8/19/2019 Jsae Paper 2007

1/8

JSAE 20077XXXSAE 2007-01-YYYY

Test of blends of hydrogen and natural gas in a lightduty vehicle

Giovanni Pede, Ennio Rossi

Italian National Agency for New Technologies, Energy and the Environment (ENEA)

Department of Energy Technologies

Maria ChiesaCatholic University of Brescia (Italy)

Environmental Physics Department

Fernando OrtenziDITS – Dipartimento di Idraulica Trasporti e Strade,

University “La Sapienza”,Rome (Italy) 

Copyright © 2007 Society of Automotive Engineers of Japan, Inc. and Copyright © 2007 SAE International

ABSTRACT

Hydrogen-enriched combustion has been studiedby several institutions and companies over the lastthree decades. The purpose of adding hydrogen toconventional fuels is to extend the lean limit of

combustion because hydrogen improves flamestability and allows a lower temperaturecombustion. Even with stoichiometric mixture,HCNG advantages had been demonstrated, sinceblends determine a reduction of noxious emissions.In the framework of an EU project called BONG-HY,bench tests with HCNG on a natural gas vehiclehad been carried on at ENEA.

Results of lab tests show a fair improvement of theefficiency and CO2 emissions as well as an overallimprovement regarding local pollutants.

INTRODUCTION

The growing sector of transports rises a big alarmeither for the day-by-day increasing number ofvehicles and for the sensible contribution to thedegradation of air quality in urban areas, as well asfor the global pollution.In Italy, with beyond 35 million of circulating vehicles,the consumption of primary energy, all coming fromfossil sources, accounts for more than 30%, whichroughly leads to a corresponding 30% increase inCO2 emissions. The European Union committed tothe goal of reducing its dependence on importedfossil fuels (oil, natural gas, coal), by using at least

20% of alternative fuels within the year 2020; thecorresponding commitment in the reduction ofGreenhouse Gases (GHG) is the well-known 8%

with respect to 1990 by 2012, as required by theKyoto Protocol.In Europe the sector of transports is responsible forthe 25% of CO2 emissions, 40% of which is relatedto the vehicles circulating in urban areas. Externalcosts due to the degradation of air quality related to

transports had been estimated in about 11.7% ofEU GDP, corresponding to an outstanding value of360 €/year per citizen.Those alarming data have to be added to thecontribution to the total emissions from theenergetic sector (carbon dioxide, natural gas,nitrogen oxides, sulphur, aromatic compounds,….),which amounts at about 50% of the totalcontribution. Deaths caused by the smog, due toparticulates and other emissions, are about 8000per year just for Italy; on the other side, the globalchange becomes a “real” problem, with anincreasing concern about GHG emissions.

Nowadays a last-generation Euro4 car emits slightlyless than 150 grCO2/km, with scarce perspectivesto be able to reduce, with fossil fuels, that valuevery much.It had been worldwide agreed that the introductionof hydrogen as a “new” fuel could have contributedto the realization of a sustainable energy system inthe long term (2050 and beyond); according to thisvision, emissions of both global and local pollutantscan be maintained under “safe” values.Even if the transition towards a hydrogen-basedeconomy will be surely very long, its sustainability isachievable since now, also considering thelimitations in the substitution of conventional fuelswith alternative ones, less polluting. Also thecontribution of the introduction of biomass-derived

8/19/2019 Jsae Paper 2007

2/8

fuels, for a limited quota of total consumption, iscounterbalanced by the still growing demand ofvehicles in the world ( see Fig. 1 at the end) [1].Even if it’s difficult to forecast the future concerningthe next decades, it has been agreed worldwide thatclimate change is closely connected with GHGemissions, so we may ask for some importantdecisions for the beyond-Kyoto years. Thestabilization of CO2  concentration at values nothigher than 550 ppm (today’s value is 380 ppm)

requires a strong emissions reduction: some of theIPCC scenarios aiming at that values shows arequired decrease of GHG of 40-60% with respectto 1990, which means a “real” reduction of 70-90%of the emissions with respect to the “business-as-usual” forecast.Such a reduction won’t ever be achieved by usingany actual available sustainable technology.Nevertheless, a “cultural shift” will be necessary, inorder to reach that goal: the introduction ofhydrogen as an energy carrier seems to be a realcontribution to that goal, making possible, in thelong term, the realization of a cleaner World.

