Professional Documents
Culture Documents
513
the particular (tropical) conditions of the site is presented and analyzed together with 3D numerical
analyses using LCPC Cesar software, permitting overall conclusions to be drawn that are of practical
and academic interest.
General Aspects of Drilled Displacement Piles
The execution, or the boring process, of this type of foundation can be summarized as follows: the
head is driven by rotation; the same machine used in the continuous flight auger type piles (CFA) may
be employed. During the descent of the drilling auger, the surrounding soil is pushed downwards and
to the side of the hole being excavated. Once drilling is completed and the stem rod has been removed,
the drilling auger units resume and a fluid concrete is injected down under pressure. Generally the
concrete has a minimum cement consumption of 400 kg/m3 with a slump test at around 240 mm.
What limits the use of this boring system (auger and stem units) is the available torque from the
drilling machine, which must be greater than 150 kN.mand the length of the rod does not generally
exceed 30 m. Currently, its diameter can vary between 0.31 m and 0.66 m. During execution,
monitoring is performed to gather information on depth, torque, penetration rate and the molding
characteristics of the concrete (rate of flow, volume etc.).
The main difference between the drilled displacement (DD) type of pile and the continuous flight
auger (CFA) relates to the transfer of soil to the surface. In addition, the former pile type does not
remove the soil, which remains compressed around the shaft of the hole. Its configuration is associated
with a conical shape and variations on the step of the auger provide unique characteristics of
downward and lateral soil movement duringexecution. Van Impe (1988) clearly shows the differences
between the two processes during execution, stating that CFA piles allow for decompression of the
soil (Figure 1a), whereas drilled displacement piles (Figure 1b) clearly enhance the compressive
characteristics of the soil, confining the shaft of the pile in the post executionphase.
The advantages of using the drilled displacement pile lie in the fast implementation process, together
with noise-free excavation. It also increases the load capacity when compared to standard CFA piles.
Daily production,over eight working hours, can range from 120 to 200 linear meters (Bustamante and
Gianeselli, 1998). Albuquerque et al. (2011) have demonstrated the high performance of drilled
displacement piles, in terms of capacity gain, when compared to other piles. This was done via load
tests on instrumented piles, with a new technique exclusively developed to insert the instrumented bar
insidethe drilled displacement auger.
514
This pile, nevertheless, has a higher consumption of concrete than standard CFAs and this can range
from 5% to 30% of overconsumption, depending on soil characteristics. A value of 15% is the most
representative (Albuquerque et al., 2011). Some further considerations regarding the execution of this
pile type can be cited, such as (Van Impe et al., 1998): (1) the shape of the drilling member leads to
beneficial penetration, although the increase in load capacity cannot be directly proven (further work
is needed); (2) the torque of the machine is of great importance for its implementation; (3) the
penetration rate depends on the diameter of the pile and on the soil type; (4) the penetration expends
more energy removing the soil than overcoming the friction between the auger and the soil; (5) there is
no problem when this is carried out in saturated, loose, granular deposits and finally, (6) no soil is
excavated during drilling, which is of great interest in brown field or contaminated sites.
Geological-Geotechnical Characteristics of the Site
The test site (experimental site) related to the present paper is located in the city of Campinas, in the
state of So Paulo, Brazil. The subsoil of the region is formed by migmatite and diabase, whichare
basic intrusive rocks that cover around 14% of the total Campinas region (with a total surface area of
98 km2). This site has a surface layer 6.5 m thick, composed of a highly porous,silty clay on top of a
layer of silt clay (residual soil) the thickness of which ranges from 10 m to 20 m. The water table
level is located at a depth of around 17 m. The average profile, together with SPT-T and CPT results,
are schematically depicted in Figure 2. These in-situ tests were performed near the piles.
SPT-T
SPT
Typical profile
CPT
T/Nspt 60
(N.m/blows/30 cm)
0 10 20 30 40 0 10 20 30 40 0
Nspt 60
(blows/30 cm)
Grain Size
Distribution
(%)
Rf (%)
qc (MPa)
fs (MPa)
25 50 75 100 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0.0
Red porous
silty clay
(Colluvium Soil)
0.2
0.4
0.6
Electrical
Mecanical
4
CL
LG'
Depth (m)
10
12
14
Clay Silt
(Residual
Diabasic)
ML
Silt
Clay
Concretion
Fine Sand
NG'
GWT
16
18
20
22
SIlty Fine
Sand
(Decomposed
Diabasic Rock)
(a)
(b)
(c)
(d)
Unified Classification System
(e)
(f)
(g)
Figure 2 Electrical and Mechanical SPT-T and CPT test results (Giacheti et al., 2004)
Results, Analyses and Discussion of the Experiments
Instrumentation and Load Tests
The testing piles had a nominal diameter of 0.37 m and were 12 m long. Concrete was used with
gravel in a cement consumption of around 400kg/m3 and slump of 240mm. The instrumentation
employed instrumented bars with strain-gages at the specific levels of 0.5m (N1 - reference), 5.0 m
(N2), 11.1 m (N3) and 11.7 m (N4) along the depth of the shaft.
