cvips_9_11

8/7/2019 cvips_9_11

1/10

Y. Y. Kim

Jong T. Lee

School of Mechanical Engineering,Sungkyunkwan University,

Jangan-gu, Suwon 440-746, Korea

J. A. CatonDepartment of Mechanical Engineering,

Texas A&M University,

College Station, TX 77840

The Development of aDual-Injection Hydrogen-FueledEngine With High Power and HighEfficiencyTo achieve high power and high efficiency in a hydrogen-fueled engine for all loadconditions, the dual-injection hydrogen-fueled engine, which can derive the advantages of

both high efficiency from external mixture hydrogen engine and high power from directcylinder injection was developed. For verifying the feasibility of the above engine, a

high-pressure hydrogen injector of ball-valve type and actuated by a solenoid was devel-oped. A systematic experimental study was conducted by using a modified single-cylinder

dual-injection hydrogen-fueled engine, which was equipped with both an intake injectorand high-pressure in-cylinder injector. The results showed that (i) the developed high

pressure hydrogen injector with a solenoid actuator had good gas tightness and finecontrol performance, (ii) the transient injection region, in which injection methods are

changed from external fuel injection to direct-cylinder injection, ranged from 59 to 74%

of the load, and (iii) the dual-injection hydrogen-fueled engine had the maximum torqueof direct-cylinder fuel injection and the maximum efficiency of external fuel mixture hy-

drogen engines. DOI: 10.1115/1.1805551

Introduction

Two major types of hydrogen-fueled engines have been devel-

oped over the years. These two types are i those using externalfuel mixture preparation and ii those using direct-cylinder fuel

injection 1 4. The use of an external fuel mixture preparationhas the advantages of a simple configuration and high efficiency

5, but low power due to backfire occurrence at high load 6,7.

On the other hand, a direct-cylinder fuel-injection engine in which

only air is induced during the intake process can obtain high

power by avoidance of backfire occurrence 8; but its thermal

efficiency becomes relatively lower due to a poor hydrogen-air

mixing rate, hydrogen leakage in the high-pressure injector, and

the losses associated with the injection system 9,10.

The final goal of the engine development is to achieve high

power and high efficiency simultaneously for practical use. It is

difficult for the case of external mixture preparation and for

direct-cylinder injection to meet these requirements because of the

above reasons 11. The dual-injection hydrogen-fueled engine,

which can derive the advantages of both high efficiency from

external mixture and high power from direct-cylinder injection,

was introduced by the authors 12. The dual-injection hydrogen-

fueled engine uses only external mixture under idling and low

load because no backfire occurs. For the case of high load, most of

the fuel is injected directly into the cylinder during the compres-

sion process and the rest, which guarantees that the intake mixture

is lean enough so that no backfire occurs, is supplied into theintake pipe to increase the mixing rate 13.

To realize the dual-injection hydrogen-fueled engine, a high-

pressure hydrogen direct-cylinder injector installed in the cylinder

head is necessary. This injector has to be capable of good gas

tightness against injected high-pressure hydrogen gas, and it also

has to accurately control the amount of fuel injection correspond-

ing with that of an intake injector under high engine speed. Mostof the hydrogen injectors used in previous hydrogen engines havenot met the above requirements.

The fuel-injection regions of dual-injection hydrogen-fueled en-gines are divided into external fuel-injection, transient-injection,and direct-cylinder fuel-injection regions. The transient-injectionregion is one in which the injection methods are varied from ex-ternal fuel injection to direct-cylinder fuel injection. If the injec-tion method is shifted rapidly in this transient-injection region, itmay affect unfavorably on the hydrogen-air mixing rate and,therefore, on the combustion performance. This may result in un-stable operation, such as higher cycle variations and more torque

fluctuations. The suitable injection conditions in the transient injection region, therefore, have to be established to obtain stablecombustion and high performance for the dual-injectionhydrogen-fueled engine.

At the beginning of this study, a high-pressure hydrogen injec-tor using a poppet valve of ball-valve type and actuated by asolenoid was developed to verify the feasibility of the dual-injection hydrogen-fueled engine development. An experimentalsingle-cylinder engine was modified into the dual-injectionhydrogen-fueled engine by equipping this engine with both thishigh-pressure injector and a low-pressure injector. The proper in-

jection timing for the direct-cylinder injection was determined bycycle analyses of the two kinds of dual-injection hydrogen-fueledengines. The suitable operating conditions, including the transient-injection region and engine performances, were established by

using the experimental dual-injection hydrogen-fueled engine.The feasibility of this engine, which can achieve high power orhigh efficiency according to load conditions, was also evaluatedby comparing with the performance characteristics of typical hy-drogen engines.

Cycle Analyses and Determining the Proper Dual Injec-

tion Cycle

Cycle Analyses of the Dual Injection Hydrogen Engines.The method of direct-cylinder injection is divided by injectiontiming into two types. The first type is the method of injectionearly in the compression process, and the other type is that of

Contributed by the Internal Combustion Engine Division of THE AMERICAN SO-CIETY OF MECHANICAL ENGINEERS for publication in the ASME JOURNAL OF

ENGINEERING FOR GAS TURBINES AND POWER. Manuscript received by the ICE

Division January 20, 2003; final revision received by the ASME Headquarters April

23, 2004. Associate Editor: D. Assanis.

