1

Marketdrivenmineplanning:Optimisingproducts
portfolio,MuzoEmeraldminecasestudy
EnriqueRubio
MiningEngineeringDepartment,UniversidaddeChile
ABSTRACT
Mine planning is the mining engineering that transforms a geological orebody into a business
promise that shall optimise shareholders value. Traditionally this process commences with the
geologicalmodellingstatingtheestimationoforeresources,thentheseresourcesaresubjecttoan
economic envelope which is after sequenced defining mining methods and finally a production
schedule is computed in order to provide the financial team the best offer to generate the sales
contracts.Ingeneral,thismethodologyworksforcommoditiesandrawmaterialsinwhichformal
contracts are set among the producers and the buyers. In the precious stone business and in
particularintheemeraldbusinessthemarketismuchmoresegregatedanddifferentbuyersattend
emerald auctions to acquired package of emeralds that will be later transform in any form of
jewellery.Thus,thebusinesschangesdependingonhowthedifferentemeraldpackagesareoffered
toasetofbuyers.Thischallengehasmotivatedtheauthortodeviceamineplanningmethodology
tointegratedifferentlevelsofoperationalhedgingtorespondtoamarketsegmentationthatcould
changeovertime.Inparticular,flexibilityhasbeendesignedandaddedto:thenumberofoperating
mines under production, the rate of development, preparation and production at any given time,
the adoption of an inclined draw point caving method, and finally the automated production
control system to capture in real time the emerald production. The device methodology is under
application by Muzo International at the Muzo underground mine, located in the province of
Boyac, Colombia. This paper describes the theoretical framework, the actual tools developed to
apply the methodology and full details regarding the mine and plant design to transform an
artisanaloperationintothefirstworldemeraldfabric.

INTRODUCTION
Themineplanningtraditionalprocessflowisasfollows:

Figure1Mineplanningtraditionalflowsheet
Inthepreciousstonebusinessatremendousfactorisplayedbytheintrinsicuncertaintycontained
in the underlying asset concentration as grades and the market in which those gemstones will be
finally commercialised. The following figure shows the market composition of the emerald
productionworldwide.

Figure2Worldwideemeraldproduction
Alsothepriceofemeraldsplayatremendousroleinvaluingdifferentmineplanningalternatives.
Thefollowingfigureshowsapricetrendovera35yearsperiod.

Figure335yearsemeraldprices,nationalgemstonecompany

Atypic

Itissh
extrem
depend
Thefol
import

From t
above,
quality
20%ar
50% is
thefina
calcommerci
hownthatthe
mely volatile.
dingongems
llowingfigur
tanttonoteth
the total mat
50% was cl
yemeraldsan
reemeraldsth
waste. The p
alprice(seeF
alisationsche
Fig
reareseveral
Typically in
squality(size
reshowsthep
hatcuttingan
Figur
erial recovere
assified as w
ndgemsname
hatcanbetre
products of th
Figure1).Crit
emetomarke
gure4Comme
lstepsfromm
the emerald
e,cut,clarity,
percentageof
ndpolishingh
re5Valuechai
ed from the m
waste. From t
edChispero,
eatedtobeco
his chain are
ticalpointsar
3
etemeraldsis
ercialisationsch
minetomark
d business the
greenningles
feachpotenti
has,inaverag
inMuzoemera
mine emerald
the remainin
thesecond20
omeartificial
emeralds gro
retherobberi
theoneshow
heme,agents
ketthatmake
ere will be a
ss,shape),Ch
ialproductfr
ge,a50%reco
aldmineprojec
d product of
ng 50%, 10%
0%arequalit
emeralds`,n
ouped in lots
esatthemine
wnbelow.

thewholeem
at least three
hispero,Morr
romthemine
veryasapro

ct
the methodo
corresponds
tyemeraldsn
namedPerma
s for auction,
e,theclassific

meraldbusine
main produ
rallaandPerm
etomarket.It
cess.
ology describ
to the high
namedMorral
andthelow
which then g
cationstagea
ess
ucts
ma.
tis
bed
hest
lla,
est
get
and

finally the auction. The engineering, development and management should aim to control these
itemstoensurethesuccessandsustainabilityofthisbusiness.
In terms of pricing, GemFieldsthe largest public emerald producing company reports high
variability as a function of time and underlying contracts existing between the intermediate agent
andthefinaljewellery.
Table1SalesperAuction,Gemfieldreport2010
AUCTIONRESULTS
SUMMARY
JULY09
AUCTION
NOVEMBER09
AUCTION
MARCH10
AUCTION
JULY10
AUCTION
DECEMBER10
AUCTION
Dates 2024July2009 2327November2009 1115March2010 1923July2010 610December2010
Location London,England Johannesburg,S.A. Jaipur,India London,England Johannesburg,S.A.
Type HigherQuality HigherQuality LowerQuality HigherQuality HigherQuality
Caratsoffered 1.36million 1.12million 28.90million 0.85million 0.87million
CaratsSold 1.36million 1.09million 22.80million 0.80million 0.75million
No.ofcompaniesplacingbids 23 19 25 37 32
Averageno.ofbidsperlot 10 13 8 18 16
No.oflotsoffered 27 19 56 27 19
No.oflotssold 26 14 49 24 18
Percentageoflotssold 96% 74% 88% 89% 95%
Percentageoflotssoldbyweight 99.8% 97.2% 78.9% 94.2% 86%
Percentageoflotssoldbyvalue 82% 76% 89% 87% 99%
Totalsalesrealisedatauction USD5.9million USD5.6million USD7.2million USD7.5million USD19.6million
Averagepercaratsalesvalue USD4.40percarat USD5.10percarat USD0.31percarat USD9.35percarat USD26.20percarat

