Micromechanics of seismic wave propagation in granular materials

This research uses sophisticated laboratory testing and advanced numerical modelling to consider the dual problem of wave propagation in granular soils and estimation of stiffness. Understanding how stress waves propagate through soils is essential for earthquake engineering analysis and correct interpretation of current in-situ and laboratory seismic investigation results while accurate description of the soil stiffness is central for predictions of ground movements during and after construction. Soils are described as continua at scales which are large in comparison with the particle size, but this continuum response is a consequence of the particulate nature of the material. Discrete element modelling (DEM) describes the macroscopic response of particulate systems from integration of the mechanical interactions between particles. Computing power now permits the 3D analysis of assemblies of spherical particles with the actual dimensions of advanced laboratory tests. Sophisticated laboratory testing will be performed in a flexible boundary Cubical Cell permitting independent control of three principal stresses with fixed principal axes. Materials to be tested will include an angular sand, a rounded sand, and glass ballotini of two sizes: a size which matches the test sands and the larger size used in the DEM simulations. The influence of the nature of the granular material will be examined by increasing the complexity of the material in phases: starting with smooth spherical ballotini, then roughened ballotini, then real, irregular soil particles.Parallel DEM and multiaxial testing will be used to develop understanding of (1) the propagation through cubical samples of different granular materials of 'seismic' waves generated by piezoceramic elements; (2) the evolution of stiffness anisotropy with general histories of principal stresses; and (3) the effect of particle size and shape on the stiffness properties of granular materials.There will be three important outcomes: (1) testing and simulating the same system will provide important information on the reliability of DEM models to simulate wave propagation and on the level of detail in the DEM model required for accurate simulation; (2) testing of a range of different granular materials will show how results of DEM simulations of ideal particles can inform prediction of real soil response of real soils (under stress conditions which cannot be attained in the laboratory for example); and (3) knowledge of the nature of wave propagation and the evolution of small strain stiffness of geomaterials will inform the interpretation of laboratory and field tests.