Abstract
The undesirable breakage of agglomerates is a problem present in almost all operations involving the handling of these materials. The response of agglomerates to mechanical loading has traditionally been related to the total porosity of the structure. However, there is at present no universally accepted model which relates the internal structure of the agglomerates with their breakage propensity. This Thesis addresses the preparation and impact breakage of agglomerates of large primary particles with well-quantified interparticle adhesion and structure, using experimental and numerical simulation techniques, namely the Distinct Element Method (DEM). The agglomerates structure is manipulated in both cases by introducing a well-known number of large voids of controlled sizes. For the experimental part of the work, this required the development of a novel agglomeration technique, which allows good control of the interparticle bond properties and void size and number. The detailed observations of the impact process and of the impact product show two main patterns of breakage for the experimental agglomerates. The first one is the local disintegration of the impact site into small debris, producing a bimodal size distribution of the impact product, with fines made of small clusters and the residual fragment. The second pattern consists of the generation of large fracture planes, which split the agglomerates into relatively large fragments. For the numerical simulation, only the local damage pattern has been observed. The local structure of the impact site has been shown to have a significant effect on the pattern and the extent of breakage. Impact breakage is shown to increase with impact velocity and macro-void number and to decrease with interparticle bond strength. The macro-void size is shown to have a smaller influence on the agglomerate breakage. Finally, experimental and DEM agglomerates with matching physical properties have been prepared and their breakage behaviours have been compared. The results show satisfactory agreement when localised damage takes place, thus confirming the potential of the combination of the two techniques, i.e. experiments and numerical simulations. There is no agreement between the simulation and experiments for the case of fragmentation. This issue needs to be addressed in future work. This work is the first attempt to compare experiments and DEM simulations on a rigorous basis. It has addressed the case of brittle bond behaviour. Other modes of bond breakage need to be addressed in the future, in order to provide a comprehensive mapping of agglomerate failure.