The structural integrity of embankment dams depends on the ability of filters to mitigate the risk of internal erosion failure. Granular filters are designed to trap finer base particles within the pore space of coarser filter particles. Using the component membership from the base particle contact network, a meso-scale investigation of the number, size, and spatial distribution of components can yield insight into filtration heterogeneities. The current work investigates the base particle components resulting from mono-disperse gravity-led DEM filtration simulations retaining 75% of base particles on average. The filtration samples investigated showed the formation of a giant component at the top of the sample, a few medium size components in top 40% of the filter and smaller components throughout the filter depth. These meso-scale insights provide a starting point for further investigations into the filtration heterogeneities and how these can be used for optimised filter design.

Internal erosion,Granular filter,Graph theory,Contact network