The sttegic nd tcticl vlue of 3D geotechnicl model fo mining optimiztion, Anglo Pltinum, Sndsloot open pit by A. Bye* J o u n l Synopsis Sndsloot open pit is situted on the nothen limb of the Bushveld Complex, 250 km noth est of Johnnesbug, South Afic. Sndsloot is cuently the lgest open pit pltinum mine in the wold nd is one of six potentil open pits in the e. The geotechnicl stength popeties of the Plteef hve posed significnt mining nd pocessing chllenges to the PPRust opetion since inception in 1992. In ode to optimize the whole business s opposed to isolted cost centes mine to mill inititive ws embked upon in 1997. Initil wok focused on oe chcteiztion nd n entie suite of dt ws collected fo ech ock type. This included minelogy, geochemisty, geotechnicl, blstbility indices, s well s metllugicl stength dt. The second phse involved the cle definition of custome equiements, nmely the optimum pefomnce citei fo the plnt nd mining deptments. This involved extensive field tils nd dt cptue to mesue the intection nd influence of the mine to mill vlue chin. Once the design tgets wee defined, 3D geotechnicl model ws constucted to ensue tht the tgets wee consistently chieved. The model is vey simil to n oe eseve model; howeve, it is populted with geotechnicl, metllugicl nd blst index dt. Though softwe queies the ppopite blst powde fcto fo ech mining block cn be obtined, theeby ensuing the coect fgmenttion is blsted fo the plnt nd mining opetion. The inititive hs deliveed significnt esults showing impovements coss the mine to mill vlue chin. These include incesed loding tes, educed electicity nd cushing consumbles, s well s highe plnt thoughput. Additionlly, the inititive hs benefited both bownfield nd geenfield pojects on the PPRust opetion Intoduction Geology nd the detiled undestnding of its popeties e fundmentl to the optiml design nd successful opetion of ny mine. To tht end, extensive fieldwok ws conducted t Sndsloot open pit (PPRust Anglo Pltinum) to collect geotechnicl infomtion both fom explotion boeholes nd in-pit mining fces. Ove fie-ye peiod, geotechnicl dt wee collected fom 29, 213 m of explotion coe nd 6 873 m of exposed mining fces. Extensive field nd lbotoy testing ws undetken in ode to define the complete set of geotechnicl popeties fo ech ock type in the Sndsloot mining e. The geotechnicl infomtion collected ws stoed in the Dtmine mining softwe pckge. The chitectue of the dtbse ws developed long the pinciples used fo geneting n oe eseve model. The geotechnicl dt, nmely mining ock mss ting system (MRMR), unixil compessive stength (UCS), fctue fequency pe mete (FF/m) nd ock qulity designtion (RQD), ws modelled using geosttistics to genete 3D geotechnicl model. Dt ws intepolted between explotion boeholes nd exposed mining fces. The modelling ws constined, using wiefmes, by ock type nd mjo stuctul fetues. The esult is block model contining 15 m3 blocks of intepolted geotechnicl infomtion. The size of the model blocks e linked to the mining bench height of 15 m. By hving detiled geotechnicl infomtion vilble in 3D model, which cn be edily ccessed nd intepeted, significnt poduction optimiztions, fesibility studies nd plnning inititives cn be implemented. Fom slope design pespective, the model is used to tget dt deficient zones nd highlight potentilly wek ock mss es. As this cn be viewed in 3D, the open pit slopes cn then be designed to * Anglo Pltinum. The South Aficn Institute of Mining nd Metllugy, 2006. SA ISSN 0038 223X/3.00 + 0.00. This ppe ws fist published t the SAIMM Semin, Sttegic vesus tcticl ppoches in mining, 19 21 Septembe 2005. P p e The Jounl of The South Aficn Institute of Mining nd Metllugy VOLUME 106 NON-REFEREED PAPER FEBRUARY 2006 97
