The Whittle Interface

Introduction to Three-dimensional Pit Optimisation Intrinsic Value Time - The Fourth Dimension Cost Assumptions How to Calculate Costs

1 String Model Interface

Overview of the Strin !odel to "hittle Interface !odel Preparation The #$tent Strin S%&PAC '((( Su)re ions !odel *loc+in S%&PAC '((( to "hittle TH&##-D S%&PAC '((( to "hittle FO%&-D "hittle to S%&PAC '((( S%&PAC '((( to "hittle TH&##-D - Case Stud, S%&PAC '((( to "hittle FO%&-D - Case Stud,

2 Block Model Interface

Overview of the *loc+ !odel to "hittle Interface !odel Preparation &e-uired Attri)utes Indicator .ri ed !odels *loc+ !odel to "hittle TH&##-D *loc+ !odel to "hittle FO%&-D *loc+ !odel to "hittle FO%&-/ Indicator .ri ed !odels to "hittle "hittle to *loc+ !odel "hittle Three-D0 "hittle Four-D and "hittle Four-/ are proprietar, open pit optimisation pac+a es written ), "hittle Pro rammin Pt, 1td2 The, are not included in S%&PAC'((( and must )e purchased separatel,2 This part of the S%&PAC'((( pac+a e provides ,ou with an interface to Three-D0 Four-D and Four-/3 it allows ,ou to prepare the input files for Three-D0 Four-D and Four-/ from ,our S%&PAC'((( eolo ical resource files or )loc+ models2 It allows ,ou to convert a "hittle Three-D0 Four-D or Four-/ result file into a )loc+ model2 It also allows ,ou to e$tract S%&PAC'((( strin files from the output of Three-D or Four-D2 4ou will then use these strin files to uide ,ou in ,our final pit desi n2 The "hittle Interfaces were first developed ), SSI in 566( to replace the internal 7'28-Dimensional9 pit optimisation pac+a e which was in S%&PAC at that time2 The '28-Dimensional pit optimisation pac+a e suffered from a num)er of limitations - it was slow and cum)ersome to use and not trul, :-Dimensional2 It has )ecame outdated and as a conse-uence has )een omitted from S%&PAC'2 !eanwhile0 considera)le software development effort has )een focused to improve the "hittle Interface0 the most recent improvement )ein the addition of an interface to the S%&PAC )loc+ modellin functions2

"e are often as+ed 7"hat is the difference )etween "hittle Three-D and Four-D;9 *oth offer three dimensional solutions0 and apart from havin a hi her price0 Four-D will find the pit with the optimal net present value2 In other words Four-D allows ,ou to e$amine the effect of time and se-uencin on the value of the optimal pit outline2 Three-D on the other hand calculates the optimal pit on the assumption that one minute the mineral and waste is in the round and the ne$t minute the waste is on a dump0 the mineral has )een processed0 sold and ,our mone, is in the )an+< This is not ver, realistic )ut ma, not )e important if ,ou are evaluatin a small pit which has a life of around one to two ,ears2 For a mine which is e$pected to have a life in e$cess of two ,ears0 ,ou are stron l, advised to use Four-D2 "hittle9s latest pac+a e0 Four-/ is ver, similar to Four-D e$cept the optimi=ation will )e calculated over multiple elements in a sin le roc+ t,pe2 Four-D calculates each roc+ t,pe as if it was composed of a sin le element2 This "hittle interface documentation should )e used in con>unction with the appropriate "hittle Three-D0 Four-D or Four-/ documentation2 The interface documentation will not tell ,ou how to use the "hittle pac+a es )ut it should ive ,ou sufficient information to understand how to prepare ,our input model correctl, for Three-D0 Four- D and Four-/2

3 Introducing Three-dimensional Pit Optimisation

A num)er of methods have )een descri)ed which aim to find the optimal outline for an open pit2 ?ot all methods can uarantee to find the a)solute optimum2 One method which is uaranteed to find the optimal pit was descri)ed ), 1erchs and @rossman in 56A80 it is on this method that )oth "hittle Three-D and Four-D are )ased2 The optimal pit outline is the three-dimensional pit outline which0 if mined out0 would ive the ma$imum currenc, return whilst still o)e,in the re-uired pit slope constraints2 Consider that ever, cu)ic metre of roc+ in a eolo ical resource model has an Intrinsic Value2 If the roc+ is of a sufficientl, hi h rade0 then the value is the price o)taina)le for the metal which can )e e$tracted from it0 less the costs of minin and processin it2 If the roc+ is waste0 then the value is ne ative and numericall, e-ual to the cost of minin it2 The optimal pit outline is thus the outline which ma$imises the total intrinsic value of the roc+ included in the outline whilst o)e,in the pit slopes2 In other words0 the optimal pit outline will mine ever,thin that is worth minin 2 "e mine ever, )it of ore which at an, time in the minin process can )e mined at a profit2 ?othin can )e added to or ta+en awa, from the optimal outline which will increase the value without )rea+in the slope constraints2 The 1erchs-@rossman method is a mathematical search techni-ue which wor+s from >ust two sources of information2 The first is a 7value9 model2 This ives the intrinsic value of each of a set of re ular rectan ular )loc+s which completel, fill the space under consideration for minin 2 The second is a list of relationships )etween these )loc+s2 These relationships are +nown as 7arcs92 #ach arc oes from one )loc+ BAC to a second )loc+ B*C and indicates that0 if A is mined0 then * must )e mined first to uncover A2 The reverse does not hold since * can )e mined without distur)in A2 This list of arcs encapsulates the re-uired pit slopes2 @iven these two sources of information0 a computer pro ram )ased on the 1erchs-@rossman al orithm will fla each )loc+ in the model as )ein inside or outside the optimal pit outline2 An, feasi)le pit outline has a Currenc, Value2 *, feasi)le0 we mean that it will o)e, safe slope re-uirements2 The optimal pit outline will )e the pit with the hi hest Currenc, Value2 &evenue can )e calculated from ore tonna es0 rades0 recoveries and product price2 Price is often the )i est un+nown )ut0 in order to desi n a pit0 ,ou must assume some price2 In eneral0 if the product price is increased0 the optimal pit ets )i er2 Similarl,0 if the costs are escalated0 the optimal pit ets smaller2 If ,ou use steeper slopes0 the optimal pit will enerall, et deeper2 Once all of these factors are fi$ed0 there is onl, one optimal pit outline2

Costs ma, )e a comple$ issue )ut with care0 ,ou can calculate the costs of minin and processin to a reasona)le level of accurac,2 Consider a simple two-dimensional e$ampleD

Tonnages for possi!le pit outlines

Pit Ore 5 ' : E 8 A F G

8(( 508(( 508(( '0((( '08(( :0((( :08(( E0((( 6(( 50A(( '08(( :0A(( E06(( A0E((

"aste 5(( E(( Total

A(( 50E(( '0E(( :0A(( 80((( A0A(( G0E(( 5(0E((

"
Pit

Pit #alues for ore $%2&''(T and )aste $%1&''(T

5 ' : E 8 A F G

Value 6(( 50A(( '05(( '0E(( '08(( '0E(( '05(( 50A((

The raph BFi ure 'C shows that small deviations from a desi n which is not optimal BAC can have si nificant effects on pit value2 @enerations of minin en ineers have e$perimented with small chan es to improve their desi ns2 Providin ,ou start from an optimal outline B*C0 small deviations in the si=e of the pit will have ne li i)le effect on the value of the pit2 The lesson here is that providin ,our practical desi n is )ased on the optimal outline0 fine-tunin will )e a waste of time2 4ou will ain far more if ,ou e$amine the sensitivit, of the optimal pit to chan es in costs or product price2 Open pit optimisation is an iterative process2 Durin sensitivit, wor+0 ,ou will e$plore the economic and slope sensitivit, of the resource model2 4ou will et a feel for the eneral scale of minin and hence the operatin costs2 4ou will then decide appro$imatel, where the haul roads are oin to )e and then ad>ust the slopes in those re ions to the avera e slopes re-uired2 There is more to pit optimisation than creatin a eolo ical resource model0 eneratin a value model and optimisin it0 and then doin the final desi n2

* Intrinsic #alue
Three-D e$pects its input model to comprise the intrinsic value for each )loc+ in the model2 S%&PAC' converts a strin model into a re ular )loc+ model and in turn applies costs0 recoveries and a metal price to calculate the intrinsic value for each )loc+2 The input model for Four-D is a re ular )loc+ model which holds data on tonna e and metal content for each parcel of material in a )loc+2 Four-D calculates the intrinsic value of each parcel and overall )loc+ internall,2 There are two )asic rules to o)serve when calculatin the value of a )loc+ for o)servation purposes2

,irst -ule

Calculate the )loc+ value on the assumption that it has )een uncovered and that it will )e mined2 "hen we mine a )loc+0 is the rade of an, mineralised material in the )loc+ hi h enou h to ma+e it worth processin ;

!a+e no allowance for avera e strippin rations )ecause this is precisel, what the pro rams sort out2 Ta+e no notice of 9)rea+-even9 cut-offs2 These are useful for manual desi n )ut inappropriate for optimisation purposes2 For more information on cut-offs0 refer to the section on model preparation2

Second -ule

Include an, on- oin costs which would stop if minin stopped2 This is )ecause when the pro ram is decidin whether to mine a )loc+0 it is effectivel, decidin whether to e$tend the life of the mine2 It must therefore )e a)le to pa, for all the costs involved in e$tendin the life of the mine2 Incremental costs such as fuel costs0 wa es etc2 must o)viousl, )e included in the cost of minin or processin 0 whichever is involved2 Overhead costs which will stop if minin stops must also )e included2 How these should )e divided )etween minin and processin costs is not alwa,s eas, to determine2 Overhead costs which will not stop if minin stops Bsuch as the repa,ment of a )an+ loanC should not )e included2 The, are in effect alread, spent2 If the pit ,ou desi n will not cover these costs0 then the pro>ect should not proceed as no amount of desi n or redesi n will ma+e it profita)le2 There are a num)er of wa,s of writin an e$pression for the value of a )loc+2 The one that "hittle Four-D and the S%&PAC' to Three-D interface use is as followsD where the part of the e$pression in parentheses is repeated for each separatel, minea)le mineralised parcel in the )loc+ for which it is positive0 and whereD

&OC. O&# !#TA1 &#COV#&4 P&IC# COST! COSTP

The total tonna e of the )loc+ - not >ust waste2 The tonna e of a particular parcel of mineralised material2 The num)er of units of product in the parcel - this is not the rade2 The proportion of product recovered when the material is processed2 The amount of currenc, o)taina)le for a unit of the product2 The cost of minin and removin a tonne of waste2 The difference )etween the cost of minin and processin ore0 and the cost of minin waste2 BThe difference )etween the cost of minin ore and the cost of minin waste is included in COSTP2C Air )loc+s must also )e iven a value so that the optimisation pro ram can differentiate them from waste2 Air )loc+s have a value of =ero2 A )loc+ can comprise parcels of ore which have a positive value and are worth treatin )ut the )loc+ can still have an overall ne ative value2 However this ne ative value will )e less ne ative than if the )loc+ were mined wholl, as waste2 The si=e and shape of an ultimate pit outline is affected more ), the economic conditions when minin ends than the economic conditions when minin starts2 If the mine life is to )e onl, a ,ear or two0 then ,ou can pro)a)l, predict the economic conditions reasona)l, well2 Otherwise ,ou have a pro)lem2 4ou can see from formula B5C a)ove that there are three relevant economic varia)les2 These are P&IC#0 COSTP and COST!2 This of course assumes that the eolo ical resource model has

)een developed ), a competent person with appropriate and ade-uate information2 If we are to do a thorou h anal,sis and e$plore the possi)le effects of var,in economic varia)les0 ,ou could loo+ at a ran e of values0 sa, 5(0 for each of these varia)les2 This re-uire 5((( individual optimisations2 This is a dauntin tas+2 The computation involved in the 5((( optimisations should not )e ta+en li htl,0 however it is possi)le2 The most si nificant pro)lem ,ou would )e faced with is how to compare and anal,se 5((( different pit outlines2

/ Time - The ,ourth 0imension

Four-D ives ,ou the a)ilit, to e$amine ,our eolo ical resource in the fourth dimension0 that is ,ou can consider the effects of se-uencin and schedulin 2

1' The effect of time on the 1alue of mone2

A unit of currenc, we have toda, is more valua)le to ,ou than the unit of currenc, that someone is oin to ive ,ou in a ,ear9s time2 There are several reasons for this some of which are as followsD

Inflation will reduce the value of ne$t ,ear9s unit of currenc, - it will )u, less2 If ,ou et the unit of currenc, now0 there is no ris+ of somethin oin wron and ,ou not ettin it2

If ,ou have not ot the unit of currenc, now0 ,ou ma, need to )orrow it and pa, interest on it until ,ou can repa, it2

The accepted method to allow for this is to discount ne$t ,ear9s unit of currenc, ), a certain amount and cumulativel, thereafter for ,ears into the future2 This means that ,ou discount future revenues and costs ), a particular discount rate and reduce them all to a 7net present value9 or ?PV2 There are two discount rates2 The first is the 9notional9 discount rate which is applied to actual revenues and costs which are li+el, to occur2 In other words the revenues and costs will follow the inflation rate2 Thus the notional discount rate will include an allowance for inflation0 and it would )e correct to use this provided ,ou inflate all ,our costs and revenues for future ,ears2 #conomists and politicians spend their lives uessin at the future inflation rate - what chance have ,ou0 a hum)le en ineer0 of uessin ri ht< For this reason it is easier0 and most importantl, pro)a)l, more accurate0 to wor+ out revenues in toda,9s unit of currenc, and then to use the second discount rate which is referred to as the 9real9 discount rate2 The real discount rateG ta+es no account of inflation2

11 The effect of mining se3uence on the optimal pit outline

This is )est e$plained with the aid of Fi ure 52 There are two e$treme approaches that ,ou can ta+e when schedulin a pit2 The first we will refer to as 9flat minin 92 This is where ,ou mine all the top )ench to the final pit outline0 then move ,our e-uipment down to the ne$t )ench and repeat the e$ercise and so on2 "aste at the top of the outer shell is mined earl, and the cost is discounted less than the revenue from the correspondin ore which is mined at the end of the pit life2 The optimal pit for flat minin is thus enerall, smaller than would )e indicated ), simple optimisation usin toda,9s costs and revenues2 The second approach is 9incremental9 minin where each shell is mined in turn and thus the related ore and waste is mined in appro$imatel, the same time period2 Fi ure 5 has introduced the concept of incremental or 9nested9 pits2 A series of value models ma, )e prepared for0 sa,0 a ran e of product prices2 The outlines o)tained from puttin these throu h a three dimensional optimiser will form a set of nested pits2

The onl, pro)lem here is that ,ou want to maintain ,our usual hi h professional standards and will therefore need to loo+ at a ran e of processin and minin costs0 and all the interactions )etween the different costs2 4ou have >ust presented ,ourself with the 5((( pit pro)lem2 A modified )loc+ value formula can reduce the si=e of the pro)lem2 If0 for a iven set of economic value0 ,ou calculate a set of )loc+ values0 in sa, dollars0 ,ou will et a particular value for each )loc+2 If ,ou then run these values throu h an optimiser0 ,ou will et a particular pit outline2 If ,ou then calculate the )loc+ value in cents0 the numeric value allocated to each )loc+ will )e one hundred times )i er2 If these values throu h an optimiser0 the pit outline we et will )e identical to the one we ot when we used dollars2 ?ow0 if ,ou calculate the )loc+ value in &ou)les2222 In other words0 we can use an, mone, as our unit of currenc, and ,ou will alwa,s et the same optimal pit outline2 "hat matters is the ratios )etween the )loc+ values0 not the values themselves2 @iven that P&IC#0 COSTP and COST! are each amounts of mone, Be2 2 dollarsHo=0 dollarsHtonne etc2C0 then an, of them could )e used as a unit of currenc,2 However0 there are particular advanta es to usin COST! as the unit of currenc,2 If we divide all the terms of our Value formula ), COST!0 it )ecomesD There are now onl, two economic varia)les2 These are BP&IC#HCOST!C and BCOSTPHCOST!C2 ?ow ,ou onl, need to do 5(( optimisations to maintain ,our professional inte rit,< If ,ou e$amine the economic varia)le BCOSTPHCOST!C more closel,0 ,ou will see that it is the ratio )etween the cost of processin ore and the cost of minin waste2 Four-D calls this cost ratio the C&ATIO and e$perience will show that this ratio is surprisin l, sta)le with time2 The )loc+ value formula )ecomesD If fuel costs increase0 the haula e costs for ore and waste will )oth )e affected2 If la)our costs chan e0 the, are li+el, to chan e for emplo,ees wor+in in )oth the processin plant and the mine2 Althou h the C&ATIO is an economic varia)le0 it is not reall, an economic un+nown if ,ou )rea+ ,our costs down correctl,2 For man, purposes the C&ATIO can therefore )e re arded as fi$ed2 This is ood news )ecause now instead of 5(( optimisations0 ,ou need onl, do 5(2 &emem)er that althou h the C&ATIO is sta)le with time0 it can certainl, var, with position in ,our pit2 It is hard to relate to the economic varia)le BP&IC#HCOST!C2 However0 if ,ou invert it to ive BCOST!HP&IC#C this is the amount of product ,ou will need to sell to pa, for the minin of a tonne of waste2 It will henceforth )e referred to as the metal cost of minin B!COST!C2 It has the units of a rade0 )ut this is not the rade which has an,thin to do with ore rades2 In a lar e open pit old mine for e$ample0 it ma, )e (25 rams per tonne or (2(: ounces per ton2 The final form of the )loc+ value formula )ecomesD C&ATIO is reasona)l, sta)le !COST! is dependent on prices and costs Althou h we have reduced the num)er of un+nown varia)les from three to one0 there is no loss in eneralit, )etween formulae B5C and BEC *, usin formula BEC for the )loc+ values and steppin !COST! over a suita)le ran e of values ,ou can enerate a series of value models which can then )e optimised with a three dimensional optimiser to produce a series of pit shells similar to the ones in the simple e$ample0 )ut for a real eolo ical resource model2 Four-D automates this steppin and optimisin process and the wa, that ,ou can use the resultant pit shells is what sets Four-D apart from Three-D2

