Advanced

FEM Modelling of Piled Raft Foundations in Two and Three Dimensions

Ryltenius, André (2011) In TVGT
Geotechnical Engineering
Abstract
In conventional design of pile foundations, all loads are taken by the piles, i.e. the contact pressure between the raft and the soil is neglected. In the last decades geotechnical engineers have started to take this pressure into account in design of pile foundation. Such a foundation,
where the raft and the piles interact to transfer the loads to the ground, is in this dissertation called piled raft foundation or piled raft.

Analysis of piled rafts requires numerical methods, due to complex soil-structure interaction. In this dissertation four different modelling approaches for analysis of piled raft foundation are compared; a full three dimensional finite element method model (FEM model) and also three plane strain FEM models (i.e.... (More)
In conventional design of pile foundations, all loads are taken by the piles, i.e. the contact pressure between the raft and the soil is neglected. In the last decades geotechnical engineers have started to take this pressure into account in design of pile foundation. Such a foundation,
where the raft and the piles interact to transfer the loads to the ground, is in this dissertation called piled raft foundation or piled raft.

Analysis of piled rafts requires numerical methods, due to complex soil-structure interaction. In this dissertation four different modelling approaches for analysis of piled raft foundation are compared; a full three dimensional finite element method model (FEM model) and also three plane strain FEM models (i.e. two dimensional models). All models are carried out by using programs developed by Plaxis, i.e. Plaxis 3DFoundation and Plaxis 2D, respectively. The plane strain models are similar but differ in the way of modelling the interaction between the piles and the soil. The first plane strain model is introduced in Chapter 3.4.2. Since this model produce questionable results, due to too weak modelling of the pile-soil interaction, two alternative models are introduced in Chapter 8.4 and 8.5 (called AM1 and AM2).

Piled raft foundations are three dimensional problems, in a two dimensional analysis one has to introduce simplifications and thereby inaccuracies. However, it could still be convenient to use this method since it is faster and the software is less expensive. The inaccuracies in a 2D model compared to a 3D model will vary depending on the characteristics of the problem. The object of the work is to study the inaccuracies and how these changes as the characteristics of the problem change.

The work starts with a study of previous master dissertations and other literature. To get a better understanding of Plaxis, a sheet pile wall is then modelled in Plaxis 2D. Subsequently, a hypothetical piled raft is analysed in Plaxis 2D and 3DFoundation to illustrate the different
modelling approaches. The piled raft is square with the piles distributed uniformly, and it is loaded vertically and uniformly. Finally, a parametric study of the different models is performed, where two parameters for the hypothetical piled raft are varied, i.e. pile spacing and the shape of the piled raft (rectangular, with different base to length relations).

When comparing the results, the 3D model is considered “true” and the maximum values are examined. In general for the piled rafts analysed, the two dimensional models (AM1 and AM2) overestimate the settlement (~30%), the raft bending moment (~30%) and the pile force (~10%). As the pile spacing decreases, the pile force from the 2D models resembles the
3D models more. While the settlement and the bending moment coincides less as the pile spacing decreases. As the length of the piled raft increases, the settlements and pile force from the 2D model converge towards the 3D results, which is expected since the problem get more two dimensional. However, a problem occurs as the raft gets non quadratic. The maximal differential settlement will then take place in the longer direction, and to calculate it, calculations have to be made for sections in the longer direction, were the plane strain condition is less satisfactory. This problem is off course less significant when the differential settlements are small as for a piled raft with a very stiff raft.

