3D free fall rock size sensor

Abstract

The energy efficiency of the grinding process is less than 1%, and it consumes about 40% of the energy in a typical open-pit mine operation (CIPEC, 2005). One of the causes of this low energy efficiency is the high variability of the run-of-mine ore hardness and size distribution, which must be measured accurately in order to develop process control strategies that will minimise their impact on mill efficiency. On-line rock hardness sensors are still unavailable and 2D rock size sensors show important limitations under uncontrolled outside lighting conditions. A rock hardness sensor based on primary crusher data analysis has provided interesting results but it requires a more accurate size measurement of the run-of-mine ore, which is a key information for evaluating the reduction ratio and compute the rock hardness index accordingly.Recently, laser 3D profiler has been identified as an interesting technology for measuring rock size distribution over a conveyor belt and they are now commercially available. This paper discusses the extension of this application to the measurement of ore size distribution as it is dumped into the primary crusher. The objectives are (1) to provide an accurate evaluation of the reduction ratio at the primary crushing stage and to allow the development of a rock hardness sensor, (2) to quantify the quality of the blast before any crushing size reduction, and (3) to address the measurement error of the fines associated with the conveyor belt system.A first prototype of the 3D rock sizing technology was evaluated in a rock quarry environment. The prototype was then improved based on the findings. The most important upgrade is a free fall rock- speed evaluation technique, which is based on the spectral analysis of a line-scan image of the rock flow. Despite some accuracy enhancement over existing 2D technologies, the 3D free fall rock-sizing sensor is still, in its industrial arrangement, limited to the measurement of a surface based rock size distribution. This limitation is due to the intrinsic depth sampling error of the 3D sensor, which sees only the front layer of the rock flow and, as a result, overestimates the percentage of coarser rocks. A full volumetric 3D free fall rock-sizing sensor was built in parallel at laboratory scale to overcome this limitation: it measures the entire volume of each piece of rock as they fall in a monolayer flow between two facing laser scanners. The resulting volumetric rock size distribution was found to be coherent to the size distribution measured by sieving, unbiased and, therefore, more accurate than any surface based rock size distribution. The volumetric 3D free fall rock-sizing sensor is proposed for upgrading commercial 2D imaging technologies, which are now used as an alternative to rock sieving in laboratory. A few industrial application scenarios, including its potential integration into an ore-sorting process, are also being investigated.