A series of bender element tests and drained and undrained monotonic triaxial compression and extension tests were performed on air-pluviated samples of Hostun sand. Samples were prepared to different initial void ratios, consolidated under various isotropic and anisotropic stress states and sheared using different stress-paths and a wide range of deformations to characterise the sand's stress-strain response. The results suggest that the sand's small-strain behaviour essentially depends on the current void ratio and mean effective stress. Within the medium to large strain range, a state parameter approach in conjunction with the critical state framework can successfully predict the distinctive states of the sand's monotonic response, namely the phase transformation, the peak stress ratio and the critical states. Furthermore, the data is used to examine a stress-dilatancy relationship often incorporated in constitutive models. The characterisation presented herein aims at assisting the efficient calibration of numerical models and provides insight into this sand's behaviour, thus supporting the interpretation of results of physical modelling involving this sand. This paper highlights the importance of characterising sand's behaviour over the full strain range and shows that accurate predictions of the critical state and small-strain stiffness are crucial to assess other aspects of the sand's behaviour.

Uniform cyclic loading is commonly used in laboratory tests to evaluate soil resistance to earthquake-induced liquefaction, even if the cyclic stresses induced by earthquakes in the field are highly irregular. This paper discusses the use of stress and energy-based approaches to evaluate the liquefaction resistance of sand under irregular loading. Results of undrained cyclic triaxial tests including a large-amplitude singular peak loading cycle are presented and compared to those obtained using uniform loading. Although samples are subjected to loading patterns which would have been deemed equivalent by conventional stress-based methods, the number of cycles required to trigger liquefaction strongly depends on the amplitude and location of the peak within the loading history.Conversely, a unique relationship exists between the accumulation of dissipated energy per unit volume, computed using stress and strain measurements, and the observed residual pore water pressure build-up for all tests, throughout the entire cyclic loading application. This demonstrates that conventional laboratory tests using uniform loading conditions can be employed to determine liquefaction resistance if their interpretation is carried out based on energy principles.