The sttegic nd tcticl vlue of 3D geotechnicl model fo mining optimiztion ccommodte the poo qulity e befoe it is excvted. It lso follows tht geotechnicl zones cn be edily identified nd the slope optimized ccodingly. Rthe thn viewing the dill nd blst deptment s n isolted cost cente nd focusing on minimizing dill nd blst costs, the study focused on the fgmenttion equiements of the comminution nd lod nd hul business es. It is well undestood tht chemicl enegy is the chepest fom of comminution nd tht mjo downstem benefits cn be deived by incesing dill nd blst expenditue. 238 blsts wee ssessed to detemine the optimum fgmenttion equiements fo oe nd wste. Bsed on the study, men fgmenttion tget of 150 mm ws set fo delivey to the cushing cicuit nd men fgmenttion of 230 mm ws set fo wste loding fom the pit. Substntil benefits hve been elised in the dill nd blst deptment by developing empiicl coeltions, which elte the MRMR vlues in the geotechnicl model to blstbility index, fgmenttion, equied powde fcto nd costs. As the geotechnicl model cn pedict chnges in geotechnicl conditions, the blsting pmetes cn be djusted in dvnce to ensue the lod nd hul nd comminution plnt s fgmenttion equiements e met. PPL opetes utogenous mills, which e sensitive to the fgmenttion pofile deliveed. The hde zones occuing in the oe zone theefoe hve mjo impct on the plnt s pefomnce. Accodingly, these zones e identified nd dditionl blst enegy is intoduced in ode to chieve the tget fgmenttion fo the plnt. The 3D geotechnicl model llows these optimiztions to be undetken poctively by the dill nd blst deptment. Geotechnicl oveview The Plteef is pyoxenite oebody hosted within the nothen limb of the bsic igneous ocks of the Bushveld Complex (Figue 1). It hs n economic stike length of 40 km nd contins pltinum goup elements, gold, coppe nd nickel. The Plteef is cpped by the Min Zone hngingwll sequence consisting of gbbonoites. This in tun is ovelin by Uppe Zone sequences of feogbbos. The mineliztion is hosted pedominntly within pyoxenite nd ppyoxenite. The ppyoxenites e confomble with the footwll of the Plteef nd e essentilly contminted metmophosed pyoxenite fomed between the cold county ock nd the Plteef intusive phse. The footwll to the Plteef in the Sndsloot open pit consists of metdolomite, known geneiclly s clc-silicte. Intection of bsic mgm with the footwll sediments of the Tnsvl Supegoup nd vying degees of ssimiltion hs esulted in unique suite of hybid ock types, which povide host of engineeing geologicl chllenges. The open pit is distubed by two noml fults nd mjo oblique sinistl fult. Thee e thee mjo joint sets tht influence the Sndsloot e. These lge folition plnes dip steeply, nd e ltelly nd veticlly continuous ove hundeds of metes. They not only pose slope stbility poblems but lso ffect dill nd blst pefomnce. Tble I documents the esults fom the extensive geotechnicl pogmme tht ws undetken fo the vious ock types occuing t Sndsloot. It is evident fom the esults tht the oe zone contins nge of ock hnesses nd elstic popeties, which hve mjo impct on dill nd blst s well s comminution pefomnce. YOUNGER FORMATIONS ACID PHASE BUSHVELD MAFIC COMPLEX PHASE OLDER FORMATIONS SANDSLOOT PIETERSBURG POTGIETERSRUS ATOK GABBRO-NORITE SANDSLOOT OPEN PIT 0 1 2km PLATREEF OUTCROP DOLOMITE FOOTWALL DOLOMITE TONGUE AMANDELBULT NORTHAM UNION SOUT AFRICA PIETERSBURG JOHANNESBURG RUSTENBURG WESTERN PLATS JOHANNESBURG 0 25 50km PRETORIA WITBANK CAPE TOWN DURBAN Figue 1 Loction of Sndsloot open pit nd genel geology of the Bushveld Complex 98 FEBRUARY 2006 VOLUME 106 NON-REFEREED PAPER The Jounl of The South Aficn Institute of Mining nd Metllugy
