Tar reduction in downdraft biomass gasier using a primary
methodEinara Blanco Machin a, Daniel Travieso Pedroso a, *, Nestor
Proenza b, Jos
e Luz Silveira a,Leonetto Conti c, Lcia Bollini Braga a, Adrian
Blanco Machin aaEnergy Department, S~ ao Paulo State University
(UNESP), Guaratinguet a, SP, BrazilbMechanical Engineering
Department, University of Camagey, CubacDepartment of Chemistry,
University of Sassari, Sassari, Italyarti cle i nfoArticle
history:Received 19 June 2014Accepted 30 December 2014Available
online 4 February 2015Keywords:BiomassDowndraft
gasierGasicationTarSwirlowabstractThis work present a novel primary
method, for tar reduction in downdraft gasication. The principle
ofthisnewtechnologyistochangethe uiddynamicbehaviourof themixture,
formedbypyrolysisproduct and gasication agent in combustion zone;
allowing a homogeneous temperature distribution inradial
directioninthisreactionzone. Toachievethechangeinthe
uiddynamicbehaviourof themixture; the entry of gasication agent to
combustion zone is oriented by means of wall nozzles in orderto
formaswirl ow. Thismodicationincombination
withtheextensionofthereductionzone, willallow, to increases the
efciency of the tar thermal cracking inside the gasier and the
extension of theBoudouard reactions. Consequently,thequantity
oftarpassingthroughthe combustionzone withoutcracking and the
concentration of tar in thenal gas, decrease signicantly in
relation with the commonvalue obtained for this type of reactor,
without affecting signicantly the heating value of the producergas.
Inthiswork ispresentedanewdesignfor15kWdowndraftgasicationreactor,
withthistech-nology implemented, the tar content obtained in the
experiments never overcome 10 mg/Nm3, with alower heating value of
3.97 MJ/Nm3. 2015 Elsevier Ltd. All rights reserved.1.
IntroductionBiomass, mainly in the form of wood, is the oldest form
of en-ergyusedbyhumans. Biomassgenerallymeansarelativelydrysolidof
natural matter that
hasbeenspecicallygrownorhasoriginatedaswasteorresiduefromhandlingsuchmaterials[1].The
thermochemical conversion of biomass (pyrolysis,
gasication,combustion)isoneofthemostpromisingnon-nuclearformsoffutureenergy.
Biomassisarenewablesourceofenergyandhasmany ecological advantages
[2]. Gasication is the key technologyof biomass based power
generation; is a high-temperature process(873e1273K) that
decomposes complexbiomass hydrocarbonsinto gaseous molecules,
primarily hydrogen, carbon monoxide, andcarbon dioxide; also are
formed some tars, char, methane, water,andotherconstituents.
Several institutionsworkingonbiomassgasication have given many
denitions of tar. In the EU/IEA/US-DOE meeting on tar measurement
protocol held in Brussels in theyear 1998, it was agreed by a
number of experts to dene tar as allorganic contaminants
[polycyclic aromatic hydrocarbon (PAH)]with a molecular weight
higher than benzene [3]. Tar is undesirablebecause of various
problems associatedwithits condensation,causing problems in the
gasication installations as well as in
theequipmentsthatusetheproducergasasfuel likeinternalcom-bustion
engines and gas turbines. The required gas quality to fuelinternal
combustionengines is normallyreachedeasilyinthemodern downdraft
gasiers, except for the content of dust and tar.Thermal, catalytic
or physical processes either within the
gasica-tionprocess(primarymethods)
oraftertheprocess(secondarymethods) can be applied to remove tars.
Primary methods have theadvantage that dispenses the use of an
expensive cleaning systemfor producer gas. In addition, cracking of
tars in the reactor couldincreases the amount of combustible gases
in the producer gas andtherefore, the overall process efciency.
There are some sophisti-cated options available, which claimed a
signicantly reduction ofthe tar content in the producer gas,
however, the method must beefcient in terms of tar removal,
economically feasible, but moreimportantly, it should not affect
the formation of useful producergas components [4].*Corresponding
author.E-mail addresses: [email protected] (E.B. Machin),
[email protected],[email protected] (D.T. Pedroso).Contents
lists available at ScienceDirectRenewable Energyj ournal homepage:
www. el sevi er. com/ l ocat e/
renenehttp://dx.doi.org/10.1016/j.renene.2014.12.0690960-1481/ 2015
Elsevier Ltd. All rights reserved.Renewable Energy 78 (2015)
478e483The catalytic cracking and electrostaticlters are two
examplesof the options, that claim a signicant tar reduction in the
producergas, but they increase the cost of the plants, especially
in the smallones. Currently,
thepreferredoptionfortarreductionisinthegasier itself
throughprocess control andthe use of primarymeasures such as
additives and catalysts which modify gasicationconditions [4e12].
