Page 1
Mike BeesEskom Ma
4.4 A lower gsurface. the hydra
Table 2:Ground
zon
Shallow groundwazone
Intermedupper groundwazone
IntermedLower groundwazone
Gus coalgroundwazone
Lower groundwazone
slaar ajuba UCG
Lower Ggroundwater No informataulic conduc
: Hydraulic Pdwater ne
D
ater Throuarea
iate
ater
Throuarea.compthrou
iate
ater
Throuarea.compthrou
seam ater
Throuarea
ater Throuarea
Groundwatr zone is asstion regardin
ctivity will be
Parameters Distribution
ughout Majuba
ughout Majuba (May be
partmentalized gh dykes)
ughout Majuba (May be
partmentalized gh dykes)
ughout Majuba
ughout Majuba
ter zone umed to be
ng piezometrlow.
of GroundwDepth
0-70m below surface
70 - 120m below surface
180 – 270m below surface
280-284m below surface
284-unknown depth.
6/41
present beloric levels hyd
water ConceHydraulic P
(HydConduc
/TransmisK=1.7 x 10-
8.6 x 10-3 mtest done ducurrent stud
K=8.0 x 10-4
(Slug test dthe current
T=0.1 to 0.9pumping dothe current
K=1.0 X10-4
X10-5 m/day
No informatavailable
ow the Gus cdraulic proper
eptual ModeParameters raulic tivity (K) ssivity (T) 1 m/day –
m/day (Slug uring the dy)
4 m/day
done during study)
9 m2/d (test one during study)
4 m/d to 1.0 y
tion
oal seam at rties is availa
el
Highly weKaroo se
Permeabdepth
Groundwfollows th
High grougenerally
Watercouareas
Fractured
Permeabfracturing
Rechargezone
Fracturedbelow the
Discharg
Fracturedwithin the
Groundwmbgl dur2008
Rechargezone
Discharg
Fracturedbelow the
• Permfractu
• Rechzone
• DischRiver
11
depths belowable but it ca
Properties
eathered/fractuediments
bility generally d
water piezometrihe topography
und between wy constitutes rec
urses and sprin
d dolerite.
bility depends ong
e from overlying
d dolerite and Ke dolerite.
ge to local base
d coal and lithole coal seam
water levels appring 2006-2007
e from overlying
ge to local base
d dolerite and Ke Gus seam
meability depenuring. Likely to bharge from ove. harge to regionr?)
1613755_Mem_22 February 20
w 284 m beloan be assume
s
red dolerite and
decreases with
ic surface gene
atercourses charge areas
gs are discharg
n the extent of
g groundwater
Karoo sediments
level (Vaal Rive
logical partings
roximately 100 and 40-60 mbg
g groundwater
level (Vaal Rive
Karoo sediments
nds on extentbe very low. erlying groundw
al base level (V
_006013
ow ed
d
rally
ge
s
er?)
l in
er?)
s
t of
water
Vaal
Page 2
ME
F
Mike Beeslaar skom Majuba UCG
Figure 3: Updated C
onceptual Hydrogeoological Model
7/41
11613755_22 Febru
_Mem_006 uary 2013
Page 3
Mike Beeslaar Eskom Majuba
5.0 NU5.1 InThe groundas the volumflow model the coal seastrata on gr
The developsuitable conis available gasifier to thand UCG pr
The objectiv
To repUCG s
To useground
Ch
Th
Th
Th
It is importadata and acassumption
5.2 MThe code sethe WASY Ian interactivdensity-couheat transpoefficiently usdesign remeFEFLOW is
5.3 NThe modelligeological athe groundwmodelled ar
5.3.1 Boundary cthey expresboundary coconditions.
Dirichl
ra UCG
UMERICAntroductiowater flow mme of groundis further utilam to surfaceoundwater in
pment of a mntaminant soa mass tran
he regional groject.