Since each shell honours the re-uired pit slopes0 ,ou can use the shells to indicate technicall, feasi)le minin se-uences2 Four-D will allow ,ou to anal,se three separate se-uencin scenarios2

12 Worst 4ase Scenario

In this0 Four-D completes the minin of each )ench )efore startin the ne$t )ench2 This is well +nown to )e the worst wa, of ma+in mone, out of a pit0 althou h in some small pits0 it ma, )e the onl, practical course availa)le2 Since it is the worst0 it sets a floor on the ?PV2 The precise se-uence used ), Four-D is to mine the top )ench of the smallest of the pit shells0 followed ), the strip )etween the smallest and the ne$t smallest etc2 until the ultimate pit has )een reached2 After this the process is repeated for the second )ench2

13 Best 4ase Scenario

In this0 Four-D mines the pit as a series of narrow push )ac+s0 often thirt, or fort,2 This is rarel, ph,sicall, feasi)le0 )ecause the push-)ac+s are usuall, far too narrow2 *ut0 if it were practical0 and if minin this wa, did not increase the costs0 it would provide ,ou with the )est possi)le earl, cash flow2 Since it is the )est0 it sets the ceilin on the ?PV of the pit2 "hen the worst and )est ?PVs are ver, similar0 as the, ma, well )e in a small mill-limited operation with a fairl, even rade0 then ,ou can conclude that the minin se-uence has ver, little effect on the ?PV and can therefore )e i nored for desi n purposes2

Specified Push-Back Scenario

The actual minin se-uence will lie somewhere )etween the worst and )est case se-uencin scenarios2 Four-D will allow ,ou to select particular pit shells as push-)ac+s2 Thus practical push-)ac+s can )e considered and the effect the, have on the cash flow and ?PV assessed2

1" 4ost 5ssumptions

There are three )asic rules to follow when ,ou calculate the value of a )loc+ for optimisation purposes2

1* The ,irst -ule

The value must )e calculated on the assumption that the )loc+ has alread, )een uncovered2 In other words no allowance should )e made for the cost of the strippin re-uired to access the )loc+0 )ecause this is precisel, what the optimiser calculates2 In particular0 an, cut-off used to define ore should reflect the cost of processin and an, e$tra cost of minin the )loc+ as ore rather than waste0 )ut not the cost of strippin 2 If an allowance for strippin is included in the costs0 the strippin will )e paid for twice2

1+ The Second -ule

The value should )e calculated on the assumption that the )loc+ will )e mined2 So0 if the )loc+ contains some ore that could )e profita)l, )e processed as well as some waste0 the value of the ore should )e added in0 even if the net value of the )loc+ is still ne ative2 The optimiser will chose to mine such a )loc+0 )ut if it has to mine it to et to a more valua)le )loc+0 the ore will help pa, for the strippin 0 as it would in practice2

1. The Third -ule

An, cost that would stop if minin stopped must )e included in either the cost of minin or the cost of processin 2 Conversel,0 an, cost that would not stop if minin stopped must )e e$cluded2 This is discussed in more detail later on2

1/ 6o) to 4alculate 4osts

It was stated a)ove that an, cost that would stop if minin stopped must )e included in the cost of minin or the cost of processin 2 The reasonin )ehind this is that0 when the optimiser adds a )loc+ to the pit outline0 it ma, effectivel, e$tend the life of the mine2 If the e$tra overheads involved in this are not included when calculatin the )loc+ value0 then the addition of a positive )loc+ ma, actuall, reduce the value of the pit2 All da,-to-da, costs such as wa e and fuel costs must0 of course0 )e included0 and it is usuall, clear whether the, should )e included in the minin or the processin cost2 However0 there are some costs where the correct treatment is not so o)vious2 Some e$amples of the handlin of various costs ma, )e helpfulD

2' Processing Mill

Consider a processin mill and mine infrastructure that costs I8( million to )uild and commission2 If the mine were to shut down0 for whatever reason on da, '0 the mill and infrastructure would have a certain salva e vale0 sa, I:( million2 In this case0 I'( million has one forever2 It is an up-front or sun+ cost that must )e su)tracted from an, optimised value of the pit itself2 4ou can deal with the I:( million in two wa,s2 If we assume that there will )e an on- oin pro ram of maintenance and capital replacement that will +eep the salva e value of the mill close to I:(0 then the I:( million is theoreticall, recovera)le when the mine closes0 and so is not a cost2 However0 the maintenance and periodic capital replacement e$penses are costs for these purposes0 )ecause the, would stop if minin stopped2 The, should )e avera ed and reduced to a fi ure per ton processed2 This fi ure should )e added to the direct costs of processin 0 such as wa e and fuel costs2 Alternativel,0 ,ou could assume that onl, essential maintenance will )e done and that the salva e value of the mill will pro ressivel, decline2 In this case0 the e$pected rate of decline should )e added on a per ton )asis to the essential maintenance0 wa e and fuel costs2 ?ote that the rate of decline is not necessaril, the rate of depreciation used ), accountants2 In most cases0 the depreciation rate is set ), a ta$ation re ime0 and ma, reduce the )oo+ value to =ero when the salva e value is clearl, not =ero2 To reduce costs to a per ton )asis0 assumptions have to )e made a)out the production rate2 If the si=e of the pit produced ), the optimisation ma+es these assumptions inappropriate0 then the costs must )e recalculated and the optimisation done a ain2 4ou should set up all ,our cost calculations on a spreadsheet0 this ma+es the recalculation much easier2

21 Trucks
If the e$pected life of the mine is shorter than the operatin life of a truc+0 then the truc+ purchase must )e treated in the same wa, as the cost of the mill0 e$cept that it will pro)a)l, )e included in the minin rather than the processin cost2 If the life of the mine is much lon er than the life of a truc+0 then the truc+s will have to )e purchased pro ressivel, to maintain the fleet0 and such purchases will stop if minin is stopped2 Conse-uentl, the cost of purchasin truc+s should )e avera ed out over the life of the mine and reduced to a cost per tonne2 %nless the life of mine is e$pected to )e ver, lon 0 some compromise )etween the a)ove two approaches is usuall, re-uired2 Contract minin companies must ta+e these factors into account when -uotin for a >o)0 and it sometimes useful to thin+ as the, do when wor+in out the cost of a compan, owned fleet0 e$cept that the contractor9s allowance for profit is not included2

22 5dministration 4osts
On-site administration will usuall, stop if minin stops2 The, must therefore )e included somewhere in the costs2 The -uestion is should the, )e included in the minin or processin cost; The answer depends on whether production is limited ), minin or ), processin 2 %suall,0 it is limited ), processin 0 and in this case onl, the minin of an ore )loc+ e$tends the life of the mine2 The ore )loc+ values should therefore include an allowance for on-site administration costs2 The cost is0 in effect0 added to the processin cost per ton2 Conversel,0 if production is limited ), minin 0 as in a heap leach operation0 ever, )loc+ that is mined e$tends the life of the mine0 so that on-site administration costs should )e added to the minin costs2 Head office administration costs ma, or ma, not stop if minin stops0 and thus ma, or ma, not )e included2

23 Bank 7oans for 8p-,ront 4osts

&epa,ment Bprincipal and interestC of a )an+ loan ta+en out to cover up-front costs will have to continue whether minin continues or not2 It should therefore not )e included in the costs used for calculatin )loc+ values2 Of course these pa,ments have to come from the cash-flows of the mine2 If the mine is not oin to produce enou h cash flow to cover them0 the pro>ect should not proceed2 These repa,ments should not )e introduced as costs in an attempt to 7improve9 the optimisation2 The result will in fact )e -uite the opposite - a smaller pit with a smaller total cash-flow2 Althou h the )an+ loan repa,ments themselves are not included0 some items that the loan was used to pa, for ma, )e included as ,ou will see )elow2

Bank 7oans for -eco1era!le 4osts

If mone, is )orrowed from the )an+ for da,-to-da, wor+in capital or for items such as the I:( million discussed in the mill e$ample a)ove0 then the salva e value can )e realised and repaid to the )an+ if minin stops2 Conse-uentl, the interest paid on such a loan is a cost that stops if minin stops2 It should therefore )e treated as an overhead li+e the on-site administrative costs discussed a)ove2

2" 9rade 4ontrol 4osts

It is usuall, necessar, to do rade control wor+ on waste as well as ore2 In this case rade control costs appl, to waste costs too2 If onl, some of the waste is rade controlled0 then ,ou can either load the cost of those particular waste )loc+s0 or ma+e an estimate of the tonnes of such waste per tonne of ore0 and load the cost of minin ore2

2* Support - 4a!le Bolts

The cost per tonne is related to the pit si=e0 which has to )e calculated ), the optimiser2 Then a cost per unit of area of wall can )e transformed into a cost per tonne of waste2 This is an iterative estimate0 )ut fortunatel, costs per tonne are usuall, low2 Cop,ri ht J '((A Surpac !ine$ @roup

The String Model Whittle Interface 2+ O1er1ie) of the S8-P542 to Whittle Interface
The "hittle Interface is part of S%&PAC'2 4ou will use it for convertin a S%&PAC' eolo ical resource model into strictl, formatted input files for either Three-D or Four-D pit optimisation pac+a es2 4ou will also use it to e$tract S%&PAC' contour outline files of an optimal pit from the result files of either a Three-D of Four-D pit optimisation2 The flow of data is summarised in the followin process chart2

"hether ,ou are oin to use Three-D or Four-D0 ,ou will perform the conversion process in two sta es2 The first sta e0 which is called Model Blocking0 is common to )oth the Three-D and Four-D interfaces2 4ou use Model Blocking to 7re ularise9 ,our eolo ical resource model2 This process cuts ,our model up into re ular0 e-uidimensional )loc+s which hold data on volumes and rades2 #ach )loc+ is iven an I0 K0 . reference address2 This is instead of the ?orthin B4C0 #astin B/C and #levation BLC used ), S%&PAC'2 The first )loc+ at the )ottom front left hand corner of ,our model will have an I0K0. address of 505052 The re ularisin pro ram in S%&PAC' ta+es a strin file for each )ench as input and outputs a file for each )ench with the file e$tension 72)l+92 The first sta e also opens a pro>ect file which stores ,our +e,)oard input and is used for su)se-uent sta es in the "hittle Interface2 If ,ou are oin to use Three-D0 the second sta e of ,our conversion process will )e to calculate a )loc+ value model file from ,our 72)l+9 files2 As part of the process0 ,ou will input economic parameters and )ul+ densities via ,our +e,)oard2 This process is called Surpac to Whittle T6-::-0 and the parameters ,ou entered at the +e,)oard will )e com)ined with the volume and rade data in ,our 72)l+9 files to calculate a value for each )loc+2 As an ancillar,0 this process updates ,our pro>ect file with ,our additional +e,)oard input0 so ,ou can rerun this sta e without ret,pin all the information2 Surpac to Whittle ,O8--0 is the step ,ou will use to convert ,our 72)l+9 level files into a sin le Four-D input model file2 In addition to the 72)l+9 files0 ,ou will use the su)re ion outlines stored in the e$tent strin file and ,our other input parameters B)ul+ densities0 cost ad>ustment factors etc2C2 It outputs the FO%&-D input model file which will )e named after ,our pro>ect and )e iven the file e$tension 72mod92 The pro ram will also add ,our input data to the pro>ect file0 thus this file serves a useful dual purpose since it now stores all the relevant parameters throu h the S%&PAC' process of ,our pro>ect2

The final sta e which is the reverse of the process0 is called Whittle to Surpac2 It ta+es output model files from either Three-D or Four-D and produces strin files of the optimal pit outlines2 4ou then use these contours to uide ,ou in final pit desi n usin the S%&PAC' interactive raphical pit desi n module2

2. Model Preparation
This is an important and often lossed over area2 4ou have )een presented with0 or developed0 a eolo ical resource model which must )e converted into a re ular )loc+ model for optimisation2 6int; To conform to uidelines issued ), most professional institutions0 it is important that ,ou onl, include mineralised material which is classified as 7measured9 or 7indicated9 in the pit optimisation process2 !aterial which falls into the 7inferred9 cate or, should )e i nored for the purpose of pit optimisation2 The eolo ical resource model files developed in S%&PAC' can ta+e two forms2 The files ma, )e eitherD

a ran e of strin files formin a strin model3 or a ran e of strin files formin centroids of a )loc+ model2

2/ The :<tent String

*efore ,ou run Model Blocking on ,our eolo ical resource files0 ,ou must prepare an e$tent strin 2 The e$tent strin is stored in the e$tent strin file2 The e$tent strin file will also store ,our S%&PAC' su)re ions which will )e e$plained in more detail later2 The e$tent strin must alwa,s o)e, the followin rulesD

)e strin num)er 53 and )e rectan ular3 and )e a closed0 cloc+wise se ment3 and it must )e normal to ,our model rid3 and the difference )etween the ma$imum and minimum / coordinates must )e an inte ral multiple of the )loc+ si=e in the / direction that ,ou specif, for the model3 and the difference )etween the ma$imum and minimum 4 coordinates must )e an inte ral multiple of the )loc+ si=e in the 4 direction that ,ou specif, for the model3 and contain the whole eolo ical resource model on each )ench2

The e$tent strin forms a )oundar, around the model and is used to provide a reference frame for each I0K0. address2 The e$tent strin limits the si=e of the model and allows the Whittle to Surpac function to e$tract /040L contour strin files from the pit optimisation result model file2 6int; One wa, to ensure that ,our e$tent strin will enclose the whole optimal pit at each level is to ma+e it ver, lar e2 However this will increase the num)er of )loc+s in the model unneccesaril, and will increase ,our computer9s wor+load dramaticall,2A )etter wa, is to -uic+l, enerate an outer)ound pit in the 9-5P6I4S Pit 0esign in S%&PAC'2 The outer)ound pit should )e desi ned 9)ottom to top9 and to mine ever, )it of mineral in ,our resource model2 It is the optimal pit which is e-uivalent to an infinite metal price2 #nsure that the e$tent strin completel, encloses the top crest of the outer)ound outline2 If ,ou follow this method0 ,our optimal pit will alwa,s )e contained within ,our model no matter how hi h ,ou set ,our metal price0 or how low ,ou set ,our minin costs2 If ,our optimal pit is not wholl, contained within the e$tent strin 0 Three-D and Four-D ma, form incorrect outlines with vertical pit walls at the ed es of the model2

3' String Outline Model

"ithin the e$tent strin )oundar,0 all material in ,our eolo ical resource model is assumed to )e waste unless it is comprised of mineral or air strin closed se ments2

The air strin s are formed ), e$tractin closed contours from ,our topo raph, DT! in the 0TM TOO7S module2 4ou then outersect ,our e$tent )oundar, on each level with the appropriate surface contour usin the Pol2gon Intersection function in ST-I=9 TOO7S2 The result should )e one or more cloc+wise air se ments which can then )e appended to ,our eolo ical resource model in ST-I=9 TOO7S 4om!ine and -enum!er Strings2 Alwa,s reserve a consistent strin num)er to represent the air in ,our models - strin num)er 6( is commonl, used to represent air2 O)viousl, it will have a )ul+ densit, of =ero2 6int; If the material within ,our e$tent strin is not defined as mineral or air0 it must )e 7default waste92 &emem)er that in some cases ,ou could have waste material which has characteristics which are distinct from default waste2 This could )e an interface )etween o$idised and non-o$idised waste0 or perhaps an e$istin surface waste dump which is ver, li+el, to have a lower )ul+ densit, than insitu waste2 4ou should model the waste dump with a different strin num)er2

31 4entroid Model
This is usuall, )ased on the output from the Ordinar2 >riging function in the 9:OST5TISTI4S module2 It is alread, in the form of a 7re ular9 )loc+ model in that each centroid represents )loc+s which have consistent dimensions in )oth the /0 4 and L directions2 The Model Blocking function e$pects to find a centroid for mineral0 non-default waste or air2 Areas of ,our strin model which are not occupied ), an, of these cate ories will )e automaticall, assumed to )e default waste2 It is necessar, to form a centroid )loc+ model of ,our air model and to then append this model to ,our mineral and non-default waste model2 6int; Create ,our air contours as for the strin outline model2 Then form a re ular rid of centroids for each level in ,our model2 %se 9-I0 TOO7S to do this tas+ and set ,our ma$imum and minimum rid coordinates to )e the same as ,our e$tent strin 2 #$tract a sin le strin file from ,our rid2 %se the ST-I=9 TOO7S 5ppl2 Boundar2 to determine which centroids fall within the closed air strin se ments2 Perform String Maths to set the L value of each air )loc+ to the strin file 9id92 Then append and renum)er Bto sa, strin 6(C these air centroids to ,our mineral and non-default waste centroid model usin the 4om!ine and -enum!er Strings function in ST-I=9 TOO7S2 Important; S%&PAC' will onl, allow ,ou to re ularise a centroid )loc+ model into )loc+s which have dimensions which are entirel, consistent with the centroid )loc+ model2

32 S8-P542 Su!regions
S%&PAC' su)re ions can )e different from "hittle Su) &e ions2 "hittle Three-D uses su)re ions to define portions within the input model which have distinct slope criteria and possi)l, distinct default waste minin costs2 "hittle Four-D uses su)re ions to define portions within the input model which have distinct slope criteria and default waste tonna e2 In the S%&PAC' to Three-D interface0 su)re ions descri)e re ions within the model which have distinct economic parameters Bminin and processin costs0 processin methods0 metal recoveriesC and )ul+ densit, parameters2 In S%&PAC' to Four-D interface0 su)re ions descri)e re ions within the model which have distinct economic parameters B!inin and Processin Cost Ad>ustment FactorsC and )ul+ densit, parameters2 S%&PAC' su)re ions are not used to define eotechnical characteristics2 S%&PAC' su)re ions ma, in fact )e the same as "hittle su)re ions0 however ,ou are enerall, more li+el, to use additional S%&PAC' su)re ions as a tool to et )etter control while settin up economic parameters2