515
Slow-maintained type load tests were carried out with successive load stages of no more than 20% of
the predicted work load for each of the tested piles. This procedure followed the prescripts of Brazilian
Standard NBR 12131. All tests were performed until the loads were approaching the limiting capacity
of the reaction system, allowing final displacements on the tested piles in the range of the capacity of
the dial gauges (close to 70mm). Maximum stabilized loads and displacements are displayed in Table
1.
Table 1 Maximum values of load and displacement for tested piles.
Pile
Load (max)
Displacement (max)
1
1545 kN
65 mm
2
1420 kN
62 mm
3
1320 kN
23 mm
Small pile displacements in the order of 5 mm were required to fully mobilize (to reach the
maximum value of) the lateral friction along the shaft. In Pile 3 it was found that a structural concrete
failure started at the top when the test displacement reached the maximum value of 23 mm, preventing
it from continuingwith the loading sequence.
In this same experimental site, previous load tests had been carried out on three mechanically bored
piles (standard procedure with bentonite mud stabilization for the open shaft). These piles were around
12 m long andhad a diameter of 0.4 m, reaching average maximum loads in the order of 682 kN. By
comparing this value with those in Table 1, one can clearly see that, as a first approximation, drilled
displacement piles do lead to load capacities well beyond those regularly found withtraditional boring
methods relating to standard bored piles in the region.
The instrumentation permitted the computation of the load distribution along the shaft and residual
values at the pile tip. Figure 3 displaysthe calculations performedusing the instrumented data in order
to obtain the experimental variation of the modulus of deformation (Es) with shaft strain.
40
30
20
10
0
120
Reference section
Level 5m
Level 11.1m
Level 11.7m
100
80
60
40
20
0
0,01
0.01
0,02
0.02
0,03
0.03
Strain (%)
0,04
0.04
0,05
0.05
120
Reference section
Level 5m
Level 11.1m
Level 11.7m
50
60
Reference section
Level 5m
Level 11.1m
Level 11.7m
100
80
60
40
20
0
0.01
0,01
0.02
0,02
0.03
0,03
Strain (%)
0.04
0,04
0.05
0,05
0.01
0,01
0.02
0,02
0.03
0,03
0.04
0,04
0.05
0,05
Strain (%)
516
was used with a finite element mesh composed of triangular elements with quadratic interpolation.
The elements were extruded at every meter in depth in order to render the 3D aspect of the problem.
Soil properties assigned to the different soil layers followed adopted failure criteria and are given by
the following parameters compiled in Table 2: specific weight (); cohesion (c); angle of friction ();
modulus of elasticity (E) and Poisson coefficient(). They were obtained through prior laboratory
tests with undisturbed samples retrieved fromthe same site, as presented by Albuquerque (1996).
Table 2 Soil parameters adopted in the numerical analyses.
Soil layer
Ei
0-6m
13
28
30
0.4
15
7 - 14 m
15
59
22
0.3
24
15 to22 m
16
70
20
0.3
28
- Specific weight (kN/m); c - cohesion(kPa); - Angle of friction (); - Poisson Coefficient;
- adhesion factor and Ei - Modulus of elasticity (MPa).
A parabolic model, valid for elements with a brittle-type behavior, was assigned to the piles structural
element, i.e., the concrete, thus with values related to their compressive resistance (Rc), tensile
strength (Rt), specific gravity and elastic parameters, such as the elasticity modulus and the Poisson's
coefficient.
The software CESAR LCPC v.5 from Itech-Soft Ltd. was employed in these analyses, which
considerably facilitatedthe use of the abovementioned model, and soil / pile characteristics and the
loading sequence. Although a contact (or joint) element between the materials with contrasting
properties, such as the piles concrete and the compacted soil surrounding its shaft, could be defined, it
was not employed here. This is due to the lack of experience, or typical properties, to be assigned to
this interface in a problem with a complex excavation process. More research should certainly be
devoted to this item in the future. Hence it should be pointed out that the numerical analyses to be
presented are considered to be an approximation of the real phenomena, given all the simplifications
considered with them.
