Development of new equipment for rapid determination of coal gas

Developmentof new equipmentfor rapid
determination of coal gas content
by
D. J. Wlliams, A Saghafi, M. S. Drummond and D. B. Roberts
CSIRO,Divisionof Coal and EnergyTechnology
ABSIRACT
In undergroundcoal mines wbere higb lerrels
of seam ga.!rmay present a safety hazard,
planning of new de'relopmentsneeds to be
based on a prior estination of gassinessof
of sea:n
the areato be mined. Measurements
gascontenthare becomea routine activity in
Australiqnundergroundcoal mines' Anottrer
area cfrere the knowledge of gas content is
inportrnt is gas utilisation. In Australia,
capture of methane fronr some deep coal
measuresis currently being consideredand
initial exploration drilling has been undertaken.
The methodcurrentiyusedby tbe AusEeliqn
coal indusuy is a modification of the US Bureau of Mines method which consists of
measuringthe volume of gas desorbed from
cored or lump srmples of coal sealed in a
ssntqinsl. The 6ein problem with this
method is the length of time required to accomplishthe test. For fast desorptionrate
an averageof one montb is needed to reach
the flat part of tbe desorptioncurve. For
slow desorption rates, tbe desorption period
can be as long as 4 montbs.The other drawbackq of this method are identilied further
dowu in this paper. The time required for
completion of a test can be a signilicant
problem particulaily in coal sea:nswitl outburst history wherc tbe assessmentof gas
contentueedsto be done in the spaceof days
to main&rina high rate of coal production.
L?Cog can be loadedand tested usingthe
threegascontenttest apparatus.The results
of tests are lnown in less tban 2-3 hours.
Core sanples frour in-seam drilling were
collectedand testedusing the USBM and the
CSIRO (rapid desorption) metlods. With
correction for the residual gas content, the
two methods!8r/€similqrresults.
Effect of coal sample size also was studied.
The mediumand smallcrusherswere usedto
determinetbe gas contentof samplesof tbe
weight 15 to 30 g and 200 to 3009 respectively.For the rangeof studiedgascontents,
no diJferencein the results of the two sizes
were observed.
INTRODUCTION
Importancc of determinatlon of gas content
In a gas control and/or prediction project,
the first most important para-neter to meas'
ure is tbe gas content of the mined seam
and, if feasible,tbc gas contentsof the ad'
jacent coal seans.This is palticulatly impor'
tart for tlose mines tbat experiencegas out'
bursting or high gas emission rates during
mining. For suchmines,basedou past experience, a gas content threshold is generally
set. This gas content identifies tbe zonesof
hazardous mining'. Wbererar tbis limit is
exceededsome kind of pretrentivemeasure
The
is undertaken to alleviate the area.
Tbis paper presents a brief descriptionof
outburst-prone mines may bave different
equipmentdesig;nfor the rapid measurement tbresbold contents tban the mines with higb
of gas coutent in coal samplesand comgas emission.In the outburstsituation, gas
pares some resulB of gas content tests obcoutent , gas pressuregadient and rate of
Fined with this metbod and those obtained desorption are all rary important. Howewr
in thc traditional way.
for higb gas emissionrates, gas content is of
parrmount
importance.Otber factors,related
Three different setsof equipmenthaw been
coal
structure, also influence gas re'
to
the
constructedc/hich can measure gas content
rariation in tbe permeFor
exa.mple
lease.
by pulwrising coal samples.Tbis allows tbe
may
create a zone of bigb
of
coalbed
ability
acceleradonof the gas desorptiontate. Coal
gradient
wben combined with
pressure
cfiich
samplesof 3G50g or 200-3009or 1000to
Townsille 19-21N ovember.1992
CoahedM ethaneSynpo sium
'r',
mechanicalproperties of coal may give rise
to an outburst when the coal face is close to
this zone. OutbursBoccurmost frequentlyin
thosemines with mixed gasand a higber per'
cent of carboir dioxide. This is to be expected
as, for tbe same pressure, coal adsorbs a
larger .olume of carbon dioxide than methane. Methane explosionscan occur when the
mine atmospherehas a concentratiouof 5 to
lSVo of methane and such a concentralionis
generallyproduced ia minss witb high methane contents in the coal seams. However,
there are casesof explosions in mines with
wry low gas content. In France a nethane
erylosion occurred in a mine with only 0.5
m'/t total gascontent(Jeger,1992). The gas
contenttest thctefore,shouldnot onlybe resened for gassyareasatrd it is a good practice to routinely dercrmine gas content dur'
ing face adrance.
