cfpbolivia

1. Design and Dynamic Testing of Precast
Piles on large Water Treatment Plant
Oscar A. VARDÉ a
; Rodolfo E. GUIDOBONOa
a
Vardé y Asociados S.A., Buenos Aires, Argentina
Abstract. In order to supply drinking water to the northern area of Buenos Aires,
AySA, the local water and sewer public company, awarded the contract to build
the North Potable Water Treatment System to the joint venture Aguas del Paraná.
The project comprises a fluvial water intake on the Paraná de las Palmas river, a
15 km long tunnel, a water treatment plant and 48 km distribution pipelines. The
main structures of the plant were founded on precast piles driven to dense silty
sands and sands using diesel and hydraulic hammers. A thorough analysis was
carried out in order to develop a systematic approach for assessing pile capacity at
given depths, acceptability of various pile hammer equipment and overall
acceptance criteria. This paper contains the description of the analysis carried out
to support this work, details of the structural design of the piles, the comparison
between the static and dynamic load tests, the pile driving acceptance criteria and
the evaluation of the setup based on CAPWAP analysis of the results obtained
from the pile driving analyzer (PDA) system.
Keywords. CAPWAP analysis, Driven piles, Pile driving analysis
1. Introduction
AySA, the local water and sewer public company, hired Aguas del Paraná (a joint
venture led by the brazilian contractor Odebrecht and the argentine companies José
Cartellone, Benito Roggio and Supercemento) to build the North Potable Water
Treatment System in Buenos Aires (Juan Manuel de Rosas Treatment Plant). A
consortium led by Halcrow (Halcrow-Latoniconsult UTE) undertook the design of the
project within a “design for design and build” scheme; Vardé & Asociados was the
geotechnical consultant to the Contractor.
The project includes a fluvial water intake on the Paraná de las Palmas River (1.5
million m3
/day), a 15 km long 3.6 m internal diameter tunnel to convey water from the
intake to the plant, a 900,000 m3
/day capacity water treatment plant and 48 km of
distribution pipelines. The aim of the project is to ensure drinking water supply for
almost three million residents of the Buenos Aires metropolitan area.

2. Figure 1. Location of the North Potable Water System Project.
The plant was built at a 15 ha site located in the vicinities of Maschwitz, 30 km
north from Buenos Aires city. The plant conforms to a classic treatment approach
consisting of three large units of flocculation, decantation and rapid gravity filtering,
two large pumping stations (for raw and treated water respectively), administration and
process buildings and electrical powerhouse (Figure 2).
Figure 2. Water treatment plant layout.
Flocculationunits
450m
Ancillary
facilities
Raw Water
Pumping
station
Treated Water
Pumping
station
Decantationunits
Filteringunits
430 m
490 m
210 m
Water intake
(Paraná de las
Palmas river) Temporary
water intake
(Luján river)
Water treatment
plant
Tunnel

3. 2. Geotechnical Conditions
The soil stratigraphy of the plant site comprises the following strata, based on 99
boreholes carried out by Vardé & Asociados: (a) a 7 m to 20 m thick upper layer of soft
sandy silts and clays; (b) stiff plastic clays; and (c) very dense sands from the
Puelchense formation, whose upper bedding surface was found at a depth between 20
m and 32 m, as shown in Figure 4. There is a layer of transitional materials – stiff
sandy clays and dense silty sands – between the stiff clays and the very dense sands.
The water table is 1 m deep across the area. Typical values of shear strength parameters
are included in the table 1.
Figure 3. Boring logs.
Table 1. Shear strength parameters
Stratum Depth (m) cu (kPa) u c’ (kPa) ’
Sandy silts 0 to 4/7 30 to 40 5 0 27
Soft clays 4/7 to 7/20 20 to 30 0 0 26
Stiff plastic clays 7/20 to 20/30 80 to 120 0 to 5 5 28
Stiff sandy clays
20/30 to 26/32
100 to 120 10 10 30
Dense silty sands - - 0 35
Puelchense sands below 26/32 - - 0 40

