A new analysis method based on serial sectioning and three-dimensional (3D) reconstruction was developed for the quantification of minerals in an iron ore pellet. The morphology and spatial distribution of the minerals in 3D space were analysed via 3D reconstructed images of an iron ore pellet. The volume fraction of the minerals in the 3D image was also calculated based on the pixel points. The results showed that the morphology and spatial distributions of hematite, magnetite and silicate varied among different pellets as well as among different positions within the same pellet. Thick plate and interconnected hematite was observed in the outer area of the pellets, whereas the inner area of the pellets mainly contained small granular and independent hematite. The hematite grains in pellet 1 were small, whereas those in pellet 2 were relatively dense. Some of the magnetite in pellet 1 was in the form of dense blocks, whereas fewer blocks were found in pellet 2, where magnetite occurred mainly in chain form scattered within the large particles of hematite. The silicate in pellet 2 was dense, granular and smaller than that in pellet 1. The quantitative analysis results for the two kinds of pellets showed that pellet 2 contained more hematite. Moreover, pellet 2 contained less unoxidized magnetite than pellet 1. These results indicate that the main bonding phase in pellet 1 was magnetite, whereas the main binding phase in pellet 2 was hematite. The greater compressive strength of pellet 2 was strongly related to higher amount of hematite interconnections.
Three-Dimensional Reconstructed Image of Electrical Tree using Serial Sectioning Method and its Fractal Characteristics