Experimental study of factors affecting the viscosity-based pore pressure generation model and the pseudo plastic behaviour of liquefiable soils

Fully liquefied soils behave like viscous fluids, and models developed within the framework of soil mechanics fail to catch their behaviour on the verge of liquefaction or after it. Several research works have shown that modelling the liquefied soil as a fluid is physically more convincing. Such an equivalent fluid can be characterised via an apparent viscosity (η) (sharply dropping when liquefaction is triggered) which can be modelled as a power law function of the shear strain rate (pseudo-plastic fluid), depending on two parameters: the fluid consistency coefficient (k) and the liquidity index (n). With this approach, it is possible to consider a simple correlation between the equivalent viscosity and pore pressure increments independent on the equivalent number of cycles, whose parameters can be calibrated from the results of stress-controlled laboratory tests. The paper investigates the effect of some relevant experimental factors (effective vertical stress, stress path, frequency and waveform of the applied cyclic load, soil fabric and pre-existing shear stress) on the apparent viscosity of soils during their transition from the solid to the liquefied state, and therefore also on the pore pressure increments generated by the stress path. To do that, the results of stress-controlled laboratory tests performed in a sophisticated simple shear apparatus, along with published data, have been interpreted in terms of the apparent viscosity. Simple correlations in terms of viscosity-based pore pressure generation and pseudo-plastic behaviour are proposed and confirmed from the results of 1D non-linear site response analysis for the case study of Scortichino (Italy).