-
Comparative Study on the Combustion Performance of Coals on
aPilot-Scale Test Rig Simulating Blast Furnace Pulverized
CoalInjection and a Lab-Scale Drop-Tube FurnaceHongyu Li,, Liza
Elliott, Harold Rogers, and Terry Wall*,
Chemical Engineering, University of Newcastle, Newcastle, New
South Wales 2308, AustraliaSchool of Materials and Metallurgy,
University of Science and Technology Liaoning, Anshan 114001,
Peoples Republic of ChinaBlueScope Steel, Port Kembla, New South
Wales 2505, Australia
ABSTRACT: Direct evaluation of the combustibility of pulverized
coals in ironmaking blast furnace pulverized coal injection(PCI) is
difficult. A pilot-scale PCI rig may be used to test the combustion
performance of a PCI coal. However, the replicationof the
conditions in the blowpipetuyereraceway region is complicated and
costly. Drop-tube furnaces (DTFs), which havebeen widely used in
coal combustion research, are seen as an alternative to such
combustion tests. This study therefore comparescoal combustion
performances in a DTF and a PCI rig, where the coal is burned at
the as-ground size distribution to assess thesuitability of a DTF
to replace a PCI rig. In addition, this study tries to establish
the methodology for ranking the PCI coalcombustion performance
through coal burnouts produced in the DTF. The measured burnouts
from both the DTF and thepilot-scale PCI rig produced a linear
relationship against the coal volatile matter (VM) content,
although the trend of the DTFburnouts have a steeper slope. The
burnouts of the two low-volatile coals (coals 1 and 2) stand
outside the band and hadsignificantly higher burnouts in the PCI
rig but not in the DTF. This is attributed to higher char
fragmentation duringcombustion in the PCI rig. Overall, the data in
this study suggest that cheap lab-scale DTF tests can be a good
substitute ofexpensive tests using a pilot-scale PCI rig for the
evaluation of PCI coals.
1. INTRODUCTION
Pulverized coal has been routinely injected into ironmakingblast
furnace tuyeres as an auxiliary fuel for around 40 years,following
the oil crisis in the 1970s. It reacts with pressurizedblast at a
high temperature until it is burnt out or the residualchar leaves
the raceway. This technology reduces costs andimproves blast
furnace productivity. However, the injection rateis limited by the
coal combustion performance. If the injectedcoal combusts
efficiently, a high injection rate may be appliedto reduce the
amount of coke required. However, unburnt charleaving the raceway
increases significantly with an increasinginjection rate because of
a decreasing coal burnout and candecrease the permeability of a
blast furnace burden column. Acombustion test is a safe alternative
to optimize coal selectionfor pulverized coal injection (PCI) in a
blast furnace andminimize negative impacts on the blast furnace
operation.Pilot-scale PCI test rigs, Aachen-type rigs, and
drop-tube
furnaces (DTFs) have been previously used to evaluate
thecombustion performance of PCI coals.13 PCI rigs are designedto
simulate the injection of coal into a blast furnace, includingcoal
injection into the blast passing through the blowpipetuyere system,
jet expansion, and combustion as it passes acrossthe void space of
the raceway. The thermal and chemicalconditions of the blast are
set to model as closely as possiblethose of actual blast furnaces
(such as blast temperature, oxygencontent, and velocity). Early
types of these rigs were describedby Kobe Steel4 and Nippon Steel5
in the development ofmodern PCI technology. Broken Hill
Proprietary, Ltd. (BHP),later BlueScope Steel, developed and
operated a number ofsuch pilot-scale test rigs for PCI coal
research between 1983
and 2008, and a large number of coals have been tested, withsome
results being presented in the public domain.610
The Aachen-type3 bench-scale test rig was developed in
the1980s11 and used to simulate the behavior of fine coal
particlesinjected into the blowpipetuyereraceway region. This
rigcomprises two furnaces in series. The first furnace produces
hotgas as blast, and the other furnace is the combustion
chamber.These rigs have limited availability, and the test results
aredifferent to the results from the pilot-scale PCI rigs.3
Alternatively, a DTF is a very common piece of equipment incoal
research laboratories and has also been used to test coalcombustion