These two sources of uncertainty create a great deal of volatility when valuating any of the main
components of the traditional mine planning process. Therefore, to set up a mine planning model
upon expected values of prices of main outcome production and the production itself would be
extremely dangerous, and there is certainly a tremendous gain potential as well as a loss. Thus, a
differentmethodologyhasbeendevicedinordertoderivethemainmineplanningdecisionssuch
as economic envelop, mining sequence and production scheduling as a result of a portfolio
optimisation exercise in which the expected return over the investment as well as its volatility
aretakenintoaccount.
PROPOSEDMETHODOLOGY
Efficient portfolio has been discussed extensively by Samis et al (2006), and Davis and Newman
(2008), using real options and quantifying the risk of different mining strategies and also reviewing
value at risk method. In this paper, the author wanted to give a fresh review at the Markowitz
method (1959) and complemented by Haugen (1990) and Merton (1990) in which he defines a
frontier efficient optimisation method to allocate resources to a portfolio of assets with different
return over investment and risk. The methodology consists of computing the cross covariance of
all the possible combination of assets in a portfolio to compute the mediumvariance space upon
which a given portfolio is efficient to be invested in. So, for instance, in the following figure the
highlighted dots represent a portfolio that is inefficient since there are combinations of assets that
couldprovideahigherreturnforthesamecomputedaveragerisk.

Note that the risk in this context is seen as the average volatilityof the underlyingasset portfolio.
Then the mining application will be to mimic several mining decisions such as mining methods,
productionrate,miningsequenceandproductionscheduleasifthesedecisionswherecomponents
of a portfolio. Then the covariances of different decisions will define the variance of a given
decisionsubjecttotheotherstatussuchasmine,productionrate,sequenceandothers.

Figure6Frontierefficientforportfoliooptimisation
Thefirststeptousethismethodologyistomodeltheprobabilitydistributionofthemainproducts
of interests, in the case of emeralds these products would be heads or Chisperos, Morralla and
Perma. The following figure shows probability density functionsfor those three prices taken for a
givenhistoricaltimeinterval.

Figure7Priceprobabilitydensityfunctionsfordifferentproducts
The following step consists of modelling the grade concentration of the main products (Chispero,
MorrallaandPerma)asaprobabilitydensityfunctionforeverydifferentminingmethodtouseas
anextractionsystem.Thefollowingfiguredepicturesthesefunctionsforagivenminingmethod.

R
e
t
u
r
n
o
v
e
r
I
n
v
e
s
t
m
e
n
t
Risk
IneficientProjects

Figure8Gradesprobabilitydensityfunctionfordifferentproductsforagivenmineandminingmethod
After defining the probability distribution of the main sources of return volatility of interest to
integrateinthedecisionmodel,theassetportfolioshouldbedefined.Inthiscaseitcorrespondsto
define for every one of the mines or sectors under study the possible mining method. In other
words,beingaminem = 1. . Handalternativeminingmethods u = 1. . u,anassetcanbedefinedas
thecombinationofamineandamethodas A
u
m
.Thenthegradeofproductkfortheminemandthe
method u canbe definedas g
k
(m, u), the expected price for productk is defined as P
k
, the mining
recovery can be defined as r(m, u), the processing recovery can be defined as R. For the different
productsapricethatisdistributedfollowingaknowndensityfunctioncanbemodelled P
k
.Thena
valuefunctionisproposedasfollows.
:(m, u) = _g
k
(m, u)
k
r(m, u)R|P
k
-c
k
] -C(m, u) _ I(m, u)
(1)
Where:
c
k
isthesellingcostofcuttingandpolishingofproductk
C(m, u)istheminingcostofminemusingextractionmethodu
I(m, u)istheproductivityofminemusingextractionmethodu

Notethatgrades,pricesandmethodproductivityareallrandomvariableswithknownprobability
densityfunction.Thusthereturnovertheinvestmentiscomputedas:

rr(m, u) =
:(m, u)
{ g
k
(m, u)
k
r(m, u)R|c
k
] -C(m, u) ]I(m, u) +Fc(m)

(2)
Where:
Fc(m)isthefixedcostofmovingonetonoforefromminem.

Then several simulations are performed over the defined random variables sampling the
probability distribution that defines theuncertainty of grades, pricesand method fora given time
periodatthemine.ThentheexpectedrateofreturncanbedefinedasE|rr(m, u)]foraportfolioof
n assets co:(i, ]) would be the covariance of asset i respect to j. Thus the average standard
deviationofagivenportfolioofcomponentsx

isdefinedas:
o = _x

x
]
co:(i, ])
]
_
0.5

Then the above formulation can be optimised by minimising the standard deviation subject to a
given minimum expected rate of return and assuming that the portions of the portfolio x

should
addatthemost1,beingx

thepercentageofcapitaltodevelopasseti.
CASESTUDY:THEMUZOSTRATEGICPLANNING
EmeraldMuzominesarelocated815moverseaintheWestDepartmentofBoyac,Colombia.The
population of this zone is about approximately 15,000 people. The zone has special
geological/metamorphic features that facilitated the emerald genesis. In spite of there was mining
from1540bySpaniardsconquers,itwasonlyinthe60sthatbiggeremeraldvolumesstartedtobe
produced.Thehistoryofminingmethodsatthiszoneisasfollows:
19601970:surfacemining.
19701985:surfaceminingwithminingloaders.
1985current:undergroundmining(tunnels,shaftsandchimneys).