33 Summar2

The purpose of definin a su)re ion is to cater for possi)le chan es in minin or processin costs in different parts of the model2 In the S%&PAC to Three-D interface0 use classified model strin s andHor su)re ions to define different minin and processin costs and )ul+ densities2 In the Four-D interface0 use su)re ion strin s to define different minin and processin cost ad>ustment factors2 The "hittle structural arc pro rams0 1@ST and FDST use different su)re ions to descri)e re ions in the model which have different slopes2 It is )est to consider a set of su)re ions to appl, minin and processin costs etc2 and another set to define slope conditions2 Hereafter0 discussions and mention of su)re ions within this documentation0 will refer to S%&PAC' su)re ions as distinct from "hittle su)re ions2

0efining Su!region Strings

Su)re ion strin s are stored in the #$tent Strin File alon with the #$tent Strin 2 4ou will normall, prepare ,our su)re ion strin s in 9-5P6I4S2 S%&PAC' su)re ions must o)e, the followin rulesD

)e rectan ular and closed0 cloc+wise strin se ments3 and su)re ion strin s must )e num)ered )etween ' and :'0((( in the e$tent file3 and if su)re ions enclose the whole area of the model0 the strin can simpl, )e a replica of the e$tent strin 53 and need not encompass the whole pit )ut if there e$ists more than one su)re ion0 each )loc+ must )elon to onl, one su)re ion3 and

!odel *loc+in re-uires that the /-4 dimensions of the )loc+s )e specified )efore the process can commence2 The method for )loc+in a model is descri)ed as follows2 The point of minimum northin and eastin of the e$tent strin )ecomes the centroid for the first )loc+2 %sin the /-4 dimensions specified0 a )loc+ is created around this point2 The )loc+in then continues in the /- direction at a distance $ from the previous point until the whole row is complete Brefer to Fi ure :C2 The first point of the second row commences at a point , a)ove the first point of the first row and so on until the whole e$tent has )een )loc+ed2

If a )ench in ,our model is made up of several su)re ions0 ,ou have to arran e the su)re ions in such a wa, so that each )loc+ is contained in onl, one su)re ion2 ?otice from the previous dia ram BFi ure :C that there is overhan of half the )loc+ si=e on each side of the e$tent strin 2 The same will )e true for each su)re ion0 which means there must )e a ap separatin each su)re ion e-ual to the )loc+ dimensions specified2 This is demonstrated in this dia ram BFi ure EC of three su)re ions defined ), su)re ion strin num)ers '0: and E2

?oteD the outer ed es of strin s '0 : and E coincide with strin 5 Bthe e$tent strin C0 )ut for clarit, the, are shown sli htl, apart in the dia ram2 "hen the )ench is )loc+ed with these three su)re ions0 ,ou will end up with )loc+s as illustrated in the dia ram BFi ure 8C2 Here ,ou can see that su)re ions '0: and E have a 7halo9 or 7frame9 around them )ut the overall effect is that the whole )ench has )een )loc+ed2 ?otice the ap )etween su)re ion strin ' and su)re ion strin s : and E is e$actl, the width of one )loc+2 Similarl,0 )etween su)re ion strin : and su)re ion strin E0 the ap is the len th of e$actl, one )loc+2 #ach )loc+ centroid )elon s to onl, one su)re ion2

A future release of the "hittle Interface in S%&PAC' will allow the use of pol, ons to define su)re ions strin s to set ,our minin and processin costs2 This will ive ,ou much more fle$i)ilit,2 However rectan ular su)re ions will still )e used to define the slope conditions since FDST and 1@ST pro rams e$pect rectan ular su)re ions2 Cop,ri ht J '((A Surpac !ine$ @roup

Model Blocking
,unction =ame?s@ ;

!OD#1 *1OC.I?@

The first step in convertin ,our S%&PAC' eolo ical resource model to a Three-D or Four-D input model is to 7re ularise9 the model2 This cuts ,our model up into re ular0 e-uidimensional )loc+s2 #ach )loc+ is iven an I0 K0 . reference address2 This is instead of the ?orthin B4C0 #astin B/C and #levation BLC used ), S%&PAC'2 The first )loc+ at the )ottom front left hand corner of our model will have an I0K0. address of 505052 The )loc+ re ularisin pro ram in S%&PAC' is Model Blocking2 It ta+es a strin file for each )ench as input and outputs a file for each )ench with the e$tension 72)l+9

3" What Model Blocking 0oes 3* If 2our model is a string model

Ima ine ,ou are la,in a re ular rid over each )ench of ,our entire model within the e$tent strin 2 The cells of the rid are centred on enerated )loc+ centroids0 and are whatever si=e ,ou elect2 The pro ram determines how much mineralised material or air lies in each rid cell2 It then stores the volume and rade of each 7parcel9 of material as a record in the appropriate 72)l+9 file2 An, remainin volume not covered ), defined mineralised =ones or air0 is written to the 72)l+9 file as waste0 with a rade of =ero2 This means that if ,our strin model is -uite comple$ or ,our rid si=e is lar e0 ,ou will finish up with multiple parcels of material in each cell or )loc+2 This is fine0 )ecause when either S%&PAC' to Three-D or Four-D FDOP calculates the intrinsic value of a )loc+0 the, first calculate the intrinsic value of each of these parcels2 Dependin upon the rade0 and economic factors a parcel will )e mined as ore or waste2 If the parcel is mined as ore it will have a positive value2 If it is mined as waste it will have a ne ative value2 A )loc+ can contain parcels which are mined as ore )ut the net intrinsic value of the )loc+ ma, still )e ne ative Bhowever it will )e less ne ative than if the whole )loc+ were wasteC2

3+ If 2our model is a centroids !lock model

The )loc+ dimensions that ,ou select in the interface must match up with the centroid )loc+ model dimensions enerated ), 9eostatistics2 *loc+ volumes which are contained ), the model )ut are not defined in the eolo ical resource model are written to the 72)l+9 file as waste0 with a rade of =ero2 6ints on Block SiAe; It is enerall, accepta)le to perform pit optimisation on a model which comprises around '8(0((( )loc+s2 #$perience has shown that the results achieved from a 5 million )loc+ model are no )etter than one which has '8(0((( )loc+s2 "hen ,ou select ,our re ularised )loc+ dimensions0 ,ou should ta+e this into consideration2 Also ,ou should )e thin+in a)out ,our minimum pit width2 There is little point in producin an optimum outline which has a )ase which is onl, ' metres wide2 If ,ou are usin a centroids )loc+ model0 this will control the si=e of ,our model )loc+s2 However0 if ,ou have too man, )loc+s0 prepare ,our "hittle input model files and then use the "hittle &e)loc+in utilities BFD&* in Four-D and 1@&* in Three-DC to reduce the num)er of )loc+s in ,our model to a sensi)le level2 &efer to the "hittle manuals for more information2

3. The ProBect ,ile

#ach time ,ou run Model Blocking0 ,ou will create or edit a pro>ect file2 The file will )e named after the pro>ect and will )e iven the e$tension 72pr>92 The pro>ect file stores information on )loc+ si=es and later su)re ions0 cost ad>ustment factors and )ul+ densities2 *ut importantl, at this sta e0 it stores the name of the e$tent strin file2 The enerall, accepted namin convention for an e$tent file is e$t52str2 If ,ou stic+ to conventions0 there is more chance of someone else understandin ,our wor+ at a later date2 The first : characters of the 72)l+9 files which Model Blocking creates are ta+en from the first three characters of the pro>ect name2 Hence a pro>ect called 7eure+a9 and with model files called mod5((0'((082str0 will name all the 72)l+9 files eur5((0'((082)l+2 Choose Model Blocking from the S8-P542 C-D Whittle menu to displa, the W6ITT7: P-OE:4T =5M: form2

3/ =ame of ProBect
#nter a convenient pro>ect name2 In eneral0 the pro>ect name identifies the files related to a iven optimisation2 A pro>ect file named after the pro>ect and with a file e$tension 72pr>9 will )e opened2 If the pro>ect file alread, e$ists0 S%&PAC' will insert the values stored in the pro>ect file as defaults in su)se-uent forms2 4ou ma, accept or overwrite these defaults2 Complete the W6ITT7: P-OE:4T =5M: form and choose 5ppl2 to displa, the MO0:7 B7O4>I=9 form2

' =5M: O, P-OE:4T

The pro>ect name defined in the previous form is displa,ed2

1 0efine The -esource Model

#nter the resource model file 7ocation0 Id -ange and String -ange2 The Id -ange of the files0 which also ives the ran e of level elevations in the model2 The ran e must )e specified from the )ottom of the model upwards2 ?oteD These elevations must represent the mid-)ench elevation of each )ench2 The String -ange of the classified resource model files2 S%&PAC' re ards the last strin specified in the ran e as default waste for the model2 S%&PAC' will allocate the last strin in the ran e to waste even if material which relates to this strin is not present in the model files2

2 9rade In Which ,ield FG or 0nnnH

@rades ma, )e stored in the L field or a descriptor field2 #nter L or D50 D' etc2 Althou h rade can )e stored in the L field it is not ood practice to do so since this field should )e reserved for elevation2

3 #aria!le S&9Is
4ou have the choice of usin a fi$ed specific ravit, BS2@2C for a material strin ran e or a strin dependant specific ravit,2 "hen usin the latter0 the varia)le S2@2 must )e stored in a particular Description field for each strin contained in the &esource !odel Bincludin air strin sC2

Which ,ield
If ,ou choose Varia)le S2@9s then ,ou will )e re-uired to enter the description field used to store the S2@2 for each strin num)er2

" Waste S&9&

If ,ou choose Varia)le S2@9s then the specific ravit, for waste is re-uired )ecause "aste will not )e defined in the &esource !odel9s strin files2

* 4entroids Model ?J(=@

#nter 7?9 if the resource model is a strin model or enter 749 if it is a centroids )loc+ model2

+ -egular Block SiAe

Specif, the )loc+ si=es ,ou want to output to the re ularised 72)l+9 files2 If ,ou are usin a centroids )loc+ model0 ,ou have no choice )ut to use the )loc+ si=es inherent in the model2 Complete the MO0:7 B7O4>I=9 form and choose 5ppl2 to displa, a second MO0:7 B7O4>I=9 form2

. =5M: O, P-OE:4T
The pro>ect name defined in the W6ITT7: P-OE:4T =5M: form is displa,ed2

/ 0efine The :<tent String ,ile

Specif, the 7ocation and Id =um!er of ,our e$tent strin file2 The e$tent strin descri)es the )oundar, of ,our model2

"' 0efine The :rror 7og ,ile

Specif, the 7ocation and Id =um!er of the error lo file2 This file will record an, errors which ma, have )een encountered durin )loc+in 2 Clic+ here for a listin of these errors2 If no errors are encountered0 the lo file will not )e created2 4ou will )e warned if errors are encountered2

?oteD If overlaps are detected0 ,ou are advised to fi$ them up and repeat the *loc+in process2 The pro ram is not meant to )e a ri orous overlap chec+er and the sensitivit, to overlaps will depend on the cell or )loc+ si=e that ,ou have selected2 Hence0 the smaller the )loc+ si=e0 the more li+el, overlappin strin s will )e detected2 A sample listin of a 72)l+9 output files is shown )elowD

The first three columns represent the I0 K and . address of the )loc+2 The fourth column shows the strin num)er of the material2 The fifth column displa,s the volume of each parcel and the si$th column shows the rade2 !ost of the )loc+s contain onl, waste0 shown ), strin '(( Ba eneral default waste strin C0 volume of E((( cu)ic metres and a rade of =ero2 *loc+ :F E8 E contains 5(:G cu2m and 8(28 cu2m of strin 5 and ' respectivel,2 The remainin material0 '65528 cu2m is waste2 6int; A common mista+e made ), users is to create the 72)l+9 files in more than one run2 This will not wor+ )ecause the lowest level in the model will )e iven a 7.9 address of 52 The value of 7.9 will )e incremented ), 5 for each level in the model startin with 52 There is a wa, around this however2 !a+e a series of dumm, B)lan+C strin files for each la,er which ,ou have alread, )loc+ed in a previous run2 Cop, ,our pro>ect file to a new name so that the 72)l+9 files which will )e created from the )lan+ strin files will not overwrite those prepared in the previous run2 A sample listin of a 72)l+9 output files is shown )elowD

I 33 34 35 36 36 36 37 37 37 38 39 40 41 41 41 42

J 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45

K 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

String 200 200 200 1 2 200 1 2 200 200 200 200 3 3 200 200

Volume 4000.0 4000.0 4000.0 430.5 293.0 3276.5 1038.0 50.5 2911.5 4000.0 4000.0 4000.0 513.5 1516.5 1970.0 4000.0

Grade .0000 .0000 .0000 .6130 .8800 .0000 .6130 .8800 .0000 .0000 .0000 .0000 1.3310 1.3760 .0000 .0000

Cop,ri ht J '((A Surpac !ine$ @roup

Surpac to Whittle T6-::-0

In this sta e ,ou appl, costs to the rade and volume data which are stored in the 72)l+9 files2 4ou will create an economic model file0 also referred to as a value model file2 The file will )e named after the pro>ect title and will )e iven a file e$tension 72eco92 4ou will also create a "hittle Three-D parameter file stu)0 which contains correctl, formatted information that "hittle needs0 and Surpac +nows a)out2 This file9s name will )e the pro>ect title0 plus a 92par9 e$tension2 Since minin and processin costs and other parameters ma, var, throu hout a model0 ,ou ma, define appropriate su)re ions within the model in which these parameters ma, )e set2 The su)re ions are stored in the e$tent strin file and for each su)re ion0 ,ou must specif, the elevation ran e which a particular su)re ion refers to in the model2 After enterin

eneral data which relate to the whole model0 ,ou input information forD

each su)re ion within the model3 each material within each su)re ion3 and each treatment method within each material2

#ach intersection volume0 or parcel0 is classified into a su)re ion and material t,pe0 )ased on its )loc+ position0 its model strin num)er0 and its rade2 The net value of the various processin alternatives Bif more than one method e$istsC is evaluated and the hi hest value chosen2 If the revenue from a parcel e$ceeds the minimum processin cost0 the parcel will )e treated as ore and processed2 If the processin cost e$ceeds the revenue0 Busuall, )ecause the rade andHor recover, are too lowC0 the parcel will )e mined as waste2 If this occurs0 the parcel will have a ne ative value2 The value of a )loc+ is the sum of the values of the parcels2 ChooseSurpac to Whittle T6-::-0 from the S8-P542 C-D Whittle menu to displa, the W6ITT7: P-OE:4T =5M: form2

"1 =ame of ProBect

#nter the pro>ect name2 Complete the W6ITT7: P-OE:4T =5M: form and choose 5ppl2 to displa, the W6ITT7: T6-::-0 9:=:-57 I=,O-M5TIO= form2

"2 =5M: O, P-OE:4T

The pro>ect name is defined in the previous form2

"3 =o& of S8-P54 Su!regions

The num)er of su)re ions ,ou want to use2 All )loc+s in the model must )elon to a su)re ion2 Su)re ion strin s are stored in the e$tent strin file and are iven strin num)ers from ' to :'0(((2

"

Product Price(8nit

#nter the price of one unit of the product2 For old this will )e price per ram or price per ounce2 For rades which are e$pressed as a percent0 it is usual to present the price per 5(+ of product2

"" 5ir String =um!er

#nter the air strin num)er2 #ven if there is no air in the model0 ,ou must specif, a strin num)er0 however choose a strin which is not present in the resource model2 It is a ood idea to adopt a standard convention for num)erin ,our air strin 2 !ost people use strin 6(2

"* :<clusion String =o&

The e$clusion strin lets ,ou define su)re ions which cannot )e mined for various reasons0 such as )oundaries or surface infrastructure2 S%&PAC' assi ns all the )loc+s in this su)re ion an e$tremel, hi h minin cost Bma+in the )loc+s have a ver, lar e ne ative valueC2 This will force the optimiser not to mine them2 If there is no e$clusion =one in ,our model0 enter =ero2

"+ ,ile ,ormat

#nter the format of the output 92eco9 file2 File formats are fi$ed0 meanin that the data is ri idl, formatted into column0 or free0 where the information is separated ), a comma2 Free format files are much smaller0 and so are the default option2 Complete the W6ITT7: T6-::-0 9:=:-57 I=,O-M5TIO= form and choose 5ppl2 to displa, the S8-P54 S8B-:9IO= 0:,I=ITIO= form2

". =5M: O, P-OE:4T

The pro>ect name defined in the first form is displa,ed here2

"/ S8B-:9IO=
This indicates the su)re ion for which ,ou are enterin data2 It is incremented ), one and will reflect the num)er of su)re ions ,ou are considerin 2

*' String =o&

#nter the strin num)er in the e$tent file which outlines this su)re ion2

*1 :le1ation -ange
#nter the elevation ran e over which the su)re ion e$tends2 Separate the 7from9 and 7to9 elevations ), a semicolon2

*2 Waste Bulk 0ensit2

#nter the default waste )ul+ densit, for the su)re ion2 4ou will not )e prompted for this field if ,ou choose to use Varia)le S2@9s in the !odel *loc+in phase2 In this case the "aste *ul+ Densit, is contained in the )loc+ files2

*3 Waste Mining 4ost ?costm@

#nter the waste minin cost BCOST!C for the su)re ion2 This will )e e$pressed in units of currenc, divided ), units of wei ht Be2 2 I per tonC2

=o& of Material 4lasses

#nter the num)er of material classes which are present in this su)re ion2 The ne$t forms will provide ,ou with the opportunit, to enter data which relate to each material class2 Complete the S8-P54 S8B-:9IO= 0:,I=ITIO= form and choose 5ppl2 to displa, the T6-::-0 M5T:-I57 0:,I=ITIO= form2