for PCI applications.2,12,13 In comparison to otherlaboratory
combustion apparatuses, such as a thermogravi-metric analyzer and a
wire mesh reactor, it can produce arelatively high temperature
(1800 K) and a high heating rate(104 K/s). The particles are fed in
a dynamic, dilute phase,which allows for individual and cloud
particle combustion. Inthis kind of reactor, both solid and gaseous
products aremeasured downstream from the feeding port. Lu et al.13
used aDTF with temperatures from 900 to 1500 C to produce charfrom
pulverized coal before characterizing it, suggesting thatchar
produced in a DTF would be similar to that in PCI. Du etal.2
studied the influences of the reaction temperature, coalparticle
size, fuel ratio (defined as the ratio of fixed carbon to
Special Issue: 4th (2013) Sino-Australian Symposium on
AdvancedCoal and Biomass Utilisation Technologies
Received: July 31, 2013Revised: October 27, 2013Published:
October 28, 2013
Article
pubs.acs.org/EF
2013 American Chemical Society 363 dx.doi.org/10.1021/ef4014967
| Energy Fuels 2014, 28, 363368
pubs.acs.org/EF
-
volatile matter), and coal blending on the burnout of
pulverizedcoal in a DTF at the temperature range of 11001400
C.Their findings indicated that the combustion behavior of
coalsdepends upon not only the gas temperature but also theparticle
size.Significant differences exist in the configuration and
combustion conditions between a PCI rig and a DTF. Nocomparison
of coal combustion results from them has beenpreviously published
in the literature. The purpose of this workis therefore to (1)
compare the combustion performance of arange of coals combusted in
the two rigs with typical particlesize distributions of PCI, (2)
establish the methodology to findif the ranking of the PCI
combustion performance can bepredicted by DTF tests on the same
coals with DTF burnoutscovering a similar range as the available
PCI burnout data, and(3) identify coals not fitting the ranking of
combustionperformance by the DTF and suggest reasons.
2. EXPERIMENTAL SECTION2.1. Coal Samples. Seven coals,
previously used in tests on the
PCI rig by BlueScope Steel, were selected. They are suitable for
PCIuse or have been considered for PCI use. These coals were
cold-storedafter the PCI tests and were made available for the DTF
tests in thisstudy, except coals 1 and 2, which were replaced with
fresh samplesfrom the same mines with matching ultimate, proximate,
andpetrographic analyses. The proximate and ultimate analyses for
eachcoal are listed in Table 1.2.2. Experimental Apparatus and
Methods. 2.2.1. Pilot-Scale
PCI Test Rig. The pilot-scale PCI test rig considered here was
operatedbetween 2001 and 2008 at BHP Billitons Research
Laboratories inNewcastle,3 and the burnouts presented in this work
were taken fromthose tests. The PCI rig consisted of a refractory
cylindrical testsection. Air, heated using a resistance heater and
a N2 plasma torch to1200 C, was injected into the combustion
chamber at up to300 N m3 h1 as blast. The blast was introduced
through a duct with areducing internal diameter from 110 to 80 mm
over a length of 800mm upstream of a tuyere, where coal is
injected. After the tuyere, theinternal diameter increased
dramatically, allowing for a free expansionof the gas jet.
Pulverized coal was injected into the blast at the center
of the tuyere inlet via a coaxial injection lance, which was
inclined at anangle of 1012 to the blast duct center line. The
coaxial lanceconsisted of a 19.05 mm (outer diameter) 1.6 mm wall
thicknessouter tube, which carried a cooling gas of air, and a 12.7
mm (outerdiameter) 1.6 mm wall thickness inner tube, through which
coal wasconveyed by a nitrogen stream. Coal injection rates of 2569
kg/hwere used, and the blast oxygen concentration was
correspondentlyincreased from 21 to 26%. The test conditions were
chosen to matchas closely as possible the conditions within the
blowpipetuyereraceway region of blast furnaces. Char samples were
collected by awater-cooled argon-quenched probe at port 5 (925 mm
downstreamof the injection point) at the center-line position and
50 mm oneither side of the center line. The estimated particle
residence time was20 ms, which is equivalent to the predicted
transit time of a coalparticle across the raceway in the blast
furnace.9 The rig layout ispresented in Figure 1.