Figure9BoyacDepartmentinColombia
The mine operation until 2009 was divided into a series of shafts: Puerto Arturo, Tequendama,
Catedral, Retorno 1 and Volver, where there were several artisanal mining contractors who did
not register neither production information nor plans or topographic features. The figure shows a
3Dmapoftheminingatthemomentoffinding.

Figure103Dlayoutoftheminebackin2009
The owners of the mine CoexminasS.A. made the decision backin 2009 to associate withaNorth
American investment consortium CrestInvestment to take over the operation through an option
contracttochangethewaythemineoperatedandpursuedamassiveemeraldproductionreducing
therobberiesfrom50%oftheactualproductiondownto10%.Inordertoimplementthisobjective
CrestInvestment contracted the Chilean engineering company REDCO Mining Consultants to
conceptually design, implement and operate a solution that could match the budgeting and time
constraintsoftheowners.
Thechallengewasdividedintofourmainareasofdevelopment:
1.Modifythecurrentdriftingminingmethodintoamoremassiveandcontrolledminingsystem
2.Implementawirelesstrackingundergroundsystem
3.Implementanopticalsortingplanttoautomatetheemeraldclassificationandcleaningprocess
4.Modifythemineplanningprocessandproductioncontrolsystems
Alternativeminingmethods
Driftting
ThemethodthatwasusedinthepastatMuzoconsistedoffollowingtheinstinctofdifferentmine
contractors without taking samples or any geological observations that could facilitate any
engineeringprocedureatthefield.
Driftandfill
BecauseoftheoccurrenceofsuccessfulproductioninTequendamamineduetotheidentificationof
a geological emerald belt, it was necessary to incur in subsequent deepening production levels,
whichconstantlyweakenedtheinfrastructureofthemineinlevelsR1Inf.,S1R1Inf.,S2R1Inf.Due
to this, was proceeded to design and build a concrete slab that could support the vertical
and horizontal forces present in the sector, the location of it and its schematic design are shown
inthefollowingfigure.

Tqdama
R2
R1
Volver
Catedral
Puerto Arturo
3D Model
Mine Shafts [m] Drifts [m]
Tequendama246 2487
Parturo 154 718
Retorno1 83 1176
Volvere 239 1744
Retorno2 215 476
Catedral 139 1145

Figure11Constructionscheme:constructionsite(left),design(right)
Besides the two activities described above, a rehabilitation programme was implemented all over
haulage tunnels, replacing timber pools support in poor condition according to a cost management
and a priorities strategy that would not interrupt the production of emeralds and kept their
stabilityoutofrisk.
Inclineddrawpointcaving
The design of the exploitation method involves the construction of loading stabs, facing
perpendicularlythegeologicalproductivezone,andconnectedthroughamainhaulagelevel.This
patternisrepeatedverticallyifageologicalandstructuralproductivecontinuityzoneisverified.

Figure12Mineralisedzoneprofile
Theplanviewoftheproposedminedesignforthisexploitationmethodisshowninthefollowing
figure.ForTequendama,theaccessshafttotheproductivelevelsindepthwillbemadethroughthe
CL04shaft,becauseofitsconvenientlocationinhardrock,whichgivesanappropriatesupportfor
undergroundminingandgoodworkperformance.

10

Figure13Productionlevelplanview
ForCatedralandPuertoArturomines,implementingthesamesystemofexploitationaccordingto
theshownandmodelledgeologicalbehaviourisplanned.
Aproductionsummary2010,withtheinformationaccumulateduntil7January,2011,thelastday
of the REDCO team work in the operation of the Muzo emerald mine. During the mentioned
period, about 280,000 carats of emerald material were produced, corresponding to approximately
350 tulas and approximately 52,800 carrs, with an average of about 32 carrs per metre of mining
advance. The main productive activities were carried out at Level Tequendama R1Inf. and
SubR1Inf,level11S1CatedralandLevel12PuertoArturo.
Table2Operationperformanceindicators,2010
Units Q12010 Q2.2010 Q3.2010 Q4.2010
ROM*Operation (Tonnes) 3,820 4,663 6,382 3,286
ROM*MxZone (Tonnes) 2,674 2,798 4,468 3,244
Drifting (m) 222 223 390 264
ROM*Production (carats) 121,244 60,637 16,850 80,760
Grade (c/t) 45,3 21,7 3,8 18,3
*ROM:runofminematerial
The following figure shows the operational performance measured in terms of drifting, carrs and
tulas,itisnotedthat50%oftheactivityisconcentratedinTequendama,35%inPuertoArturoand
15%inCatedralandVolver.

