DYNAMIC PROPERTIES OF SAND UNDER TRUE TRIAXIAL STRESS STATES FROM RESONANT/COLUMN TORSIONAL SHEAR TEST

The purpose of this research is to study the effect of three-dimensional stress states on dynamic properties of dry sand. A uniform washed mortar sand was selected for study. The dynamic properties of concern were the shear modulus and material damping ratio at shearing strains between about 0.0001 and 0.1%. To perform this study, an operational resonant column/torsional shear (RCTS) apparatus, with which biaxial loading $(sigmasb{1}spprime > sigmasb{2}spprime$ = $sigmasb{3}spprime)$ could be applied, was modified so that hollow cylindrical specimens could be tested with internal cell pressures set independently from external cell pressures. With the modified apparatus, sand specimens were then subjected to various true triaxial states of stress ($sigmasb{1}spprime > sigmasb{2}spprime > sigmasb{3}spprime$) prior to dynamic testing. In addition, computer-aided excitation, monitoring, and data analysis systems were developed to facilitate experimental testing.;The dynamic shear modulus of this sand under isotropic loading agreed closely with the results of previous researchers. However, under biaxial and true triaxial loadings, the low-amplitude shear modulus was found to depend about equally upon the principal stresses in the direction of wave propagation and particle motion and to be relatively independent of the stress in the out-of-plane direction. A modified Hardin (1978) equation was developed to predict the shear modulus under anisotropic loads. Furthermore, the modified equation was also used to predict high-amplitude shear moduli under anisotropic loads. The study shows that by combining this equation with the Ramberg-Osgood equation, high-amplitude shear moduli under anisotropic loads can be well predicted.;Low-amplitude material damping under isotropic loading decreased with increasing mean effective principal stress in a fashion similar to that predicted by previous researchers. However, low-amplitude damping ratio under anisotropic loads depends on both the applied stress state and stress ratio. Material damping at high amplitude strains depends not only on stress state and shearing strain but also on the measurement method used. This study shows that high-amplitude damping ratios determined from resonant column tests exhibit the lowest values, while damping ratios obtained from the half-power bandwidth method exhibit the highest values. In addition, high-amplitude damping ratios predicted from shear moduli measured by the resonant column test combined a Ramberg-Osgood model and Masing criteria agreed well with those determined from torsional shear tests.