ANCE &

UR Rapid Load Testing of

V ER
LITY ASS
SPECIAL

IFICATIO
ISSUE: Foundation Piles in Barcelona
UA

Q N

COVER STORY

Verifying the capacity of deep foundations is an important part of An essential part of this test is determining the correct location of
the construction phase of a project. Various pile load test methods the jack assembly such that the resistance above the assembly equals
can be used to verify capacity; two different test methods were used the resistance below it. Otherwise, this test may underestimate the
for such purposes at a project in Barcelona, Spain. axial resistance of the foundation element, as the capacity of the
The construction of an important project in Barcelona began in most resistant section may not be mobilized. Unfortunately, this is
2017. Given the size of the large diameter (1.5 m [5 ft]) cast in-situ exactly what happened with the first bi-directional test performed
piles, specified service loads (between 6 and 8 MN [1,350 and on the project. It was then decided to perform the other load tests
1,800 kip]) and results of the geotechnical site characterization, it using a different method with the understanding that the tests had to
was decided to perform full-scale load tests to ensure that the piles be performed in such a way to not delay the project.
provided the necessary axial resistance. Axial compression load
testing using the bi-directional method was selected for this project. Pile Load Test Methods
With this type of load test, an embedded jack assembly is attached Of the foundation load test methods available, top-down static load
to the steel reinforcement and positioned and cast within the shaft. tests are considered the most reliable. These tests apply a static load
Once the concrete has obtained sufficient strength, hydraulic to the pile, either through weights (i.e., dead load) placed on a
pressure is applied to the jack assembly, which then pushes the two reaction structure on the pile head or, more generally, by anchoring
parts of the shaft (above and below the jack assembly) in opposite the test beam to anchor piles adjacent to the pile being tested. A
directions. The pressure increases gradually until one of the hydraulic cylinder then applies the loading gradually to the pile
sections reaches geotechnical failure or until the assembly’s thrust head, while both the applied load and resulting movement of the
limit (i.e., stroke) is reached. At that time, the test ends and the axial pile head are monitored and recorded. Performing a top-down
capacity of the pile is determined to be at least twice the load that static load test is a long and cumbersome process, both for the
caused the failure. construction of the reaction system and for the test itself. This

Nicolás Moscoso, Allnamics España; Carlos Fernández Tadeo, CFT Asociados; Marcel Bielefeld, Allnamics; and
AUTHORS Gerald Verbeek, Allnamics USA

14 • DEEP FOUNDATIONS • MAY/JUNE 2019

means that top-down static load tests have an associated high cost In 1991, just a few years after it was developed, the first
and, in addition to mobilizing the large reaction elements, often Statnamic test was performed in Japan. Intrigued by the reliability
interferes with the site logistics. of this new testing method, the Tokyo Institute of Technology, led
In the 1960s, an alternative to top-down static load tests was by Prof. Osamu Kusakabe, grouped a research team to perform
developed, the high strain dynamic test (HSDT), more commonly comparisons of this new pile load test method with top-down static
referred to as a dynamic load test (DLT). The main characteristic of load tests. The results of that investigation clearly demonstrated
a DLT is the productivity with this method and the elimination of a that that results from both methods were equivalent; the results of
reaction system. To perform a DLT, a single blow on the pile head is the comparison were presented at the 1st and 2nd International
required with an impact ram that is equivalent to approximately Statnamic Seminars in 1995 and 1998, held in Vancouver and
2% of required mobilized static capacity. This impact generates Tokyo, respectively, which led to the incorporation and
pressure waves that are monitored by accelerometers and strain implementation of this new testing method in the Japanese
gauges mounted near the top of the pile, and the recorded data is standards in 2002 (Japanese Geotechnical Society, 2002), being the
then analyzed using signal matching techniques to estimate the first country in doing so. In the U.S., ASTM released a standard for
pile’s axial resistance. the Statnamic test method in 2008 (ASTM D7383, Standard Test
However, the advantages of cost savings and short test duration Methods for Axial Rapid Load (Compressive Force Pulse) Testing of
are offset by a reduction in the accuracy of the test results. In Deep Foundations), which was recently updated and released as a
addition, DLTs include viscous and inertial resistances that 2019 standard. In 2011, the Dutch introduced their own standard
dissipate energy; therefore, greater impact energy must be applied for this method (CUR-H410, Rapid Pile Load Test). In 2016, the
to mobilize the test load, which can damage a concrete pile. To European Committee for Standardization, CEN, approved the
avoid these drawbacks, it was decided to perform a rapid load test rapid loading tests as Standard EN-ISO-22470-10:2016
using the Allnamics StatRapid equipment. (Geotechnical investigation and testing – Testing of geotechnical
structures - Part 10: Testing of piles: rapid load testing), which was
applicable to the project in Spain.
Given the environmental and safety regulations in many
countries around the world, the fuel and transport thereof for a
Statnamic test can be problematic. As a result, alternate rapid load
testing devices were developed that do not involve combustion,
such as Jibanshikenjo’s Hybridnamic and Matsumoto’s Spring
Hammer Test, both developed in 2004, and, more recently,
Allnamics’ StatRapid in 2012. These devices, which follow
Procedure B in ASTM D7383, generate a load on the pile by
releasing a drop mass equivalent to 5% to 10% of the test load in
free fall. By varying the drop mass and drop height, a wide range of
test loads can be applied with a particular device. A specially
developed (relative soft) spring system is then placed on top of the
pile to extend the load duration. Finally, to prevent any rebound, a
catching mechanism is provided that is activated after the drop
mass impacts the pile head.
Specifics of dynamic and rapid load testing
Apart from extending the load duration, the specially-
developed spring system also greatly reduces the stresses in the pile
Rapid Load Testing
head. Therefore, it is a particularly convenient test method for cast
During a rapid load test, a load is applied to the foundation element
in-situ piles, which generally use lower strength concrete than
under quasi-static conditions. One of the first applications was the
prefabricated piles. For the project in Barcelona, it was determined
StatnamicTM device, an amalgamation of STATic and dynNAMIC,
that a dynamic load test producing the required test load would
which was developed in Canada and the Netherlands in 1985. The
have likely damaged the piles, which was one of the reasons why
Statnamic device consists of a reaction mass on top of a combustion
dynamic load testing was not selected.
chamber. Once solid fuel is ignited in the combustion chamber, the
During a rapid load test, the test load is applied fast and yet slow
pressure gradually increases, lifting the reaction mass while, during
enough to minimize dynamic effects (i.e., 100 to 200 msec for a
the process, generating an opposite downward reaction force onto
rapid load test compared to 5 to 15 msec for a dynamic load test).
the pile head. Once the fuel has burned, the lifted reaction mass
As the load duration increases, the dynamic effects are reduced
drops down within the container, where a bottom gravel layer acts
thereby avoiding the need to estimate complex dynamic soil
as a catching system.