11

Figure14Productionanddrifting

Figure15Miningoperationandproductivity

Figure16Operationalperformance

10.000
20.000
30.000
40.000
50.000
60.000
70.000
0
20
40
60
80
100
120
140
160
180
Dec,
09
Jan,
10
Feb,
10
Mar,
10
Apr,
10
May,
10
Jun,
10
Jul,
10
Aug,
10
Sep,
10
Oct,
10
Nov,
10
Dec,
10
Jan,
11
P
r
o
d
u
c
t
i
o
n

(
c
a
r
a
t
s
)
D
r
i
f
t
i
n
g

D
e
v
e
l
o
p
m
e
n
t

a
n
d

P
r
e
p
a
r
a
t
i
o
n

(
m
)
CA(m) VO(m) PA(m) TQ(m) Production(Carats)
0
1.000
2.000
3.000
4.000
5.000
6.000
7.000
Dec,
09
Jan,
10
Feb,
10
Mar,
10
Apr,
10
May,
10
Jun,
10
Jul,
10
Aug,
10
Sep,
10
Oct,
10
Nov,
10
Dec,
10
Jan,
11
O
p
e
r
a
t
i
o
n

A
c
t
i
v
i
t
y

(
c
a
r
r
s
)
TQ(carrs) PA(carrs) VO(carrs) CAT(carrs)

12

Notethattowardstheendoftheyearthemineproductionwasmoreconcentratedintothemineralised
zone, which is not wider than 2.5 m. Thus in November and December of 2010, there was a higher
emerald production andless carrs moved. This indicates that the trace of lithologiesand metasomatic
evidencefoundonthehangingwallaretremendouslyrelevantfortheemeraldproduction.
Undergroundproductionmanagementsystem
This project involves the installation of an underground network based on fibre cables and
extremelyresistantwiresforelectricity,opticalsealedswitchesandwirelessaccesspoints,inPuerto
ArturoandTequendamamines,inordertoprovidevideomonitoringservice,trackingpeopleand
carrs,IPtelephonyandsensornetworksystemsinsidethemine.
The system consists of a set of networking equipment, cameras, special cables, tags and electronic
tags, sensors and other devices, all with maximum protection standards for underground mining
conditions and high quality and continuity of service. Approximately one hundred metres of
compoundcable,and500metresofredcablewithspecialcoverage,asetofapproximately30high
resolutionvideocameras,200tagsfortracking,20wirelessaccesspointsandantennashighfingertips
will be implemented. Fixed cameras will be installed at intersections of movement corridors, fixed
cameraswithvariablefocusinarrivalplaces,shaftsandextractionpointsandeasymountingcameras
forextractionpointsandadvancementtunnels,whichwillbeilluminated byinfrared,forlowlight
conditions. The same wireless network infrastructure allows having seven mobile phones, three
insideofeachmineandoneinthesurfacetomaintaincommunicationswiththeengineersincharge
of the operation at all times. Particularly in Figure 17 it is possible to appreciate the location of the
networkcomponentscontrolinTequendamaandPuertoArturomines.

Figure17ProductioncontrolsystematTequendamaandPuertoArturomines

13

AcontrolroomisplannedtobeimplementedonthesecondfloorofthePuertoArturooffices,with
a video surveillance server, which will monitor and record the information of the cameras. There
will be a computer displaying the position of carrs and people in real time (online), through a
trackingapplicationthatwillusethesignalemittedbythetagsinsidethemine.
Theprojectalsoincludesthefutureadditionofsensorstoexpandthenetworkinfrastructurecoverage.
The information that they will deliver will be processed through an application, optimising the
operationalactivities.Thesystemhasbeendesignedinamodularfashion,aimingtomeettheneedsof
coverageasthemineexpands,anditisthereforeconsideredtohavepartsforimmediateextensions.
ProcessingplantfortheMUZOoperation
InordertofacilitatetheclassificationandcleaningprocessofemeraldsfromthePuertoArturoand
Tequendamamines,itwillbenecessarytoestablishamanagementsystemtostockmobilemineral
containers, a storage silo, discharge grills and pick hammers for the oversize and a feed belt
totheprocessingplant.
10tonnescontainers
Near the entrance accesses of Tequendama and Puerto Arturo mines 10 tonnes containers will be
located, to be carried by trucks equipped for that purpose and taken to the feed silo. After being
unloaded,thecontainerswillgobacktobefilledattheoutputofbothmines.
Roadconnectionwithsilofeeder
Tocarrythecontainersfromthetwomentionedmines,theroadsconnectingtheminesandthesilo
mustbeenabledandrepaired.
PuertoArturo:Theroadbehindtheminefacilitiesmustbeconstructedtoconnecttothefeedsilo.
Tequendama:Itisnecessarytorepairtheroadtothefeedsilo,becauseoftheimportantslope.
Surfacesilo
The construction of a silo to store the ore extracted from Tequendama and Puerto Arturo is
planned.Thesilomustbecapableofgivingautonomytotheprocessofopticalsortingforatleast
threehours,anditwillalsoensureanadequateconstantsupplyforthenextprocess.TheSiloisalso
usedtostorethematerialincasedamagemightoccurintheprocessingplant.
Gridsizeselection
Thesortingplantwillprocessmaterialoflessthanfourinchesinsize,soitisnecessarythatthegrid
has that aperture setting. This will ensure the mineral moves the following process having the
appropriateparticlesize.
Secondaryreductionhammer
Asthegridwillhavefourinchesaperture,aflexiblesecondaryreductionsystemwithahydraulic
hammer is necessary. The hammer shall reduce the size of all particles exceeding four inches. Its
actionisdirectlyonthegrill.Itisimportanttomentionthatitisexpectedthatonly5%oftheoreis
greaterthanfourinches.Thefollowingfigureshowsthehammer:

14

Figure18Hydraulichammer
Silowall
The storage silo was located immediately below the wall where the trucks unload the ore from
mines Tequendama and Puerto Arturo and the wall has been built to provide maximum security
forthetruckstodownloadandtobepartofthesilodescribedabove.
Extractionbelt
Tocarrytheorefromthesilofeedertotheopticalsortingplant,itisnecessarytoinstallabeltformineral
movement,asshowninthefigurebelow.Thismainbeltshouldhavetwostraps,onethatcarrytheore
tothenewprocessingplantandanothertosendtheoretotheexistingprocessingplant.
Opticalsortingplant
TheMuzoprocessingplantwasdesignedforatreatmentcapacityof250[tpd]consideringthatthe
plant operates ten hours per day, being operated and monitored from a control room located in a
differentplantequipmentsector.Itmainlyconsistsoftwoopticalsortingequipment,twoscreeners
forselectionofsize,twowashinganddryingtrays,andtransferbelts.Somesmallequipmentwhich
purchaseispending,are:belts,transferchutes,generators,compressors,blowers,amongothers.

Figure19Longitudinalviewofthesortingplant

Thepla
be obs
sorting
Sorters
The or
materia
10
37
19
9.5
GEMf
Thisso
andits

GEMl
Thisso
andits

antwillbelo
erved from th
gplant.
s
re granulome
alintofourfr
01[mm] +37
7[mm] +19,
9.1[mm] +9,5
5[mm] +4,8
fineCOLOR
orterclassified
weightis8,00
argeCOLOR
orterclassified
weightis8,00
catedinsidea
he control ro
etry of the m
ractionstocon
[mm]
,1[mm]
5[mm]
8[mm]
thetwosmalle
0kg.Machine
R
dthetwobigge
00kg.Machine
acompletely
oom. The follo
mine is quite
nsider:
erfractionsan
eof5KVA.
Figure20
erfractionsan
eof5KVA.
Figure21
15
lockedsheda
owing is a br
variable; the
ndtheircapacit
0GEMfineCO
nditscapacity
GEMlargeCO
andcontrolle
rief descriptio
erefore it wa
tyis8[TPH].I
OLOR
yis8[TPH].It
OLOR
edbyvideoca
on of equipm
as determined
Itsdimensions

tsdimensions

amerasthatw
ment used in t
d to divide t
sare4.6x2.3[m
are4.6x2.3[m
will
the
the
m],
m],

16

Screeners
These machines split the mineral in four fractions, the four machines correspond to the Sandvik
LF1030screener,1.0x3.0[m],andtotalweightof3,000kg.Eachonehastwoenginesof6.6[kW].

Figure22Screenerslocation(screener1=H1)
Airblower
It is necessary to wash the mineral to remove dirt and dust, and then proceed to blow the same
product to remove water from the washing process. So need two machines are needed to deliver
500[cfm],withapressureof100[mbar].Eachofthesemachineshasacapacityof7.5[HP].

Figure23Lowcapacityblower

Airext
Comm
moistu
connec

Genera
Below
tractor
modas also pr
ureproductth
ctedtothetw
alinfrastruct
isamapofth
roposed the
heparticlesej
oopticalsort
ture
helocationof
installation o
ject,thismach
tingequipmen
Figure
ftheelements
Figure25G
17
of an exhaus
hinemustbe
nts.
24Airextract
sdescribedab
Generalinfrastr
st fan to con
elocatedonth
tor
bove:
ructure
ntrol airborne
heroofofthe

e dust and t
ehangarand

the
be

18

Productionsecurityarea
The production security area is a highly secure area within the mining complex in which
productiontakesplace.Thisareawasconceivedtobecontrolledbycamerasatalltimesandgated
lockingtheaccessforpeoplenotrelatedtotheoperation.Themainareasare:
Mainaccessgateandpersonnelscreening
ControlaccesstoTequendamaandPuertoArturo
Plantwarehouse
Engineeringandproductioncontrolroom
THEMINEPLANNINGPROCESS
Ascanbeseen,thebusinessvaluechaincanbesetinacircleconnectingthestagesofexploration,
design,productionplanning,miningoperations,productioncontrol,sorting,polishing,andfinally
themarketsale.Becauseofthestrategicobjectivesofthecompanythatownsthemine,theguiding
operation should be based on the strategy of market positioning, conditioning the exploration,
operationandtherestofthechain.

Figure26Strategicmethodology
OneimportantaspectofplanningMuzowastofindthegradedistributionofemeraldssurrounding
ageologicalcontactformedinahydrothermalandpostmetasomaticmetamorphicprocess.Below
is the genetic model in the Muzo emerald mine and the conceptual summary of the mines
geologicalmapping.

19

Figure27Geneticmodel
AstudyoffieldgeologyhasbeenperformedforeachmineoftheMuzoComplextoidentifytheareas
wheretheemeraldsarelocated.Basedonthisdata,thefollowingprovisionoflithologieswasidentified:

Figure28GeologicalprofileMuzoemeraldmine
CarbonateBlackShale(BS):
Carbonatedshalewithaveragehardness.
Itshowsnosignsofposthumoustectonicsandmineralisation.
Typicallyknownasliso

20

FragmentedBlackShale(FS)
Shalewithfolding
Veinswithcalciteandpyrite
Thepyritezonescanbeseenindisseminatedform.
Stockwork(SK)
Blackshalecarbonate
Veinletsofalbite,calcite
Disseminatedpyrite
Veinsupto20[cm]thickwithalbite/calcite(averagethickness10[cm])
Maycontainemeralds
ClayZone(CZ)
Grayrockwithhighlyalteredrocktexturedofbreccia
Areseenhealthyandalteredcrystalsofalbiteanddolomite
Crumblesinyourhand
Disseminatedpyrite
Somelaminarorbandedareasavailable
Youcansubmitmmangularclastsofblackshale
Somecarbonatedareas
Carbonatedbreccia(CB)
Verysoftblackshalethatcrumblesinyourhand
Albiteveins(alteredtoclay)anddisseminatedpyrite
Claymatrix

Overayearofmineproductiongeologicalmapsweresetupinordertofindthemainconcentration
of emerald production at different levels of each mine: Puerto Arturo, Catedral and Tequendama.
ThefollowingfigureillustratesthefindingsatTequendamamine.

