are fairly low in comparison with those in the

conventional diesel combustion

4 CONCLUSIONS

Characteristics of DME HCCI combustion and

composi-tion of the combuscomposi-tion product were experimentally

investi-gated, while various quantities of DME fuel were

introdu-ced by different methods in the combustion chamber of an

RCEM Major findings of the study are as follows

(1) The knock-limited rich limit of HCCI operation is

ex-panded by directly injecting fuel

(2) A homogeneous DME-air mixture is associated with

shorter combustion duration and higher peak pressure

than a mixture formed by direct fuel injection

(3) The timing of LTR is slightly delayed when fuel

injec-tion timing is retarded

(4) Imep increases with delayed timing of fuel injection

(5) The main components of THC are unburned fuel (DME),

formaldehyde and methane, and more fuel remains

unburned when fuel is supplied premixed

(6) More THC and CO are formed for any equivalence

ratio tested when fuel is introduced premixed than when

it is directly injected at BDC

(7) As fuel injection is retarded, ignition occurs later with

less THC and more CO formed

(8) As equivalence ratio is raised, THC emission decreases

in the low range of equivalence ratio (0.1 to 0.2), but it

increases (fuel premixed) or levels off (fuel injected) in

the higher range of equivalence ratio (0.2 to 0.3) More

nitrogen oxides are generated as the equivalence ratio

increases, but their absolute levels are quite low

ACKNOWLEDGEMENT−This research was conducted as a

part of “Development of Basic and Practical Technology for

HCCI Engine” under the financial sponsorship of the Ministry of

Knowledge Economy

REFERENCESCho, S H., Kim, K S and Lim, M T (2004) Development

of a rapid compression expansion machine and ssion ignition combustion of homogeneous premixtures

Chung, J W., Kang, J H., Kim, N H., Kang, W and Kim,

B S (2008) Effects of the fuel injection ratio on theemission and compression performances of the partiallypremixed charge compression ignition combustion engine

applied with the split injection method Int J Automotive

2006-01-Gray, A W and Ryan, T W (1997) Homogeneous charge

compression ignition (HCCI) of diesel fuel SAE Paper

No 971676.

Hamada, K., Niijima, S., Yoshida, K., Shoji, H., Shimada,

K and Shibano, K (2005) The effects of the ssion ratio, equivalence ratio, and intake air temperature

compre-on igniticompre-on timing in an HCCI engine using DME fuel

SAE Paper No 2005-32-0002.

Iida, N., Yamasaki, Y., Sato, S., Kumano, K., Kojima, Y.(2004) Study on auto-ignition and combustion mechanism

of HCCI engine SAE Paper No 2004-32-0095.

Inagaki, K., Fuyuto, T., Nishikawa, K., Nakakita, K andSakata, I (2006) Dual-fuel PCI combustion controlled

by in-cylinder stratification of ignitability SAE Paper

compression machine SAE Paper No 2006-01-3319.

Longbao, Z., Hewu, W., Deming, J and Zuohua, H (1999).Study of performance and combustion characteristics of

a DME-fueled light-duty direct-injection diesel engine

SAE Paper No 1999-02-3669.

Ogawa, H., Miyamoto, N and Yagi, M (2003) kinetic analysis on PAH formation mechanisms of oxy-

Chemical-genated fuels SAE Paper No 2003-01-3190.

Sato, S and Iida, N (2003) Analysis of DME

homogene-ous charge compression ignition combustion JSAE Paper

No 20030236.

Sjöberg, M and Dec, J E (2006) Smoothing HCCI release rates using partial fuel stratification with two-

heat-stage ignition fuels SAE Paper No 2006-01-0629.

Sjöberg, M., Dec, J E and Cernansky, N P (2005) tial of thermal stratification and combustion retard forreducing pressure-rise rates in HCCI engines, Based on

Poten-multi-zone modeling and experiments SAE Paper No.

EXPERIMENTAL INVESTIGATION OF CHARACTERISTICS OF PRESSURE MODULATION IN A FUEL INJECTION SYSTEM

D HUANG1,2)* and M.-C LAI1)

1)Mechanical Engineering Department, Wayne State University, Detroit, Michigan 48202, USA

2)Mechanical Engineering Department, Shanghai University, Shanghai 200072, China

(Received 25 July 2007; Revised 24 April 2008)

ABSTRACT−A piezoelectric atomization device achieves fuel pressure modulation through vibration of a piezoelectricpressure modulator As a consequence, the fast alternating and slow moving streams collide with each other and further break

up the fuel drop In this paper, an experimental investigation was carried out to study the fluid dynamic characteristics of thespray atomization process of automotive port fuel injectors with a piezoelectric pressure modulator The investigation mainlyfocuses on: (a) the coupling characteristics between the piezoelectric stack and the hydraulic as well as the transfercharacteristics of pressure modulation from the piezoelectric modulator to the point above the orifice; (b) the time history ofthe pressure dynamic response at the point above the orifice under a typical modulation frequency, which reflects the variation

of pressure modulation while the fuel injector is working; and (c) the time-variation characteristics related to mechanicalstructure and fluid dynamics The experimental results expose some important dynamic characteristics of pressure modulation,which will be very significant and lead us to greatly improve the fuel injection system, optimize the control parameters andimplement spray atomization with a high quality performance in the near future

KEYWORDS : Pressure modulation, Fuel injection system, Dynamic characteristic, Spray atomization

1 INTRODUCTION

Spray atomization is very important to reduce the fuel/air

mixing time and minimize the attachment of liquid fuel to

port surfaces in automotive and turbo fan engines

Well-atomized gasoline spray has a high potential to reduce

hydrocarbon emissions and improve the engine cold

start-ing properties Dressler developed a pressure modulator

based on the piezoelectric principle (Dressler, 1993) The

pressure modulator is installed inside the fuel line and

generates the pressure modulation It efficiently enhances

the atomization characteristics of gasoline spray injectors

(Zhao et al., 1996; 1995; 2002; Sipperley et al., 1998;

Schiller et al., 2006; Kim at al., 2004)

To optimize the performance of the pressure modulation

fuel system in spray atomization (Hu and Wu, 2001a;

2001b), presented mathematical modeling of an individual

injector and an entire fuel injector system by considering

one-dimensional, unsteady Bernoulli’s equation and loss

factors of kinetic energy Kf and Ko based on a discrete

segment of the injector (loss factors Kf and Ko are used to

account for the losses of kinetic energy as fluid enters the

injector through the filter at the top and discharges through

the orifice at the bottom, respectively) (Miller, 1990) They

predicted the dynamic response of an automotive fuel

system under low frequency pressure fluctuation (around

330 Hz) produced by a pressure modulation machine.Although different models (Ren and Mally, 1996) havebeen developed to describe the fuel injection system, fewcan accurately predict the transient response of a fuel injec-tion system, especially its internal pressure modulation Inorder to fill the defect of the numerical computation, thepresent paper is experimentally investigating the fluiddynamic characteristics in a fuel injection system, whichcontains a piezoelectric modulator This study was motivat-

ed by the success of earlier work on fuel spray atomization

in the automotive port injector After carefully observingthe coupling phenomenon of the piezoelectric-hydraulicsystem and the dynamic response of the fuel injectionsystem with pressure modulation, the characteristics of pre-ssure modulation in a fuel injection system are summari-zed Along with this summary, this manuscript also de-scribes the experimental setup, procedures and signal pro-cessing employed in the experiments Experimental resultsand future work are then discussed

2 EXPERIMENTAL SETUPFigure 1 shows the scheme of an experimental pressuremodulated port fuel injector system The pressure modu-lator, as shown in Figure 1(a), is composed of a piezo-electric driver mounted inside a circular housing and bolted

to one end The other end of the circular housing is closed

by a rigid endplate upon which the fuel injectors are

*Corresponding author e-mail: hdishan@shu.edu.cn

mounted by means of an adapter The piezoelectric driver

consists of a basemount, a pair of piezoelectric disks, and a

piston A hollow bolt clamps these items together into a

cylindrical assembly The liquid from the fuel tank,

pre-ssurized by high pressure nitrogen gas, is introduced into

the piezoelectric driver through the basemount It flows

through the hollow bolt and a passage in the piston and

then enters the fluid manifold surrounding the piston The

fluid flows through a hole in the end plate and the injector

adapter, and then enters the gasoline port injector

The piezoelectric elements receive the electrical signal

and convert it into a longitudinal motion Such a process

produces a transient variation in the pressure of the fluid as

it passes through this gap, allowing for pressure

modu-lation of the fuel system Therefore, the device here is

referred to as a pressure modulator It should be denoted

that even though there is significant pressure perturbation

inside the fuel line, the liquid flow rate remains nearly

constant While the device can operate at many

frequen-cies, the optimized operating point can be determined in

order to minimize energy loss and maximize pressure

per-turbation The device has a small size and can be installed

easily between the port fuel injector and the fuel line There

is no extra control difficulty associated with this system

since the fuel injection rate and injection timing are

1992) The CPI injector has a single spray hole with adiameter of 500 μm, and is connected to the pressuremodulator through an adaptor The baseline case is thespray with static pressure of 276 kPa (40 psi) regulated bythe pressure of the nitrogen gas The pressure modulatorreceives a sinusoidal signal from a wave generator operat-ing between 3 and 30 kHz, and the signal is amplified by anamplifier before it is used as a modulation voltage driving.The final signal delivered to the crystal stacks is a 25~250

V sine wave When the CPI injector is controlled by a pulsewave, it alternatively opens and closes and the fuel passesthrough the orifice as instructed The pulse width is set at

16 ms for the pressure modulation test

To measure the fluid dynamic pressure in the electric pressure modulator and injector, three KulteXT-123C-100 pressure sensors are installed The first one islocated at the point above the injector orifice and itsresponding pressure is P1, which directly affects sprayatomization The second one is near the entrance of theinjector chamber and its responding pressure is P2 Thethird one is located in the middle of the pressure modulatorand its responding pressure is P3, which directly reflectsthe source of pressure modulation All sensors are connect-

piezo-ed to computer data acquisition

A computer collects data from the measuring points andprocesses the data for different statuses With this setup,one can simulate various working conditions with differentmodulation frequencies, extract the dynamic characteristics

of pressure modulation in the fuel injection system andobserve the fuel spray atomization