Apart ftom safety concerns in underground
coal mines, I new area of gas utilisation re'
quires e$ensilt measurementof in-situ gas
conteut. An accurate and rapid method of
gas coDtent test can be used to guide the
6fuillingprogutr and produce a less expensiw aud morc acculate estimationof tbe gas
reser!?.
D.J. Willians et aL
tle methane desorption into the collection
cylinder occurring with a methane partial
pressureof one atmospbere(Kissell et aL
1973). In the French coal researcbcentre
(CERCI{AR) method, adoptedby the Coal
Directorate of the Commissionof the European Communities (Handbook for the
'desorbable
QsalminiagIndustry, 1980),the
gasconteat'of coal is definedas the quantity of gas given off whea the coal passes
from equilibrium wi& gas at seam pressure
to equilibrium with gas at atmosphericpressurein normal conditions.The termsQ r, Qa
ud Qs are defined as follows (Bertard et
al., 1970), Qr is the gas wbich is released
(lost) betweenthe momentof coal e)craction
and wheu the sample is sealedin a canister
for transport. Qz is the gas releasedinto the
canisterduring transport ftom underground
to the surface and then to the laboratory.In
the laboratory tbe sampleis put in a con[ainer, flushed with gas and crusbed at atmosphericpressure. The releasedgasis des'
ignated as Qr . The desorbablegascontent
is definedas the sum of Qr, Qz ana qr.
The situation becomes more complicated
when tbe coal sample containsmi:ad gas.
The USBM method makes no distinction'
whereas the CERCHAR metbod does, by
Deflnltlon of gas content
defining thc total gas as the quantity of gas
gasrela'
values of the gas content of coal samples releasedif coal is crusbedunder a
for a
is
that
reason
tiw pressure of nil. The
appear frequently in techhical reports and
gas
be'
content
meetings.Ilowever it seens that a universal mi:<edgas the definition of
value
measured
the
definition is uot lrt accepted or understood. comes less accurate and
Recently, the Australian StandardsAssocia- may corespond to any partial pressure' The
tions has publisheda'Guide to the determi- partial pressure for each component is less
than one atmosphereduring tbe wbole test
nation of desorbable gas content of coal
tbe partial pressure for eacb gas
and
seems-Directmethod' (1991),In that reporL
the gascotrtentof a coal sanple is presented cbangesduring the test dependingon initial
composition in the sanple and kinetics of
alr -a) lost gas, v/tich is tbe gas lost during
desorption for each gas. The gascontent test
d5illing or prior to any mealiurement, -b)
measr:rablcgas,or thc gas e\rch€d at atmos- with the traditional method for mi:<edgas
pheric pressurcfton tbe non-pulverisedsa.nn- therefore carry the inaccuracy of a pressure
defrnition as the gas compositioncbanges
ple, and -c) residual gas,qihich is releasedif
during desorption.
srmple
pres1trs
is crushed at atmospheric
sure.The symbolsQt' Qz and Qr respectir/tly But for this paper we adopt tbe Ausralian
are used to identify tbese tbree components Standardsdefinition and use the term 'total
'desorbablegasconof the gas content. The
desorbablegas content' as tbe sum of Qt,
tent' is defined as tbe sum of Q1 and Qz,
Qz, and Qr releasedat atmospbericpressure
wbile the 'total desorbable gas content' is
and without distinction between methane
gir,enas Qr, Qz and Qr. Tbis approacbcan
and mixed gas.
be compared with tbose carried out by tbe
USBM and in Europe. In the USBM
metbod tbere is Do Q: determination with
CoalbedM ethuteSympo sium
Townsville19-21N ovember, 1992
Tladtdonal method of gas content test
(ModltredUSBM method)
The in-situmethanecontentof coal seans is
traditiodally measuredby uking a section of
a core, typically 5-10kg in weigbt placing it
in a container and monitoring tbe release of
methane by collecting the gas in a bell'jar.