4. Figure 4. Level of upper bedding of puelchense sands.
The soil strata vary greatly in thickness and depth over the area of the plant. Given
this heterogeneity, 5 different geotechnical zones were defined (Figure 5).
1) Zone 1: there are no stiff clays present. Transitional dense silty sands
underlie the upper soft cohesive soils.
2) Zone 2: dense sands (Puelche formation) were found at 25 m deep, with a
layer of dense silty sands 3 to 4 m thick above them. The thickness of soft
silts and clays deposits are approximately 15 m.
3) Zone 3: stiff clays are encountered at depth of 7 meters followed by
transitional soils. The Puelche formation is at a depth of 25 to 30 m.
4) Zone 4: similar to Zone 3, but with a much wider range of depths where
the puelchense sands begin between 20 m to 30 m deep.
5) Zone 5: transition area between Zone 1 (no intermediate stiff clays
between the soft soil and the sands) and Zone 4.
Puelchense
sands
Depth
32 m
30 m
25 m
20 m

5. Figure 5. Geotechnical zones (A-A profile is shown in Figure 8).
3. Foundations
The foundations of the main structures of the plant consist of 4500 centrifuged annular
piles driven to the dense silty sands and very dense sands. The piles were built by
SCAC in Brazil and transported by ship to Zárate port. Pile driving was carried out by
Pilotes Trevi, San Pedro Fundaciones and Arpile. The main structural features of the
piles are included hereinbelow:
 External diameter = 500 mm; thickness = 110 mm; 12 m long segments.
 Concrete: H-38; ’bk = 38 MPa; R = 27 MPa; sulfate resistant cement (ARS)
 Reinforcement: ADN-500 steel bars, 35 mm concrete cover
Longitudinal: 11 bars 12.5 m diameter (1.16 % of reinforcement)
Transversal: 6 mm spirals; 125 mm center to center, and 62.5 mm center to
center at the top and bottom 1 m of the segments.
 Joints: steel rings, 250 MPa yield stress, with total penetration welding.
Diameter = 500 mm; height = 120 mm; thickness = 6.35 mm

6.  Allowable pile capacity = 1750 kN
 Allowable tensile load = 260 kN (for an extraordinary load state assuming the
processing units are empty and the external water level is 4 m above ground
level)
Delmag D36 hammers were selected for this project. These hammers have a
nominal driving energy or 52.5 kJ (at energy level 2) and 72.5 kJ (at energy level 3)
and efficiencies between 70 % and 90 %.
4. Load Tests
Static and dynamic load tests were carried out in order to define the pile capacity.
Given the scale of the work involved in terms of the number of piles to be driven, the
execution of a quantity of static loads enough to be representative to ensure the quality
of the process was prohibitive in terms of program. Therefore an approach had to be
developed, envisaged at testing an adequate number of piles within a timeframe
compatible with the overall construction program. This consisted in executing a
minimum number of static load tests coupled with a more intensive program of
dynamic tests which can be executed almost in parallel with the production program as
these testing do not involve any sacrificial piles. The success of this approach lied in
the calibration between both types of tests, as described in the following sections.
4.1. Static Load Tests
Two static load tests were carried out on piles driven near the borings H-6 and H-39
(Figure 6).
Figure 6. Location of static load tests.
H-6
H-39