METHANE-HYDROGEN MIXTURES

 A good opportunity in the short term can berepresented by the utilization of blends of hydrogenwith other fuels, first of all with natural gas (HCNG).When used in an Internal Combustion Engine (ICE),even the addition of a small amount of hydrogen tonatural gas (5-30% by volume, that means ~1.5-10% by energy) leads to many advantages,because of some particular physical and chemicalproperties of the two fuels.

HCNG15

CH4

lambda=1

HCNG15

CH4

lambda=1

 Fig.2

Methane has a slow flame speed while hydrogenhas a flame speed about eight times higher; whenthe air/fuel ratio (lambda) is much higher than forthe stoichiometric condition the combustion ofmethane is not as stable as with HCNG. As aconsequence of the addition of hydrogen to naturalgas an overall better combustion had been verified,even in a wide range of operating conditions(lambda, compression ratio, etc.), finding thefollowing main benefits:

• a higher efficiency• lower emissions

Because of the characteristics of hydrogen, HCNG,despite its higher LHV per kg, has a lower LHV perNm

3, depending on the hydrogen content. Therefore,

a natural gas engine, when fuelled with HCNG,shows a lower power output, while maintaining itsbetter efficiency.

0

10

20

30

40

50

60

0 10 20 30 40% volume H2

   L   H   V

LHV (MJ/kg) LHV (MJ/m3)

 

Fig.3

In case of turbocharged engines, power output can

be increased again by a simple increase of thecharging pressure, possible even because of thehigher reluctance to detonation of hydrogen. Additionally, CO2 emissions had been reduced notonly as a result of the substitution of CNG byhydrogen. The special properties of hydrogen as acombustion stimulant can produce leverage factorsmuch greater than 1 by improving fossil fuels--not just displacing them. Hydrogen leverage is definedas the following ratio : (% Emissions Reduction)/(%Energy Supplied as Hydrogen).The increased efficiency makes this value higherthan 1. An obvious benefit of the leverage effect is

that a CO2  reduction is possible even if neededhydrogen is produced by natural gas without any“sequestration” of CO2.

STATE OF THE ART

First experiences of methane-hydrogen mixture,with vehicles, were carried on in the framework of aprogramme financed by DOE and NREL, inColorado, the "Denver Hythane Project”, from 1991to 1993, whose results are shown in the tablebelow:

Table 1 Denver Hythane Project

NMHC(g/mile)

CO(g/mile)

NOx(g/mile)

Gasoline 0.59 14.1 2.2

ULEV 0.04 1.7 0.2

Natural gas 0.01 2.96 0.9

Hythane 0.01 0.7 0.2

In a next phase, according to an exhaustive reviewby R. Sierens and E. Rosseel, by University of Gent,Belgium that relates to laboratory testings in thesecond half of 90’, many other experiences arerecorded:

8/19/2019 Jsae Paper 2007

3/8

« Hoekstra et al. (1994, 1995) examined a V8 Chevrolet350 engine at one particular speed (12.7 kW, 1700 rpm)with different hydrogen enriched compressed natural gasmixtures, to simulate a light-duty truck……. They found

extremely low NOx values at  λ  =1.6 ( φ  = 0.625) for the 28and 36 percent H2 blendsSwain et al. (1993) and Yusuf et al. (1997) made testswith a 20 percent hydrogen–80 percent natural gas blendon two engines (2L Nissan and 1.6L Toyota) under lightload conditions...For blended fuel, a 10 to 14 percentimprovement in the brake thermal efficiencies over

