Sensor-based Sorting
A growing number of mining operations are implementing sensor-based sorting technologies to pre-concentrate excavated material. Sensor technologies can be mounted on shovels, hoppers and conveyors, and sorted as bulk lots of material or on a particle-by-particle basis. Wyman (1966) presented a list of the necessary elements for successful mechanical sorting, they include: 1) a means of presenting the mineral pieces for examination; a sensing device for selecting the pieces to be removed; 3) an electronic system to act upon the information provided by the sensing device; and 4) a means of removing the selected pieces.
Sensor types for mineral applications include prompt gamma neutron activation analysis (PGNAA), optical, X-Ray transmission (XRT), X-Ray Luminescence (XRL), near-infra red, X-ray fluorescence (XRF), electro-magnetic and radiometric sensors (Klein & Bamber, 2018 and Kobzev, 2014).
At Newcrest’s Ridgeway Deeps mine, trials of a nuclear magnetic resonance (NMR) sensor for rapid (20 second) bulk sensing of chalcopyrite, the main copper-bearing mineral, showed that the technology has potential for use in a bulk sorting system through integration with a diverter (Coghill et al., 2018). Teck trialled shovel-mounted XRF sensors at their Highland Valley Copper operations to measure grade information in real time and instruct operators whether loads are to be sent to ore or waste stockpiles (Teck, 2018). Sensor technologies can be mounted on shovels, hoppers, conveyors and slurry systems (Van Haarlem, 2017), and sorted as bulk lots of material or on a particle-by-particle basis.
At the Priargunsky Mine in Russia, both bulk and particle sorting systems are used for beneficiation of uranium ore. On a bulk scale, the grade of material inside rail cars or dump trucks is measured using XRF sensors; material that is above cut-off grade is fed to XRF-based particle sorters for further beneficiation while low grade material is directed to a heap leaching site (Kobzev, 2014)
Ore sorting at the Khumani Iron ore operation is another example where both bulk and particle sorting technologies are combined. At Khumani, the grade of run-of-mine (ROM) material is measured with a PGNAA online analyser. Low grade ROM material is sent to waste stockpiles, intermediate grade material is sent to washing, screening and jig beneficiation circuits, and product grade material bypasses treatment and is sent directly to product stockpiles (Matthews & du Toit, 2011). An advantage of this arrangement is the reduction in capacity requirements of the beneficiation plant. Furthermore, the bulk sorting system reduces fluctuations in the grade of material being sent to the beneficiation plant, allowing the jig beneficiation circuit to be operated more effectively.
PGNAA grade sensors are also used at the Sepon copper-gold operations in Laos. Kurth (2017) describes the technology used at both Sepon and Khumani as having a source of neutrons, Californium-252 located under the conveyor belt, and the emitted neutrons are absorbed by elemental nuclei in the material being transported on the conveyor belt. Each excited nucleus generates a gamma ray having an energy level related to the element from which it has been emitted; detector arrays positioned above the conveyor belt measure the energy of received gamma rays (Kurth, 2017). The measured elemental content of transported material is output to the plant control system that operates diverter gates within the material handling system.
The effectiveness of particle sorting systems relies on the use of grade sensing technologies that are suitable for the feed material (Rule et al., 2015 and Tong et al., 2015). In particle sorting applications, XRF sensors use the interaction of x-rays with rock surface material to determine its elemental composition. Sensor data is used in algorithms to infer the grade of the rock. A sorting decision is made based on the relation between the inferred grade and the setpoint cut-off grade. Sorting is typically carried out through use of blasts of compressed air, or mechanical paddles. XRF-based sorters have been used in precious metal, base metal, ferrous metal, industrial and rare earth ore applications (Tong et al., 2015). The capacity constraints of XRF-based particle sorters support the use of bulk sorting to reduce the quantity of material requiring beneficiation and thereby decrease the number of sorter modules operating in parallel.
Bamber et al. (2008) describe four key components of evaluating the feasibility of sensor-based sorting: ore heterogeneity, sensor response evaluation, sorting analysis and feasibility, as presented in the following figure.
The natural grade heterogeneity of the material being assessed for sorting controls the limit of sorting potential (Duncan, 2016, Kobzev 2018 and Mazhary & Klein, 2015). Gy (2004) defines heterogeneity as occurring in two main forms:
Constitutional Heterogeneity (CH): The variation in the content of a certain component, for example copper or elements related to gangue minerals within individual rock fragments which make up an ore domain. Blending does not affect CH; however, CH increases with comminution.
Distributional Heterogeneity (DH): The variation in the content of a certain component, such as copper, within individual groups of ore (which report to draw-points or surface in a caving application). The sum of all of the groups makes up a lot. The DH parameter is considered to represent the potential to sort ore types at a bulk scale. DH can be reduced by blending ores but is not affected by comminution stages.
CH has been used as an indicator of particle sortability by Robben (2014), and Mazhary and Klein (2015), while DH shows potential for assessment of bulk sorting systems.
As a cave matures, it is expected that the large number of mixing events that have occurred within caved ore results in a lower variation in grade on a bulk scale, represented by a lower DH value, thereby reducing the potential for bulk sorting.
The nature of cave operations makes them ideal for incorporation of sensor-based systems to reject rock and allow the operation to become more dynamic. An integrated mine and mill approach will be necessary to coordinate mine schedules and beneficiation processes so that a satisfactory level of throughput and recovery can be achieved.