21

Figure29EmeraldfindingsandgeologyatTequendama
Thetypicalgeologicalcrosssectionsoftheminesarepresentedasfollows:

Figure30Geologicalmapping
Geophysics
The aim of this study was to test the resistivity geophysical techniques, IP and resistivity
tomography, to determine its applicability in the exploration of emerald mines, and thereby help
reducetheuncertaintyofoccurrenceofemeralds.
WithresistivityandIP,twostudieswereconducted,500[m]longeach.Datawasrecordedinthetime
domain, dipoledipole configuration, the distance between the electrodes was 50 metres (a = 50),
progresswasmadeatevery25metres,therewere6levelsdeep(n=6)(110mdepthofinvestigation)
andanintegrationtimeof2secondswasused.Intotaltherewere950metreslinearsurfacecovered.
In resistivity tomography 10 lines of 45 to 90 metres were studied. Data was recorded in the time
domain,dipoledipoleconfiguration,thedistancebetweenthestakeswas5metres(a=5),upto13
levelsdeepwererecorded(n=13)equivalentto15to20metresdepthofinvestigationandusean
integrationtimeof0.5seconds.Intotaltherewere630metreslinear,ofwhich505metresoccurred

22

in the interior of tunnels and 125 metres at the surface, these last 90, plus 35 test and calibration
samples. We used a team scores IRIS SYSCALPro model. Below there is one of the specific
outcomefiguresforTequendamamine,themostproductiveminein2010.Itispossibletoidentify
the lithological contact between structures of different hardness, allowing to identify sectors
concentratingthehighprobabilityofoccurrenceofemeralds.

Figure312DresistivitymodelTQR1Infsector.Carbonaceousshaleandwet(pink),shale,
drierandmoreestablished(blue,yellowandorange)
Geochemistry
Takingthemainobjectiveofdefiningasystemofeffectiveprospectingandexplorationthatallows
increased emerald recovery with respect to the number of blocks in production, it is of interest to
findarelationshipbetweenthegeophysicalresults,mineralogy,chemicalelementsandoccurrence
of emeralds. Based on this idea, a complete search for equipment that can help obtaining the
spectrum of mineralogical and chemical elements in rock samples, to create a complete system
characterisationandidentificationofareasofhighemeraldprobabilitymustbeconducted.Theuse
of technologies such as fluorescence and Xray diffraction (XRF and XRD) has been considered.
Someoftheresultsfor12samplesanalysedinthelaboratory,arepresentedinthefollowingtable.
Table3Samplesdetailandcriticalelementscontent
N Sample Si
(%)
Al
(%)
Fe
(%)
Ca
(%)
Mg
(%)
S
(%)
Na
(%)
K
(%)
Ti
(%)
P
(%)
Mn
(%)
Sr
(%)
Zn
(%)
Cu
(%)
Cr
(PPM)
Ba
(%)
1 TQDR1i_SASN 36,4 10,3 4,6 2,6 6,3 5,7 0,3 0,4 0,3 0,1 0,0 0,2 0,0 937,0 0,1
2 TQDR1iB03 20,3 7,2 4,8 26,4 5,5 7,3 1,7 1,1 0,3 0,1 0,1 0,0 0,1 0,0 786,0 0,0
3 TQDR1iB01 40,0 11,0 3,2 12,9 4,5 3,0 6,3 0,1 0,5 0,4 0,1 0,0 0,0 0,0 458,0 0,1
4 TQDR1iSAS 39,2 10,5 2,2 14,6 4,1 1,6 6,3 0,2 0,4 0,6 0,1 0,0 0,0 0,0 511,0 0,0
5 CATN11Cx01
(ClayZone)
0,0 0,2 3,3 29,9 19,4 0,6 0,0 0,0 0,0 0,0 0,1 0,0 0,0 0,0 162,0 0,0
6 CATN11Cx01
(ClayZone)
0,0 0,3 6,5 28,1 18,6 6,5 0,0 0,0 0,0 0,0 0,1 0,0 0,0 0,0 481,0 0,0
7 TWDR1iSAS
NCx01
25,1 7,2 6,5 24,4 1,5 8,6 4,1 0,1 0,3 0,2 0,1 0,0 0,0 0,1 519,0 0,0
8 B03(LeftWall) 33,0 8,2 4,5 14,3 8,0 5,0 4,1 0,2 0,3 0,4 0,1 0,0 0,1 0,0 739,0 0,1
9 B03(RightWall) 22,2 6,3 6,2 26,9 3,3 7,9 3,5 0,1 0,2 0,1 0,1 0,0 0,1 0,0 459,0 0,0
10 R1iB03(1) 40,8 10,1 4,6 14,9 3,6 5,7 6,1 0,1 0,5 0,3 0,1 0,0 0,0 0,1 815,0 0,1
11 R1iB01 29,1 9,6 7,0 16,3 6,4 11,2 3,0 1,4 0,4 0,1 0,1 0,0 0,7 0,0 939,0 0,2

