Simulation studies of the packing of multi-sized particle systems

Abstract

Particulate materials are ubiquitous in nature and of fundamental importance to many industries. As the simplest status, the packing of particles are involved in many industrial processes, and its structure greatly affects the performance of these processes. However, the knowledge of packing structure is limited due to the difficulty in obtaining experiments, especially for multi-sized systems. This problem tends to be more challenging as many factors could affect the packing structures, such as the cohesive force and interstitial medium between particles. Considering most powder materials are distributed, there is a great need to acquire the knowledge with regard to packing structures of multi-sized particle mixtures. In this work, numerical models based on the Discrete Element Method (DEM) were developed to simulate the packing of multi-sized particle systems. In the simulation, the gravity force and interparticle forces were explicitly considered, and the dynamic processes of the packing were well reproduced. The systems studied include the packing of particles with binary, ternary and lognormal size distribution. The simulation results of binary and ternary systems were firstly compared with experimental measurements of parallel systems in terms of coordination numbers. Good agreement between the simulations and experiments has validated the current numerical model at particle-scale. The investigations of local packing structure were then fulfilled based on validated packings by means of radical tessellation. The effects of volume fractions on local structural properties were discussed and found the radical properties are strongly dependent on the volume fractions. For the packing of particles with a lognormal distribution, the effects of volume fractions are replaced with those of the standard deviation. By comprehensive consideration of the conclusions derived from different particle mixtures, the local structures defined by radical tessellation were found strongly depending on the solid density of polyhedra, indicating that there is a universal rule for the packing of multi-sized particle systems. Lastly, we examined the scope of application for some typical analytical models using simulation results and found that these models were built with certain criteria and assumptions, thus their capability are condition dependent, and improvements or new concepts are required for more practical applications.