Massmin 2004 491 1 INTRODUCTION Cullinan Diamond Mine (previously called Premier Mine) started mining diamonds in 1903. The kimberlite pipe, the largest in South Africa is cut by a flat dipping 75 m thick gabbro sill at approximately 400 m below surface, shown in figure 1, together with the position of current mining blocks. Mining above the sill was initially open cast mining, later long hole benching (early 60âs) and block caving (late 60âs). Below the sill, resources in the BA5 and BB1E mining blocks are currently exploited by retreat panel caving. BA5 and BB1E presently have a combined production of approximately 13,000 tons per day. In BA5, mining started in 1988; 130 m below the gabbro sill and this mining block has a current life of mine until 2005. It is anticipated that the BB1E, where production started in 1996 approximately 230 m below the sill, will cease operations in 2009. In 2005, the mine plans to start producing kimberlite ore from the BB1E Advanced Undercut Cave providing interim tons until the Centenary Cut (previously referred to as C-Cut) commences mining at approximately 900 m below surface and starts production in 2009. Draw Control in the current block caves BA5 and BB1E has always been regarded as strategic (Rood & Bartlett, 1994, Bartlett & Nesbitt, 2000, Nesbitt & Vorster, 2000). A great deal of effort and capital have been spent over the years on infrastructure, computer programs and other tools to monitor the drawn volumes from the cave blocks. There are several reasons to understand why Draw Control is considered important, not only at Cullinan Diamond Mine but also at most caving operations worldwide. 2 VITAL REASONS FOR DRAW CONTROL Cullinan Diamond Mine hasnât experienced serious mud rushes or seismicity and therefore these two supplementary risks dictating Draw Control at many other sites are not discussed. 2.1. Avoid premature gabbro dilution Draw Control is very crucial in avoiding early ingress of the overlying gabbro waste as this premature dilution would reduce the overall ore recovery and shorten the life of cave. Right from the Feasibility Study stage for BA5 and BB1E onwards, fragmentation of the kimberlite was predicted to be coarse and that of the overlying gabbro sill to be fine (Rood and Bartlett, 1994). At Cullinan Diamond Mine, it is anticipated that a recovery of 85 percent of the in situ ore can be achieved if proper Draw Control is exercised. Poor draw practice results in a much lower ore recovery rate as drawpoints will be forced to close earlier than anticipated for two possible reasons. Firstly, cut-offs due the influx of gabbro into the drawpoints adjacent to âoverdrawnâ drawpoints occurs much earlier than planned. Secondly, inconsistent draw practice stimulates migration of material over considerable vertical and horizontal distances, inducing the premature mixing of waste with ore, especially as gabbro fragments are fine and would thus move through the column quickly, (Bartlett, 1998). To compound matters, it was found that the kimberlite fines percentage (a function of kimberlite accelerated weathering) was underestimated at the time of initial block cave feasibilities in the mid 80âs (BA5) and early 90âs (BB1E). 2.2. Minimise gabbro into plant headfeed The gabbro not only sterilises the drawn ore but also causes problems at the Dense Media Separator in the diamond recovery plant due to its high specific density. 2.3. Reduce risk of ore recompaction Lack of good Draw Control results in static columns of cave rock, resulting in ore re-compaction, generating point Santiago Chile, 22-25 August 2004 Abstract Cullinan Diamond Mine (previously known as Premier Mine) currently employs two mechanized block caves and plans another extraction level at approximately 900 m below surface. As with all major block cave operations, challenging situations and difficult events occurred during the implementation phase, unforeseen throughout the project feasibility study. The panel caves in the kimberlite pipe have experienced problems of various types that caused deviations from the original planned mining sequences. In combination with (the traditional) production pressures, numerous problems resulted in not achieving good Draw Control practice. A drastic change to the weekly Draw Control planning was introduced with the buy-in from production personnel and mine management. In conjunction with the improved Draw Control practice, an innovative but simple waste determination process has been developed and implemented. Both the Draw Control and the Waste Management at Cullinan Diamond Mine are currently based on âback to basicsâ principles, straightforward and not controlled (yet) by impressive computer programs.This paper describes some of the geotechnical and practical related difficulties that were encountered during the implementation phase and how those were tackled