2.2.2. DTF. An Astro model 1000 DTF was used for the burnouttest
experiments. Furnace operation is controlled by a Honeywellmodel
DCP 511 temperature controller/program. A schematicdiagram of the
experimental setup for the DTF is shown elsewhere.14
The central tube (50 mm inner diameter) is made from
recrystallizedalumina and is heated externally using a graphite
heating element (62mm inner diameter and 300 mm in length). The hot
zone of thefurnace is 255 mm in length and is able to be maintained
at a relativelyuniform temperature (50 C). The oxygen concentration
used in theexperiments varied between 21 and 26%. Mass flow
controllers wereused to produce the correct flow rate of oxygen and
nitrogen beforethe two gases were mixed. The experiments were
completed withoxygen concentrations that matched those in the PCI
tests. When 21%oxygen was required, air from a Kaeser Ask27
facility compressor wasused. The gas was then split between primary
(4.5 L/min) andsecondary (5.3 L/min) gas flows. Coal was fed at a
rate of around4 g/h through a vibrating plate enclosed in an
airtight perspex boxabove the DTF. The primary gas assisted feeding
the coal into the topof the furnace through a small funnel and
water-cooled feeding probe.The secondary air stream was preheated
as it entered the furnacethrough the annulus between the central
tube and the Kaowool heatshield on the feeding probe. Char samples
were collected at the base ofthe furnace hot zone by a water-cooled
collection probe with nitrogengas quenching. A vacuum was applied
to withdraw the char, quenchgas, and combustion gases out of the
furnace. The char sample was
Table 1. Coal Properties
coal 1 2 3 4 5 6 7
proximate analysis (%) M (ad) 1.4 1.6 1.1 2.2 2.6 3.5 6.7VM (db)
13.0 13.6 18.4 27.5 28.4 36.2 41.7FC (db) 77.4 78.0 71.2 63.1 68.9
55.6 53.4ash (db) 9.6 8.4 10.4 9.4 2.7 8.2 4.9
ultimate analysis (%) C (db) 80.8 81.6 80.3 76.4 81.4 76.9 77.9H
(db) 3.6 3.7 4.0 4.3 4.7 5.1 5.5N (db) 1.5 1.8 1.4 1.6 1.2 1.8 1.4S
(db) 0.4 0.5 0.3 0.4 1.2 0.4 0.3O + errors (db) 4.1 4.0 3.6 7.9 8.8
7.6 10.0
Figure 1. Schematic layout of the combustion test
apparatus.16
Energy & Fuels Article
dx.doi.org/10.1021/ef4014967 | Energy Fuels 2014, 28,
363368364
-
removed from the gas stream via a cyclone and aerosol filter.
Theorganic matter in the sample was determined by burning a portion
ofthe sample with air at 815 C in a muffle furnace until a
constantweight was attained. Then, the weight of the remaining ash
wasmeasured, and the unburnt organic matter content of the
charproduced in the DTF was calculated.2.3. Major Differences of
the Two Rigs. The major differences
between the two rigs are summarized in Table 2. In this study,
the gas
temperature in the DTF was set at 1450 C. The furnace had
acalculated residence time of 210 ms and a heating rate of 24
104K/s. On the other hand, the maximum gas temperature in the
PCIplume combustion is estimated to be more than 2000 C, with
acalculated heating rate and a residence time of 105 K/s and 20
ms,respectively.15,16
The oxygen partial pressure provided during combustion
differedgreatly in the two rigs. Stoichiometry of pulverized coal
combustion isbased on reaction 1.
+ =C O CO2 2 (1)
The O/C ratio, which is based on the atoms of oxygen fed per
atom ofcarbon, is often used to describe combustion conditions.
This has avalue of 2 for stoichiometric combustion, greater than 2
for excessoxygen, and less for sub-stoichiometric conditions.In the
DTF, operating with a dilute feed of coal, significantly excess
oxygen exists. The oxygen concentration does not reduce
significantlyduring combustion; all of the coal particles burn in a
consistent gasatmosphere. The O2 concentration of the feed gas
therefore uniquelydefines the combustion conditions, these being 21
(i.e., air), 22.6, and26% in the present study. The conditions in
the PCI rig were assignedto duplicate those of an operating blast
furnace. The O/C ratio wasaround 3.2 (2.83.6 for the range of
coals) when firing with air, about2 with an O2 concentration of
22.6% of the feed gas (1.72.5 for therange of coals), and
approximate 1.4 with an O2 concentration of 26%of the feed gas
(1.21.6 for the range of coals). The decline of O/C isbecause the
coal injection rate was increased without a change in thevolume of
the blast (i.e., gas). Therefore, for the PCI rig, the
O2concentration of the feed gas does not uniquely define the
combustionconditions.The flame conditions were also significantly
different in the two rigs.