ves of the mo
present, in nusite and surro
e the model tdwater syste
hanges in nat
e coal seam
e groundwat
e potential im
ant to note thccordingly is s. The mode
Modelling elected for coInstitute for Wve groundwapled, thermoort in subsursed to descrediation strat
s used worldw
Numerical ing area wasand structurawater systemrea is approx
Modelling onditions ex
ss the conditionditions resBoundary co
et Type (or c
L GROUNon model has bedwater inflowised to assee as well as nflow rates in
mass transpoource term fosport model
groundwater
odelling inclu
umerical formounding area
to simulate thm, including
tural groundw
m hydrostatic
ter inflow rat
mpact in grou
at the groundpresented as
el will be peri
Code onducting th
Water Resouater modellingo-haline or unrface water reribe the spatitegies and towide as a hig
Model Pros selected baal control. Bom. Most of theximately 32 k
Boundariepress the waons of know
sult in differeonditions in a
constant hea
NDWATER
een constructw that may bess water leveto assess thento the gasifi
ort model usinr the gasifiercan be utilissystem durin
ude:
m, the principas; and
he gasificatio:
water levels;
pressure; an
e into the ga
undwater lev
dwater flow ms a preliminaodically upda
e modelling urces Planning system forncoupled, vaesources wital and tempo
o assist in degh-end groun
operties ased on a cooundaries of te boundarieskm, and the a
es ay the considn water flux,nt solutions h
a groundwate
d) boundary
8/41
R MODELL
ted and updae expected toel drawdowne impact of per.
ng the flow mr which is pre
sed to assessng the opera
ple hydrogeo
on activities a
;
nd
asifier.
vel as result
model has bary model usated once ne
is FEFLOW.ng and Syster three and twariably saturath or without oral distributesigning alterndwater num
nceptual mothe numericas were selectarea is 32,16
dered domain or known vahence the imer flow mode
y conditions o
LING
ated to addreo impact on tn impacts in tpossible goa
model as a besently unavs the possibleational and po
logical flow p
and hence a
of goafing.
een developsing a concepew data beco
. This is a finems Researcwo-dimensioated, transienone or multiion of groundrnatives and
merical mode
odel and alsoal model werted along the69,352 m2.
n interacts wariables, suc
mportance of el can be spe
or:
ess key operthe UCG prothe various afing and sub
ase is subjecailable. Oncee migration oost closure p
processes oc
ssess potent
ed on the baptual model wome availabl
ite element pch, Ltd. Berlinnal, areal annt or steady sple free surfadwater contaeffective molling tool.
combinationre chosen to e watersheds
with its enviroh as piezomstating the c
ecified either
1161322 F
rational quesocess. The gaquifers extebsidence of th
ct to the avae sufficient inof contaminaphases of the
ccurring at th
tial impacts o
asis of spatiawith significale.
package devn, Germany.
nd cross-sectstate flow, maces. FEFLO
aminants, to onitoring sch
n of topograpreflect the g
s. Perimeter
onment. In otetric head. D
correct boundas:
3755_Mem_006February 2013
stions such roundwater nding from he overlying
ilability of a nformation ants from thee gasifier
he Majuba
on the
ally limited ant
veloped by FEFLOW is
tional, fluid mass and OW can be plan and emes.
phical, eometry of of the
her words, Different dary
63
e
Page 4
Mike Beeslaar Eskom Majuba
Neuma
A mixt
The model dgasifier to m
The boundanumerically boundary co
5.3.2 The hydrogmodel, fieldvariable, anrepresent th
Table 3: MoAquife
Shallow Aquifer
IntermediaAquifer
Coal seam
Lower Aqu
5.3.3 A finite elemelement grid174,512 elenumerical m
ra UCG
an Type (or s
ure of the ab
domain was minimise imp
aries of the nby what is re
ondition).
Model Layeological set data and ge
nd is controllehe different a
odel Layers r Mod
Layer Layer
ate LayerLayerLayer Layer
m Layer
uifer Layer
Finite Elemment networkd was compiements and 9model area.
specified flux
bove.
designed to act from UC
numerical moeferred to as
yers tting is repreeology as dised by the geoaquifers as d
in relation tel Layer
r 1 r 2
TW
r3 r4 r 5 r6
CUSL
r 7 C
r 8 B
ment Mesh k (grid) was dled by FEFL
99,828 nodes
x) boundary c
ensure the eG activities,
odel are shows a “no-flow”
sented by anscussed in thological inforescribed in T
to Geology
Topsoil Weathered d
Contact doleUnweatheredSugary dolerLower Sedim
Coal seam la
Base sedime
designed to pLOW, which fs. Figure 5 il
9/41
conditions; o
edge of the mthus avoidin
wn on Figureboundary co
n eight-layerhe earlier secrmation receTable 3.