7. The piles for the H-6 test were driven between March 30th
and April 4th
, 2009; the
test was performed on April 21st
and 22nd
, 2009 (set up time of 22 days).
The piles for the H-39 test were installed between April 6th
and 8th
, 2009 and the
test was performed on April 28th
and 29th
, 2009 (set up time of 22 days).
The ultimate capacity of the piles was estimated using Chin [1], Brinch Hansen
80 % [2] and Van der Veen [3] methods obtaining ultimate load values of 4550 kN,
4700 kN and 5000 kN.
The settlement for the service load was in both tests less than 5 mm.
Figure 7. Static Load Tests.
4.2. Dynamic Pile Tests
A total of 96 dynamic load tests was performed. A first battery of 6 tests was carried
out; on each of these piles at least two dynamic load tests were executed: (a) right after
finishing the driving and (b) 1 to 7 days afterwards. The tests on piles H-6 and H-39
were used to calibrate the results from dynamic tests against the static load tests.
A pile driving analyzer system was used in the dynamic tests to assess the pile
capacity and evaluate driving stresses and hammer energy and efficiency. The data
were obtained from two strain transducers and two accelerometers attached to the pile,
1.5 m below its top. Each couple of sensors was installed diametrically opposed in
order to compensate flexural effects.
The bearing capacity of the pile, the resistance distribution along its length and the
point resistance were evaluated by CAPWAP analysis of the data. This software
resolves the wave equation that governs the structurally dynamic behavior of piles by
representing them with a series of mass elements in series connected by weightless
springs. Table 2 summarizes the results obtained in a series of dynamic tests; for each
pile, the first line of values report the results for the test during driving, whilst the
second line of values are associated to the test carried out several days after driving.

8. Table 2. Dynamic load test results. L is the pile length; refusal (mm/10 blows); QT, QF and QP are the total,
friction and point capacities; setup time (days); c and t are the compression and tension stresses; E is the
hammer nominal energy; ET is the energy transferred to the pile.
Pile L
(m)
Refusal QT
(kN)
QF
(kN)
QP
(kN)
Setup
time
c
(MPa)
t
(MPa)
E
(kJ)
ET
(kJ)
ET/E
(%)
H-6 26.6 33 3186 1859 1327 - 29.5 0.0 52.5 47.7 91
4 4095 3730 365 5 26.7 3.2 52.5 33.6 69
H-9 22.4 29 3013 2413 600 - 28.4 2.3 52.5 46.7 89
5 4000 3400 600 3 26.2 0.2 52.5 35.9 68
H-25 23.2 17 3673 2699 974 - 27.5 5.0 52.5 43.9 84
2 4186 3874 312 7 28.1 6.5 52.5 38.1 73
H-29 34.6 23 2113 753 1360 - 22.8 3.6 52.5 37.8 72
15 3340 2432 908 1 31.6 1.3 52.5 44.9 86
1 4099 3282 817 7 26.8 2.6 52.5 45.0 86
H-39 29.3 30 2300 874 1426 - 26.3 3.6 52.5 46.1 88
15 3150 2633 517 1 28.5 4.9 52.5 47.9 91
H-43 27.0 15 2713 902 1811 - 26.8 2.4 72.5 57.7 80
5 3694 3105 589 6 25.3 3.9 52.5 37.7 72
A pile driven in cohesive soils gains capacity after driving has been completed;
this is the process normally referred to as “set up”. During driving large pore pressures
develop in saturated cohesive soils, hence reducing the soil shear strength. After the
driving takes place, a consolidation process begins and the shear strength and therefore
the pile friction increase over an unknown period of time.
The energy transferred to the pile in the driving is mostly used to mobilize point
resistance due to the reduced friction along the shaft. Conversely during the restrike the
shear strength has been increased and the percentage of energy used to mobilize the
friction is much larger than the fraction used to mobilize the point resistance. This
thixotropic behavior is a plausible explanation for the decrease of the value of point
load capacity in the redriving load tests.
As part of the overall testing program, it was needed a means of comparing the
results from the dynamic tests with those obtained from the static tests that had a
proven degree of confidence. The above physical explanation permitted to conclude
that the total load capacity (QT) could be approximated by the sum of the friction load
capacity reached in the redriving load test (QFmax) and the point load capacity obtained
in the initial load test (QPmax), as is presented in Table 3.
Table 3. Dynamic load test results.
Pile QFmax (kN) QPmax (kN) QT (kN)
H-6 3730 1327 5057
H-9 3400 600 4000
H-25 3874 974 4858
H-29 3282 1360 4642
H-39 2633 1426 4059
H-43 3105 1811 4916
The total load capacity obtained from these 6 test ranges from 4000 kN (H-39) to
5000 kN (H-6) and compares well with the extrapolated values estimated from the
static load tests. The friction capacity ranges from 65 % to 85 % of the total load
capacity of these piles.
Finally, dynamic testing presents one further complexity, as the results, and hence
the degree of approximation to the values, obtained through static tests, are largely