In the DTF, coal was fed as a disperse phase in the primary
(feeding)gas stream. Particles were expected to burn individually,
whereas in thePCI rig, coal was injected through the blowpipe into
the blast,producing a highly turbulent plume. These differences of
conditionsand configurations significantly impact the combustion
performance ofcoals.2.4. Analysis of Particle Size Distribution.
The particle size
distribution of coals and chars collected from the PCI rig and
the DTFwere measured by a Malvern 2600 particle size analyzer.
3. RESULTS AND DISCUSSIONThe ash tracer method was used to
determine coal burnout inboth the DTF and the PCI rig. It assumes
that the mass of ashin a coal is conserved as the volatile portion
of the coal is
released and the carbon of the coal is combusted. The
coalburnout (combustion efficiency) was determined by eq 217
=
B
AA
AA
11
1100i
i
0
0 (2)
where B is the burnout and A0 and Ai are the ash content incoal
and char, respectively.During the DTF experiments, five samples of
each coal were
fed and chars were collected and ashed separately. Figure 2
shows the average coal burnout with the range of
experimentalvalues at 21% O2 as a function of the coal volatile
matter (VM)content. The burnout increases with an increasing VM
[drybasis (db)] content. The highest VM coal presents themaximum
burnout of 86.8%, and the minimum value of24.1% was produced by the
lowest VM coal.The burnouts of coals produced in the DTF at
different O2
concentrations from 21 to 26% are shown in Figure 3. These
results increase monotonically with an increasing coal VMcontent
at all O2 concentrations; that is, VM is a dominateindicator of the
coal combustion performance in DTFconditions. In addition, burnout
increases with enhancementof the O2 concentration in the DTF
experiments, but theamount of increase varies among the coals. Coal
2 (13.6% VM)and coal 5 (28.4% VM) show a large increase. However,
coal 4(27.5% VM) and coal 6 (36.2% VM) are weakly influenced by
Table 2. Combustion Conditions in the DTF and the PCIRig
technique DTF PCI rig
particle heating rate(K/s)
104 (ref 2) 105 (ref 15)
peak gas temperature(C)
1450 >2000
residence time (ms) 210 20 (ref 16)flame conditions laminar
dispersed
particlesturbulent, high particledensity coal plume
stoichiometry and O2concentration
significant excessoxygen
approximately stoichiometric
Figure 2. Burnout of coals in the DTF experiments at 1450 C
and21% O2.
Figure 3. Coal burnouts in the DTF as a function of VM at
differentO2 concentrations at 1450 C and () 21% O2, () 22.6% O2,
and() 26% O2.
Energy & Fuels Article
dx.doi.org/10.1021/ef4014967 | Energy Fuels 2014, 28,
363368365
-
the O2 concentration. Thus, coal burnout is influenced by
boththe coal VM content and O2 concentration.The burnouts produced
in the PCI rig are shown in Figure 4.
They vary between 41.3 and 78.7% with an increasing VM.
Coal 3 produced the lowest burnout at 22.6% O2, and thehighest
value was obtained by coal 7 at 21% O2. Coals 1 and 2show excellent
combustion performance, which cannot beexplained by the proximate
analysis. Furthermore, the figurealso shows that the PCI burnouts
decrease with an increasingoxygen concentration. The higher
burnouts were commonlyobtained when the oxygen concentration was at
21%, and thelower results were mostly obtained at the highest
O2concentration (26%). This is due to the effect of O/C at
thedifferent O2 levels on burnout detailed in section 2.3. The
O2supply is not expressed by the O2 concentration but the ratio
ofO/C. A high O2 concentration in the PCI rig tests does notresult
in more O2 available because of a simultaneous increasein the coal
injection rate with the same blast volume. Thus, thehigher burnouts
occurred at 21% O2 (high O/C).