Lit
dolerite, Sed
erite, Sedimed, unfracturerite, Sedimen
mentary unit.
ayer
entary unit
provide a higfacilitated thelustrates a th
or
modelling dog boundary e
e 4. These boondition (zero
ed model bactions of the ived from Es
thology Unit
imentary uni
entary unit. ed dolerite unntary unit.
gh resolutione constructiohree-dimens
main is sufficeffects on the
oundaries areo specified flu
sed on the ureport. The tskom UCG. T
t
t
nit.
of the numen of a trianguional view of
1161322 F
ciently distane model resu
re representeux Neuman
updated concthickness of tThe model la
erical solutionular mesh cof the finite ele
3755_Mem_006February 2013
nt from the ults.
ed Type II
ceptual the layers is
ayers
n. The finite onsisting of ement
63
Page 5
Mike Beeslaar Eskom Majuba
Figure 4: Num
ra UCG
merical Modell Boundaries ((Blue dots indi
10/41
icate the drainnage system o
over the mode
1161322 F
elling domain)
3755_Mem_006February 2013
63
Page 6
Mike Beeslaar Eskom Majuba
Figure 5: Thr
5.4 M5.4.1 The steady-activities. Thsubsequent
The model wgeological uunder/overemodel calib
The simulatconductivity
ra UCG
ree dimension
Model CaliSteady-sta-state pre-ophis period wat stages.
was calibrateunits and effeestimated, thration. Table
ted water levy or recharge
nal view of finit
ibration ate pre-opeperational staas modelled
ed using the ective aquifeis can be co
e 4 indicates
vel distributioe values can
te element net
erational caage compriseto provide a
observed bor recharge. Smpensated bthe calibrate
on is comparebe altered u
11/41
twork
alibration es the natura
a baseline an
orehole wateShould the aby the adjusted hydraulic c
ed to the mentil an accep
al groundwatd appropriat
er level, the hverage aquiftment of the conductivitie
easured headptable correla
er flow regime initial cond
hydraulic confer thicknesshydraulic cos specified in
d distribution ation betwee
1161322 F
me prior to Uditions for mo
nductivity of ts therefore beonductivity van the model.
and the hyden measured
3755_Mem_006February 2013
CG odelling of
the e alues during
draulic and
63
Page 7
Mike Beeslaar Eskom Majuba
simulated hhydraulic co
Table 4: HyLayer
Layer 1
Layer 2
Layer 3
Layer 4
Layer 5
Layer 6 Layer 7
Layer 8
5.4.1.1 Evapotransalready ove
5.4.1.2 For the purpreasonable model was
5.4.2 The results calibration tDeviations fshould be ragroundwatethe simulate
Table 5: FinBH ID Ob
WMS1
WMS2 WMS3 WMS4
P1 P2 P3
PZ1 PZ2 PZ4
ra UCG
eads is obtaonductivity (K
ydraulic conr
WeatSedimContaSedimUnwedoler
SugaSedim
LoweCoal Base
Evapotrapiration was
ercome the e
Rechargpose of this mand likely tocalibrated us
Steady Stafrom the ste
the observedfrom the straandomly dister levels (Anded and obser
nal calibratibserved WL (
(mamsl) 1675.3
1677.2 1689.8 1680.8 1621.9 1630.4 1624.9 1606.5 1600.7 1601.0
ained. The caK) and recha
nductivity (KLithology U
Topsoil thered doleritmentary unit act dolorite, mentary unit.eathered,unfite unit.
ry dolerite, mentary unit.
er Sedimentaseam layer sedimentary
anspirationnot specifiedffect of evap
ge modelling stu
o result in simsing all availa
ate Calibrateady-state cad groundwateaight line, whtributed indicderson and Wrved groundw
on measure(2005) Simu
(m1
111111111
alibration prorge within a
K) values spUnit
t
te,
ractured
ary unit.
y unit
n d in the mod
potranspiratio
udy, a constamilar regionalable groundw
tion Resultalibration are er levels are ich is the pe
cating that theWoessner 19water elevati
es for the stulated WLmamsl) (1672.2
1677.1 1679.2 1673.8 1619.2 1619.2 1619.5 1613.5 1621.2 1620.7
12/41
ocess was donarrow rang
pecified in thAverage
hickness (m20
37
3
107
3
100 4.5
120
el. It is assuon losses fro
ant rechargel predictions water levels.
