An experimental study on the kinetics of a free-standing retaining wall under seismic excitation

A multi-purpose centrifuge testing system for soil-rigid structure interaction was developed to study the seismic behavior of gravity retaining walls. A Centrifuge Modeled Rigid Structure (CMRS) models static and dynamic behavior of a rigid body. Based on the principles of the theory of rigid body kinematics, the structure's plane motion is characterized from direct measurements of CMRS's displacements and accelerations. Transducers were developed to measure the magnitude, angle, and position of the soilstructure interaction forces. The measurements provide a complete description of the kinetics in terms of phasing between the CMRS's motion and the interaction forces, as well redundancy needed for consistency checks. A series of tests were conducted on the Rensselaer Polytechnic Institute centrifuge where CMRS modeled a free standing retaining wall that is supported and backfilled with dry sand strata. The observations made during tests when the wall underwent excessive permanent deformations were used to verify the validity of modeling assumptions in the seismic design of gravity retaining systems both under failure and maximum displacements defined limit states.; The developed testing system proved to be an effective toot for experimental studies on kinetics of rigid structures interacting with the soil. The new measuring approach that includes consistency checks along with employed signal processing techniques ensured high confidence in the results. For the first time, experimental evidence was obtained of the kinetics of free standing walls failing by coupled sliding and rotation. The base shear resistance failures resulted in wall sliding with constant horizontal accelerations. The moment bearing failures generated excessive permanent rotations, characterized by threshold angular accelerations. The experimental evidence substantiated the validity of most of the modeling assumptions. Two important postulates were found incorrect: (1) the wall's inertia can be neglected (stability governed design), and (2) the wall yields outward only by sliding (displacements governed design).