Discussion Graphs for the Design of Laterally Loaded Piles in Clay by Alem and Benamar University of Le Havre, France Discussion by Mete Oner Oklahoma State University, OK, USA

Discussion

This paper uses what is now known as the conventional Soil-Structure Interaction, SSI, method. This method is based on the well-known beam-on-elastic-foundation analysis of classical mechanics, also known as the classical SSI.

Terzaghi has shown, in his landmark 1955 Géotechnique paper, how realistic soil parameters might be chosen for use in such an analysis. In this paper he has demonstrated the application of the method to the analysis of a number of geotechnical problems, including grade beams, sheet pile walls, and laterally loaded piles.

After Terzaghi's paper the method has enjoyed popularity for the analysis of such problems, as the most sophisticated geotechnical analysis tool in 1960's and 1970's, and at several places well into the 1980's and 1990's. The University of Texas efforts have given a longer life to the conventional SSI method: they have improved the SSI method by modeling soil nonlinearity into the system by introducing limits on soil contact pressures, and in some cases experimentally-determined nonlinear curves. This improved model, with nonlinear springs, is definitely superior to the classical SSI. This is the method used in Alem and Benamar (2002) ejge paper.

The finite element method (FEM), being hard to use, has failed to replace the conventional SSI method, in spite of the fact that it is not based on such shaky assumptions as SSI, and the soil is actually represented by the 2D or 3D elements depending on the model. This is especially true in the laterally-loaded pile problem due to the 3D nature of the geometry: In the laterally-loaded pile problem the lateral load is not axisymmetric, and thus stress and deformation fields are non-axisymmetric. Nonlinearity of the soil, of course, makes it even more non-axisymmetric. By contrast, the vertically loaded pile problem is axisymmetric, thus permitting a 2D analysis.

A well-known Foundation Engineering book also complicated the matters by calling the numerical solution of the SSI equations "the FEM" just because it used matrix formulation similar to the FEM. This writer strongly believes that one should not call any analysis model, where the soil is represented by springs and not by at least the same order of finite elements as the structure, "SSI," let alone "FEM." This confusion still persists today in some people's minds.

New methods are being developed to remedy the situation. The Shear Ring method was developed in 1990's for sheet pile walls (including floodwalls) by the writer. This method yields results essentially the same as a fully blown FEM analysis, though requiring a fraction of the effort. Another paper to appear in EJGE in this volume addresses the laterally-loaded pile problem by an improved SSI formulation, called Discrete Layer Analysis.

This paper is presented by EJGE as it represents a good reference point for the best conventional SSI has to offer.