ts summarizedplotted againrfect match bere is no bia
992). A correions.
teady state mME(m)
(WLo-WLs)i3.2
0.1 10.6 7.0 2.7
11.2 5.3 -6.9
-20.4 -19.7
∑=-6.9
one by adjuste compatible
he model
m) K
Kx,Ky=0
Kx,Ky=0
Kx,Ky=0
Kx,Ky=0
Kx,Ky=0
Kx,Ky=0Kx,Ky,K
Kx,Ky,K
med that them the system
e rate approxcompared to
d in Table 5 anst the simulbetween the
as toward oveelation coeffic
model MAE(m)
|(WLo-WLs)3.2
0.1 10.6 7.0 2.7 11.2 5.3 6.9 20.4 19.7
∑=87.2
ting the mode with the hyd
value (m/d)
0.3; Kz=0.01
0.01;Kz=0.00
0.03;Kz=0.01
0.0015;Kz=0
0.01;Kz=0.00
0.0152;Kz=0Kz=0.0001
Kz=0.001
e recharge spm.
imately 5mmo spatially dis
and Figure 6lated water leobserved an
er or under pcient of 96%
)i|RMS(
|(WLm-W10.1
0.0113.48.67.412528.547.9416.388.
∑=118
1161322 F
el parametedrogeologica
)
15
0104
15
0.0003
05
0.001
pecified in th
m per year is stributed rec
6 To show thevels in Figund simulatedpredicting thewas obtaine
(m) WLs)i|2 1
0 .4 6 4 5 5 9 .8 .8
86.4
3755_Mem_006February 2013
rs for al situation.
e model has
considered harge. The
e level of ure 7 values,
e ed between
63
s
Page 8
Mike Beeslaar Eskom Majuba
BH ID Ob
WL=Water MAE=MeanRMS= Root
Table 5 shoSociety for T(1992); Spitwas calculamethods: th(Anderson a(WLs) watewater level overall mod
The root meis generally
Figure 6: Com
ra UCG
bserved WL ((mamsl)
level; ME=Me
n Absolute Ert Mean Squar
ows the RMSTesting and tz and Moren
ated for each he mean erroand Woessnr levels. In kechange acro
del response
ean squared not under o
mparison betw
(2005) Simu(m
ean Error; rror; red Error
S error, whichMaterials (A
no (1996). Th observation
or (ME), the mer 1992). Theeping with s
oss the mode(Anderson a
(RMS) errorr over-predic
ween observed
ulated WLmamsl) (
R
h is a CalibraSTM) guidelhe difference borehole. T
mean absolue ME is the standard prael domain. If tand Woessne
r of the grouncting groundw
d water level v
13/41
ME(m) (WLo-WLs)i
1/n=-0.7
RMS% of wat
ation Error Anines for calib
e between thThe error in thute error (MAmean differe
actice, the RMthe ratio is ser, 1992).
ndwater levewater elevati
vs simulated w
MAE(m)|(WLo-WLs)
1/n=8.7
ter level rang
nalysis calcubration of thee simulated he calibration
AE) and the roence betweenMS error wasmall, the erro
el range is 11ons.
water level
)i|RMS(
|(WLm-W1/n=10
SQRT=10e=11.7
ulated accorde model. (Andand the obsen was expresoot mean sqn measured s evaluated aors are only
.7%, which i
1161322 F
(m) WLs)i|2 07.9 0.4
ding to the Aderson and Werved hydraussed by threequared (RMS
(WLo) and sas a ratio to a small part
indicates tha
3755_Mem_006February 2013
merican Woessner ulic head e common
S) error simulated the total of the
at the model
63
Page 9
Mike Beeslaar Eskom Majuba
Figure 7: Cor
Figure 8 shotowards the
ra UCG
rrelation betwe
ows the steae drainage ar
een observed
ady state watreas and tow
water level vs
ter level contwards the nor
14/41
s simulated wa
tours from thrthwest of the
ater level (R2=
he model oute UCG trial m
=0.96)
put. Generalmining area.