The rarc of desorptionis such that it can take
sewral weeksto obtain an accurate determi'
nation. Cteady, a more rapid procedure
wouldha.,pbenefitsin speedingup decisions
with regard1emins develoPment.
closedby an on/off valw (Whitey) mounted
about15mm abovethe cap.
3(X)g contalner
This consistedof a 300 mm x 75 run OD x 3
mm ',rnll Vs tube, closed at one end and internally threaded at tbe other to accept a
screw-incap. The final 6 mm of the tube was
recessedto form a sealing surface for an oring mounted on a correspondingsboulder
on the cap; this constitutedthe gas-rigbtseal.
The usefulinternal length of the tube was 150
m:n uhen the cap was fitted. The gas outlet
was via a 3 mm flexible hose closed by a
Whiteymlve.
METEODOLOGY
125I<g contalner
Rapld desorpdonmethod
'rapid desorptionmetlod' is based on
The
the fact that tbe rate of releaseof gas from
coal is determinedby diffr:sion of gas witbin
the coal particleand is tberefore dependent
on ttre particle size.If a diffusion coefficient
of 10l0-cm2/sis assumedfor the gas, then
tbe theoreticalsolution of tbe diffitsion equation for a spherical sample ( Gunther' 1965)
suggeststhat if tbe diameter of coal grains is
reduced to 1 micron, then only 2 secondsis
neededfor release of. 904oof gas contained
in the sample.Howercr if the grain size werc
I cm then 15 yearswould be needed to re'
leasethe same ,mount of gas (network of
fractures within the coal sample keep the
desorptionrimo to much lessthan 15 years).
This was identical to the 300 g container, except thc lengthwasincreasedto 530mm witb
a usefulinternallengtbof 500mm.
Three sets of equipment were constructed,
eachoperating on the sameprinciple, namely
crusbingthe coal sample with a steel ball by
shekingit, vertically, inside a stainlesssteel
',esselTbe difference betweenthe tbree sets
vas size of the sample container and hence
the wigbt of tbe coal sa:nple used for the
metbane dercrmination. The sa.nple sizes
were 30 g, 300 g and 1.25kg and brief descriptionof tbesecontainersfollows:
30 g contalner
ffus snqlls5tcontainerconsistedof a 138sun
x 30mm O.D. with 3mm wall stainlesssteel
tube, closedat one end and threadedon the
outsideat the other end to uke a screw-on
cap.The gas-tigbtsealbetweentbe tube and
the cap consistedof a Teflon wasber.The gas
outlet consisted of a small bole in tbe cap
leadingto a 3 -- O.D. stainlesssteel tube
TownrvilJe
19-21N ovember,1992
Crushtngprocedure
The 30 g and 300g conlainerswere mounted
on top of reciprocatingpistons.Speciallyde'
signedholders were bolted to the tops of the
pistous to firmly retain the containers. For
tbe ssraller one, a cylinder/piston assembly
fton a small air compressorwas used,whilst,
for the larger one, a water-cooledcar engine
block wasused.Tbe strokeswere 60 srm and
80 mm respectirrclSand both were driven by
electricmotors generallyat a ftequencyof a
few Hz. The largestcontainer was mounted
on a ftamework whicb, in turn, was mounted
via linear bearingson two slides.The assembly was connected in tbc usual way to an
electricmotor to provide reciprocatingmotion with a ftequency of about 5 Ha the
strokebeing 200mm.
The crushing,which was effected by steel
balls, was carried out continuously for the
ssrallestcontainer but was intcrrupted at in'
terrals throughout the crusbing cycle for the
otber two. Tbe purpose of this was to avoid
carry over of fine coal dust witb tbe evol'aed
seam gas. Ifence, whilst tbe crushing mills
were in action, the on/0ff val'reswere closed.