Theoretically, producer gas with low tar contentcan be obtained if
a high-temperature zone can be created, wherethe gaseous products
of pyrolysis are forced to reside the necessarytime to undergo a
secondary gasication. Previous works have beendeveloped in order to
design a downdraft gasier, able to increasethe efciency of tar
reduction in the producer gas during gasica-tion process. Bui et
al. [13] developed a multi-stage reactor designthat separates
theaming-pyrolysis zone from the reduction zone.Inthat design,
thetar vapours generatedinthe rst zoneareburned or cracked to
simple molecules by high temperature in thesecond zone, improving
the gas quality and conversion efciency.The minimum content of
gravimetric tar obtained with this designwas 92 mg/Nm3. Susanto and
Beenackers [14] developed a down-draft moving bed gasier with
internal recycle and separate com-bustion of pyrolysis gas with the
aim of reduce a tar content in theproducer gas; in their
experiments a minimumof 48 mg/Nm3of tarwas obtained.On this
background, the main objective of this work is to pro-pose a
newdowndraft gasier design, able to generate the producergas
withlowtar concentrationusinganovel primarymethodwithout decreasing
signicantly the heating value of the producergas.2. Process
principleIn the Imbert design of downdraft gasier, the gasication
agentis fed above a constriction (throat) by nozzles uniformly
distributedon the wall of the combustion chamber, oriented toward
the centreof the circle, that describe the perimeter of the
combustionchamber.In this design, some cool zones are created near
to thenozzles, where the temperature is not sufciently higher to
permitthe thermal cracking of the tar present in the mixture and to
un-dergo its secondary gasication [15]. This is one of the reasons
forthe presence of tar in the producer gas. If tarry gas is
produced fromthis type of gasier, is common practice reduce the
centralconstriction area, until a gas with low tar content can be
produced.However, this area dimensions also play an important role
in thegas production rate.In order to avoid the formation of cool
zones, it is proposed
inthisworktomodifytheuiddynamicbehaviourofthemixtureformedbythepyrolysisgasesandthegasicationagent
inthecombustion chamber.2.1. The combustion chamberSwirlows are
widely used to intensify the process of heat andmass transfer
between solid particles and airow in vortex cham-bers,
theadvantagesof swirl
owshasbeendeeplystudiedbyseveralauthors[16e20]. Theswirl
owofthemixturecouldbecreated changingtheentry
angleofthegasicationagentto thecombustionchamber. Thenewanglemust
be different of thestandard 90
in the Imbert design. This modication allow that
thecirculationG(Equation(1))of thevelocityvectorV(ro,t)of
anyelement of the uid at any position r s0 in the plane in which
thenozzles are located, or any other parallel plane below this
until thediaphragm, is different from zero (G s0).G ILVr0; tdl
(1)The circulation of the vector V(ro, t) combined with thedownward
movement of theuid, caused by absorption from thebase of the
chamber through the diaphragm, generates a swirl ow.This uid
dynamic behaviour would allow to increase the mixing ofthe
gasifying agent with the pyrolysis gases [21,22]; homogenizingthe
temperature inside the combustion chamber, diminishing
theformationof cool areasbetweenthenozzlesasmainresult.
Inadditionthis modicationincrease
theresidencetimeofthegasinsidethecombustionchamber;therebyincreasingthethermalcrackingof
thetar inthis zone, minimizingits passagetothereduction
zone,decreasing the tar concentration in the producergas. Swirl
numberSmayeffectivelycontrol theresidencetimedistribution of the
gas mixture, which is function of theuid entryangle [18]. The
increase of the residence time has the undesirableeffects of
decreasing the efciency and productivity of the gasier,as described
by Susanto [13]. Fig. 1 shows a top view of the com-bustion chamber
of the reactor,illustrating the inclination oftheinlet nozzles of
gasication agent.3. Experimental approach3.1. Investigated
samplesThe gasicationtests wereperformedusingthree
differentwoodybiomasses, suppliedbyawoodprocessingfactory.