1161322 F
l groundwate
3755_Mem_006February 2013
er flow is
63
Page 10
Mike Beeslaar Eskom Majuba
Figure 8: Ste
5.5 O5.5.1 A transient model baseoperational
ra UCG
eady state wat
OperationaIntroductiocalibration h
ed on the watphase.
ter level conto
al Phase Ton as been undter levels me
urs
Transient C
dertaken usineasured in bo
15/41
Calibratio
ng the pre-moth shallow b
on
ining steady boreholes an
state calibrad deep bore
1161322 F
ated groundweholes during
3755_Mem_006February 2013
water flow g their
63
Page 11
Mike Beeslaar Eskom Majuba
5.5.2 The mine pl
Curren
In 2014produc
According to
The model h9.
Gasificationmodel. Eachcompleted.
The determmine develo
Table 6: UC
From the UCwas calcula10 indicated
Panel
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
ra UCG
Mine plan lan informati
ntly mining o
4 the miningction for 6 ye
o this very b
has been co
n panels of 22h panel’s wo
ination of groopment will p
CG Mine Pla
CG water proated from thed the average
Starting(day
01224364860728496
108120132144156168180192204216228240252264
on provided
n the wester
will continueears.
road mine pl
nstructed to
20m by 100morking progre
oundwater inprobably be d
an specified
oduction date average wae water prod
g time ys)
E
0 0 0 0 0 0 0 0
80 00 20 40 60 80 00 20 40 60 80 00 20 40
by Eskom (S
n side of gas
e to the rema
an it is assu
simulate min
m were definession is 120
nflow into thedifferent from
in the Mode
a provided bater productioduction receiv
Ending time (days)
120 240 360 480 600 720 840 960
1080 1200 1320 1440 1560 1680 1800 1920 2040 2160 2280 2400 2520 2640 2760
16/41
September, 2
sifier 1 (P1-P
ainder of gas
med that gas
ne progress
ned to assign days (4 mon
e workings mm the mining
el
by Eskom, anon from P5 aved from Ma
Panel
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
2010) is sum
P30).
sifier 1 and g
sifier 1 and g
as described
n the operationths) and by
may thereforeassumptions
n estimated wand P6, from ajuba UCG.
Startin(da27283031323334363738394042434445464849505154
mmarised as f
asifier 2. Thi
gasifier 2 will
d in Table 4a
onal boundar2020 gasifie
e be conservas incorporate
water producJanuary 200
ng time ays) 760 880 000 120 240 360 480 600 720 840 960 080 200 320 440 560 680 800 920 040 160 400
1161322 F
follows:
is gas field w
be complete
and illustrated
ry conditionser 1 and 2 wi
ative; since ted into the m
ction of 16 to09 to July 20
Ending time
2880300031203240336034803600372038403960408042004320444045604680480049205040516052805400
3755_Mem_006February 2013
will be in
ed by 2020.
d in Figure
s in the ll be
the actual model.
ns/day that 10. Figure
e (days)
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
63
Page 12
Mike Beeslaar Eskom Majuba
Figure 9: Illus
Figure 10: Av
5.5.3 The calibratobservationwas conducobtained fro
ra UCG
stration of min
verage water p
Transient Ction process
nal data whilected for the som monitorin
ne plan used in
production fro
Calibrationconsists of o
e adjusting thsteady-state g.
n the model
m P5+P6 (Esk
n obtaining a mhe aquifer papre-operatio
17/41
kom).
match betweearameters in nal model us
en the simulathe model w
sing the grou
ated model rwithin a realisundwater leve
1161322 F
results and stic range. Cael measurem
3755_Mem_006February 2013
alibration ments
63
Page 13
Mike Beeslaar Eskom Majuba
The transienPZ2 and PZof the mininindicate sim
5.6 ImImpacts on and gradienusing the as
5.6.1 Scenario 1 wduring the sscenario, w
5.6.1.1 No further doutlined mingasifier 1 an
The gasifierspecifying cthe constanschedule.
