The liberation and exposure of valuable mineral grains in ore particles are of significant importance in understanding the efficiency of flotation separations. A flotation feed sample from the Kensington concentrator was scanned by micro X-ray computed tomography (micro-XCT), to determine the liberation and exposure of auriferous pyrite grains. The analysis results suggest a satisfying extent of liberation for particle sizes smaller than 212 μm. Theoretical flotation recovery of pyrite for selected particle size fractions was predicted from the three-dimensional liberation and exposure analysis. The plant flotation tail was also characterized by micro-XCT to identify liberated and partially liberated pyrite particles that might not be collected into the concentrate. However, liberation analysis of the tail sample revealed that the majority of liberated pyrite was recovered in the plant flotation circuit. Trainable Weka, a machine learning segmentation approach, was compared with the Watershed thresholding segmentation for the 106 × 45 μm particle size fraction scanned at a voxel size of 1.85 μm. Improved segmentation of small particles was found using the machine learning method.
Full-text paper:
Mining, Metallurgy & Exploration (2023) 40:1621–1630, https://doi.org/10.1007/s42461-023-00852-9
