Professional Documents
Culture Documents
ICDF Vol 1 Conference Program - Complete Book PDF
ICDF Vol 1 Conference Program - Complete Book PDF
3 BOLIVIAN
rd
PROCEEDINGS
VOLUME 1
VOLUME 1
VOLUME 1
Invited Lectures
Edited by
Bengt H. Fellenius
K. Rainer Massarsch
Alessandro Mandolini
Mario Terceros Herrera
Design, execution, monitoring
and interpretation of deep foundation methods
ORGANIZING COMMITTEE
Conference Chairman:
Mario Terceros Herrera (Bolivia)
© 2017, 3° C.F.P.B.
Copyright Information
Manuscripts are published according to an exclusive publication
agreement between the author and the conference organizer. Authors
retain copyright to their works.
Volume 1
Preface ........................................................................................................................v
Invited Lectures..........................................................................................................1
Pile Integrity Testing: History, Present Situation and Future Agenda ........................................................ 17
Amir, J.M., Israel
Best Practice for Performing Static Loading Tests. Examples of Test Results with
Relevance to Design ................................................................................................................................... 63
Fellenius, B.H., Canada
Simple Approach to Static and Seismic Design of Piled Rafts ................................................................. 107
Mandolini, A, Di Laora, R. and Iodice, C., Italy
Effects of Installation Processes on the Axial Capacity of Pile Foundations in Sand .............................. 161
Prezzi, M., USA and Basu, P., India
Expander Body and Toe-Box: Expansion Devices for Deep Foundations Enhancement ........................ 209
Terceros A. M. and Terceros H. M.A., Bolivia
The 3rd International Conference on Deep Foundations is held April 27 – 29, 2017 in Santa Cruz
de la Sierra, Bolivia. It follows two successful conferences held in 2013 and 2015. The conference
is organized with the support of INCOTEC SA in association with the Society of Engineers of
Bolivia, the Bolivian Society of Soil Mechanics and Geotechnical Engineering and the Chamber
of Construction of Santa Cruz. It is held at the UPSA Campus (Universidad Privada de Santa
Cruz), the main private university of the city and arranged with the support of the International
Society of Soil Mechanics and Geotechnical Engineering (ISSMGE), Technical Committee 212,
“Deep Foundations”.
The principal objective of the conference is to bring together local engineers and international
experts in order to facilitate the exchange of experience and to introduce to the region new design
concepts, methods and equipment for the application to deep foundations. The conference program
is composed of invited lectures, discussions, a field demonstration, and a pile testing prediction
event where international experts have been invited to predict the load-movement response of piles
in static loading carried out prior to the conference.
During the first two days of the conference, speakers of international repute have been invited to
present papers on specific topics, covering different aspects of deep foundations. The third day of
the conference is devoted to the presentation and discussion of tests a comprehensive pile testing
program. The Bolivian Experimental Site for Testing Piles (B.E.S.T.) was adopted by ISSMGE
TC 212 as a reference site for investigations on piles and pile groups. B.E.S.T. offers a unique
possibility to enhance the understanding of the performance of different pile types and pile groups
when subjected to load. The geotechnical conditions at the B.E.S.T. site have been documented
by detailed investigations, using state-of-the art testing and interpretation methods. The results of
the field testing programme, including interpretation of in-situ methods and results of the pile
loading tests will be presented during the third day of the conference.
Volume 1 of the proceedings comprises the papers presented at the conference. All papers have
been reviewed by at least two members of the Review Committee. The dedicated work by the
reviewers and their valuable contributions is gratefully acknowledged.
Volume 2 contains a description of the geological setting and the results of comprehensive
geotechnical investigations carried out at the B.E.S.T. site. It is the intention of the Conference
Organizers to make available all data from the B.E.S.T. site investigations and pile tests in digital
format at the conference web platform for use in future investigations, in cooperation with
ISSMGE TC 212.
Volume 3 includes a description of the test piles and the loading test programme. The predictions
as well as a presentation of test results will be published in a Volume 3 after the conference.
Conference Program
April 27th 2017
8:00- Registration
8:45-9:45 SESSION 1
Paul Mayne (United States of America)
Recent developments and applications in geotechnical field Investigations - Session
Chairman: Jorge Alva (Perú)
15:00-16:00 SESSION 3.
Bengt Fellenius (Canada): Discussion on best practice for performing static loading tests.
Examples of test results and relevance to design.
Session Chairman: Luciano Decourt (Brazil)
16:30-17:30 SESSION 4
Joram Amir (Israel): Testing pile integrity - past present and future.
Session Chairman: Victor Hugo Alvarez (Bolivia)
8:30-9:30 SESSION 6.
Dan Brown (United States of America): State of art and state of practice in deep foundations.
Session Chairman: Oscar Vardé (Argentina).
10:30-11:30 SESSION 7.
Alessandro Mandolini (Italy): Design options for piled rafts - An overview.
Session Chairman: Juan Carlos Rojas V (Bolivia)
Mario Terceros Arce (Bolivia): Practical application of new expansion devices for pile
improvement in sandy soils.
Session Chairman: Walter Paniagua (Mexico)
14:30-15:30 SESSION 8.
Franz Werner Gerresen (Germany): Conventional methods and recent developments in retaining
walls.
Session Chairman: Carlos Medeiros (Brazil)
15:30-16:30 SESSION 9.
K. Rainer Massarsch (Sweden): Recent developments in vibratory driving and soil compaction.
Session Chairman: Bengt H. Fellenius (Canada).
9:00-10:15 Session 1 – Soil Investigation Results. - Memorial Session in Honour of Prof. Silvano Marchetti
SCPT: Peter Robertson (Canada)
SDMT: Diego Marchetti (Italy)
SPT-T and DPSH: Luciano Decourt (Brazil)
PMT: Roger Frank (France)
GEOPHYSICS: K. Rainer Massarsch (Sweden)
Session Chairman: Paul Mayne (United States of America)
Albuquerque, P.J.R.(1)
(1)
University of Campinas - Unicamp, Campinas, São Paulo, Brazil <pjra@fec.unicamp.br>
ABSTRACT. The purpose of this paper is to describe some instrumentation techniques for steel
piles and for cast-in-place piles by means of strain gages. A brief history of strain gages is
presented as well as the theoretical principles that the technique involves, showing the types of
connections used in pile instrumentation, indicating the advantages and disadvantages of the use
of each type. Determining the load distribution of a pile is the key objective of instrumentation.
Therefore, three ways are proposed to analyze the stress data collected. Also procedures for
instrumentation and installation of sensors in depth are presented. Instrumentation via strain gage
is a reliable technique commonly used in all areas of knowledge. However, it is necessary to adapt
the parameters one wishes to obtain to the technique to be used.
1. INTRODUCTION
The use of instrumentation to get parameters of the behavior of geotechnical structures is very
frequent nowadays in works such as: dams, slopes, foundations, containments, etc. For each type
of work and purpose of the data to be collected and analyzed, a different type of instrumentation
is used. According to Dunnicliff (1993), there are two general categories of measuring instruments.
The first category covers instruments used for in-situ determination of existing conditions of rock
and soil, such as resistance, compressibility, and permeability to be used in projects for which
purpose cone tests, vane tests, pressuremeter, etc, are used. The second category is used for
monitoring changing conditions during the construction phase and operation, such as water
pressure, total stresses, displacements, loads, and deformations.
Over the last decades, instrumentation equipment manufacturers have developed a variety of
materials and instruments for geotechnical monitoring. However, it is important to point out that
users must be aware of the desired parameters and of the techniques involved in each type of
instrument to get reliable and accurate information.
Several types of instruments may be used. In general, the most commonly used are based on
change of strain, employing vibrating wire gages, electrical resistance gages, and optical fiber
gages. Foundation engineering is constantly trying to learn about the behavior of deep foundations
in terms of load transfer. To this end, among the aforementioned techniques, strain gages are used.
This article is a study that aims to describe the technique of instrumentation of deep
foundations with the use of strain gages. The article will include both a theoretical approach and
practical experience.
Electrical extensometers, also called strain gages, are not recent tools. In the 1930's, Edward
Simmons (California Institute of Technology, Pasadena, CA, USA) and Arthur Ruge
(Massachusetts Institute of Technology Cambridge, MA, USA) working separately, were the first
where:
∆𝐿𝐿 = variation in length both of the extensometer wire and of the specimen
𝐿𝐿0 = initial length
𝛼𝛼 = material expansion coefficient, or resistivity coefficient
∆𝑡𝑡 = variation in temperature
∆𝐿𝐿 ∆𝑅𝑅⁄𝑅𝑅
𝜀𝜀 = = (2)
𝐿𝐿 𝐾𝐾
Using Eq. 2 and the principle of the Wheatstone bridge, we get to the generic Eq. 3:
Δ𝑅𝑅
= 𝐾𝐾 (𝜀𝜀1 − 𝜀𝜀2 + 𝜀𝜀3 − 𝜀𝜀4 ) (3)
𝑅𝑅
Among the characteristics of electrical resistance strain gages are high measuring accuracy,
excellent dynamic response, and excellent linearity; they can be used immersed in water or in
corrosive gas atmospheres provided that the suitable treatment is made; remote measurements can
be made, etc. Because of such characteristics, this sensor has many applications in experimental
studies.
• ¼ bridge with 2 wires: In this type of connection, the strain gage signal could be affected
by variations in resistance of the cable because of temperature and change in length.
Depending on the cable length, the instrument will not be able to balance the bridge.
Calibration will be incorrect: the more incorrect, the greater the cable resistance. The
advantage of this type of connection is the use of a smaller number of cables. The
application of this connection is indicated for places with controlled temperature, short
cables with a large diameter, and short-term tests. For this type of circuit, absence of
bending and temperature change are required.
This type of connection is suitable to situations of long-term tests with large thermal variation.
In this type of connection, the effect of temperatures and axial strength is eliminated. The readings
of strains will be due to bending only. The signal is amplified two times.
• This type of connection is the most suitable one to be used in situations of long-term tests
with large thermal variation of strain gages where it is necessary to eliminate the bending
effect. In this type of connection, the temperature effect is also eliminated.
By placing gages at selected depths in the pile, we obtain load at various depths and the load
distribution. Using Hooke's Law, we get the Young modulus in the pile's reference section and,
from that value on, the loads at each instrumented level are obtained. The slope of the load versus
strain diagram shown in Figure 5 is the pile axial stiffness, EpA. N.B., the stiffness is obtained
without input of the pile cross section, which is a variable for cast-in-place piles.
For the example of Fig. 5 in which product EpA is equal to 3,142,917 kN, the effective
diameter of the pile (bored with mechanical auger) was 0.44 m in average (A = 0.152 m2). This
way, the Young modulus is about 20.7 GPa.
Another way to analyze the instrumentation data is given by Fellenius (2016), using the
method of tangent modulus as indicated in Eq. 9.
dσ
Mt = = aε + b
dε (9)
From the results shown in Fig. 6, a constant modulus of the order of 24.8 GPa (average) was
obtained.
As the strain gage level lies close to the pile, it is possible for determining both secant and
tangent modulus. The Fig. 7 shows that the tangent plot a bit of scatter, and the secant modulus is
less sensitive to such variations with a smoother curve, even though requires a well-established
zero level (Fellenius, 2016).
