Properties of force networks in jammed granular media

Abstract

Discrete element method simulations are conducted to investigate the effects of applied uniaxial compressive load, and bidisperse particle size distributions on force networks within jammed granular media. The differences between the strong and weak networks are examined through investigating the spatial correlation and distribution of contact angles, and emergence of chainlike structures. The simulation results show that the chainlike structures are more prevalent in the strong network due to the larger cumulative probabilities of contact angles, but not all the contacts belonging to the strong or weak networks are able to constitute the chainlike structures. Although the contacts of coarse-fine particles are dominant for the bidisperse systems, the contacts of coarse–coarse particles dominate the strong network, as well as the linear chainlike structures. Upon increasing the pressure from very low to high, the probability of contact orientations with respect to the compression direction in the strong network increases for contact orientation less than \(60^{\circ }\) and decreases for contact orientation greater than \(60^{\circ }\), while the opposite trends are observed in the weak network. The tails of normalized normal contact forces distributions are quantified by \(\hbox {P}(\hbox {f}) = \hbox {exp}(-\hbox {cf}^{\mathrm{n}})\), and it is found that the value of n depends on the applied pressure and particle size distribution. Statistical analysis shows that the degree of homogeneity of contact force increases with increasing pressure, which is also validated by participation number.