The sttegic nd tcticl vlue of 3D geotechnicl model fo mining optimiztion Mining pocess The Plteef oebody t Sndsloot is tbul in geomety, dips t 45 nd is ppoximtely 50 m in width. These popeties llow the oebody to be excvted by open pit mining methods, which is considebly chepe thn conventionl undegound mining. The stipping tio is the tio between the wste nd oe mined o the mount of wste tht hs to be emoved in ode to ccess the oe. Due to the 45 dip of the oebody, the stipping tio is eltively high t 8:1, when comped with open pits such s Phlbo which hve much lowe stipping tio of 1.5:1. Sndsloot open pit is the lgest open pit pltinum mine in the wold. In single month the mine cushes ove 425 000 tons of oe nd excvtes 50 million tons of oe nd wste nnully. The open pit is in the pocess of thid cut, which hs depth of 200 m below sufce. Subsequent cutbcks will extend this to mximum of 320 m. The benches e 15 m in height nd mining blocks e 100 m x 50 m. The pit is scheduled to expnd in seies of phsed cut-bcks until mining ceses duing 2010, when it is pedicted tht the economic cut-off between open pit nd undegound will be eched. The poductivity of mining equipment depends lgely on the blst fgmenttion size. An economic blnce hs to be found between the vey high loding tes poduced fom highly fgmented ock mss, nd the dill nd blst costs ssocited with poducing such fgmented ock mss. Add into this the cushing nd milling benefits ssocited with vey fine blsting of the oe nd it cn be seen tht chieving minimum mining cost is not s simple s optimizing isolted cost centes such s the quntity of explosive used. The pocess is, howeve, simplified by hving detiled geotechnicl, geomechnicl nd minelogicl infomtion fom the ock mss. Fom this infomtion, mining blocks cn be ssigned optimum dill nd blst configutions tht not only impove comminution nd loding tes but lso the oe concenttion pocess. It follows tht n integted ppoch to mining, which involves ll cost centes, is needed to successfully educe ovell mining costs nd impove poductivity. In ode fo mining compny to sty competitive in the moden economy, it is essentil tht it opetes t the lowest possible cost. Mining compnies e theefoe constntly stiving to educe the opeting costs of the mining equipment by impoving the equipment s pefomnce. By poviding compehensive geotechnicl infomtion, in the fom of 3D model, equipment equiements cn be ccutely defined nd theefoe pefomnce nd mining efficiencies cn be optimized. Dt collection The development of detiled geotechnicl dtbse t Sndsloot ws bsed upon extensive field mpping using line suveys, s well s geotechnicl fce-mpping nd logging of explotion boeholes. Ove five-ye peiod, geotechnicl dt wee collected fom 29 213 m of explotion coe nd 6 873 m of exposed mining fces. Geotechnicl mpping involved the visul seption of mining fce into simil geotechnicl zones bsed on ock type nd stuctue. Ech zone ws then mpped individully nd ll the dt equied to te the zone, using Lubsche s (1990) mining ock mss ting system (MRMR), ws collected. This included line suvey nd mjo stuctul infomtion such s ock qulity designtion (RQD), joint oienttion, oughness nd continuity. Lbotoy smples, schmidt hmme nd point lod edings wee used fo clcultion of unixil compessive stength (UCS). The dimond dill explotion boeholes e logged fo the sme infomtion s descibed bove. Thee is theefoe common set of dt between the open pit fces nd the explotion boeholes fom which intepoltion nd pedictions cn be mde. The dt collected ws evluted nd the deived geotechnicl infomtion then stoed in Dtmine dtbse. Dtmine is one of the numeous mining softwe pckges vilble tht e designed to model oe eseves nd pln mining opetions. J o u n l P p e Tble I Geotechnicl popeties of the Sndsloot ock types Hngingwll Oe zone Footwll Rock type Noite Pyoxenite P-pyoxenite Sepentinized ppyoxenite Clc-Silicte UCS (MP) 190 160 200 250 130 Tensile stength (MP) 9.5 8.0 10.0 13.5 7.0 Young s modulus (GP) 82 72 134 108 50 Poisson tio (m) 0.232 0.173 0.218 0.264 0.288 Density (kn/m3) 2.9 3.2 3.3 3.1 2.9 RQD% 80 65 75 70 55 FF/m 9 13 10 11 16 MRMR 55 48 51 56 47 MRMR clss (III A) (III B) (III A) (III A) (III B) Dillbility clss Low Medium-high Vey-low Low Medium-high Absive clss Medium-high Low High Medium-high Low The Jounl of The South Aficn Institute of Mining nd Metllugy VOLUME 106 NON-REFEREED PAPER FEBRUARY 2006 99