methane was found.Larsen and Wallace (1997) and Cattelan and Wallace(1994) tested a turbocharged 3.1L V6 engine under midand high load conditions with a 15 percent H2 hythaneblend and found similar trends (for the exhaustconcentrations in ppm) as the light load tests by Swain etal. (1993) and Yusuf et al. (1997). Raman et al. (1994)described lean burn and stoichiometric combustion testswith a three-way catalyst. For the lean burn (5.7L GM V8)engine it was again shown that hydrogen extends the leanlimit of natural gas, thereby enabling lower NOx emissionswithout excessive THC. When the BMEP advantage ofhythane is sacrified by retarding the spark advance untilmethane and hythane produce equal BMEP, the NOx

concentrations drop significantly. Bell and Gupta (1997)described tests with lean mixtures of natural gas blendedwith 5, 10, and 15 percent hydrogen on a 4 cylinder 2.5LGM engine at 2200 rpm and 50 percent WOT….Again thesubject of the research was to extend the lean operatinglimit of the engine and to investigate the performance andemissions characteristics of the SI engine at these

conditions. At the natural gas lean operating limit λ  =1.56

( φ =  0.64) hydrogen addition allowed an increase in power(up to 47 percent again with 15 percent H2) due to anincrease in the average flame speed maintaining asufficient heat release rate for good combustionquality……Brake thermal efficiencies (15 percent H2)were higher than for the other fuelling cases at

corresponding equivalence ratios..» During the last years, also a number of fleet testingswere carried on. The recent Hythane® ((24.8% vol.Hydrogen, Frank Lynch, Hydrogen Components,Inc., HCI), bus demonstration project at Sunlinetransit in California used a 7% hydrogen by energyformula and the NOx emissions were reduced by50%. Based on success with Hythane® buses, andthe cost-effectiveness of Hythane® compared toavailable fuel cell technology, a number of projectsare currently carried on all around the world, like theBeijing Hythane Bus Projet, whose demonstrationphase will be to adapt 30 natural gas engines for

Hythane operation. In Sweden, operation withHythane® and natural gas had been compared fora heavy-duty natural gas engine and the study hadrevealed a small increase in efficiency.Subsequently, a couple of buses had been testedon the road with blend with a 8% hydrogen content(by volume). Tests during full load and constantload demonstrated a 20-30% reduction in HC-emissions and higher power with mixture.Transients bring to 50% less emissions both for HCand CO but a 50% increase of NOx. From theenergy point of view, there is a 14% fuel reduction.

EXPERIMENTAL SET UP

In the framework of an EU Interreg IIIC projectcalled BONG-HY (parallel application of Blends OfNatural Gas and Hydrogen in internal combustionengines and fuel cells) bench tests on a natural gas

vehicle had been carried on in one lab of theCasaccia Reasearch Center of ENEA. The mainItalian partners involved in the project had been theMunicipality of Brescia (lead partner), ASM SPA(the energy multiutility of Brescia), the CatholicUniversity of Brescia , the Universities of Rome “TorVergata” and “La Sapienza” and ENEA.The light duty commercial vehicle under test hadbeen a Daily, belonging to the ASM fleet, that hadbeen mainly modified in the control system (ECU)

for the test. Its engine was a 2.8 L NG fuelled,manufactured by IVECO (see Tab.2 forspecifications).

Tab.2 Engine specifications

Displaced volume 2800 cm3 

Compression ratio 12.2

Bore 94.4 mm

Stroke 100 mm

Rated power 78 kW @ 3800 Rpm

Max Torque 220 Nm @ 2200 Rpm

Emission Standard Euro III

The roller bench comes from APICOM; its controlsystem has been recently upgraded by Assing.The cycle adopted for the characterisation hadbeen the urban part of NEDC (repeated at least 20times) and the value for the vehicle mass had beenset to 3500 kg; therefore the test had been moresevere than the homologation one. Therefore, theresults are not directly comparable with OEM data,but surely more significative for the guide cycle ofthe ASM vehicles that could use these blends in thefuture (waste collecting vehicles).

