Analysis of piles and piled raft foundation under horizontal load

Abstract

It is well known that there are only few experimental evidences and analysis methods for piled raft foundations subjected to lateral loads. On the contrary, a lot of studies are focused on piled rafts under vertical loads and it was seen how more rational design methods can led to great benefits in terms of performance and in saving of construction materials. Piled raft foundations have a response to lateral loading that makes use of different mobilization mechanisms and experimental evidences show how the contribution offered by the raft can be significant and how in serviceability conditions can help to reduce the pile bending moments. The evaluation of horizontal displacements and load sharing requires numerical analyses and, even today, the most used analysis methods for pile foundations are Winkler-based approaches, while FEM (Finite Element Method) analyses, despite the big potential, have strong disadvantages and require high computational costs. In this work, some FEM analyses were carried out on a piled raft model in order to understand better the mechanism of the response under lateral loading. It was found that not always the piled raft performance is better compared to the raft alone. It is then presented the development and the validation of a ‘Hybrid BEM - p-y curves’ method for the analysis of single piles and of a BEM-based method (Boundary Element Method) for the analysis of pile groups and piled rafts under static horizontal loads. The latter can consider the contribution offered by the raft-soil contact, the interactions between all the foundation system elements using the Mindlin’s and Cerutti’s solutions and includes: a) the non-linear behaviour of the soil; b) the non-linear response of reinforced concrete pile sections, taking into account of the influence of tension stiffening; c) the influence of suction, using the so-called ‘Modified Kovacs Model’; d) the modelling of the shadowing effect, using a similar approach as presented in the so-called ‘Strain Wedge Model’. The proposed analysis methods result in saving computational costs compared to more sophisticated FEM codes and can provide reliable results using as input, data that come from a standard site investigation. The reliability of these simplified methods, for the analysis of single piles, pile groups and piled-rafts, was verified by comparing the results with data from well documented full-scale, 1-g and centrifuge tests.