The sttegic nd tcticl vlue of 3D geotechnicl model fo mining optimiztion Figue 2 Flow digm illustting the model development pocess The boehole nd fce-mp dtbse wee combined nd the dt ws then geosttisticlly intepolted between the boehole nd fce-mp infomtion. The fce-mps wee essentilly teted s hoizontl boeholes. The dt wee nlysed using histogms nd semi-viogms to detemine the most suitble intepoltion method to use fo ech modelled pmete, nmely MRMR, UCS nd RQD. Due to the bell cuve ntue of the dt sets nd the dt being eltively evenly spced, the invese distnce intepoltion method ws used s fist pss. As the dtbse expnds, futue pocessing my equie moe dvnced methods such s kiging. The models fo ech ock type nd model pmete wee combined into single model contining ll the infomtion. The esult is 3D model with 15 m3 blocks tht contin intepolted geotechnicl infomtion such s MRMR, UCS, RQD vlues s well s the ock type. Empiicl clcultions nd ppliction of the model Rthe thn viewing the dill nd blst deptment s n isolted cost cente nd focusing on minimizing dill nd blst costs, the study focused on the fgmenttion equiements of the pocessing plnt nd lod nd hul business es. It is well undestood tht chemicl enegy is the chepest fom of comminution nd tht mjo downstem benefits cn be deived by incesing dill nd blst expenditue. Ove the lst thee yes blst pttens hve been djusted in ode to find n economic blnce between dill nd blst costs nd ovell mine poductivity. Using the Modul Mining Tuck Disptch System, 238 blsts wee ssessed to detemine the optimum loding tes fo the equipment t Sndsloot. The method used ws instntneous loding tes, which mesues the effective time tht the shovel bucket ws in the muckpile nd is elted to tons pe hou. In othe wods, it is the ese with which blsted muckpile cn be loded. The tget loding te fo the equipment ws set t 3200 t/h. The study eveled tht fom 169 wste blsts powde fcto of 1.17 kg/m3 yielded n vege lod te of 3 182t/h nd 69 oe blsts yielded nd vege loding te of 3 316 t/h fom n vege powde fcto of 1.56 kg/m3. The loding tes wee then elted to fgmenttion nd selection of the blsted muckpiles wee ssessed using the SPLIT desktop digitl fgmenttion nlysis softwe. Fom the fgmenttion nlysis it ws detemined tht the blst pttens yielded men fgmenttion of 150 mm in oe nd 230 mm in wste. The powde fctos nd chieved fgmenttion in the oe blsts my seem excessive but this is due to the equiements of the utogenous milling system t PPL. Bye (2000) discussed in detil the substntil finncil benefits tht hve been deived t PPL by hnessing chemicl enegy in the pit to incese the thoughput in the plnt. Bsed on this detiled study, men fgmenttion tget of 150 mm ws set fo delivey to the cushing cicuit nd men fgmenttion of 230 mm ws set fo wste loding fom the pit. In ode to incopote this infomtion into the geotechnicl model, Cunninghm s (1986) fgmenttion equtions wee utilized. Cunnighm s eqution (Eqution [1]) gives pediction of the men fgmenttion bsed on the explosive enegy nd ock popeties. Howeve, 100 FEBRUARY 2006 VOLUME 106 NON-REFEREED PAPER The Jounl of The South Aficn Institute of Mining nd Metllugy
The sttegic nd tcticl vlue of 3D geotechnicl model fo mining optimiztion t Sndsloot thee ws detiled geotechnicl model nd known fgmenttion tgets. It theefoe mde sense to ewok the Cunnighm eqution in ode to poduce equied enegy fcto eqution (Eqution [2]) bsed on the inheent ock popeties nd fgmenttion tgets. In ode to obtin equied enegy fcto pe geotechnicl model block, Lilly s (1986) blstbility index (BI) ws deived fom the geotechnicl infomtion contined in the model. The BI (Eqution 3) ws then used to clculte the ock fcto fo input into the equied enegy fcto eqution (Eqution 2). The equied enegy fcto eqution (Eqution 2) is deived fom the tget fgmenttion, ock fcto, explosive eltive weight stength, ock volume nd mss of explosives. 