Fig. 7. Comparison between stiffness determined from tangent and secant modulus approaches
The experience is vast in the use of instrumentation via strain gages in pile foundations. In general
this technique involves instrumenting a steel bar that may or not be part of the pile cage in the case
of pre-cast or cast-in-place piles. For steel piles, the instrumentation must be made directly on the
metal profile. Below is a simplified presentation of the instrumentation processes.
As a general rule, the technique of installing instrumented bars inside the pile is used for this type
of piles, the difference being the executive procedure of the pile. In general, the instrumentation is
carried out with construction steel bars with 12.7 mm or 19.0 mm diameter, depending on the
diameter of the cross section of the pile. It is not appropriate using large cross section steel bars in
piles with small cross section, because the installation process can be arduous.
The manufacturing sequence of full-bridge bars is shown below. When preparing the bars, it
is important to know at what levels the bars will be placed so as to have appropriate cable lengths.
The cables to be used have 4 paths (full bridge connection), and must have a minimum cross
section area of 0.5 mm2 and an outer protective coating (PVC) and an inner protective lining with
twisted tin copper (Figures 9 - 14).
Fig. 11. Gluing the strain gage Fig 12. Wheatstone bridge connection
Fig 13. Checking the bar on a tensile testing Fig 14. Instrumented bar guide
press
Fig. 15. Fastening the instrumented bar in the pile reinforcing cage
Gages can be installed in continuous auger flight piles or full displacement piles by inserting
a steel tube (smooth or corrugated; with external diameter of approximately 50 mm). For
continuous auger flight piles and full displacement piles, the tube is inserted before the pile is
drilled. Figure 16 shows the tube to be placed into the shaft of the continuous flight auger pile.
Figure 17 shows the tube extending above the pile head after completed installation (Albuquerque
2001).
Fig. 16. Tube being placed in the pile Fig 17. Tube after the pile is constructed
Once the bar installation is completed, the step of forming and casting of the pile cap starts.
Fig. 20 shows the details. We can see the black side tube used as exit for the instrumentations
cables. Fig. 21 shows the cables coming out on the side of the pile crowning block at the time of
the loading test.
Fig. 20. Preparing the crowning block. Fig. 21. Completed block
For this steel piles, the instrumentation must be carried out on site by gluing the strain gages to the
pile section (Albuquerque and Melo 2014; Albuquerque et al. 2016). These spots must be arranged
so as to supply information from different depths as required.
The sensors are directly adhered (by cyanocrylate) to the core, after treating the surface with
a degreasing product. Strain gages are guarded against humidity and mechanical shocks by means
of a silicone resin, adhesive tape, and electrical guarding resin, besides epoxy resins and metal
fishplates in the section made of metal profiles.
The metal profile must be installed in several steps:
Welding of metal fishplates at the joint between the base and the core of the metal profile in
order to provide protection against mechanical shock when the metal profile is driven (figure
22) and cables are passed through;
Cleaning and treating the surface of the metal profile to be fastened to the extensometer later
on;
Fastening the sensors (Figure 23) and applying resin to protect strain gages from humidity and
mechanical shock;
Closing the instrumentation levels between the metal fishplates with epoxy resin for protection
against mechanical shock when driving the piles;
Fastening electrical cables, which are connected to the strain gages along the instrumentation
levels of the metal segment in the upper part of the metal profile.
Lifting and driving the metal profile must be performed with utmost care so as not to break the
cables.
Fig 22. Guarding fishplates. Fig 23. Strain gage fastened to the surface
450
455
460
465
470
475
With this type of piles, the strains are obtained during and after driving. The value of residual
load can be obtained prior to the execution of the slow maintained load test.
4. FINAL CONSIDERATIONS
• The instrumentation technique via strain gages is widely used from medicine through
engineering, which proves that the technique is reliable. However, lack of theoretical and
practical knowledge may lead to unreliable results;
• Using strain gages to get strains and loads in deep foundations is a reliable technique that
can be applied to practically all types of piles;
• It is important to stress that the type of connection to be used (full bridge, ½ bridge or ¼
bridge) must be carefully considered so as to get highly reliable data;
• The instrumentation in cast-in-place piles or pre-cast piles (steel or concrete) is more
effective when the bars have been previously instrumented at the laboratory, and are
installed on site. In the case of continuous flight auger piles or another similar type of
concreting, a different installation technique should be used;
• In the case of steel piles, direct instrumentation in the profile proved to be appropriate;
however the procedure must be carried out on site and very carefully since preparation of
the surface, cleaning and gluing the sensors are key considerations for success. Another
factor that requires attention is during pile driving, since cables may get damaged and
jeopardize the entire process.
Amir, J.M.(1)
(1)
Piletest.com Ltd., Herzlia, Israel <jmamir@piletest.com>
ABSTRACT. Deep foundations have served humanity for the last few millennia but really fast
progress in piling systems and equipment had to wait until the 20th century. Today, piles can be
produced in all soil and rock formations, in diameters reaching four meters and depths of 150 m
or more. Moreover, it has become apparent that even the most advanced piling technology cannot
assure perfect products. As a results, the 1960's triggered the new discipline of integrity testing of
deep foundations. Present methods of integrity testing are either not-intrusive (mainly acoustic) or
intrusive, the latter requiring the installation of access ducts during casting. Currently, the analysis
of test results is invariably an inverse problem, thus, a unique accurate solution is still unavailable.
Future research will have to concentrate on the integration of all testing methods with common
interface and on advanced digital analysis methods.
1. HISTORY
1.1 Piling Technology
Since prehistoric times, humankind has looked for lakeside and riverside dwelling sites that offered
both ample water and protection from attack. To support their dwellings above high water level in
the muddy soil, driven timber piles were the obvious solution. Modern radiocarbon dating
technology established that timber piles recently discovered in London (Figure 1a) were more than
6,000 years old and are still in reasonable shape (Milne et al. 2010). This foundation method was
successfully practiced over the millennia in cities like Venice and Amsterdam. A painting by
Maximilien Luce (Figure 1b) proves that manual piledrivers were used to drive timber piles as late
as the twentieth century even in a developed city like Paris.
a b
Fig. 1: Wood piles - (a) London 4,600 BC, (b) Paris1905.
When we hit the pile head with a hammer, we create a compressive wave that travels
downwards along the pile. Βy combining Newton's second law with Hooke's law (Vincke and van
Nieuwenburg 1987) we get the one-dimensional wave equation:
𝜕𝜕2 𝑢𝑢 𝜕𝜕 2 𝑢𝑢
= 𝑐𝑐 2 (1)
𝜕𝜕𝑡𝑡 2 𝜕𝜕𝑥𝑥 2
Where u is the displacement and 𝑐𝑐 = √(𝐸𝐸/𝜌𝜌) is the wave speed in the rod. In concrete piles, for
instance, the harder the concrete the faster the waves. The general relation between concrete
compressive strength fc and wave speed is given by Amir (1988):
Clearly, c is determined by the physical properties of the pile and does not depend on the
strength of the blow. Actually, it is exactly the same for a large pile driver and a small plastic
hammer. For concrete grades used in piling it will typically vary between 3,600 and 4,400 m/s.
Particle velocity v, on the other hand, is a function of the stress, σ, applied to the head. The ratio
between the two, v/c, is equal to σ/E. Therefore, the force P in a given section is given by:
P = σA = (EA/c)v = Zv (3)
Where L is the pile length, D is the pile diameter, 𝜌𝜌𝑠𝑠 and 𝜌𝜌𝑐𝑐 respectively are the soil and pile
densities, vs is the shear wave speed in the soil and c the wave speed in the pile. For a pile with a
typical slenderness ratio L/D of 25 in a soil with a shear wave speed of 250 m/s, the total attenuation
A is 4.7 Neper or e4.7 = 110. For modern digital instruments, it is a simple matter to amplify the
reflections by this amount, but at the same time, any noise present will also be amplified.
• P-Waves – longitudinal wave in which particle motion coincides with the direction of
propagation
• S-Waves – transverse wave in which particle motion is perpendicular to the direction of
propagation.
• Rayleigh waves – P-Waves that advance close to a free surface in which the particles move
is both longitudinal and transverse directions.
For good-quality concrete, P-waves move at a speed of 4,470 m/s. S-Waves and Rayleigh
Waves are typically slower by 42% and 48%, respectively. In comparison, wave speed in a rod of
the same concrete material is 4,080 m/s.
The Impact Echo test is fast and inexpensive, with typically less than one minute needed to
test a given pile (Amir & Amir 2008). The main drawbacks of the Impact Echo method are:
• The wavelength produced by the plastic hammer is in the order of 3-4 meters. Therefore,
the information it provides regarding the part of the pile close to the head is limited.
• The lengths reported are a function of the assumed wave speed that is usually unknown.
• It is influenced by skin friction, thus its effective penetration depth varies between 10
diameters in very hard soils and 60 diameters in very soft soils.
• The pile head must be accessible.
The testing methods described in the following sections have tried to address some of these
shortcomings.
z = c/(2∆f) (6)
In addition, the inverse of the slope of the initial part of the mobility plot is equal to the
apparent low strain rigidity of the pile head. This method did not gain popularity because of the
bulky vibrator and the elaborate preparations it required. This situation has changed in the 1970's
• The method is sensitive to inhomogeneity in the surrounding soil, such as saturated sand
or ground water flow.
• The method is sensitive to variations in the amount of retarder used, especially in large
piles cast by several truck mixers.
• The method may totally miss a discontinuity of small vertical dimension.
• The test must be performed once the pile is hot enough, and before it had time to cool
down too much. Once this time windows was missed for any reason, there is no second
chance. This limitation can be overcome by using embedded strings of thermistors
connected to an automatic data logger, but this option is obviously more costly.
The first assumption is trivial, but the second one is not. While for driven steel piles the one-
dimensional wave equation may be applicable, for the purpose of integrity testing, all the
assumptions on which it is based are more or less wrong. A pile with flaws, for instance, is neither
prismatic nor homogeneous, and the soil profile for a specific pile is known only approximately.
Under the circumstances, the best we can sometimes offer is a reasonable estimate of the pile
length. Evidently, the wave equation approach does not deliver and the most sophisticated
analyses will produce little more than an axisymmetric pile profile with no direct bearing on the
structural properties of the pile (Figure 8).
By definition, intrusive methods are much more informative than non-intrusive ones but they
too have their deficiencies. The engineer specifying the test has therefore to make a compromise
based on type of the piles, equipment availability and budget. In case of uncertainty, the prudent
engineer will muster some redundancy: hire another laboratory to repeat the test (a practice adopted
in Hong Kong after the 1999 fraud) or test with other methods. The cross-hole ultrasonic test, for
example, can be supplemented by single-hole testing to investigate the external part of the pile.
Fig. 9 illustrates the case of a pile with a 15% flaw at 4 m depth tested by impact echo with no
anomaly and by single hole ultrasonic clearly showing increased FAT and attenuation. Of course,
redundancy may produce conflicting results, where sound engineering judgement is unavoidable.
Fig. 9: A pile with a 15% flaw at 4 m – impact echo (left) and single hole ultrasonic (right).
• It can model piles with any shape and physical properties (Fig. 10)
• It can accurately model the surrounding soil profile, even if irregular, with proper
geotechnical parameters
• It can model the superstructure in cases where there is no access to the pile head
• The input force can be static or dynamic, transient or steady state, axial or lateral, concentrated
or distributed.
• The output displacements can be monitored at any convenient location.