Sea.mgaswas only allowed to escapeinto the
gasmeasuringslstem after the dust had been
allowed to settle in the container for about a
minute or so. For the 300 g and 1.25kg size
equipment,sinteredmetal filters were saUed
becauseof tbe greaterquantitiesof coal and
gas erolution. Thesc were mounted on tbe
Coah edM ethane SYnPosium
u
D.J. Wiltiamset aL
screw-in end caps. The filters nominally remor,eddust greater than 7.5 pm in dia:neter
and could be cleaned by back-flushingwith
compressednitrogen. Tbe intermittent nature
of the crushtng prelEnts tlese filters from
cloggiag.
Gas msurern
nt
The containers were connected via flexible
tubing and on/off viahrcsto inwrted measuring cylinders fille6t with acidulated warer.
During the interruptions in the crushing cycle, tbe ralrcs were graduallyopenedto aUo*
ssamttr to escapeand bubble into the cvlinders. Provision was made for taking a sample
of the gsrm gas for analpis by gas chromatography.
Empirically, it was found that the diameterof
the steel balls for effecti.ae crushing was lg
mm and 25 mm for the 30 g and 300 g conteinsls respectirely. For the largest con_
triner, 2 X 25 rnn lrlk g/e19Uggd.
canisterresealedagainand connectedto tbe
inverted cyiinder. The sub-samplewas then
crushedand gas rolume measured. In Fig.l,
the procedure is shown for one of the snmples. At point A, the gas erolved from the
sample sealed in a desorption canister was
Qz = 2.86mr/t,,a sub-sarnple
wastakenand
crushedwhile the rest of tbC sa:nple was allowed to 1s6nin in the desorptioncanister.
Th: Cat evolred during crushing was el =
2.38 m'lt, rhereforethe total desorbablj gas
content for, sa.mple4 is equalsto ez +
el
= 5.?Am'lt. We eryect thereforeto reacb
this .talueif we let the sample in the canisrer
desorbfor a sufEcientlengrhof time. In Fis
f. it can be seenchatthe line C = 5.?1 m3i
can be consideredas tbe aslmptote of the
desorptioncurve.
In Table 1, the resultsof four testson borecoresare shown.The rolume of gasdesorbed
before the first crushing is identified by
ez ,
the gasrolume evolrcd during crushingis
e3
. Qz + Qs is therefore'total desorbablegas
The time required to crush the coal samples,
content'. The sa.ureprocedurewas repeated
sucb tbat there was uo further gas emission,
on
the sameu.mple after a few days:another
was20 - !Q min for the 30 g containerand I sub-samplewas,taken from the canister and
2h fot the ot[ers. The time nried with coals
was crusbed.Q z in Table I is the volume of
from different s€lms. After tbesetimes it was
gasdesorbed from start of desorption(after
found that 70% of the coal saurplehad been
flsilling) uutil tbe start of this new crushing.
puharised to lessthan 90 pm.
Tle gas,desorbedduring crushingis named
Qs. Qz+ Ql isanothermeasurementof
the 'total desorbablegas content' . To verify
RESULTS AIYD DISCT'SSION
tbe method is accurateone should compare
's wio
's.
Comparlson of results from rapld
Qz + Q:
ThesJare
Q'z + Q'l
presentedin Table 1.
desorpdon wlth tradldonal methods
A better measure of the
Comparison has been made of fte gas con,accuragyof the
teuts measlued with the above equipment metbod is to plot C2 = Q z + e.r against
and thosc using tbe traditional desorption Cr = Qz + Ql and estimatethe distanceof
tbe points from line of equal-content,ie the
equipment (USBM method). In-sea.srdri[ings iu four difierent sites were undertaken line C2 = Cr. In Fig. 2 tbe graph of measured couten6 and tbc line of equal-content
and cores ccre taken deep into tbe sean.
is given. The total deviation from ,the line of
Seam gas in all sites consisted of mi:ad gas.
equal gascontentis dev = i X(Cr - Cz)210J
At each driiling site, two canisters *!r"
Slled witb 2 parB of the sa.ne borecore. = ^ 0.79m"/t for an alrrage deviation of 0.2
.3/t.