9. dependent on the energy applied to the piles and this in turn depends on the equipment
used and its level of efficiency. Table 4 presents the results of 25 dynamic load tests
on piles driven using Delmag D36 hammers; these tests were carried out at least 5 days
after the piles were driven, no dynamic tests were executed during the installation of
these piles. In all these cases the ultimate load was higher than 3500 kN, ensuring a
factor of safety of at least 2 with respect to the target service load of 1750 kN, as well
as confirming the suitability of this equipment for the production process. The
following section addresses the use of other pile driving equipment.
Table 4. Delmag D36 Hammers. Dynamic load test results. L is the pile length; refusal (mm/10 blows); QT,
QF and QP are the total, friction and point capacities; setup time (days); c and t are the compression and
tension stresses; E is the hammer nominal energy; ET is the energy transferred to the pile.
Pile L
(m)
Refusal QT
(kN)
QF
(kN)
QP
(kN)
Setup
time
c
(MPa)
t
(MPa)
E
(kJ)
ET
(kJ)
ET/E
(%)
P0330 24.3 9 4851 4685 166 15 33.3 0.0 72.5 55.8 77
P0518 18.2 20 3903 2301 1602 15 24.8 0.0 52.5 35.2 67
P0519 18.7 14 4330 2246 2084 15 23.8 0.6 52.5 31.9 61
P0247 25.3 12 3810 3271 539 17 27.2 3.4 52.5 36.0 69
P0274 23.5 2 3506 3163 343 27 23.7 3.3 52.5 35.3 67
P0639 30.4 26 4025 3283 742 13 25.6 1.6 52.5 44.8 85
P0640 30.1 7 3714 3167 547 13 25.2 2.1 52.5 39.3 75
P0655A 26.8 22 3509 2646 863 28 21.0 3.5 52.5 38.1 73
P0655B 27.6 27 4189 2601 1588 28 24.4 0.6 52.5 41.9 79
P0763A 25.5 19 3517 2284 1233 18 24.0 2.7 52.5 37.1 71
P0763B 26.7 22 3520 3080 440 18 26.2 3.5 52.5 42.4 81
P0714 27.0 2 4035 3356 676 11 38.9 0.4 52.5 43.1 82
P0715 20.7 11 4266 3005 1261 10 29.2 4.8 52.5 32.1 61
P0854 18.8 20 3650 2424 1266 34 24.5 0.4 52.5 34.1 65
P0857 18.6 25 3501 2636 867 43 25.6 0.0 52.5 32.0 61
P0862 20.3 19 3517 2190 1327 5 26.8 1.1 52.5 42.6 81
P0863A 19.8 20 3600 2544 1056 43 25.0 1.1 52.5 33.1 63
P0031 20.2 23 3960 3306 654 8 25.4 1.5 52.5 39.0 74
P0034 19.3 11 3540 3062 478 7 25.3 0.5 52.5 36.5 70
P0039 20.9 30 3882 2717 1165 5 26.6 1.3 52.5 36.1 69
P0142 18.8 20 3800 2135 1665 24 22.8 1.9 52.5 37.1 71
P0143 18.8 18 4266 3005 1261 25 28.6 1.7 52.5 45.2 86
P0192 22.9 20 4500 3803 697 21 24.8 3.2 52.5 37.8 72
P0206 18.2 14 3750 2178 1572 21 26.5 0.8 52.5 42.1 80
P0223 18.7 15 3700 1870 1830 19 26.3 0.8 52.5 43.4 83
5. Validation of other hammers
Due to a tight schedule set by the Client and the fact that there was a limited amount of
D36 hammers in the local construction market, it was imperative to use other brands
and even different kind of equipment, such as hydraulic hammers.
In order to check the suitability of the different hammers an approval procedure
was established which consisted in performing two dynamic load tests on a single pile:
(a) a first test after driving the pile with the hammer to be validated; and (b) a second
test executed at least 120 hours later, but applying the restrike with a D36 hammer.
Table 5 shows the results of the tests carried out to validated a Banut hydraulic hammer
(weight 62 kN, drop height 0.8 m, impact energy 49.6 kJ).