3 COUPLING CHARACTERISTICSThe measurement of the coupling characteristic betweenthe piezoelectric stack and the hydraulic system is carriedout with the port injector being closed The single frequ-ency sine wave is applied to the piezoelectric stack, and as

a result, the pressure modulation is generated in the fuelinjection system By recording the pressure P1 at a pointabove the orifice of the injector and its modulation frequ-ency, and using the signal filtering process to extract thepressure modulation component, the responding amplitude

is given for different driving voltages Repeating this cedure for a series of modulation frequencies, the measure-ment result for the coupling characteristics is obtained, asFigure 1 Schematic of experimental setup

pro-shown in Figure 2, where the driving voltage varies from

25 to 250 V and the frequency range is from 1 to 17 kHz

The experimental data analysis shows a strong nonlinear

phenomenon in the piezoelectric stack and hydraulic

coupl-ing, which depends on the driving voltage over the

piezo-electric stack and modulation frequency At several

frequ-ency points, the magnitude of pressure modulation is not

sensitive to driving voltage When the driving voltage is

greater than 50 V, the increment of pressure modulation

almost halts In this application, using this nonlinear

characteristic, the biggest magnitude of pressure

modu-lation can be attained at some special frequency points with

smaller driving voltages For example, around frequencies

of 8 kHz, 9 kHz and11.5 kHz, satisfying satisfactory level

of pressure modulation can be obtained through a driving

voltage of 50 V The optimized control parameters can save

the driving power in the fuel injection system Although

different characteristics will occur in different structures of

the piezoelectric stack, through experimental investigation,

we can determine the optimal driving voltage for pressure

modulation

In fact, saturation status of the coupling characteristic of

pressure modulation is almost attained for the given

piezo-electric stack while the driving voltage is close to 250 V in

the range of working frequencies from 6 kHz to 17 kHz

There is no benefit to using a higher voltage as a driving

power Therefore, a voltage of 250 V is employed in

ex-amining the coupling characteristic of pressure modulation

in the fuel injection system study to obtain the best output

property over the given frequency range

The coupling characteristics of pressure modulation are

very sensitive to minor changes in the mechanical structure

of the fuel injection system While the port injector is open,

the coupling characteristics vary with flowing fuel A

detai-led representation of the change in coupling characteristics

during the injector opening is presented in Figure 3, where

the driving voltage is 250 V Under the given voltagecondition, the coupling characteristics at P1 are greatlyaltered below a frequency of 7 kHz while the port injector

is opened, and keeps the same in the higher frequencyrange Notably, the curve of the coupling characteristic tends

to be smooth over the whole working frequency range inthis situation However, the maximum output magnitude ofpressure modulation is worth considering, and it can bedetermined from the given coupling characteristics At afrequency of 5.4 kHz, the maximum output value is about

22 kPa (rms value) For the same reason, a similar omenon can be observed at measuring point P3

phen-The coupling characteristics of pressure modulation arerelated to the static pressure in the fuel injection system.The magnitude of pressure modulation at P1 is proportional

to the static pressure when the port injector is closed, and

Figure 2 Measurement result of coupling characteristic

between the piezoelectric stack and the hydraulic, where

static fuel pressure is 276 kPa (40 psi) on the CPI injector

and the sine wave driving voltage is 25~250 V

Figure 3 Comparison of coupling characteristics withclosed and opened CPI injector, where static fluid pressure

is 276 kPa (40 psi) and the driving voltage is 250 V

Figure 4 Relationship between the magnitude of pressuremodulation and static pressure, where the driving voltage is

150 V over the piezoelectric stack while the CPI injector isclosed

this is confirmed by the detailed results shown in Figure 4.

It also can be seen that different modulation frequencies

have different magnitudes under a constant driving voltage

of 150 V

4 TRANSFER CHARACTERISTICS

The term transfer characteristics, in particular, indicate the

transmission properties of pressure modulation from the

source of the pressure modulation to the point above the

injector orifice It also expresses the pressure modulation

difference between P1 and P3 The transfer characteristics

mainly depend on the route through which the fuel flows,

working status (whether the port injector is closed or open),

and the quantity of fuel sprayed per second While the

pressure signals P1 and P3 are sampled at the same time,

the transfer characteristics of the pressure modulation can

be obtained by signal processing, as shown in Figure 5,

where the driving voltage is 250 V Figure 5(a) is the port

injector closed case, and Figure 5(b) is the port injector

opened case

Though the route between the measuring points P1 and

P3 is very short, wherever the injector is closed or open,

due to the complicated inner space structure of the CPI

injector, the transfer characteristics of the pressure

modu-lation cannot remain constant in the fuel fluid and take a

multiple harmonic form

5 RESPONSE CHARACTERISTICS

Response characteristics mainly involve two aspects of the

fuel injection system and pressure modulation One part is

the dynamic response that relies on the structure of the fuel

injection system, the physical property of fuel fluid and the

static pressure The other is the variation of pressure

modulation, which we focus on in this paper As the port

injector is working, the pressure signal P1 is recorded.Some typical records, such as the modulation frequencies

0, 2.2, 4.0, 8.0 kHz reflecting the variation of pressuremodulation, are plotted in Figure 6, where the transientdynamic response and a 16 ms width pulse control signalcan be seen clearly To make a comparison with anotherresult, the time history records at the measuring point P3are shown in Figure 7

No modulation case (Modulation Frequency is 0 Hz):

While the injector is open, the pressure drops immediatelybecause of rarefaction in the fluid When the injector issuddenly closed, compression occurs in the fluid system,which causes a strong pressure surge known as the “waterhammer” effect on the fuel system This dynamic responseoscillates, decaying exponentially to the ambient level due

to fluid viscosity It should be mentioned that while theinjector is open, the pressure signals P1 and P3 also exhibit

a dynamic response, but the observing time is too short for

us to distinguish them clearly

Modulation case: To estimate the variation in the

Figure 5 Transfer characteristics of pressure modulation

while the CPI Injector is (a) closed and (b) opened, where

the static fluid pressure is 276 kPa (40 psi) and the driving

voltage is 250 V

Figure 6 Time history record of pressure signal P1 whilethe CPI injector is working, where the static fluid pressure

is 276 kPa (40 psi) and the driving voltage is 155 V

Figure 7 Time history record of pressure signal P3 whilethe CPI injector is working, where the static fluid pressure

is 276 kPa (40 psi) and the driving voltage is 155 V

pressure modulation, the pressure signal is processed using

a band pass filter The harmonic component of modulation

is plotted in Figure 8

As the modulation frequency is maintained at 2.2 kHz,

the magnitude of pressure modulation becomes smaller

when the injector is open versus when the injector is

closed In fact, its amplitude is reduced to about 65% The

coupling characteristic occurs in a soft manner for the

output property While the port injector is open, the static

pressure mean decreases to 150 kPa, and the ratio of the

peak to peak values of the pressure modulation to the mean

static pressure Pm/so is around 10% As such, in the case

when pressure frequency is 4 kHz, the magnitude of

pressure modulation becomes bigger when the injector is

opened versus closed The amplitude is 6.5 times as much

as it is the closed case The coupling characteristic occurs

in a harder manner for the output property In this situation,

the ratio Pm/so is around 23% However, whether the injector

is opened or closed, the modulation magnitude is the same

in the modulation frequency range from 7 to 13 kHz for the

output property For example, at a frequency of 8.0 kHz,

the modulation magnitude is 17.5 kPa (rms value), and the

ratio Pm/so is around 30% Its output property remains

con-stant However, at the moment the injector opens or closes,

there is an instantaneous loss in the amplitude of pressure

modulation, and tow gaps are formed in the response

6 SPECTRAL ANALYSIS

FFT is used to analyse the harmonic components in the

pressure signal As shown in Figure 9, the spectral analysis

of pressure signal P1 is carried out The time history

records come from the case of 0.0 kHz and 2.2 kHz in

Figure 6, and the data analysis commences at the moment

the injector closes Without modulation, the dominant

component is 45 Hz and is the natural frequency

compo-nent reflecting the response of the inner structure In themodulation case, the modulation frequency can be found as

an additional component of 2 kHz in the spectral picture.Through spectral analysis, we know that there are two mainharmonic components in the fuel injection system when theport injector is working However, for the components with

a very short time response, especially instantaneous teristics, there is no solution shown in the spectral analysis.STFT (Short Time Fourier Transform), as shown inFigure 10, is employed to observe instantaneous magnitudeand frequency while the injector is working To see it clear-

charac-ly, its harmonic components with low frequency (naturalfrequency in the fuel injection system) are removed fromthe 3D picture With the help of the 3D picture, we canintuitively learn that the pressure modulation varies indetail with time as well as the soft and hard output proper-ties At the moment the injector is opened or closed, themagnitude of pressure modulation has instantaneous loss.The phenomenon of instantaneous loss can easily beobserved from Figure 10(c), where there are two gaps inthe component of pressure modulation Besides this, someinstantaneous harmonic components, both the frequencyand the amplitude varying with time significantly, such as 3kHz and 8.5 kHz, are also revealed in the 3D picture, asshown in Figure 10(b), which are given out by a stepexcitation while the injector is opened and closed, and isprobably related to the inside structure of the port injector.Indeed, they are interesting components, and we still need

to study their dynamic reaction

CWT (Continuous Wavelet Transform) is used to scribe the power spectral density in the time-frequencydomain It has an advantage in depicting time-variationcharacteristics for non-stationary signal Minor disturbing

de-of the modulation frequency during the transition can beobserved from the CWT spectrogram shown in Figure

Figure 8 Variation of pressure modulation at P1 while

the CPI injector is working (Unit: kPa), where the static

fluid pressure is 276 kPa (40 psi) and the driving voltage is

155 V

Figure 9 Comparison of spectral analysis on pressuresignal P1 while CPI injector is closed where the static fluidpressure is 276 kPa (40 psi)

11(a) in the contour form, where the pressure modulation

with a frequency of 4 kHz at P3 is analyzed At the moment

the injector opens, the modulation frequency instantly

de-creases In the inverse case, it instantly inde-creases This

phenomenon can be explained by the non-stationary fuel

flowing inside the pressure modulator At that moment, the

flowing fluid may tend to the vortex case, which has an

influence on the coupling characteristics between the

piezoelectric stack and the hydraulic

7 CONCLUSIONS(1) The results of experimental investigation show that thecoupling characteristics between the piezoelectric stackand the hydraulic possess nonlinear features in a fuelinjection system The coupling characteristics willattain saturation status in the higher frequency rangewith a driving voltage beyond 250 V It will be changedwith an open injector Moreover, the curve of couplingcharacteristics tends to be smooth over whole workingfrequency range With the given piezoelectric modu-lator, the magnitude of pressure modulation will bereduced within the frequency range of 2~3 kHz and5~7 kHz while the port injector is open; but themagnitude of pressure modulation will be enlargedwithin the frequency range of 3~5 kHz; it will remainthe same within the frequency range of 7~13 kHz.The transfer characteristics from the source of pressuremodulation to the nozzle are sensitive to the injectioncondition Different properties will appear while theinjector is opened or closed

