981
Centrifuge modelling of bored piles in sands
Modélisation en centrifugeuse de pieux forés dans le sable
Williamson M.G.
University of Cambridge and Ove Arup & Partners
Elshafie M.Z.E.B., Mair R.J.
University of Cambridge
ABSTRACT: As part of a series of experiments to investigate the effects of tunnelling on bored piles carried out at the Cambridge
University Geotechnical Centrifuge, a new and novel model pile design is presented. The pile is semi-circular in cross section
allowing the sub surface displacements around the pile toe to be monitored using particle image velocimetry (PIV). The new pile
design, along with the loading mechanism, ensured the load was transmitted predominantly though the pile centroid, which reduced
the bending effects which have previously caused significant errors in these types of problem. The pile load cells at the head and the
base were also placed along the pile centroid to minimise the effects of bending on the load measurement. The paper presents the
results of a centrifuge pile loading test, illustrating the excellent response of the load cells within their working range and the high
quality PIV data which was obtained through this novel modelling approach.
RÉSUMÉ : Dans le cadre d’une série d’expériences réalisée dans centrifugeuse géotechnique de l’Université de Cambridge pour
examiner les effets tunnels sur les pieux forés, un nouveau modèle de conception de pieu est présenté. Le pieu est semi-circulaire en
coupe transversale permettant aux déplacements sous la surface autour de la pointe du pieu d’être surveillés par imagerie de la
vélocimétrie des particules (PIV). La nouvelle conception du pieu et le mécanisme de chargement assurent que la charge soit
transmise à dominante par le centroïde du pieu, ce qui a réduit les effets de flexion qui avaient précédemment causé des erreurs
significatives dans ce type de problème. Les cellules de charge fixées en tête et en pointe des pieux étaient aussi placées le long du
centroïde des pieux pour minimiser les effets de flexion sur la mesure de la charge. L’article présente les résultats d’un test de
chargement de pieux en centrifugeuse, montrant une excellente réponse des cellules de charge dans la zone de travail et la grande
qualité des données PIV obtenues par cette nouvelle approche de modélisation.
KEYWORDS: pile settlement centrifuge bored tunnelling base PIV modelling
1 INTRODUCTION
The ability to predict the effect of tunnelling on piles is
increasingly important no only from a safety, but also economic
point of view. To provide a better understanding of the
mechanisms influencing the effects of tunnelling on bored piles
in sands centrifuge modelling has been carried out at the
Cambridge University Geotechnical Centrifuge (Schofield
1980).
A significant amount of research work at Cambridge and
worldwide in recent years has been carried out using particle
image velocimetry (PIV) developed at Cambridge (White et al.
2003) as a tool to understand subsurface mechanisms for soil-
structure interaction problems. The ability of this method
allows both 2D plane strain and 3D plane of symmetry
problems to be modelled. The latter of these methods was used
in this research using plane strain (2D) tunnelling movements
and applying these to a non plane strain (3D) pile loading setup.
As part of this research work a novel pile and pile loading
system were designed at Cambridge University to model bored
pile behaviour in sands and it is the pile design aspect of the
research which is described in detail within this paper.
1.1
Background
Accurate modelling of plane of symmetry pile behaviour in the
centrifuge has been attempted by previous researchers (Marshall
2009 and Lu 2010) however a major drawback has been the
difficulty in providing an accurate measurement of the axial pile
load.
The cross section of a plane of symmetry pile is semi-
circular; this creates a variation in the flexural stiffness of the
pile in its two principal bending modes. This results in the piles
having a propensity to bend towards or away from the plane of
symmetry (its minor axis) rather than parallel to the plane of
symmetry (its major axis), see Figure 1.
This bending can lead to significant errors when strain
gauged load cells (Marshall 2009) are used to measure the axial
load of the piles such that their accuracy cannot be relied upon.
Loading these piles has also proven to be difficult owing to
their position within a centrifuge package (extremely close to
the plane of symmetry). This has previously prevented
researchers from loading these piles within their centroid
accurately and hence led to greater bending problems. Loading
along the centroid of the piles at the pile head and placement of
the strain gauged load cells along the pile centroid would allow
a significant proportion of this error to be mitigated through
bending compensation within the strain gauge bridges.
Placement of strain gauges on the centroid of a plane of
symmetry pile has been attempted previously (Lu 2010)
however as the strain gauged bridges were only ‘quarter’
Wheatstone Bridges these still suffered from a lack of true
bending and temperature compensation. However the results
were shown to be a significant improvement on the attempts to
position strain gauged load cells away from the pile centroid.