with varying degrees of success. The importance and impact of Draw Control and Waste Management on the overall mine performance is illustrated. Status of draw control practice and waste Management at Cullinan diamond mine Gert van Hout, De Beers Consolidated Mines, Technical Support Services, Operations Geotechnical Stephen Allen, Mark Breed, John Singleton, De Beers Consolidated Mines, Cullinan Diamond Mine 492 Massmin 2004 loads on the extraction level that damage the production infrastructure. In kimberlite, if the draw of ore occurs more than three months after the undercut has taken place, the material in the drawbell and above the major apex may compact. This ore re-compaction creates uneven loading conditions on the extraction level or excessive localised stresses that adversely affect tunnels. This becomes apparent as support integrity worsens or when tunnels collapse. Experience in the BA5 and BB1E caves shows localised re-compaction of kimberlite becoming a major issue when the drawpoint was closed for rehabilitation for more than six weeks. Re-opening of compacted drawpoints requires a high level of secondary blasting, further damaging the brow area and in some cases it has taken as long as eight weeks to re-open a drawpoint. 2.4. Maximum Draw Zone Interaction In both the BA5 and BB1E, the offset herringbone drawpoint layout (figure 2) was used with the drawpoint spacing across the major apex ranging between 24.2m to 24.7m, possibly creating interaction problems across the major apex in some areas as the Isolated Draw Zone (IDZ) varies from 9m to 11m, depending on the kimberlite rock type. Figure 2 shows the zones of draw at the three different draw interaction modes according to Laubscher (2000). If drawpoints draw in isolation, (the horizontal section of) the zone of material affected by this draw can be approximated by a circle with a diameter equal to the IDZ. If there is even draw between the two drawpoints within the same drawbell, the affected area has an elliptical footprint, one and a half times larger than the IDZ. When there is interaction between adjacent drawbells, the zone of influence enlarges another one and half times. 2.5. Optimal ore fragmentation Uneven draw or drawing at too rapid a rate may lead to very coarse fragmentation, poor grade control and severe waste dilution. As mentioned before, coarse ore fragmentation and fine waste fragments induce rapid gabbro waste movement from the top, through the column, to the drawpoints, thereby sterilising the drawpoint prematurely and creating problems at the recovery plant. It has often been observed that drawpoints with coarse fragmentation were subjected to draw, far above the call allowed by the maturity rules. If material is drawn too quickly, there is not enough âresidence timeâ for the kimberlite material: there is a lack of mechanical interaction between the ore fragments and insufficient communition of primary fragments as material gravitates down the cave column. Figure 2: Interaction modes (a: isolated, b: interaction within drawbell and c: interaction across minor apex) Large blocks cause high hang-up frequencies in the drawpoints, create problems in the ore handling system and have a negative effect on the productivity as well as on the operating costs. Removal of drawpoint hang-ups may result in long down-time and are costly because of the secondary breaking requirements. Oversized ore blocks in the drawpoints must be reduced to reasonable sizes that can be handled by the load-haul-dump (LHD) machines. Back analysis of hang-up data during the period between July 1998 and May 2000 (Rahal and Smith, 2000) revealed that in any given shift, 34 percent of the available Santiago Chile, 22-25 August 2004 Figure 1: Diagrammatic plan and section of the kimberlite pipe showing geology and mining blocks. Massmin 2004 493 drawpoints are hung up and that the vast majority of these hang-ups were cleared within one day. 3 THE HISTORY OF DRAW CONTROL SYSTEMS EMPLOYED AT CULLINAN DIAMOND MINE To impose Draw Control on the current mining blocks, various systems and software packages were implemented. Some systems had more success than others. A comprehensive Draw Control system should at least consist of an integrated system of three major components: an accurate and dynamically updated ore resource database, a reliable vehicle monitoring system and a planning system. 