Journal of Engineering for Gas Turbines and Power JANUARY 2006, Vol. 128 203Copyright 2006 by ASME

8/7/2019 cvips_9_11

2/10

injection late in the compression process. Since direct-cylinderinjection is adopted at high load for the dual-injection hydrogen-fueled engine, there are also two types of the dual-injection

hydrogen-fueled engines injecting at early compression processand at late compression process. These engines are called thedual-injection hydrogen engines with early direct-cylinder injec-tion and with late direct-cylinder injection, respectively.

For the purpose of investigating the proper cycle for dual-injection hydrogen engines, both cycles of early direct-cylinderinjection and late direct-cylinder injection were examined. Thisexamination is based on theoretical, idealized cycles.

Early Direct-Cylinder Injection. Figure 1a shows the p-Vdiagrams of the theoretical cycle of the dual-injection hydrogenengine with early direct-cylinder injection. For the case of lowload, hydrogen fuel is only supplied into the intake pipe and thefuel-air mixture is burned under near-constant volume combus-tion, like the 23 processes. The cycle diagram for this case isequal to the Otto cycle. For high load, the in-cylinder pressure is

increased, due to the supplied energy, by direct-cylinder injection;hydrogen and air are pre-mixed and then burned under near-constant volume combustion. So, the theoretical cycle of the dual-injection hydrogen engine with early direct-cylinder injection isthe same as for the typical Otto cycle for all load conditions.

Because hydrogen gas is induced quickly by direct-cylinder injection, the temperature and the pressure of the working fluid

during the constant volume process, 1 1 , are changed under anadiabatic process. Define Rf e mass of external fuel/mass oftotal fuel as the ratio of the amount of fuel supplied into intakepipe to the amount of total fuel supplied in the cycle. A value ofzero for Rf e means only direct in-cylinder injection, and a value of100% for Rf e means only external mixture. The theoretical ther-modynamic thermal efficiency de of the dual-injection hydrogenengine with early direct-cylinder injection is given by

de11k

B k1de1

B k1de1 (1)

where,

de1Hu f

a fCvaCvfT1k1B k1

Bm 1

m 1

a f1

a fRf e

where is compression ratio, k is the specific heat ratio, Hu f is thelower heating value of hydrogen, Cva and Cvf are the specificheats at constant volume for air and hydrogen, a f is the air-fuelmass ratio, and T1 is the initial temperature.

The theoretical mean effective pressure pmi (de ) is

pmi de dep1

kB k1de1

S ck1 1 (2)

where p1 is initial pressure, and Sc(a fCvaRf eCv)/ (a fCvaCvf).

Late Direct-Cylinder Injection. The theoretical p-V diagramof the dual-injection hydrogen engine with late direct-cylinder in-

jection is shown in Fig. 1b. The cycle for low load is the same asfor the early direct-cylinder injection due to only using externalmixture. But, for high-load operating conditions, a lean hydrogen-air mixture is induced into the cylinder during the intake process.This mixture undergoes a near-constant volume premixture com-bustion process and most of the hydrogen fuel injected into thecylinder undergoes a near-constant pressure-combustion process.The dual-injection hydrogen engine with late direct-cylinder in-

jection, therefore, theoretically works on the Dual cycle havingboth constant volume and constant pressure combustion processesat high load.

The pressure ratio, dl (p2 /p2), and the shut off ratio,

(V3 /V2), meaning the quantity of supplied heat energy in the

2 2 and the 2 3 processes are concerned with the ratio of theamount of fuel injection Rf e . The theoretical thermal efficiencyfor the above engine dl is as follows:

dl11k dl

kSR1

SCdl1 kdl SR1 (3)

where

dl1Rf eHu f

a fCvaRf eCvfT1k1

1 1Rf eHu fa fCva1Rf eCvfT1

k1dl SR

SRa fRaRf eRfa fR aRf

And the theoretical mean effective pressure p mi (dl ) is given by

p mi dl dlp 1

kdl1 SC1dl SR1

k1 1 (4)

Considering hydrogen and air as ideal gases of diatomic mol-ecules, the theoretical specific heat ratios of hydrogen and air areequal in Eqs. 3 and 4. So, SR is equivalent to SC .

Determining the Proper Dual-Injection Cycle. Compres-sion ratios of conventional SI engines are mainly limited by knockoccurrence, but compression ratios for hydrogen engines dependon backfire as well as knock due to the fast burning velocity, low

ignition energy, and wide flammability of hydrogen gas.Backfire phenomenon is generally attributed to a fresh mixturein the cylinder being ignited by any ignition source while theintake valve is open. For the case of low load, fuel injection intothe intake pipe is used, and then the fresh mixture related to back-fire exists in the intake pipe and in-cylinder for both types ofdual-injection hydrogen engines. This means that the compressionratios for these engines are subject to backfire occurring underexternal injection region, though the compression ratio of the Dualcycle is basically higher than that of the Otto cycle. So, compres-sion ratios for the two types of the dual-injection hydrogen-fueledengines may be similar to each other. For the same compressionratio, the dual-injection hydrogen engine with early direct-

Fig. 1 Theoretical p-V diagrams of the dual-injection hydrogenengine cycle for early and late direct-injection timing

204 Vol. 128, JANUARY 2006 Transactions of the ASME

8/7/2019 cvips_9_11

3/10

cylinder injection is more favorable than that with late direct-cylinder injection because the thermal efficiency of the Otto cycleis higher than that of the Dual cycle.