One of the canisterswas used for the raditional test with tbe gas desorbedrolume be_
To compare tbe results obtained &om Oe
ing continuouslymonitored. Tbe results of
rapid desorptiontechniqueand the resultsof
thesetestsare listed in Table l. The second.
tbe taditional methodsthe gascontentresult
canisterwas used to enluate the new method
in USBM column in Table 1 were adjusted
of gas content testing. Initially this canister
to representthe 'toal desorbablegas con,
was conuected to an inrerted cylinder and
tent'. Based on some residual gas conte-nt
erolved sas was measuredby water displacetests on similar sanples a ralue of 0.5 m1t
ment. After a few days canister was opened,
was consideredappropriatefor tbis sample.
s 5u6-semplevas rqLen for crushing and the
CoalbedMetltaaeSyn posium
Townsville
19-21November1992
In Table 2 tbe 'total desorbablegas content'
for each sample using two methods is given.
The gas content for rapid desorptionmethod
is the a"rrage of Cr and Cz as described
abo,,e.The a'reragedeviation in the results between
the two methods was 0.4 m3/t over the four
tests (only one of the results from 4 and 5
wasused in this analysis).In terms of relative
error using each of this methods produces
0.4/6.58= 6% difference in respect to the
other nethod (tbe a't/eragegas content for
the four srmplesis 6.58m'lt). In Fig 3. the
results of tbe two methods are compared
againstthe equal-contentline.
The a.reragedifference of 6 Vo in results using the two techniquesis very small and even
if we were using the sa:netechniquetwice on
the samesemplewe could not expecta rariation less than this nlue. Although the number of test in parallel is small , we think,
basedon or€rsets experienceand our knowledge of gas desorption from coal sanples,
tbat thc results of the abore tests demonstrate the validity of the rapid desorption
metbod.
Efiect olsample slze on gas content
determlnatlon
Constructionof three ses of equipmentwitb
tbe capability of crushing coal samples of
substantiellydifferent sizesallowed tbe investigation of the effect of the sanple size on
gascontentdetermination.
Theoretically, for a given coal, one can expect that a cbange in size would produce
some scale effect and gas contents results
could rraryaccording to certain rules which
may be determinedstatistically.
If the moistue content of the coal does not
vary wry much, the reasonsfor any scaleeffect can be either, - due to tbe effect of fissure/cleatnetvork or, - due to the \adation
of ash content distribution acrossa segment
of coal seam;in our caseacrossthe borecore.
Tbe eftect of fissures density can be owrcome if the size of sanrpleis larger than the
a\rrage distanccbetween fssures or cleats.
For exanple if the cleat planes naturally di,hf coal sample in cubes of. 2.5 x 2.5 x
:9t
2.5 cm', then we can expectthat for a coal
sempleof weigbt higber rJ,an22 g (coal
Townnille 19-21N ovember,1992
densityis assumedto be 1.4,the effect of size
sboulddisappear).
Allowance for variation in the ash contenr
distribution is difficult. In case of in-seam
drilling one can expect that along a small
length of core (300 mm) irs \ariation should
not be substantial. However, for sross-searr
drilling the ash contenr may vary largely
acrossthe tbichess of coal seam. We suggest that whenercr a gas content is determined, an ash content test be also undertqken and the gas contentbe defined on a
ftee-of -ashbasis.
In Table 3, the resultsof gas contentsdetermined using 30 g and 300 g samplesirre pre:
sented.Fig. 4 comparestle two gascontents
by visualising the positionsof eryeriment
points and the line of equal-content.
CONCLUSIONS
Three setsof equipmentwere constructedto
pulrcrise di-fferentsize of coal samplesand
measuretbe 'total desorbablegas conten!',
The results obtained show that the rapid gas
desorptionmethodcan be used to determine
tbe total desorbablegas content. A maximum of 2-3 hours is neededto determinethe
gas content compared to few weeks by the
traditional method. Contrary to tbe traditional metbod, the gas content determined
witb the rapid desorptionmethod is not dependentupon the sizeof coal grains.Moreover in the case of a mi:red gas, Oe rapid
desorptionmethod is more accuratebecause
tbe results of the tests are not affected by
different rates of desorption of methane and
carbon diodde.
The coal sanples of 2 diJferent sizes were
tested usitg the medium and small size
crusbers.The results of gas content tests
sbowno differencein ralues obtained for tbe
gancontentsin the range of 0.4 to 3 m3/t.