10. Table 5. Approval load test for a Banut hydraulic hammer. L is the pile length; R is the refusal (mm/10
blows); QT, QF and QP are the total, friction and point capacities; setup time (days); c and t are the
compression and tension stresses; E is the hammer nominal energy; ET is the energy transferred to the pile.
Pile L
(m)
R Hammer QT
(kN)
QF
(kN)
QP
(kN)
Setup
time
c
(MPa)
t
(MPa)
E
(kJ)
ET
(kJ)
ET/E
(%)
P0016A 19.3 16 Banut1 2975 1834 1141 - 21.3 3.5 49.6 38.1 81
D36 3750 2410 1340 7 20.7 3.6 52.5 27.9 53
P0117 19.9 19 Banut2 3900 2444 1456 - 22.7 7.9 49.6 46.8 94
D36 4471 2968 1503 5 29.9 1.5 72.5 52.5 72
Further results of dynamic load tests using different hammers are included in
Tables 6, 7 and 8 for the following equipment: Banut, Kobe K35 and Delmag D30
hammers respectively. Only one dynamic load test were carried out on these piles, at
least 3 days after they were driven.
Table 6. Banut Hammers. Dynamic load test results.
Pile L
(m)
Refusal QT
(kN)
QF
(kN)
QP
(kN)
Setup
time
c
(MPa)
t
(MPa)
E
(kJ)
ET
(kJ)
ET/E
(%)
P0202 19.9 13 3520 2885 635 15 19.3 5.6 48.0 30.0 63
P0311 29.0 13 3993 3726 267 8 22.1 1.6 48.0 32.6 68
P0851 18.4 12 3515 2362 1153 3 20.2 3.2 48.0 33.0 69
P0852 18.7 3 3530 2512 1018 3 189 2.9 48.0 29.6 62
P0855 18.4 20 3560 2951 608 14 20.7 2.1 48.0 34.5 72
Table 7. Kobe K35 Hammers. Dynamic load test results.
Pile L
(m)
Refusal QT
(kN)
QF
(kN)
QP
(kN)
Setup
time
c
(MPa)
t
(MPa)
E
(kJ)
ET
(kJ)
ET/E
(%)
P0528 19.0 27 3550 2461 1089 9 20.6 1.2 59.5 24.5 41
Table 8. Delmag D30 Hammers. Dynamic load test results.
Pile L
(m)
Refusal QT
(kN)
QF
(kN)
QP
(kN)
Setup
time
c
(MPa)
t
(MPa)
E
(kJ)
ET
(kJ)
ET/E
(%)
P0522 22.5 9 4791 3927 864 17 26.8 2.4 55.6 35.4 64
P0523 21.8 9 3600 2169 1431 17 20.6 2.2 55.6 24.4 44
P0029 32.5 15 3140 2218 922 4 20.8 7.7 30.0 24.3 81
Results above show that the use of this suite of hammers was validated as all of
them permitted to mobilized the strength required to guarantee the piles will reach the
service load required as shown by the calibration tests. In particular, the pile P0029
yielded a lower value but it was accepted as it was part of the foundation of a runway
beam and had a service load of 1500 kN, being the ultimate load 3140 kN and the
associated factor of safety 2.1.
6. Pile Driving Protocol
A pile driving protocol was defined for pile installation in order to assist the production
process to ensure the piles were driven up to the design standard; basically achieving
the strength required to deliver the projected service load. Three criteria were
developed:

11. 1) The pile was driven to refusal (50 mm/10 blows) and reached a specified
tip elevation (penetration in puelchense sands).
2) The pile was driven to refusal but tip elevation was above the puelchense
sands upper bedding. PDA results were extrapolated to determine the
relationship between depth and pile capacity (the red dotted line in Figure
8 indicates the elevation where piles reached a capacity of 3000 kN and
4000 kN). This analysis was used to infer whether the pile had reached
the minimum required capacity.
Figure 8. Static Pile Tests.
If the pile did not comply with neither of the first two criteria, then:
3) A detailed geotechnical analysis was carried out and, eventually an
additional dynamic load test was performed on the pile.
7. Bearing Capacity Parameters
The piles were driven into dense to very dense silty sands and sands, through a layer of
stiff plastic clays. The embedment into competent soils varied from 10 to 20 meters.
The skin friction builds up after driving as the dissipation of excess pore pressure
progresses, reaching final values larger than the undrained shear strength. This
phenomenon occurs because as the pile is driven, the displacement of the surrounding
soil leads to its compression and subsequent gain of strength.
A comparison of the clay cohesion (from undrained triaxial tests) with the average
skin friction developed along the piles as obtained from the load tests carried out
several days after the piles were driven, shows that the ultimate friction is mostly 20 %
to 60 % larger than the cohesion, with an average increase of 40 % (Figure 9).
Soft cohesive soils
Stiff clays
Dense silty sands
Very dense
puelchense sands

12. Figure 9. Skin friction from static Pile Tests vs. cohesion from undrained triaxial tests.
The point resistance results obtained from these tests are very scattered, with a
minimum of 0.9 MPa, a maximum of 10.6 MPa and an average value of 5.5 MPa.
The friction load carried by the pile along its length is 60 % to 80 % of the total
pile capacity. It is probable that in most cases, if not all, due to the set up process the
driving energy reached at the tip was not enough to mobilize the entire bearing point
capacity.
8. Conclusions
The work presented herein demonstrates the geotechnical challenges faced as part of
the design and build program of the North Potable Water System, requiring to deal with
precast pile driving production of un-precedent scale in Argentina.
In order to support a very tight construction schedule, which required 4500 piles
(total length of 120,000 m) to be driven between August, 2009 and February, 2010, an
adequate approach had to be developed to support a production program that addressed
several important aspects, such as: the need for a sound geotechnical campaign, the use
of precast piles that had to be driven at different depths of varying strengths and the
need to develop a testing protocol that could encompass the rhythm of construction and
the availability of equipment at the same time of ensuring the required quality was met.
This was specifically accomplished by using dynamic testing on a statistically
representative number of piles according to normal practice; this in turn lead to the
need to calibrate its results with those obtained from the more standard static tests, of
an already proven confidence. This process of calibration also helped to validate the
various driving equipment that had to be used.
Finally, the analysis of the results permitted to obtain further insight into the actual
values of pile skin friction and tip resistance to be used in design in comparison with
values normally inferred from the application of the standard available formulae.

13. References
[1] F.K Chin, Estimation of the ultimate load of piles from tests not carried to failure, Proceedings of the
Second Southeast Asian Conference (1970), 83-91
[2] J.Brinch Hansen, Discussion on hyperbolic stress-strain response – Cohesive Soils, ASCE, Journal of
Soil Mechanics and Foundation Engineering 89 (1963), 241-242
[3] T.L.M Van der Veen, The bearing capacity of a pile, Proceedings of the Third International Conference
on Soil Mechanics and Foundation Engineering, Zürich II (1953), 84-90