3.1. Ore Resource database The very first version of PC-BC, developed for Cullinan Diamond Mine in 1988, was cumbersome to use: it took a long time (2 to 3 hours) to deplete all drawpoint loads on a daily basis and it did not have the fine graphics displayed by the present version. After some time, mine personnel moved towards spreadsheet type applications to store daily production data from the drawpoints. Later, a Microsoft Access database was employed to store the drawn volumes in combination with the status of each drawpoint. A user interface was then developed to present this data in a graphical format and to allow management to extract comprehensive summary reports. This application called BLOCINFO, also permitted the user also to reconcile the drawpoint production figures with the tons recorded by the treatment plant. In 1999, MinRAS an SQL based product that also contained the mixing algorithms, enabling the Mineral Resource Manager to derive a more accurate and auditable ore reserve statement (Guest et al, 2000). 3.2. Vehicle Monitoring System Reconciling tons, calculated from the loads that are recorded by the vehicle monitoring system (VMS) with the tons from the weightometers at the plant is essential in any good planning or effective Draw Control program to obtain a correct and representative mineral resource database. Various VMS options have been trialled at Premier including a beacon system using micro-wave technology (Nesbitt & Vorster, 2000) followed by a gyroscopic based monitoring system. None of the systems were a hundred percent satisfactory due to technical problems, as well as resistance from production personnel who saw it as a management policing tool. A combination of the radio-based voice communication (utilising the leaky feeder to establish contact between control room at surface and underground) and a manual recording based information system, currently in use, yields the best results. However, it requires production crews to buy into Draw Control. As Cullinan Diamond Mine has different size LHD units, there is an issue with the average bucket factor applied to derive âtons minedâ from âbuckets loadedâ at each drawpoint and until recently, the total tons hoisted were consistently larger than the value based on the recorded buckets. 3.3. Planning System Up until 1998, Production planning at Cullinan Diamond Mine was founded on empirical geotechnical guidelines (Bartlett and Nesbitt, 2000) and did not take into account the full effect of resource and equipment availabilities. The empirical guidelines were derived from extensive block cave experience at Cullinan Diamond Mine and have been described by Bartlett, 1998. A more pro-active approach was initiated in 1998 when the weekly planning also incorporated mining constraints such as LHD availability, ore pass capacity, haulage and hoisting capability. The next draw control related concept established was that of the âideal depletion surfaceâ. A predetermined ideal depletion sequence up until the life of draw determines the short-term (week) draw schedule in that the plan attempts to come as near to the ideal depletion state as possible. At any stage, mining is in function of the ideal depletion profile and the call per drawpoint is influenced by its draw history (corrective call when there was poor draw in the past). Planning on a monthly and weekly basis was also carried out using spreadsheets. These files produced satisfactory output as it catered for the input of all parties involved in Draw Control. It was rather cumbersome to use but has served its purpose to stimulate interaction between geotechnical, mining and mineral resource management departments. Regular meetings were held between these departments to agree upon the monthly production plan produced by these excel files but adherence to this plan could be improved. 4 DISCUSSION OF CURRENT DRAW CONTROL PRACTICE 4.1. The Integrated Draw Control System In 1999, the âIntegrated Draw Control Systemâ (IDCS) with a Mixed Integer Linear Programming (MILP) component described by Guest et al , 2000 was introduced on mine. The long and medium term schedule program based on the MILP was originally developed at Koffiefontein Mine for its Front Cave (Hannweg and van Hout, 2001), and is currently being used in the BB1E Advanced Undercut at Cullinan Diamond Mine. This scheduling module is the first planning optimisation tool in block cave operations that is able to optimise over the life of mine as well as over multi- time periods. It incorporates all geotechnical, mining and financial constraints, in a unique way. Long term production planning for the older BA5 and BB1E blocks is still done by the use of spreadsheets and does not incorporate all ore flow constraints or geotechnical rules, except for the draw rate limits. The reason for the implementation of the excel files instead of MILP is threefold. Firstly, those panel caves are considered too depleted to optimise according to the principles within the MILP. Secondly, the main focus of those caves currently is avoiding tunnel collapses occurring from vertical loading and keeping porosity in the areas that hasnât experienced destructive stress levels. And lastly, the MILP version allowing remote access only became available late 2003. The current Draw Control system at Cullinan does not cater for