In general, the thermal efficiency of a direct-injection hydrogenengine becomes relatively lower compared to an external-mixturehydrogen engine due to the short hydrogen-air mixing durationand less mass of hydrogen gas 14. To enhance the poor mixingrate due to the above reasons, a portion of the hydrogen gas isinduced by auxiliary external injection into the intake pipe in thecase of high load for the two types of dual-injection hydrogenengines. Though the enhancement effect of the mixing rate isaided by the auxiliary injection, the mixing of hydrogen and air inthe cylinder for the direct-cylinder injection mainly relies on the

mixing duration, which is the period after fuel injection and upuntil ignition. The dual-injection hydrogen engine with earlydirect-cylinder injection, which has a longer mixing duration, isalso more favorable than that with late direct-cylinder injectionwith respect to the mixing rate.

Furthermore, the higher injection pressure for the dual-injection hydrogen engine with late direct-cylinder injection hasdisadvantages in a practical application due to the difficulties ofkeeping gas tightness, use of a high-pressure hydrogen gas, andfuel supply system. The early direct-cylinder injection, therefore,is employed as the proper direct-injection method for the dual-injection hydrogen-fueled engine for the above reasons.

Experimental Apparatus and Methods

High-Pressure Hydrogen Injector of Ball-Valve Type Actu-ated by a Solenoid. A high-pressure hydrogen injector for thedual-injection hydrogen engine should be capable of keeping gastightness against high-pressure hydrogen gas. Its driving parts,which are in contact with the hydrogen gas, must not wear out bythe nonlubricant property of hydrogen. In addition, a high-pressure hydrogen direct-cylinder injector interlocking with an in-take injector has to control the amount of fuel injection underhigh-speed operation.

A high-pressure hydrogen injector of the ball-valve type andactuated by a solenoid was developed to meet the above require-ments. Figure 2 shows a schematic cutaway diagram of this high-pressure hydrogen, direct-cylinder injector. It consists of a sole-

noid actuating part, push-rod, and injection valve part of ball typeand a holder with injection holes. The design concept of this in-

jector is based on those of poppet-valve-type and ball-valve-typehigh-pressure hydrogen injectors, which are actuated by a hydrau-lic driving system. This type of injector has been developed by theauthors in preceding research, which was concerned with thedirect-cylinder injection hydrogen engine 15,16.

The solenoid actuator, compared to the hydraulic actuator, caneasily control the injection parameters and has a faster responseunder high-speed operation. Due to the high pressure of the sup-plied hydrogen gas and the high spring tension acting on the pop-pet valve, the driving force generated from a conventional sole-noid is not enough to actuate a high-pressure hydrogen injector.For adopting a solenoid actuator, it is necessary to decrease thedriving force required by the hydrogen injector. This driving force

has been reduced by using the principle of pressure-offset due tothe area difference between the poppet valve fillet and the pistonfor generating the differential pressure. The piston for generatingthe differential pressure is installed at the valve stem end, and itsarea is a little larger than that of the poppet valve fillet.

The differential pressure due to the area difference automati-cally closes the poppet valve. The O-ring is inserted in the pistonto prevent hydrogen leakage. The valve spring installed below thepiston has the primary role of keeping the hydrogen injector clos-ing, even though the hydrogen supplement is shut off, and it hasthe secondary role of helping the poppet valve to return back tothe valve seat. The wear between the valve stem and the stemguide due to the nonlubricant property of hydrogen gas will causethe valve center to be eccentric and results in poor gas tightness.In general, the ball valve is always apt to be fitted with the valve

seat because it has a spherical geometric surface. Furthermore, therotation of the ball valve during its operation has an effect topolish the valve seat. This can lead to better gas tightness. Tofurther improve the gas tightness of the high-pressure hydrogendirect-cylinder injector, the poppet valve face is formed into a ballshape as shown in Fig. 3. To rotate the poppet valve, the blade isinstalled in a valve-head direction of the valve stem. The pistonand the valve stem end are linked by a spherical pair structure toprevent the valve stem from being eccentric. The valve rotationforce is generated by the kinetic energy of the hydrogen jet flowpassing through the blade in the valve stem. The blade is located10 mm from the valve head. Its geometric configurations are 1.5mm wide, 1.5 mm high, 3.5 mm long and 30 deg deflection angle.

Fig. 2 Schematic cutaway diagram of the high-pressure hy-drogen injector of ball-valve type with solenoid actuator

Fig. 3 The poppet valve face with a shape of ball in the high-pressure hydrogen injector

Journal of Engineering for Gas Turbines and Power JANUARY 2006, Vol. 128 205

8/7/2019 cvips_9_11

4/10

The shape of entrance passage toward the blade is a nozzle typefor the conversion of the hydrogen pressure energy to kinetic en-ergy.

The combustion and performance of the hydrogen engine withdirect-cylinder injection mainly depends on the homogeneity ofthe hydrogen-air mixture, which is affected by the injection pres-sure, the configuration of the injection holes, the combustionchamber shape, and the in-cylinder flow pattern. The hydrogenmomentum decreases rapidly after injection due to its low density.For the cases of constant engine operating conditions, therefore, a

configuration of the injection holes should be a very importantfactor to enhance the in-cylinder mixing rate of the. The visual-ization of the mixing process and the analysis of engine perfor-mance have been carried out to determine the suitable configura-tion of the injection holes in the previous studies. The nine-holetype, which is the most favorable among the proposed five typesas shown in Fig. 4 17, is used in this hydrogen injector.