By a rough estimate, this project may take at least ten years and several millions of Dollars
to complete. Initially the system could be developed in two dimensions (axisymmetry) and later
in full three dimensions. Once complete, the system will be capable of positively detecting at least
90% of important flaws. This will certainly lead piling technology towards better, safer and more
economical foundations.
Fig. 10: Piles with bulges (left) and finite element simulation (right).
Brown, D.(1)
(1)
President, Dan Brown and Associates, Sequatchie, TN, USA <dbrown@dba.world>
ABSTRACT. The last few decades have seen significant advances in construction practices for
drilled foundations, and these developments have greatly affected foundation engineering
regarding the selection and use of deep foundations. Advances in construction have come in
response to greater challenges of 21st century infrastructure development, leading to technological
advances in equipment, materials, and testing. This paper describes some of the more significant
developments reflected in the author’s experiences in North America and is intended to provide
foundation engineers with increased awareness of modern capabilities and potential applications.
1. INTRODUCTION
The modern-day world presents an increasingly complex situation for foundation engineering. The
many challenges include congestion in the urban environment, with replacement or enhancement
of existing infrastructure while maintaining service. New structures are planned for increasingly
difficult geologic settings, since the “easy sites” are already occupied. Modern foundation designs
are subject to increased complexity and demands from seismic and other extreme event loadings.
Environmental constraints present further challenges to construction, as the public demands that
we minimize noise, vibration, and impacts to the natural environment. Reduced public tolerance
for pile driving activity has encouraged greater use of drilled foundations. Project delivery has
modified contractual relationships and financing aspects of projects, with the inevitable pressure
on foundation engineers and specialty contractors to take on risks and deliver faster.
Industry has responded to the market pressures associated with these challenges with
significant advances in construction technology, and thus foundation engineers have an incredible
variety of drilled foundation options at their disposal. This situation is both a blessing and a curse;
it presents opportunities for innovation and creativity, but also presents challenges in identifying
the optimal solution for a given situation. Constructors are more often engaged in the foundation
type selection and must therefore understand design issues and similarly designers are required to
have greater understanding of construction techniques and capabilities.
This paper describes some of the more significant advances in the construction of drilled
foundations. It is not intended to be an exhaustive “state-of-the-practice” paper, but rather a
reflection of the more notable advances. Brief descriptions of selected case histories from the
author’s experience are provided to illustrate the implementation of modern drilled foundation
technology. The author’s experience derives primarily from his practice in North America, and it
is recognized that the worldwide implementation of these technologies varies.
Although the definition of different technologies varies around the world, for purposes of this
discussion drilled foundations will be generally described using four categories in accordance with
the drilling techniques used. Micropiles are small diameter drilled piles installed using lightweight
drilling equipment in which the cuttings are generally flushed to the surface with a circulating
fluid. Continuous Flight Auger Piles (CFA piles) are installed by advancing the continuous auger
into the soil and then placing the concrete through the hollow stem as the augers are withdrawn.
Reinforcement is subsequently placed into the fluid concrete. Drilled displacement piles are
2. MICROPILES
Micropiles are most often 300 mm or less in diameter and often selected for use because of the
advantages provided by the lightweight and maneuverable equipment available to install these
piles and overcome site access difficulties. The use of these foundations has grown dramatically
during this period, and the popularity of this type of drilled foundation is largely the result of the
versatility of the equipment (as shown in Figure 1) used to install them. Micropiles are used for
applications including underpinning and seismic retrofitting and in locations and ground
conditions where more conventional deep foundations would be difficult or impossible to
construct. The use of micropiles has evolved from small, lightly loaded “root piles” into the current
practice where micropiles are designed to support substantial individual pile axial forces in the
same way as other types of piles.
a) Foothills Bridge, East Tennessee b) World Trade Center, New York City
Fig. 1. Micropile Drill Rigs in Restricted Access Locations.
Besides the ability to overcome difficult site access, micropiles can be drilled to provide good
foundation support into materials which are impossible to penetrate with driven piling or which
represent extremely difficult drilling conditions with larger diameter drilled foundations.
Examples include piles through boulders, fills including rubble or other hard debris, and karstic
formations in hard limestone. The micropiles may be advanced through porous layers with casing
In a typical application, the structural loads dictate the size of the steel element and then the
embedded length is determined to provide the geotechnical resistance necessary for the transfer of
load from the steel through the grout to the soil or rock. The transfer of axial load is typically
accomplished through side resistance in the portion of the pile below the casing (the bond zone),
with no reliance upon side resistance in the permanently cased zone or in end bearing.
The unit side resistance in the bond zone is not only affected by the type of soil or rock, but
may be strongly affected by the type of construction practice used for drilling and for grouting.
Because of the strong influence of construction technique, the average nominal unit grout-to-
ground bond strength is usually estimated empirically and verified through site-specific load
testing, with the final micropile geotechnical design performed by the specialty contractor. By
working in this way with either performance-based specifications or a design-build type of
contract, the contractor has the ability and responsibility to select the most appropriate and cost-
effective drilling and grouting techniques. The constructor thus typically has some design
responsibility and incentive to improve geotechnical performance, and static load tests are
routinely employed to provide verification of axial resistance.
A major factor in the broad acceptance and increased use of micropiles within the last 15 years
has been the work of industry groups to develop and promote standards, share knowledge and
expertise, and transform the technology into a more universally accepted foundation option. The
practice in the U.S. has been shaped by the joint micropile committee of the Deep Foundations
Institute and ADSC: The International Association of Foundation Drilling, whose members have
worked with the Federal Highway Administration and the National Highway Institute to produce
design and construction guidelines and training materials. An FHWA reference manual was
published in 2000 (Armour, et al. 2000) and subsequently updated (Sabatini, et al. 2005), which
provides a widely-used reference for the design and construction of micropiles. Micropiles were
only recently incorporated into the AASHTO Bridge Design Specifications in 2007 and the
International Building Code (IBC) in 2006.
Pier
Residual Soil
Sound
Weathered
Rock
Rock
Fig. 4. Micropile Foundation Design for a Typical Pier, Foothills Parkway Bridge No. 2.
Analyses of the pile group foundation for bridge foundation loads from the pier were used to
determine shear, moment, and axial demands on the individual piles and the design completed in
accordance with the AASHTO 2007 LRFD guidelines. The shear and moment demand is resisted
by the grout-filled permanent casing, which is 244 mm (9-5/8 inch) diameter, 12 mm (0.472 inch)
wall thickness, 550 MPa (80 ksi) yield strength, and extends through the overburden soils and
weathered rock zone. The casing was also installed so that no joint was located within 2.4 m (8 ft)
of the top of the pile beneath the footing. The piles include a No. 18 center bar (57 mm, or 2-1/4
inch diameter) with 414 MPa (60 ksi) yield strength.
The maximum factored axial load demand of 1,380 kN (310 kips) is resisted by the 203 mm
(8 in.) diameter uncased portion of the pile which extends 4.6 m (15 ft) into the rock. This socket
is designed for a nominal unit side resistance of 690 kPa (100 psi) and a resistance factor of 0.7.
The axial resistance was confirmed by load tests.
The key factor in utilizing micropiles for the Foothills Bridge was the ability to position a
small rig into place to install piles (shown in Figure 1a) with a minimum impact on the rugged and
Crossover
Sub-Seal
Crossover
Sub Port
Shock
Absorber
Hammer
Pilot Bit
a) Side loading the auger due to excessive b) Soil loosening due to excessive rotation
rotation (from Fleming, 1995) measured by CPT (from Van Weele, 1988)
Fig. 10. Effects of Excessive Rotation of CFA Auger.
a) positive pumping pressure is maintained at the point of discharge at the bottom of the
augers,
b) a structural defect or neck in the pile does not result from pulling the auger string too fast,
and
c) wasteful pumping of excess concrete or grout does not occur, particularly in soft soils
where overconsumption would provide little or no benefit.
Through much of the history of the use of CFA piles, the skill of an experienced drill rig
operator has been recognized as a critical component, because the “feel” of the operator was always
so important to both advance the drill effectively and withdraw the drill during concreting in the
correct way. The use of modern electronic controls, shown in the photos of Figure 11, provides
the operator with direct feedback measurements on the critical parameters and also the ability to
document that the pile has been constructed in accordance with good practices. These days, rig
operators, having grown up playing electronic games, are quite comfortable operating a joystick
and using a graphical electronic display. For the constructor, the monitoring can also provide a
measure of productivity, since some systems provide a minute-by-minute log of the activity of the
drill rig. Equipment maintenance requirements represent another common function that may be
included as a part of the on-board computer system.
The most important control parameters include the rate of penetration and sometimes the
applied torque and crowd (down force) on the tools, rotation rate, the concrete or grout pressures,
and the volume of concrete or grout pumped as a function of the elevation of the auger tip and the
theoretical volume required to that point. When these parameters are calibrated to site-specific
loading tests the use of automated monitoring provides a high level of quality control and quality
assurance. The monitored parameters are recorded and documented in a production log that
provides a record of the successful completion of each pile.
The torque and crowd required to construct a drilled displacement pile is substantial, and the
modern fixed-mast hydraulic drill rigs are typically used for these piles. Because the pile fully
displaces the soil, there are no issues related to over-rotation of the auger and potential loosening
of the soil as described for CFA pile construction. The energy required to install the pile is related
to the resistance of the soil to the displacement, and so the piles are often installed to a depth that
is controlled by the capabilities of the drilling rig. The potential effect of lateral displacement or
heaving on nearby structures may also be a consideration.
With the monitoring equipment capabilities described previously for CFA piles, it stands to
reason that the measure of torque and crowd as a function of penetration might logically be related
to the axial resistance of the completed pile in a manner similar to a CPT sounding. Variations in
stratigraphy are readily detected, i.e., the penetration into a denser stratum is immediately evident
by the measured torque and crowd required to maintain penetration. Although a broad
Heavy line
is post
installation,
light gray
line is pre-
installation
Fig. 13. Effect of Drilled Displacement Pile Installation on Cone Tip Resistance in a Sandy Soil
(from Siegel et al. 2007).
Because of the ground improvement benefits with drilled displacement pile equipment, these
piles are popular for construction of pile raft foundations. The delineation between what is to be
called a “pile” as opposed to a “rigid inclusion” or “column” in terms of ground improvement
technology has become obscured and the terminology used may often reflect the design approach
with respect to building code requirements. When used primarily as a ground improvement
technique, the structure may in fact be designed to bear on spread footings that are not connected
-10
Displacemetn, mm
-20
-30 Test 1
-40 Test 2
Test 3
-50
-60
Fig. 14. Load Tests of Drilled Displacement Columns Supporting a Spread Foundation.
The presence of a fully cased hole also provides a reduced risk of soil caving during concrete
placement, and therefore improved reliability for construction in applications such as bridge
foundations where flexural demands require the use of large diameter drilled shafts. Where artesian
groundwater conditions are present, the water table inside the casing can be readily maintained at
an elevation well above the ground surface to provide sufficient head within the shaft excavation
to counterbalance the artesian condition.
Katzenbach et al. (2007) reviewed available load test information on drilled shafts constructed
using the oscillator and rotator segmental casing method and report that the results are comparable
and in some cases favorable to other installation techniques. One factor that favors the performance
of drilled shafts constructed using this method is the fact that the teeth that are used on the cutting
shoe at the bottom of the casing tend to produce a roughened surface at the concrete/soil interface
as the casing is extracted. An opportunity to examine the surface of drilled shafts constructed using
this construction method was provided recently at the Huey Long Bridge in New Orleans (Brown
et al. 2010). The 2.8 m (9 ft) diameter drilled shafts were exposed within the sheet pile cofferdam
after placement of the seal slab and during construction of the footing. These foundations were
constructed prior to excavation of the cofferdam, with a corrugated metal pipe used as a temporary
Fig. 16. Exposed Texture on the Drilled Shaft Surface, Huey P. Long Bridge.
