Statistical Physics Group

Hysteresis, manifest in loading-unloading cycles of three-dimensional polydisperse granular packs, is explored numerically from its microscopic underpinings. In the loading regime one can decompose the force network in a weak and strong subnetworks carrying a negative and positive anisotropy, respectively. On unloading the underlying force network undergoes a switching in anisotropy, where the anisotropies of the weak and strong subnetworks change sign. The network switching is then the underlying granular scale mechanism for hysteresis controlled by friction and surface elasticity. Two special behaviours of the unloading stage are analysed one corresponding to a local memory effect and a hydrostatic point.