(2) The maximum output of pressure modulation can bedetermined for the fuel spray atomization from thecoupling characteristics Based on the coupling charac-teristics, we can select the optimal driving voltage toobtain the satisfying pressure modulation At specialfrequencies, smaller driving voltages can result inbigger output magnitudes of pressure modulation tooutput, which will save driving power in engineeringapplications

(3) To analyze the instantaneous characteristics in the fuel

Figure 10 STFT analysis of pressure modulation P1 while

the CPI injector is working, where the static fluid pressure

is 276 kPa (40 psi) and the driving voltage is 155 V

Figure 11 Contour of CWT spectrogram for pressuremodulation while the CPI injector is working, where themodulation frequency is 4.0 kHz, the static pressure is 276kPa and the driving voltage is 155 V

injection system, STFT and CWT are employed to

study the response at P1 and P3 These are good signal

analysing tools that are well-suited to observe

time-variation characteristics in the pressure modulation, by

which not only instantaneous harmonic components of

3 kHz and 8.5 kHz at P1 are captured from STFT, but

also negative changes in gaps of modulation

compo-nents are observed while the injector is opened or

closed On the other hand, a hidden variation, such as

instantaneous variation in the modulation frequency at

P3, is also discovered from CWT Due to

compli-cations in the mathematical model of the fuel injector,

so far these time-variation characteristics are

diffi-cult to be numerically simulated by the Computational

Fluid Dynamics method, and their fluid dynamic

properties are also difficult to fully explain Therefore,

it is necessary for researchers to study fluid dynamics

in the fuel injector through experimentation and

modern signal processing

(4) At modulation frequencies of 23.3 kHz or 27.2 kHz,

extremely strong, but unsteady pressure pulsation is

observed at the measuring point P3 An example is

shown in Figure 12, where the static pressure is 276

kPa and the sine wave driving voltage is 155 V This

phenomenon may be explained as the resonance in the

mechanical structure of the piezoelectric pressure

modulator The pressure modulation could not be

directly used in spray atomization

(5) Experimental investigation should expand the

fre-quency range of pressure modulation up to 50 kHz in

the future research, in which more dynamic

characteri-stics could be captured in the fuel injection system

(6) To clarify the pressure modulation effect on spray

characteristics, Figure 13 shows the comparison of

spray performance, including the histograms of droplet

size and velocity between the baseline and pressure

modulation CPI injector based on the PDA

measure-ments It is clear that the weak bimodal distribution

appearing in the baseline spray size histogram

dis-appears completely for the pressure modulation spray.This observation is also confirmed in size-velocitycorrelation Almost no large droplets are observed fromthe pressure modulation spray, and the droplet size isgreatly reduced in the pressure modulation mode Thefuel spray is more uniform in space with the pressuremodulation technique compared to the baseline condi-tion spray Moreover, the percentage of high velocitydroplets shows the apparent increase in the velocityhistogram of the pressure modulation spray Thedependence of the atomization characteristics on driv-ing power and frequency of this pressure modulationtechnique indicates its potential application in controll-ing spray performance of automotive port fuel injec-tors More detailed test results regarding the atomi-zation improvement with pressure modulation techni-

ques have been presented in the literature (Kim et al.,

system J Sound and Vibration, 245, 815−834.

Hu, Q and Wu, S F (2001) Modeling of dynamic response

of an automotive fuel rail system, Part 1: Injector J.

Kim, H., Im, K.-S and Lai, M.-C (2004) Pressure lation on micro-machined port fuel injection performance

modu-Int J Automotive Technology 5, 1, 9−16

Miller, D S (1990) Internal Flow System 2nd edn

BHRA Information Services Cranfield Bedford UK

Figure 12 Pressure in measuring point P3 with a driving

frequency of 27.2 kHz, where the static pressure is 276

kPa, the driving voltage is 155 V and the injector is closed

Figure 13 Comparison of spray performance for the CPIinjector at a fuel injection pressure of 276 kPa

at atmospheric pressure ILASS-Americas, Sacramento,

COMBUSTION DEVELOPMENT OF A BI-FUEL ENGINE

O S ABIANEH1)*, M MIRSALIM2) and F OMMI1)

1)Department of Mechanical Engineering, Tarbiat Modares University, Chamran High Way,

P.O Box 14115-111, Tehran, Iran

2)Department of Mechanical Engineering, AmirKabir University, 424 Hafez Ave, P.O Box 15914, Tehran, Iran

(Received 8 August 2007; Revised 20 July 2008)

ABSTRACT−Environmental improvement and energy issues are increasingly becoming more important as worldwideconcerns Natural gas is a good alternative fuel that can help to improve these issues because of its large quantity and cleanburning characteristics This paper provides the experimental performance results of a Bi-Fuel engine that uses CompressedNatural Gas as its Primary fuel and gasoline as its secondary fuel This engine is a modification of the basic 1.4-liter gasolineengine Generally, on the unmodified base engine, torque and power for CNG fuel are considerably lower than gasoline fuel

In this paper, the influence of fuels on wall temperature, performance and emissions are investigated

KEY WORDS : Bi-fuel engine, CNG fuel, Gasoline fuel, Fuel consumption, Emission, Engine wall temperature

1 INTRODUCTION

The IKCO Company has been developing an engine that

can run on both CNG and gasoline Some companies have

not adopted CNG dedicated systems because these vehicles

have lower mileage and, thus, need more refueling

Environ-mental improvement and energy issues are becoming more

and more important as worldwide concerns Natural gas is

a good alternative fuel to help improve these issues because

of its large quantity and clean burning characteristics Since

the fuel system of CNG vehicles is completely closed, fuel

evaporative emissions are practically eliminated The

evapo-rative emission control system commonly used in gasoline

vehicles is unnecessary for CNG vehicles In addition,

in-use total reactive organic gas (ROG) emissions from fuel

storage and refueling of CNG vehicles are small (Gas

Research Institute, 1994) Therefore, CNG vehicles are

high-ly beneficial for environmental protection relative to other

engines In order to achieve a higher catalytic conversion

ratio, an A/F ratio sensor on the upstream of a

close-coupled converter (to control to stoichiometric A/F ratio),

and a heated oxygen sensor on the downstream of an under

floor converter (to precisely compensate the A/F ratio)

were employed The engine is also designed to pass the

Euro 5 emissions standard.A precisely controlled air/fuel

(A/F) ratio and a higher catalytic conversion using three

way catalysts are necessary to reduce exhaust gas emissions

For precisely controlled A/F, a sequential multi-port gaseous

injection system (MPI) was chosen Consequently, the

injectors and pressure regulator are newly developed.As

shown in Table 1, natural gas has a higher calorific value

per mass compared with gasoline fuel However, for anatural aspirated engine, the volumetric efficiency for agaseous fuel like CNG is more reduced than for a liquidfuel (Figure 1) As a result, the output of the engine burninggasoline is higher than if the CNG fuel was used.In order toprevent abnormal wear in the engine, the hardness of thevalves and valve seats was increased Additionally, in order

to reduce deformation, high-strength pistons and a fied water jacket design were employed in the cylinderhead

modi-In order to take advantage of the high octane rating, theengine compression ratio was raised However, for theengine to work acceptably with gasoline the compressionratio was then decreased to 10.8

As shown in Table 2, the CNG base engine is a cylinder 1.4 liter-DOHC 16-valve engine

4-Thermocouples are mounted 2 mm below the surface formeasuring surface temperature, as shown in Figure 2~3

2 IGNITION AND FLAME PROPAGATIONMethane has a lower unstretched and stretched laminarburning velocity when the gas temperature is higher than

450 K (due to Markestein numbers) (Liss and Thrasher,1991; Goodwin and Whiston, 1991; Duan, 1996; Bradley

et al., 1992) This proves that the total combustion

dura-tion is prolonged compared to diesel and gasoline fuel.Turbulent burning velocity is approximately proportional

to u L0.6 (unstretched laminar burning velocity) and inversely

related to Le0.3 (Lewis number) As the compressionincreases, the Lewis number decreases Therefore, there isless observed difference between the turbulent burningvelocities of natural gas and gasoline in the final stages of

*Corresponding author e-mail: samimiomid@gmail.com

combustion (Bradely et al., 1992; Jones and Evans, 1985).

However, some experiments suggest (Sharma et al., 1981)

that when one to five percent of methane is replaced with

ethane, there is a slight increase in burning velocity In our

study, 4.8% of the test fuel is ethane (Table 1)

A comparison of the heat release for both fuels for

research engine is shown in Figure 4~7

The heat release formula that is used in this paper is

displayed below: (Equation 1)

n : interval (1 deg of crank angle)

Figure 1 Volumetric efficiency

Table 2 Engine specifications

IVO/IVC@1m lift 7°ATDC/197°ATDC intake

Combustion chamber pent roof

Figure 2 Cylinder block sensor position

Figure 3 Cylinder head sensor position

Figure 4 Crank angle difference between position ofmaximum pressure and position of 5% Heat release (αmax−

pressure−α5%), full load, lambda=0.96 for CNG and 0.9 forgasoline

As shown in Figure 5, the duration between 5% heat

release and ignition timing of the CNG is longer than the

duration for gasoline fuel in total speed range

The duration between 50% heat release and 5% heat

release is nearly the same for both fuels, as shown in Figure

6 However, as shown in Figure 7 for the CNG fuel, the

duration between 90% and 5% heat release is more

pro-longed than for gasoline by approximately 10 to 20° of CA

These results prove that the total combustion duration of

CNG fuel is prolonged compared with gasoline fuel

Methane has a higher isentropic coefficient (Heywood,

1998) than other fuel, which results in a higher final

compression temperature Additionally, the more advanced

timing for the increase of power output and efficiencygives more combustion time and, hence, a higher temper-ature and pressure The heat transfer to the walls is, how-ever, not greatly changed for these two fuels, as shown inFigure 8 Due to the higher isentropic coefficient of CNG,the temperature of the gas in the expansion stroke (workingstroke) for CNG fuel decreases more rapidly than for gaso-line fuel Finally, the exhaust gas temperature of gasolinefuel is higher than CNG exhaust gas fuel, as shown inFigure 9

The CNG’s lack of latent heat of evaporation increasesthe temperature of the chamber wall during the inductionstroke

If the total mass of fuel (gasoline) evaporates, it can coolthe air by up to 20ºC (Equation 2)