18 In the PCIrig, coal burnout is a function of the coal
injection rate and O/C ratio rather than the O2 concentration.
However, the O/Cratios used in the PCI rig could not be replicated
in the DTF.To clarify the combustion performances clearly within
the
two rigs, the burnouts from the two rigs were compared at
acorresponding oxygen concentration. A comparison of coalburnouts
at 21% O2 is shown in Figure 5a. Two linear bandswith different
slopes were obtained. Burnouts in the DTFincrease linearly from
24.1 to 86.8% with an increasing VM(db) from 13.0 to 41.7%,
producing a steeper slope than thatfrom the PCI rig. The burnouts
from the PCI rig increase from52 to 78.7% in the corresponding
range of VM. Lower VMcoals produced lower burnouts in the DTF than
the resultsfrom the PCI rig, but higher VM coals presented
highercombustion performance. Coal 1 produced good performancein
the PCI rig but sits at the bottom of the region whencombusted in
the DTF. Coal 2 stands outside the highlightedrange of the PCI
results, sitting higher than expected for a coalwith its VM. When
it was combusted in the DTF, this coal sitswithin the expected
range. Coals 3 and 5 sit along the bottomof the PCI band when
combusted in the PCI rig, but they arealong the top of the DTF
combustion region. The burnouts ofhigh VM coals (coals 6 and 7) in
the DTF are higher than thosein the PCI rig and stand at the top of
the trend.
The comparisons of the burnouts from the two rigs at 22.6and 26%
O2 are given in panels b and c of Figure 5,respectively. Two bands
with different slopes can be drawn inthese panels. Coal burnouts
from the DTF increase withincreasing O2 concentrations. However,
those from the PCI rigdecrease. Thus, the shaded band of the DTF
results movesupward, and the band of the PCI results goes downward.
Thetwo bands almost merge together at 22.6% O2. The band of theDTF
burnouts is mostly over that of the PCI results at 26% O2.The slope
of the DTF band of burnouts deceases slightly as theoxygen
concentration increases. However, there is little changein the
slope of the shaded band of the PCI results.The condition
experienced by coal particles in the PCI rig
combined the characteristics of a higher temperature,
higherheating rate, less residence time, and O/C ratios close
tostoichiometry. Coal particles experienced a longer residence
Figure 4. Burnouts produced in the PCI rig at different
O2concentrations as a function of VM at () coal injection rate
(CR)= 24.9 kg/h, O/C = 3.2, and 21% O2; () CR = 40.6 kg/h, O/C =
2,and 22.6% O2; and () CR = 64.9 kg/h, O/C = 1.4, and 26% O2(some
of these results appeared in refs 3, 16, and 18).
Figure 5. Comparison of burnout as a function of VM between
theDTF at 1450 C and the PCI rig at O2 concentrations from 21 to
26%with shaded operating bands: (a) 21% O2, (b) 22.6% O2, and
(c)26% O2.
Energy & Fuels Article
dx.doi.org/10.1021/ef4014967 | Energy Fuels 2014, 28,
363368366
-
time, excess O2, and lower temperature and heating rate in
theDTF. Consequently, different operation bands were
obtained.Because different combustion efficiencies of the same coal
wereobserved in the two rigs, the differences may be attributed
tothe different configurations and combustion conditions
inthem.19
Burnouts produced by coals 1 and 2 (low-volatile coals)during
combustion in the PCI rig are significantly higher thanexpected
with their VM contents. This may be attributed to thesmaller
particle size distribution of the raw coal and higherfragmentation
during combustion in the PCI rig.It is common knowledge that
particle sizes impact the coal
combustion performance. Less time is taken by small particlesto
reach a given burnout than large particles;20 i.e., smallparticles
produce better combustion. The particle sizedistributions of coals
are demonstrated in Figure 6. As seen,
coal 5 had the coarsest particles among these coals. Samples
ofcoals 1 and 2 used in the PCI rig were aged and replaced bynewly
mined samples for DTF experiments. These coals hadthe finest
particle size distribution. Other coals had median
sizedistributions for these samples. The particle size
distributions ofresulting chars are shown in Figure 7 (chars from
coals 4 and 5were consumed in other tests and were not available).
