Analysis of a Yielding Pile-Column in Nonlinear Soil
Pile Response to Dynamic Loading
First, a preliminary investigation of the influence of pile nonlinearity on soil-pile-structure response has been conducted through parametric non-linear analyses. Eight different structure configurations founded on four different soil materials were excited by seismic strong motions. Results indicate that the increased soil deformability and the potential formation of plastic hinge on the pile may reduce the ductility demand and consequently lead to more economic design of the structure.
Then, a case history of kinematic pile damage is used as a yardstick to validate the phenomenological BWGG model (Gerolymos and Gazetas, 2005). The case history refers to the Konan Junior High School in Atsumatown, southeast of Sapporo, in the Japanese island of Hokkaido, which was severely damaged during the 2003 MJMA 8.0 Tokachi-Oki earthquake. A large number of shear cracks was observed on the wall of the building, indicating differential settlement and not cyclic shearing − the usual seismic deformation damage. Detailed surveys showed that this damage was caused by failure of foundation piles. Pile damage was unveiled after the piles were extracted from the soil, and can be summarized as follows : (a) all of the investigated pile foundations exhibited flexural cracks at the pile heads ; and (b) damage to piles occurred not only at their head but also at a depth of 20 m, where a large crack reached from the outer to the inner surface of the pile. Luckily, a downhole strong motion array had been installed just 50 m away from the building, making this an ideal case for the purposes of this work. The recorded strong motion at 153 m depth was used as seismic excitation. The problem was analysed through a nonlinear Winkler-type model, taking account of the inelastic response of both the soil and the pile. The nonlinear reaction of the soil was modelled realistically by BWGG interaction springs and dashpots, with due consideration to separation (gapping) of the pile from the soil, radiation damping, and loss of strength due to pore–water pressure development. It has been shown that the proposed phenomenological Winkler-based model is capable of reproducing the observed pile response and kinematic damage.
Pile Response to Kinematic Loading
A new hybrid method for analysis of slope stabilizing piles was developed, combining the accuracy of rigorous 3D finite element simulation with the simplicity of widely accepted analytical techniques. A comprehensive validation was presented against published experimental, field, and theoretical results from fully coupled 3D non‐linear FE analyses. It was shown that the proposed method provides a useful computationally efficient tool for analysis and design of slope stabilizing piles. The developed analysis method was then used in a parametric investigation to gain deeper insights on the factors influencing the response of piles and pile−groups. Axis‐to‐axis pile spacing, S, thickness of stable soil mass, Hu, depth, Le, of pile embedment, pile diameter, D, and pile group configuration were the addressed problems parameters. It was shown that S = 4D is the most cost effective pile spacing, since it is the largest spacing that can still generate soil arching between the piles. For relatively small pile embedment, pile response is dominated by rigid‐body rotation, without substantial flexural distortion : “short pile” mode of failure. In such a case, the structural capacity of the pile cannot be exploited, and such a design will not be economical. The critical embedment depth to achieve fixity conditions at the base of the pile is found to range from 0.7Hu to 1.5Hu depending on the relative strength of the unstable ground compared to that of the stable ground (i.e. the soil below the sliding plane).
A new methodology was developed for analysis of pile group distress under conditions of liquefaction-induced lateral spreading, behind a displacing quay-wall (a very critical problem indeed). To overcome the 3D nature of the problem, the analysis is conducted in two separate steps : (i) a 2D plane-strain dynamic effective‐stress analysis of a vertical section of the soil–pile system ignoring the presence of the piles, and (ii) a 2D plane-strain static analysis of a horizontal slice of soil–pile–quay‐wall system in the middle of the liquefiable layer. In the first case, the response of the system in terms soil‐movement behind the quay‐wall is obtained. In the latter step, the output of the first (soil displacement) is used as input to conduct an equivalent quasi-static analysis of the pilegroup. For this purposes, horizontal out of plane springs are attached to the piles, representing the stiffness of the whole pile, controlled largely by its boundary conditions at the top and reflecting the stiffness of the pile cap and the kinematic constraints of the superstructure. The proposed methodology was successfully validated against centrifuge model test results of experiments that were conducted at the Institute of Technology of Shimizu Corporation, in Japan.
Relevant Publications: C5 [Tasiopoulou et al., 2009].