Reverse-circulation drilling is another technique that has been increasingly used in recent
years to construct drilled shafts to large diameters and depths. This drilling technique provides full
face rotary cutting at the base of the excavation with the drilling fluid used to remove cuttings via
air-lift pumping up through the center of the drill pipe. This closed system avoids the need to cycle
in and out of the hole to remove cuttings from an auger and can also be very effective in excavating
rock.
The photos on Figure 17 illustrate the equipment used with this technique. The system in
Figure 17a is mounted onto a casing that was installed with a rotator, and is working in the space
beneath an existing bridge on I-90 in Connecticut to install 2.8 m (9 ft) diameter drilled shaft
foundations into the bedrock for the replacement bridge structure prior to demolition of the old
one. The drill removes the cuttings by pumping the cuttings and fluid up through the center drill
pipe, through the swivel at the top, and on to a spoil container via the discharge hose in the
foreground. Drilling fluid is simultaneously pumped into the top of the excavation through a return
The Wolf Creek Dam project in Kentucky is an example of the advancement of drilled shaft
equipment and technology to overcome challenges in a way that was not possible years ago.
Seepage through the underlying limestone bedrock below has threatened the stability of the earth
dam that retains Lake Cumberland, the largest reservoir east of the Mississippi. A previous cutoff
wall had been constructed into the bedrock in the late 1970’s using the best available technology
at that time, but the seepage problem was not successfully resolved by that effort. Seepage has
found new paths under and around the wall, leading to sinkholes and soft wet areas downstream
as well as high measured pore water pressures in the embankment. The Wolf Creek Dam was in
critical need of remediation to correct the problem.
The key component of the repair to the dam is the construction of a secant pile cutoff wall,
and the construction of this wall utilizes reverse circulation drilling to construct drilled shafts to
very great depths. After lowering the reservoir and completing an initial grouting program, the
cutoff wall is constructed through the dam from a bench on the upstream face, as shown in the
photo of Figure 18.
First, a 1.8 m (6 ft) wide concrete diaphragm wall is constructed through the embankment to
the top of rock at a depth of around 25 to 30 m (80 to 100 ft). The secant pile wall is then
constructed to depths of up to 84 m (275 ft) through the concrete diaphragm wall and the karstic
limestone and into a sound limestone layer. The secant pile excavation is started using
conventional drills with rock augers to open a hole to a depth of around 15 m (50 ft) into the
diaphragm wall, and then completed using reverse circulation drilling as illustrated in the photos
of Figure 16. In order to maintain the alignment on such deep drilled shafts and ensure that the
secant piles overlap to form a water-tight cutoff wall, a pilot hole is first installed using directional
drilling techniques. The reverse circulation drill is equipped with a “stinger” to follow the pilot
hole and maintain the alignment during drilling.
Fig. 19. Reverse Circulation Drilling for Secant Pile Cutoff Wall.
An example of the use of this technique on a major project is described by Dapp and Brown
(2010) for the John James Audubon Bridge over the Mississippi River in Louisiana, which utilized
drilled shafts founded in dense alluvial sand. Each of the two pylon foundations for the cable-
stayed bridge included 21 drilled shafts which were 2.3 m (7.5 ft) diameter and approximately 60
m (200 ft) deep. The base grouting was accomplished via a sleeve-port system (tubes-a-manchette)
that utilized the crosshole sonic logging tubes. The eight tubes were connected in pairs across the
base of the drilled shaft to form four separate U-shaped circuits, as shown in the photo of Figure
20a).
The project included load tests using the a bi-directional load cell at the base of both grouted
and ungrouted drilled shafts to provide a comparison of performance for full scale foundations.
Figure 21 illustrates the improvement in base resistance achieved by base grouting to a pressure
of approximately 5 MPa (750 psi). The data shown are measured load at the base of the drilled
-1 -25.4
Displacement of O-cell plate (inches)
-3 -76.2
-4 -101.6
-5 -127.0
T3 2.1m (7ft) dia;
(Not Grouted)
-6 all others are -152.4
2.3m (7.5ft) dia
-7 -177.8
0 500 1,000 1,500 2,000 2,500 3,000 3,500
O-cell load (tons)
Fig. 21. Measured Base Resistance from the Base-Grouted Drilled Shafts at the John James Audubon
Bridge, Louisiana.
Specifications for polymer slurry construction often have evolved from those used for
bentonite, but the differences in performance of polymers require several modifications.
Experiences with polymer slurry construction indicate that designers should be aware of several
factors that affect practice.
The upper limits on viscosity used for bentonite are too restrictive for polymer; there is a need
to limit the viscosity of bentonite to avoid excessive filter cake buildup, but polymers can utilize
significantly higher viscosity in order to provide effective stabilization.
Unlike bentonite slurry, the density of polymer slurry will not be much higher than that of
water, and so where groundwater levels are very near the top of the drilled shaft it is critical that a
positive head of 2 m or more is maintained at all times. Where groundwater levels are near or
above the ground surface, it will be necessary that the contractor extend the casing above grade to
provide this head, plus additional distance to allow for fluctuations and working freeboard within
the casing.
Because there is no bentonite filter cake, it is not necessary to limit the exposure time of the
soil to slurry and/or require agitation of the sidewall. There is also substantial evidence that drilled
shafts constructed using polymer slurry result in higher values of unit side resistance in sands and
silts than similar foundations constructed using bentonite (Brown et al. 2002; Brown 2002, Meyers
1996).
Even where natural clays or shales are encountered in the soil, the polymers as shown in Figure
22b tend to stabilize these soils and prevent mixing of the clay with the drilling fluid. Many
contractors like to employ a small amount of polymer slurry when drilling through clay because it
reduces the tendency for clay to stick to the auger. There is also evidence that polymer drilling
fluids may reduce wetting of some shales and thereby reduce the tendency for degradation of shale
when the excavation is open. This behavior can result in improved side resistance for drilled shafts
socketed into shale, especially for large drilled shafts where several days may be required to
complete construction.
Where fine sands and silts are present, polymer slurry can present a challenge from the
standpoint of cleaning the slurry. These sands and silts will not stay in suspension and will tend to
settle out slowly after completion of excavation. The de-sanding units used with bentonite slurry
construction do not work with polymer because the polymer molecules would be destroyed by the
shearing process in the de-sander and polymers will also tend to clog the screens. De-sanding of
polymer is normally accomplished by adding flocculants to help promote the settling of solids, a
process that requires that the slurry be maintained in a calm environment so that the sands can
settle out. Flocculation can occur either in the borehole (followed by pumping from the base of the
hole to remove solids) or in a weir tank after removal and replacement of the fluid in the hole with
clean slurry. In a very deep drilled shaft excavation filled with sand-contaminated polymer, solids
can rain out of suspension for days and if left untreated could result in contamination of concrete
during placement. Although flocculants can be used in the borehole to accelerate the process, the
most reliable approach is to fully exchange the fluid by pumping slurry from the base of the shaft
to a holding tank while adding clean slurry into the top of the excavation.
CSL Tube
anomaly
Conventional static top down load tests are still occasionally used with drilled shafts, and load
tests of up to 50 MN (11,000 kips) have been performed with static reaction systems. Other
advancements with drilled shaft load testing include the use of rapid load testing and high strain
dynamic testing with signal matching, as would be performed on a driven pile. The rapid load
testing method is most often employed using the Statnamic® device, which launches a reaction
mass upward with about 20g of acceleration resulting in a downward thrust onto the drilled shaft.
This test method offers a relatively economical means of verifying axial resistance from the top
down without the need for a reaction system, and can often be performed on production
foundations. The equipment available to perform rapid load testing is currently limited to a
maximum applied force of around 45 MN (10,000 kips), and the maximum static resistance which
can be mobilized is slightly lower due to inertial and rate-of-load effects.
The major implications of the advancements in testing for high capacity drilled shaft
foundations are:
• Designers now have the means to obtain measurements that will provide the feedback
necessary to improve design practices.
• Alternative forms of project delivery such as design-build can now include
performance measurements for verification, and the availability of such testing can
allow for performance-based specifications to be employed in the design-build
process.
An example of the use of load testing for verification in design-build is the Honolulu Transit
project currently under construction. The first phase of this project includes approximately 10 km
(6 miles) of elevated guideway to be constructed in a tight space within existing right-of-way. A
single drilled shaft foundation at each pier provides maximum support in the minimum footprint.
O-cells
Fig. 24. Load Test Shaft for the Mississippi River Bridge, St. Louis.
6. ACKNOWLEDGMENTS
The author wishes to acknowledge the contributions of the following individuals with respect to
the information conveyed in this paper: John Turner, Robert Thompson, Steve Dapp, Paul Axtell,
and Tim Siegel of the author’s engineering firm; John Wolosick of Hayward-Baker; Erik Loehr of
the University of Missouri; Andy Burns of Intercoastal Foundations and Shoring; Willie NeSmith
of Berkel and Company; Heinrich Majewski and Peter Faust of Malcolm Drilling Company;
Brannin Beeks of Raito, Inc.; Jim Holtje of PCL Construction; Steve Saye and Luis Paiz of Peter
Kiewit and Sons; Fabio Santillan of Trevi Icos Corp.; and John Kelley of Massman Construction.
American Association of State Highway and Transportation Officials, AASHTO 2008, AASHTO
LRFD Bridge Design Specifications, Customary U.S. Units, 4th Ed., Section 10,
‘Foundations’, Washington, D.C.
Armour, T., Groneck, P., Keeley, J., and Sharma, S. (2000). Micropile Design and Construction
Guidelines Implementation Manual Report No. FHWA-SA-97-070, June, 376 p.
Atlaee, A., Burns, A. and Shah, H. (2010). Hammer-Grout Piles at the Bronx-Whitestone Bridge,
Proceedings of the Deep Foundations Institute 35th Annual Conference, Hollywood, CA, 11p.
Axtell, P., Thompson, R., and Brown, D. 2009. Drilled Shaft Foundations for the kciCON Missouri
River Bridge. Proceedings of the Deep Foundation Inst. 34th Annual Meeting, Kansas City,
10p.
Brown, D. (2002). The Effect of Construction on Axial Capacity of Drilled Foundations in
Piedmont Soils, J. of Geotechnical and Geoenvironmental Engineering, 128(12), pp. 967-973.
Brown, D., Axtell, P., and Kelly, J. (2011). The Alternate Technical Concept Process for
Foundations at the New Mississippi River Bridge, St. Louis, Proceedings of the Deep
Foundation Institute 36th Annual Conference, Boston, pp. 169-178.
Brown, D., Bailey, J. and Schindler, A. (2005). The Use of Self-Consolidating Concrete for Drilled
Shaft Construction: Preliminary Observations from the Lumber River Bridge Field Trials,
Proceedings of the GEC3 Conference, Dallas, TX, pp. 437-448.
Brown, D. and Chancellor, K. (1997). Instrumentation, Monitoring and Analysis of the
Performance of a Type-A INSERT Wall – Littleville, Alabama, Final Report RP 930-335,
Highway Research Center, Auburn University, 105p.