Figure 5 shows the leakage flow rate of hydrogen gas as afunction of the number of injections. The leakage flow rate in-creased to about 800 injections and then rapidly decreased as thenumber of injections is further increased. The leakage rate be-comes as low as 50 cc/min for over 1600 injections. This may becaused by the fact that the gas tightness of the hydrogen injector isimproved by the polishing effect of the valve seat due to valverotation. This leakage rate is negligible as it is estimated to be lessthan about 0.5% of the lean flammable limit of the hydrogenmixture at 1000 rpm engine speed. The increasing and decreasing

tendencies of the leakage rate mean that the polishing effect by thepoppet valve rotating contributes to the improvement of the gas

tightness of the hydrogen injector.

Dual Injection Hydrogen Fueled Engine. Figure 6 showsthe experimental dual-injection hydrogen-fueled engine. It is asingle-cylinder 500 cc engine with 9:1 compression ratio, andmade by modifying conventional engine parts. Table 1 lists themain engine specifications for this engine.

For proper use of the external mixture or the direct-cylinderinjection according to the load conditions, the low-pressure intakeinjector and the high-pressure direct-cylinder injector as men-tioned above are installed in the intake pipe and cylinder head,respectively. The low-pressure hydrogen injector is modified froma conventional CNG injector SPI. The installed location of thelow-pressure intake injector is 100 mm from the intake valvestem, and the installed location of the high-pressure direct-cylinder injector is at the flywheel end of cylinder head with a 60

deg slope.Because of the low ion density of hydrogen-air mixture flames,

much residual electric energy may remain in the ignition systemof a hydrogen engine after normal ignition 18. Discharging thisresidual energy abnormally during the intake process or valveoverlap would lead to backfire occurrence 19. For preventingabnormal ignition, a high value of earth resistance is connected toa secondary coil of a full transistor-type ignition system.

The coolant is supplied into the cylinder head, the cylinderblock and the oil cooler separately. The coolant temperature iscontrolled by using a solenoid valve installed at the coolant pas-sage outlet.

Experimental Apparatus and Method. Figure 7 is a sche-matic diagram of the experimental apparatus, including the dual-injection hydrogen-fueled engine, the hydrogen supply system,

Fig. 4 Five types of the injection hole configurations

Fig. 5 The leakage flow rate of the high-pressure hydrogeninjector as a function of the number of injections

Fig. 6 Photograph of the dual-injection hydrogen engine

Table 1 Specifications of the dual injection hydrogen engine

Engine Single cylinderBore 86.0 mmStroke 86.0 mmDisplacement volume 500 ccCombustion chamber Pent-roof typeCompression ratio 9:1Ignition system Full transistor typeCooling type Water coolingSpark-plug Cold rating type


8/7/2019 cvips_9_11

5/10

the dynamometer unit, the injector control system, and the dataacquisition system. Hydrogen gas decompressed from a high-pressure cylinder by a two-stage regulator is induced into thelow-pressure intake injector and high-pressure direct-cylinder in-

jector. The lift of the hydrogen injector is measured by a gapsensor Ono-Sokki, VE231 installed in the back of the solenoid

armature plate. The injection timing and injection duration of theinjectors are varied by the electronically controlled injection sys-tem with 8255I/O detecting TDC and angle signals. For prevent-ing the solenoid from overheating and improving its response,electric current with peak-hold wave form is supplied to the sole-noid. The coil suppressor, consisting of a resistance and a diode, isinstalled in each end of the solenoid in parallel to remove thecounterelectromotive force, which occurs when the solenoid isshut off.

The experiment is to measure whether backfire occurs or notand to determine the engine performance as a function of theinjection timing, the injection pressure, the fuel-air equivalenceratio, and the ratio of the amount of fuel supplied into the intakepipe to the amount of total fuel supplied in the cycle Rf e . Thepressures of the intake injection and direct-cylinder injection are

varied from 3 to 5 bar, and from 20 to 30 bar, respectively. Thefuel-air equivalence ratio is changed from 0.3 to 1.2 by 0.1. TheRf e is increased from 0% to its maximum value, where no back-fire occurs. For each experiment, the coolant temperature, enginespeed, intake throttle valve, and spark timing are fixed with valuesof 70 C, 2000 rpm, WOT Wide Open Throttle, and MBT Mini-mum spark advance for Best Torque, respectively.

Suitable Operating Conditions of Dual Injection

Injection Timing and Injection Pressure

For the Intake Injection. Figure 8 shows the torque and thebackfire limit BFL equivalence ratio for the case of the externalmixture as a function of the intake injection pressure. The BFLequivalence ratio is defined as the upper limit of the fuel-air

equivalence ratio where the hydrogen engine can be normally op-erated without backfire occurrence. The injection timing is set tointake TDC according to the above results. The torque is almostconstant with all intake injection pressures. But the BFL equiva-lence ratio is decreased as the intake injection pressure is in-creased. This may be caused by the possibilities that any ignitionsource causing backfire occurrence and burning velocity are in-creased due to the difference of mixing rate and the increase offlow velocity of in-cylinder mixture according to variation of in-

jection pressure. The thermal efficiency of the dual-injection hy-drogen engine with external mixture is higher than direct-cylinderinjection so that it is reasonable that the BFL equivalence ratio isexpanded to increase the thermal efficiency of the dual injection

as high as possible. From the above results, intake TDC for intakeinjection timing and the 3 bar for injection pressure are adopted,respectively.