08. X A V 0167. RWS = x x Q x Q 115 0. 633 X = Pedicted men fgmenttion dimete (cm) [ ( )] = ( 0. 167 ) (( ) 0633. ) EF 1/ X / AxQ x RWS / 115 125. EF = Requied enegy fcto (kg/m3) X = Requied men fgment dimete (cm). A = 0.12 x BI (ock fcto) Q = Mss of explosive pe blsthole (kg) V = ock volume o yield (m3) RWS = Reltive weight stength of explosive (ANFO = 100) [1] [2] BI = 0.5( RMD + JPS + JPO + RDI + S) [3] BI = Blstbility index RMD = Rock mss desciption JPS = Joint plne spcing JPO = Joint plne oienttion RDI = Rock density influence (25 x density-50) S = Rock stength (0.05 x unixil compessive stength). It becme ppent tht the model could be used to pedict costs nd, bsed on the stndd dill nd blst costs used, Sndsloot coeltion ws developed between the equied enegy fcto (Ef, kg/m3) nd cost pe cubic mete (Eqution 4). These costs included dilling, explosives, lbou nd mintennce. Figue 3 illusttes plnned blst outline ovelin on the geotechnicl model. A set of ctegoy filtes cn be seen, which elte the model block colous to the blstbility index, equied enegy fcto nd costs. Figue 4 illusttes gphiclly the effect of the blstbility index on the equied enegy fcto nd the ssocited dill nd blsts costs. 3 Cost pe m = Ef x 2.946 3 Ef = Requied enegy fcto ( kg/m ) [4] An dditionl function ws built into the model to clculte stble inte-mp ngle bsed on the MRMR vlues within the model. The eqution (Eqution 5) fo deiving stble slope ngle ws tken fom the design cht J o u n l P p e Figue 3 Pln view of blst outline with colou coded infomtion on blstbility index (BI), equied enegy fcto (EF) nd cost/m3 deived fom the model The Jounl of The South Aficn Institute of Mining nd Metllugy VOLUME 106 NON-REFEREED PAPER FEBRUARY 2006 101
The sttegic nd tcticl vlue of 3D geotechnicl model fo mining optimiztion developed by Hines nd Tebugge (1991). A fcto of sfety of 1.2 ws used to clculte 100 m high, stble intemp ngle fom the MRMR vlues contined in the block model. Slope ngle = 0.4456 x MRMR + 37.226 [5] MRMR = mining ock mss ting vlue Stewt nd Kennedy (1971) showed tht it ws not only the steepness of the ultimte slopes in n open pit mine tht hd n influence on the ovell pofitbility of n opetion. They contended, on the bsis of csh flow clcultions, tht thee is fequently consideble economic dvntge to be gined fom using steep slopes duing the initil stipping pogmme. This is pticully the cse t Sndsloot whee the stipping tio plys lge ole in pofitbility. The model cn be used to optimize the pit slopes the thn pplying single design to the entie pit. A one degee slope optimiztion will genete n dditionl $46 million in evenue duing the life of Sndsloot pit. A Visul Bsic font end in Dtmine is used to quey the model fo enegy fctos, costs, etc. The plnned dill ptten e cn then be evluted ginst the model to give summy of ll the model infomtion fo tht plnned mining block. Additionlly dxf file cn be expoted fom the model so tht ptten nd blst design cn be undetken in othe softwe pckges. Mine to mill esults Open pit mining involves pocess of contolled destuction of the ock mss so tht the wste my be stipped nd the oe extcted. The blsting enginee is fced with the conflicting equiements of poviding lge quntities of wellfgmented ock fo the pocessing plnt, educing dill nd blst costs nd minimizing the mount of dmge inflicted on the ock slopes left behind. A esonble compomise between the conflicting demnds cn only be chieved if the blsting enginee hs vey sound undestnding of the fctos tht contol ock fgmenttion, highwll dmge nd slope stbility. This undestnding ws significntly enhnced though the use of 3D geotechnicl model. Mnul infomtion systems used fo design equie significnt dediction nd time commitments nd cn be oneous to continully updte. They often ely on the commitment of single individul nd e theefoe not sustinble. The 3D geotechnicl model is use-fiendly nd sustinble