P-Y Curves for Rigid Walls Retaining Granular Soil

Abstract

The mobilization of lateral stresses behind retaining walls constitutes a typical soil structure interaction problem particularly under cyclic loading conditions. There is interest in quantifying the relationship between the lateral earth pressure and the wall displacement. One of the methods used for this purpose is the p-y method, which is widely used in the analysis of piles. Soil-structure interaction methods that are based on the p-y curves method aim at replacing the soil behind the wall by a series of springs that mimic the soil behavior. The use of p-y curves in complex numerical analyses of buildings with underground stories is gaining interest in the structural engineering community. This poses a significant challenge in relation to the selection of realistic and simplified p-y models to be used as input in the structural numerical model. This study aims at investigating lateral earth pressure behind rigid walls in the context of p-y curves. The main target is to advance knowledge of the main mechanisms that govern the p-y response of basement walls supporting granular backfill. The objective is achieved using robust numerical and experimental tools. The numerical program entails developing active and passive p-y curves using a finite element model of a rigid wall retaining granular backfill. For the at-rest to active p-y response, the numerical results were used to derive a simplified hyperbolic p-y model for non-frictional walls and a bilinear model for the development of frictional shear stresses at the soil-wall interface. For the at-rest to passive p-y response, numerically-derived passive p-y curves were used to propose a truncated hyperbolic p-y model that is defined by a limit state passive pressure and an initial slope that is expressed using a depth-dependent soil stiffness model. Validation experiments indicated that the utilization of the proposed simple p-y models for both active and passive states provides realistic estimates of the p-y response of frictional walls supporting sands. P-y curves were also determined using an experimental program that involved building a large-scale rectangular soil-retaining-system in the laboratory. One side of the retaining system was specifically designed to act as a rigid wall that is hinged at its bottom and free to translate horizontally at its top. Friction forces at the sidewall were carefully addressed, measured, and minimized. Pressure sensors and LVDTs were mounted on the retaining system to measure wall displacements and soil pressures. Experimental p-y curves for static and cyclic loading conditions were determined and analyzed. The experimentally derived p-y curves were highly nonlinear and cannot be adequately represented by existing simple elastic-perfectly plastic models in the literature. More importantly, repeated cycles of loading resulted in a process of densification that affected the p-y response by significantly increasing the stiffness and maximum passive pressure after 10 loading cycles. These cyclic p-y curves are the first in the literature and will provide the basis for future studies that target the development of p-y curves for soil-structure interaction problem involving buildings with basements.