planning based on maximising NPV as the grades in the ore columns are set to an average column value and the mixing algorithms do not cater for vertical mixing within an ore column. The choice of an average column grade is justified, as there is generally very little resolution in the assumed grade profile vertically across the massive orebody. An optimal (and in South Africa legally prescribed) plan for all mining operations should be based on maximum ore recovery, thus maximising tons instead of maximising NPV, the last being a method relying rather on grade and revenue per carat. Block cave mining is a massive mining operation where principles of selective mine planning, based on financial parameters, cannot determine the production plan. These principles may constrain the schedule but an optimal plan is primarily based on geotechnical considerations. The mixing algorithms mentioned in section 3.1 were developed on site to simulate the ore movements within the cave and the parameters were calibrated successfully: the predicted time of drawpoint closure in the BA5 cave was within one month from the actual date of closure. These Santiago Chile, 22-25 August 2004 494 Massmin 2004 mixing algorithms are still being used but it is expected that REBOP,(Pierce et al, 2004), will be used as the standard tool to update the ore resource database. This tool should then also provide an accurate production waste ingress and grade profile (or carat production graph). 4.2. The back to basics approach Towards the end of 2002, a major drive from management, the âback to basicsâ principle occurred, not only in terms of âkeeping it simpleâ but of increased adherence to the basic Draw Control rules outlined in the Cullinan Diamond Mine Code of Practice, (Bartlett, 2003) and ameliorating interaction between Draw Control and Production Departments. Some of the simple but critical procedures for a successful Draw Control Practice are discussed in this section. With issues of interaction being inherited from design (see section 2), the main aim of the current BA5 and BB1E Draw Control system is to draw equally throughout the caves with all drawpoints in production at one time and to minimise dilution from the gabbro sill. Future optimising codes such as the MILP are seen to be not applicable simply due to the maturity status of the BA5 and the BB1E and past Draw Control practices. The key to the whole short-term Draw Control process (figure 3) is the accumulation, storage, processing and presentation of data. The key relationship is between Draw Control and Production. The culmination of the whole process is presented at the weekly meetings that are held where all parties "strategise" key loading and tunnel or drawpoint rehabilitation patterns around Draw Control. Production and short-term planning understand that Draw Control needs alignment to the official annual production plan, therefore by applying correct Draw Control principles long-term cave management and production targets can be achieved. Figure 3: Process flow chart on Draw Control at Cullinan Diamond Mine. Draw Control Observers gather the occurrence and type of drawpoint hang-ups and record it into the Draw Control system daily (Bartlett & Nesbitt, 2000). This information, in combination with daily production data, enables the Draw Control engineer to analyse the effect of the draw rates and draw history on the frequency and type of hang-up as well as waste entry parameters. Extraction rate limits at Cullinan Diamond Mine have hence been determined in terms of tonnage per day using a maturity classification, based on the accumulative production or life of draw from a drawpoint. Four classes were established and the associated draw rates increased from a maximum of 50 tons per day for a new drawpoint to a maximum of 200 tons per day for a mature drawpoint. As can be seen in figure 4, since the end of 2002, adherence to Draw Control has improved significantly (the average deviation in 2000 was approximately 73%). The graph displays the deviation (actual production - planned), averaged over all BB1E drawpoints with the weighting factor being the planned tons. A value around 10% is considered to be very good. A Production Factor (van Hout & Guest, 2000), based on this data has not been implemented at the mine as it is felt that average weighted deviation is adequate enough to communicate how well Draw Control is adhered to. Figure 4: Weighted Average Deviation (datapoints and trendline) from the weekly Draw Control plan for BB1E cave. Cullinan Diamond Mine personnel are confident that the present recording, storage and presentation of actual mined tonnages in the MinRAS is satisfactory. Initially, the actual production per drawpoint was entered into the system on a weekly basis. This data is now imported on a shift to shift basis, in combination with other information