Thebiomasses used were Peach (Prunus persica), Olive (Olea
europaea)andPine(Pinuspinea). Thepropertiesofthe
woodybiomassareshowninTable1. Theelemental
compositionsweredeterminedusingaCHNS-OElementar VarioGmbHEL III
andtheHigherHeatingValue(HHV)usingacalorimeterIKAC-5000(ASTMD-3286-91a).
Themoistureandashcompositionweredeterminedusing the ASTM E-871-82
and ASTM D-3174-82. The results weresimilartoliteraturevalues.
Fortheexperiments,
thebiomasseswerechoppedinsquare-basedprismpieceswithdimensionsofabout
2 1 1 cm. The size and shape are very important for thebehaviour of
biomass in the downdraft gasier as far as its move-ment,
andbridgingandchannellingformations. Inaddition, theheight of
theoxidationzoneandthepressuredropinsidethereactor, depend on these
characteristics.3.2. Experimental setupThe scheme of the downdraft
wood gasier is showin Fig. 2. Thegasier unit is constitutedof
twocylindrical coaxial structuresconstructed using a mild steel
sheet. An insulating material coatstheexternal one,
whiletheinternal cylinder is providedwithadditional heat
recuperation surfaces to improve the efciency ofFig. 1. Nozzles
inclination in the combustion chamber.E.B. Machin et al. /
Renewable Energy 78 (2015) 478e483 479the gasication process (Fig.
2). The internal capacity is 0.452 m3,the height of the gasier is
1.02 m and the internal radius at thedryingepyrolysiszoneis0.30m.
Thedimensionsofreductionzoneareenlargedtoboost therateof
theBoudouardandthewateregas reactions, in order to increase the
concentration of COand H2 in the producer gas and also decrease the
gas temperature.The gasication agent for the experiments (air) is
supplied using anelectric blower with control valve, capable of
supply the requiredair for the gasication process.The lines are
heated up to 453 K in order to prevent conden-sation of the
producer gas compounds inside the conducts and themeasurement
device.The producer gas sample isltered, cooledand drained, before
be analysed in the Gasboard-3100P mobile gasanalyser. The
temperatureare measuredby mean ofsix thermo-couples (type K)
located at different height of the reactor bed. Airand gas ows are
measured with an orice and differentialmanometer. All the
experimental data is recorded by data logger in5 min intervals. The
simplied experimental setup for the test ofthe modied reactor is
presented in Fig. 3.3.3. Tar sampling principleThe principle ofthe
test method for gravimetric tar measure-ment is based on the
continuous sampling of a gas stream,containing particles and
organic compounds (tar) under isokineticconditions; according to
the methodology described in DD CEN/TS15439:2006
[23].Thedeterminationiscarriedoutintwosteps:samplingandanalysis.
The equipment for sampling shown in Fig. 4, consists of
aheatedprobe(module1), aheatedparticle lter(module2),
acondenser,aseries of impinger bottles containing
asolvent(iso-propanol) for tar absorption (module 3), and equipment
for pres-sure and owrate adjustment andmeasurement (module
4).Upstream ofthe condenser,the tubes connecting these
partsareheated in order to prevent tar condensation. Temperatures
of thecondenser
andtheimpingerswereproperlyselectedtoensurequantitative collection
of the tars (1, 2, and 4 is between 308 and313 K, and 3, 5 and 6 is
between 258 and 253 K). Tar collectionoccurs both by condensation
and by absorption in the condenser, inthe impinges, and by
capturing of aerosols in glass frits. The analysisof the samples is
carried out according to the methodologydescribed in Ref. [23].3.4.
Processow descriptionThe gasier system was run nine times, for
periods between 2.5and 4 h. To start the gasier, initially the fuel
biomass is loaded uptothereactormaximumcapacityandisclosed.