Brown, D., Dapp, S., Thompson, R. and Lazarte, C. (2007). Design and Construction of
Continuous Flight Auger Piles, Geotechnical Engineering Circular No. 8, Federal Highway
Administration Office of Technology Application, Office of Engineering/Bridge Division,
294p.
Brown, D.A. and Drew, C. (2000). Axial Capacity of Augered Displacement Piles at Auburn
University. Geotechnical Special Publication No. 100, ASCE, pp. 397-403.
Brown, D., Faust, P. and Santos, J. (2010). Construction of the Drilled Shaft Foundations for the
Huey P. Long Mississippi River Bridge, New Orleans, Proceedings of the Deep Foundation
Inst. 35th Annual Meeting, Hollywood, CA, 8p.
Brown, D. and Loehr, E. (2007). A Simple Solution for Slope Stabilization Using Micropiles Proc.
of the 32nd Annual Conf. on Deep Foundations, Deep Foundations Inst. Oct., pp. 93-104.
Brown, D., Muchard, M., and Khouri, B. (2002). The Effect of Drilling Fluid on Axial Capacity,
Cape Fear River, NC. Proceedings of the Deep Foundation Inst. 27th Annual Meeting, San
Diego, 5p.
Brown, D., Turner, J., and Castelli, R. (2010). Drilled Shafts: Construction Procedures and LRFD
Design Methods, FHWA/NHI Publication 10-016, Reference Manual and Participants Guide
for National Highway Inst. Course 132014, 972p.
Dapp, S. and Brown, D. (2010). Evaluation of Base Grouted Drilled Shafts at the Audubon Bridge.
Geotechnical Special Publication 199, Advances in Analysis, Modeling and Design, ASCE,
10p.
Esrig, M.I., Leznicki, J.K.,and Gaibrois, R.G. (1994). Managing the Installation of Augered Cast-
in-Place Piles, Transportation Research Record 1447, Transportation Research Board, pp. 27-
29.
Fellenius, B.H.(1)
(1)
Consulting Engineer, Sidney, BC, Canada. <bengt@fellenius.net>
ABSTRACT. When determining "capacity" from the result of a static loading test, the
profession has no common definition of "capacity". Thus, a group of professionals will come up
with an array of capacity values from the same test results. To improve reliability, piles are
usually instrumented with strain-gages at selected depths. The analysis of the strain records apply
the axial secant stiffness of the pile, which is best determined applying the direct secant modulus
method to the records from a gage level near the pile head and applying the indirect method, the
incremental stiffness method, to records of gage levels down the pile. Examples demonstrate the
necessity that a test be carried out with no unloading-reloading events included and that all
increments be equal and the load levels be held constant for an equal duration not shorter than 10
minutes. If so, the analysis of the records will produce the load distribution along the pile and the
pile toe load-movement response. When combined with the calculated soil settlement around the
piles, the designer will be able to determine the settlement of the piled foundation. The approach
is far more reliable and appropriate for a piled foundation design than one based on some
perceived value of "capacity" with the settlement estimate just referenced directly to the load-
movement of the test.
1. INTRODUCTION
In countries with a stagnant and code-driven piled foundation system, loading tests are rarely
performed. In contrast, in countries with an advanced and variable foundation industry, testing is
a fundamental part of the engineering process. Sometimes, the tests are performed as a part of a
design effort, sometimes they are performed for proof testing during or after construction. Most
of the times, the tests are conventional head-down test with no instrumentation down the pile.
Testing an uninstrumented pile has little value for design, however. Moreover, including
unloading-reloading steps in a test or uneven length of load-holding will adversely affect the
strain records. The pile axial stiffness, EA, can be determined from the measured strains by
applying direct secant and incremental stiffness methods, as is illustrated in this paper. However,
determining the axial loads imposed by the test can be difficult even when the test is properly
planned and executed.
2 CAPACITY
The dominant approach to assessing the results of a static loading test is to determine a pile
capacity from the pile-head load-movement curve. The term "capacity" implies an ultimate
resistance and is, in its purest form, defined as a continued movement for no increase of load
once reached, i.e., plastic response after an initial linear ("elastic") response. Figure 1 shows the
test results of a 14 m long, 400 mm diameter, bored pile equipped with a telltale to the pile toe:
the pile-head load is plotted against the movements of the pile head and the pile toe. The profile
consists of about 4 m of sand on 7 m of clay on a thick layer of dense sand. No obvious "kink" in
1,500
LOAD (kN)
1,000
500
0
0 5 10 15 20 25 30 35 40 45 50
MOVEMENT (mm)
Fig. 1 Pile-head load-movement curve
What definition to employ differs widely within the geotechnical profession, as illustrated
by the following example from a prediction event organized by the Universidade Federal do Rio
Grande do Sul in the Araquari Experimental Testing Site, Brazil in 2015. The event comprised
a 1,000-mm diameter, 24 mm long bored pile constructed in bentonite slurry a in sand deposit.
The premise of the prediction was that the test be carried to a capacity defined as the load that
resulted in a pile head movement determined as 10 % of the pile diameter (100-mm), a definition
of "capacity" taken from the EuroCode that has its root in a misconstrued recommendation by
Terzaghi (Likins et al. 2012). Moreover, the definition does not consider obvious aspects such as
whether the pile is driven, bored, advanced by CFA or by full displacement methods, installed in
open borehole, if the hole was maintained by means of slurry, the pile constructed by full-
displacement method, or if the soils are clay or sand.
The task was to predict the pile-head load-movement curve for the test pile until the 10-%
defined maximum load was reached. After the test results had been published, I contacted all
predictors and asked them to tell me, using their own definition, what capacity the actual test
curve demonstrated. Twenty-nine, about half of the total, replied and Figure 2 compiles the
capacities received. The values diverged considerably. Seven accepted the organizers' assertion
that the capacity was the load that gave a movement equal to 10 % of the pile diameter, whereas
the others indicated values that were as low as two-thirds of the maximum—with a 21-mm
movement, as opposed to the 100 mm value stipulated by the organizers. Compilations from
other cases have been published that show similar diversity (Fellenius 2016b, Fellenius and
Terceros 2014). It is obvious that to make use of results of static loading tests performed by
others, a researcher cannot take stated "capacities" at face value, but needs to re-assess the results
in order to develop a consistent data base.
An additional phenomenon affecting the shape of the pile-head load-movement curve and
the assessment of "capacity" is presence of residual force, which is an environmental axial force
introduced in the pile during or after installation or construction of a pile. It is almost always
present in a driven pile and, on occasion, also in a bored pile. For example, jacked-in piles have a
more or less fully developed residual force. The residual force is always zero at the pile head or
at the ground surface, it then increases due to accumulated negative skin friction to a maximum
value at a "neutral plane" where it is in equilibrium with the below this point developing positive
shaft resistance and, usually, some residual toe force.
Figure 3 compares the pile-head load-movement curves of two piles, one with fully
developed residual force and one with no residual force. The piles and soil are otherwise
identical in all respects, including using the same ultimate resistance and load-movement
response for the pile elements. The figure includes the pile "capacities" determined according to
the in North America common offset-limit method (Davisson 1972, Fellenius 2017). Other
definitions show a similar difference between the capacity values determined from the two
curves. Obviously, presence or not of residual force will affect the evaluation of the results of a
test in addition to the fact that the capacity evaluated from the pile-head load-movement curves
are very different. The example emphasizes the uncertainty of the concept of "capacity".
1,000
800
Residual
Load present
LOAD (KN)
600
No Residual
Load present
400
200
= OFFSET
LIMIT LOAD
0
0 5 10 15 20 25 30
MOVEMENT (mm)
2,500
y = -0.071x + 2,617
2,000
1,500
0 100 200 300 400 500 600 700 800
STRAIN, με (---)
Fig. 4. Pile stiffness for a 400-mm CFA pile (after Fellenius 2012).
The direct secant modulus plot for the gage records indicates an initial pile axial stiffness
(EA) of 2,600 MN reducing with increasing strain as the load increases. The decrease with
increasing strain is minute for the case. The important observation is that the stiffness is obtained
directly from the measurements and does not depend on a frequently uncertain value of the pile
cross section and a modulus obtained from calculations or testing of a separate specimen.
The direct secant method depends very much on the accuracy of the load and strain records.
As mentioned, if the pile has been unloaded and reloaded—whether this was intentional or not
matters little, the strain readings will be adversely affected by the so-introduce hysteresis effect.
Figure 5 shows the secant analysis of strain-gage records from a gage level in a 600-mm
diameter, octagonal, 33 m long, prestressed concrete pile driven into a sandy silt. The gage level
is located about 1.5 m below the pile head and 1.0 m below the ground surface. The pile was
loaded in twenty-six 150-kN increments to 3,900 kN maximum load. The night before the test,
an unprescribed check of the test set-up was carried out involving loading the pile to about
1,000 kN without taking records. As seen, the incident resulted in a disturbance of the initial part
of the secant stiffness line.
30
SECANT STIFFNESS, EA (GN)
20
EAsecant (GN) = 11.2 - 0.005µε
y = -0.0053x + 11.231
10
40
INCREMENTAL STIFFNESS (GN)
P3 Stage L1
SGL11
35
Tangent Stiffness, EtA (GN) = 32.6 - 0.020 µε
gives the Secant Stiffness, EsA (GN) = 32.6 - 0.010 µε
30
y = -0.0195x + 32.587
R² = 0.7729
25
20
0 100 200 300 400 500
STRAIN (µε)
The loads translated from the strain records apply the secant stiffness times the strain, Es A ϵ.
The y-intercept (stiffness at zero strain) is the same for the secant and incremental plots and the
slope of the straight line in the incremental plot is twice that of the secant line. Thus, for the
shown plot, the secant stiffness, EsA (GN), is 32.6 - 0.010µϵ. Because the incremental method is
based on differentiation, small inaccuracies in the records are exaggerated and an incremental
stiffness plot is always more scattered than a direct secant plot. Thus, provided that the tests is
carried far enough so the applied load has imposed sufficient strain in the pile for a meaningful
plot of the data, the records can be used to determine the stiffness of a test pile at the various
gage levels down the pile (Fellenius 1989).
Figure 7A shows the load-movement curve from a bidirectional test on a 1.85 m diameter,
65 m long bored pile, for which accidental hydraulic leak necessitated an unloading and
reloading cycle (Thurber Engineering Inc., Edmonton; personal communication 2016). Figure
7B shows the incremental stiffness curves from the initial and reloading records.
(GN)
-60
2L
70 y = -0.0215x + 81.827
R² = 0.7967
-80
60
-100 50
0 10,000 20,000 30,000 40,000 50,000 60,000 0 100 200 300 400 500 600 700
A
LOAD (kN) B STRAIN (µϵ)
It is clear from the examples illustrated in Figures 5 and 7 that using the load and strain
records to determine the pile stiffness requires that the data must not be affected by unloading-
reloading cycles. Many practitioners incorporate intentional unloading-reloading cycles in a
static loading test and they frequently keep the applied load constant for a longer duration at one
load level or other. I have many times tried to have users of such actions explain to me what they
expect to gain from this extraneous imposition on the test procedure, but nobody had ever been
able to tell me anything else than "this is what we always do" or "this is what I think the code
says I must do". The fact is that nothing is gained by this and such deviation from the simple
direct incremental procedure will instead result in that the investment in the instrumentation is
wasted.