associated with the drawpoint status (produced volume, hang-up type, waste content, remedial support status, etc.). Presentation of the draw data is usually in graphical format, accessible for all people involved in Draw Control to get a clear idea of the mined tonnage profile over a user defined period as well as of the actual drawpoint status. Figure 5 shows an example of the weekly Draw Control plan that has been derived on a Friday afternoon, after consultation between Draw Control and Production. Copies of this sheet are given to the Draw Control Officers (who enter it into MinRAS) and Production Mine Overseers (who distribute it to the Shift Bosses). The Mine Overseers also write this information onto âDraw Control loadingâ whiteboards and compare the planned tons with the actual production data on a shift to shift basis. Listed below are some of the more important Draw Control production principles developed with the "back to basics approach" that are vigorously implemented with the aim of achieve correct block cave management: Santiago Chile, 22-25 August 2004 Massmin 2004 495 ⢠Drawpoints closest to the orebody perimeter (highly depleted or not), are continually loaded, at a reduced rate if necessary. This allows for porosity in the caves and prevents the movement of groundwater from the contact areas to the centre of the ore body. ⢠A strict minimum (120 tons) and maximum (1,200 tons) weekly call per drawpoint is implemented with a maximum daily call of 300 tons. This prevents the loading of all planned production from a drawpoint in a single shift, with no loading during the remainder of the week. ⢠A target for the weighted average deviation (figure 4) has currently been fixed at ten percent with acceptance from both production crews and senior management. ⢠It is strived to achieve equal draw across the major apex and continuous production per drawpoint throughout the week. This ensures achieving a maximum zone of influence (as in figure 2c). ⢠Frequent interactions between Draw Control and production to communicate clearly the availability, waste content and status of drawpoints to anticipate correct loading calls. ⢠Reducing the closure/maintenance time from six to three weeks where possible, to avoid re-compaction. Figure 5: Weekly Draw Plan for the BB1E block cave. The implementation of above principles should dramatically improve fragmentation across the caves, creating further reductions in secondary blasting and the possible achievement of monthly "production" targets at acceptable waste percentage. A secondary function of Draw Control was developed in Feb 2002 and consists of a qualitative assessment of the physical state of drawpoints. The system involved a monthly rating of each drawpoint on the basis of condition, stress damage, water and LHD damage. The ratings are transferred onto mine plans for future analysis. Fragmentation data in line with Laubscherâs, (2000), Rock Mass Rating classification were added to the data collection. The data is currently used in back analysis for the fragmentation analysis programs in order to model future block caves within the same ore body. Another Draw Control function is to maintain waste levels at twelve percent or below. Past sampling practices did not record levels of gabbro in each drawpoint. A waste management system developed by the Mineral Resource Management department assists in the prediction of waste tons and is discussed below. 5 WASTE CALCULATION ON SURFACE 5.1. Current practice For the last five years the official measurement of global waste percentage has been determined by taking a sample of approximately 50 kilograms with a plough sampler per shift on the Sortex tailings stream. The size fraction on this belt is -65mm, +32mm. The sample is washed and hand- sorted into three different piles: internal waste, external waste and kimberlite. The piles are then weighed and their relative percentages calculated. Different tests have been carried out to ascertain that this process of waste determination is appropriate as the following underlying assumptions could be questioned: ⢠The waste percentage in the -65mm to +32mm range is representative for the total size distribution of the headfeed. ⢠One sample of 50kg material per shift is an adequate representation of the entire volume fed into the plant during that shift. ⢠The waste percentage does not depend on the person who performs the test. The first test involved the waste analysis in the +4mm size distribution classes. It is impossible to quickly and accurately recognise rock types in fragments smaller than 4mm by manual visual methods. Results for this analysis were very similar to those of the -65mm to +32mm range. It can therefore be concluded that waste is evenly distributed across the different size fractions and the current fragment size range is adequate. During the second test, samples