Subsequentlyisintroduced a propane gas duct by the air entrance to
the reactor, tocreate aameinsidethecombustion chamber,
thenthevacuumpump was turned on and the propane gas feed is
removed. In lessthan 15 min or when the temperature in pyrolysis
zone (TC 2 andTC 3) reaches 573 K the ignition step is completed
and the record ofthe prole of reactor temperatures and the gasesow
starts. Theproducer gas analysis starts when the preset temperature
prole inthe reactor is reached, due to the high concentration of
condensablegases in the producer gas composition during the
ignition process.The tar sampling process starts at the same time
of the producergas analysis, with the installation shown in Fig. 4;
each tar samplingtakes 45 min.4. Results and
discussionTable2andTable3showntheperformanceof thebiomassgasier
system and the composition of the producer gas during
theexperiments, at regular intervals of 5 min.Fig. 5 shows a
typical behaviour of the temperature prole in
thereactorduringtheexperiments. Asit isobserved,
thereareanoscillation of the temperature value in all the bed
section during alltheexperiments, withtheexceptionof
thetemperatureof theproducer gas, where the temperature remain more
stable. The mainreasonof
thisvariationisbiomassmovementinsidethereactorduringthegasication
process. Thetemperature oftheproducergas remains in the range of
410e430 K, lower than the typical rangeof 700e720 K reported for
this type of reactor.The HHV of the producer gas is calculated from
the concentra-tion of the combustible components. For all the
experiments, theHHV obtained was higher to 3.50 MJ/Nm3, and the
higher valueswere obtained in the experiments using Peach as fuel,
where themean value was 3.97 MJ/Nm3.Thesevaluesare lower thanthe
theoreticaland experimentalresults reported in the literature;
Zainal et al. [14] report 4.72 and4.85
MJ/Nm3respectivelyforsamecapacityandtypedowndraftgasier.These
results are because the mediumcontent of H2, COand CH4in the
producergas obtained in the experiments with the testedreactor was
slightlylower thanthetypical compositionof theproducer gas reported
by several authors [2,3,13,14,24,25]. The O2concentration has the
same behaviour, showing an increase in theTable 1Elemental
composition and HHV of the studied biomasses.Biomass C%wt dbH%wt
dbN%wt dbO%wt dbAsh%wt dbMoisture%wtHHVMJ/kgPeach 48.06 5.83 0.55
44.03 1.53 9.8 18.74Olive 46.43 5.63 0.55 44.91 2.48 10.6 17.80Pine
48.18 5.71 0.15 43.89 2.07 9.0 18.67Fig. 2. Reactor's scheme.E.B.
Machin et al. / Renewable Energy 78 (2015) 478e483 480combustion
rate of the fuel gas in the reactor as negative effect ofthe
modications implemented.The mean tar content of the producer gas
obtained in the ex-periments was 9.10 mg/Nm3for Olive, 4.07
mg/Nm3for Peach and8.73 mg/Nm3in the case for Pine. Fig. 6 compares
the tar content inthe producer gas obtained by several authors
19e35 mg/Nm3[26],5 mg/Nm3[25], 97 mg/Nm3[27], 50 mg/Nm3[28] and 10
mg/Nm3[29]; withthecontent obtainedinthestudiedreactor.
Thegasquality is comparable with the obtained in experiments with
theoptimizedtwostagesgasier, developedbyBentzen[25](5 mg/Nm3), but
with higher HHV. Burhenne et al. [29] reported similargas quality,
with a minimum tar content of 10 mg/Nm3and
HHVbetween4.85and4.48MJ/m3usingamulti-stagedgasicationtechnology.The
CO/CO2 and H2/CO ratios are constant; the heating value ofthe gas
is a direct consequence of its chemical composition, whichdepends
on the reaction conditions, rather than the heating valueof the
entering biomass, equal for all those experienced.The increase of
the residence time of the gas mixture in reactorasconsequenceof
themodicationinthecombustionchamberalsohastheundesirableeffectsof
decreasingtheefciencyandFig. 3. Experimental installation
setup.Fig. 4. Modular sampling train of tar.Table 2Operating
parameters.Biomass Olive Peach PineMean process time (h) 3.80 2.50
3.10Mean temperature error 1.0 K (K)T1513 473 503SD 18 20 22T2531
491 521SD 49 21 18T3880 780 853SD 30 25 22T41193 1173 1143SD 60 65
61T51123 1153 1103SD 68 73 62T6417 425 408SD 7 9 5Biomass fed (kg)
8.74 7.6 7.75FlowsAir (Nm3/h) 5.74 5.3 5.4Gas 28.9 18.4 21.3E.B.