*" =5M: O, P-OE:4T

The pro>ect name defined in the first form is displa,ed here2

** S8B-:9IO=
The su)re ion for which ,ou are enterin data is displa,ed here2

*+ M5T:-I57
The material num)er is displa,ed2 "hen ,ou choose 5ppl20 the ne$t T6-::-0 M5T:-I57 0:,I=ITIO= form is displa,ed and the material num)er is incremented ), one2

*. Mineral String -ange

#nter the strin ran e of mineralised material2 Do not include default waste in this ran e2

*/ Material Bulk 0ensit2

#nter the )ul+ densit, of the mineralised material2 4ou will not )e prompted for this field if ,ou choose to use Varia)le S2@9s in the !odel *loc+in phase2 In this case the !aterial *ul+ Densit, is contained in the )loc+ files

+' Mining 4ost ?costm@

#nter COST! for the mineralised material0 this is the cost of minin the mineralised material as waste2

+1 =o& of Treatment Options

#nter the num)er of technical feasi)le treatment options for mineralised material2 There is a ma$imum of five different treatment options per material t,pe2

+2 TreatmentK Processing 4ostK -eco1er2

#nter for each processin method0 the processin cost BCOSTPC and the recover, factor as a fraction of 52 6int; 4ou ma, have material in ,our model which is not mineralised )ut is also not default waste0 such as a waste dump2 For the purpose of calculatin a value0 assume that it is mineralised material )ut ensure that the rade is set to =ero2 Since the revenue less the cost of treatment will )e ne ative0 the material will )e mined as waste2 Cop,ri ht J '((A Surpac !ine$ @roup

Surpac to Whittle ,O8--0

The interface to Four-D produces the strictl, formatted Four-D input model file from the intermediate 72)l+9 files2 4ou are re-uired to enter information on )ul+ densities for the material t,pes in the model2 After enterin

eneral data which relate to the whole model0 ,ou input information forD

each su)re ion within the model3 and each material within each su)re ion2

The pro ram will also add ,our input information to ,our pro>ect file0 thus this file serves a useful dual purpose since it now stores all the relevant parameters throu h the S%&PAC' process of ,our pro>ect2 Also output from this interface is a stu) "hittle Four-D parameter file2 It contains all information re-uired ), "hittle that Surpac +nows a)out2

+3 ProBect 4ost 5ssumptions

The interface to Four-D is simpler than that for Three-D2 %nli+e the Three-D interface0 the Four-D interface does not input cost data directl,0 these costs are presented to the Four-D model file independentl, of S%&PAC' throu h the Four-D parameter file2 However the S%&PAC' to Four-D interface uses su)re ions to revise minin and processin costs with position in the model2 It does this ), assi nin 7Cost Ad>ustment Factors9 to particular )loc+s in the Four-D input model file2 There are two t,pes of cost ad>ustment factorsD

!inin Cost Ad>ustment Factors Processin Cost Ad>ustment Factors

Cost ad>ustment factors can )e e-ual to or reater than 52 %suall, ,ou var, costs with depth0 however ,ou can also var, ,our cost ad>ustment factors across a )ench2 4ou do this ), suppl,in more than one su)re ion for a particular )ench Bor ran e of )enchesC2 Future versions of S%&PAC' will allow ,ou to use pol, onal su)re ions2 4ou can also achieve much the same effect now if ,ou use different strin ran es to reflect different material t,pes Bmore a)out this laterC2 Also the su)re ions will not have the one )loc+ width space )etween them2 Choose Surpac to Whittle ,O8--0 from the S8-P542 C-D Whittle menu to displa, the W6ITT7: P-OE:4T =5M: form2

=ame of ProBect

#nter the pro>ect name2 Complete the W6ITT7: P-OE:4T =5M: form and choose 5ppl2 to displa, the W6ITT7: ,O8--0 9:=:-57 I=,O-M5TIO= form2

+" =5M: O, P-OE:4T

The pro>ect name defined in the previous form is displa,ed here2

+* =o& of S8-P54 Su!regions

The num)er of su)re ions ,ou want to use2 All )loc+s in the model must )elon to a su)re ion2 Su)re ion strin s are stored in the e$tent strin file and are iven strin num)ers from ' to :'0(((2

++ 5ir String =um!er

#nter the air strin num)er2 #ven if there is no air in the model0 ,ou must specif, a strin num)er0 however choose a strin which is not present in the resource model2 It is a ood idea to adopt a standard convention for num)erin ,our air strin 2 !ost people use strin 6(2

+. ,ile ,ormat
#nter the t,pe of file to output2 Fi$ed format files have the data arran ed in specific columns0 separated ), spaces2 Free format files have the data separated ), commas2 Free format files will )e much smaller0 and this is the default settin 2 Complete the W6ITT7: ,O8--0 9:=:-57 I=,O-M5TIO= form and choose 5ppl2 to displa, the S8-P54 S8B-:9IO= 0:,I=ITIO= form2

+/ =5M: O, P-OE:4T
The pro>ect name defined in the first form is displa,ed here2

.' S8B-:9IO=
This indicates the su)re ion for which ,ou are enterin data2 It is incremented ), one and will reflect the num)er of su)re ions ,ou are considerin 2

.1 String =o&
#nter the strin num)er in the e$tent file which outlines this su)re ion2

.2 :le1ation -ange
#nter the elevation ran e over which the su)re ion e$tends2 Separate the 7from9 and 7to9 elevations ), a semicolon2

.3 4ost 5dBustment ,actor; MiningK Processing

#nter the minin and processin cost ad>ustment factors which relate to the su)re ion2

Waste Bulk densit2

#nter the default waste )ul+ densit, for the su)re ion2 4ou will not )e prompted for this field if ,ou choose to use Varia)le S2@9s in the !odel *loc+in phase2 In this case the "aste *ul+ Densit, is contained in the )loc+ files2

." =o& of Material 4lasses

#nter the num)er of material classes which are present in this su)re ion2 The ne$t forms will provide ,ou with the opportunit, to enter data which relate to each material class2 The ma$imum num)er of material classes for a su)re ion is presentl, set at 8(2 ?oteD It is important that ,ou use the same num)er of material classes for each su)re ion2 For e$ample0 if ,ou have three material classes in one su)re ion comprisin o$ide0 transition and sulphide ores0 S%&PAC' will automaticall, assi n material class num)ers of (((50 (((' and (((:2 If o$ide is the first material0 it ma, comprise a num)er of strin s Bsa, 506C and the material code assi ned to o$ide in the model file will )e (((52 However0 if in another su)re ion there is no o$ide ore Bi2e2 strin s 506 are a)sentC0 ,ou must still enter an o$ide material class2 The material class which is defined first in each su)re ion will alwa,s )e assi ned the code (((50 the second (((' and so on2 Complete the S8-P54 S8B-:9IO= 0:,I=ITIO= form and choose 5ppl2 to displa, the ,O8--0 M5T:-I57 0:,I=ITIO= form2

.* =5M: O, P-OE:4T
The pro>ect name defined in the first form is displa,ed here2

.+ S8B-:9IO=
The su)re ion for which ,ou are enterin data is displa,ed here2

.. M5T:-I57
The material num)er is displa,ed2 "hen ,ou choose 5ppl20 the ne$t ,O8--0 M5T:-I57 0:,I=ITIO= form is displa,ed and the material num)er is incremented ), one2

./ Mineral String -ange

#nter the strin ran e of mineralised material2 Do not include default waste in this ran e2

/' Material Bulk 0ensit2

#nter the )ul+ densit, of the mineralised material2 4ou will not )e prompted for this field if ,ou choose to use Varia)le S2@9s in the !odel *loc+in phase2 In this case the !aterial *ul+ Densit, is contained in the )loc+ files2

/1 Material =ame
!aterials in "hittle have a name which is up to four characters in len th2 #nter here the name of a material which is to )e output to the "hittle file2 Since materials can )e in several su)-re ions0 duplicate names are allowed in material definitions2 Cop,ri ht J '((A Surpac !ine$ @roup

Whittle to Surpac
This part of the interface ives ,ou the tool to e$tract contour strin s of optimal pit outlines from Three-D and Four-D result files2 4ou can view these outlines in 9-5P6I4S2 Also ,ou can use the contours as a uide in S%&PAC' interactive 9-5P6I4S Pit 0esign2 Choose Whittle to Surpac from the S8-P542 C-D Whittle menu to displa, the W6ITT7: P-OE:4T =5M: form2

/2 =ame of ProBect
#nter the pro>ect name2 This will cause the pro ram to see+ the pro>ect file name2 The pro>ect file stores information on the )loc+ dimensions and also the e$tent strin used to )loc+ the model in the first place2 "ith this information it is possi)le to relate the cell addresses of the )loc+s in the "hittle result file to /4L real world coordinates for further processin in S%&PAC'2 The pro>ect also stores whether ,our pro>ect has used Three-D or Four-D optimisation pac+a es2 The pro>ect file also stores information re ardin the file format of the model to )e input2 This is either Free or Fi$ed0 and must )e correct for the "hittle to Surpac function to wor+ correctl,2 If ,ou are usin Four-D the ,O8--0 TO S8-P54 form is displa,ed2 If ,ou are usin Three-D the T6-::-0 TO S8-P54 form is displa,ed2

/3 =5M: O, P-OE:4T
The pro>ect name defined in the previous form is displa,ed here2

Whittle -esults ,ile

#nter the name of the Three-D results file2

/" Output ,ile 7ocation

#nter a location and name for the file which will store the optimal outline2

/* Multiple Phases L
If ,ou want to e$tract specific or multiple phases from the results file ,ou should answer 749 to this -uestion2

/+ -ange Id
If ,ou choose 7?9o to the !ultiple Phases -uestion0 enter the &an e Id of the file which will store the optimal outline2

/. Phase -ange

If ,ou choose 749 to the !ultiple Phases -uestion0 enter the phase num)ers to e$tract from the results file2 The phase num)ers should )e specified as a ran e2

// Include 5ir Blocks

The outlines will include air )loc+s if the appropriate air fla s were set )efore ,ou ran 1@:D2 @enerall, ,ou should set these as followsD

T,pe A M ' T,pe * M :

1'' String Position

Strin contours of the "hittle )loc+s can )e ta+en at an, one of three positions on the )loc+2 These outlines can either )e on the toe of the )loc+0 the crest0 or the midpoint2 This settin will ma+e a sli ht difference to the final output of ,our model2 &efer to the Three-D documentation for more information2 If ,ou are usin Four-D the ,O8--0 TO S8-P54 form is displa,ed2

1'1 =5M: O, P-OE:4T

The pro>ect name defined in the first form is displa,ed here2

1'2 Whittle -esults ,ile

#nter the name of the Four-D results file2

1'3 Outline ,ile

#nter a 7ocation and -ange of Pit =um!ers to e$tract outlines which are stored in the Four-D results file2

1' Include 5ir Blocks

The outlines will include air )loc+s if the appropriate air fla s were set )efore ,ou ran 1@:D2 @enerall, ,ou should set these as followsD

T,pe A M ' T,pe * M :

1'" String Position

Strin contours of the "hittle )loc+s can )e ta+en at an, one of three positions on the )loc+2 These outlines can either )e on the toe of the )loc+0 the crest0 or the midpoint2 This settin will ma+e a sli ht difference to the final output of ,our model2 &efer to the Four-D documentation for more information2 The outlines will include air )loc+s if the air fla is set correctl, )efore ,ou run FDOP2 The air fla should )e set to ' or :2 &efer to the Four-D documentation for more information2 6int; If ,ou re)loc+ ,our "hittle input model usin either the 1@&* or FD&* utilities0 ,ou will need to ad>ust the model information B)loc+ si=e and num)er of )loc+s in the /0 4 and L directionsC held in pro>ect file to reflect this2

1'* :rrors
Blo ! at entroid "#$nort%ing$ &#$ea'ting$ %a' $(er entage$) o*erla( o+ 'tring'

An overlap of mineral parcel strin s was detected2 4ou must chec+ ,our model or use lar er )loc+ si=es
Blo ! $",id$ $&,id$ $-,id$ ould not .e lo ated in a 'u.region

The indicated )loc+ does not )elon to an, of the defined su)re ions
/entroid Blo ! 0odel +ile i' not a multi(le o+ t%e .lo ! 'i-e

The )loc+ model )ein used can not fit an even num)er of )loc+s in the iven e$tent strin
/ould not '%rin! (it outline +or 'tring $'tring num.er$

One of the pit outline strin s was found to )e open and hence0 not created correctl,
1le*ation .et2een la"er' doe' not mat % 3,dimen'ion' o+ .lo !

The distance )etween )enches does not match the hei ht the specified as the L dimension of the )loc+s
1rror reating ore (ar el S45

The pro ram was una)le to create a 9S%&PAC "or+ Area9 to )e used for storin material0 tonnes and metal for each different mineral t,pe in a )loc+2
1rror inter'e ting 'egment'

The pro ram was una)le to calculate the contents of a )loc+ usuall, )ecause of a peculiarit, in the model strin file
1rror o(ening 1&tent +ile $'tring +ile name$

The pro ram was una)le to open the specified e$tent file Busuall, 9e$t52str9C
1rror o(ening 6ro7e t +ile $+ile name8.(r7$

The pro ram was una)le to open the specified pro>ect file
1rror o(ening 4%ittle 9e'ult' +ile # $error num.er$

The results file Bwith 92res9 e$tensionC could not )e found or access to the file was not possi)le
1rror reading data in 6ro7e t +ile $filename.(r7$

The contents of the pro>ect file do not contain the e$pected format Bi2e2 data commencin in column '5C
1rror reading +rom re'our e model $'tring +ile name$

The rade or the strin num)er Bcentroid )loc+ models onl,C could not )e read from the specified strin file description field
1&tent 'tring mu't ontain 5 (oint'

#ach e$tent and su)re ion specified in the e$tent file must )e rectan ular and hence contain five points in the correspondin strin num)er
:ini'%ed ... $+un tion name$

Indicates that the selected function was completed successfull,

Grade mu't .e in t%e 3 or ; +ield

"hen enterin the location of the rade field ,ou incorrectl, specified the / or 4 field or an invalid description field
In'u++i ient data in 6ro7e t +ile $+ilename.(r7$

The first seven lines are not all present in the pro>ect file or the, do not contain the e$pected format Bdata commencin in column '5C Invalid Description field used for varia)le S2@9s If ,ou are usin varia)le S2@9s then the values must )e stored in one of the description fields in the strin file2 Invalid S@ detected in D field 7num)er9 The value in the description field used to store the S2@9s was not a valid num)er2
<::S1= +rom 1&tent String to /entroid Blo ! 0odel i' not a multi(le o+ t%e Blo ! 'i-e

The distance from the first )loc+ in the model to the )ottom left hand corner of the e$tent strin is not a multiple of the )loc+ si=e
6oint $num.er$ in +ile $'tring +ile name$ i' out'ide 0odel 1&tent'

The e$tent file must contain all points in the model Point 7point num)er9 is not in sorted se-uence 4M ,-coord9 /M9$-coord$ "hen processin a centroids model0 this error usuall, indicates there is a duplicate point at the specified coordinates2
6ro e''ing Ben % ... $.en % id$

Informational messa e indicatin what )ench is currentl, )ein processed Processin Phase 222 7phase num)er9 Informational messa e indicatin what phase is currentl, )ein processed
6ro e''ing 6it ... $(it num.er$

Informational messa e indicatin what pit is currentl, )ein processed

9ange mu't go +rom ><41S= u( to ?IG?1S= 1le*ation

The la,er ran e for a su)re ion specified in the form min3 ma$ must have the lower of the two elevations specified first
9ange mu't %a*e t%e +orm min3 ma$

4ou must specif, the su)re ion elevation ran e as rlmin3rlma$ even when the su)re ion onl, applies to one la,er Bi2e2 rlmin e-uals rlma$C
String model $'tring +ile name$ i' (oorl" +ormed

An overlap of mineral parcel strin s was detected2 4ou must chec+ ,our model or use lar er )loc+ si=es
Su.region $num.er$ i' not multi(le o+ Blo ! 'i-e

The width0 hei ht or depth of the su)re ion strin must )e contain a whole num)er of )loc+s in each direction

1'+

-eferences

"hittle Three-D &eference !anual - "hittle Pro rammin Pt, 1td

"hittle Four-D &eference !anual - "hittle Pro rammin Pt, 1td The Theor, )ehind Pit Optimisation - Keff "hittle Cop,ri ht J '((A Surpac !ine$ @roup

S8-P54 to Three-0 4ase Stud2

The president of Alli ator @old !ines is wonderin whether to e$tend the life of the compan,9s #ure+a !ine2 The deposit has )een wor+ed as a small scale operation over a num)er of ,ears2 The present minin operation comprises two pits0 ?orth Pit and South Pit2 !ost of the potential for e$tension appears to lie )eneath and )etween the e$istin two pits2 The president +nows that the remainin eolo ical resource would )e sufficient to e$tend the life of the mine for a ma$imum of another two ,ears at current production rates2 This is an ideal pro>ect to e$amine usin "hittle Three-D2 The data are stored on a dis+ette la)elled 7#ure+a Three-D Pro>ect92 Cop, the contents into a director, called #%&#.A on ,our hard dis+2 First step is to familiarise ,ourself with the dataD

,ilename top52str #$istin topo raph,

0escription

eurE(E08:(0A2str @eolo ical resource model E(Em&1 to 8:(m&1 )enches Am apart e$t52str eure+a2pr> master2pr> mod)loc+2cmd formNeco2cmd eure+a2res ta+eNpit2cmd #$tent strin file #ure+a pro>ect file *ac+up of #ure+a pro>ect file !acro to convert strin model into a re ular )loc+ model !acro to calculate intrinsic value model to input to Three-D Three-D results file B,ou need Three-D if ,ou want to ma+e ,our ownC !acro to e$tract optimal pit outline from Three-D result file2

