Liquefaction resistance of loose Hostun sand considering partial drainage at increased overburden stress

抄録

Historically, liquefaction triggering methodologies have relied on the use of cyclic resistance curves to predict the initiation of liquefaction. These curves are informed by the undrained hypothesis which assumes that loading during an earthquake occurs too quickly for void redistribution to occur. However, recent research challenges the validity of this assumption, especially in layered soil deposits with varying permeability characteristics, where partially drained conditions are more appropriate. In parallel, the current body of earthquake case histories contains only data from shallow depths and relies on empirical correlations to extrapolate to greater depths. This is of concern for structures such as dams, levees, and embankments, which impart much greater effective overburden stresses than are accounted for by current simplified methodologies. Additionally, many such structures have historically been constructed with little consideration on the liquefaction potential of their foundation soil stratigraphy, where partially drained conditions can lead to localised volumetric strains. In this study, loose (D_R= 30%) Hostun sand specimens were subjected to both undrained and partially drained triaxial cyclic loading. Undrained experiments were used to create a baseline cyclic resistance curve. Partially drained tests at varying volumetric strain rates revealed the sensitivity of liquefaction resistance to volumetric strain. Volume contraction was associated with increased cyclic resistance, even at very small strain rates, while volume expansion resulted in decreased cyclic resistance. Effects were more pronounced under higher overburden stresses.