There isno evidence to suggest that one rig produces consistently
finerchars. The char particle size is a function of the coal
particlesize, but different behaviors were observed from different
coals
within the two rigs. The particle size distributions of coals 1
and2 and their chars are compared to the results of coal 7 and
itschars in Figure 8.
It can be seen that coal 7 experienced greater swelling in
thePCI rig, producing larger char than in the DTF, while
theopposite trend was observed for coals 1 and 2. Fragmentationof
chars from coals 1 and 2 in the PCI rig is believed to causethe
chars to be substantially finer. Within the
mass-transfer-controlled regime expected at a high temperature, the
charburning time is inversely proportional to the particle
size,20
which may explain why the burnouts of coals 1 and 2 in thePCI
rig are significantly higher than expected. Lessfragmentation of
the same coals in the DTF resulted in lowerburnouts, falling within
the band with the other coals.Considering laboratory-scale rigs for
coal testing, a
comparison of characteristics in these rigs is shown in Table3.
In comparison to common laboratory-scale rigs, theadvantages of low
cost, easy operation, and good predictionare obviously observed in
a DTF. Although there is hot airinjection in the Aachen-type rig,
the measured burnouts arequite different compared to the results
from the PCI rig.3 Thepeak temperature (normally 1273 K) and
heating rate
Figure 6. Particle size distribution of coals.
Figure 7. Particle size distribution of chars obtained from the
DTF at1450 C and 21% O2 and from the PCI rig at 21% O2.
Figure 8. Particle size distribution of (a) coals 1, 2, and 7
and (b)chars from the DTF and the PCI rig. The same coal sample was
usedin the PCI rig and the DTF for coal 7, but different coal
sizedistributions for coals 1 and 2. Chars were obtained from the
DTF at1450 C and 21% O2 and from the PCI rig at 21% O2.
Energy & Fuels Article
dx.doi.org/10.1021/ef4014967 | Energy Fuels 2014, 28,
363368367
-
(maximum of 50 K/min) of thermogravimetric analysis (TGA)are
much lower than that in the PCI condition, and thecombustion occurs
in a fixed bed. Therefore, a laboratory-scaleDTF is a good
substitute of a pilot-scale PCI rig in theevaluation of PCI
coals.The limited number of coals presently used to obtain the
operating region has not outlined the band definitely. Datafrom
more coals are required to evaluate the use of a DTF forthe
prediction of performance of any new coals for PCI.Further
investigations of volatile release under PCI conditionsand the
resulting char burnout are required. An on-goingmodeling study by
the authors will further clarify the differencein combustion
performance of coals in the two rigs and will bereported in the
future.
4. CONCLUSION(1) The burnout of coals over a range of VM
contents duringcombustion in both the DTF and the PCI rig increases
almostlinearly with an increasing coal VM content. The burnout
ofcoals during combustion in the DTF is more sensitive to coalVM
content. In this study, DTF tests can provide a
reasonableindication of coal combustion performance in the PCI rig
formedium (18.4%, db) to high (41.7%, db) volatile coals. (2)
Theburnouts of the two low-volatile coals (coals 1 and 2)
standabove the band for other coals, have significantly
higherburnouts in the PCI rig than expected, but fit well in
theburnout band in the DTF. This is attributed to charfragmentation
during combustion in the PCI rig, resulting insmaller char particle
size distributions and, hence, greater charburnout.
AUTHOR INFORMATIONCorresponding Author*E-mail:
[email protected] authors declare no competing
financial interest.
ACKNOWLEDGMENTSThe authors thank BlueScope Steel for financial
support andprovision of PCI combustion results, chars, analysis
data, andcoals. The authors also thank the Australian Research
Council(ARC) and the Australian Coal Association Research
Program(ACARP) for financial support.
REFERENCES(1) Ueno, H.; Yamaguchi, K.; Kenji, T. ISIJ Int. 1993,
33, 640645.(2) Du, S. W.; Chen, W. H.; Lucas, J. A. Energy 2010,
35, 576581.(3) Rogers, H.; Wall, T. Review of Ironmaking Blast
Furnace PulverisedCoal Injection Combustion Testing; Australian
Coal AssociationResearch Program (ACARP): Brisbane, Queensland,
Australia, 2011;Project C19049.
(4) Suzuki, T.; Hirose, R.; Morimoto, K.; Abe, T. Symp.