Figure 5 Crank angle difference between position of 5%

Heat release and ignition position (α5%−αignition), full load,

lambda=0.96 for CNG and 0.9 for gasoline

Figure 6 Crank angle difference between position of 50%

and 5% Heat release (α5%−α5%), full load, lambda=0.96 for

CNG and 0.9 for gasoline

Figure 7 Crank angle difference between position of 90%

and 5 % Heat release (α90%−α5%), full load, lambda=0.96

for CNG and 0.9 for gasoline

Figure 8 Cooling water temperature difference (T outlet,water−

T inlet,water), at full load

Figure 9 Exhaust gas temperature and temperature

differ-ence between CNG and gasoline exhaust gasses (TGasoline exhaust gas−TCNG exhaust gas), 2000 rpm, lambda=1, 90ºC coolant

Some types of matter affect the temperature of the

cylinder block, including radiation, latent heat of

evapo-ration of gasoline and heat transfer from gases, which was

explained previously

As shown in Figures 12 and 13, the temperature

differ-ence between CNG and gasoline fuels for the exhaust side

of cylinder block wall at TDC and 1/4 stroke is lower than

intake side

The temperature of the cylinder block at 1/4 stroke onthe exhaust side is lower than on the intake side because ofthe oil return line

as can be seen in Figures 11 to 16, the difference in thetemperature of the walls between the fuels in cylinder headand cylinder block are the same

Figure 17 shows the influence of lambda variation oncylinder head temperature for CNG fuel The maximum

Figure 10 Temperature and temperature difference of

cylinder block material (TCNG cylinder block−TGasoline cylinder block),

cylinder 3 at TDC, intake side, 2000 rpm, lambda=1, 90ºC

coolant

Figure 11 Temperature and temperature difference of

cylinder block material (TCNG cylinder block−TGasoline cylinder block),

cylinder 3 at 1/4 stroke, intake side, 2000 rpm, lambda=1,

90ºC coolant

Figure 12 Temperature and temperature difference of

cylinder block material (TCNG cylinder block−TGasoline cylinder block),

cylinder 3 at TDC, exhaust side, 2000 rpm, lambda=1,

90ºC coolant

Figure 13 Temperature and temperature difference of

cylinder block material (TCNG cylinder block−TGasoline cylinder block),cylinder 3 at 1/4 stroke, exhaust side, 2000 rpm, lambda=1,90ºC coolant

Figure 14 Temperature and temperature difference of

cylinder head (TCNG cylinder block−TGasoline cylinder block), cylinder 3,exhaust side, 2000 rpm, lambda=1, 90ºC coolant

Figure 15 Temperature of cylinder head material andBMEP, on cylinder 2, between exhaust valves, at Full load,90ºC coolants

temperature for CNG fuel at full load occurred at lambda

=0.94~0.96 At this point, flame speed and engine torque

are at their maximum

The influence of water temperature on wall temperature

and nitrogen’s hydrocarbons and oxides is shown in Figures

18 and 19

As shown in Figure 18, when the temperature of the

cooling water is decreased by 10ºC, then the temperature of

the cylinder head is also reduced by 10ºC This relationship

is nearly linear also for other tested points Figure 19 shows

that when the temperature of the cooling water is increased,the amount of hydrocarbon pollutant is decreased This isbecause of the thickness of the quench area and the heattransfers to the walls are decreasing but the amount of NOx

is increasing because of the higher combustion chambertemperature resulting from lower heat transfer

3 IGNITION AND FLAME PROPAGATIONThe minimum spark energy required for methane ignition

is markedly higher than for other hydrocarbons As a result,the conversion of an engine to natural gas requires a high-performance ignition system (Guibet, 1999) In the fueltested in this study, other hydrocarbons (e.g ethane) arepresented in Table 1 In the case of ethane, the coil canignite the CNG at 38 (mJ) energy It was also observed thatthe required energy for pure methane is 100 to 120 mJ(Guibet, 1999)

Methane also has a wider flammability range than otherhydrocarbons This allows an engine to operate on a leanmixture, which is advantageous in some applications, such

as industrial vehicles

Methane combustion is relatively slow This can causedeterioration in performance due to an increase in heattransfer to the combustion chamber walls

An interesting observation is that methane's slowercombustion contributes to less combustion noise due to aless aggressive pressure gradient

4 FULL LOAD INVESTIGATION4.1 Torque and Power

The higher compression ratio increases torque and power

in CNG mode In addition, the closed timing intake valvewas advanced in the CNG mode with CVVT2 As a result,torque at lower engine speeds was nearly restored to levelsfound during gasoline use

Torque and Power for CNG decrease compared togasoline (Figure 20 and 21) In the worst case, power at

Figure 16 Temperature of cylinder head material, on

cylinder 2, between exhaust valves, at Full load, 90ºC

coolants

Figure 17 Temperature of cylinder head and cylinder block

material, between exhaust valves, on cylinder 2 and 3,

2000 rpm, BMEP=9 bar, 90ºC coolant, for different

lambda in CNG mode

Figure 18 Temperature of cylinder head and cylinder block

material, cylinder 3, 2000 rpm, lambda=1, BMEP=9 bar,

for different coolant temperatures in CNG mode

Figure 19 Pollution, 2000 rpm, lambda=1, BMEP=9 bar,for different coolant temperatures in CNG mode

2Continues Variable Valve Timing

6000 rpm for CNG decreases by 6% compared to gasoline.

Torque was found to decrease 22% at 3500 rpm

Torque, Volumetric efficiency and Power for CNG

de-crease compared to gasoline because of some reasons:

− The volumetric efficiency is reduced because of the

gaseous state of the fuel Gaseous fuel occupies a

larger volume per unit energy than liquid fuel

− The gasoline fuel mixture temperature is reduced

because of liquid fuel latent heat of evaporation and

therefore the volumetric efficiency is increased

com-pare to gas fuel

− The other main factor that reduces the power output is

the low flame speed of natural gas, which requires

more advanced spark timing to achieve complete

com-bustion within the correct portion of the engine cycle

(Figure 7) This can cause a further reduction in the

power output of the engine

4.2 BSFC

As shown in Figures 22 and 23, the BSFC (Brake Specific

Fuel Consumption) of CNG is lower than the BSFC of

gasoline In addition, the normalized BSFC of CNG is

lower than the normalized BSFC of gasoline, but the

differ-ence is less than BSFC

The formula for the normalized BSFC is: (Equation 3)

− As shown in Figures 24 and 25, because of the hightemperatures found at high rpm values at full load, thefuel/air ratio must be richer to decrease the temperaturebelow the limit (800°C), this limit comes from materialtemperature strength As a consequence, the BSFC isincreased

4.3 Peak Pressure

As shown in Figure 26, the position of peak pressure for

HHV -

Figure 20 Corrected torque

Figure 21 Corrected power

Figure 22 BSFC

Figure 23 Normalized BSFC

Figure 24 Lambda variation at full load, 90ºC coolanttemperature

CNG is between 12-16 CA-ATDC For gasoline, this

position is between 16-28 CA-ATDC because of knocking

4.4 Pollution

Pollution consists of Carbon monoxide, unburned

hydro-carbons and nitrogen oxides; and these pollution from an

engine with CNG fuel follow a pattern similar to gasoline

fuel The position of emissions sampling is before the

catalyst converter

4.5 Hydrocarbons (HC)

The amount of unburned hydrocarbons for CNG fuel is

lower than for gasoline fuel However, it is very much

dependent on engine specifications, like compression ratios

and running conditions (Duan, 1996)

Substantial oxidation of the hydrocarbons that escape the

primary combustion process by any of the processes can

occur during expansion and exhaust The amount of

oxida-tion depends on the time course of the temperature and

oxygen concentrations of these HCs as they mix with the

bulk gases (Duan, 1996), as shown in Figure 27 The O2

concentration in the exhaust of CNG fuel is higher than

gasoline fuel This is a result of the lean CNG mixture used

in full load conditions and because the unburned

hydro-carbons for CNG fuel can find oxygen and burn during the

expansion and exhaust stroke

As shown in Figure 28, the amount of HC pollution in

exhaust gas is lower for CNG fuel than gasoline fuel in the

conditions tested in this study

Theoretically, the HC emissions from gas engines should

be lower due to the gaseous form of the fuel, whichprovides excellent mixing However, in some engines withearly exhaust valve opening and improper ignition timing,because of the slow flame speed and slow reaction ofmethane, the combustion was not completed before theexhaust valve opened This can be attributed to a high level

of HC

It should be noted that UHC (Un burned Hydro Carbon)from gas engines mainly consists of methane (approxi-mately 85%), which is not photochemically reactive There-fore, the non-methane hydrocarbons (NHMC) from gasengines are extremely lower than from gasoline engines.4.6 NOx

As shown in Figure 29, the amount of NOx in the exhaustgas of CNG fuel is higher than for gasoline fuel The most important engine variables that affect NOemissions are the fuel/air equivalence ratio, the burned gas

fraction of the in-cylinder unburned mixture, and the spark

timing The burned gas fraction depends on the amount ofdiluents, such as recycled exhaust gas (EGR) and theresidual gas fraction, used for emission controls Fuel pro-perties will affect burned gas conditions The effect ofnormal variations in gasoline properties is, however, modest

Figure 25 Exhaust gas temperature

Figure 26 Position of peak pressure

Figure 27 O2, full load condition, and

Changes in the time course of temperature and oxygen

concentration in the burned gases during the entire

com-bustion process and early part of the expansion stroke are

the important factors

The peak pressure of CNG fuel is higher than the peak

pressure of gasoline fuel because of the advanced timing of

the ignition This causes the amount of NOx to increase

As the burned gases cool during the expansion stroke,

the reactions involving NO cease and leave NO

concen-trations far in excess of levels corresponding to equilibrium

at exhaust conditions The isentropic coefficient of CNG is

higher than for gasoline and this causes the temperature of

the gas in the expansion stroke in CNG mode to decrease

more rapidly than the gasoline fuel Therefore, the NOx

freezes rapidly in the CNG mode Furthermore, advanced

ignition timing and a faster heat release rate in the

spark-ignited gas engine further increase peak temperatures

(Duan, 1996)

4.7 Carbon Monoxide (CO)

The carbon monoxide for CNG fuel is lower than for the

gasoline engine for all loads and speeds, due to its low

carbon content, and leaner mixture in full load condition, as

shown in Figures 30 and 31

Carbon monoxide also forms during the combustion

process in rich fuel-air mixtures In these mixtures, there is

insufficient oxygen to fully burn the fuel In addition, inhigh-temperature products and lean mixtures, dissociationensures there are significant CO levels Later, in the ex-pansion stroke, the CO oxidation process also ceases as theburned gas temperature falls For fuel-rich mixtures, COconcentrations in the exhaust increase steadily with increas-ing equivalence ratios, as the amount of excess fuelincreases