Figure 9 illustrates the torque and the intake air flow rate in thedual-injection hydrogen engine using only the external mixture asa function of intake injection timing. The highest torque is shownat intake TDC injection, and the torque has decreasing tendency asthe intake injection timing is advanced or retarded from intakeTDC. But the intake air flow rate is gradually decreased withretarding the intake injection timing. It may be due to the decreaseof entrainment duration entraining an air into hydrogen jet in-

jected with high speed. With the advance of injection timing, thetorque is decreased in spite of the increase of intake air flow, sincea small portion of the induced hydrogen in cylinder is leaked outto the exhaust pipe during the valve overlap. The torque is de-

creased by the decrease of the amount of intake air flow with theretard of the injection timing from intake TDC.

Fig. 7 Schematic diagram of experimental apparatus

Fig. 8 The torque and the intake air flow of the dual-injectionhydrogen engine with external mixture

Fig. 9 The torque and the BFL equivalence ratio of the exter-nal mixture for each intake injection pressure


8/7/2019 cvips_9_11

6/10

For Direct Cylinder Injection. Figure 10 shows the torque andthe thermal efficiency using only direct-cylinder injection as afunction of injection timing for three injection pressure condi-tions. The torque shows a little increase until the injection timingis retarded to the intake valve close timing 55 deg aBDC, butthen remains unchanged in spite of further retarding the injectiontiming. The decrease of the torque before intake valve close tim-ing is caused by a restriction of the induction of intake air massflow due to the increasing of in-cylinder pressure related to directinjection.

The thermal efficiency shows an increasing tendency with theadvance of injection timing because the hydrogen-air mixing du-ration becomes relatively longer. As hydrogen is injected beforeintake valve close, a portion of the hydrogen flows back intointake pipe. This hydrogen-air mixture that flowed back into the

intake pipe could cause a backfire occurrence. It is concluded thatthe proper direct-injection timing is, therefore, 120 deg bTDCretarded by 5 deg from the intake valve close timing.

As shown in Fig. 10, the torque and the brake thermal effi-ciency are essentially unchanged in spite of increasing the direct-cylinder injection pressure, and higher injection pressures havethe disadvantages of more difficult gas tightness and practical use.So, for direct injection pressure, 20 bar is confirmed as the propercondition.

Transient Injection Region Changing Injection Method.The injection regions of the dual-injection hydrogen-fueled enginecan be divided according to the fuel supply method into i theregion of external mixture preparation, ii the transient-injectionregion of changing injection method, and iii the region of direct-

cylinder injection. The transient-injection region is defined as theshift region of injection method where the amount of external fuelmixture is decreased and that of direct-cylinder injection is in-creased, simultaneously.

The brake thermal efficiencies and the occurrence point ofbackfire with engine loads are shown for the hydrogen enginewith external mixture and direct-cylinder injection in Fig. 11. Us-ing external mixture injection has higher brake thermal efficien-cies compared with direct-cylinder injection, but normal operationis restricted to under 62% of load because of backfire occurrence.The start point of transient injection region, therefore, has to beprior to 62% of load. Determining the end point of that, it is worthregarding how to increase the relative low-brake thermal effi-

ciency of direct injection as high as possible. The low thermalefficiency of direct injection is due to the shorter hydrogen-airmixing duration. As mentioned previously, supplying a portion ofhydrogen into the intake pipe of the dual injection engine beforedirect-cylinder injection could enhance the mixture homogeneityand lead to higher brake thermal efficiency. To evaluate the effectof an auxiliary supply of external mixture on the brake thermalefficiency, the brake thermal efficiency of direct cylinder injectionis investigated with the amount of hydrogen gas supplied into theintake pipe, and the results are also shown in Fig. 11. The direct-cylinder injection with some external mixture has a higher brakethermal efficiency than only direct injection as a whole. The brakethermal efficiency shows further increased tendency as the amount

of the external mixture is increased. The more auxiliary externalmixture is supplied with direct injection, the higher its brake ther-mal efficiency is increased. But the amount of external mixturesupplied with direct-cylinder injection depends on backfire occur-rence. The upper limited value of engine load operated withoutbackfire occurrence is around 75% in 47% of Rf e , and 77% in22% of Rf e . Under 10% of Rf e , nonbackfire appears and a littleincrement of brake thermal efficiency is shown. This means thatthe transient-injection region should start around 62% and endaround 78% of load. Using 5% to the above range for a safemargin for maintaining reliable normal operation, the transientinjection region ranged from 59 to 74% of load.

The brake thermal efficiency varied little for up to 10% supple-ment of hydrogen gas into intake pipe. Since the auxiliary supplyof the external mixture in the direct-cylinder injection region con-

tributes to the enhancement of the mixture homogeneity and theutilization of the low-pressure hydrogen gas, it seems to be betterthat about 10% of external mixture to total fuel amount is suppliedin the region of direct-cylinder injection.

The heat release per cycle and the fuel-air equivalence ratio ofthe dual-injection hydrogen engine operated in the transient-injection region are shown in Fig. 12. The heat release continu-ously increased with the increase of load. The fuel-air equivalenceratio is increased until the end point of the region of externalmixture, but adversely decreased from the transient-injection re-gion and then increased again from that of direct-cylinder injec-tion. Although the heat release is increased in the transient-injection region, the fuel-air equivalence ratio is decreased by the

Fig. 10 The torque and the brake thermal efficiency of direct-cylinder injection for each injection pressure as a function ofinjection timing

Fig. 11 The brake thermal efficiency and where backfire oc-curs for external mixture and direct-cylinder injection with thechange of the amount of external mixture


8/7/2019 cvips_9_11

7/10

increase of intake air flow due to reducing of the amount of intake

hydrogen injection. The fuel-air equivalence ratio for direct-cylinder injection begins to increase in accordance with the in-crease of the amount of fuel mass acquired because the amount ofintake air mass flow is fixed. For these reasons, the transient-injection region from 59 to 74% of load is equivalent to be from0.7 in the region of external mixture to 0.65 with 10% of

Rf e in the region of direct-cylinder injection.