DEEP FOUNDATIONS • MAY/JUNE 2019 • 15

parameters and, consequently, the test results resemble more and the velocity of the pile equals zero, at which point the pile is
closely those of a static load test. To ensure this is the case, both the assumed to behave quasi-statically. Therefore, from [Eq. 1], only the
ISO and ASTM standards specify a minimum load duration of 10 static and inertia terms remain (Fdynamic =0). Furthermore, given the
L/c where L is the pile length and c is the speed at which the stress equilibrium of forces, the measured force from the rapid load test
wave propagates through the pile. (FRLT) can be expressed as shown in [Eq. 2].
[Eq. 2] FRLT=Fsoil =Fstatic +Finertia =(k∙u)+(m∙a)
Rapid Load Test Analysis
Rearranging [Eq. 2], the static resistance (Fstatic ) can be computed
When a pile is subjected to a load test, the total pile resistance can be
using [Eq. 3].
split into three components: static, dynamic and inertial. As shown
in [Eq. 1], the static resistance is represented by the term k∙u where k [Eq. 3] Fstatic =FRLT-Finertia =FRLT -(m∙a)
is the soil stiffness and u is the pile displacement. The dynamic Measuring acceleration (using accelerometers), displacement (using
resistance is represented by C∙v where C is the soil damping constant an optical sensor) and FRLT (using load cells), the unloading point
and v is the pile velocity. The inertial resistance is represented by m∙a can be defined. Using the load and acceleration at the instant the
where m is the pile mass and a is the pile acceleration. unloading point occurs (tUPM), the static resistance Fstatic can be
[Eq. 1] Fsoil=Fstatic+Fdynamic+Finertia=(k∙u)+(C∙v)+(m∙a) determined. As an additional check, the measured displacement can
be compared with the computed displacement determined from
Given the extended load duration during rapid load testing, the pile double-integrating the acceleration signal.
can be assumed to behave as a rigid body, where the velocity and By comparing the results determined using the UPM with
acceleration are the same along the length of the entire pile. On that results from top-down static load testing, it became clear that the
basis, in the late 1980s, Middendorp developed the Unloading Point effect of the induced pore water pressures cause the mobilized
Method (UPM), which is the commonly used method to analyze resistance of a pile to be overestimated. Therefore, the value of
data obtained from a rapid load test. The UPM is straightforward Fstatic determined using [Eq. 3] must be corrected to take these
and, more importantly, is independent of the person performing the effects (i.e., loading rate effects) into account, as shown in [Eq. 4]
analysis, and centers around the point on the measured load- where η represents the loading rate effects based on soil type.
displacement curve where the displacement of the pile is maximal Typical values for η are 1 for rock, 0.94 for sand and 0.66 for clay.
[Eq. 4] Fstatic,corrected =η∙Fstatic

Test Device
The StatRapid test device used on the Barcelona project consisted of
a main frame, a modular hammer and a system of specially-
designed springs, and was capable of generating a test load up to 12
MN (2,700 kip). The entire system was transported to the site on
three trucks. Once the trucks arrived on site, the test device was
assembled in about three hours due to its modular design.

