Collapse behavior of very loose dry sand

The objective of this thesis is to advance the understanding of the processes governing the collapse of very loose sands leading to flow liquefaction. Current methods of flow liquefaction analysis assume that once liquefaction is triggered the subsequent deformation is undrained. This is a gross over-simplification of the process since it does not take into account differing rates of drainage. For example with gravel the rate of dissipation of pore pressures could be commensurate with the rate of pore pressure generation. This would result only in volume changes without any loss of strength required for full scale failure, i.e., flow liquefaction.;Material properties and parameters for the state boundary surface have been obtained by testing very loose dry sand under a variety of triaxial stress paths. Necessary monitoring of volume changes was accomplished by direct measurement of axial and lateral deformations. For measurement of lateral deformations a new resistance wire transducer was developed and is described in this study.;An effective stress based theory is introduced to calculate the pore pressures developing during flow liquefaction is proposed in this thesis. The theory emanates from the state boundary surface concept that provides a constitutive relation between volume change and stress change during the collapse of very loose sand. It is shown that material instabilities exist on the state boundary surface that manifest themselves as increments of volumetric creep. In this study both the structural contraction along the state boundary surface and the structural collapse phenomena have been incorporated into a simple theory of pore pressure generation.