Injection Conditions in the Transient-Injection Region.Figure 13 shows the torque and brake thermal efficiency in thetransient-injection region as a function of the Rf e . Here, 0% in

Rf e means only direct injection and 100% in Rf e means onlyexternal mixture. The brake thermal efficiency is highest for afuel-air equivalence ratio of 0.6 and decreases slightly with theincrease of the fuel-air equivalence ratio. As the Rf e is increased,the brake thermal efficiency is also increased by the tendency of

the enhancement of mixture homogeneity, but the torque is de-creased due to the decrease of the amount of intake air flow in

spite of the increasing effect of the brake thermal efficiency. Thereduction rate of torque is increased by the decrease of intake airflow mass. The torque reduction rate is about 6.2% at the fuel-airequivalence ratio of 0.65, and about 12% at an equivalence ratioof 0.7.

Three processes A, B, and C in Fig. 14 of changing the fuelsupply during the transient-injection period are examined. Thesethree processes change the fuel-air equivalence ratio and the Rf ein the transient injection region as shown in Fig. 14. These threeprocesses have to avoid the upper limited Rf e to avoid backfireoccurrence in the beginning stage of the transient-injection region,hence the process of the transient-injection region has to start withthe increase of d/dR f e .

The A-process begins to change injection method for conditionof the largest d/d Rf e than any other process. It decreases boththe fuel-air equivalence ratio and the Rf e up to the lower fuel-airequivalence ratio than that of the end point of the transient injec-tion region, and then increases the fuel-air equivalence ratio anddecreases the Rf e . The A-process can obtain higher thermal effi-ciency in the transient injection region because it passes near to0.6, where thermal efficiency is relatively higher. But if thed/dR f e is increased so much, it is difficult to control the amountof fuel injection and could cause the fluctuation of torque.

For the case of the B-process, the fuel-air equivalence ratio andthe Rf e are decreased until its fuel-air equivalence ratio is thesame as that of the end point of the transient-injection region, andthen the Rf e is decreased with the constant fuel-air equivalenceratio. As the Rf e is decreased under the constant fuel-air equiva-

lence ratio, the amount of the direct-cylinder injection is increasedby the increase of intake air mass flow due to the reduction of theamount of intake fuel injection. This may result in the torquefluctuation of the dual-injection hydrogen engine; hence, this pro-cess has a demerit in that the intake air flow is controlled by thethrottle valve installed in the intake pipe.

The C-process is to decrease the fuel-air equivalence ratio andthe Rf e gradually in the transient injection region. Comparing withthe other processes, it is easy to control the amount of fuel and theengine output, and the unstable operation should rarely occur. Thethermal efficiency of the C-process is lower than that of theA-process, but its difference is very slight and the C-process has abetter controllability on the amount of fuel injection. Therefore,

Fig. 12 The heat release per cycle and the fuel-air equivalenceratio of the dual-injection hydrogen engine

Fig. 13 The torque and the thermal efficiency in the transientregion with the change of Rfe

Fig. 14 Three types of processes how to change injectionmethod in the transient region


8/7/2019 cvips_9_11

8/10

the C-process is confirmed to be the suitable injection condition inthe transient injection region from the above results.

Characteristics of Performance in the Dual Injection

Hydrogen Engine

Combustion and Operation Characteristics. Figure 15shows the comparison of the in-cylinder pressure of the dual-injection hydrogen engine with that of direct-cylinder injectionhydrogen engine for the same fuel amount. The conditions for thistest are that the Rf e of the dual injection is 30% and the sparktiming is 10 deg bTDC. The maximum pressure of the dual injec-tion is 48 bar, which is a little higher than the 45 bar of thedirect-cylinder injection. The appearance point of maximum pres-

sure is also advanced about 2 deg more than for the direct injec-tion case. This seems to be caused by the fact that the combustionin the dual injection is promoted due to the enhancement ofhydrogen-air mixing rate by the auxiliary hydrogen gas injectioninto the intake pipe during the intake process.

To clarify these tendencies, the combustion durations are inves-tigated as a function of Rf e , and the results are shown in Fig. 16.Here, constant fuel mass is supplied into the cylinder though Rf eis changed. The combustion duration is obtained from an en-semble average of the in-cylinder pressure data from 100 cycles.

The combustion duration, which is the duration of 0100%mass burned rate, has a decreasing tendency with the increase ofthe Rf e . This tendency may be explained by the fact that thecombustion promotion depends on the increase of the fuel-airequivalence ratio by the decrease of the intake air mass flow and

the enhancement of mixing rate. The duration of 0 10% massburned rate shows almost the same 8 deg crank angle regardlessof the change of the Rf e , the duration of 1090% is decreaseddue to the enhancement of mixing state with the increase of the

Rf e , and the duration of 90100% shows a slight decreasing ten-dency.

In general, higher maximum in-cylinder pressures increasesNOx emission so that a dual-injection hydrogen engine may beinclined to have increased NOx emissions. But a dual-injectionmethod may have a stratifying effect on the hydrogen-air mixture.Potentially, this could lead to the decrease of the maximum tem-perature and, thus, NOx emission would be decreased with theretard of direct injection timing 20. Lean burn, EGR, and ex-

haust aftertreatment are considered the possible strategies toachieve a near-zero emission level for a dual-injection hydrogenengine as would be true for a conventional hydrogen engine. NOxemission is one of the important remaining problems that must besolved before the hydrogen engine may be put into production.Therefore, possible strategies for reducing NOx emission for thedual-injection hydrogen engine will have to be investigated in thenext research step.

