Tips & Tricks

Free-Standing Pile Groups in PLAXIS

Creating a free standing pile group in PLAXIS (i.e. a pile group where the pile cap is not in contact with the ground surface like a pile raft foundation) requires a couple of tricks when using volume piles in PLAXIS.

In order to install a pile, PLAXIS first needs a volume of soil to convert to the properties of concrete during the pile installation phase. It can be difficult, however, to have these piles extend above the ground surface where there is no soil volume.

One trick to get around this is by creating an extra layer of soil above the ground surface (a 'dummy' layer). The user then converts the soil within the pile cross-section to the properties of concrete. The top dummy layer is then excavated so that the piles are left extending above the soil surface.

Thus the stages of analysis for a free-standing pile group are as follows:

1. Initial stress generation

2. Pile installation by changing the soil within the volume pile to the properties of concrete

3. Excavation of the dummy material around the piles to the required depth (i.e. to the original ground surface)

4. Pile cap installation by using 'floor' elements

5. Pile group loading by placing a compressive (or tensile) uniform distributed load on the top surface of the pile cap.

One final tip - if the adopted soil model is a nonlinear soil model (for e.g. the Hardening Soil model), the excavation of the dummy material phase may alter the state of the soil from the original properties (i.e. the OCR and the soil stiffness). In order to prevent this stage having any effect on the adopted soil properties, it is advisable that the unit weight of the dummy material is set to 0.0 kPa so that no changes in stress are generated.

Large Group Sizes in PLAXIS

The number of piles in a group possible using volume piles in PLAXIS is approximately 40 piles. For pile groups larger than 40 piles, however, it is possible to model these groups by using a couple of tricks.

First of all, taking advantage of the symmetry boundaries in PLAXIS can greatly reduce the number of piles you need to model and thus the number of elements in your finite element model (assuming the pile group is symmetrical).

The pile group can be modelled as half a group by using one line of symmetry or can be modelled using a quadrant of a group by using two lines of symmetry (i.e. placing your group in the corner of your computational domain). While it is possible to create a half-pile in the pile designer, it is not possible to create a quarter of a pile.

To allow more than ~40 clusters or piles to be modelled, this can be carried out by sectioning your group into a 'box' by creating rectangular 'rings' around your piles using geometry lines.

Pile Rafts with Pinned Connections

When using volume piles to model a pile raft in PLAXIS, the connection between the top of the volume piles and the raft (i.e. the floor in PLAXIS) is fixed. To the writer's knowledge, there is no way of modelling a pile raft with pinned conditions at the pile heads using volume piles.
When using embedded piles, however, the conditions of the top of each embedded pile can be selected as free, fixed or pinned. Thus to model a pinned-head pile raft, it is recommended to use embedded piles in PLAXIS.
Since embedded piles are an obvious simplification of pile behaviour - it is highly recommended that predictions determined using embedded piles are validated against measured data and/or volume pile behaviour. Reference can be made to Sheil and McCabe (2012)* for the validation of embedded piles in a free-standing pile group.

*Sheil, B.B. & McCabe, B.A. (2012) Predictions of friction pile group response using embedded piles in PLAXIS. The 3rd International Conference on New Developments in Soil Mechanics and Geotechnical Engineering, 28-30 June 2012, Near East University, Nicosia, North Cyprus

Validating PLAXIS Predictions

When using predictions determined by PLAXIS, it is first necessary to validate the adopted soil model and finite element model against measured field data to ensure that PLAXIS predicts true behaviour with sufficient accuracy.

An example of finite element predictions determined using the Hardening Soil model validated against measured data is documented by Sheil and McCabe (2012)**.

**Sheil, B.B. & McCabe, B.A. (2012) A 3-D finite element study of pile groups in soft clay. The 9th International Conference on Testing and Design Methods for Deep Foundations, 18-20 September 2012, Kanazawa, Japan.