PARTICLES 2023

Structuring a micro-scale term for size-driven segregation for the particles upscaled to the same size parcels with DEM

  • Stasiak, Marta (Mines Saint-Étienne)
  • Martin, Sylvain (Mines Saint-Étienne)
  • Emam, Sacha (Itasca Consultants S. A. S)
  • Dedecker, Fabian (Itasca Consultants S. A. S)

Please login to view abstract download link

This scientific contribution is a numerical investigation build upon Discrete Element Method (DEM) with an addition of the particles upscaling technique also referred to as coarse-graining. In order to exploit the numerical efficiency as much as possible, the discrete model has been constructed with Same Size Parcels (SSP) approach. The numerical campaign has been conducted with Particle Flow Code (PFC) of our industrial partner ITASCA Consultants S.A.S. Both the reference bi-disperse mixture and its mono-disperse coarse-grained representation have been modelled under the dense flow conditions. Such an inertial regime has been obtained by simulating a gravity-induced motion of chute down the inclined plane. As a consequence, the upscaled assembly has been deprived form the granular segregation, more precisely, the size-driven aspect of the phenomenon causing the large grain to be squeezed toward less sheared zones of the assembly by the small particles falling into dilated free space. To compensate this loss, each parcel represents both phases through two local solid-volume fractions updated after every time increment. Adopting a well-established theoretical, macroscopic framework to a local context, a discrete mass flux rate has been constructed as its analogy yet in a discrete fashion. Respecting the mass conservation, those fractions are exchanged between the colliding parcels. Within this context, the main novelty of this work is to limit the consideration to a single collision, re-define the triggering variables locally and establish the scaling laws. Although, the applicability has been limited to the dense flow conditions, the method quantifies the size-driven segregation properly with a significant computational time gain being reported.