Figure 17 represents the minimum spark advance for the besttorque MBT as a function of load. As the load of the dual-injection hydrogen engine is increased, the MBT is retarded in theregion of external mixture, but advanced slightly in the transient-injection region and retarded adversely in the region of direct-cylinder injection. One of the most important parameters in deter-

Fig. 15 The comparison of in-cylinder pressure of the dualinjection with that of direct cylinder injection

Fig. 16 The combustion duration of the dual-injection hydro-gen engine at each Rfe

Fig. 17 The minimum spark advance for best torque of thedual-injection hydrogen engine as a function of load


8/7/2019 cvips_9_11

9/10

mining the MBT is the combustion duration, which is essentiallydependent on the fuel-air equivalence ratio. Since the fuel-airequivalence ratio is increased according to the increase of theengine load, the MBT is retarded as a whole. The fuel-air equiva-lence ratio is decreased by the reduction of intake air flow rate inthe transient injection region so that the MBT is a little advanced.Although a point of inflection of the MBT exists in the transient-injection region, this may not be a problem to control spark igni-tion timing because the differences are negligible.

Figure 18 shows the fuel mass flow rate and the COV imep withthe increase of the load. The COVimep is defined as the coefficientof variation of the indicated mean effective pressure. The numbersof consecutive sample cycles are 100. The figure shows the totalfuel amount of the dual injection is nearly proportional to the load,and its gradient of increment has the increasing tendency from90% of the load. For the transient injection region, the fuelamounts of intake injection and direct-cylinder injection are var-ied with a little higher deviation in the beginning stage and gradu-ally in the later stage.

It can be also seen that the COVimep of the dual-injection engineis a little higher than that of the external mixture, but lower thanthat of the direct-cylinder injection as a whole. The COV imep ofthe transient-injection region is slightly increased by the decreaseof the external mixture and the increase of the direct-cylinderinjection. The COVimep of the dual injection does not exceed about5% that could cause unstable engine operation. This means thatthe dual-injection hydrogen engine can be operated with goodstability for all load conditions. No combustion noise or vibrationwere observed in the transient-injection region where the injectionmethods are changed.

Effect of Dual Injection on Improving Performance. Thetorque and brake thermal efficiency of the dual-injection hydrogenengine as compared to external mixture and direct-cylinder injec-tion are shown in Fig. 19. As the mass flow rate of injected fuel isincreased, the torque of external mixture and the dual injection areincreased. But the maximum value of external mixture is toward0.75 due to backfire occurrence, whereas that of the dual in-

jection continues to be increased about up to 1.1. The maxi-mum torque of the dual injection is higher by 60% than that of theexternal mixture, and it is almost equal to that of the direct-cylinder injection. The torque of the dual injection in the transient-injection region is increased proportionally with the increase ofthe amount of supplied energy without its fluctuation.

The brake thermal efficiency of the dual injection, as comparedto direct-cylinder injection, is improved for all load conditions.The improvement in low load is remarkable. The incremental rateof the brake thermal efficiency by adopting the dual injectionrepresents about 22% for low load and about 5.3% for high loadconditions, compared to direct injection.

Considering the above results, it was verified that the dual-injection hydrogen-fueled engine can be operated with stabilityand achieve both maximum power of direct-cylinder injection hy-drogen engine and the maximum efficiency of an external-mixturehydrogen engine.

Summary and Conclusions

Several tasks were completed to develop the dual-injectionhydrogen-fueled engine with high power and high efficiency.First, a high-pressure hydrogen injector of ball-valve type actu-ated by a solenoid was developed. This high-pressure injector andan intake injector were installed on a modified single-cylinderdual-injection hydrogen engine. The systematic experimentalstudies on the proper direct-cylinder injection method, the suitableoperating conditions, including the transient injection region, andthe characteristics of combustion and power were investigated byusing the above experimental engine. In addition, the feasibility ofachieving high power and high efficiency were evaluated by com-paring to typical hydrogen engines. In summary,

1. The gas tightness of the high-pressure hydrogen injectorwith a solenoid actuator was improved by the use of a pop-pet valve face with a spherical shape, the linkage of spheri-cal pair structure between the valve stem end and the piston

for generating differential pressure, and the rotation of thepoppet valve.2. Considering the theoretical cycle analyses and the facilities

of practical use, the dual injection with early direct injectionduring compression process was more reasonable than thatwith late direct injection.

3. The transient injection region, in which the injection meth-ods are quickly varied from external fuel injection to direct-cylinder fuel injection, ranged from 59 to 74% of load. Thebest fuel supply method to obtain stable operation in thetransient injection region was determined to be when boththe fuel-air equivalence ratio and the Rf e were decreasedgradually.

Fig. 18 The fuel mass flow rate and COVimep of the dual-injection hydrogen engine as a function of load

Fig. 19 The comparison of the torque and the thermal effi-ciency of the dual-injection engine with that of external mixtureand that of direct injection


8/7/2019 cvips_9_11

10/10

4. The maximum torque of the dual-injection hydrogen enginewas increased by about 60% compared to a hydrogen engineusing external mixture preparation, and the brake thermalefficiency was higher by about 22% at low load comparedwith direct-cylinder injection hydrogen engine.