1'. Step 1 Introduction to the :ureka Propert2

&ecall top52str strin file into @raphics2 This shows the e$istin surface topo raph, of the #ure+a !ine2 4ou can recall the eolo ical resource la,ers into @raphics either separatel, or to ether and familiarise ,ourself with the eometr, of the resource e$tension2 The resource model has )een classified into o$ide0 sulphide and transition materials2 Strin s 50E are o$ide0 5505E are transition and '50'E are sulphide2 *ecause #ure+a have classified their materials li+e this0 the, have enormous fle$i)ilit, to appl, different )ul+ densities0 minin costs0 processin methods and costs and recoveries to individual strin s0 or ran es of strin s0 later on in the anal,sis2

1'/ Step 2

11' 4on1ert the S8-P54 String Model into a -egularised Block Model
This sta e is common to Three-D and Four-D pro>ects2 The model produced has )loc+s which measure 8m B/ directionC0 5(m B4 directionC and Am BL directionC2

111 Macro Pla2!ack

The macro mod)loc+2cmd will run the process automaticall,2 All macros in this tutorial should )e run from the main menu2 The macro will pause at each input screen until ,ou press a +e, to continue2

112 Manual Pla2!ack

4ou can run the process manuall, since all the +e,stro+es have )een stored on the eure+a2pr> file2 If ,ou enter an, chan es0 the values stored in the eure+a2pr> will )e chan ed2 Pressin the #nter +e, will accept the e$istin information or an, chan es that ,ou ma, have made2 The file master2pr> is a )ac+up of ,our eur2pr> file in case ,ou ma+e accidental chan es2

113 Step 3 11 4alculate #alue Model from -egularised Block Model

This process creates the strictl, formatted value model which is input into "hittle Three-D2 The pro ram allows ,ou to specif, su)re ions which ive ,ou further control appl,in costs0 densit, and recover, data2 The #ure+a !ine considers three su)re ions2 The first su)re ion comprises all )loc+s which are in the ?orth Pit from E(Em&1 to EEAm&12 This material has a hi her )ul+ densit, than an,where else on the propert,2 The roc+ is a little harder and drill and )last costs are a little hi her2 The second su)re ion covers the South Pit and e$tends and over the same elevation ran e as the first su)re ion2 Finall,0 the third su)re ion covers the whole pro>ect area and starts where the first and second su)re ions finish and runs ri ht throu h to surface2 !inin costs in the third su)re ion are a little lower due to shorter haul distances2 The )est wa, to present the input parameter is as a ta)le2 This is a summar, print of Alli ator @old !ines9s cost anal,sis spreadsheet2

4a'te <&ide =ran'ition Sul(%ide String 9ange 98 1@4 11@14 21@24 Su.,region Ao 1 String Ao 2 9> 9ange 404B446 Bul! ;en'it" 2.30 2.30 2.35 2.50 /<S=0 1.60 1.80 1.90 2.00 /<S=6 81C19.00 20.00 22.00 /<S=6 82C10.00 14.00 9e o*er" 81C0.92 0.91 0.89 9e o*er" 82C0.68 0.65 4a'te <&ide =ran'ition Sul(%ide String 9ange 98 1@4 11@14 21@24 Su.,region Ao 2 String Ao 3 9> 9ange 404B446 Bul! ;en'it" 2.20 2.20 2.30 2.40 /<S=0 1.55 1.75 1.85 1.95 /<S=6 81C19.00 20.00 22.00

/<S=6 82C10.00 14.00 9e o*er" 81C0.92 0.91 0.89 9e o*er" 82C0.68 0.65 4a'te <&ide =ran'ition Sul(%ide String 9ange 98 1@4 11@14 21@24 Su.,region Ao 3 String Ao 4 9> 9ange 452B530 Bul! ;en'it" 2.20 2.20 2.30 2.40 /<S=0 1.45 1.70 1.80 1.90 /<S=6 81C19.00 20.00 22.00 /<S=6 82C10.00 14.00 9e o*er" 81C0.92 0.91 0.89 9e o*er" 82C0.68 0.65

11" Macro Pla2!ack

The macro formNeco2cmd will activate the S%&PAC to Three-D function and enerate an intrinsic value )loc+ model from the 72)l+9 files produced in the previous sta e2 The eure+a2eco file is read, to run throu h "hittle Three-D2 If ,ou have the Three-D pac+a e0 use the parameter file supplied0 enerate a structural arcs model and then run the pit optimisation2 These steps will ta+e a few minutes each0 refer to the Three-D manual for uidance2 "e have alread, prepared a Three-D results file which is called eure+a2res2

11* Step Whittle to Surpac

#$tract the Optimal Pit Outline Contours from the Three-D &esults File2 This process is ver, strai ht forward2 4ou must ensure that the followin files are presentD

eure+a2res Three-D results file eure+a2pr> pro>ect file e$t52str e$tent strin file

11+ Macro Mode

From the main menu run the macro etNpit2cmd2 This will recall the three files listed a)ove and produce a strin file pit52str which will store the optimal pit outline for our particular technical and economic scenario2

11. Manual Mode

#nter the followin D

4ou will usuall, show air strin s in ,our output model )ecause this function will allow ,ou to see the relationship of the new pit with the e$istin topo raph,2 In the case of Alli ator @old !inin 0 this means the e$istin #ure+a !ine2

4ou can then separate the optimal pit outline into individual la,ers usin Strin Tools - Split Strin Files2 The individual la,ers can )e pulled into the Pit Desi n pac+a e and used as a template to uide ,ou with ,our final practical pit desi n2 4ou ma, also wish to section the optimal pit outline2 This is done usin ST-I=9 ,I7: -:PO-TS :nd 5rea #olumes2

Cop,ri ht J '((A Surpac !ine$ @roup

S8-P54 to ,our-0 4ase Stud2

Sim)a !inin Corporation0 which owns the Casa)lanca copper pro>ect have identified a eolo ical resource of 'A million tonnes at an avera e rade of 52'FO copper2 The Casa)lanca deposit su)-outcrops alon a rid e of hills which rises sharpl, from a plain2 The hills run north-south and are continuous over man, +ilometres2 Sim)a have conducted an e$tensive diamond drillin campai n which has identified mineralisation in several lodes which dip around :( de rees to the east to an appro$imate depth of '(( metres )eneath the level of the plain2 The pro>ect has a stri+e len th of around 528 +ilometres2 The eolo ical resource would indicate a mine life in e$cess of G ,ears2 Sim)a reco nise that the pro>ect will )e difficult to schedule - the, are oin to need all the help the, can et2 An unsettled world econom, and unsta)le commodit, prices prevail2 The political part, in power have a favoura)le polic, towards resource investment )ut the, are not firml, in control and this could chan e2 These factors have convinced Sim)a9s C#O to pursue a development polic, which will minimise the compan,9s capital e$posure while still ensurin a health, net present value2 The Sim)a en ineers have )een instructed to set a)out their tas+ and to use "hittle Four-D as their pit optimisation and evaluation tool2 The compan, have surve,ed the topo raph, and have formed a di ital terrain model in S%&PAC'2 A thorou h eotechnical assessment of the area has )een done and Sim)a were luc+, to inherit some old adits into the hillside from which mappin was carried out2 This also provided an opportunit, to e$tract some )ul+ samples for metallur ical test wor+2 Create a director, called 7sim)a9 and cop, the contents of the dis+ette mar+ed 7Casa)lanca Four-D Copper Pro>ect9 into this director,2 4ou should familiarise ,ourself with the data )elow2 ,ilename top'2str #$istin topo raph, 0escription

mod'G(80::(80'( @eolo ical resource model 'G(8m&1 to ::(8m&1 )enches '(m apart e$t'2str sim)a2pr> master2pr> mod)loc+2cmd formNmod2cmd sim)a2res sim)a2prm ta+eNpit2cmd copeval2spd fdanNcop2prn #$tent strin file Casa)lanca pro>ect file *ac+up of Casa)lanca pro>ect file !acro to convert strin model into a re ular )loc+ model !acro to convert re ular )loc+ model into Four-D format )loc+ model2 Four-D results file B,ou need Four-D if ,ou want to ma+e ,our ownC Casa)lanca parameter file for Four-D !acro to e$tract optimal pit outline shells from Four-D result file2 7Spreadsheet9 te$t file which is output from Four-D Anal,sis Print file output from Four-D Anal,sis2

11/ Step 1

12' Introduction to the 4asa!lanca ProBect

&ecall top'2str strin file into @raphics2 This shows the surface topo raph, of the Casa)lanca pro>ect2 4ou can recall the eolo ical resource la,ers into @raphics either separatel, or to ether and familiarise ,ourself with the eometr, of the lode structures2 The resource model has )een classified into low rade Bstrin 5C0 mid rade Bstrin 'C0 and hi h rade Bstrin :C2

121 Step 2 122 4on1ert the S8-P54 String Model into a -egularised Block Model
This sta e is common to Three-D and Four-D pro>ects2 The model produced has )loc+s which measure '( metres B/ directionC0 '( metres B4 directionC and '( metres BL directionC2

123 Macro Pla2!ack

The macro mod)loc+2cmd will run the process automaticall,2 All macros in this tutorial should )e run from the main menu2 The macro will pause at each input screen until ,ou choose 5ppl2 to continue2

12 Manual Pla2!ack
4ou can run the process manuall, since all the +e,stro+es have )een stored in the sim)a2pr> file2 If ,ou enter an, chan es0 the values stored in the sim)a2pr> will )e chan ed2 Pressin the #nter +e, will accept the e$istin information or an, chan es that ,ou ma, have made2 The file master2pr> is a )ac+up of ,our sim)a2pr> file in case ,ou ma+e accidental chan es2

12" Step 3 12* 4on1ert the -egularised Block Model to ,our-0 ,ormat
This process creates the strictl, formatted model which is input into "hittle Four-D2 The pro ram allows ,ou to specif, S%&PAC' su)re ions which ive ,ou control appl,in costs and densit, data2 The Casa)lanca pro>ect considers three su)re ions2 #ach su)re ion covers the entire evaluation area2 The first su)re ion starts at 'G(8m&1 and e$tends to :((8m&10 the second oes from :('8m&1 to :5'8m&10 and the last from :5E8m&1 to the hi hest natural surface at ::(8m&12 The minin costs will var, )etween within the same material t,pe )etween su)re ions2 Sim)a have used S%&PAC' to assi n e$plicit minin cost ad>ustment factors to )loc+s in each su)re ion of the model2 The minin costs increase with depth2 The )est wa, to present the input data are in a ta)le2 This is a summar, print of Sim)a9s9 cost anal,sis spreadsheet2
0ineral 'tring range 0aterial la'' 69I/1 8DE10!g o((er 9e o*er" /95=I< 0/<S=0 Su.,region 'tring 1le*ation range Bul! den'it" 2a'te Bul! den'it" mineral /<S=0 /<S=6 0ining /5: 6ro e'' /5: 1@3 1 25 90) 13.04 0.0368 Su.,reg 1 2 2805B3005 2.89 2.89 1.05 12.00 1.14 1.00

Su.,reg 2 3 3025B3125 2.79 2.79 0.97 12.00 1.05 1.00

Su.,reg 3 4 3145B3305 2.72 2.76 0.92 12.00 1.00 1.00

12+ Macro Pla2!ack

The macro four-d2cmd will run the process and will pause at each input screen until ,ou choose 5ppl22 The macro must )e run from the main menu2

12. Manual Pla2!ack

The +e,stro+es ma, )e entered manuall,2 Activate the function 7S%&PAC to Four-D90 ,ou will )e prompted for the pro>ect name0 enter 7sim)a92 The ne$t form will prompt ,ou for num)er of S%&PAC su)re ions Benter :C and the air strin num)er Benter 6(C which is used in the model2 Then for each su)re ion ,ou will )e prompted for su)re ion strin num)er in the e$tent strin 0 the &1 ran e of the su)re ion0 the minin and processin cost ad>ustment factors0 the waste )ul+ densit, and the num)er of material classes2 All this information is held in the ta)le a)ove2 For each material class BSim)a have onl, one material class0 so ,ou will see this form onl, onceC ,ou must enter the mineral strin ran e and mineral )ul+ densit,2 The sim)a2mod file is read, to run throu h "hittle Four-D2 If ,ou have the Four-D pac+a e0 use the parameter file supplied0 enerate a structural arcs model and then run the pit optimisation2 The optimisation will ta+e a)out :( minutes to run2 &efer to the Four-D manual for uidance2 "e have alread, prepared a Four-D results file which is called sim)a2res2

12/ Step Whittle to Surpac

#$tractin the Optimal Pit Outline Contours from the Four-D &esults File2 This process is ver, strai ht forward2 4ou must ensure that the followin files are presentD

sim)a2res Four-D results file sim)a2pr> pro>ect file e$t52str e$tent strin file

13' Macro Mode

From the main menu run the macro etNpit2cmd2 This will recall the three files listed a)ove and produce a strin file for the particular optimal pit shell which ,ou have nominated2 4ou can chose from )etween 5 to :( and produce them individuall, or specif, a ran e Be2 2 805F0'C2 #ach strin file will store the optimal pit outline for a particular value of !COST! - refer to the "hittle Four-D manual for more information on this2

131 Manual Mode

#nter the followin D

4ou will usuall, show air strin s in ,our output model )ecause this function will allow ,ou to see the relationship of the pit with the e$istin topo raph,2 The ne$t process is to understand which optimal pit outline shell is the most appropriate for Sim)a9s re-uirements2 Four-D provides a powerful anal,sis tool +nown as FDA?2 FDA? has allowed Sim)a to nominate values for minin cost0 copper price0 minin fleet andHor treatment plant capacit,0 interest rate and man, other factors and it has pic+ed the shell which ives the optimal value for a particular set of conditions2 This process too+ moments and the results of such an anal,sis are ta)ulated )elowD
Grand total'F

6I= 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

9</KE= 15522259 101956085 103693003 103736567 111816825 127123290 133826540 140064929 140246146 146176577 151098773 153118868 153856468 169556074 180632533 196305562 196374122 207263965 207468845 260105327 261890205 264677616 266532031 280198042 285610436 296267991 297398072 303513002 361402528 364174711

0I>>E= 0I>>EG >I:1EB 2343657 10806742 11006290 11009588 11701046 12724547 13303363 13833023 13841922 14084323 14456655 14605392 14642651 15520484 16214490 17065768 17068320 17714287 17719984 19938038 20000493 20143074 20209844 20737172 20846571 21276577 21322787 21504686 23284240 23353093 1.18 1.17 1.17 1.17 1.17 1.19 1.18 1.18 1.18 1.19 1.19 1.19 1.19 1.19 1.19 1.19 1.19 1.18 1.18 1.19 1.19 1.19 1.19 1.19 1.19 1.19 1.19 1.19 1.18 1.18 2.34 10.81 11.01 11.01 11.70 12.72 13.30 13.83 13.84 14.08 14.46 14.61 14.64 15.52 16.21 17.07 17.07 17.71 17.72 19.94 20.00 20.14 20.21 20.74 20.85 21.28 21.32 21.50 23.28 23.35

S=9I6 5.62 8.43 8.42 8.42 8.56 8.99 9.06 9.13 9.13 9.38 9.45 9.48 9.51 9.92 10.14 10.50 10.51 10.70 10.71 12.05 12.09 12.14 12.19 12.51 12.70 12.92 12.95 13.11 14.52 14.59

/5S? ;IS/<GA=1; :><4 B1S= /5S1 V5>G1 V5>G1 24401751 22682227 80819296 46485429 81938610 46994532 81958976 47003171 85894341 48672508 91963488 50834986 94642187 51760220 96788653 52539272 96828868 52553868 97922783 52890169 98782662 53178852 99038893 53264876 99079126 53278383 99617136 53242958 99643857 53154815 99051733 52949708 99047006 52948452 96938923 52388312 96901409 52378343 84121630 48401747 83621433 48287768 82631843 48079659 82004193 47947666 76628768 46817226 74705427 46412752 69776144 45334917 69181486 45219240 66053956 44610856 34434990 38327629 32877557 38068409

;IS/<GA=1; 4<9S= /5S1 V5>G1 22682227 43757369 43829662 43870309 43866011 42534243 42009453 41301262 41294767 40073221 39071815 38701374 38574614 34519197 31681132 27158142 27141251 23564279 23515343 4550496 3968341 3042872 2309642 ,2568952 ,4820076 ,8638165 ,8999379 ,11233479 ,33904728 ,34816975

Instead of allowin "hittle FDA? to choose a particular pit0 Sim)a have tested one of their scenarios on all the shells in the output model2 FDA? has calculated the cash flow value and )est and worst discounted cash flows of each shell and presented the results in a hand, format2 A raph of &OC.HT Btotal roc+ tonna e or pit si=eC a ainst net present value loo+s li+e )elowD

The C#O of Sim)a was a)le to use the raph to help ma+e the most appropriate investment decision2 The raph of the discounted )est case versus pit si=e shows that ma>or chan es in pit si=e a)out the characteristicall, flat pea+ Ba)out 5(( million tonnes in this caseC have relativel, little effect on the net present value2 The ma$imum value for the cash flow0 discounted )est case and discounted worst case scenarios showed that the ma$imum cash flow will )e represented ), Pit 582 The )est case discounted schedulin scenario has selected Pit 5: and the worst case discounted se-uence Pit E2 In realit,0 the Sim)a en ineers have advised their C#O that the actual schedulin scenario for Casa)lanca will lie somewhere )etween the )est and worst case2 The C#O considers that the corporate polic, will )est )e served if the )oard were to select a final pit outline somewhere )etween Pit 8 and Pit A2 Pit 8 has a discounted )est case value of IEG2F million versus Pit 5: which is on the pea+ at I8:2: million2 FDA? can actuall, simulate the actual schedulin scenario0 the en ineers were a)le to specif, in which shell se-uence the push )ac+s should ta+e place2 The, showed that push )ac+s from Pit 5 to Pit ' and then to Pit 8 will )e practical and will produce a ?PV of IEG28 million2