(Int.)Combust., [Proc.] 1984, 14191425.(5) Masakazu, N.; Kojima,
K.; Hara, Y.; Kase, M. Influence of cokequality on blast furnace
performance. Proceedings of the 38thIronmaking Conference; Detroit,
MI, March 2528, 1979; pp 1827.(6) McCarthy, M. J.; Mathieson, J.
G.; Nomura, S.; Rogers, H.Combustion of pulverised coals under
simulated blast furnaceconditions. Proceedings of the International
Conference on Coal Science;Sydney, New South Wales, Australia, Oct
2831, 1985; pp 423426.(7) Keating, J.; Mason, M.; Rogers, H.
Combustion testing ofpulverised coal for blast furnace tuyere
injection. Proceedings of the JointConference of SCENZ/FEANZ/EMG;
Auckland, New Zealand, April910, 2001.(8) Haywood, R. J.; McCarthy,
M. J.; Truelove, J. S.; Mason, M. B.;Thomson, A. D. An experimental
and theoretical investigation ofpulverised coal combustion in blast
furnaces. Proceedings of the FourthAustralian Flame Days; Adelaide,
South Australia, Australia, Nov 910,1995.(9) Mathieson, J. G.;
Trueove, J. S.; Rogers, H. Fuel 2005, 84, 12291237.(10) Scaife, P.;
Mathieson, J.; McCarthy, M.; Rogers, H.; Nomura, S.Replacement of
Oil by Coal Injection at the Blast Furnace; NationalEnergy
Research, Development and Demonstration Council(NERDDC): Canberra,
Australian Capital Territory, Australia, 1983;NERRD Project
79/9185.(11) Machado, J. G. M. S.; Osorio, E.; Vilela, A. C. F.;
Babich, A.;Senk, D.; Gudenau, H. W. Steel Res. Int. 2010, 81,
916.(12) Osorio, E.; de Lourdes Ilha Gomes, M.; Vilela, A. C.
F.;Kalkreuth, W.; de Almeida, M. A. A.; Borrego, A. G.; Alvarez, D.
Int. J.Coal Geol. 2006, 68, 1429.(13) Lu, L.; Sahajwalla, V.; Kong,
C.; Mclean, A. ISIJ Int. 2002, 42,816825.(14) Li, X.; Rathnam, R.
K.; Yu, J.; Wang, Q.; Wall, T.; Meesri, C.Energy Fuels 2009, 24,
160164.(15) Haywood, R. J.; Truelove, J. S.; McCarthy, M. J.
Modelling ofpulverised coal injection and combustion in blast
furnaces. Proceedingsof the Ironmaking Conference of AIME;
Warrendale, PA, 1994; pp 437442.(16) Mathieson, J. G.; Rogers, H.;
Somerville, M. A.; Jahanshahi, S.ISIJ Int. 2012, 52, 14891496.(17)
Su, S.; Pohl, J. H.; Holcombe, D.; Hart, J. A. Prog. EnergyCombust.
Sci 2001, 27, 7598.(18) Rogers, H.; Mathieson, J. G.; Mason, M. B.
The impact of coalpetrographic composition on the combustion of
pulverised coals undersimulated blast furnace tuyere injection
conditions. Proceedings of theCHEMECA 2011; Sydney, New South
Wales, Australia, Sept 1821,2011.(19) Li, H.; Elliott, L.; Rogers,
H.; Austin, P.; Jin, Y.; Wall, T. EnergyFuels 2012, 26,
46904695.(20) Field, M. A.; Gill, D. W.; Morgan, B. B.; Hawksley,
P. G. W.Combustion of Pulverised Coal; British Coal Utilisation
ResearchAssociation (BCURA): Leatherhead, U.K., 1967.
Table 3. Comparison of Coal Test Rigs
type scale operationoperation
cost comments
PCI rig pilot difficult andcomplicated
high closest simulation of PCIand linear increasingtrend
Aachentype
bench normal medial level trend, different toPCI rig3
DTF bench easy low linear increasing trendTGA bench easy quite
low low-temperature
fixed-bed combustion
Energy & Fuels Article
dx.doi.org/10.1021/ef4014967 | Energy Fuels 2014, 28,
363368368
mailto:[email protected]