Note, translating strain measurements to load requires that the measured strains cover an
acceptable range. There is little sense in investing in instrumentation if the induced strain at the
maximum load are smaller than 200 µϵ. Designing the pile and test toward achieving strains in
excess of 500 µϵ is preferable. If a calculation check shows that the strain are small, and it is
important to determine the load distribution, it is much better to do the test on a smaller diameter
pile and cautiously "extrapolate" the result to the larger diameter pile. For example, a pile with
half the diameter will show four times larger strain for the same load. In the process, it might be
realized that the original pile design is too conservative and a smaller diameter pile will suffice.
Of course, the transfer of the analysis results for the smaller diameter pile to the larger diameter
pile must be carried out with care and with some conservatism.
Furthermore, to obtain records that can be used for determining the pile axial stiffness and
load distribution, requires that a static loading test, be it a head-down test or a bidirectional test,
be carried out by applying equal increments of load and time. At each level, the load must be
maintained (held constant) for an equal length of time. The load-holding time can be short or
long, but an interval shorter than 10 minutes or longer than 20 minutes is impractical. The
frequently applied 5-minute load-holding interval is not suitable when testing piles with strain-
gage or other instrumentation down the pile. The reason is that it takes a few minutes after
adding an increment for the pile to react, i.e., for the gages down the pile to register the load
change. Therefore, the applied load and the strain-gage records are best combined after at least
ten minutes of constant load at the pile head (or bidirectional cell). And, such "waits" must be the
same for all load levels.
4 SGL 3
DEPTH (m)
8 SGL 2
10
12
SGL 1
14
0
Toe Movement
10 PILE TOE:
Measured movement
(mm)
LOAD (kN)
Figure 8 Load distribution determined from strain gage records.
10 10
Af ter Con- Soil Settlement
DEPTH (m)
struction
15 Clay 15
Qn Neutral Plane
20 20
25 25
Till
Rt
30 30
0 0
LOAD (kN)
2,000 2,000
3,000 3,000
4,000 4,000
0 1,000 2,000 3,000 4,000 5,000 0 50 100
LOAD (kN) PILE TOE PENETRATION (mm)
5. CONCLUSIONS
The scatter of capacity values resulting from the various definitions in vogue in the profession
and the effect of residual force on the evaluation of capacity are addressed. It is shown that
including an unloading-reloading event in a static loading test will adversely affect the
calculation of the pile axial stiffness. The analysis of the strain-gage records aim to determine the
load distribution and pile toe-load-movement, which, when combined with the soil settlement in
an interactive analysis, will show the settlement of the piled foundation. Designing for settlement
makes the conventional "capacity approach" redundant and provides a more reliable design.
6. REFERENCES
Davisson, M.T., 1972. High capacity piles. Proc. of Lecture Series on Innovations in Foundation
Construction, ASCE Illinois Section, Chicago, March 22, pp. 81-112.
Fellenius, B.H., 1975. Test loading of piles—Methods, interpretation, and proof testing. ASCE
Journal of the Geotechnical Engineering Division 101(GT9) 855-869.
Fellenius, B.H. and Atlaee, A., 1995. The critical depth – How it came into being and why it
does not exist. Proceedings of the Institution of Civil Engineers, Geotechnical Engineering
Journal, London, 113(2) 107 111.
Fellenius, B.H., 2004. Unified design of piled foundations with emphasis on settlement analysis.
Honoring George G. Goble—Current Practice and Future Trends in Deep Foundations. Geo-
Institute Geo-TRANS Conference, Los Angeles, July 27-30, 2004, Edited by J.A. DiMaggio
and M.H. Hussein. ASCE Geotechnical Special Publication, GSP125, pp. 253-275.
Fellenius, B.H., 1989. Tangent modulus of piles determined from strain data. ASCE,
Geotechnical Engineering Division, the 1989 Foundation Congress, F.H. Kulhawy, Editor,
Vol. 1, pp. 500-510.
Fellenius, B.H. 2012. Critical assessment of pile modulus determination methods. Discussion.
Canadian Geotechnical Journal, 49(5) 614-621.
Fellenius, B.H., 2015. Field Test and Predictions. Segundo Congreso Internacional de
Fundaciones Profundas de Bolivia, Santa Cruz May 12-15, Lecture, 22 p.
Fellenius, B.H., 2016a. The unified design of piled foundations. The Sven Hansbo Lecture.
Geotechnics for Sustainable Infrastructure Development – Geotec Hanoi 2016, edited by
Phung Duc Long, Hanoi, November 23-25, pp. 1-26.
Fellenius, B.H., 2016b. Fallacies in piled foundation design. Geotechnics for Sustainable
Infrastructure Development – Geotec Hanoi 2016, edited by Phung Duc Long, Hanoi,
November 23-25, pp. 41-46.
Fellenius, B.H., 2017. Basics of foundation design—a textbook. Electronic Edition,
www.Fellenius.net, 451 p.
Gerressen, F-W.(1)
(1)
BAUER Maschinen GmbH, Schrobenhausen, Germany <Franz-Werner.Gerressen@bauer.de>
ABSTRACT. A retaining wall is a structure designed and constructed to resist the lateral
pressure of soil when there is a desired change in ground elevation that exceeds the angle of
repose of the soil. The walls must resist the lateral pressures generated by loose/soft soils or, in
the case of existing ground water, water pressures as well. Very typical is the use not only for the
stabilization of slopes, but of course for the use of excavation pits. The paper gives a brief
overview of existing methods, but will then focus on recent developments which improve the
opportunities of using the retaining wall system.
1. INTRODUCTION
Retaining walls are very important structures in the field of special foundation business. Not only
for the use of stabilization of slopes, but mainly as supporting structure for excavation pits.
Using retaining walls for excavation pits safes a lot of space and allows the installation of
excavation pits also close to adjacent structures and under the ground water table. There are
several different systems of retaining wall types existing as shown in figure 1.
All of these systems have advantages and limitation as described above in terms of
deflections, water tightness and durability, but this is at that stage only the comparison in terms
of their functionality. Obviously, there are advantages and limitations for the various systems in
2. RECENT DEVELOPMENTS
2.1 Secant pile walls using Cased Continuous Flight Auger piles (CCFA)
2.1.1 Construction principle
The fundamental principle of double rotary drive systems is a continuous flight auger in
combination with an outer casing, that are drilled simultaneously, but in the opposite direction,
into the ground. These characteristics resulted to the common term Cased Continuous Flight
Auger (CCFA) pile.
2.1.2 Quality control system of CCFA method for secant pile walls
The CCFA method is mainly used for the installation of secant pile walls. The order of
installation reduces many of the sequence issues experienced with traditional methods. The
quality standard for pile walls is necessarily very high as the costs in case of repairs would be
significant and repair work time consuming. It is therefore essential to employ an accurate
quality surveillance system.
As primary requirements, a stable working platform with less than 3% inclination and a
guide wall as shown in figure 3 are highly recommended.
The guide wall ensures the correct starting location of every pile and facilitates the set-up of
auger and casing. The verticality of the casing is checked manually with a water level. However,
the verticality of the mast is both measured and recorded by electronic sensors. Verticality
corrections are easily made in the x & y directions either manually or automatically, ensuring
that the mast and tools are kept vertical at all times. For quality control all drilling data and data
relative to the verticality of the mast are available in real time.
Many factors contribute to the efficient drilling and subsequent concreting of CCFA piles. To
assist and allow the operator to monitor all the various parameters, drill and extraction assistant
programs have been developed. During the drilling phase the optimal performance is affected by
the applied torque and crowd forces and rotation speed. These will need to be varied for each
particular soil conditions encountered. In addition to these rig-related factors the geometry of the
The verticality of the pile depends on an accurate alignment of the casing, the casing guide and
the stiffness of the drilling tools. The relative flexural rigidity of a casing is min. 100 times
higher than the rigidity of a CFA auger. The opposed drilling direction of casing and auger
increases the stability additionally as any deflection is compensated and the drill string is
stabilized. The high grade of verticality and pile quality could be seen impressively on a jobsite
where standard CFA and CCFA piles were used next to each other:
During the execution of the pile wall by the CFA method, the operator noticed already a
deflection in the verticality of the piles. Especially when drilling a secondary pile, the auger
drifted out of direction. Therefore the piling method was changed to CCFA. The result was an
imposing improvement of pile quality in the same soil conditions, which could be seen after the
excavation of the pit. With the CCFA method a 1 in 200 verticality tolerance could be achieved,
whereas the secondary CFA piles show significantly higher deviations.
The major application for the CCFA system is the construction of secant pile walls. In
comparison to traditional systems like cased bored piles using Kelly drilling or standard CFA
piling, the CCFA system has advantages in costs and time against the cased bore piles and in
quality against the standard CFA system. Therefore, the CCFA system can be seen as a
significant improvement for the installation of secant pile walls.
The Cutter Soil Mixing (CSM) System differs essentially from traditional soil mixing techniques
in that the in situ mixing of the existing soil with self-hardening slurry is performed by mixing
tools rotating around horizontal axis rather than the traditional vertical axis. The in-situ soil
mixed with self-hardening slurry produces a rectangular shaped panel element, which e.g. takes
on the role of a cut-off and/or structural retaining wall.
The cement and bentonite content and the water/cement ratios of the mixing slurry are
determined based on the strength and/or permeability requirements of the project and the
properties of the soil being mixed. In general, for a stronger wall, cement content is increased
and water/cement ratio is lowered. Typically, sandy soils will require a larger amount of
bentonite in the slurry than clays. At some clay sites where enhanced resistance to permeability
is not required, acceptable liquefaction of the soil can be achieved without the use of bentonite.
A typical construction sequence as shown in Fig. 6 consists of the following steps:
One-phase system
During the penetration (down-stroke) phase, cutting, mixing, liquefying and homogenising is
performed while pumping the binder slurry into the soil. Adding compressed air is recommended
for assisting the down-stroke phase. As a rule of thumb about 70 % of the total slurry volume is
pumped during this phase. The backflow of soil and binder slurry is collected in the pre-
excavated trench or stored in a settling pond to be removed later off the site. After reaching the
design depth air flow is stopped. In the upstroke phase the remaining volume of binder slurry is
blended into the soil. The speed of extraction can be high as the majority of the binder slurry has
already been mixed with the soil in the down-stroke phase.
Two-phase system
The soil is liquefied and homogenized in the down-stroke phase by pumping of bentonite slurry
into the soil. The mixing process can be supported by adding compressed air. The backflow of
soil and bentonite can be pumped to a de-sanding plant where the sand is separated from the
slurry which is then pumped back to slurry storage area. When the backflow becomes too heavy
for pumping, it can be removed by a backhoe or a scratching belt. A scratching belt is a belt
which is placed on top of the trench, immerging into the trench and conveying the bentonite/soil
mixture automatically up-wards into a dewatering screen Using a hose pump the fluid fraction is
then pumped to the de-sanding plant unit for further treatment. After reaching the design depth,
the flow of bentonite is stopped and replaced by cement slurry. On the upstroke movement
cement slurry is mixed thoroughly with the liquefied soil. The speed of extraction and flow of
binder are adjusted to ensure that the total calculated quantity of binder is blended with the soil.
The most important elements of the CSM unit are the cutter gear drives. They are driven by
hydraulic motors which are located in a water-tight housing. The slurry is introduced into the soil
directly between the mixing wheels. During construction, the counter-rotating mixing wheels and
vertically mounted plates are effectively acting like a forced action mixer. Because of this
mixing principle, CSM is suitable for mixing cohesive soils. For loosening and mixing the soil
different types of mixing wheels were developed.