The inaccuracies occurring in the alternative plane strain models are more or less significant in all cases and the 2D models should be used carefully. Especially when calculating differential settlements or the bending moment in the longer direction. The plane strain models generate similar settlement and pile force as a 3D model when pile spacing is narrow and the piled raft shape is greater than 1:2, i.e. when the problem is more two dimensional. However, the piled rafts analysed are simple and as the problems get more complex the conclusions drawn here could be questionable, and the choice of model should lean towards a
3D model. (Less)
Please use this url to cite or link to this publication:
author
Ryltenius, André
supervisor
organization
year
type
H3 - Professional qualifications (4 Years - )
subject
publication/series
TVGT
report number
TVGT-5046
ISSN
0281-6679
language
English
id
5204468
date added to LUP
2015-04-20 14:36:10
date last changed
2015-04-20 14:36:10
@misc{5204468,
  abstract     = {In conventional design of pile foundations, all loads are taken by the piles, i.e. the contact pressure between the raft and the soil is neglected. In the last decades geotechnical engineers have started to take this pressure into account in design of pile foundation. Such a foundation,
where the raft and the piles interact to transfer the loads to the ground, is in this dissertation called piled raft foundation or piled raft.

Analysis of piled rafts requires numerical methods, due to complex soil-structure interaction. In this dissertation four different modelling approaches for analysis of piled raft foundation are compared; a full three dimensional finite element method model (FEM model) and also three plane strain FEM models (i.e. two dimensional models). All models are carried out by using programs developed by Plaxis, i.e. Plaxis 3DFoundation and Plaxis 2D, respectively. The plane strain models are similar but differ in the way of modelling the interaction between the piles and the soil. The first plane strain model is introduced in Chapter 3.4.2. Since this model produce questionable results, due to too weak modelling of the pile-soil interaction, two alternative models are introduced in Chapter 8.4 and 8.5 (called AM1 and AM2).

Piled raft foundations are three dimensional problems, in a two dimensional analysis one has to introduce simplifications and thereby inaccuracies. However, it could still be convenient to use this method since it is faster and the software is less expensive. The inaccuracies in a 2D model compared to a 3D model will vary depending on the characteristics of the problem. The object of the work is to study the inaccuracies and how these changes as the characteristics of the problem change.

The work starts with a study of previous master dissertations and other literature. To get a better understanding of Plaxis, a sheet pile wall is then modelled in Plaxis 2D. Subsequently, a hypothetical piled raft is analysed in Plaxis 2D and 3DFoundation to illustrate the different
modelling approaches. The piled raft is square with the piles distributed uniformly, and it is loaded vertically and uniformly. Finally, a parametric study of the different models is performed, where two parameters for the hypothetical piled raft are varied, i.e. pile spacing and the shape of the piled raft (rectangular, with different base to length relations).

When comparing the results, the 3D model is considered “true” and the maximum values are examined. In general for the piled rafts analysed, the two dimensional models (AM1 and AM2) overestimate the settlement (~30%), the raft bending moment (~30%) and the pile force (~10%). As the pile spacing decreases, the pile force from the 2D models resembles the
3D models more. While the settlement and the bending moment coincides less as the pile spacing decreases. As the length of the piled raft increases, the settlements and pile force from the 2D model converge towards the 3D results, which is expected since the problem get more two dimensional. However, a problem occurs as the raft gets non quadratic. The maximal differential settlement will then take place in the longer direction, and to calculate it, calculations have to be made for sections in the longer direction, were the plane strain condition is less satisfactory. This problem is off course less significant when the differential settlements are small as for a piled raft with a very stiff raft.

The inaccuracies occurring in the alternative plane strain models are more or less significant in all cases and the 2D models should be used carefully. Especially when calculating differential settlements or the bending moment in the longer direction. The plane strain models generate similar settlement and pile force as a 3D model when pile spacing is narrow and the piled raft shape is greater than 1:2, i.e. when the problem is more two dimensional. However, the piled rafts analysed are simple and as the problems get more complex the conclusions drawn here could be questionable, and the choice of model should lean towards a
3D model.},
  author       = {Ryltenius, André},
  issn         = {0281-6679},
  language     = {eng},
  note         = {Student Paper},
  series       = {TVGT},
  title        = {FEM Modelling of Piled Raft Foundations in Two and Three Dimensions},
  year         = {2011},
}