Measurements used to determine needed to determine static Offloading components of the testing device
resistance using the Unloading Point Method (UPM)

16 • DEEP FOUNDATIONS • MAY/JUNE 2019

 The footprint of the test device used for
this particular test was about 4 m by 4 m
(13 ft by 13 ft), small enough to cause
minimal, if any, interference with other
ongoing site activities. Only when the
device had to be moved using the crane was
site work interrupted for about 15 minutes,
which achieved the minimal interference
requirement mentioned previously and
was appreciated by the project developer
and the construction contractor.

Test Preparation and Execution

For this project, 4 different types of piles for
a total of 25 piles had to be tested. The
required test loads varied between 6 and 8
MN (1,350 and 1,800 kips), which was
easily accommodated by the selected
testing equipment. The pile lengths ranged
from about 28 to 35 m (92 to 115 ft);
therefore, a specific load duration was
required for each pile type since the
minimum required load duration was a

Moving the fully assembled and loaded test device

Once the device was ready, the test itself

took approximately 10 to 15 minutes per
pile to perform. At the end of the test, the
device could be moved to the next pile to be
tested in several ways: completely erected,
with either all or some of the drop mass still
in place, or with the device completely
disassembled to be reassembled at the next
test location. When moved completely
erected, three to six piles could be tested
per day, depending on the distance
between the test piles, configuration of the
test device and the capacity of the crane.
For this project, a 400-ton crane was
chosen to move the device fully assembled
and loaded, allowing the testing of up to six
piles within a single day and about an
average of just over three piles per day.
Load cells installed on the pile head
monitored the applied load, while the
movement of the pile head was monitored
by a Reyca optical system and accelero-
meters. As mentioned previously, the
acceleration measurements were also used
to analyze the quasi-static behavior of the
pile, as prescribed by the UPM and as required
by the applicable Eurocode to derive to the Performing a rapid load test using the StatRapid test device
load-displacement curve.

DEEP FOUNDATIONS • MAY/JUNE 2019 • 17

 Conclusion
Using the testing procedure described, 25 load tests of 10 MN each
were performed within 9 working days, with an average of 3 tested
piles per day. In general, the mobilized resistance was about 20%
higher on average than the required value, with the range of
mobilized resistance ranging between 10% and 50% greater than
the required resistance. Excellent coordination between the
general contractor and the testing contractor was key to the
successful completion of this testing work with minimal
interference to the ongoing site activities and without any impact to
the overall project schedule. Thus, after an inconclusive bi-
directional load test at the beginning of the testing campaign, rapid
load testing was successfully applied to confirm the axial resistance
of the 25 test piles, which, in doing so, achieved the objective of the
Cyclic load-displacement diagram from a rapid load test testing program.

Nicolás Moscoso is a geotechnical engineer with Allnamics, and

specializes in impact and vibratory pile driving predictions and analysis.
Carlos Fernández Tadeo is a private geotechnical consultant based in
Barcelona, Spain, who specializes in pile testing. Marcel Bielefeld is
commercial director at Allnamics, who has more than 30 years of
experience with various types of foundation testing. Gerald Verbeek
heads Allnamics USA and is actively involved in sharing European
experiences in foundation and soil testing in North America.

Corresponding static load-displacement diagram derived from

the rapid load test

function of pile length. To configure the device, including the drop

mass and, especially, the spring system, computer simulations were
performed in advance for each pile type to optimize the device
configuration and to ensure that the rapid load test requirements
would be met.
Rapid load tests using a cushioned drop mass are performed in
accordance with a standard test protocol, usually consisting of two
to five cycles with a gradually increasing test load, from which the
load-displacement diagram is derived. For this particular project,
five loading cycles were used for the initial tests to establish a good
understanding of the soil response. Given the small plot size and
limited site variability, subsequent tests could then be optimized to
limit the load cycles to just two: the first one to confirm the soil
stiffness and the second to mobilize the desired resistance. In those
cases where the soil stiffness differed from the anticipated value,
additional load cycles were performed to precisely define the load-
displacement curve for that particular pile.

18 • DEEP FOUNDATIONS • MAY/JUNE 2019