Fig.4

Engine ECU Tuning tool

The tool used to modify the engine maps is calledRACE2000, elaborated by Dimensione Sport. Thissoftware is able to transform data contained in theoriginal eprom in easily modificable electronicmapping. Along with the software cited above, theMET16 simulator, able to modify the spark advance,

the injection time and other parametersinstantaneously, had been used too.The programming of the final EPROM had beenobtained thanks to the EMP21 EPROM programmerproduced by Needham’s Electronics.

8/19/2019 Jsae Paper 2007

4/8

 

MEASUREMENT SYSTEM

Cylinder pressure

 A single cylinder head had been equipped with apiezo-electric pressure transducer hosted on aspecial spark plug and signals had been processedby a twochannel amplyfier while the angular position

had been measured with a inductive crank-anglecalculator module  for on-line indicatingmeasurements. The equipment had beenmanufactured by AVL, all data had been collectedby a Yokogawa DL 716 digital scope. In the picturebelow, an example (1500 r.p.m., 50 % at mediumloads) of preliminary testing with pure methane isgiven, for the engine model validation.

Fig.5

Emissions

Emissions had been measured using a HORIBAOBS-1300 integrated system. This is composed bya MEXA-1170HNDIR (Dispersive Infra-redDetectors) that measures in real time CO, CO2 andHC and by a MESA-720NOx (ZrO2 technology), forthe evaluation of nitric oxides concentrations andair-fuel-ratio (AFR); furthermore, a heated flowmeter(pitot type) mounted on the sampling probe permitsto calculate the mass ( see Fig. 6).

Fuel Consumption

Two blends had been tested, characterised by 10and 15% by volume in hydrogen (HCNG10 andHCNG15) and used as a fuel for the “urban part” ofthe ECE-15 driving cycle. Pure methane of certifiedcomposition and certified mixtures had been usedfor the tests. For characterization tests, singlecylinders had been used. To assure the requestedprecision with regards to energy consumption

measures, they were weighed before and after drivetests (ECE 15) on our roller bench that lasted asignificative time.

Fig.7

EXPERIMENTS

The main parameters that had been investigatedare lambda (with values of 1 and 1.4), differentspark advance angles and different values for theenrichment of the blends during transients. As main exhaust parameter which had beenconsidered as a constraint that had not to beovercome in case of stoichiometric set up is NOxemission. Actually, hydrogen addition implies ahigher laminar combustion speed and this causesan increase of combustion temperature andtherefore higher NOx emissions . On the contrary,CO and HC emissions are always lower , thanksboth to the lower quantity of carbon and to theimproved combustion process.For lean-burn mixtures, not only NOx , but also HCmonitoring had been a decisive parameter that hadbeen taken into consideration. Actually, also HCemissions can raise due to the fact that laminarcombustion speed decreases remarkably. Thisproduces a not complete oxidation of HC. Moreoverhigher gas cooling during expansion delays the fueloxidation, in particularly near cylinder’s walls and inthe most hidden parts of the combustion chamber.The first tests using mixtures to feed the engine,carried without any modification of the injectioncontrol map, had confirmed the foreseen NOxemissions increase related to the increasingcombustion speed. This fact had required amodification of the spark ignition time. In Fig. 8 it ispossible to see that a spark advance reduction ofonly 3 degrees (which means a little retardcompared to the case of pure methane) brings to alarge decrease of NOx emissions, without torquereduction.

8/19/2019 Jsae Paper 2007

5/8

0

100

200

300

400

500

600

700

800

900

-8 -6 -4 -2 0 2 4 6

 Advance variation degree

   N   O  x  p  p  m

500

520

540

560

580

600

620

640

660

680

700

   T  o  r  q  u  e   N  m

NOx

Torque

Fuel and CO2 reduction, %

0,00

5,00

10,00

15,00

20,00

HCNG10λ=1 HCNG10λ=1,4 HCNG15λ=1 HCNG15λ=1,4

   %

Fuel reduction CO2 reduction

Energy saving, %

0,0

2,0

4,0

6,0

8,0

10,0

12,0

HCNG10 λ=1 HCNG10 λ=1,4 HCNG15 λ=1 HCNG15 λ=1,4

    %

 Fig.8

Notwithstanding the above reported spark ignitiontime correction, engine performances with methane-hydrogen blends had remained not acceptableduring ECE driving cycle in terms of emissions, dueto the too high NOx emission values, compared tomethane. A more detailed examination of the

