A network-based investigation on the strong contact system of granular materials under isotropic and deviatoric stress states

The contact network is essential in linking the macroscale and microscale behaviors of granular materials. Particles and contacts are equivalent to nodes and links in the framework of the complex network graph, and then the network-based metrics can be applied to assess the network topology of granular materials. In this paper, several series of 3D DEM simulations are carried out, including isotropic compression and deviatoric loading tests for assemblies with different particle size distributions. The topological features, such as cluster size, connectivity, network distance and betweenness centrality, are captured for strong force subnetworks with different force thresholds in granular materials. It is found that the particle size distribution does not influence the general topological characteristics of granular materials, and the percolation is observed in the strong network with a proper force threshold for all granular samples. Chained features are consistently identified for strong networks extracted by around 1.5 times the average contact force at all stress states, and this force threshold also supports a giant cluster spanning the overall granular system. Assemblies under deviatoric states exhibit more identical features in strong network topology than that under isotropic states, giving a hint to understanding the structural uniqueness of granular mechanics.