ACKNoIIILEDGMEIYTS
The authors would like to thank Dr R. Williqrns(egs6as Ltd), Mr T Sbarkey( Meuopolitan Colliery) and Mr S Battino (Gastrade) for their assistanceduring resringof
the rapid desorptionequipment.bur $anks
CoabedM ethaneSynposium
26
are erended to Dr R. Lama (Kembla Coal
and Coke Pty. Ltd.) for usefuldiscussions.
RTF'ERENCTS
dg56elirn St^"dard, 1991.Guide to the determination of desorbablegas content of coal
serms-Direct method'.AS 3980.
Bertrad C, Bru;et B and Guutber J, 1970. '
Deternination of desorbablegas concentration of coal (direct metlod)'. Int. J. Rock
Mech. Min. Sct Vol.7. pp 43-65.
Coah edM etltancSynposium
D.J.lVillianset aL
KissellF N, McCullochC M, and Elder C H,
L973. The direct method of determining
methanecontent of coalbedsfor lentilation
design'.USBM Report of Investigation'7'767.
Guntber J, 1965. Etude de la liaison gazcharbon'. Journal de I'industrie minerale.
France.
Jeger C, 1992.Application of the European
experiencein longwall gas emissiou control
to the NSW longwallmines'. Syurposiumou
coalbedmethaneresearchand development
in Australia.
Townsville19-21Novemben1992
A
v
4.5
a
8
4
aa
t (
,,<
Figurc1. Gas desorption curve for a samplefrom wniO
the emountof gesreleasedduring crushing.
It
lt
bo l0
.El
a
e e
5 s
E
0
6 6
o
8 s
4
6
1
Cas content, fir*
Figure2. Comparison
of gascon
Townsville19-21N ovember. 1992
1
9
cndring (m3/t)
.
Coahed M ethaneSyn posium
28
D.J. Williams et aL
u
6 D
E
9
E s
E
c
o
e
5
rY
4
7
8
_ 6
9
Cas conrent (rn3i), Rapid mahod
Figure3.
Sample
I
conteut determination,trr
Qr.
2.16
(3 davs)
J
Al
5'
Qe
2.30
(2 davs)
7.43
(33davs)
2.86
(1 dav)
2.86
(l dav)
4.t2
Q:.+ Qe
O.ZU
767o CIlt
1.96
Q'r.
4.37
(33 davs)
4.26
J,5 I
2.97
54% Clll;
10.40
238
57%&l^
2.38
SLVo
&ll
(7 davs)
9.18
(3 davs)
5.24
J.Z4
USBM
Q,1
2.08
U% C.H
0;58
0.99
80%CH,
5.01
0.91
(15davs'l 89%er
s.03
(19davs)
0.90
90%c*r,
Q z + Q r method
6.45
3.95
10.17
4.94
(35 davs)
4.29
(15 davs'l
10.50
(48 davs)
5.9r
s.93
4.24
(8davs)
4.24
(8 davs)
r samples4 and5 areiaken
fromttresameborccore
andsealedin nvodiffercntcanisteni
Table1. Resultsof gasron
CoalbedM ethane Synposium
Townsville
19-21Novemb
er I99i
Sample
I
2
a
4.5
Rapid
6.37
4.11
USBM
5.34
4.69
t0.29
10.90
5.58
4.64
TableZ Comparisoubetween'total desorbablegascontent'obtainedby rapid desorptionand
traditiooal techniques.
15-30g
200-300g
Composition
sample
samDIe
CHzl(COr+CIIa)
2.09
2.4
84Vo
1.99
r.9z
2.98
2.95
54?o
0.61
0.62
9270
2.39
2.36
5lVo
0.92
0.91
0.51
0.42
1007o
0.49
0.43
89%
LAOVo
Table3. Qsmparison
of gascontentusingsnaUffi
Townnille 19-21N ovember,1992
CoabedM ethaneSynposium
30
D.J. Williantset aL
-g
!t
Et3
-
!o
8
)x
(\
- 2
7
E
E
8 r
3
A
uas content(r8/t),
Coalbed Methattiifi posrum
15_30g samplc
TownsitteI 9-21NoiinberJ W