The amount of CO in the exhaust gas of CNG is lessthan in gasoline fuel because of its leaner mixture (Figure31)

5 CONCLUSIONSNatural gas is a type of hydrocarbon fuel with a simplechemical structure It is a clean fuel that is very suitable foruse in spark ignition engines Its large reserve around theworld further encourages its application in transport andpower generation markets

As a result, the following conclusions were reached:(1) The decrease in power output usually found in the CNGengine, resulting from the use of gaseous CNG fuel,was minimized by increasing the high compressionratio, adding increasing valve lift, optimizing valve timingand reducing engine backpressure However, the poweroutput needs to be improved further to equal that ofgasoline

(2) The durability of the base gasoline engine is insufficientfor use with CNG because of the characteristic nature

of natural gas Therefore, improvements were made tothe pistons, cylinder head, valves and valve seats forboth intake and exhaust systems These characteristicsare not related to the high temperature of the com-bustion chamber wall, since a large difference betweenthe wall temperature of the two fuels, CNG and gaso-line, was not seen in this study

CNG is not an oily fuel and therefore causes damage tothe seat material, increasing the blow by (the oil andfuel vapor) of the engine This problem can be solved

by adding a passage for lubrication of the seat valve or

by changing the seat material

Figure 29 NOx emission, full load conditions,

(3) Combustion in full load mode of the engine using CNG

fuel is leaner than the combustion of the engine using

gasoline fuel

(4) The CO and NMHC (Non Methane Hydro Carbon)

emissions from the engine running on natural gas are

lower than those of an engine running on gasoline

under similar conditions

(5) It was demonstrated that, with careful design, natural

gas engines can achieve good performance, low

emi-ssions and are efficient enough to be realistic

alter-natives to diesel and gasoline engines

(6) Decreasing the water temperature causes a decrease in

the temperature of the combustion chamber and

cylin-der walls Most importantly, it causes a decrease in the

amount of NOx emitted and increases the amount of

HC pollutant

(7) The wall temperature of CNG fuel is hotter than for

gasoline fuel but not enough to damage the cylinder or

cylinder head materials

(8) The combustion duration of CNG fuel is more

pro-longed than for gasoline fuel

(9) The normalized brake specific fuel consumption of the

engine tested with CNG fuel is lower than with

gaso-line fuel

REFERENCES

Bradely, D., Lau, A K C and Lawers, M (1992) Flame

stretch rate as a determinant of turbulent burning

velo-city Phil Trans R Society LOND, A338, 357−387.

Bradley, D., Lawes, M., Sheppard, G G W and Woolley,

R (1996) Department of mechanical engineering

univer-sity of leeds, UK, methane as an engine fuel Proc.

Institution of Mechanical Engineers, Part D: J mobile Engineering S410/002/96

Auto-Duan, S Y (1996) Cosworth engineering limited, ampton, UK laboratory experience with the use of

north-natural gas fuel in IC engines IMechE Seminar

Publi-cation.

Goodwin, M J and Whiston, P J (1991) Analysis of thecombustion of methane in a spark ignition internal

combustion engine Proc I Mech E Seminar, IC Engines

Research at Universities, Polytechnics, 55−60

Guibet, J C (1999) Fuels and Engines Editions Techip.

Prais

Heywood, J B (1998) Internal Combustion Engine

Fund-amental MacGraw Hill New York.

Jones, A L and Evans, R L (1985) Comparison of

burn-ing rates in a natural-gas-fueled spark ignition engine J.

Liss, W E and Thrasher, W H (1991) Natural gas as a

stationary engine and vehicular fuel SAE Paper No.

912364

Sharma, S P., Agrawal, D D and Gupta, C P (1981) Thepressure and temperature dependence of burning velo-

city in a spherical combustion bomb 18th Symp Int.

Topical Report GRI-Gas Research Institute (1994) Light

Duty Vehicle Full Fuel Cycle Emissions Analysis Chicago.

USA

Ningbo HOYEA Machinery Manufacture Co., Ltd, Ningbo 315131, China

(Received 4 September 2007; Revised 26 July 2008)

ABSTRACT−As the internal combustion engine moves into the 21st century, fully flexible valve actuation systems are beingproposed as an enabling technology for advanced internal combustion engine concepts Electro-hydraulic valve actuatorsystems are being considered as a potential variable valve technology Compared to the servo control system, the system using

a proportional valve has the advantages of low price, high anti-pollution ability and high reliability Our research focuses onexploring the dynamic characteristic of the electro-hydraulic variable valve system, which is based on three-way proportionalreducing valve In this paper, the structure and working principles of the system are described The dynamic mathematicalmodel of the system is derived From the analysis of a linearized model and dynamic simulation, it is demonstrated that thesystem will be stable only if the proportional reducing valve has a positive opening Some structural factors that affect thesystem’s dynamic characteristics, such as input signal, the stiffness of the return spring and the pre-tightening force of thereturn spring, are studied using AMESim The experimental results coincide with the theoretical and simulated analyses.Further study shows that the dynamic response can be improved effectively by adopting closed-loop control of valve lift

KEY WORDS: Electro-hydraulic, Variable valve, Dynamic characteristics, Three-way, Proportional reducing valve

NOMENCLATURE

P c : input pressure of the single-rod hydraulic cylinder

K I : constant

U : input voltage signal

A p : area of the hydraulic cylinder

K : stiffness of the return spring

X p : displacement of the piston

X0 : initial reduction length of the return spring

F L : load force

K u : pulse duty factor

L : coil inductance

R : resistance of the coil and amplifier

K e : velocity back electromotive force coefficient

I : input current

X v : displacement of the valve spool

F m : magnetic force of the proportional electromagnet

m v : mass of the armature and valve spool

D : viscous damping coefficient of the valve spool

F v : load force of the valve spool

K i : current-force gain of the proportional electromagnet

K v : displacement-force gain of the proportional

electro-magnet

Q L : flow rate of the load

K q : flow rate gain coefficient

K c : flow rate-pressure coefficient

C ip : inner leakage coefficient of the hydraulic cylinder

V c : volume of the controlled chamber of the hydraulic

chamber

βe : bulk modulus of elasticity

M p : total mass of the piston and load

B p : viscous damping coefficient of the piston

K ce : total flow rate-pressure coefficient

C d : flow rate coefficient

y0 : expected valve lift

y f : feedback of valve lift

1 INTRODUCTIONValve timing, lift and duration of conventional internalcombustion engine driving systems, which use a cammechanism to control the intake and/or exhaust valve,

*Corresponding author e-mail: bjin@zju.edu.cn

cannot be adjusted because the border line of the cam is

fixed Flexible intake and/or exhaust valve motions can

greatly improve the fuel economy, emissions and torque

output performance of internal combustion engines (Chen

and Cui, 2002) Flexible valve actuation can be achieved

using mechanical, electromagnetic and electro-hydraulic

valve mechanisms (Dresner, 1989) The cam-based

mech-anisms can offer limited flexibility of the valve event The

Fiat 3D cam variable valve mechanism (Titolo, 1991) can

implement variable valve lift and timing, but only within in

a certain range The Honda VTEC mechanism (Takefuml,

1991) is a multiple-step device that can switch between two

discrete cams and offer discontinuous change of valve

timing To achieve continuous valve timing, lift and duration,

camless engines, which include electromagnetic and

electro-hydraulic mechanisms, are proposed The electromagnetic

mechanisms, such as the GM R&D Magnavalue (Theobald

et al., 1994), FEV (Boie et al., 2000), Siemens (Hartke and

Koch, 2002), Aura (Schneider, 2001) and Magneti Maralli

(Cristiani et al., 2002), can generate continuous variable

valve timing and duration, but these devices cause high

valve-seating velocities The electro-hydraulic systems, such

as the Ford camless (Wright et al., 1994) and the Sturman

systems (Sturman, 1997), can provide fully flexible control

of valve events The advantages of the electro-hydraulic

mechanisms are low energy consumption and high

reliabi-lity

This paper is organized as follows: Section 2 presents

the structure and working principle of the electro-hydraulic

variable valve system, which is based on a three-way

pro-portional reducing valve Section 3 presents the structural

factors which influence the dynamic performance of the

system which is analyzed using a linearized mathematical

model, and the stability of the system is also analyzed in

this section Section 4 addresses the simulation model of

the system, which is built by AMESim Through

simu-lation, the dynamic performance of the electro-hydraulic

system, which is affected by the structural factors, is

analy-zed Section 5 presents the experimental results of

electro-hydraulic variable valve system Section 6 presents future

work on improving the dynamic performance Section 7

contains a summary of the work

2 STRUCTURE AND WORKING PRINCIPLE

OF ELECTRO-HYDRAULIC VARIABLE VALVE

SYSTEM

The schematic diagram of the electro-hydraulic variable

valve system, which is based on the three-way proportional

reducing valve, is shown in Figure 1 As shown in Figure 1,

the system consists of a proportional relief valve, oil

reser-voir, hydraulic pump, high-speed proportional reducing

valve, motor, single-rod hydraulic cylinder, return spring

and valve

In the steady state, the input pressure of the single-rod

hydraulic cylinder is proportional to the input voltage signal

of the high-speed proportional reducing valve Theexpression can be defined as follows

(1)

Where, P c is the input pressure of the single-rod hydraulic

cylinder, Pa; K1 is the constant; U is input voltage signal, V.

The relationship between the output hydraulic force ofthe hydraulic cylinder and the load force in the steady statecan be described by

(2)

Where, A p is the area of the hydraulic cylinder, m2; K is the stiffness of the return spring, N/m; X p is the displacement of

the piston, m; X o is initial reduction length of the return

spring, m; and F L is the load force, N

According to Equations (1)~(2), one can write

(3)

It is clear from Equation (3) that the valve lift, timingand duration can be changed by varying the input voltagemagnitude, triggered time and pulse width, respectively

3 ANALYSIS OF THE STABILITY OF THE ELECTRO-HYDRAULIC VARIABLE VALVE SYSTEM

3.1 Dynamic Mathematical Model of the Electro-hydraulicVariable Valve System

In order to determine how structural factors influence thedynamic performance of electro-hydraulic variable valvesystems, the dynamic mathematical model of the system is

required (Merritt, 1967; Wong et al., 2008).