5. From the above results, it was shown that the dual-injectionhydrogen engine is feasible. This engine can derive the ad-vantage of both high efficiency from external mixture andhigh power from a direct-cylinder injection hydrogen en-gine, and has stable operation performance for all load con-ditions. Further studies on the dual-injection hydrogen en-

gine will be needed for the continued refinement of thisconcept.

Acknowledgment

The authors would like to gratefully acknowledge generoussupport of The Korean Energy Management Corporation.

References

1 Furuhama, S., 1991, Trend of Social Requirements and Technical Develop-ment of Hydrogen-Fueled Automobiles, JSME Rev., 13, pp. 413.

2 Blarigan, P. V., Paradiso, N., and Goldsborough, S., 1998, HomogeneousCharge Compression Ignition With a Free Piston: A New Approch to Ideal

Otto Cycle Performance, SAE Technical Paper 982484.3 Sxwabowski, S. J., Hashemi, S., Stockhausen, W. F., Natkin, R. J., Reams, L.,

Kabat, D. M., and Potts, C., 2002, Ford Hydrogen Engine Powered P2000

Vehicle, SAE Technical Paper 2002-01-0243.

4 Nakaijma, Y., Yamane, K., Shudo, T., Hiruma, M., and Takagi, Y., 2000, Re-search and Development of a Hydrogen-Fueled Engine for Hybrid Electric

Vehicles, SAE Technical Paper 2000-01-0993.5 Lutz, A., and Keller, J., 2000, Hydrogen Fueled Engine in Hybrid Vehicles,

SAE Technical Paper 2000-01-0546.

6 Lynch, F. E., 1974, Backfire Control Techniques for Hydrogen Fueled Inter-nal Combustion Engines, Hydrogen Energy, Part B, Proc. World Hydrogen

Energy Conference, Miami, International Association for Hydrogen Energy,

pp. 686696.7 Lee, J. T., Kim, Y. Y., Lee, C. W., and Caton, J. A., 2001, An Investigation of

a Cause of Backfire and Its Control Due to Crevice Volumes in a Hydrogen

Fueled Engine, ASME J. Eng. Gas Turbines Power, 123, pp. 204210.

8 Furuhama, S., 1977, Combustion Improvement in a Hydrogen Fueled En-gine, Int. J. Hydrogen Energy, 2, pp. 329340.

9 Kim, J. M., Kim, Y. T., Lee, J. T., and Lee, S. Y., 1995, Performance Char-acteristics of Hydrogen Fueled Engine With the Direct Injection and Spark

Ignition System, SAE Technical Paper No. 952488, pp. 162175.

10 Tsujimura, T., Mikami, S., Achiba, N., Tokunaga, Y., Senda, J., and Fujimoto,H., 2003, A Study of Direct Injection Diesel Engine Fueled With Hydrogen,


11 Lee, J. T., 1996, Current Status and Characteristics of Hydrogen Fueled En-gine, Trans. KSAE, 18, pp. 2952.

12 Kim, Y. Y., Park, J. B., and Lee, J. T., 1997, A Study on Development ofHydrogen Fueled Engine With High Power and High Efficiency, Proc. 4th

KHES-HESS Joint Symposium, Yokohama, Aug., Hydrogen Energy SystemsSociety, pp. 6271.

13 Choi, H. K., Ahn, J. Y., Kim, Y. Y., and Lee, J. T., 1999, A Basic Study on theHydrogen Fueled Engine With Dual Injection, Proc. 5th KHES-HESS Joint

Symposium, Taejon, Nov., Korea Hydrogen Energy Society, pp. 237248.

14 Kwon, B. J., Lee, J. Y., Lee, J. T., and Lee, S. Y., 1993, The Effect ofCompression Ratio on Combustion and Performance Characteristics of Direct

Injection Spark Ignition Hydrogen Fueled Engine, Trans. KSAE, 1, pp. 17

26.

15 Lee, J. T., Lee, S. Y., Kim, Y. H., and Lee, J. C., 1988, A Basic Study on theDevelopment of Hydrogen Fueled Engine, Proc. KSAE Spring Conference,

Korea Society of Automotive Engineering, pp. 5963.

16 Nam, S. W., Park, J. B., Choi, K. H., and Lee, J. T., 1995, Cooling Losses andHeat Flux of Hydrogen Fueled Spark Ignition Engine With Inner Injection,Proc. of 3rd KHES-HESS Joint Symposium, Sep., pp. 8391.

17 Hong, H., Lee, J. T., and Lee, S. Y., 1992, As Experimental Study on theSuitable Configuration of Injection Hole in Direct Injection Hydrogen Fueled

Engine, KSME-JSME 2nd Thermal Engineering Conference, Vol. 2, Oct., pp.

110.

18 Kondo, T., Hiruma, M., and Furuhama, S., 1996, A Study on the Mechanismof Backfire in External Mixture Formation Hydrogen Engines, Proc. WHEC,Vol. III, Stuttgart, June, International Association for Hydrogen Energy, pp.

15471556.

19 Kim, Y. Y., Ryu, T. H., and Lee, J. T., 1998, Backfire Occurrence by Abnor-mal Electric Discharge in Hydrogen Fueled Engine, Proc. of KHES Autumn

Annual Meeting, Jinju, May, Korea Hydrogen Energy Society, pp. 105115.20 Eichlseder, H., Wallner, T., Freymann, R., and Ringler, J., 2003, The Poten-

tial of Hydrogen Internal Combustion Engines in a Future Mobility Scenario,



Documents

cvips_9_11