were taken every 15 minutes from the -65mm +32mm stream. The results from this test did not indicate that the current practice of one sample per shift needed to be adjusted. The last test consisted of an identical series of samples given to four different laboratory assistants. As can be seen in figure 6, results may vary depending on the lab assistant but would not justify the extra cost of additional personnel. Figure 6: Analytical bias due to different lab assistants 5.2. Research and future technologies Research was undertaken to identify technology that could recognise waste accurately in - and possibly remove it from - the ore flow within a short time span. Providers of technology based on Microwave Attenuation, Infra red, Laser Induced Fluorescence, Natural Gamma Emission were approached and three different optical sorters underwent testing and extensive assessments. Santiago Chile, 22-25 August 2004 496 Massmin 2004 All applications were successful in differentiating between waste and kimberlite to some degree on surface (conveyor belts) but none were suitable for underground application. Optical sorter technology, widely used in the food and glass industries, proved to be the most effective method in recognising and ejecting waste from an ore stream. The ability to eject waste from an ore stream offers obvious, additional advantages in ore processing, allowing improved flexibility, an increased ore extraction ratio by recovering more diamonds from previously considered non payable drawpoints, lower crushing costs and ultimately improved revenues. 6 WASTE CALCULATION UNDERGROUND To manage waste effectively, an accurate means of measuring the waste percentage in each drawpoint must be established but the following factors complicate this process: ⢠Dust, generated by LHDâs, makes visual observations extremely difficult. ⢠Washing the ore in the drawpoint with water to get rid of the dust settled on the muckpile can result in rapid disintegration of the ore, thereby biasing the sample taken to determine the waste content. ⢠There are different types of waste and different percentages of waste within different size fractions of ore, making the derivation of an average waste value rather difficult. ⢠The finer the fragment size the more difficult it is to differentiate between the different waste and kimberlite types. The ore fragment sizes depend largely on the rock type and the maturity of the cave. At Cullinan Diamond Mine waste is classified as either internal or external waste. Internal waste is that which slumped back into the pipe during volcanic emplacement of the kimberlite pipe and consists mainly of felsite, Waterberg quartzite, norite and metasediments. The internal waste also includes barren late stage carbonatite dykes. In the metallurgical process these "floats" are separated from the ore stream via a process of dense media separation, as these rock types have a lower specific gravity. External waste consists of the country rock norite and the gabbro sill. These two rock types are mineralogically similar and can differ only slightly in texture. The norite enters the cave through the boundary drawpoints when country rock blocks detach, slump down, and cuts off or sterilizes parts of the resource It also enters the ore flow through the tipping of waste development into ore passes. External waste fragments are known in the metallurgical process as "sinks". Their specific gravity is similar to that of diamond bearing kimberlite, and it therefore reports to the concentrate of the dense media separation. Before 2002, the waste content was measured by visually estimating the waste in each drawpoint on a daily basis. Loss of historical waste data combined with the inaccuracy of visual estimates made it impossible to compare waste estimates from underground with the surface measurement of waste. The test work described in section 5 showed that the sampling methodology used on surface yields adequate waste percentage values. It was therefore decided to change the underground waste determination process from the visual estimation to one similar as applied on surface. This involves samples taken from a drawpoint, transported to a lab analysed and results are entered into a database. 