tool, which cn be edily updted nd theefoe does not suffe fom the limittions of mnul system. Figue 5 illusttes the vege loding tes fom 1999 to 2003. Thee is cle distinction in pefomnce befoe nd fte the fgmenttion model ws pplied. The vibility in loding tes is clely evident pio to 2002 nd this is due to the blst designs not tking into ccount the vitions in ock mss conditions. Afte 2002 the loding te is bove the design tget evey month. The ppliction of the fgmenttion model to blst design esulted in n 8.5% nd 8.8% impovement in loding nd milling tes, espectively, fom 2001 to 2003. It must be stessed tht these e ctul poduction figues mesued ove two-nd--hlf ye peiod nd theefoe epesent significnt ecod of pefomnce. A moe detiled nlysis of ll the custome pefomnce mesues ws undetken fo the peiod fom Jnuy to June 2003 (Figues 6 nd 7). Thee is cle impovement coss ll the pefomnce indictos, which include the following: Ef (kg/m3) o Cost/m3 Effect of blstbility index on cost/m3 nd Ef (kg/m3) fo set fgmenttion tgets 6.0 5.7 TARGET FRAGMENTATION: Ef-Wste ORE P50 = 150mm 5.4 WASTE P50 = 230mm 5.1 COST/m3-Wste 4.8 4.5 Ef-Oe 4.2 3.9 COST/m3-Oe 3.6 3.3 3.0 2.7 2.4 2.1 1.8 1.5 1.2 0.9 0.6 40 42 44 46 48 50 52 54 56 58 60 62 BLASTIBILITY INDEX Figue 4 Effect of blstbility index on costs nd equied enegy fcto fo set fgmenttion tgets 102 FEBRUARY 2006 VOLUME 106 NON-REFEREED PAPER The Jounl of The South Aficn Institute of Mining nd Metllugy
The sttegic nd tcticl vlue of 3D geotechnicl model fo mining optimiztion Instntneous Loding Rte (t/h) 4,400 4,200 4,000 3,800 3,600 3,400 3,200 3,000 2,800 2,600 2,400 2,200 2,000 Actul Loding Rte (t/h) Tget Loding Rte (3,200 t/h) Line (Actul Loding Rte (t/h) Poo blst design not tking into considetion chnges in geotechnicl conditions Loding tes consistntly bove tget duing 2002 Appliction of blsting model (Jn - 02) Jul-99 Aug-99 Sep-99 Oct-99 Nov-99 Dec-99 Jn-00 Feb-00 M-00 Ap-00 My-00 Jun-00 Jul-00 Aug-00 Sep-00 Oct-00 Nov-00 Dec-00 Jn-01 Feb-01 M-01 Ap-01 My-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Dec-01 Jn-02 Feb-02 M-02 Ap-02 My-02 Jun-02 Jul-02 Aug-02 Sep-02 Oct-02 Nov-02 Dec.02 Jn-03 Fb-03 M-03 J o u n l P p e Figue 5 Loding tes showing the impct of the fgmenttion model fo blst design Impct of blst powde fcto on utonenous gind milling tes (2003) 180 172.7 2.0 170 166.8 1.9 Milling Rte (t/h) 160 150 140 130 120 145.6 143.8 164.6 157.7 Milling Rte (t/h) Powde Fcto (kg/m3) 1.8 1.7 1.6 1.5 1.4 1.3 1.2 Powde Fcto (kg/m3) 110 1.1 100 Jn-03 Feb-03 M-03 Ap-03 My-03 Jun-03 1.0 Figue 6 Loding nd milling pefomnce fom 2001 to 2003 Milling Rte (t/h) 300 280 260 240 220 200 180 160 140 Avege Plnt Milling Rte (t/h) [18% Impoved] Avege AG Milling Rte (t/h) [16% Impoved] Avege Instntneous Lod Rte Oe & Wste (t/h) [13% Impoved] Avege Instntneous Lod RteOe (t/h) [11% Impoved] 3,900 3,700 3,500 3,300 3,100 2,900 2,700 Instntneous Loding Rte (t/h) 120 Jn-03 Feb-03 M-03 Ap-03 My-03 Jun-03 2,500 Figue 7 Gph illustting the pefomnce of the dill nd blst deptment s two customes duing 2003. The loding tes include oe nd wste nd the milling tes include both the AG nd bll mills The Jounl of The South Aficn Institute of Mining nd Metllugy VOLUME 106 NON-REFEREED PAPER FEBRUARY 2006 103
The sttegic nd tcticl vlue of 3D geotechnicl model fo mining optimiztion Avege plnt milling te (AG nd Bll mills) 18% impovement Avege AG milling te 16% impovement Avege instntneous loding te (oe nd wste) 13% impovement Avege instntneous loding te (oe) 11% impovement. These pefomnce impovements epesent substntil vlue dd to the ovell business nd the ssocited finncil benefits e significnt in tems of millions of nd pe month. The impovement in the utogenous gind milling pefomnce fo the 18-month model ppliction peiod fom Jnuy 2002 to June 2003 ws ecoded. The vege milling te fo 2001 ws 156 t/h, while this impoved by 5.5 t/h to n vege of 161.5 t/h in 2003. The