12 MO

Q
Q
Q
ORALLA 5
Quadrant1:M
Quadrant2:Re
Quadrant3:M
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1
DolomiteCaMg(CO3
CalciteCaCO3
BerylBe3Al2(SiO3)6
5,4 1,7 2,8
F
EmeraldId=
Modelandreal
ealitywithem
Modelandreal
2 3
3)2
6
44,5 1,6 3
igure32Grap
Figure33
=(1.957*10^3
lityagreewit
meraldoccurr
lityagreeont
4 5
AlbiteNaAlSi3
UraloliteCa2B
Pyrite FeS2
23
3,7 0,5 0,0
hXraydiffrac
3Predictedmo
3xNa/K)+(3
themeraldso
renceofMorr
thenonoccur
6 7
O8
Be4(PO4)3(OH)35(H2O)
0,0 0,0 0,
ctionresults
odel
.01*10^2xA
ccurrence.
ralla,butthem
rrenceofeme
8 9
Illite(K,
Bavenite
Quartz
1 0,0 0,0

Albite[%])
modeldoesn
eralds.
10 11
H3O)(Al,Mg,Fe)2(Si,Al)4O
eCa4Be2Al2Si9O26(OH)2
SiO2

0,0 477,0 0

notmatch.
12
10[(OH)2,(H2O)]
2
0,0

24

After the research process, we recommend testing equipment SAX (Bruker distributor), IGMO
(distributorofFEICompany)andSpectralInternationalInc,performingtestsonsamplessentfrom
themine,thendotheanalysisofwhattechnologyisbestbasedonthequalityandquantityofthe
results,andfinallydefinewhatisthemostappropriateequipmentfortheconditionsofuse.
MINEDESIGNANDPRODUCTIONPLANNINGBASEDIN
PORTFOLIOOPTIMISATION
Basedonthegeologicalgeneticmodelspresentedbeforeaprobabilitydistributionofgradesforthe
differentproductswereconstructedforthedifferentminesatdifferentwidths,withthemainaxis
being the metasomatic contact of hard and soft rock. Every single mining width represents a
different mining method. So for the artisanal method the width happens to be 1.5 m, for the drift
andfill2.5mandforInclinedDrawPointCaving4.0m.Thefollowingtableshowsthelognormal
gradedistributionofChispero,MorrallaandPermaforallthreeminesfordifferentminingwidths.
Table4Lognormaldistributionofgradesforthedifferentminesanddifferentminingmethods
GradesM1 GradesM2 GradesM3
Chispero A1 A2 A3 A1 A2 A3 A1 A2 A3
Media 1.10 0.41 0.36 1.44 0.92. 0.22 1.79 1.39 0.41
StandardDeviation 0.11 0.04 0.04 0.29 0.18 0.04 0.26 0.20 0.06
Moralla A1 A2 A3 A1 A2 A3 A1 A2 A3
Media 1.79 0.69 0.69 2.13 1.03 0.36 2.48 1.39
StandardDeviation 0.36 0.14 0.14 0.43 0.21 0.07 0.50 0.28 0.20
Perma A1 A2 A3 A1 A2 A3 A1 A2 A3
Media 1.57 0.47 0.92 1.91 0.81 0.58 2.26 1.16 0.22
StandardDeviation 0.31 0.09 0.18 0.38 0.16 0.12 0.45 0.23 0.04

Thepricedistributionoveraoneyearperiodwastakenfrompublicreportsaswellasthe2010sales
performedbyCoexminas.Thepricedistributionperproductsisshowninthetablebelow:
Table5Lognormaldistributionofpricestakenfromayearofmineproduction
Prices/products Chispero Morralla Perma
Media 6.68 5.01 3.91
StandardDeviation 0.07 0.10 0.06

Based on the above, several simulations were performed to analyse at least 50 combinations of
mining methods (drifting, drift and fill and inclined draw point caving) for the different mining
areas.Thefollowingparametersofcostsandproductivitybymethodbymineareusedtocompute
therateofreturnofeverycombination.

25

Table6Parametersfortheportfoliooptimisation
M1 M2 M3
ProductivityMethod1(t/year) 18,000 12,000 12,000
ProductivityMethod2(t/year) 60,000 42,000 32,400
ProductivityMethod3(t/year) 72,000 60,000 72,000
CostMethod1($/t) 600 400 500
CostMethod2($/t) 200 300 350
CostMethod3($/t) 100 140 180
MiningRecoveryofMethod1 0.3 0.3 0.3
MiningRecoveryofMethod2 0.4 0.4 0.4
MiningRecoveryofMethod3 0.5 0.5 0.5
ProcessRecovery 0.35 0.35 0.35
Miningfixedcost($) 3,000,000 1,500,000 500,000
Sellingcost($/c) 20

Finally, the efficient frontier is computed for the Muzo Emerald mines for the three different
operating mines andfor three alternative mining methods. There wasan integer constraint added
tothemodeltoavoidsolutionssuchthatinaminetherecouldbetwocoexistentminingmethodsat
thesametime(Norstand,1999).Thefollowingchartshowstheresult.