The en ineers separated the optimal pit outline into individual la,ers usin ST-I=9 TOO7S Split String ,ile2 The individual la,ers when pulled into Pit 0esign0 were used as a template to uide the final practical pit desi n2 The incremental shell pits were also used to uide the en ineers with the two ma>or sta e pits2 Cop,ri ht J '((A Surpac !ine$ @roup

Block Model to Whittle Interface

,unction =ame?s@ ;

*! O%TP%T "HITT1#

Outputtin "hittle Four-D or Four-/ model file Outputtin "hittle Three-D economic file Outputtin "hittle from an Indicator .ri ed model Importin a "hittle file The )loc+ model interface is desi ned to convert a )loc+ model into the re-uired format necessar, for "hittle to perform an optimisation0 usin "hittle9s Three-D0 Four-D or Four-/ optimisation pac+a es2 Once the optimisation is complete0 it is also )e possi)le to read the results of the optimisation )ac+ into a )loc+ model0 either ), creatin a new )loc+ model and fillin it with the appropriate data0 or ), readin "hittle data )ac+ into an e$istin )loc+ model2 For e$ample0 it is possi)le to output a "hittle file0 optimise it with one of "hittle9s products0 and then read the results )ac+ into the ori inal )loc+ model2 The data can then )e visualised usin standard )loc+ model functions2 Due to the nature of a )loc+ model )ein ver, similar to the )loc+ concept of "hittle0 the interface itself is ver, simple to use2 Several attri)utes for specific values need to e$ist in the )loc+ model0 )ut once these attri)utes e$ist and are filled with the correct values0 outputtin the data as a "hittle file is a simple e$ercise2 "hen outputtin a model0 ,ou will )e as+ed to input the attri)utes and values needed for the calculation0 and the )loc+ si=e of a "hittle )loc+2 The )loc+ model will then )e traversed one )loc+ at a time2 If the )loc+ currentl, )ein traversed is smaller than the "hittle )loc+0 then the current )loc+ will )ecome a parcel of the "hittle )loc+2 If the current )loc+ is lar er than the "hittle )loc+ Ba super )loc+C0 then the super )loc+ will )e divided into it9s component "hittle si=e )loc+s2 The si=e of the "hittle )loc+ will alwa,s )e related to the resolution of the )loc+ model2 Please note that the )loc+s in the "hittle model file output from this interface will not )e in sorted numerical order2 This is due to the nature of the )loc+ model0 and is no pro)lem at all to "hittle2 However0 if ,ou wish to sort ,our model file0 use "hittle9s re)loc+in tools0 F/&*0 FD&* or 1@&* without re)loc+in to sort the model2

132 Model Preparation

To output a reasona)le "hittle file0 correct and useful data must first e$ist in the )loc+ model2 "ithout the data in the )loc+ model )ein reasona)le and correct0 the "hittle model will not ma+e sense0 or will )e incorrect2 Therefore0 ood model preparation is essential2 The three "hittle pro rams need sli htl, different data to output successfull,0 )ut there are several common attri)utes which must e$ist2 Fillin these attri)utes with reasona)le data is the most difficult part of the interface2 *ut here are some su estions2 !an, values in "hittle chan e )etween what "hittle calls su)re ions2 The )loc+ model interface does not care a)out su)re ions as such0 )ut will use the values ta+en from a )loc+ in a specific area2 The )est wa, to fill the values in these su)re ions is to create a constraint for each su)re ion ,ou wish to have0 and simpl, fill the appropriate attri)utes to the appropriate values inside that constraint2 Attri)utes such as !inin Cost

Ad>ustment Factors0 Processin Cost Ad>ustment Factors0 and Cost of Processin and &ecover, values all are dependant on su)re ions2

133 -e3uired 5ttri!utes

Firstl,0 there must e$ist an attri)ute containin a rade value for ever, )loc+ in the model2 These values must e$ist as a fraction of PunitsP H tonnes0 with a =ero rade usuall, denotin an air or waste )loc+2 The metal contents produced will then )e in PunitsP2 If an air or waste )loc+ has a rade reater than =ero0 the rade will >ust )e i nored2 For a "hittle Four-/ output20 several rade fields are needed0 one for each element to process2 Second0 there must e$ist an attri)ute with a roc+ code in it2 This attri)ute can )e either a character or an inte er attri)ute2 This roc+ code is used to determine whether the )loc+ is ore0 waste or air2 If the )loc+ is an ore )loc+0 the roc+ code is used to determine the material name to output to the "hittle file BFour-D or Four-/ onl,C2 "hittle roc+ codes have a ma$imum of four characters0 so if the code in the )loc+ model is lon er than four characters0 then the names are truncated2 Please ma+e sure that all roc+ codes used in the )loc+ model are uni-ue over the first four characters2 If the roc+ code is numeric0 then the num)er will )e converted to a =eropadded character strin 2 Therefore0 an, numeric roc+ code lar er than 6666 will also )e truncated2 All )loc+s must have a roc+ code value0 otherwise the, will )e assumed to )e waste2 Air )loc+s must have a roc+ code value2 The third common field is a specific ravit, attri)ute2 This attri)ute contains the specific ravit, for the )loc+2 If )loc+ is to )ecome a parcel for the "hittle )loc+ and the )loc+ is not air0 it9s specific ravit, value will )e used in calculatin a wei hted avera e of the specific ravit, for the entire "hittle )loc+2 There are several attri)utes which also must e$ist for some "hittle products )ut not others2 "hittle Four-D and Four-/ re-uire a !inin Cost Ad>ustment Factor B!CAFC2 All )loc+s must contain an !CAF value2 However0 if all )loc+s in the model contain the same !CAF0 then no attri)ute is re-uired0 the sin le value can )e entered on the e$port whittle form2 As cost of minin onl, applies to )loc+s containin roc+0 !CAF can )e =ero for air )loc+s2 See Outputtin "hittle Four-D or Four-/ model file for more information on how !CAF is used and calculated in su))loc+ed models2 A Processin Cost Ad>ustment Factor BPCAFC must also e$ist for a Four-D or Four-/ file2 A ain0 however0 if the PCAF value is the same for all )loc+s in the model0 no attri)ute is re-uired0 the value can )e entered on the e$port whittle form2 The cost of processin a )loc+ o)viousl, onl, applies if the )loc+ is to )e processed0 so air and waste )loc+s do not need a meanin ful PCAF value2 See Outputtin "hittle Four-D or Four-/ model file for more information on how PCAF is used and calculated in su)-)loc+ed models2 For a "hittle Three-D file0 !CAF and PCAF values are not re-uired0 )ut other processin costs are used to calculate a )loc+ value2 There must e$ist a Cost of !inin 2 Other costs are determined ), a calculation usin the recover, factor of the material0 and a Cost of Processin associated with that material2 #ver, material can have several or these CostPHrecover, pairs2 Therefore0 for ever, different processin t,pe0 two attri)utes must )e created0 one containin a recover, factor Bas a factor of 5C and a cost of processin 2 These values a ain can )e filled usin constraints0 as the cost of processin and recover, will )e used for all )loc+s of the same material t,pe2 So simpl, assi n a value to the recover, and cost of processin where the roc+ code attri)ute e-uals the re-uired material2 If all the a)ove attri)utes are correctl, filled0 and this can )e raphicall, chec+ed usin the )loc+ model functions0 then the output "hittle file should )e simple to output2 Onl, one )loc+ model t,pe can ive a sli htl, different answer0 and that is a model with attri)utes filled with Indicator .ri in 2

13 Indicator >riged Models

Indicator .ri ed models are different in that the, will not contain a rade as such2 "hat the, do contain a cutoff and a percenta e2 The cutoff is a rade value2 The percenta e determines how much of the )loc+ is contained )elow that cutoff2 A mean value is determined from the cutoff and the previous cutoff0 and this value is ta+en as the rade2 #ach cutoff will ma+e a new parcel of the "hittle )loc+0 with a parcel value )ein determined ), multipl,in the mean ), the specific ravit, ), the volume2 The volume is calculated as the )loc+ volume times the percenta e )elow the cutoff2 In this wa,0 all cutoffs will )ecome parcels of the )loc+2 This is potentiall, dan erous0 as "hittle has a ma$imum of 8( parcels per )loc+2 If each )loc+ contains 5( cutoff levels0 then a ma$imum of 8 su) )loc+s can contain ore0 so for an Indicator .ri ed model0 the resolution used is critical2 Cop,ri ht J '((A Surpac !ine$ @roup

Block Model to Whittle Three-0

"hittle :d uses an #conomic file B2ecoC which contains onl, a )loc+ inde$ and a value for that )loc+2 To calculate the value of a )loc+0 values are needed for the rade0 densit,0 costs of minin and processin 0 recover, and price2 As all of these values0 e$cept price0 can chan e for each )loc+0 material t,pe or su)re ion of the model0 each value must come out of an attri)ute of the )loc+ model2 These values must )e completel, filled for the entire model )efore an, attempt is made to convert it to a "hittle Three-D model2 Choose :<port Whittle from the B7O4> MO0:77I=9 main menu to displa, the :MPO-T W6ITT7: ,I7: form2

13" Whittle file t2pe

Choose "hittle Three-D2 The form will modif, itself accordin l, so that values will )e entered for "hittle Three-D2

13* Block model t2pe

This field ives ,ou the opportunit, to output values from either a standard or an indicator +ri ed )loc+ model2 If an indicator +ri ed model is selected0 ,ou will not )e a)le to enter a rade on the :MPO-T W6ITT7: ,I7: form0 )ecause the rade will )e entered later2 Otherwise0 the form will )e unchan ed2 See Indicator .ri ed !odels to "hittle2

13+ -ock code attri!ute

The roc+ code attri)ute is a character attri)ute which contains a value for each )loc+0 desi natin the t,pe of roc+ found in that )loc+2 Air and waste )loc+ must also contain a roc+ code2 An, )loc+ in the model where a roc+ code is not present is assumed to )e waste2

13. Specific gra1it2 attri!ute

#ach )loc+ must also contain a specific ravit, attri)ute2 All roc+ )loc+s0 includin waste must have a meanin ful specific ravit,2

13/ 4ostM attri!ute

This attri)ute contains a cost of minin for each )loc+2 This is necessar, as the cost of minin ma, chan e throu hout the model2 Be 2 as the pit ets deeper0 the truc+ fuel costs increaseC2 A cost of minin is necessar, for ever, roc+ )loc+0 )ut can )e i nored for air )loc+s2

1 ' 9rade attri!ute

#ach )loc+ will contain a rade value2 @rade values in waste and air )loc+s will )e i nored2

1 1 Material price

To calculate a value for the )loc+0 a price for the mined material must )e +nown2 This is a numeric value0 not an attri)ute name0 as a sin le price will affect the entire model2

1 2 Waste rock code

This character value is the value contained in the roc+ code attri)ute which denotes a waste )loc+2 An, )loc+ which is defined as waste must have the value entered here as the value in it9s roc+ code attri)ute2

1 3 5ir rock code

An, )loc+ in the model which is an air )loc+ must have this value in it9s roc+ code attri)ute0 or it will not )e processed as air2

Model file name

#nter the name of the model file Bwithout the 2eco e$tensionC to which the model will )e written2 If the file e$ists0 another form will appear as+in for confirmation of the model name2

1 " Parameter file name

#nter the name of the "hittle parameter file stu)2 "ritten to the parameter file are the values which S%&PAC'((( +nows a)out0 such as )loc+ si=e0 num)er of )loc+0 ori in etc2 If the file e$ists0 another form will appear as+in for confirmation of the parameter file name2

1 * Ore !od2 id num!er

This attri)ute is optional2 "hittle allows for a third field to )e output which contains an ore )od, identification num)er2 If ,ou wish to have this field in the economic file0 then create an attri)ute to contain it0 fill the attri)ute with the re-uired values0 and then enter the name of that attri)ute here2 ?o processin will )e done with this value0 is will simpl, )e directl, output to the economic file2

1 + -eco1er2 attri!ute
Different materials can have different processin methods2 However0 rather than definin su)re ions and materials so that processin methods can )e input0 ,ou must have two attri)utes in the model for each of these processin methods2 4ou then fill these attri)utes with the re-uired values0 for materials and su)re ions ), usin )loc+ model constraints2 If some materials have two processin methods0 whilst other materials have onl, one0 then there will )e two pairs of attri)utes throu h the entire model2 If this is the case0 then set the recover, value to =ero0 and the second processin method will )e i nored for the material with onl, one processin method2 So this field contains the first half of the processin method pair0 the recover, value2 #nter the attri)ute for the recover, value in the first processin method2 Then for the second processin methodQetc2 %p to five processin methods can )e entered2

1 . 4ostP attri!ute
Processin methods are calculated ), ettin a recover, value0 and a cost of processin 2 The costp attri)ute is the second half of this0 and so a cost of processin must )e entered for each processin method in the model2

1 / ,ile format
"hittle files can either )e free format with commas delimitin each field0 or fi$ed column format2 Comma delimited files will )e much smaller0 and therefore faster to read2 Historicall,0 however0 fi$ed format files were used0 )ut these are no lon er recommended2

1"' 8ser !lock resolution

"hittle )loc+s have a specific si=e2 Due to the nature of the )loc+ model0 it is impossi)le to allow the fle$i)ilit, of )loc+ si=es that "hittle allows2 Therefore0 the )loc+ si=e can )e chan ed in relation to the resolution of the )loc+ model2 1owerin the resolution will increase the )loc+ si=e of the model0 as indicated ), the )loc+ si=e section of the form2 However0 as an, )loc+ in the model smaller than this user )loc+ resolution is ore0 it will )ecome a "hittle parcel2 This ma, lead to more )loc+s )ecomin parcels2 "hittle has a ma$imum of 8( parcels0 and so this interface will output a ma$imum of 8( parcels and truncate the rest Bi nore them totall,C2 Therefore0 whilst decreasin the )loc+ si=e will speed up the interface0 the model will not )e as accurate0 and there is a ris+ of oin over the 8( parcel limit2 It is recommended that ,ou output the file at the default resolution Bset to the user )loc+ resolution of the )loc+ modelC and perform an, re)loc+in usin "hittle9s 1@&* pro ram2

1"1 4onstrain model

Constrainin the model will mean that it is possi)le to output some part of the model2 For e$ample0 usin constraints it is possi)le to output to the "hittle file onl, calculated and indicated rades0 and i nore inferred rades2 If constraints are selected0 the standard )loc+ model constraints form will appear2 Cop,ri ht J '((A Surpac !ine$ @roup

Block Model to Whittle ,our-0 ( ,our-M

This function will ta+e a )loc+ model0 and output a "hittle Four-D or Four-/ model file B2modC2 To correctl, output this file0 values are needed for rade0 specific ravit,0 minin cost ad>ustment factor and processin cost ad>ustment factors2 If outputtin a "hittle Four-/ file0 then several rade fields ma, )e re-uired2 Since these values can all chan e an,where in the model0 the, all have to e$ist as attri)utes of the )loc+ model0 so that effectivel, all of these values can chan e for each )loc+2 These attri)utes must )e completel, filled with correct data )efore an, outputtin can ta+e place2 Choose :<port Whittle from the B7O4> MO0:77I=9 main menu to displa, the :MPO-T W6ITT7: ,I7: form2

1"2 Whittle file t2pe

Choose "hittle Four-D or "hittle Four-/2 The form will modif, itself accordin l, so that values will )e entered accordin to the "hittle file t,pe2

1"3 Block model t2pe

This field ives ,ou the opportunit, to output values from either a standard or an indicator +ri ed )loc+ model2 If an Indicator .ri ed model is selected0 ,ou will not )e a)le to enter a rade on the :MPO-T W6ITT7: ,I7: form0 )ecause the rade will )e entered later2 Otherwise0 the form will )e unchan ed2 See Indicator .ri ed !odels to "hittle2 If0 however0 ,ou wish to output an indicator +ri ed model0 then due to the nature of Indicator .ri in 0 a "hittle Four-D file will alwa,s )e output0 even if the file t,pe field is set to "hittle Four/2

1" -ock code attri!ute

The roc+ code attri)ute is a character attri)ute which contains a value for each )loc+0 desi natin the t,pe of roc+ found in that )loc+2 Air and waste )loc+ must also contain a roc+ code2 An, )loc+ in the model where a roc+ code is not present is assumed to )e waste2 &oc+ codes in "hittle have a ma$imum of four characters2 Due to this0 an, roc+ code in a )loc+ model lon er than four characters will )e truncated at the fourth character2 Therefore0 ma+e sure that an, roc+ codes ,ou enter in the model are uni-ue for the first four characters0 or the, will )e assumed to )e the same roc+ code2 It is possi)le to use an inte er attri)ute for the roc+ code2 The attri)ute value will then )e converted to a =ero-padded character roc+ code2 Please ma+e sure that an, numeric roc+ codes are less than 66660 or the, will also )e truncated2

1"" Specific gra1it2 attri!ute

#ach )loc+ must also contain a specific ravit, attri)ute2 All roc+ )loc+s0 includin waste must have a meanin ful specific ravit,2 Specific ravit, of air )loc+s should alwa,s )e set to (2

1"* M45, attri!ute

This attri)ute contains a !inin Cost Ad>ustment Factor for each )loc+2 This ad>ustment factor ma, chan e throu hout the model0 )ut meanin ful values are a)ove 52 Therefore0 it is recommended when creatin this attri)ute that a )ac+ round value of 5 )e set2 Air )loc+s ma, have a value of (2

If ,ou wish to have a sin le !CAF value for the entire )loc+ model0 simpl, enter the value instead of an attri)ute name0 and that value will )e used in ever, )loc+ in the resultant whittle model2