5.6.2 Groundwatethe modelleafter 2640 d
Figure 11: Sc
5.6.3 Figure 12 re13 represenperiod. Figu
ra UCG
nt model wasZ4 during theng from the smilar water lev
mpacts onthe groundw
nts, drawdowssumed gasi
Operationawas run usin
steady and trhereas the B
Modellindetailed gasifne plans indind gasifier 2
r mine plan foconstant heant head cells
Simulated er inflow intoed total inflowdays (7.5 yea
cenario 1: Sim
Impact on epresents thents the shalloure 14 and F
s calibrated e beginning otart to 2010. vels to the o
n the grouwater flow regwn and inflow
fication plan
al Phase Sng the calibraransient stateB5 dolerite si
ng Methodofier mine placated the opas discusse
or each timed cells at thein the model
Inflow intoo the UCG mw during the oars) while the
mulated water
groundwae shallow aqow aquifer waigure 15 sho
by using the of gasifier wo
The results bserved one
undwater fgime include
w rates to the described e
cenario 1: ated steady se calibration.ll acts as a b
ology for thns are availa
perational phad in previous
e period and e bottom elevl were theref
o UCG mineine workingsoperation phe average wa
inflow into UC
ter level inuifer water leater level con
ow the coal se
18/41
recorded waorking panel obtained fro
es.
flow regime changes in e gasifier. Theearlier.
No Goafinstate water le. The updatebarrier.
he Calculatable to descrase was sims section.
panel has bevation of the fore specified
e during ths depends onhase of the Uater inflow in
CG mine
n mining arevel contoursntours after 5eam water le
ater level me(panel 1-panm the transie
me the natural ge impacts on
g or Subsievels with theed conceptua
tion of Infloribe how min
mulated for a
een incorporgus coal sea
d exactly acc
he operation the mining
UCG. The manto UCG mine
rea over thes after 26405400 days (1evel contours
easurements nel4). This inent state cali
groundwater n the groundw
dence e hydraulic pal model was
ow Rates ining will progrperiod of 540
ated into theam (layer 7).cording to the
on phase (5panel progre
aximum inflowe is 211m3/d
e 15 year mdays (7.5 ye
15 years) at ts after 2700
1161322 F
s obtained frodicates the pibration simu
levels (hydrawater were s
parameters ds used as ba
nto the UCress. As note00 days (15
e flow model . The floor ele mine worki
5400days) ess. Figure 1w of 252m3/d
d over the mi
mining periears) of mininthe end of mand 5400 da
3755_Mem_006February 2013
om PZ1, progression ulations
aulic heads) simulated
determined ases for this
CG Mine ed earlier years) for
by evations of ng
1 shows d occurs ning period.
od ng. Figure
mining ays. Figure
63
Page 14
Mike Beeslaar Eskom Majuba
16 shown athe mining p
The followin
No imp
After 2southe
After 5the sou
Figure 12: W
ra UCG
section fromperiod.
ng can be co
pact on the s
2700days, theeast of the UC
5400 days, thuthwest from
Water level con
m west to eas
oncluded from
shallow aquif
e maximum CG mine wo
he maximumm the UCG m
ntour map of sh
st through th
m Figure 12 t
fer water leve
distance of torkings within
distance of mine workings
hallow aquifer
19/41
e mining are
to Figure 16
els is seen a
the 1m drawdn the coal sea
the 1m draws within the c
r after 2700da
ea and indica
below:
s a result of
down cone isam aquifer.
wdown cone hcoal seam aq
ays of mining:
ates the chan
the UCG min
s seen a dist
has is seen aquifer.
Scenario 1
1161322 F
nge in water
ning
tance of 1423
a distance of
3755_Mem_006February 2013
level over
3m to the
f 1317m to
63
Page 15
Mike Beeslaar Eskom Majuba
Figure 13: Sh
ra UCG
hallow grounddwater contour
rs after 5400 d
20/41
days of miningg: Scenario 1
1161322 F
3755_Mem_006February 2013
63
Page 16
Mike Beeslaar Eskom Majuba
Figure 14: Co
ra UCG
oal seam piezzometric heads
s after 2700 d
21/41
days mining: SScenario 1
1161322 F
3755_Mem_006February 2013
63
Page 17
Mike Beeslaar Eskom Majuba
Figure 15: Co
ra UCG
oal seam Piezzometric head contours afte
22/41
er 5400 days: SScenario 1
1161322 F
3755_Mem_006February 2013
63
Page 18
Mike Eskom
Figur
Beeslaar m Majuba UCG
re 16: Change in waater level during UCCG mining Scenario 1
23/41
11613755_Me22 February
em_006 y 2013