The dynamic differential equation of the coil current in aproportional electromagnet can be defined as follows (Sabri,2006)

-magnet, N; m v is the mass of the armature and valve spool,

Kg; D is the viscous damping coefficient of the valve

spool, N/(m/s); and F v is the load force of the valve spool,

N

The dynamic output force of the proportional

electro-magnet can be described by (Guan, 2003)

(6)

Where, K i is the current-force gain of the proportional

electromagnet; and K y is the displacement-force gain of the

proportional electromagnet

The Laplace transform expressions of Equations (4)~(6)

can be described by

(7) (8)(9)According to Equations (7)~(9), one can write

(10)The linearized discharge expression of the valve can be

defined as follows

(11)

Where, Q L is the flow rate of the load, L/min; K q is the flow

rate gain coefficient; and K c is the flow rate-pressure

coeffi-cient

The continuity expression of the working chamber of the

hydraulic cylinder can be defined as follows

(12)

Where, C ip is the inner leakage coefficient of the hydraulic

cylinder; V c is the volume of the controlled chamber of the

hydraulic chamber, m3; and βe is the bulk modulus of

elasticity, Pa

The dynamic relationship between the output force of

the piston and the load force can be described by

(13)

Where, M p is the total mass of the piston and load, Kg; and

B is viscous damping coefficient of the piston, N/(m/s);

(17) (18)

Where K ce is the total flow rate-pressure coefficient.According to Equations (10)~(17), one can write

(19)

It is clear from Equation (19) that K u , K, F L and U are the

main structural factors that influence the dynamic mance of the electro-hydraulic variable valve system.3.2 Analysis on Stability of the System in Diffierent Open-ing Modes

perfor-In this system, the three-way proportional reducing valve isthe key component whose opening mode determines thestability of the electro-hydraulic variable valve system

In positive opening mode, the flow rate gain coefficient

K q and flow rate-pressure coefficient K c can be defined asfollows

(20) (21)

Where, C d is the flow rate coefficient; w is the area gradient; Z is the positive opening size; ρ is the oil density;

and P s is the system pressure

In zero-opening mode, the flow rate gain coefficient K q and flow rate-pressure coefficient K c can be defined asfollows

(22) (23)The system parameters are listed in Table 1 and the poles ofthe system are listed in Table 2

The system is unstable because there are two poles in theright half S-plane in the zero-opening mode However, in

ρ -+

K c= 2 -C2 d wZ ρ -(P s1–P c) ρ1P c

+

K q =C d w 2 P( s–P c)

ρ -

2ρ(P s–P c) -

the positive opening mode, the system is stable because all

poles are located in the left half S-plane

As shown in Figure 2, the simulation model is built

by AMESim and the simulation parameters are listed in

Table 3

The simulation results are shown in Figure 3 and Figure 4

The output pressure of the proportional reducing valve

P c is unstable in the zero-opening mode, whereas the output

pressure of the proportional reducing valve P c is stable inthe positive opening mode

In order to validate the simulation results, the mental system is built (as shown in Figure 19) and theexperimental results are shown in Figure 5 and Figure 6 It

experi-is clear in Figure 5 and Figure 6 that the experimentalresults agree well with the simulation results

Table 1 System parameters of the electro-hydraulic

vari-able valve system

Figure 3 Output pressure of the proportional reducing

valve P c in zero opening mode

Figure 4 Output pressure of the proportional reducing

valve P c in positive opening mode

4 SIMULATION ANALYSIS OF THE

DYNAMIC PERFORMANCE OF THE

ELECTRO-HYDRAULIC VARIABLE VALVE

SYSTEM

It would appear that structural factors that affect the dynamic

performance of the electro-hydraulic variable valve system

were found They are K u , K, F L , and U In order to analyze

the dynamic performance of the system, a simulation

model was built using AMESim

In the following simulation results, the opening time is

defined as the necessary time when the lift rises from 10%

of the stable value to 90% of the stable value, and the

closing time is defined as the necessary time when the lift

falls from 90% of the stable value to 10% of the stable

value

4.1 Simulation Result of Input Voltage Magnitude

In order to analyze the influence of the input voltage

magnitude on the dynamic performance of the system, a

simulation is adopted The simulation results are shown in

Figure 9

It is obvious in Figure 7 that the valve lift increases

with increasing input voltage This result agrees well with

Equation (3)

Figure 8 shows that the output pressure of the

propor-tional reducing valve P c increases and the hydraulic force,which acts on the piston of the hydraulic cylinder, is

proportional to P c Hence, P c increases with increasinginput voltage This result agrees well with Equation (1)

In Figure 9, the input signal pulse width is 47 ms, thesystem pressure is 6 MPa and the stiffness of the returnspring is 8000 N/m with a 52 N pre-tightening force It isobvious that valve lift increases from 4 mm to 9 mm whenthe input voltage magnitude increases from 1 V to 1.5 V

Figure 5 Experimental output pressure of the proportional

reducing valve in zero-opening mode

Figure 6 Experimental output pressure of the proportional

reducing valve in positive opening mode

Figure 7 Valve lift for different input magnitudes

Figure 8 Output pressure of the proportional reducingvalve for different input magnitudes

Figure 9 Influence of variable input voltage magnitude

In order to analyze the average velocity when the valve

is opening and closing, the average velocities of 0.5 mm

~3.5 mm lift for different input magnitudes are analyzed

and the simulation result is shown in Figure 10 It is clear

in Figure 10 that the opening average velocity increases,

because the driving force (hydraulic force minus the sum

of the spring and pre-tightening forces) increases with

increasing input voltage And closing average velocity

increases slightly because the return spring force increases

with increasing input voltage magnitude

4.2 Simulation Result of Input Pulse Width

In order to analyze the influence of input pulse width on the

dynamic performance of the system, the simulation is

adopted The simulation results are shown in Figure 11

The system pressure is 6 MPa, the input voltage is 1.5 V

and the stiffness of return spring is 8000 N/m with a 52 N

pre-tightening force It is clear in Figure 11 that the valve

lift increases from 3 mm to 8 mm when the pulse width

increases from 9 ms to 39 ms with the input voltage

magni-tude at 1.5 V

In order to analyze the average velocity when the valve

is open and closed, the average velocities of 0~3 mm lift

with different input pulse widths are shown in Figure 12

Figure 12 shows that the average opening velocity increases

slightly because the K u and input voltage are increasing.However, the average closing velocity is increasing whenthe input pulse width is increased This phenomenon iscaused by the increased spring force

4.3 Simulation Result of the Pre-tightening Force the of

Return SpringThe results of the pre-tightening force on the dynamic

Figure 10 Average velocity curves of different input

magnitudes

Figure 11 Influence of variable input pulse width

Figure 12 Average velocity curves of different pulsewidths

Figure 13 Valve lift curves of different pre-tighteningforces

Figure 14 Influence of the pre-tightening force of thereturn spring

performance of the system are shown in Figure 14 and

Figure 15

The input signal pulse width is 47 ms with the input

voltage magnitude at 1.25 V, the system pressure at 6 MPa

and the stiffness of return spring at 8000 N/m

Figure 13 shows that the valve lift decreases when the

pre-tightening force of the return spring increases This

result agrees well with Equation (3)

It is clear in Figure 14 that the valve lift decreases from

9 mm to 5.5 mm when the pre-tightening force of the return

spring increases from 36 N to 68 N

In order to analyze the average velocity when the valve

is open and closed, the average velocities of 0.6 mm~5.4

mm lift with different pre-tightening forces for the return

spring are shown in Figure 15 Figure 15 shows that the

average opening velocity decreases because the driving

force (hydraulic force minus the sum of the spring and

tightening forces) decreases with the increase of the

pre-tightening force However, the average closing velocity is

increasing due to the increasing sum of the pre-tightening

and spring force

4.4 Simulation result of the stiffness of the return spring

The results of the stiffness of the return spring on thedynamic performance of the system are shown in Figure 17and Figure 18 The input signal pulse width is 47ms withthe input voltage magnitude at 1.25V, the system pressure

at 6 MPa and the pre-tightening force at 52 N

Figure 16 shows that the valve lift decreases when thestiffness of the return spring increases This result agreeswell with Equation (3)

In order to analyze the influence of the stiffness of thereturn spring on the dynamic performance of the systemwhile maintaining the pre-tightening constant, the simu-lation is performed The simulation result is shown inFigure 17 It is clear in Figure 17 that the valve lift de-creases from 7 mm to 4 mm when the stiffness of the returnspring increases from 8000 N/m to 14000 N/m

In order to analyze the average velocity when the valve

is opening and closing, the average velocities of 0.4 mm~4

mm lift with different stiffnesses of the return spring areshown in Figure 18 It is clear in Figure 18 that the averageopening velocity is decreasing because the driving force(hydraulic force minus the sum of the spring and pre-tightening forces) decreases with the increase of the stiff-ness The average closing velocity has an optimal valuewhen the stiffness of the return spring is increasing becausethe valve lift decreases when the stiffness of the return

Figure 15 Average velocity curves of different

pre-tighten-ing forces

Figure 16 Valve lift of different stiffnesses of the return

spring

Figure 17 Influence of the stiffness of the return spring

Figure 18 Average velocity curves of different stiffnesses

of the return spring

spring increases Hence, the spring force is at its maximum

value when the stiffness of the return spring is increasing

5 EXPERIMENT OF THE

ELECTRO-HYDRAULIC VARIABLE VALVE SYSTEM

In order to validate the simulation results and analyze the

dynamic performance, the following experiments were

carried out

The experimental system of the electro-hydraulic

vari-able valve system, which is based on the three-way

reduc-ing valve, is shown in Figure 19

The pressure transducer is used for measuring the input

pressure of the single-rod cylinder The displacement

trans-ducer, which can linearly convert a 0~25 mm displacement

signal to a 0~5 V voltage signal, is used for measuring the

displacement of the valve The oscilloscope is used for displaying and storing the input, pressure and displacement

magni-N pre-tightening force Figure 21 shows that the valve liftincreases from 3.8 mm to 8.8 mm when the input voltagemagnitude increases from 1 V to 1.5 V Figure 21 showsthat the opening time is shortened and closing time isslightly shortened at the same lift interval when the inputvoltage magnitude increases

5.2 Experimental Results of the Variable Input VoltagePulse Width

The experimental results of the different input pulse widthare shown in Figure 22 In the experiment, the input signalmagnitude is 1.5 V, the system pressure is 6 MPa and thestiffness of the return spring is 8000 N/m with a 52 N pre-tightening force Figure 22 shows that the valve liftincreases from 2.5 mm to 7.5 mm when the pulse width isFigure 19 Block diagram of the experimental system