Machin et al. / Renewable Energy 78 (2015) 478e483 481productivity
of the gasier; that is why these parameters are lowerthan in
commercial gasiers. According to this, more experimentsare required
to determinate the optimum angle to achieve a bal-ance between all
these effects in order to obtain a clean gas withoutdiminish
signicantly the overall efciency of the gasicationprocess.
Furthermore the small size of experimental model and
itsproportionallyhigherheatloss,
inuencesintheoverallprocessefciency.These results have been
obtained applying additionally, acleaning systemtruly simple and
inexpensive, for particlesremoving.5. ConclusionsAcleanproducer gas
was obtainedwithanovel downdraftgasier. A modied combustion chamber
that prevents theformationof cool
zonesinsideitandincreasesthethermal ho-mogenization in this
reaction zone was developed. This modica-tiontogether
withanextensionof the reductionzone allowsdiminishing the tar
content in the producer gas. The mean values ofthis parameter in
all the experimental tests were lower than 10 mg/Nm3.
Thelowtarandparticlecontent makestheproducergasobtained in this
reactor suitable to the use in cycle Otto
engines.AcknowledgementWearegrateful
totheCoordinationfortheImprovementofHigher Education Personnel
(CAPES) (process 5993105), from theBrazilian Ministry of Education
(MEC) and to the National Councilfor Scientic and Technological
Development (CNPq) (process162633/2013-0) fromthe Ministry of
Science and Technology(MCT) for their generousnancing support to
this research.References[1] GiltrapDL, McKibbinR, BarnesGRG.
Asteadystatemodel of gas-charre-actions in a downdraft biomass
gasier. Sol Energy 2003;74:85e91.[2] Babu BV, Chaurasia AS.
Modeling for pyrolysis of solid particle: kinetics andheat transfer
effects. Energy Convers Manag 2003;44:2251e75.[3] Neeft JPA, Knoef
HAM, Onaji P. Behaviour of tar in biomass gasication sys-tems. Tar
related problems and their solutions. Novem ed. 1999. Nederland.[4]
Devi L, Ptasinski KJ, Janssen Frans JJG. A review of the primary
measures for tareliminationinbiomassgasicationprocesses.
BiomassBioenergy2003;24:125e40.[5] Akay G, DogruM, CalkanOF.
Biomasstorescue. ChemEngLond2006;786:55e7.[6] Huang J, Schmidt KG,
Bian Z. Removal and conversion of tar in syngas fromwoody biomass
gasication for power utilization using catalytic hydrocrack-ing.
Energies 2011;4:1163e77.[7] Jordan CA, Akay G. Speciation and
distribution of alkali, alkali earth metals
andmajorashformingelementsduringgasicationof fuel canebagasse.
Fuel2012;91:253e63.[8] Abu El-Rub Z, Bramer E, Brem G. 2004 review
of catalysts for tar elimination inbiomass gasication. Ind Eng Chem
Res 2004;43:6911e9.[9] Miskolczi M,Borsodi N, Buyong F,Angyal
A,Williams PT.Production of py-rolytic oils by catalytic pyrolysis
of Malaysian refuse-derived fuels in contin-uously stittred batch
reactor. Fuel Process Technol 2011;92:925e32.[10] Elbaba IF, WuC,
Williams PT. Hydrogenproductionfromthe pyrolysis-gasication of
waste tyres with a nickel/cerium catalysts. Int J Hydrogen En-ergy
2011;36:6628e37.[11] Wu C, Wang L, Williams PT, Shi J, Huang J.
