Search

Site improvement by Deep Soil Mixing (DSM) to form in-ground shear walls has been used to remediate against the potential effects of earthquake-induced liquefaction. The grid pattern of soil-cement walls act as a confined shear-box which can provide additional shear stiffness and strength for sites to withstand strong ground motion. Current design practice for DSM grids commonly relies on the strain compatibility assumption, wherein the DSM walls and confined loose soil are assumed to experience the same shear strain. In this paper, the distribution of shear stresses and strains in liquefiable soil deposits treated with DSM grids are investigated using 3-D linear elastic finite element analyses of unit cells with the OpenSeesPL platform. From the analyses, it is found that the assumption of shear strain compatibility can be unconservative because portions of the DSM system can deform in flexure and the shear strains within the enclosed soil can exceed those within the DSM walls. The effects of input motion frequency content, wall stiffness, and area replacement ratio are evaluated. A revised design equation is proposed that accounts for the identified limitations in the strain compatibility assumption.

Inelastic behavior of large diameter extended pile shafts subjected to earthquake shaking and liquefaction-induced lateral spreading is investigated. A series of Nonlinear Dynamic Finite Element Analyses (NDA) are performed covering a range of soil, pile, and ground motion conditions. Results showed that combined effects of lateral spreading and inertia produce larger demands that cannot always be enveloped by designing the pile for each load case separately. The NDA results were used to develop an Equivalent Static Analysis (ESA) method.

Two case histories of liquefaction at down-hole array stations were analyzed using one-dimensional dynamic site response procedures with three different constitutive models for liquefiable soils. The two sites were the Port Island Array in the 1995 Kobe Earthquake and the Wildlife Liquefaction Array in the 1987 Superstition Hills Earthquake. The three constitutive models represent a range of modeling complexity and were run as user-defined models in the finite difference program FLAC. Procedures for calibrating the non-linear constitutive models against field data and the numerical solution procedures are described. Sensitivity of the computed responses to parameter uncertainties was evaluated. The capabilities and limitations of these methods for simulating the observed responses at these sites are discussed.

The potential of a novel, bio-mediated soil improvement to increase resistance to liquefaction triggering and to reduce the consequences of liquefaction if it occurs was evaluated. Microbial induced calcite precipitation (MICP) binds sand particles together through calcite crystal formation at particle-particle contacts. This results in an increase in the small-strain stiffness and strength of treated loose sand. Geotechnical centrifuge tests were used to evaluate the increased resistance of MICP treated sand relative to untreated loose sand when subjected to seismic shaking. Results of one model with a structure founded on sand treated to a moderate level of cementation and another model with the structure founded on loose untreated sand are compared. The centrifuge models were subjected to ground motions consisting of sine waves with increasing amplitudes. The accelerations, pore pressures, and shear wave velocities measured in the soil during shaking are presented. The resistance to liquefaction and deformation in the MICP treated model showed significant increases, as evidenced by substantial decreases in excess pore pressure ratios and vertical strains beneath the structure.