engine behaviour during transients shows that fuelenrichment (as mapped in ECU) had been too low,

therefore actual λ  reaches values comprisedbetween 1.1-1.2. As a result, NOx emissions hadincreased too much. For this reason, a mapcorrection has been adopted for accelerationphases. This had been allowed by a dedicatedfunction of electronic injection unit. In this way it hadbeen possible to decrease NOX emissions todesired values, lower than figures obtained withmethane ( see Fig.9).For a lean burn blend, research of better AFR (wewanted to reach a value of 100 ppm, the same ofpure methane) was limited from λ= 1 to λ= 1,45.Furthermore, the increase of AFR values causesvery important power losses, as shown in Fig. 10.Therefore, λ=1,45 had been the maximum valueinitially fixed (this value substantially reduces NOx)and a series of tests had been produced changingthe spark ignition advance to optimize the controlstrategy in order to increase the performances.Unfortunately, NOx had grown in an exponentialway, while power gain hadn’t been significative. Ithad also been decided that it is more convenient toadopt a lower AFR value without changing the sparkadvance instead of setting a high λ  together withoptimal advance timing.

0

500

1000

1500

2000

0.95 1.05 1.15 1.25 1.35 1.45

Lambda

   N   O  x  p  p  m

0

100

200

300

400

500

600

700

   T  o  r  q  u  e   N  m

Nox

Torque

 

Fig.10

RESULTS

In the following pictures the obtained values of thefirst 6 months of lab tests for both the fuelconsumption and the pollutants and CO2 emissions(for different operating conditions) are represented .Figures 11 and 12 shows fuel consumption andCO2  emissions for different hydrogen contents inHCNG while Fig. 13 represents the local emissionsfor the same blends.

Fig. 11 

Fig. 12

CONCLUSIONS

The analysis of the results of the lab tests can beconsidered encouraging for the continuation of theactivities, even if the tests had just covered a shortperiod of time (a few months).It’s well known that the reduction of fuelconsumption brings to the enhancement of thepollutants and greenhouses emissions; therefore,

the optimum condition can be found dealing with theproblem with different approaches, i.e. with the useof lean blends in the first case ( for the fuelconsumption reduction) and with the use ofstoichiometric blends in the second case( emissions reduction). Actually, these two approaches correspond to thedifferent conceptual ideas adopted by VOLVO(whose engines are mounted on the urban buses ofthe “Malmo Hythane project”) and by IVECO (themanufacturer of the DAILY vehicle used for ourexperimentations) for the realisation of their naturalgas motorisation. In our case, dealing with an

IVECO engine, that had been designed in order towork in stoichiometric conditions, the adoption of a“lean combustion” strategy had brought us tounsatisfactory results: actually, since nochangements in the engine hardware had beenmade, as for example the compression ratio

8/19/2019 Jsae Paper 2007

6/8

enhancement and/or the engine overfuelling, withrespect to the vehicle’s behaviour, a reduction ofthe engine specific power had been verified, beyondto the reduction of the power due to the inferiorenergy content in volume ( - 11% for the blend witha 15% hydrogen content by volume).Though, for both the combustion strategies adopted,the vehicle had succeeded in doing the urban cycle,the engine had worked regularly and the obtainedresults had been encouraging in terms of both

consumptions reduction and pollutants and CO2 emission reduction as demonstrated even by theforeign lab and road tests.The pollutants emission reduction had been verypromising mainly working in stoichiometricconditions.Finally, even considering the Swedish works ( thatdon’t present problems of power losses with the useof blends) we can say that, starting from the existentmotorisations, the strategies that have to beadopted for the modifications must be coherent withthe base constructor philosophy; in our case, wehad taken into consideration the IVECO philosophy.