Hydrogen production from biomassgasication with Ni/MCM-41
catalysts: inuence of Ni content. Appl Catal BEnviron
2011;108:6e13.[12] Jordan CA, Akay G. Effect of CaO on tar
production and dew point depressionduringgasicationoffuel
canebagasseinanovel downdraftgasier. FuelProcess Technol
2013;106:654e60.[13] Bui T, Loof R, Bhattacharya SC. Multi-stage
reactor for thermal gasication ofwood. Energy
1994;19(4):397e404.Table 3Tests results.Biomass Olive Peach
PineInputsGasier conditionsFeed (kg/h) 3.3 3.05 2.5Gasier air (20
C, 1 bar) (kg/h) 6.79 6.20 6.45OutputsDry gas (kg/h) 9.02 8.60
8.17Water (g/Nm3) 114.5 96.5 102.3Char e ash (kg/h) 0.160 0.085
0.128Tar (mg/Nm3) 9.10 4.07 8.73Error 0.01SD 0.19 0.19 0.19Dry gas
analysisCO (% vol.) 17.4 17.7 16.0H2(% vol.) 13.2 15.0 12.1CO2(%
vol.) 12.4 13.5 11.4CH4(% vol.) 0.8 1.2 0.2O2(% vol.) 1.3 0.9
0.9N2(% vol.) 54.9 51.7 59.4Dry gas HHV (MJ/Nm3) 3.55 3.97 3.65Gas
density (kg/Nm3) 1.183 1.167 1.191Operating ratiosO2/dry biomass
0.45 0.44 0.44CO/CO21.40 1.31 1.40H2/CO 0.76 0.85 0.76Mass balance
and energy efciencyMass in/mass out 1.01 0.98 0.99Cold gas efciency
0.61 0.78 0.58Fig. 5. Temperature prole along the reactor height in
the 3rd experimental test usingOlive.Fig. 6.
Comparisonbetweenthegasqualityobtainedbydifferent
authorsandthepresent study.E.B. Machin et al. / Renewable Energy 78
(2015) 478e483 482[14] Susanto H, Beenackers AACM. A moving-bed
gasier with internal recycle ofpyrolysis gas. Fuel
1996;75:1339e47.[15] Reed TB, Das A. Handbook of biomass downdraft
gasier engine systems. TheBiomass Energy Fundation Press; 1998.[16]
Volchkov EP, Lebedev VP, Lukashov VV. The LDA study ofow
gas-dynamicsin a vortex chamber. Int J Heat Mass Transf
2004;47:35e42.[17] GuoHui-Fen, ChenZhi-Yong, YuChong-Wen.
3Dnumerical simulationofcompressible swirlingow induced by means of
tangential inlets. Int J NumerMethods Fluids 2009;59:1285e98.[18]
Guo Hui-Fen, Chen Zhi-Yong, Yu Chong-Wen. 3D tangentially
injectedswirling recirculatingow in a nozzle with a
slotted-tubedeffects of grooveparameters. Int J Numer Methods
Fluids 2010;63:1256e69.[19] SyredNJ, Be erM.
Combustioninswirlingows:areview. CombustFlame1974;23:143e201.[20]
Lilley DG. Swirl ow in combustion: a review. AIAA J
1977;15:1063e78.[21] Catrakis H, Aguirre R, Mason J. Physical
modeling of turbulentuid interfacesandow regions at large
Reynoldsnumbers. World Sci Eng Acad Soc TransMech Eng
2005;2:1e19.[22] Tabak EG, Tal FA. Mixing in simple models for
turbulent diffusion. Comm PureAppl Math 2004;57:563e89.
http://dx.doi.org/10.1002/cpa.20012.[23] DD CEN/TS 15439. Biomass
gasication- tar and particles in products gases-sampling and
analyze. UE. 2006.[24] Zainal ZA, Ali R, LeanCH, SeetharamuKN.
Predictionof performanceof
adowndraftgasierusingequilibriummodellingfordifferentbiomassmate-rials.
Energy Convers Manag 2001;42:1499e515.[25] Bentzen JD. Optimized
two-stage gasier. In: Proceedings ofrst world con-ference on
biomass for energy and industry; 2000.[26] BhattacharyaSC, Mizanur
R, SiddiqueAHM, PhamHL. Astudyonwoodgasication for low-tar gas
production. Energy 1999;24:285e96.[27] WalkerM, JacksonG,
PeacockeGVC. In:BridgwaterAV, editor. Small
scalebiomassgasication:developmentofagascleaningsystemforpowergen-eration.
Progress in thermochemical biomass conversion. Oxford, UK:
Black-well Scientic Publications; 2001. p. 441e51.[28] Reed TB,
Levie B. Understanding operating, and testingxed bed gasier.
In:Bioenergy 084, Proceeding of World Conference, Goetborg, Sweden,
June 21.Elsevier; 1985.[29] BurhenneL, RochlitzL, LintnerC,
AicherT. Technical demonstrationofthenovel Fraunhofer ISE biomass
gasication process for the production of a tar-free synthesis gas.
Fuel Process Technol 2013;106:751e60.E.B. Machin et al. / Renewable
Energy 78 (2015) 478e483 483