Scaling law for correcting the gravel content effect due to scalping techniques by DEM investigations

Abstract

Scalping techniques are commonly employed to handle oversized particles in gravelly soils for performing laboratory tests, which will cause significant gravel content (GC) effect on the predicted mechanical behaviors of prototype soils. To solve this obstacle, this study tries to establish the scaling law for correcting the GC effect, which controls the same equivalent skeleton void ratio between the prototype and model soils. The proposed scaling law characterizes the mechanical parameters using the newly introduced indexes including the degree of heterogeneity and the scaling factor for parameter A of the Harding equation. Then a series of DEM simulations of element tests were performed on sand-dominant gravelly soils to parameterize and check the applicability of the proposed scaling law. The simulation results show that both the converted shear responses and shear modulus reduction curves through the proposed scaling law are highly consistent with that of the prototype soils. This suggests that in dealing with deformation problems from small to medium strains, Rocha’s assumption is highly applicable for correcting the GC effect. When the adopted scalping techniques control the change in GC (ΔGC) below 20 %, the predicted liquefaction resistance will be very close to that of the prototype soils regardless of the earthquake magnitudes. However, when ΔGC is larger than 40 %, the prediction error in liquefaction resistance will be larger. Based on these simulation results, it is recommended that when selecting scalping techniques for laboratory tests on gravelly soils containing oversized particles, it is preferred that ΔGC be kept below 20 %.