(250d) Pressure Distribution under Simulated 3d Granular Piles of Monosized Particles

Experiments in the mid seventies and early eighties observed, contrary to simple intuition, that the maximum normal stress beneath a granular pile is not located directly beneath the apex of the pile but instead there exists a dip in the normal stress distribution at this point. However other works have shown that this dip is not always observed.

Following these results, there has been an upsurge in the analysis of the stress distribution in granular piles, more specifically the pressure distribution on the base under granular piles. Most of this work has focused on predicting when this dip will occur. It is generally agreed that the dip occurs due to particles forming arches within the piles which shield the centre of the base from some of the weight of the pile, however what leads to this arches forming is still a hotly debated topic. Early theoretical work suggested that arching can only occur in a granular pile when a deformation of the supporting surface occurs, whereas a flat surface produces no arching. However recent experiments have shown that a dip can occur without any base deflection.

Because experiments on granular piles are inevitably created using multi sized particles, this has also been investigated as reasoning for the dip. Simulated piles, consisting of multi sized particles, did observe this dip but only when a relatively high degree of segregation is also observed in the granular piles. Piles that are relatively well mixed did not show the dip. Therefore it has been suggested that segregation is a factor that leads to the dip, but this reasoning has also had its critics.

The latest trend has been to look at the methods the piles are constructed. Granular piles are usually constructed from pouring the grains from a source above a surface. How the grains are introduced to the forming pile is often varied from pile to pile. Two construction methods that have been investigated are: the localized source procedure and the raining procedure. In the localized source procedure, the pile is formed using a source with an outlet that is much smaller than the final pile diameter, whereas in the raining procedure the outlet is significantly larger. Results showed that piles created using the localized source procedure did produce the dip; however the raining procedure was unable to reproduce the dip.

In this paper the results of ongoing work on the numerical simulations of the formation of granular piles by the Discrete Element Method (DEM) are reported. Three-dimensional granular piles (up to approx. 160,000 particles) are simulated using monosized spherical particles (radius 2.5mm and density 2,700 kg/m ), from which the pressure distribution on the base is measured. A variety of granular piles (both conical piles and wedge shaped piles) are simulated to test the various factors, noted in the literature, that are believed to lead to this dip in normal stress distribution. Our results show that for piles of monosized particles the dip in the pressure distribution is only observed for conical piles when the localized source procedure is used.