The size of individual panels is determined by the type and size of equipment being
deployed. Panels can be constructed in lengths ranging from 2.4 m to 2.8 m and wall thicknesses
of 0.55 m to 1.2 m. The mixing unit (Fig. 6) is either mounted on a guided Kelly bar or on a wire
rope-suspended cutter frame equipped with special steering devices. The standard set-up is the
"Kelly-guided" setup, capable of reaching depths down to 53 m (Fig. 7). "Rope-suspended"
systems are particularly suited for construction of deep soil mix walls. The greatest depth to date
at which a wire rope-suspended unit has successfully installed a soil mix wall is 80 m, using a
compact unit (Fig. 7).
The CSM process has significant advantages over conventional techniques, e.g. secant pile walls
or sheet piling walls. These are:
• The existing soil is utilised as construction material.
• Very little spoil material is generated; this renders the technique particularly suited for
work on contaminated sites.
• CSM is an ideal alternative to the traditional system, a soldier beam wall with timber
lagging ("Berlin wall") for the use in high groundwater conditions, or to sheet pile
walls in soil formations unsuitable for pile driving or in close proximity to vibration-
sensitive buildings.
• CSM is a vibration free method.
• No delivery of ready mixed concrete is necessary.
In comparison to traditional deep soil mixing methods CSM has the following advantages:
• A high degree of verticality of wall panels is achieved by the counter-rotating cutter
wheels.
• The cutter principle ensures construction of clean and trouble-free joints even between
wall panels of different construction age e.g. after weekend breaks or prolonged
stoppages on site.
• Harder soil formations can be easily penetrated, broken down and mixed by using the
cutter wheels as cutting and mixing tool.
• Homogenises the cohesive soils and self-harden slurry through horizontal mixing.
• In relation to small base units, high daily output and high panel depth may be achieved.
2.2.3 Applications
There are many possible applications for the CSM method. The main applications are:
• Cut-off walls
• Retaining walls (often with cut-off wall function)
• Foundations
• Slope stabilization
• Soil improvement / soil stabilization
• Liquefaction mitigation
As this paper focuses on the application of retaining walls, the following references will give
an overview of the use of CSM. The system is commonly used to construct water-tight soldier
beam retaining wall system for excavation pits especially near sensitive inner-city buildings
where vibration-free systems are essential. CSM provides an alternative solution to conventional
methods such as secant pile walls and/or sheet piles.
An example for an excavation pit shows a jobsite executed in Sydney, Australia. For the new
construction of some residential buildings a retaining wall was executed with CSM method. The
first project stage consists of a 13-storey building with one additional underground car park
floor; see Fig. 9.
The soil mix wall was used as retaining wall during construction of the building as well as
permanent external wall for the car park floors. The mixing depth was between 10 and 16 m and
the wall thickness 55 cm. The total wall area for this building was about 4,000 m². After
finishing the soil mixing wall the pit was excavated incrementally to a depth of about 6 m. To a
depth of 10.5 m, the wall acts as a retaining wall, below 10.5 m as a cut-off wall. The required
28-day unconfined compressive strength (UCS) was 5 MPa for the retaining wall function and 3
MPa for the cut-off wall function. The requirements on the water tightness of the wall were
relatively high. The client insisted that no leaks are visible on the wall after excavation. The
permeability of the wall was allowed to be maximum 10-8 m/s after 28 days.
One main reason for the decision to use the CSM method on this site was that it was
possible to produce soil mixing walls with a relatively smooth surface. Hence it is possible to use
the produced walls as permanent face concrete walls without costly wall surface cutting.
Soil conditions
The subsoil investigation indicated a subsurface profile generally comprising filling to depths of
0.6 m to 1.4 m over quaternary alluvial deposits, which typically comprise fine to medium
grained "marine" sands with podsols natural sand, clayey sand, sandy clay. The permeability of
the natural sands was between 2x10-4 and 5x10-5 m/s. These sediments overlie the Hawkesbury
Sandstone, the surface of which was found in depth between 10 and 19 m. The rock surface was
quite uneven. The determined rock strength varied between extremely low (UCS < 0.6 MPa) and
medium strong (UCS 6 < 20 MPa). For sealing the excavation pit it was necessary to socket the
wall a minimum of 1 m into the sandstone or 2 m into the clay layer.
The Hawkesbury Sandstone, which is also known as Sydney sandstone or Yellowblock,
forms the bedrock for the most areas of the region of Sydney. This sandstone is well-known for
its durability, which is caused by a high to very high quartz content. The quartz content varies
depending on the rock layer and can be up to 95 percent. Two exemplary borehole logs are
shown in Fig. 10.
During the first stage of this project about 4,000 m² were constructed. Because of the mixing
depth the 1-Phase method was used. The soil mix wall was executed by using the “hard-into-
hard” sequence with an overcut of minimum 20 cm.
The CSM wall was installed using a BCM 5 mixing head mounted on a RTG RG 19 T base
carrier equipped with a round shaped Kelly bar, see Fig 11.
Fig. 11. RTG RG 19 T and excavator for pre-excavation and backflow handling.
This tooth has carbide cutting edges on both sides. It can cut the sandstone much more
aggressively and allows a change of direction of mixing wheel rotation without increased tooth
wear. By using the BAUER SB 38 HR DC teeth the penetration speed in the sand could be
improved by about 28 % and the penetration speed into the sandstone by about 143 %.
ø 7 cm/min
Sandstone 10.2 m
Sand;
ø 22 cm/min partly silty,
medium
ø 80 cm/min dense –
extremely
dense,
partly
cemented
0.0 m
Fig. 13. Exemplary depth-time-diagram with average mixing speeds and the associated soil
profile for a panel with 1 m socketing into rock.
The CSM technique offers a great diversity of possible applications, such as cut-off walls,
structural retaining walls, foundation elements and numerous others. The capacity to reach big
depths offers an enormous potential for the construction of deep walls. The described example
demonstrated that this construction technique, which combines the advantages of the cutter and
the soil mixing technique, can be used in soil conditions where common mixing methods are at
its limit. Even in these soil conditions the CSM elements showed a very high quality. Thus, the
CSM method is a useful supplement to traditional soil mixing methods and an interesting
alternative for the use as retaining wall.
3. SUMMARY
Looking to recent developments it can be seen that there is an ongoing process to move the
borderlines for the use of retaining wall systems, e.g. in terms of cost efficiency, execution
quality, depth and soil limitations. Obviously, there are also limitations for these new systems,
but they are offering advantages and therefore alternative solutions resulting in time and cost
saving. Obviously, still all factors must be taken into consideration to achieve the best solution in
terms of feasibility and cost efficiency, but the new developments provide more alternatives in
the future.
ABSTRACT. Helical piles and anchors are factory manufactured deep foundation and anchor
elements that are provided as segmental modular units and assembled and installed in the field.
The shape and size of the central shaft, the number, size and spacing of the helical plates vary
depending on the soil and load requirements of the project. Even though helical piles and anchors
were introduced over 170 years ago into the civil engineering profession, they have only been
used in the modern era for about the last twenty years to any significant degree. Currently, the
application and use of this technology is one of the fastest growing markets in deep foundation
and anchor work in the world. A summary of the recent technological developments related to
helical piles and helical anchors is presented.
1. INTRODUCTION
The 2015 International Building Code, Section 1802.1, defines a Helical Pile as:
This technology is not new. In the mid 1800’s, helical piles (“screw-piles”) were introduced by a
blind, self-taught Irish engineer, Alexander Mitchell, as an effective method for mooring ships
and supporting iron lighthouses in shallow offshore environments (Mitchell 1849). As a result of
these successes, the use of helical foundations and anchors grew rapidly and were used
throughout the world on large-scale construction, especially to support oceanfront shipping and
pleasure piers and railway bridges (Lutenegger 2011a). Most consisted of a single helical plate
with a solid round shaft or were fabricated as a “screw-cylinder” with a shaft diameter of 2 to 4
ft. in diameter. A decline in the use of helical piles occurred around the beginning of the 20th
century coinciding with the development of other deep foundation methods and installation
equipment, especially pipe driving equipment.
During the 2nd World War and shortly after, helical piles and screw-cylinders were
“rediscovered” as a viable deep foundation technology for the rapid reconstruction of large
wharves damaged or destroyed during the war. In the 1950’s, the expansion of rural electrical
transmission lines in the U.S. resulted in extensive used of helical anchors to support
transmission towers and utility poles from overturning. The current use of helical piles and
anchors now has impacted nearly every segment of civil construction around the world. As with
all foundations, the design of helical piles and anchors must be based on sound principles of soil
mechanics and field observations and to some degree the experience of the Engineer with local
ground conditions. In modern times, helical piles and anchors have been “reinvented”; steel
Most helical piles consist of a lead section with helical plates and extension sections that
are usually just additional lengths of shaft to allow the installation to extend to any desired depth.
The length of lead sections and extension sections vary depending on the needs of the project and
the availability of the individual components. Lead sections with lengths from 5 to 20 ft. (1.5 to
6.1 m) are typical. Like other deep foundation systems, the behavior of the foundation depends
on the quality of the installation provided by the Contractor. Helical piles and anchors are a
manufactured modular foundation/anchor system with predetermined dimensions produced at the
factory.
Traditionally, helical piles have been installed using conventional construction equipment, such
as track excavators and skid steers, equipped with a high torque, low speed hydraulic head. In
recent years, there has been a growing trend to use the same equipment but fitted with a fixed
mast to be able to more carefully control the installation, maintain alignment, and reduce wobble
of the shaft. This greatly improves the quality of the installation and provides better contact
between the shaft and soil for round shaft helical piles.
Simultaneously, there is movement to provide continuous installation monitoring of
helical piles, recording the depth, torque, rotation speed, crown force, and rate of advance. This
is a natural development in line with the technology currently used for monitoring ACIP and
ACIPD piles. Both of these changes have the potential to make significant improvements in the
use of helical piles and provide Engineers with additional confidence and quality control on the
installation and behavior.
The installation of nearly all deep foundations produces some level of soil disturbance. The
degree of disturbance and influence on behavior depends on the type and geometry of the pile,
the method of installation and the soil conditions. Most manufacturer’s literature suggests that
helical piles produce little disturbance and generally does not account for disturbance in design
procedures. The degree of disturbance to the soil during installation may be especially important
for saturated fine-grained soils where the disturbance from installation may produce a reduction
of the undrained shear strength.
Lutenegger et al. (2014) presented direct evidence that the installation of Helical Piles in
clay produces disturbance to the clay and a reduction in shear strength. Field vane tests
conducted directly over the top of the path of the helical plate showed a reduction in undrained
shear strength as compared to the undisturbed soil. Moreover, the results showed that additional
helices along the shaft produced additional soil disturbance over a single-helix pile. The
measured undrained shear strength after installation was not the fully remolded value but was
lower than the peak value, closer to post-peak strength. Additional results also demonstrated that
poor installation by the Contractor could produce more disturbance than an ideal or “perfect”
installation in which the advance of the pile was carefully controlled according to the pitch of the
plate.