dditionl evenue geneted by this incesed efficiency fo the eighteen-month peiod ws $ 4.5 million. This esech illusttes the impovement in business efficiencies tht hs been elized t Sndsloot, not fom estuctuing but by ssessing the compny s totl business pocess nd defining custome focus fo the dill nd blst deptment. This custome focus ws fcilitted by the use of fgmenttion model. Conclusions A 3D geotechnicl model hs ppliction to ny mjo civil o mining ventue tht equies detiled undestnding of the vibility in ock mss conditions. A geotechnicl model does not popose to genete solutions by ceting infomtion fom limited dt set. It does, howeve, give the enginee tool wheeby he cn ssess the sptil vibility of the ock mss infomtion nd theeby identify dt-deficient o high-isk es. Thee e numeous cse histoies detiling the filue o significnt ove-expenditue of civil, tunnelling nd mining pojects cused by lck of knowledge of the vibility of the in situ ock mss. The 3D geotechnicl model povides infomtion well hed of the mining fce, which cn then be used fo ock qulity pediction, poduction optimiztion, slope evlution nd design, s well s plnning nd costing. Using simil quey function s the oe eseve model, mining slots cn be evluted nd not only gde nd tonnge figues deived but pedictions of penettion tes, powde fctos, pesplit nd blst designs, s well s equipment nd explosives equiements. The mining costs could be boken into dilling, blsting, cushing nd milling costs, bsed on expected powde fctos, penettion, cushing nd milling tes, theeby futhe optimizing pit plnning nd expenditue. Moe detiled costing nd budgeting cn be undetken especilly in espect to comminution nd dill nd blst costs. Thee is the potentil fo simil geotechnicl pogmme nd the ppliction of geotechnicl model to be eqully pplicble to undegound opetions. Additionlly, thee is consideble scope fo the implementtion of these methods s tool fo mine evlution nd fesibility ssessments of new oe deposits. The oe eseve model hs gined widesped cceptnce s n invluble tool to mining opetion. Cetinly most finncil ogniztions will not invest in mining poject tht does not hve n oe eseve model. Thee is the potentil tht the sme cceptnce s vitl tool to the mining pocess will follow the development of geotechnicl models. In the ce fo educed mining costs nd incesed poductivity, the development of geotechnicl model povides cost-effective tool to impove poductivity nd educe mining costs. The Sndsloot cse study illusttes this with $4.5 million in dditionl evenue geneted ove n eighteen-month peiod though the mine to mill cmpign. Refeences BYE, A.R. The ppliction of geotechnicl infomtion to dill nd blst opetions t Sndsloot open pit. Bulk Mining Explosives, Annul Confeence, Petoi. 2000. CUNNINGHAM, V.B. The Kuz-Rm model fo pediction of fgmenttion fom blsting. Poc. 1st inten. Symposium on ock fgmenttion by blsting, Lule, Sweden. 1986. pp. 439 52. HAINES, A. nd TERBRUGGE P.J. Peliminy estimtion of ock slope stbility using ock mss clssifiction systems. Poc. 7th Int.Cong. Int. Soc. Rock Mech. Achen. Blkem, Rottedm, vol. 2, 1991. pp. 887 892. LAUBSCHER, D.H. A geomechnics clssifiction system fo the ting of ock mss in mine design. J. S. Afi. Inst. Min. Metll., vol. 90, 1990. pp. 257 273. LILLY, P.A. An empiicl method of ssessing ock mss blstbility. The Aus.IMM-IE Aust. Newmn Combine Goup, Lge Open-pit Mining Confeence. Octobe, 1986, 1986. pp. 89 92. Özcęlik, Y. Effect of discontinuities on fgment size distibution in open-pit blsting cse study. Tns. Intsn. Min. Metll. (Sect. A: Min. Industy), 107. 1998. pp. A146 A151. STEWART, R.M. nd KENNEDY, B.A. The ole of slope stbility in the economics, design nd opetion of open pit mines. Poc. 1st Symposium on Stbility in Open pit Mining, Vncouve 1970. A.I.M.E. New Yok, 1971. pp. 5 21. 104 FEBRUARY 2006 VOLUME 106 NON-REFEREED PAPER The Jounl of The South Aficn Institute of Mining nd Metllugy