4.3 TEST RESULTS ON END-BEARING SINGLE PILE
4.3.2 Stage 2: In-flight Pile Installation
The need for pile installation in-flight has long been recognized for an accurate simulation of prototype pile behavior in sand. For piles in clay, Craig (1984) postulated that the necessity for in-flight installation is less critical compared to that in sand.
However, there are evidence that piles installed at 1g in clay and subsequently tested at high g has a compression shaft capacity of about 15% less than that of the pile installed and tested in flight (Tomas, 1998). Lee (2001) also concluded after his centrifuge model tests that model piles should be installed in-flight for an accurate simulation of NSF on piles. In the present study, the pile was installed without stopping the centrifuge for a realistic replication of prototype stress levels during pile installation. Another advantage of in-flight pile installation is that the generation and dissipation of excess pore pressure
100
due to pile installation can be simulated which enable the study of NSF mobilized due to soil re-consolidation after pile installation to be studied effectively.
The model pile was jacked in using displacement control with a constant penetration rate of 2 mm/sec until the pile toe went through the soft clay and come in contact with the underlying acrylic plate (see Fig. 4.1). It should be noted that while the length scaling is straightforward with a scaling factor of 1/N from model to prototype (N is the g level of a centrifuge model test), the time scaling in the centrifuge modeling is a bit tricky. For a dynamic process, the time scaling from model to prototype is 1/N, while for a consolidation process, the time scaling from model to prototype is 1/N2. If one argues that, strictly speaking, certain consolidation occurs during the pile installation process, then the pile installation rate in prototype should be amplified by N times in the model simulation, which would turn out to be too fast to be properly controlled in the centrifuge simulation.
However, if one makes an approximate assumption that the pile installation process is a more-or-less undrained process, then the scaling of installation speed is 1/1 from model to prototype, and the installation rate of 2 mm/sec adopted in the present test represents an installation of 16 m of pile through top 2-m thick sand layer and 14-m underlying thick soft clay in about 2.2 hours in prototype, which is not that unrealistic.
The variation of measured forces at some selected levels along the pile shaft as the pile was jacked into the soil is shown in Fig. 4.6. In particular, gauge level-1 is at the lowest elevation and located at 0.88 m from the pile tip, while gauge level-9 is at the uppermost elevation and essentially flushes with the ground surface when the pile was installed to its final position (see Fig. 4.1(a)). Only readings of 3 levels of strain gauges are presented in Fig. 4.6 to avoid cluttering of curves. It can be seen that before pile
101
installation, tension forces developed along the pile shaft as the pile was at that moment fixed to the vertical actuator and suspended above the model ground. The tension load registered by each level of gauge was essentially the self-weight of the pile segment below that particular level. The tension recorded by the load cell before pile installation was the maximum as it registered the total self-weight of the whole pile as well as the pile cap. As the pile was pushed down and touched the top sand layer, resistance to the pile began to pick up quickly. The soil resistance first peaked at a penetration depth of about 1 m into the sand with subsequent very minor increment of resistance during the penetration of the lower 1 m sand layer. This should be due to the effect of the extension of the stress bulb beneath the pile tip into the underlying soft clay. Within the clay domain, the soil resistance increased gradually until the pile reached the underlying acrylic plate. It should be noted that although the vertical actuator was disengaged from the pile and rest half-way within the vertical slot of the coupling connector at the end of pile installation, the load on the pile head was non-zero in the present case. The coupling connector and pile cap act their self-weights to the pile head. The total self-weight of the above components is about 350 gram at 1g, which scaled up to be equivalent to about 1790 kN in prototype scale acting on the pile head, as registered by the topmost gauge level-9 in Fig. 4.6.
Fig. 4.7 shows the development of excess pore pressure as the pile penetrating through the model ground. The PPTs were embedded at various elevations within the soft clay but at a constant distance of 3 m from the center of the pile, as schematically shown in Fig.
4.1(a). It can be seen that in general, the deeper the PPT embedded in the ground, the later it started to feel the intrusion of the pile. For example, the topmost PPT5 which was located at 2.4 m below the clay surface, began to sense the excess pore pressure the
102
moment the pile toe touched the clay. Whereas the lowest PPT1 which was embedded 12 m below the clay surface, began to pick up excess pore pressure only after the pile penetrated about 4 m below the ground surface. It is evident from Fig. 4.7 that each PPT sensed the maximum excess pore pressure when the pile tip reached its respective elevation. Subsequently, after the pile tip passed that elevation, the excess pore pressure began to dissipate. These observations are very similar to those observed by some other researchers like Lee et al. (2004) in their centrifuge modeling of casing insertion for installation of sand compaction piles and Leung et al. (1991) in a field test of driven piles.
The maximum excess pore pressure registered by various PPTs embedded along the depth of the clay are plotted in Fig. 4.8 as solid circle symbols.
When a pile is driven into soft clay, the lateral displacement of clay will cause an increase in mean total stress in the clay around the inserted pile shaft, which in turn will generate excess pore pressures. On the other hand, pure shearing (without change in mean total stress) during the pile installation will also induce excess pore pressure in the clay.
Randolph et al. (1979) showed that by discarding the contribution of pure shearing, the excess pore pressure can be equated to total mean stress increment due to the pile installation. Assuming the soil as an elastic perfectly plastic material, a closed-form solution for the excess pore pressure can be expressed as (Randolph et al, 1979):
2 ln( ) r C R
u= u (for r0 ≤r≤R) (4.2)
where Cu is the undrained shear strength of the clay; r0 is the radius of the pile; R is the radius within the region where the soil becomes plastic. Beyond R, the soil remains elastic.
R can be expressed as
103
cu
r G
R= 0 (4.3)
The calculated maximum pore pressures at the location of the PPTs based on Eqs. (4.2) and (4.3) are shown as dash line in Fig. 4.8. It should be noted that for the kaolin clay used, Eu can be approximately evaluated by 150Cu according to Leung et al. (2006) or Ong et al.
(2006). Thus G/Cu = Eu/2/(1+υu)/Cu = 50. The measured undrained shear strength after soil self-weight consolidation as shown in Fig. 4.2(a) is also utilized for the calculation in Eq. (4.2). It appears that the theoretical predictions compare favorably with the measured values, although it tends to underestimate the measured data at the great depths. The neglect of the component contributed by pure shearing may be the main culprit for the under-estimation.
Besides outward displacement of the soil to accommodate the pile installation, Fig. 4.7 also shows certain ground heave during pile installation as detected by the potentiometer installed 12 m away from the center of the pile. No appreciable vertical movement was detected until the pile penetration exceeded 5 m, which is about 4 times the pile diameter.
Cooke et al. (1979) observed that that ground heave occurred earlier and was bigger closer to the pile. Further away from the pile, the ground heave occurred much later and the magnitude is much smaller. In the present study, a maximum ground heave of about 6 mm was registered by the potentiometer located 12 m from the center of the pile at the end of pile installation.