Figure34EfficientfrontierfortheMuzoEmeraldmine
It was very interesting to see that for every combination of return and volatility the production
scheduleandminingmethodsperminechangeaccordingly.Thefollowingchartshowsthestrategy
of mining method and proportion of production contributing to the schedule per mine for different
rateofreturnandvolatility.

0%
10%
20%
30%
40%
50%
60%
70%
0% 5% 10% 15% 20% 25%
E
x
p
e
c
t
e
d

R
a
t
e

o
f

R
e
t
u
r
n
ReturnStandard Deviation

26

Forinstancethefollowingconclusionscanbederivedfromtheanalysis:
1. Forahighreturnandhighvolatilityapproach,theproductionscheduleshouldconcentrate
atTequendamamineusingdriftandfillmethodconcentratingover80%ofproduction.
2. For an intermediate return and medium volatility, the production schedule should
concentrate at Puerto Arturo with Drift and Fill, Catedral with Drift and Fill and
TequendamawithInclinedDrawPointCaving.
3. For low return but also low risk approach the schedule should concentrate at Puerto
ArturowithDriftandFillmethod.
The following figure shows different mining combinations, methods and schedule that could be
usedatMuzofordifferentreturn/riskapproaches.

Figure35Differentstrategiesandproductionschedulesdependingonthereturn/volatilitydecision
Basedotheaboveguidelinetheschedulefor2011ispresentedasfollows:

Figure36Finalproposedproductionschedule
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
A1_M1 A2_M1 A3_M1 A1_M2 A2_M2 A3_M2 A1_M3 A2_M3 A3_M3
%

o
f

p
r
o
d
u
c
t
i
o
n

u
n
d
e
r

t
h
e

o
p
t
i
o
n
3% 12% 20% 25% 30% 50% 60% 61%
0
20
40
60
80
100
120
140
160
0
10,000
20,000
30,000
40,000
50,000
60,000
j
a
n
,

1
1
f
e
b
,

1
1
m
a
r
,

1
1
a
p
r
,

1
1
m
a
y
,

1
1
j
u
n
,

1
1
j
u
l
,

1
1
a
u
g
,

1
1
s
e
p
,

1
1
o
c
t
,

1
1
n
o
v
,

1
1
d
e
c
,

1
1
R
O
M

(
t
p
d
)
/
H
e
a
d

g
r
a
d
e

(
c
/
t
)
P
r
o
d
u
c
t
i
o
n

(
c
a
r
a
t
s
/
m
o
n
t
h
)
Carats ROM(tpd) Grade(c/t)

27

CONCLUSIONSANDRECOMMENDATIONS
The main conclusion that can be obtained from the approach presented in this paper is that
uncertaintymodellingopensanewwayofperformingstrategicmineplanning.Itisnotpossibleto
integrate uncertainty into our production planning discipline if there are not clear and
understandablefinancialtoolsthatcanfacilitatethedecisionsmakerstoseethevalueofvariability.
Currently,thereisplentyofuncertaintymodellingmethodsavailablethatendupbeingusedasa
sensitivity analysis of a fixed production schedule. This ongoing research has shown that when
adding uncertainty to the modelling of grades and prices the structural mining decisions may
changeaccordinglybasedontheacceptanceofriskandreturn.
In terms of the specific results for the Muzo mine it is interesting to outline that a couple of
modelling techniques together with a financial well known approach could contribute to the
delineation of the ore body, sequence, mine design and production schedule from the valuing of
optionsdowntogeology.
It is expected that the mining industry understands that the new paradigm of strategic planning
wouldbetoconcentratemuchonthemarketandhowthefinancialpositionofshareholderscould
facilitatethedelineationofourmineplanningdecisionsandnottheotherwayaround.
ACKNOWLEDGEMENTS
Theauthorwouldliketothankfirstofalltheorganisationsthatsupportedtheventuresummarised
inthispaperstartingwiththeUniversidaddeChileMiningDepartmentandtheAdvancedMining
TechnologyCentreforsupportingthetechnologyappliedinthisproject.Theauthorwouldalsolike
to thank the engineers of REDCO Mining Consultants that took over the project in particular
Gabriel Pais, Pamela Castillo, Daniela Siuela, Jorge Aros, Claudio Guzmn and Ignacio Muoz.
Also, many thanks to the authors graduate students at the time Marcelo Vargas and Fernando
Peirano for their help in many aspects of the work presented in this paper. Finally, the author
wouldliketothankallhisundergradstudentsthathelpedoutwithmanyshiftsattheMuzomine
andcontributedinagreatdealtothesuccessofthisproject.
REFERENCES
Haugen,RobertandNardinBaker,DedicatedStockPortfolios,JournalofPortfolioManagement,Summer1990,
pp.1722.[1]
Markowitz,Harry,PortfolioSelection:EfficientDiversificationofInvestments,JohnWiley&Sons,Inc.,1959.[2]
RobertC.Merton.ContinuousTimeFinance.Blackwell,1990.[3]
JohnNorstad.Anintroductiontoportfoliotheory.http://homepage.mac.com/j.norstad/finance,Apr1999.[4]
Samis, M., Davis, G. A.,Laughton, D., andPoulin, R., 2006, Valuing uncertain asset cash flows when there are no
options:arealoptionsapproach,ResourcesPolicy30:285298.[5]
Davis,G.A.,NewmanA.M.,2008.ModernStrategicMinePlanning.ColoradoSchoolofMines.Workingpaper.[6]