7 WASTE MANAGEMENT UNDERGROUND AND BENIFTS THEREOF As mentioned above, the waste percentage of the headfeed needs to be kept below a critical level, above which the plant recovery would be less optimal. Waste at Cullinan Diamond Mine has become increasing challenging over the last few years, as the current block caves become older and the reserves in the block are depleting fast. Keeping the average waste content below a threshold has historically been addressed by stopping underground drawpoints with waste content higher than 15% but this has sterilised large portions of the kimberlite resource. Waste problems can be controlled to some degree by blending, ore from various sources underground and on surface can assist to draw higher waste drawpoints longer, thereby maximising extraction. Prediction and control of waste tons is a task performed by Draw Control. Multiplying drawn tons per drawpoint with a waste percentage and adding this for all drawpoints across the two caves represents a value that indicates the expected waste tons. Initially (2002) there was an extremely close match between the predicted waste tons and the actual recorded waste tons. But of late, differences of 4% - 5% have been observed, with the predicted values based on the underground results always exceeding the on surface determined results. Benefits of the current waste management system are listed below: ⢠The greatest benefit from a mineral resource perspective is that more accurate forecasts and estimates can be made. Financial contribution of drawpoints can be established and assist in the decision making process, together with geotechnical factors, whether or not drawpoints or tunnels need to be closed. ⢠When production planning is done, the expected waste tons can be calculated for each draw scenario. On the basis thereof, optimal draw and waste percentage can be derived. The estimated waste percentage is of great value to the metallurgical department as unexpected and excessive high density waste negatively affects recovery efficiencies. ⢠A third benefit lies in the improved understanding of the migration of gabbro sill material as the cave depletes. The drawpoint waste data recorded since 2002 is still too scarce but correlation between the column depletion status and waste percentage will be analysed to determine waste ingress curves that can be used in future ore depletion planning. 8 SUMMARY AND CONCLUSIONS Cullinan Diamond Mine has experienced Draw Control challenges, resulting in early waste ingress and serious loading onto production tunnels. After a drive from senior management for âback to basicsâ, Draw Control adherence and ground conditions have improved drastically. Several Draw Control procedures are outlined in this paper. Waste is and always will be a mining and treatment constraint requiring constant management to ensure optimum extraction of ore. The introduction of a system that recognises and can eject waste from the ore flow will reduce the constraint of waste on Draw Control and treatment efficiencies. Cullinan Diamond Mine and the greater De Beers Group are vigorously pursuing the implementation of optical sorting technology, having completed extensive testing. A full feasibility study of the project is being undertaken. ACKNOWLEDGEMENTS The authors are grateful to all their colleagues who helped and supported them during the development and implementation of this work, in particular AR Guest, M Preece, C Baltus, HP Grobler and PJ Bartlett. The authors acknowledge the permission of the Director Operations and Santiago Chile, 22-25 August 2004 Massmin 2004 497 the Cullinan Diamond Mine, General Manager to publish this technical paper. REFERENCES ⢠Bartlett, P.J. 1998. Planning a mechanised cave with coarse fragmentation in kimberlite. PhD. Thesis. University of Pretoria, South Africa. ⢠Bartlett, P.J. & Nesbitt, K. 2000. Draw Control at Premier Mine. Proc. MassMin200, Brisbane. Vol 1: pp, 485-493. ⢠Bartlett, P.J. 2003. Block Cave Code of Practice for Cullinan Diamond Mine. ⢠Guest, A.R., van Hout, G.J., von Johannedis, A & Scheepers, L.F. 2000. An application of linear programming for block cave Draw Control. Proc. MassMin2000, Brisbane. Vol 1: pp, 461-468. ⢠Hannweg, L.A. & van Hout, G.J 2001. Draw Control at Koffiefontein Mine. Proc. VIth International Symposium on Mine Mechanisation and Automation, SAIMM, 2000, Vol 1:pp, 93-96. ⢠Laubscher, D.H.L 2000. Block Cave Manual. ⢠Nesbitt, K & Vorster, J.A. 2000. Premier mine Draw Control and underground vehicle monitoring and control system. Proc. VIth International Symposium on Mine Mechanisation and Automation, SAIMM, 2000, Vol 1:pp, 93-96. ⢠Pierce, M., van Hout, G.J. & Singleton J 2004. Draw Control of the BA5 cave Cullinan Diamond Mine: Back Analysis with REBOP. Proc. MassMin2004, Santiago. ⢠Rahal, D.C. & Smith, M.L. 2000. De Beers Site Visit, unpublished internal report. Corporate Head Quarters, De Beers Consolidated Mines. ⢠Rood, H.R. & Bartlett, P.J. 1994. Mechanized Caving at Premier Mine. Proc. XVth CMMI Congress, Johannesburg, SAIMM, 1994, Vol 1: pp, 219-225. ⢠van Hout, G.J. & Guest A.R. 2000. Production Factor and Draw Control Factor, presentation to the International Caving Study. ⢠Malope, P. 2001, Waste distribution in different size fractions of the kimberlite ore at Premier Mine, unpublished internal report, Premier Mine, De Beers Consolidated Mines. Santiago Chile, 22-25 August 2004
Comments
Report "Status Del Control de Trazado y El Desperdicio en La Administración"