Figure 4 presents the load vs. displacement curves obtained through the experimental load tests and
the numerical simulation (for a typical pile equivalent to those experimentally tested). Indeed, as
previously stated, the numerical results allowed an approximation of the phenomena, given the
essential simplifications built into the analyses. As one can clearly see, the numerical simulation was
lacking in terms of an adequate acquisition of an ultimate (bearing) capacity response equivalent to
what has been measured in the field. An average underestimationwas found, in the range of 20 %,
together with an overestimation with pile displacement under working conditions (considering a safety
factor (SF) of two, the numerical result was almost 4x the experimental predictions).
517
Load (kN)
0
200
400
600
800
0
Displacement (mm)
10
20
30
40
Pile 1
Pile 2
Pile 3
Num.
Qmax = 1545 kN
Qmax = 1420 kN
Qmax = 1320 kN
Qmax = 1177 kN
max = 65 mm
max = 62 mm
max = 23 mm
max = 62 mm
50
60
70
Pile 1
Pile 2
Pile 3
Numerical
Figure 4 Load vs. Displacement curves of the tested piles versusthe numerical prediction.
Nevertheless, in spite ofthe simplifications, it is possible to recognize that some qualitative pattern of
behavior can be educed (and is of interest) from the numerical results. For instance, Figure 5 shows
the load transfer mechanism that could be developed during the loading sequence of this particular
foundation.
518
predictions (Figure 5) and from experimental data available from a previous series of tests at the same
site (i.e. load at standard bored piles, as previously mentioned).
This comparison is presented in Figure 6, where it can be seen that:
Comparisons between numerical and experimental load transfer curves at both maximum and
working conditions are surprisingly good for the drilled displacement piles. Given the discrepancies
found in Figure 4, the present set of results are remarkably close, underlining the fact that the
numerical analyses could qualitatively comprehend some of the real effects of interaction between the
piles and the surrounding soil in spite of inherent simplifications;
In both cases, maximum and working conditions, less than 10% of the load applied at the top
of the pile reaches, and is supported by, its base. It is surprisingthento know that, for the tested
conditions, the drilled displacement pile has behaved essentially as a floating foundation;
The bored pile demonstrates contrasting behavior when compared to the drilled displacement
piles. Although similar in terms of load transfer mechanism, i.e. essentially floating, the bored pile
has achieved a maximum load of around half of the equivalentloads with the drilled displacement
piles. In other words, this traditional (and cheaper) excavation/molding procedure leads to a pile with
half the capacity andworking capabilities, when compared to the more sophisticated (and more
expensive) drilled displacement pile execution procedure.
Load (kN)
0
200
400
600
800
1.000
1.200
1.400
1.600
0
1
2
3
Depth (m)
4
5
6
Qmax (FEM)
Qmax (DD 1)
Qmax (DD 2)
Qmax (DD 3)
Qmax (bored)
1/2 Qmax (FEM)
1/2Qmax(DD 1)
1/2Qmax (DD 2)
1/2Qmax (DD 3)
1/2Qmax (bored)
7
8
9
10
11
12
13
Figure 6 Comparison between the numerical predictions and experimental transfer curves from all
the drilled displacement piles and a typical standard bored pile at this site.
Indeed, the differences between the drilled displacement pile and the traditional bored pile behavior
areastonishing. Using the data in Figure 6 and calculating the average (assuming a constant along the
depth) unit friction stress on the shaft, and the unit tip stress on the base, one can derive the values
displayed in Table 3. This comparison clearly demonstrates that the execution process of piles does
improve the soil characteristics and the overall pile behavior, during ultimate as well as during
working load conditions.
Table 3 Ultimate values of avg. unit lateral frictionand tip stress for all studied cases
Pile
Shaft Friction (kPa)
Tip (%)
Drilled displacement Pile
86
14
Bored
40
2
FEM
78
11
519
Table 4 shows the results of applying methods for estimating load capacity based on SPT and CPT
tests and load test results.
Table 4 Values estimated by the study methods and load test results.
Method
In-Situ Test
Qshaft(kN)
Qtip (kN)
Cabral et al. (2000)
SPT
324
280
1073
701
SPT
175
1011
CPT Begemann
Bustamante and Gianeselli
120
1011
CPT Delft
(1998)
221
1129
CPT - electrical
Pile 1
1315
230
Pile 2
1000
440
Load Test
Pile 3
1133
187
Qtotal (kN)
604
1774
1186
1131
1350
1545
1440
1320
Table 4 shows that the method used by Bustamante and Gianeselli (1998), based on electrical
CPT,gave the best predictions of load capacity. The method of Cabral et al. (2000) showed a much
lower value. In general, it can be observed that the calculations performed with the SPT tests did not
produce good results.
Once the load test on the drilled displacement piles was completed, a manual excavation was carried
out around the shaft of one of these piles in order to exhume and remove it from the ground. Duringthe
excavation process,it was possible to retrieve soil samples from the region of the stem / auger cuttings
and their surroundings. By performing tactile and visual analyses with these disturbed samples, it was
possible to note that the soil extracted 5 cm from the auger surface had a much more compact
appearance than those at 50 cm. Both were taken to a geotechnical laboratory to be characterized in
terms of standard soil tests.