1"+ P45, attri!ute

This attri)ute contains a Processin Cost Ad>ustment Factor for each )loc+2 This ad>ustment factor ma, chan e throu hout the model0 )ut meanin ful values are a)ove 52 Therefore0 it is recommended when creatin this attri)ute that a )ac+ round value of 5 )e set2 Air and waste )loc+s ma, have a value of (2 If ,ou wish to have a sin le PCAF value for the entire )loc+ model0 simpl, enter the value instead of an attri)ute name0 and that value will )e used in ever, )loc+ in the resultant whittle model2

1". #olume ,actor

The volume factor ad>usts the volume of a particular )loc+2 This effectivel, creates a partial ore )loc+2 The volume factor is a percenta e attri)ute whose value should alwa,s lie )etween ( and 52 This value is then muliplied ), the )loc+ volume to ive an ad>usted volume for the ore in the )loc+2 See the ta)le at the )ottom of this pa e to see a description of how this information is used2

1"/ Waste rock code

This character value is the value contained in the roc+ code attri)ute which denotes a waste )loc+2 An, )loc+ which is defined as waste must have the value entered here as the value in it9s roc+ code attri)ute2

1*' 5ir rock code

An, )loc+ in the model which is an air )loc+ must have this value in it9s roc+ code attri)ute0 or it will not )e processed as air2

1*1 Model file name

#nter the name of the model file Bwithout the 2mod e$tensionC to which the model will )e written2 If the file e$ists0 another form will appear as+in for confirmation of the model name2

1*2 Parameter file name

#nter the name of the "hittle parameter file stu)2 "ritten to the parameter file are the values which S%&PAC'((( +nows a)out0 such as )loc+ si=e0 num)er of )loc+0 ori in2 Also output will )e all roc+ codes used in the model0 which will )e determined ), the interface as it processes2 If the file e$ists0 another form will appear as+in for confirmation of the parameter file name2

1*3 -ock t2pe num!er

This attri)ute is optional2 "hittle allows for a field to )e output which contains an roc+ t,pe num)er2 If ,ou wish to have this field in the model file0 then create an attri)ute to contain it0 fill the attri)ute with the re-uired values0 and then enter the name of that attri)ute here2 ?o processin will )e done with this value0 is will simpl, )e directl, output to the model file2

1* 9rade attri!ute
#ach )loc+ will contain a rade value2 This rade will )e multiplied ), the parcel tonna e0 to determine the units of material Bor elementC in the output file2 @rade values in waste and air )loc+s will )e i nored2 If outputtin a "hittle Four-/ file0 then ever, rade entered in the scrollin re ion will )e assumed to )e a different element2 The, will )e output to the model file in the order that the, appear in the scrollin re ion2 The element name will )e the first four characters of the rade attri)ute used2

1*" ,ile format

"hittle files can either )e free format with commas delimitin each field0 or fi$ed column format2 Comma delimited files will )e much smaller0 and therefore faster to read2 Historicall,0 however0 fi$ed format files were used0 )ut these are no lon er recommended2

1** 8ser !lock resolution

"hittle )loc+s have a specific si=e2 Due to the nature of the )loc+ model0 it is impossi)le to allow the fle$i)ilit, of )loc+ si=es that "hittle allows2 Therefore0 the )loc+ si=e can )e chan ed in relation to the resolution of the )loc+ model2 1owerin the resolution will increase the )loc+ si=e of the model0 as indicated ), the )loc+ si=e section of the form2 However0 as an, )loc+ in the model smaller than this user )loc+ resolution is ore0 it will )ecome a "hittle parcel2 This ma, lead to more )loc+s )ecomin parcels2 "hittle has a ma$imum of 8( parcels0 and so this interface will output a ma$imum of 8( parcels and truncate the rest Bi nore them totall,C2 Therefore0 whilst decreasin the )loc+ si=e will speed up the interface0 the model will not )e as accurate0 and there is a ris+ of oin over the 8( parcel limit2 It is recommended that ,ou output the file at the default resolution Bset to the user )loc+ resolution of the )loc+ modelC and perform an, re)loc+in usin "hittle9s FD&* or F/&* pro rams2

1*+ 4onstrain model

Constrainin the model will mean that it is possi)le to output some part of the model2 For e$ample0 usin constraints it is possi)le to output to the "hittle file onl, calculated and indicated rades0 and i nore inferred rades2 If constraints are selected0 the standard )loc+ model constraints form will appear2

1*. Whittle Output 4alculations 1*/ Block header line

If ,ou have a )loc+ model with su) )loc+in 0 or are outputtin the "hittle model at a lower resolution B), usin the features of the output whittle functionC then a "hittle )loc+ is made up of man, smaller )loc+s2 "here these smaller )loc+s are ore0 the, are shown as "hittle parcels2 "here the, are air or waste0 the, simpl, affect the overall properties of the lar er )loc+2 *elow is a description of the calculations used to calculate )loc+ and parcel information used in the output of a "hittle ED or E/ file2 Output 1alue M45, 4alculation for non su!-!locked model 4alculation for su!-!locked model

&ead directl, from the !CAF attri)ute Calculated as the mass-wei hted avera e of all in the )loc+ model )loc+s which are not air2 Calculated as the mass-wei hted avera e of all &ead directl, from the PCAF attri)ute )loc+s which are not air or waste Band hence in the )loc+ model must )e processedC2 *loc+ tonna e is a simple calculation of Same calculationBs R )loc+ volumeC0 e$cept the the specific ravit, attri)ute value ), s is calculated ), a volume wei hted avera e the )loc+ volume Bcalculated from the of all su) )loc+s Bincludin air and wasteC of the )loc+ si=eC2 whittle )loc+2

P45,

Block Tonnage

1+' Block parcel line

Output 1alue 4alculation for non su!-!locked model Parcel Tonnage specific ravit,Bfrom )loc+ attri)uteC R )loc+ volume Bfrom )loc+ dimensionsC R volume ad>ustment factor Bfrom )loc+ attri)ute - not used if the volume ad>ustment attri)ute is not definedC

Parcel Metal Simple calculation of the a)ove parcel tonna e R rade for a particular element2 In a

4ontent

E/ model0 the same calculation is performed on a num)er of rade values ta+en from attri)utes2

Cop,ri ht J '((A Surpac !ine$ @roup

Block Model to Whittle ,our-0 ( ,our-M

This function will ta+e a )loc+ model0 and output a "hittle Four-D or Four-/ model file B2modC2 To correctl, output this file0 values are needed for rade0 specific ravit,0 minin cost ad>ustment factor and processin cost ad>ustment factors2 If outputtin a "hittle Four-/ file0 then several rade fields ma, )e re-uired2 Since these values can all chan e an,where in the model0 the, all have to e$ist as attri)utes of the )loc+ model0 so that effectivel, all of these values can chan e for each )loc+2 These attri)utes must )e completel, filled with correct data )efore an, outputtin can ta+e place2 Choose :<port Whittle from the B7O4> MO0:77I=9 main menu to displa, the :MPO-T W6ITT7: ,I7: form2

1+1 Whittle file t2pe

Choose "hittle Four-D or "hittle Four-/2 The form will modif, itself accordin l, so that values will )e entered accordin to the "hittle file t,pe2

1+2 Block model t2pe

This field ives ,ou the opportunit, to output values from either a standard or an indicator +ri ed )loc+ model2 If an Indicator .ri ed model is selected0 ,ou will not )e a)le to enter a rade on the :MPO-T W6ITT7: ,I7: form0 )ecause the rade will )e entered later2 Otherwise0 the form will )e unchan ed2 See Indicator .ri ed !odels to "hittle2 If0 however0 ,ou wish to output an indicator +ri ed model0 then due to the nature of Indicator .ri in 0 a "hittle Four-D file will alwa,s )e output0 even if the file t,pe field is set to "hittle Four/2

1+3 -ock code attri!ute

The roc+ code attri)ute is a character attri)ute which contains a value for each )loc+0 desi natin the t,pe of roc+ found in that )loc+2 Air and waste )loc+ must also contain a roc+ code2 An, )loc+ in the model where a roc+ code is not present is assumed to )e waste2 &oc+ codes in "hittle have a ma$imum of four characters2 Due to this0 an, roc+ code in a )loc+ model lon er than four characters will )e truncated at the fourth character2 Therefore0 ma+e sure that an, roc+ codes ,ou enter in the model are uni-ue for the first four characters0 or the, will )e assumed to )e the same roc+ code2 It is possi)le to use an inte er attri)ute for the roc+ code2 The attri)ute value will then )e converted to a =ero-padded character roc+ code2 Please ma+e sure that an, numeric roc+ codes are less than 66660 or the, will also )e truncated2

1+ Specific gra1it2 attri!ute

#ach )loc+ must also contain a specific ravit, attri)ute2 All roc+ )loc+s0 includin waste must have a meanin ful specific ravit,2 Specific ravit, of air )loc+s should alwa,s )e set to (2

1+" M45, attri!ute

This attri)ute contains a !inin Cost Ad>ustment Factor for each )loc+2 This ad>ustment factor ma, chan e throu hout the model0 )ut meanin ful values are a)ove 52 Therefore0 it is recommended when creatin this attri)ute that a )ac+ round value of 5 )e set2 Air )loc+s ma, have a value of (2

If ,ou wish to have a sin le !CAF value for the entire )loc+ model0 simpl, enter the value instead of an attri)ute name0 and that value will )e used in ever, )loc+ in the resultant whittle model2

1+* P45, attri!ute

This attri)ute contains a Processin Cost Ad>ustment Factor for each )loc+2 This ad>ustment factor ma, chan e throu hout the model0 )ut meanin ful values are a)ove 52 Therefore0 it is recommended when creatin this attri)ute that a )ac+ round value of 5 )e set2 Air and waste )loc+s ma, have a value of (2 If ,ou wish to have a sin le PCAF value for the entire )loc+ model0 simpl, enter the value instead of an attri)ute name0 and that value will )e used in ever, )loc+ in the resultant whittle model2

1++ #olume ,actor

The volume factor ad>usts the volume of a particular )loc+2 This effectivel, creates a partial ore )loc+2 The volume factor is a percenta e attri)ute whose value should alwa,s lie )etween ( and 52 This value is then muliplied ), the )loc+ volume to ive an ad>usted volume for the ore in the )loc+2 See the ta)le at the )ottom of this pa e to see a description of how this information is used2

1+. Waste rock code

This character value is the value contained in the roc+ code attri)ute which denotes a waste )loc+2 An, )loc+ which is defined as waste must have the value entered here as the value in it9s roc+ code attri)ute2

1+/ 5ir rock code

An, )loc+ in the model which is an air )loc+ must have this value in it9s roc+ code attri)ute0 or it will not )e processed as air2

1.' Model file name

#nter the name of the model file Bwithout the 2mod e$tensionC to which the model will )e written2 If the file e$ists0 another form will appear as+in for confirmation of the model name2

1.1 Parameter file name

#nter the name of the "hittle parameter file stu)2 "ritten to the parameter file are the values which S%&PAC'((( +nows a)out0 such as )loc+ si=e0 num)er of )loc+0 ori in2 Also output will )e all roc+ codes used in the model0 which will )e determined ), the interface as it processes2 If the file e$ists0 another form will appear as+in for confirmation of the parameter file name2

1.2 -ock t2pe num!er

This attri)ute is optional2 "hittle allows for a field to )e output which contains an roc+ t,pe num)er2 If ,ou wish to have this field in the model file0 then create an attri)ute to contain it0 fill the attri)ute with the re-uired values0 and then enter the name of that attri)ute here2 ?o processin will )e done with this value0 is will simpl, )e directl, output to the model file2

1.3 9rade attri!ute

#ach )loc+ will contain a rade value2 This rade will )e multiplied ), the parcel tonna e0 to determine the units of material Bor elementC in the output file2 @rade values in waste and air )loc+s will )e i nored2 If outputtin a "hittle Four-/ file0 then ever, rade entered in the scrollin re ion will )e assumed to )e a different element2 The, will )e output to the model file in the order that the, appear in the scrollin re ion2 The element name will )e the first four characters of the rade attri)ute used2

1. ,ile format

"hittle files can either )e free format with commas delimitin each field0 or fi$ed column format2 Comma delimited files will )e much smaller0 and therefore faster to read2 Historicall,0 however0 fi$ed format files were used0 )ut these are no lon er recommended2

1." 8ser !lock resolution

"hittle )loc+s have a specific si=e2 Due to the nature of the )loc+ model0 it is impossi)le to allow the fle$i)ilit, of )loc+ si=es that "hittle allows2 Therefore0 the )loc+ si=e can )e chan ed in relation to the resolution of the )loc+ model2 1owerin the resolution will increase the )loc+ si=e of the model0 as indicated ), the )loc+ si=e section of the form2 However0 as an, )loc+ in the model smaller than this user )loc+ resolution is ore0 it will )ecome a "hittle parcel2 This ma, lead to more )loc+s )ecomin parcels2 "hittle has a ma$imum of 8( parcels0 and so this interface will output a ma$imum of 8( parcels and truncate the rest Bi nore them totall,C2 Therefore0 whilst decreasin the )loc+ si=e will speed up the interface0 the model will not )e as accurate0 and there is a ris+ of oin over the 8( parcel limit2 It is recommended that ,ou output the file at the default resolution Bset to the user )loc+ resolution of the )loc+ modelC and perform an, re)loc+in usin "hittle9s FD&* or F/&* pro rams2

1.* 4onstrain model

Constrainin the model will mean that it is possi)le to output some part of the model2 For e$ample0 usin constraints it is possi)le to output to the "hittle file onl, calculated and indicated rades0 and i nore inferred rades2 If constraints are selected0 the standard )loc+ model constraints form will appear2

1.+ Whittle Output 4alculations 1.. Block header line

If ,ou have a )loc+ model with su) )loc+in 0 or are outputtin the "hittle model at a lower resolution B), usin the features of the output whittle functionC then a "hittle )loc+ is made up of man, smaller )loc+s2 "here these smaller )loc+s are ore0 the, are shown as "hittle parcels2 "here the, are air or waste0 the, simpl, affect the overall properties of the lar er )loc+2 *elow is a description of the calculations used to calculate )loc+ and parcel information used in the output of a "hittle ED or E/ file2 Output 1alue M45, 4alculation for non su!-!locked model 4alculation for su!-!locked model

&ead directl, from the !CAF attri)ute Calculated as the mass-wei hted avera e of all in the )loc+ model )loc+s which are not air2 Calculated as the mass-wei hted avera e of all &ead directl, from the PCAF attri)ute )loc+s which are not air or waste Band hence in the )loc+ model must )e processedC2 *loc+ tonna e is a simple calculation of Same calculationBs R )loc+ volumeC0 e$cept the the specific ravit, attri)ute value ), s is calculated ), a volume wei hted avera e the )loc+ volume Bcalculated from the of all su) )loc+s Bincludin air and wasteC of the )loc+ si=eC2 whittle )loc+2

P45,

Block Tonnage

1./ Block parcel line

Output 1alue 4alculation for non su!-!locked model Parcel Tonnage specific ravit,Bfrom )loc+ attri)uteC R )loc+ volume Bfrom )loc+ dimensionsC R volume ad>ustment factor Bfrom )loc+ attri)ute - not used if the volume ad>ustment attri)ute is not definedC

Parcel Metal Simple calculation of the a)ove parcel tonna e R rade for a particular element2 In a

4ontent

E/ model0 the same calculation is performed on a num)er of rade values ta+en from attri)utes2

Cop,ri ht J '((A Surpac !ine$ @roup

Block Model ?Indicator >riged@ to Whittle

Indicator +ri in is a eostatistical method which estimates the distri)ution of rades within a )loc+2 It does this ), determinin the percenta e of ore in the )loc+ which is )elow a iven cutoff2 In terms of "hittle0 this means that a certain percenta e of the )loc+ has a iven rade BcutoffC2 Therefore0 each cutoff in an Indicator .ri ed model will )ecome a "hittle parcel of that )loc+2 All that is needed is to suppl, the cutoffs Bwhich will )e consistant throu hout the modelC and the percenta e0 which will chan e for ever, )loc+0 and a useful "hittle model can )e created2 If the model has )een filled usin the )loc+ model9s indicator +ri in function0 then the percenta e attri)utes needed will )e of the formD Ni+cSTattri)ute nameUSTcutoffU2 Due to this0 the :=T:- I=0I45TO>-I9I=9 #578:S form ma, alread, )e filled with the correct values2 Complete the :MPO-T W6ITT7: ,I7: form with the )loc+ model t,pe set to 9Indicator .ri ed9 and choose 5ppl2 to displa, the :=T:- I=0I45TO- >-I9I=9 #578:S form2

1/' 4utoff
The cutoff is a rade value2 A mean value will )e determined once a value for the second and su)se-uent cutoffs are entered2 4ou ma, chan e the mean value for a iven cutoff2 #ach cutoff will ma+e a new parcel of the "hittle )loc+0 with a parcel value )ein determined ), multipl,in the mean ), the specific ravit, ), the volume2 The volume is calculated as the )loc+ volume times the percenta e )elow the cutoff2 In this wa,0 all cutoffs will )ecome parcels of the )loc+2 This is potentiall, dan erous0 as "hittle has a ma$imum num)er of parcels per )loc+2 For "hittle ED0 this ma$imum is 8(2 For "hittle E/0 the ma$imum is 662 If each )loc+ contains 5( cutoff levels0 then a ma$imum of 8 su) )loc+s can contain ore for a "hittle ED optimisation0 and a ma$imum of 6 su) )loc+s for a "hittle E/ optimisation is allowed2 For an Indicator .ri ed model0 the resolution used is critical2 %sin too man, cutoffs or too man, su) )loc+s can result in an e$cessive amount of parcels per )loc+2 If the num)er of parcels per )loc+ e$ceeds the ma$imum a)ove0 then the information for the e$tra parcels will )e i nored ), the function2

1/1 Mean
*ecause a cutoff is a threshold value0 it will not necessaril, )e the value to the rade for that interval2 It is much mode li+el, that the mean of a cutoff and the previous cutoff is closer to the true rade for that interval0 and so a mean is determined2 However0 if ,ou wish0 ,ou can chan e the mean to the rade value for that cutoff2 The mean value is the rade used to perform the calculations0 not the cutoff value2

