Statistical and regression analyses of sands stiffness in triaxial tests and application of the results

The density and stress state significantly impact on the sand stiffness. Many of hardening soil models used for geotechnical computation are based on Duncan-Chang model and do not consider the influence of density on the soil stiffness. In course of triaxial compression of very dense or loose sands, the shear strains rise induces significant changes in density. In order to evaluate the effects of grain size distribution, density, and stress state on stiffness, the results of 962 isotropic triaxial tests on soil samples from 15 Moscow and Minsk construction sites were processed using statistical and regression analysis. As a result, empirical equations enabling evaluation of the effects of density and stress state on stiffness of sands with different particle size distribution were proposed. Comparative analysis of tests performed on alluvial and continental soils from Europe, India, and the United States sites showed that the sand stiffness is in the same range as sands from Moscow and Minsk sites. Proposed equations can be applied for preliminary estimation of the stiffness parameters for finite element method calculation and also can be used in geotechnical models that allow variability, horizontal and vertical distribution of stiffness to be taken into account. Additionally, the semi-empirical relationship based on the Duncan-Chang model, is proposed. The relationship provides more realistic results for loose and extra dense sands affected by large deformations and/or complex loading paths, when the changes in density influence soil stiffness. Generally, geotechnical engineers may utilize the obtained results to apply them to design of complex soil models.