Laboratory investigation on the liquefaction behaviour of a natural volcanic soil

Soil liquefaction is one of the main causes of severe damage during earthquakes. The liquefaction assessment of sandy deposits is typically conducted using semi-empirical methods developed for hard-grained sands, based on in-situ penetration test results. However, these methods are not reliable for all soil types. Volcanic grains are highly porous, lightweight, and have irregular surfaces, which often result in grain breakage and reduced asperities, leading to a different liquefaction behaviour compared to hard-grained sands. This study investigates the undrained cyclic behaviour and liquefaction resistance through laboratory tests of a volcanic silty sand containing non-plastic to low-medium plasticity fines, retrieved in proximity to the foothill of an embankment dam in Southern Italy. An extensive experimental program was carried out, consisting of drained and undrained monotonic tests, as well as undrained cyclic triaxial and cyclic simple shear tests on both undisturbed and reconstituted specimens. The monotonic behaviour was studied to primarily determine the steady-state conditions and evaluate the liquefaction potential of the tested volcanic soil using the state parameter, within the framework of steady state theory. Cyclic tests were conducted to derive the cyclic resistance curves and to study the cyclic mechanical response of these soils through stress-strain hysteresis loops. Special attention was given to the generation and accumulation mechanisms of excess pore water pressure, as well as the development of shear strains. The results show that volcanic soils exhibit higher liquefaction resistance than hard-grained sands, emphasising the importance of the soil fabric. Most specimens exhibited cyclic mobility, due to the dilative behaviour of the soil, confirming their low liquefaction potential. Furthermore, one of the main purposes was to investigate the influence of the various state and compositional factors on volcanic soil behaviour. A substantial portion of the experimental program focused on how these factors affected the soil response in both cyclic triaxial and simple shear tests. Additionally, the behaviour of the soils under investigation was compared with that of other volcanic soils and hard-grained sands, confirming the peculiarity in their mechanical response. Finally, the energy-based approach was also used to investigate the liquefiable behaviour through the concept of the cumulative dissipated energy. Cyclic test results were used to determine the cumulative dissipated energy up to liquefaction, defining the soil capacity. The trends of excess pore water pressure and dissipated energy were examined with respect to various influencing factors. The results confirmed that a higher amount of energy compared to hard-grained sands was required to reach liquefaction, confirming that volcanic soils undergo significant changes in the soil skeleton under cyclic loading. The outcomes of this study demonstrated that volcanic soils must be examined with caution, as the procedures developed for typical sandy soils are not directly applicable, thus contributing to expand knowledge in this still limited field.