A Novel experimental methodology for studying cyclic compressional soil response

Abstract

Vertical ground motions are comprised of both upward-propagating P-waves and inclined SV-waves. However, in common practice of site-response analysis- either via the VS30 term in a GMPE or a 1D site-response analysis- the P-wave

effects are typically not incorporated. To effectively represent such effects in site response analysis of vertical ground motions, a better understanding of soil response to cyclic compressional loading is required. In this paper, we study soil response to cyclic compression in the lab, in which pure compression is achieved by constraining radial deformations. Radial deformations can be constrained by loading within a rigid cell, such as an oedometer, creating 1D compressional boundary conditions. However, wall friction within the rigid cell can cause rotation of principal stresses, and hence ultimately lead to shear within the sample. Alternatively, in the present study, we suggest loading in a triaxial device under Ko-conditions, thus obtaining pure compressional loading. In the developed methodology, radial deformation is monitored during loading and constrained via adjustments of the cell pressure. Such an approach requires extremely slow loading and close monitoring of various parameters. We compare the results obtained with the suggested approach to cyclic loading within a rigid cell on a uniaxial compression device. Stress-strain loops are found to be comparable but not identical.