Introduction

Piles are true 3D elements and as such, it is not possible to model piles highly accurate in a 2D model. Moreover, it should be realised that typical engineering aspects of pile behaviour (such as load-displacement behaviour, ultimate bearing capacity, ratio of load transfer to shaft and pile tip) should, in general, be explicitly checked by the user before running the main analysis. As such the pile behaviour should be considered a USER INPUT instead of a MODEL OUTPUT or RESULT. When the resistance of the pile is not directly required as input parameter (e.g., when using a plate element or solid elements), it is, in general, recommended to compare the mobilised resistance with that estimated from other methods (e.g. analytical methods to estimate ultimate load capacity).

In this article, a number of reasons are indicated for the aforementioned behaviour.

See the related article on pile modelling in a 2D plane strain model for an overview of the different pile modelling options and their possibilities and limitations.

Installation effects

Currently, it is not possible to accurately capture installation effects in a finite element analysis.

A pile installation effect may, for example, be the lateral stress increase due to soil displacements during installation. The effects of installation may have a very significant influence on pile behaviour. However since at this moment we cannot model the actual installation process we have to choose to neglect these effects or to simulate them in a simplified way.

A possible method here may be to use a volume expansion in the pile cluster to simulate the soil displacement and increase of lateral stresses. Note that additional numerical tools, which may help in simulating pile installation effects, are currently subject of scientific research.

3D reality vs 2D model

Using a 2D model while reality is 3D causes some unavoidable changes/simplifications in geometry, also see Figure 1. As a result, the cross-sectional area, skin and pile tip area will change, directly causing changes in pile behaviour.

When modelling piles with volume or plate elements we will need to manually correct the stiffness properties of these elements and adjust the interface behaviour (i.e. R_inter) to find realistic load-displacement behaviour. Experience, however, shows that it will not be possible this way to accurately model all relevant pile engineering aspects. So you may, for example, have realistic axial load-displacement behaviour but not realistic lateral load-displacement behaviour.

Note that with the embedded pile row element these drawbacks have been overcome to a large extent. The embedded pile row element automatically determines the required stiffness properties to account for the discontinuity along the out-of-plane direction.

Figure 1. 3D Reality vs 2D model

Soil-structure interaction

When using volume elements or plate elements to simulate pile row behaviour an interface will be required to simulate soil-structure interaction. Realize however that the interface is also a continuous element in the out of plane direction (just like the volume elements and/or plate element), also see Figure 1.

Now, in reality, there is soil-structure interaction at the pile location and a soil-soil interaction in between the piles but in the 2D model, there is only soil-structure interaction. As a result of the latter, an unrealistic shear plane may be introduced in the model at the interface location due to (un)loading of the soil when using an R_inter < 1. This behaviour may especially become relevant when there is a significant amount of lateral loading acting on the pile row. Also see the related article on soil-structure interaction for a more detailed description of the standard interfaces.

Note that the mentioned behaviour is not relevant for the embedded pile row element. Due to the use of a special “out of plane” interface, this unrealistic shear plane cannot occur here.

Recommended approach when modelling piles in 2D

From the previous, it has become clear that modelling piles in a 2D model is not a straightforward process.

It is recommended in all situations (also when working with the embedded pile row) to set up a simple test model prior to the actual (more complicated) model and to test the (relevant) behaviour of the modelled piles. Where necessary the model should be adjusted to obtain the desired engineering properties which are known from pile load testing, codes of practice, experience, engineering judgment, etc.