Seismic loading on earth retaining structures

Chul Min Jung, Purdue University

Abstract

An analytical and numerical investigation is conducted for the seismic response of earth retaining structures. The investigation consists of a two-level approach. In the first level, a fundamental understanding of soil-wall interaction is gained by developing a complete analytical solution to obtain the seismic load behind a wall with an elastic backfill. In the second level, the seismic force on the wall is obtained from numerical simulations using a non-linear plastic soil model developed for this research. A complete analytical elastic solution is proposed for the seismic response of a wall that can bend, translate horizontally and rotate about its base. Comparisons between predictions from the analytical solution and those obtained from numerical analyses show that the analytical solution provides results that are similar to those from the numerical model. In addition a series of numerical analyses are performed to identify key parameters that affect the seismic response of a retaining wall. The identified parameters are: (1) seismic input motion; (2) flexural stiffness, rotation and translation of the wall; and (3) type of the interface between the wall and the backfill such as high or low frictional or fixed interfaces. A three-dimensional cyclic non-linear plastic soil model is developed to investigate the effect of elasto-plastic response of soils on the seismic earth pressures. The soil model includes the pressure-dependence of the initial stiffness of the soil and its evolution with strain, as well as a yield criterion that is represented by the Drucker-Prager formulation. Results from the soil model are compared with those from five laboratory tests performed on diverse soils under different loading conditions. In addition the model is evaluated by comparing its predictions from laboratory tests on small-scale wall tests conducted by Koseki et al. (1998) and Watanabe et al. (2003). The comparisons indicate that: (1) the results from the numerical model agree with the laboratory measurements; (2) multiple failure surfaces may form in the backfill at high seismic loads; (3) vertical failure surfaces are observed behind a cantilever wall, which may limit the volume of the soil driving the loading; (4) sliding of the wall results in a significant reduction of the seismic force; and (5) the Mononobe-Okabe method overestimates forces at high seismic loads. A parametric numerical analysis using the cyclic non-linear plastic soil model shows that the seismic forces on a retaining wall that is built on a stiff foundation depend, to a large extent, on the flexural relative flexibility of the wall to the backfill (dB) and on the type of interface between wall and foundation. If the foundation is soft, the seismic forces are much less affected by dB and type of interface, and are much smaller than those for a wall on a stiff soil foundation. A comparison between the elastic analytical solution and the non-linear plastic soil model suggest that: (1) the analytical solution may be used to estimate the seismic force acting on a wall built on a soft foundation (V sF≤300 m/s); (2) the analytical solution may underestimate the seismic force on a wall resting on a stiff foundation, if the wall does not slide; if the wall slides, then the analytical solution gives a reasonable estimate of the seismic force.

Degree

Ph.D.

Advisors

Bobet, Purdue University.

Subject Area

Civil engineering

COinS