1/2 Percentage attri!ute

#ver, cutoff has a percenta e attri)ute2 This value stores the percenta e of the )loc+ whose rade falls )elow the cutoff2 This means that the percenta es stored in the attri)utes are cumulative2 For e$ample0 if the percenta e for the first cutoff is (280 then the percenta e for the second cutoff must )e reater than (282 Values in this percenta e attri)ute must )e a fraction )etween ( and 52 If the model has )een filled usin the )loc+ model9s indicator +ri in function0 then the percenta e attri)utes needed will )e of the formD Ni+cSTattri)ute nameUSTcutoffU2

1/3 :<ample calculation

The followin is an e$ample of the calculations that will )e performed for the a )loc+ whose si=e is '$8$5(0 and which is located at the lower left )ottom corner of a model Bthe )loc+ with the minimum /0 minimum 40 and minimum L of model e$tentsC2 "hen e$ported0 this )loc+ will have "hittle i>+ values of 505052 This e$ample will assume that the followin indicator +ri ed values e$ist for this )loc+D Ni+cSdataS52((( Ni+cSdataS'2((( Ni+cSdataS:2((( Ni+cSdataS5(2((( (2((( (2'EA (2EAA 52(((

"hen e$ported to whittle0 the followin values were enteredD Cutoff 5 ' : 5( !ean (28 528 '28 E2(

?ote that for the first three cutoffs0 the numerical avera e of the cutoff and its predecessor were used2 However0 for the last cutoff0 the numeric avera e of : and 5( would )e B5(V:CH'MA280 )ut this value was not used2 As previousl, stated0 the user ma, enter an, value for the mean2 In this e$ample0 we are assumin that we +now from previousl, runnin the function *ASIC STATISTICS on the data set used to perform the indicator +ri in estimation that the data )etween : and 5( has a mean of E2(2 "e will assume in this e$ample that the specific ravit, of the )loc+ is '2(2 "hen all of this data is used to e$port the )loc+ to "hittle0 the followin data is written to an output R2!OD file2 In this e$ample0 we will assume that the output R2!OD file containsD 1,1,1,4,1.000,1.000,2.0 1,1,1,rock,0.0,0.0 1,1,1,rock,49.2,73.8 1,1,1,rock,44.0,110.0 1,1,1,rock,106.8,427.2 The first lineD 1,1,1,4,1.000,1.000,2.0 is comprised of the followin dataD i,j,k,number_of_parcels,mcaf,pcaf,sg The ne$t E linesD 1,1,1,rock,0.0,0.0 1,1,1,rock,49.2,73.8 1,1,1,rock,44.0,110.0 1,1,1,rock,106.8,427.2 have the followin formatD i,j,k,rock_code,parcel_tonnage,contained_metal

The calculation for the parcelNtonna e and containedNmetal for each cutoff are as follows Bcff M cumulative fre-uenc, functionCD parcelNtonna e M BcffBthis cutoffC - cffBprevious cutoffCC R )loc+Nvolume R s cffB/C M value of the cumulative fre-uenc, function for cutoff / For this )loc+D cffB(C M (2((( Balwa,s a constant value of =eroC cffB5C M (2((( cffB'C M (2'EA cffB:C M (2EAA cffB5(C M 52((( The value for cffBlast cutoffC should alwa,s )e 52(((2 If not0 the last cutoff is not hi h enou h0 and the "hittle R2!OD output file will not contain all of the tonna es that it should2 The last cutoff value should )e a)ove ,our ma$imum data value2 Also0 for each cutoffD containedNmetal M parcelNtonna e R rade The calculations for each of the cutoffs is detailed )elowD 4utoff N 1 parcelNtonna e M BcffB5C - cffB(CC R )loc+Nvolume R s parcelNtonna e M B( - (C R 'R8R5( R '2( parcelNtonna e M ( containedNmetal M parcelNtonna e R rade containedNmetal M ( R (28 containedNmetal M ( i,j,k, rock_code, parcel_tonnage, contained_metal 1,1,1,rock,0.0,0.0 4utoff N 2 parcelNtonna parcelNtonna parcelNtonna parcelNtonna e M BcffB'C - cffB5CC R )loc+Nvolume R s e M B(2'EA - (C R 'R8R5( R '2( e M (2'EA R 5(( R '2( e M E62'

containedNmetal M parcelNtonna e R rade containedNmetal M E62' R 528 containedNmetal M F:2G i,j,k, rock_code, parcel_tonnage, contained_metal 1,1,1,rock,49.2,73.8 4utoff N 3 parcelNtonna parcelNtonna parcelNtonna parcelNtonna e M BcffB:C - cffB'CC R )loc+Nvolume R s e M B(2EAA - (2'EAC R 'R8R5( R '2( e M (2'' R 5(( R '2( e M EE2(

containedNmetal M parcelNtonna e R rade containedNmetal M EE2( R '28 containedNmetal M 55(2( i,j,k, rock_code, parcel_tonnage, contained_metal 1,1,1,rock,44.0,110.0 4utoff N 1' parcelNtonna parcelNtonna parcelNtonna parcelNtonna e M BcffB5(C - cffB:CC R )loc+Nvolume R s e M B52((( - (2EAAC R 'R8R5( R '2( e M (28:E R 5(( R '2( e M 5(A2G

containedNmetal M parcelNtonna e R rade containedNmetal M 5(A2G R E containedNmetal M E'F2' i,j,k, rock_code, parcel_tonnage, contained_metal 1,1,1,rock,106.8,427.2 The output R2!OD file in this e$ample is thereforeD 1,1,1,4,1.000,1.000,2.0 1,1,1,rock,0.0,0.0 1,1,1,rock,49.2,73.8 1,1,1,rock,44.0,110.0 1,1,1,rock,106.8,427.2 ?ote that the sum of the tonna es for all parcels in the )loc+ is e-ual to the )loc+ volume R specific ravit,2 Sum of parcel tonna esD ( V E62' V EE2( V 5(A2G M '(( tonnes *loc+ volume R specific ravit,D ' R 8 R 5( R '2( M '(( tonnes As previousl, stated0 if the sum of the tonna es for all parcels is less than the )loc+ volume R specific ravit,0 the last cutoff was pro)a)l, lower than the lar est value in the data set used durin estimation2 A hi her value for the last cutoff should )e chosen0 and estimation of the cumulative fre-uenc, function needs to )e performed a ain2 Cop,ri ht J '((A Surpac !ine$ @roup

Whittle to Block Model

,unction =ame?s@ ;

*! I!PO&T "HITT1#

"hen readin a "hittle file )ac+ into the )loc+ model0 it is attempted to ma+e the conversion as simple as possi)le2 To read a "hittle model into the )loc+ model0 ,ou will need )oth a model file B92mod90 92eco90 92res9C and a valid "hittle parameter file B92par9C2 The parameter file is needed to determine the )loc+ si=e0 ori in0 num)er of )loc+s in i0 > and + directions0 and also determine the "hittle file t,pe0 whether "hittle Three-D0 Four-D or Four-/2 From this point0 it is possi)le to o two wa,s2 Firstl,0 ,ou can create an entirel, new )loc+ model of the si=e re-uired Bthis will )e done automaticall,C0 or ,ou can read the "hittle data into an e$istin )loc+ model0 assumin the models are compati)le2 *, compati)le0 the ori ins must )e either e$actl, the same0 or the "hittle model can )e offset ), $ )loc+s in an, direction0 where $ is an, positive inte er2 To )e compati)le the )loc+ si=e also must )e identical0 or can )e >oined to ether in ), $ )loc+s in an, or all directions0 as lon as $ is an inte er num)er of )loc+s2 This inte er is calculated on the user )loc+ si=e of the )loc+ model2 For e$ample0 if the user )loc+ si=e is 5($5($5(0 then it is valid to import a "hittle model with )loc+ si=e of '($:($5(2 It is not possi)le to import a "hittle model with )loc+ si=e of 58$5($5(2 The "hittle model also cannot o )e,ond the reach of the model space2 If these conditions are not met0 the interface will not attempt to read the "hittle model into the iven )loc+ model2 As with the :MPO-T W6ITT7: ,I7: form0 the IMPO-T W6ITT7: ,I7: form will modif, itself dependin on condition2 These modifications will ta+e place dependin on the "hittle file t,pe and whether ,ou are creatin a new model or readin into an e$istin file2 *ecause of this0 please read the form carefull,2 There are two places where ,ou can import a "hittle file2 Therefore0 choose either the B7O4> MO0:77I=9 )utton from the S%&PAC'((( main menu0 or select the Whittle -D Block Model )utton from the S8-P542 -D Whittle menu2 #ither of these )uttons will )rin up the S:7:4T MO0:7 form2

1/ Model =ame
#nter the model name that ,ou wish to convert2 V#&4 I!PO&TA?TD !a+e sure that ,ou enter not onl, the name )ut the 92mod90 92res9 or 92eco9 e$tension2 For e$ample if ,ou enter PmodelP0 it will either not find the "hittle file Bif ,ou are comin from the S8-P542 -D Whittle menuC0 or attempt to read Pmodel2mdlP if ,ou are comin from the Block Modelling )utton2 Valid names for "hittle model conversion are Pmodel2modP0 Pmodel2ecoP or Pmodel2resP2 Appl,in the S:7:4T MO0:7 form with a valid "hittle name will displa, IMPO-T W6ITT7: ,I7: form2

1/" Parameter file name

To correctl, import a "hittle file0 ,ou must have a valid parameter file to o with it2 This parameter file0 at minimum0 must contain line t,pes 5 and '2 These determine the )loc+ si=e0 ori in and the num)er of )loc+s in each direction0 which is in turn used to calculate the e$tents of the model2 Dependin on which t,pe of "hittle file ,ou are importin 0 other lines are also necessar,2 If importin a "hittle Three-D file0 line t,pes F0 G or 6 must e$ist2 If importin Four-D0 at least one line t,pe 58 must e$ist2 If Four-/0 at least one line t,pe '( must e$ist2 Otherwise0 the interface cannot determine the file t,pe0 and thus cannot import the file2

1/* Model file name

#nterin input here will attempt to read the "hittle file into the e$istin model entered as input2 If this field is left )lan+0 a new model will )e created0 with the name of the "hittle model file2 If the model entered here is not compati)le with the "hittle model0 then an error messa e will )e displa,ed and this field will )e cleared2

1/+ Model description

This field allows ,ou to enter a description of a model0 or allows ,ou to chan e the description of an e$istin model2 #nter the new description in this field2

1/. Block siAe - =um!er of !locks - Origin - Input file t2pe

These four fields displa, information ta+en from the parameters file2 These fields cannot )e edited0 )ut please chec+ that the data contained in them is correct for the model that ,ou are importin 2 It is essential that this information is correct for the model to )e imported without error

1// -eference !lock coordinates

At times0 ,ou ma, wish to import a portion of a model0 rather than a complete model2 If this is the case0 enter the i0>0+ coordinates for the )loc+ that ,ou wish to )ecome the lower left front corner of the )loc+ model2 An, coordinates lower than these input coordinates will simpl, )e i nored2 If the "hittle model ,ou are readin is lar er than 85' )loc+s in an, direction0 then a su)-model of the "hittle model0 of si=e 85'0 will )e created0 usin the input reference )loc+ as the startin position2

2'' -etrie1e parcel information

This field will onl, appear if ,ou are readin a "hittle Four-D or Four-/ file into a new model0 or are readin a model Bas opposed to resultC file2 It will allow the readin of all information out of the "hittle file2 If no0 then onl, the !CAF0 PCAF and pit num)er fields will )e read - an, parcel information will )e i nored2 If ,es0 then ,ou will )e iven the option of readin in values for each material Bif readin "hittle Four-DC or each element Bif readin "hittle Four-/C2 See &etrievin parcel information )elow for more information a)out retrievin the parcel information2

2'1 Pit num!er attri!ute

If ,ou are readin a "hittle Four-D or Four-/ model )ac+ into an e$istin model0 it is assumed that the model ,ou are readin )ac+ into was the model which output the "hittle file in the first place Bas if ,ou output a 92mod9 file0 ran it throu h "hittle and are tr,in to read the results )ac+ into the ori inal )loc+ modelC2 Therefore0 the onl, value that will chan e )etween a 92mod9 and a 92res9 file is the pit num)er which is returned in the "hittle file2 Due to this0 onl, one attri)ute is re-uired to store the information2 This attri)ute name should )e entered in this field2 If readin into an e$istin model0 then and the attri)ute e$ists0 all current values will )e cleared2

2'2 Block 1alue attri!ute

"hittle Three-D onl, has two pieces of information0 the )loc+ value and the phase num)er2 Therefore0 whether creatin a new model or readin into an e$istin model0 this attri)ute must )e entered2 The )loc+ value will then )e read directl, into the attri)ute iven2 If readin into an e$istin model and the attri)ute e$ists0 then all current values will )e cleared2

2'3 Phase num!er attri!ute

#nter the )loc+ model attri)ute to store the phase num)er from a "hittle Three-D file2 If readin into an e$istin model and the attri)ute e$ists0 then all current values will )e cleared2

2' -etrie1ing parcel information

If this form is applied while the &etrive Parcel Information field is visi)le0 then the I=P8T W6ITT7: #578:S form will )e displa,ed2 In this form0 the values possi)le to read from the "hittle file are displa,ed0 alon with the name of an attri)ute in which to store that value2 Simpl, select a value to )e imported0 and enter the name of the attri)ute to import the value into2 If ,ou are retrievin parcel Bmaterial and elementC information0 ,ou ma, wish to enter a cutoff2 If the units of metal of a particular parcel are )elow that cutoff0 then the parcel is i nored durin processin 2 An, values in the list that ,ou do not wish to read into the )loc+ model0 simpl, remove that line from the scrollin re ion B), ri ht clic+in on the row title and then selectin remove from the pop-up menuC2 That value or material in the "hittle model will then )e i nored2 If ,ou set the &etrive Parcel Information to )e active0 then it is possi)le to read in values for materials and elements into the )loc+ model2 There is also another field active as+in if ,ou wish to store volume factors for all roc+ t,pes0 which will )e e$plained later2 For materials and elements0 two Bor optionall, threeC attri)utes are created for each material2 One will )e called Tmaterial nameUSmetal and the other Tmaterial nameUS rade2 The optional attri)ute would )e called Tmaterial nameUSvolumeNfactor2 See )elow for details of calculations performed when fillin these fields2

2'" Whittle Input 4alculations

The followin outlines the calculations performed when importin each value from a "hittle EDHE/ 2res or 2mod file2 Imported #alue 0escription of #alue 4alculation

M45,

!CAF Bminin cost ad>ustment factorC is a "hittle value used in the optimisation !CAF is read directl, from the process2 If ,ou wish to e$port the model "hittle file )ac+ into "hittle0 then ,ou will need !CAF values2 PCAF Bprocessin cost ad>ustment factorC is a "hittle value used in the optimisation PCAF is read directl, from the process2 If ,ou wish to e$port the model "hittle file )ac+ into "hittle0 then ,ou will need PCAF values2 After the "hittle optimisation0 all )loc+s The pit num)er is read directl, from are sorted into PpitsP which define the the "hittle 2res file optimal minin order2 #ffectivel,0 this pit num)er is the result of the optimisation0 and is therefore the value that is most important to import2 Pit num)ers onl, e$ist

P45,

pitOnum

in 2res files0 so cannot )e imported from 2mod files2 The roc+ num)er is an optional field which can )e stored in a "hittle 2mod file2 It is the roc+ num)er is read directl, from ver, seldom used0 )ut ,ou ma, wish to the "hittle output file import it2 It is not availa)le in the 2res output file2 The specific ravit, of a )loc+2 It assumes the materials are completel, homo enous throu hout the )loc+0 as each )loc+ will onl, contain one S@ re ardless of the num)er of parcels2 The volume factor is a calculated factor which ives an appro$imation of the percenta e of a )loc+ which is ore Bnot waste or airC2 This value is particularl, useful to allow the transfer of data from another software pac+a e to Surpac usin the "hittle interface2 Several other software pac+a es have the concept of partial )loc+s and this information can )e transferred throu h the "hittle interface0 with this value )ein the same as the partial )loc+ tonna e from the other pac+a e2 !ost useful if )loc+s onl, contain ore parcels of the same roc+ t,pe2 The S@ is calculated as the )loc+ tonna e Bread from the "hittle fileC divided ), the )loc+ volume Bcalculated from the )loc+ dimensionsC2

rockOnum

S9

1olumeOfactor

The calculation for the volume factor is the total )loc+ tonna e Bread from the "hittle model fileC divided ), the total sum of the tonna es of all ore parcels in the )loc+2 The value will alwa,s )e a real num)er )etween ( and 52

Cmaterial nameDPmetal

The metal value is contained in the parcel line of a "hittle file2 This The metal content of a "hittle parcel value is therefore >ust a simple descri)es how much of a particular summation of the metal content of all materialHelement is in the model2 This parcels of a particular roc+ t,pe in value can )e in an, units0 defined ), the ED0 or the summation of the metal ori inal model and the "hittle parameters2 content for a particular element in all parcels for a "hittle E/ file2 This value is calculated ), ta+in the Smetal value calculated a)ove and dividin it ), the total tonna e of all parcels Btonna e is stored in all parcels of a )loc+C of a iven roc+ t,pe Bin EDC or all parcels Bin E/C2

Cmaterial nameDPgrade

The rade is reported as the Bunits of metalHtonneC of the )loc+2

#$actl, the same as the volumeNfactor defined a)ove0 the onl, difference )ein This value is calculated as the total that this value is simpl, a percenta e for a Cmaterial tonna e of a particular particular material or element0 not all ore nameDP1olumeOfactor materialHelement divided ), the total as a whole2 The summation of this attri)ute tonna e of the )loc+2 for all materialsHelements should e-ual the volumeNfactor outlined a)ove2 Cop,ri ht J '((A Surpac !ine$ @roup