Abstract
The formation and evolution of heaps of granular material has in recent years received ever-increasing research attention. As with other aspects of the granular physics field, much of this attention has focused on the use of numerical simulations, including the discrete element modelling technique. Before advancements in computing technology made numerical methods a viable option, assemblies of granular materials were typically rather poorly characterised by a limited range of bulk properties, such as the angle of repose of a heap. Such properties were found to be rather insensitive to the characteristics of the individual particles and as such were of limited practical use as the basis for process design. Consequently, problems in granular materials storage, handling and flow were typically tackled using semi-empirical approaches relying on long experience of similar process situations. The availability of computing resources has resulted in the spawning of a widely diverse range of numerical simulation approaches for the solution of bulk solids handling problems. However, this development in itself has given rise to further problems. The poor characterisation of assemblies by bulk properties has made experimental confirmation of numerical simulation techniques difficult, and this is perhaps partly responsible for the injudicious use of inappropriate numerical techniques within the research community. By a systematic study of the mechanics of granular heaps using the discrete element technique, this dissertation establishes that the choice of mathematical model and model parameters at the heart of any numerical method is of crucial importance for the realistic simulation of granular assemblies. The angle of repose is established as being rather insensitive to most single particle properties. The dissertation establishes the usefulness of the granular dynamics simulation method by demonstrating how internal 'microstructural' properties of granular heaps can be computed, and examines how simulation can complement relatively new non-invasive techniques for measuring such properties experimentally. Simulation and experiment are also used as the basis for a tentative mathematical model for the kinetics of segregation and stratification processes in poured heaps.