The results of these tests are shown in Table 4where it is in fact demonstrated that the samples closer
to the auger suffered a process of compaction to a higher degree than those further away. It is
important to notice that the void ratio of the virgin soil before pile execution was 1.60. This aspect is
clearly noticeable through all depicted experimental state variables such as the natural weight, the void
ratio and the porosity. A slight decrease in moisture content is also noted, perhaps given the heat
exchange process (accentuated at distances closer to the auger) and localized evaporation phenomena.
In the literature we can find several papers whichdemonstrate soil improvement when performing this
type of foundation (Siegel et al., 2007a; 2007b; 2008; 2010).
Table 4 - Physical parameters of soil samples taken around the pileat a depth of 11 m.
Distance from Auger
w (%)
e
n (%)
nat (kN/m3)
Sample - 5 cm
17.53
31.1
1.23
55.1
Sample - 50 cm
16.10
33.5
1.48
59.7
nat = specific natural weight of the soil, w = moisture content, e = void ratio, n = porosity
Conclusions
Based on the general observations gathered throughout the results and the discussions presented
previously and the personal experience acquired with thecurrent exercise, it was possible to
outlinesome general guidelines or conclusions.
It should be pointed out, however, that all load tests were performed in a localized manner (spatial
position, soil type, foundation characteristics etc.), and the observations are actually valid for a
520
limitedset of data. Hence, although not definitive, the main conclusions of this paper can be seen
rather as tendencies that must be further confirmed or explored in future investigations.
Therefore, given the aforementioned aspects, one may conclude from this paper that:
The executive processintrinsically related to drilled displacement piles has indeed improved
the soil characteristics around the shaft of the pile and enhanced pile performance when compared to
other similar foundations, such as standard bored piles (or continuous flight auger piles, yet to be
confirmed). The drilled displacement pile execution has imprinted, in the surrounding soil,certain
beneficial characteristics, such as increasing its density and overall shear resistance. It has probably
also increased the level of horizontal locked-in stresses in the soil confined around the shaft of the
auger, although this aspect should be researched further.The gain in load capacity in comparison to
typical bored piles is astonishing (2x in the present series of analyses) and demonstrates that a higher
productivity in the field, in terms of capacity, and a lower number of piles in the design, can
eventually match, or even outweigh, the higher costs associated with the execution of this particular
type of foundation;
There is a considerable improvement in terms of residual loads absorbed in the pile tip in
comparison to common results observed for standard bored piles. Nevertheless, in both cases, it seems
that these foundations have predominantly behaved as floating piles rather than end-bearing piles.
More research effort must also be affordedin this context;
There is still some difficulty with the numerical simulations of drilled displacement type piles,
which can be associated with a series of factors, such as (a) the lack of adequate knowledge of the soil
parameters (and the behavior of the soil constituents) around the shaft and the auger of drilled
displacement piles; or (b) the lack of proper rheological models and numerical tools capable of
envisaging and takinginto account all the complex nuances of the boring process with the special
drilled displacement auger; or finally (c) the lack of adequateexperience and experimental comparative
data in Brazil, and elsewhere,with this type of foundation, that would allow researchers to forecast
with more certainty the closer to reality field phenomena associated with this type of pile.
Acknowledgements
The authors would like to acknowledge the grants, scholarships and research related funds provided by
the Government agencies CNPq and CAPES that made possible this and many other investigations at
graduate and undergraduate levels at both the University of Braslia and the University of Campinas.
They are also gratefulfor all travel-related funds that are provided by local institutions in Braslia and
So Paulo, that allowed the participation of the authors in eventssuch as the one in this paper. In this
regard, all the funding support provided by the geot. graduate programs of their universities, as well as
from the institutions Finatec and Fapesp, is much appreciated.
References
Albuquerque, P.J.R.; Massad, F.; Fonseca, A.V.; Carvalho, D.; Santos, J.; Esteves, E., 2011. Effects
of the construction method on pile performance: evaluation by instrumentation. Part 1: experimental
site at the State University of Campinas. Soils & Rocks, Vol. 34:1, pp. 35-50.
Albuquerque, P.J.R., 1996. Analysis of thebehavior ofsmalldiameterprecast pile instrumented in
diabaseresidual soil in theCampinas Region. MSc Thesis, University of Campinas, 170p. (in
Portuguese).
Brazilian Association For Technical Procedures (ABNT), 1992. NBR 12131: Piles: Static load tests.
Rio de Janeiro (in Portuguese).
521
522