Consequently, we consider the obtained results withstoichiometric conditions as the base results of apossible development of the project that foreseesthe field experimentation of vehicles with an IVECOmotorisation (for example, industrial vehicles for thetransports of goods or waste collecting vehicles).In this framework, the reduction of the energeticconsumptions increases with blends with a 15%hydrogen content by volume, in a way more thanproportional with the increase of hydrogen content,while for the pollutants emissions a significativedifference between the use of blends with a 10% or15% hydrogen content by volume hadn’t been

verified.

REFERENCES

[1] The Sustainable Mobility Project, Mobility2030: Meeting the challenges tosustainability, World Business Council forSustainable Development, Switzerland,2005

[2] Riddell, B., Malmo Hydrogen andCNG/Hydrogen filling station and Hythane

bus project, Carl Bro Energiekonsult AB,Sweden, 2004[3] Tunestal P., Einewall P., Stenlaas O.,

Johansson B., Possible short termintroduction of hydrogen as vehiclefuel/fuel additive, P. Duret and EditionsTechnip, Paris, 2004,

[4] Karner D., Francfort J., Freedom car &vehicle technologies program – Advancedvehicle testing activity- Arizona publicservice- Alternative fuel (Hydrogen) pilot plant, US DOE, 2003 

[5] F. Linch, Hythane: a bridge to an ultraclean

renewable hydrogen energy system, Attidel Workshop IEA in Denver, 1991 

[6] James Cannon, Paving the way to NaturalGas Vehicles, INFORM, Inc., 1993

[7] R. Sierens, E. Rosseel, VariableComposition Hydrogen/Natural Gas

mixtures for increased engine efficiencyand decreased emissions, Journal ofEngineering for Gas turbines and Power,2000,

[8] Hoekstra R. L., Collier, K and Mulligan N.,Demonstration of Hydrogen Mixed GasVehicles, Proceedings, 10 

th  World

Hydrogen Energy Conference, CocoaBeach, Vol. 3, anno 1994 

[9] Hoekstra R. L., Van Blarigan P., and

Mulligan N., NOx emissions andefficiency of Hydrogen, Natural Gas andHydrogen/Natural Gas blended fuels,SAE Paper 961103, 1996 

[10] Swain M. R., Yusuf M., Dulger Z., andSwain M. N., The effects of hydrogenaddition on natural gas engine operation,SAE Paper 932775, 2003

[11] Larsen J. F., and Wallace J. S.,Comparison of emissions and efficiencyof a turbocharged lean-burn natural gasand hythane-fuelled engine, ASMEJournal of Engineering for gas turbines

and power, Vol. 119, 1997 [12] Yusuf M., Swain M. R., Swain M. N., and

Dulge Z.,  An approach to lean burnnatural gas fuelled engine throughhydrogen addition, Proceedings, 30 

th 

ISATA Conference, Florence, paper97EL081, 1997

[13] Raman V., Hansel J., Fulton J., Lynch F.,and Bruderly D., Hythane – an ultracleantransportation fuel, Proceedings, 10 

th 

World Hydrogen Energy Conference,Cocoa Beach Vol. 3 , pp. 1797-1806,1994

[14] Bell S. R., and Gupta M., Extension of thelean operating limit for natural gas fuellingof a spark ignited engine using hydrogenblending, Combustion Science andTechnology, Vol. 123, pp. 23-48, 1997

8/19/2019 Jsae Paper 2007

7/8

 

Fig. 1

Fig. 6

8/19/2019 Jsae Paper 2007

8/8

0

200

400

600

800

1000

1200

0 10 20 30 40 50 60 70 80 90

Time s.

   N   O  x  p  p

  m

-10

0

10

20

30

40

50

60

   V  e   h   i  c   l  e   S  p  e  e

   d   K  m   /   h

NOx After 

NOx before

Speed After 

Speed Before

 

Fig. 9

Emissions, g/km

0,00

0,20

0,40

0,60

0,80

1,00

1,20

NG HCNG10 λ=1 HCNG10

λ=1,4

HCNG15 λ=1 HCNG15

λ=1,4

  g   /   k

  m

CO HC NOX

 

Fig. 13