Figure 20 Picture of the experimental system

Figure 21 the experiment results of variable input voltagemagnitude

Figure 22 the experiment results of variable input pulsewidth

increasing from 9 ms to 39 ms It is clear in Figure 22 that

the opening time is basically unchanged and closing time is

shortened over the same lift interval when the input pulse

width increases

5.3 Experimental Results of Different Pre-tightening Force

The experiment results of the different pre-tightening forces

of the return spring are shown in Figure 23 In the

experi-ments, the input signal magnitude is 1.25 V with a 47 ms

pulse width, a system pressure of 6 MPa and a stiffness of

the return spring of 8000 N/m Figure 23 shows that the

valve lift decreases from 8.5 mm to 4.5mm when the

pre-tightening force of the piston increases from 36 N to 68 N

It is clear in Figure 23 that the opening time is increasing

and the closing time is shortened over the same lift interval

when the pre-tightening force of the return spring is

increased

5.4 Experimental Results of Different Stiffness of the

Return Spring

The experimental results of different stiffness of the return

spring are shown in Figure 24 In those experiments, the

input signal magnitude was 1.25 V with a 47 ms pulse

width, a system pressure of 6 MPa and a pre-tightening

force of 52 N Figure 24 shows that valve lift is decreasedfrom 7 mm to 3.5 mm when the stiffness of the returnspring is increased from 8000 N/m to 14000 N/m It is clear

in Figure 24 that the opening time is increasing at the samelift interval when the stiffness of the return spring isincreasing However, the closing is not shortened with aincrease in the stiffness It has an optimal value

6 FUTURE DEVELOPMENTS

It is obvious from the experimental results that the risingand falling time of the valve is too long in the open loopsystem It is not suitable for a high-speed engine As shown

in Figure 10, increasing the input signal can accelerate thevalve speed Hence, in order to improve the dynamicperformance of the system, study of the closed-loop of thevalve lift needs to be conducted

A PID controller is added and the transfer function isshown as follows

(24)

Where, K p is the proportional coefficient; K D is the

differ-ential coefficient; K i is the integral coefficient; e is the error

whose expression is defined in Equation (25)

It is obvious that the rising and falling times are

shorten-ed by adopting closshorten-ed-loop control of the valve lift It cangreatly improve the dynamic performance of the system

7 CONCLUSIONSThis paper presents an electro-hydraulic variable valvesystem which is based on the three-way reducing valve

Figure 23 the experiment results of the pre-tightening

force of return spring

Figure 24 the experiment results of the stiffness of return

spring

Figure 25 Dynamic performance of the open and closedloop system

The key factors affecting the dynamic performance of the

system are discussed in detail The conclusions are as

follows

(1) The system is more inexpensive than using the servo

valves because it uses high-speed proportional valves

(2) The system is stable only if the proportional reducing

valve has a positive opening mode

(3) The valve lift, timing and duration can be controlled by

changing the input signal of the proportional reducing

valve

(4) The experimental results agree with the simulation

results This model can be used for optimizing system

parameters such as the pre-tightening force and

stiff-ness of the return spring

(5) By using the closed-loop PID controller, the system

dynamic performance can be greatly improved

The work in this paper is just a beginning In order to

meet the demand of high-speed motors (6000 rpm or

more), much more work must be done in the future

ACKNOWLEDGEMENT−This paper is supported by the

nature science foundation of Zhejiang province, China Project

Number: Z106543

REFERENCES

Boie, C., Kemper, H., Kather, L and Corde, G (2000)

Method for Controlling a Eclectromagnetic Actuator for

Activating a Gas Exchange Valve On a Reciprocating

Internal Combustion Engine US Patent, 6,340,008 B1.

Chen, Q X and Cui, K R (2002) A survey of variable

valve system for engine Vehicle Engine, 3, 1−5.

Crisitiani, M., Marchioni, M and Morelli, N (2002)

Electro-magnetic Actuator with Laminated Armature for the

Actuation of Valves of an Internal Combustion Engine.

European Patent, EP01114908

Dresner, T (1989) A review and classification of variable

valve timing mechanism SAE Paper No 890674 Guan, J T (2003) Electrohydraulic Control Technique.

Tongji University Press Shanghai China

Hartke, A and Koch, A (2002) Method for controlling a

Electromechanical Actuating Drive for a Gas Exchange

of an Internal Combustion Engine US Patent, 6,371,064

B2

Merritt, H E (1967) Hydraulic Control System Wiley

Press New York

Sabri, C (2006) Mechatronics Wiley Press New York Schneider, L (2001) Electromagnetic Valve Actuator with

Mechanical End Position Clamp or Latch US Patent,

6,267,351 B1

Sturman, O E (1997) Hydraulic Actuator for an Internal

Combustion Engine US Patent, 5,638,781.

Titolo, A (1991) The variable valve timing

system-appli-cation on a V8 engine SAE Paper No 910009.

Takefuml, H (1991) Development of the variable valve

timing mechanisms SAE Paper No 910008.

Theobald, M., Lequesne, B and Henry, R (1994) Control

of engine load via electromagnetic valve actuator SAE

Wright, G., Schecter, N M and Levin, M B (1994)

Integ-rated Hydraulic System for Electrohydraulic Valvetrain and Hydraulically Assisted Turbocharger US Patent,

5,375,419A

(Received 7 June 2007; Revised 8 September 2008)

ABSTRACT−Mechanisms with “negative” stiffness are a unique tool used to significantly improve vibration isolation for avehicle driver via upgrade of the seat suspension However, connection of such mechanisms to the suspension results in errors

in the function generation process, and in most cases, makes motion impossible An approach to type synthesis is presented

in order to make this process more predictable, easy and thus more practical for the upgrade process Structural classification

of the suspensions is presented, and -an atlas of function-generating mechanisms for suspensions that reveals the effect of

“negative” stiffness is completed All of the function-generating mechanisms appear in the atlas as result of enumeration.Structural properties of novel and existing mechanisms are compared Finally, some advantages in practical use of novelsuspensions with “negative” stiffness are illustrated

KEY WORDS : Type synthesis, Function-generating mechanism, Seat suspension

1 INTRODUCTION

The structure of the function-generating mechanism (FGM)

plays an important role in achievement of sufficient

vibration isolation by applying the effect of “negative”

stiffness The FGMs can provide (a) reasonable layout of a

suspension, and (b) treatment of interaction of all parts

required to transform the input motion of the suspension

into a specified output motion Evolution of the FGMs for

suspensions with “negative” stiffness began in the 1960s

By now, many ingenious designs have been created,

although they differ from one another in aspects of minor

importance only (Alabuzhev, 1989) Therefore, there is no

guarantee that engineers and researchers in the vibration

isolation area have identified all of the useable FGMs

Furthermore the motion of such kinds of suspensions is

chaotic (Lee, 2004), showing inevitable unpredictable

behavior of the FGMs Generally, this empiricism is one of

the main obstacles to full exploitation of the practical merit

of the “negative” stiffness concept as mentioned in (Rivin,

2003)

This paper presents an approach to the type synthesis of

FGMs The type synthesis refers to the kind of mechanism

selected and deals with enumeration of the links and joints

required to obtain the required mobility The approach is

demonstrated in the study through upgraded seat

sus-pensions First, a survey of regular suspensions and their

FGMs is presented Then, a structural classification of

mechanisms with “negative” stiffness is considered Anatlas of candidates for the FGM is enumerated for thesuspensions which reveal “negative” stiffness Structuralcharacteristics of the designed FGMs are also compared.Finally, we demonstrate selected upgraded seat suspen-sions and improved vibration isolation, especially in theinfra-frequency range

2 MOTIVATION FOR DEVELOPMENT OF THE APPROACH TO SUSPENSION DESIGN

A vehicle operator is most sensitive to infra-frequencyvibrations External vibration excitation to a vehicle, due tothe road surface or, operation of aggregates and equipment,

is also the most intensive in the infra- and infra-low quency ranges Hence, the vehicle operator is permanentlysubjected to multiple-frequency resonances This can beharmful and dangerous to the operator health and activity,

fre-especially in the frequency range of, f = 0.5-5 Hz Hz

(Paddan and Griffin, 2002)

Therefore, vibration isolation in the infra-frequency

range is essential for ride comfort of both drivers andpassengers This goal is unattainable through common use

of accepted concepts consisting in practice of heavy fluiddampers with passive, semi-active and active controls

(Rivin, 2003; Sun et al., 1995; Herzog, 1995) At the same,

this goal is theoretically possible if the natural frequencies,

f0, of a vibration isolating system are shifted below the

frequencies of external excitation, i.e., if f0< f This can be

achieved by extreme reduction stiffness for the system

*Corresponding author e-mail: cmlee@ulsan.ac.kr

elastic restraints This description becomes clearer when

considering a single-degree-of-freedom (1-DOF) system

with natural frequencies described by:

(1)

where k and M are the system spring stiffness and sprung

mass; is the coefficient of gravitation,;

mm/s2; z0 is the travel of the FGM, and ; c a is the

asymmetry parameter of travel depending on the system

height control with respect to mid-ride

Equation (1) shows that, minimizing k is the only real

way to provide infra-frequency vibration isolation for drivers

and vehicles subject to earth’s gravitation This minimum

is unattainable by means of the springs (metal, pneumatic,

etc.) common to the framework of the usual concept of

elasticity, despite the presence of a control mode The

reason for this is the insoluble contradiction between the

dimensions and high compliance, leading to loss of system

load-capacity when such the springs are in use (Frolov and

Furman, 1990) However, the minimum value in question

could be obtained by applying the concept of “negative”

stiffness that can be generated for some materials (Wang et

al., 2004), structures, and mechanisms (Alabuzhev, 1989;

Chyuprakov, 1987)

The “negative” stiffness in the suspensions can be

pro-duced via:

(a) global instability of the redundant kinematic chain

(RKC) that is designed and connected to the main

kine-matic chain (MKC) of an initial FGM (Alabuzhev, 1989;

Chyuprakov, 1987);

(b) local instability of the MKC, which has links

con-strained with a load-capacity spring and a redundant spring

with “positive” stiffness (Alabuzhev, 1989; Chyuprakov, 1987);

(c) global instability of the redundant stiffness

mini-mizing mechanism (SMM) equipped with a spring of

“negative” stiffness and then properly connected to the

FGM (Goverdovskiy and Lee, 2003; Goverdovskiy et al.,

2003)

Hereafter, “local instability” means that motion of the

system as a whole is stable but motion of some elements

can be unstable at certain instants in time “Global

instabi-lity” means that motion of the entire system can be unstable

at a given instant in time We also introduce the concept of

a "redundant spring” to emphasize that such new members

of the system do not change the load-capacity but are used

for stiffness control In other words, the introduced spring

does not interfere with the purpose of a load-capacity

spring, which is to keep the system in a position in

con-formity with the load

Coupling the SMM with a spring of “negative” stiffness

k2 located in-parallel with a load-capacity spring with

“positive” stiffness k1 yields the following set of k-minima:

(2a-2c)

where Φ is the coupling function between the FGM and the

SMM (Goverdovskiy and Lee, 2003), q is a generalized coordinate (DOF of the FGM), and ; q * is an idle freedom

of the FGM due to the coupling

Condition (2a) means that k is extremely small but

“positive” In this way, we reduce the stiffness of regular

N/m This leads to a reduction of naturalfrequencies of the vibration isolation system to 0.1−0.2 Hz, and thus presents perfect vibration isolation,including the infra-frequency range Condition (2b) means

that k can be reduced to zero In this case, the object

protected becomes motionless despite the amplitude andfrequency of vibration excitation Condition (2c) implies

that k can be extremely small but “negative” which is

adequate for the damping increase in the system transientmotion Thus, by controlling the value of one may

minimize the effect of stiffness k and invert it automatically

near zero in the dynamics We have proven these effectsthrough upgrade of regular mechanical and pneumatic seatsuspensions Before the upgrade, these suspensions failed

to achieve vibration isolation Moreover, they operated asamplifiers of input vibration in most parts of the infra-frequency range After the upgrade, the suspensions increasethe quality of vibration isolation to a large extent, as studied

in (Lee et al., 1989; Goverdovskiy, 1997; Goverdovskiy et

al., 1998; Lee and Goverdovskiy, 2002; Lee et al., 2006;

The structure of FGM numerically refers to the number

of movable links that transform motion of the input linkinto a motion prescribed for the output link Despite thedesign variety, we recognize six distinct types of planar andspatial FGMs as most representative; Figure, these arearranged in order of increasing structural complexity inFigure 1(b)

The structure can be conveniently represented as labeledgraphs in which edges represent joints (kinematic pairs)and vertices represent the links (Figure 1(c)) In each

graph, the vertex In represents a fixed link of the FGM, the vertex Out is an output link, edges P and R are the prismatic and revolute (low) pairs, and finally the edges H

represent centroid pairs (higher pairs) Classification of thepairs is given by Reuleaux: a low pair has surface contactbetween the pair elements, and a higher pair has line orpoint contact (Shigley and Uiker, 1995) At the top end ofthe spectrum (line 1) is the FGM, where each branch of theMKC consists of a single prismatic pair connecting input

and output links This is followed by the HRRH-linkage

(line 2), where each branch contains two higher and two

low pairs Next comes the FGM (line 3), in which the

branch is a four-bar RRRR-linkage A linkage of the

RRRH-type follows, constrained by redundant revolute pairs

to minimize unwanted relative motion of intermediate and

output links (line 4) A similar mechanism is completed

with two revolute pairs in order to transform sliding

friction into rolling friction Thus, line 5 shows a variant

for mechanisms of the previous type Due to a redundant

higher pair at the output link, this FGM has an idle degree

of freedom Motion of its output link can be planar or

spatial if the higher pair is planar or spatial Finally, line 6

demonstrates a FGM of (RRR) N -type, where is the

number of branches

4 STRUCTURAL CALSSIFICATION OF SEAT

SUSPENSIONS WITH “NEGATIVE” STIFFNESS

All design approaches that could provide perfect vibration

isolation by using the “negative” stiffness concept may beclassified into three groups:

(1)-group: Synthesis of a RKC, with links joined to thelinks of the MKC and constrained with a load-capacityspring and/or redundant spring with “positive” stiffness.(2)-group: Synthesis of a load-capacity spring that wouldreveal “negative” stiffness under post-buckling “in thesmall”

(3)-group: Synthesis of a SMM equipped with a springthat would reveal “negative” stiffness under post-buckling

“in the large” and then synthesis of a RKC to connect theSMM and an initial FGM

The concepts “in the small” or “in the large” refer to anobservation from stability theory that a buckled systemreveals comparatively small or large displacements withinpost-buckling behavior (Lukasiewicz, 1979)

Considering the anthropometry of seated vehicle drivers(Pheasant, 1990), stiffness control in a plane that coincideswith the sagittal plane is the goal of the design Thus, werestrict enumeration with 1DOF planar and spatial FGMs

Figure 1 Survey of regular seat suspensions and their FGMs: symbols and nomenclature Shown are the MKCs consisting

of fixed, input, intermediate, and output links 1-4; — vertex (fixed link), — vertex (a movable link), — edge (a pair:

prismatic P, revolute R, or centric H).

using the following structural formulae:

(3a) (3b)

where m1=1 is the DOF (mobility) of the FGM in the

direction of suspension motion; n is the total number of

links, and (n−1) of n are movable; p3, p4 and p5 are the

numbers of 3-DOF, 2-DOF, and 1-DOF pairs of links,

respectively, in the pair-relative motion

Table 1 presents type schemes and structural numbers of

suspensions with “negative” stiffness equipped with planar

FGMs All of the FGMs have 1-DOF, i.e., m1=1 Some may

have an idle degree of freedom, m2=1, due to the

redun-dancy (bolded members) In the schemes, positions 1-3

refer to fixed, input, and output links of the MKC, while

4-7 refer to the links of RKC The symbol; refers to a

load-capacity spring, and; and refer to the

redun-dant springs Here, the subscripts indicate whether spring

stiffness is “positive” or “negative” in motion

Type synthesis by the (1)-group methods presented in

(Alabuzhev, 1989; Chyuprakov, 1987) consists of design of

an RKC of one or more branches and connection of the

RKC to an initial FGM The RKCs are either made up of

low pairs (schemes 1-3 in Table 1), or of low and higher

pairs that result in m2={0; 1} (schemes 4-7) The numbers

m2=0 or m2=1 indicate that the RKC connected to the initial

FGM either does not change the mobility of the resulting

FGM or it increases it by 1 The enumeration of the initial

FGM shown in scheme 2 serves as an example It consists

of fixed link 1 and movable link 2 (n=2) generating a

revolute pair (p5=1) Enumeration by Equation (3a) results

in m1=0 Coupling the RKC of links 3 and 4 generates three

revolute pairs and the initial FGM does not change the

mobility of new FGM; thus, m2=0 which is shown clearly

by using Equation (3a) again In scheme 6, coupling the

RKC of links 3 and 4, generating one revolute pair between

them and two other pairs (one centroid and one prismatic

pairs), with an initial FGM increases the mobility of the

new FGM by 1, and thus m2=1 However, this redundancy

does not change the number of input parameters controlled

independently to bring the mechanism into a work position,

and m2=1 remains Subsequent dimensioning of links 4-7

and pre-loading (by tension, compression, etc.) of springs

and/or may result in “negative” stiffness for the

suspension in motion for link 2 at neutral position Only

joints for the MKC and RKC that might provide operability

for resulting FGM are shown to exist

Type synthesis by the (2)-group methods presented in

(Alabuzhev, 1989; Rivin, 2003; Chyuprakov, 1987) deals

with load-capacity springs under small-amplitude

post-buckling For instance, ropes, elastic rods, beams, and

similar springing elements under longitudinal deformation

could reveal “negative” stiffness in the same or in opposite

directions These approaches require the use of an FGM in

order to transform the indifferent equilibrium of such a

spring into a stable output motion Therefore, the (2)-group

methods are a variation of the (1)-group

Type synthesis by the (3)-group methods consists ofdesign of a redundant SMM and synthesis of a coupler

(RKC of RH-type) to connect the SMM to an initial FGM

(scheme 8) Links of the SMM are constrained with spring revealing “negative” stiffness before the coupling.Therefore, no matter what kinematic structure of an initialFGM, the performance of resulting FGM is determined bythe performance of the initial one, and it is indifferent toconnection of a redundant structural member (SMM).Additionally, spring , coupled in-parallel to a load-capacity spring, can control the stiffness of a seat suspen-sion up to the above minima (Formulae 2) during outputmotion of the resulting FGM (Goverdovskiy and Lee,2003; Lee and Goverdovskiy, 2002)

5 ATLAS OF FGMS FOR SEAT SUSPENSIONS WITH “NEGATIVE” STIFFNESS

In the FGM design, we assume in-parallel initial FGMs inwhich the MKC has several branches of links and pairsconnecting the fixed and output links This structure waschosen because seat suspensions must withstand largeforces during motion under infra-frequency vibration andimpulse excitation The resulting FGMs are intended to beinvariant to the structural features of the initial ones FromTable 1, connection of the RKC designed by the (1)-groupmethods or the SMM designed by the (3)-group methods(as mentioned above) do not change the number ofindependent input parameters (DOFs) required to bring theinitial FGM into a work position:

(4) where is the DOF the of resulting FGM after con-necting the RKC or SMM

Connection of the RKC or SMM also does not increase

the number of redundant constraints q RC, (classified here byReuleaux (Shigley and Uiker, 1995)), in the resulting FGMsince the RKC or chain of the SMM coupler is open:

(5)where is the total number of the pairs

(main constraints) of ith mobility in their relative motion, for a planar (s=3) or spatial (s=6) FGM Through enume-

ration of structural numbers by Formula (5), one may

obtain the condition, q RC=0

We modify each type of 1-DOF FGM from Figure 1 byjoining to it the RKC or SMM listed in Table 1, assuming

q RC=0 Then, we obtain variable FGM candidates for eachpermissible in-parallel configuration using the number per-mutations in Formula (5) Finally, we obtain an atlas of theresulting FGMs for seat suspensions Some are presented

in Table 2; here, each row meets the initial FGM fromFigure 1, and each column meets the RKC or SMM from

Table 1 In some graphs, one pair is shown with a dotted

line representing the variable stiffness of the pair-mating

surfaces (like the mechanisms of scheme 5 from Table 1)

Historically, the FGMs resulting from the (1)-group

methods have been designed, as mentioned above, from

successful compatibility of selected schemes from Table 1

and the corresponding initial FGMs listed in Figure 1 This

is the reason why many cells in Table 2 are void; it meansthat designers working empirically did not recognize agreat number of practical candidates The existing FGMsdesigned empirically for seat suspensions are shaded inTable 2 In contrast to empiricism, all of the FGMs (includ-ing the existing ones), appear in our atlas as naturalproducts of the enumeration process In addition, the atlas

2 Synthesis of a RKC including links 3-4 that are constrained with link 2 of the

4 Synthesis of a RKC including at least link 4 that is constrained with links 2-3

of the MKC and spring

5 Synthesis of a RKC including at least link 3 that is constrained with link 2 of

the MKC and springs and

7 Synthesis of a RKC including at least link 3-4 that are constrained with link 2

of the MKC and elastic link

8 Synthesis of a SMM equipped with a redundant spring and synthesis of a

RKC to connect the SMM and MKC