J.M. Castro Codelco, Chile M. Larraín Codelco, Chile

Transcription

1 Sraegic versus Tacical Approaches in Mining 2011 Y. ovin (ed) 2011 Ausralian Cenre for Geomechanics, erh, ISBN hps://papers.acg.uwa.edu.au/p/1108_04_casro/ J.M. Casro Codelco, Chile M. Larraín Codelco, Chile The common pracice of he medium- and long-erm mine planning is o se he producion plans in erms of he maximal Ne resen Value (NV) as a arge by he use of opimisaion sofware packages. These compuaional ools use simplisic represenaions of he whole producive sysem as a flow nework expressed on a yearly basis. As such, hese ools require of echnical parameers relaed o producion capaciy and flexibiliies of he differen processes, for example, energy consumpion consideraions of he crushing and milling processes, and he use of availabiliy and uilisaion facors as a mainenance programme effecs esimaion. Therefore, when relaively shor sockpiles exis beween mine haulage, crushing and milling processes wih respec o he whole sysem capaciy, he mine and meallurgical planning requires o be suppored by a more deailed and sysemic analysis of hese processes. This paper repors on he acical way El Teniene Mine esimaes is sraegic boleneck process capaciy. This approach inegraes he Teniene 8 rail haulage sysem and Colón processing plan producion plans wih he criical equipmens mainenance programme and exising sockpiles effec ino he whole sysem producion capaciy. El Teniene Mine is a high-onnage underground copper mine locaed in Chile, close o Rancagua ciy. The sraegic Business and Developmen lan (ND) liss he basic requiremens of he producion plans for he differen mines and processes using a mixed ineger-linear programming (MIL) opimisaion model (Caro e al., 2007) like oher mining companies (Trou, 2005). Therefore he operaional feaures, and variey and complexiy of he producive process, he Ferrocarril Teniene 8 haulage sysem requires an inegraed and deailed analysis o esimae he producion capaciy of he overall mine producive sysem (Casro e al., 2008). Currenly, all hese requiremens are assessed by he use of simulaion like oher complex underground operaions (Boha e al., 2008). Afer he las mine-plan expansion (royeco Desarrollo Teniene) implemened in 2005, he El Teniene Mine reached an ore producion rae close o big open pi operaions bu wihou heir sockpiling faciliies. The underground mine has a limied sockpiling capaciy a he ore pass. This is relaed o he difficulies o assure he sabiliy of large excavaions in he high sress rockmass condiions. The Colón crushing plan has he same sockpiling limiaions relaed o is locaion a 1,980 m above sea level in he Andes Mounains, surrounded by cliffs wih difficul access o is railway faciliies a he discharge zone. Any criical equipmen deenion could mean a reducion of he whole sysem capaciy performance as a one-line process. According o hese condiions, mine planning mus assess he effec of hese sockpiling limiaions on he mine-plan inegraed process capaciy as his process represens he sraegic boleneck of he producive sysem. This means ha a deailed mainenance plan analysis of he differen criical equipmen (crushers, mills, conveyor bels) is required. Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia 31

2 A acical way o suppor he long-erm mining plans producion capaciies J.M. Casro and M. Larraín This work aims o show he analysis ha suppors he curren mine planning approach used o inegrae he Teniene 8 rail haulage sysem and Colón processing plan producion plans, wih he criical equipmen mainenance programme and exising sockpiles effec on he whole sysem producion capaciy. A general view of he Colón processing plan faciliies is provided in Figure 1. The firs process is he rail sysem. This sysem carries ou he ore from he underground mine o he primary crushing bin by rains wih coarse ore size cars (run-of-mine ore under 1 m size) o he secondary crushing bin by he use of rains wih fine ore size cars (in-mine crushed ore under 0.3 m size). The fine ore cars can discharge o boh bins, bu he coarse ore cars can only discharge o he primary crushing bin. The design parameers of he railroad equipmen are shown in Table 1. This is followed by he primary crushing process and he ransference ower. The ower faciliaes disribuion of he crushed ore o he SAG-1 grinding circui sockpile, SAG-2 grinding circui sockpile and he secondary crushing bin according o he whole grinding process requiremens. The secondary and eriary crushing process follows he ball mill circui secion 1-12 and secion 13 sockpiles. This crushing process mus supply boh sockpiles according o he whole ball mills circui requiremens. Following he milling processes, is he final, frohing concenraion process. rocess Equipmen Feaures Commens Coarse ore rains 130 Schalke locos ower: [H]; max. speed 60 [Km/h] 48 cars 80 capaciy; max. ore size 1 m; laeral discharge Fine ore rains 130 Schalke locos ower: [H]; max. speed 60 [Km/h] 38 cars 100 capaciy; max. ore size 0.5 m; cenral discharge 5 rains 19 cars per rain 2 rains 19 cars per rain 32 Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia

3 Sraegic lanning and Scheduling The design parameers of he main crushing and grinding equipmen are shown in Table 2 and he bin and sockpile capaciies are shown in Table 3. rocess Equipmen Dimensions ower [H] rimary crushing Giraory crusher N m 500 Giraory crusher N m 800 Sec. and er. crushing Sec. cone crushers H Ter. cone crushers H Ter. cone crushers symons 7SH SAG-1 grinding SAG-1 mill m 15,000 Ball mills m 2 6,000 SAG-2 grinding SAG-2 mill m 26,000 Ball mills m 2 15,000 Ball mills (1 12) Hardinge mills m 7 2,500 Hardinge mills m 1 3,500 Fuller ball mills m 4 3,800 Ball mill (13) m 15,000 Iem Live Capaciy [] rimary crushing bins 2 4,000 SAG-1 sockpile 25,000 SAG-2 sockpile 50,000 Secondary crushing bins 1 8, ,000 Ball mills (1-12) 20,000 Ball mill (13) 5,600 The proposed mahemaical model of he whole sysem producion capaciy is a linear programming model and is feaures are described in he following paragraphs. The objecive funcion is o maximise he whole onnage processed in he evaluaion period, as shown in Equaion (1). TON = T I = 1 i= 1 i, (1) Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia 33

4 A acical way o suppor he long-erm mining plans producion capaciies J.M. Casro and M. Larraín TON = oal processed onnage,. i, = producion processed in he line i a period,. This problem is subjec o he mass balance consrains and echnical consrains ha define he feasible soluions. The mass balance consrains can be classified as equipmen flow balance (Equaion (2)) and bins/sockpiling balance (Equaion (3)). IN j, = OUT j, j, (2) IN j, = producion supplied o he line j a period,. OUT j, = producion processed in he line j a period,. IN S j, = S j, 1 + j, j, OUT j, (3) S j, = sock of maerial in line j a period,. The echnical consrains can be classified as maximal design equipmen capaciies by energy/power consideraions and he minimal capaciies in accordance wih operaional consideraions, as shown in Equaion (4). MIN MAX j, j, j, j, (4) MIN j, = minimal producion capaciy of he line j a period,. MAX j, = maximal producion capaciy of he line j a period,. The maximal design equipmen capaciies are esimaed by Equaion (5) for he ball mills, according o he Bond s model and Equaion (6) for he SAG mills. MAX j, f = 1 10 * W j, * k 80 1 F 80 j, (5) f k = modificaion facors of he Bond s model. W j, = Work index (Bond s model) of he ore supplied o he line j a period, KWh/c. 34 Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia

5 Sraegic lanning and Scheduling 80 = 80% ore size passing produc, microns. F 80 U = 80% ore size passing feed, microns. MAX j, = 24* * U * T ( E * f ) = elecric power of he SAG grinding circui, KWh/. = SAG grinding circui uilisaion, %. p esc e j, (6) T e = SAG grinding circui ype, as fracion. E p = ore specific energy consumpion, KWh/. f esc = scale facors for he SAG grinding circui pilo plan es o indusrial plan. The SAG grinding circuis consider he minimum capaciy of each SAG mill and ball mills circui. The las consideraion is he mainenance programme of he criical equipmen (mills, crushers, conveyor bels) according o he producion requiremens. The mahemaical model consiss of a linear programming model and i is coded and solved in he L Solve opimisaion package sofware using AML (A Mahemaical rogramming Language). The model has been se o suppor a shor-erm producion plan, i.e. a one-year period a differen imeframes: daily, weekly and monhly. The haulage capaciy is provided by he curren simulaor of he Teniene 8 rail operaion and he crushing and grinding process capaciies are provided by inernal sudies of hese processes. The mainenance programme of hese processes is se according o he sandard annual programme provided by he operaions mainenance eam. The firs sep of his analysis is he Teniene 8 haulage capaciy assessmen. Currenly, his sysem is assessed using a discree even simulaion sofware and muliple replicans of a Monecarlo simulaion. For he purpose of his sudy, i was necessary o esimae a deerminisic and represenaive value of he sysem s capaciy. Figure 2 shows he daily producion of a one-year simulaion of he Teniene 8 haulage sysem in curren condiions. This resul shows values beween 111 and 151 k/d, 131 k/d as average and 135 k/d as mos frequen value. Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia 35

6 A acical way o suppor he long-erm mining plans producion capaciies J.M. Casro and M. Larraín % Frequence Cumulaed 100% % Frequence % Cumulaed 60 40% 40 20% Haulage Capaciy, k/d 0% On he oher hand, he haulage capaciy was esimaed using a differen deerminisic approach (wihou random simulaion). The resul was 135 k/d, equivalen o he mos frequen value according o he simulaion resuls. Neverheless, his deerminisic assessmen esimaes he hourly performance of he rails during a one-day operaion and compares i wih he grinding process hourly performance as shown in Figure Haulage Milling 8000 Capaciy, /h Hour of he Day Figure 3 shows a comparison beween he hourly haulage and grinding capaciy. On one hand, he haulage shows a variable capaciy during he day mainly caused by he average cycle of he rains (close o 1.5 hours 36 Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia

7 Sraegic lanning and Scheduling per cycle), he effec of he mine operaions work sysem (3 shifs, 8 hours duraion per day) and he daily railway mainenance (14:00 o 16:00 h). On he oher hand, he grinding shows a sable capaciy only affeced by changes in he ore size disribuion and ore hardness (boh considered fixed in his analysis). The oher parameer affecing he grinding capaciy is relaed o he mainenance of he criical grinding equipmen. In relaion o he difference beween he haulage and grinding hourly capaciy, i is necessary o manage he sockpiles in he sandard operaion of he whole process. The effec of he crushing and grinding process mainenance programme will be assessed for a one-year period. The following analysis of he haulage and processing plan producion capaciy considers wo main issues: he sockpile managemen sraegy analysis and he imeframe analysis. For he sockpile managemen sraegy analysis, we assess differen objecive funcions, hence he model was se o assess hree differen scenarios. Opion 1: maximise he crushing sysem oupu; Opion 2: maximise he milling sysem oupu; and Opion 3: maximise he sum of he crushing and milling oupu. The purpose of his analysis is o find he operaional crieria ha maximise he whole sysem capaciy in differen imeframe schedules. The resuls are shown in he Figure Opion #1 Opion #2 Opion #3 Annual Tonnage, M Daily Weekly Monhly Time Scale Schedule According o he resuls shown in Figure 4, Opion 3 maximises he whole producion capaciy for all imeframe schedules assessed which presens a beer sockpile managemen alernaive. In oher words, he broader he imeframe schedule, he lower he benefi of he sockpile managemen given he daily imeframe schedule shows a value 1.8% lower han he monhly one (almos 1 M per year). Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia 37

8 A acical way o suppor he long-erm mining plans producion capaciies J.M. Casro and M. Larraín Wihou Sockpiles Wih Sockpiles Annual Tonnage, M Hourly Daily Weekly Monhly Time Scale Schedulling Basis Figure 5 shows he analysis of he differen imeframe schedules and he sockpiling managemen effec considering he use of Opion 3 sraegy. As in he previous resul, he lower he imeframe schedule, he lower he producion capaciy. Alhough for an hourly imeframe schedule, which represens he mos deailed feasible one, he resul is very close o he daily imeframe wih sockpiling managemen. This resul is caused for he sockpiling capaciy, which is close o one-day operaion capaciy of he whole process, and for his reason, a more deailed schedule does no reduce he esimaion of he process capaciy. In summary, he Teniene 8 haulage and Colón plan are he sraegic boleneck of he El Teniene Mine in he long-erm. Addiionally, given heir limied sockpiling faciliies, he producion plans mus also consider a more deailed analysis of heir processes as an inegraed sysem. The Teniene 8 haulage sysem is analysed using a deerminisic approach o obain a represenaive and comprehensive capaciy esimaion, insead of he sandard simulaion analysis of his process. By doing his, we can undersand he sockpiling managemen required in he daily operaion of he Colón plan. The inegraed Teniene 8 and Colón plan sysem was analysed for a one-year period, aking ino accoun he mainenance programme of he criical equipmen of boh he processes and differen imeframe schedules using he same daa. The resuls show he necessiy of a leas a daily imeframe scheduling of he whole process given he limied sockpiling capaciy of he sysem, close o a one-day operaion. A monhly imeframe schedule overesimaes he capaciy of he sysem by 1.8% (1 M per year) in comparison wih a daily schedule. The auhors sincerely hank heir colleagues of El Teniene Mine who helped o develop his work. Also, he auhors graefully acknowledge he permission graned by Codelco Chile o publish his echnical paper. 38 Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia

9 Sraegic lanning and Scheduling Boha, J., Wason, S., Arkadius, T. and Samosir, E. (2008) Simulaion applicaions a T Freepor Indonesia s DOZ/ESZ block cave mine, in roceedings 5h Inernaional Conference and Exhibiion on Mass Mining, MassMin 2008, H. Schunnesson and E. Nordlund (eds), 9 11 June 2008, Luleå, Sweden, Luleå Universiy of Technology ress, Luleå, pp Caro, F., Caalán, J., Epsein, R., Goic, M., Sanibañez,. and Weinraub, A. (2007) Long erm opimisaion of invesmen and producion plans in open-pi and underground copper mines, in roceedings 33rd ACOM Symposium, Saniago, Chile, pp Casro, J., Henriquez, J. and San Marín, J. (2008) Simulaion model of Ferrocarril Teniene 8 haulage sysem, El Teniene Mine, in roceedings 1s Inernaional Congress on Auomaion in Mining Indusry AUTOMINING, Saniago, Chile. Trou, L.. (2005) Underground Mine roducion Scheduling Using Mixed Ineger rogramming, 25h ACOM Conference, Brisbane, Ausralia, pp Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia 39

10 A acical way o suppor he long-erm mining plans producion capaciies J.M. Casro and M. Larraín 40 Sraegic versus Tacical Approaches in Mining 2011, erh, Ausralia