On the small strain stiffness of some Scandinavian clays and impact on deep excavation

Doctoral thesis, 2016

This thesis presents the results of a comprehensive study on the small strain stiffness of several Swedish clays and its tentative effect on the design of retaining structures. A combination of field measurement techniques, seismic dilatometer and surface seismics, are complemented with bender element testing of the retrieved samples. The overall trend in the data is that the values found for the small strain shear modulus are larger than currently recommended in the Swedish Transport Authority guidelines, TKGEO (2013) (based on empirical correlations). Furthermore, the results indicate that multichannel analysis of surface waves (MASW) works best for acquiring a 2D profile of the small strain shear modulus within the top 10 meters of the subsoil, inferred from shear wave velocities. The seismic dilatometer is more appropriate for larger depths providing higher accuracy at the expense of losing the 2D spatial information. Additionally, it was found that the best laboratory based procedure was to measure the shear wave velocity on carefully extracted and transported samples (minimal disturbance) which were stored < 2 days and brought back to in situ anisotropic stress level. The horizontal stress component was obtained from the in situ dilatometer measurements while the vertical stress component was obtained from direct measurement of density (sampling levels) and from an existing correlation based on shear wave velocity. For determination of effective in situ stresses pore pressure profiles based on the yearly average were used. Different sample quality assessment methods have been compared but it was found no one method could be used to identify the best quality samples. Using a multiple method approach this was possible. Stiffness degradation with strain is significant for all the clays tested. Many empirical relations exist however these tend to relate to shear strain amplitude, rather than shear strain (relevent in deep excavations), thus agreement is generally poor. Laboratory determined stiffness degradation appears particularly sensitive to the sample timelime (disturbance and reconsolidation procedure). Finally, the influence of the small strain shear moduli on the design of excavations with embedded retaining walls is elaborated by means of simulations of a theoretical 10 m excavation. Not surprisingly, the higher stiffness leads to large differences in structural response and more realistic deformations in the far field. It would seem prudent to insist that for critical deep excavation projects where Finite Element Analysis is performed the consistency of the model parameters are demonstrated with element test simulations on high quality test data. In this way the suitability of the chosen parameter set can be demonstrated.