Figure 4 shows results of field vane tests were conducted over the top of two round shaft
(RS) open end Helical Piles and a square shaft (SS) Helical Pile with a single 14 in. (356 mm)
helical plate with a thickness on 3/8 in. (9.5 mm) and a pitch of 3 in. (76 mm) attached at the
end. The reduction in undrained shear strength is not uniform but on average is about 75% of the
undisturbed values. It can also be seen that there was a large reduction in strength indicated for
the 2.875 in. (73 mm) round shaft pile between depths of about 5 to 7 ft. (1.5 to 2.1 m) with the
undrained shear strength approaching the remolded value. This would indicate excessive
disturbance and almost complete remolding of the soil.
Depth (ft.)
5
10
0 1000 2000 3000 4000 5000 6000
Undrained Shear Strength (psf)
Fig. 4. Results of field vane tests over single-helix round shaft and square shaft helical piles.
(Note 1000 psf = 48 kPa)
0 0
RS2875
1 RS450 1
SS5
2 2
3 3
RS2875
4 4 RS450
Depth (ft.)
Depth (ft.)
SS5
5 5
6 6
7 7
8 8
9 9
10 10
0 500 1000 1500 2000 2500 3000 3 4 5 6 7 8 9
Torque (ft.-lbs.) No. of Revolutions/ft.
Fig.5. Installation torque and advance for single-helix round shaft and square shaft helical piles.
(Note: 1000 ft.-lbs. = 1356 N-m)
6
Depth (ft.)
10
11
12 FV1
FV2
13 SS5-12
SS5-12/12
14
SS5-12/12/12
15
0 1000 2000 3000 4000 5000 6000
Undrained Shear Strength (psf)
0 0
RS2875 SCG
1 RS2875 P
1
2 2
3 3
RS2875 SCG
4 4
Depth (ft.)
RS2875 P
Depth (ft.)
5 5
6 6
7 7
8 8
9 9
10 10
0 500 1000 1500 2000 2500 3000 3 4 5 6 7 8 9
Fig. 7. Installation torque and advance for two 2.875 in. (73 mm) round shaft helical piles.
(Note: 1000 ft.-lbs. = 1356 N-m)
Results of tension tests performed on these two piles are shown in Figure 8 and demonstrate the
influence of disturbance on capacity. The additional disturbance produces a lower stiffness and
lower capacity.
20000
Load (lbs.)
15000
10000
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Displacement (in.)
Fig. 8. Load-displacement curves between two 2.875 in. (73 mm) round shaft helical piles.
(Note: 1000 lbs = 4.5 kN; 1 in. = 25.4 mm)
In 1950 Skempton presented a valuable discussion to the paper of Wilson (1950) on the
behavior of screw-piles in clays (Skempton 1950) noting that for multi-helix screw-piles it was
important to recognize that the clay beneath the upper screws had been remolded by the passage
of the first screw. However, Skempton (1950) further noted that all of the clay contributing to the
bearing capacity of the upper screws would not be fully remolded and, as a rough approximation,
suggested that it might be reasonable to assume that the average shear strength of the clay would
be equal to:
where:
This observation and suggestion by Skempton brings into importance the Sensitivity of the clay
in considering installation disturbance.
Lutenegger et al. (2014) defined the Installation Disturbance Factor, IDF, as the ratio of
actual measured installation to the ideal or “perfect” installation:
where:
3
RS2875 SCG
4 RS2875 P
Depth (ft.)
10
0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Installation Disturbance Factor
Fig. 9. Installation Disturbance Factors for two round shaft Helical Piles.
Most manufacturer’s literature suggests that the capacity of multi-helix piles/anchors may be
obtained as a direct multiplier of the capacity of a single-helix pile/anchor. Detailed test results in
both clay and sands now show that this is not true, That is, the Efficiency of co-linear multi-helix
piles/anchors is less than 100%, which is likely related to disturbance. Efficiency, E, of a multi-
helix pile/anchor may be defined as:
Indirect evidence of Efficiency can be seen by comparing the installation torque and the
load-displacement behavior of single-helix and multi-helix piles. The attachment of additional
helical plates behind a lead plate is used to increase the load capacity of the pile/anchor, however
most tests show that this increase is not directly proportional to the number of helices or the
increase plate area. That is, a double-helix pile does not provide twice the capacity of a single-
helix anchor with the same diameter helical plate nor does a triple-helix anchor provide three
times the capacity of a single helix anchor. Figure 10 shows a comparison of the installation
torque measured for single and multi-helix piles. One might expect that in a uniform soil the
torque may be directly proportional to the number of helices for a central shaft with a fixed
diameter. In this case it can be seen that the torque does not increase proportionally with the
increased number of helices and suggests Efficiency less than 100%.
10
Depth (ft.)
12
14
16
RS2875-12
18 RS2875-12/12
RS2875-12/12/12
20
22
24
0 1000 2000 3000 4000
Torque (ft.-lbs.)
Fig. 10. Installation torque of single-, double-, and triple-helix piles in clay.
(Note: 1000 ft.-lbs. = 1356 N-m)
So it might be argued that the degree of disturbance increases as each successive plate passes
through the same soil as previously indicated in Figure 3. That is, the capacity of each successive
plate is reduced progressively by the preceding disturbance so that while the 1st plate may have
100% Efficiency, the second may have only 60 to 80% Efficiency and the 3rd 40 to 60%
efficiency, etc. depending on the soil. Figure 11 presents results from five different sets of tests
100
SS5-12
90 SS5-12
RS2875-10
80 RS2875-12
RS350-12
Trend
Efficiency (%)
70
60
50
40
30
20
1 2 3 4
Number of Helices
For the last 10 years there has been a growing trend for using larger diameter (6 in. to 16 in. (152
to 406 mm) diameter) round shaft helical piles to support higher loads, such as from bridges.
These piles typically have either one or two large diameter helical plates, about 16 in. to 36 in.
(406 mm to 914 mm) in diameter. The behavior of these piles may be very different than the
behavior of smaller helical piles and, therefore, the design procedure must accordingly be
different. The pipe shaft is usually open-ended with a wall thickness and strength of steel
sufficiently large to withstand the torsional forces applied during installation. So, in some
respects, a round-shaft helical pile is simply a pipe pile with a helical plate attached to the base
that is installed by rotation rather than driving.
Depending on the length of the shaft and the type of connection between additional pipe
sections if any, the contribution of capacity between the shaft and helical plates may vary. In
some cases, it is possible that the helical plates only act as a construction technique to install
what is otherwise a traditional pipe pile, i.e., installation by rotation rather than installation by
driving or vibration. The installation progresses because the pitch of the helical plate advances
during rotation and “pulls” the pipe pile downward as it advances. When the shaft resistance
developed along the outside of the pipe exceeds the capacity of the helical plate, advance will
stop. In some cases, multiple helices are used to help advance the pile to deeper depths to
develop higher capacity. Because of the similarities, the design approach for helical pipe piles
should be similar to the traditional design approaches currently being used for driven piles in
either sand or clay.
30000
25000
Uplift Load (lbs)
20000
15000
10000
Fig. 12. Results of tension tests on helical piles with different pipe sizes in sand.
(Note: 1000 lbs = 4.5 kN; 1 in. = 25.4 mm)
15000
10000
5000
0
0.0 0.5 1.0 1.5 2.0 2.5
Displacement (in.)
Fig. 13. Comparison between uplift behavior of plain pipe pile and helical pile in sand.
(Note: 1000 lbs = 4.5 kN; 1 in. = 25.4 mm)
If the load distribution is evaluated at an even lower load level, corresponding to a more
realistic working or serviceability load condition, it might be convenient to apply a global Factor
of Safety of 2 to the 10% load, in this case giving a safe working load of 8,675 lbs. (35.6 kN) and
the distribution of load between the shaft and helix now becomes 3,300 lbs. (14.7 kN) carried by
the pipe shaft and 5,375 lbs. (23.9 kN) carried by the helix; i.e., 38% shaft, 62% helix. As the
trend toward the use of larger diameter pipes continues, a smaller portion of the load is
transferred to the helix.
Figure 14 shows the results of three helical piles tests conducted on different diameter
round shaft piles with the same diameter helical plate (12 in. (305 mm)) installed in clay. These
results again clearly show that capacity increases as pipe diameter increases and illustrate the
importance of the pipe shaft to overall behavior. Results may also be expressed in terms of the
ratio of helix diameter to pipe diameter, DH/DP to illustrate this behavior.
Figure 15 shows a similar comparison to Figure 4 between a helical pile and plain pipe
pile in stiff clay. The results obtained for both sands and clays are similar. For the same
geometry piles installed to the same depth in different soils, the distribution of load varies
considerably and in this case the shaft diameter is more important in clay than in sand. Both the
pipe shaft and the helical plate contribute to capacity but in different ways.
There is a long history in the profession of using the installation torque to estimate the axial
capacity of Helical Piles and Anchors, beginning as early as the late 1800s, through a Torque
Factor, KT. The basic premise behind such correlations is that both the axial capacity, Qult, and
installation torque, T, are a function of the specific geometry of the pile/anchor and soil
properties, i.e., soil strength and therefore:
28000
24000
20000
Uplift Load (lbs.)
16000
12000
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Displacment (in.)
Fig. 14. Tension tests on helical piles with different pipe sizes in clay.
(Note: 1000 lbs = 4.5 kN; 1 in. = 25.4 mm)
20000
16000
Uplift Load (lbs.)
12000
8000
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Displacement (in.)
Fig. 15. Comparison between uplift behavior of plain pipe pile and helical pile in clay.
(Note: 1000 lbs = 4.5 kN; 1 in. = 25.4 mm)
Manufacturers of helical piles and anchors often suggest different values of KT for
different size round shaft elements. A simple analysis of the undrained behavior of plain pipe
piles installed by rotation in uniform clay shows that the value of KT for a plain pipe is only a
function of pipe diameter. Naturally, the addition of a helical plate or multiple plates to the pipe
shaft and the direction of loading complicate this relationship. However, this approach simply
illustrates how one factor, shaft diameter, may affect a torque-to-capacity correlation. The actual
correlation will be very complex since both shaft diameter and helix diameter may vary and
different combinations of shaft and helix diameter will affect the true distribution of components
in developing load capacity in different soils.
A further complication may be that installation torque, T, is measured during installation
while Qult is measured at some time after installation. This delay time allows the soil to adjust
(pore water pressure to dissipate and soil to regain strength) to the installation effects. Some
variation in Torque correlations may also develop if non-uniform soils are present and the shaft
is embedded in one soil layer while the helix is embedded in another type of soil. It is clearly
5. SUMMARY
Helical Piles offer the Geotechnical Engineer a number of potential advantages for projects,
including: rapid installation; immediate load capacity; installation in high groundwater
conditions; installation with traditional equipment; no soil cuttings; minimal site disruption and
cleanup; installation in limited access & low headroom areas; installation monitoring;
removable/reusable; installation at any orientation; minimal disruption to adjacent structures;
pre-screening of soil conditions at each location; minimal installation noise and vibration; wide
range of soil applicability and rapid field modifications.
6. REFERENCES
Clemence, S.P., Crouch, L.K. and Stephenson, R.W., 1994. Prediction of uplift capacity for
helical anchors in sand. Proceedings of the 2nd Geotechnical Engineering Conference, Cairo
University, 332-343.
Clemence, S.P. and Lutenegger, A.J., 2015. Industry survey of state-of-practice for helical piles
and tiebacks. Journal of the Deep Foundations Institute, 9(1): 21-41.
Hoyt, R. and Clemence, S.P., 1989. Uplift capacity of helical anchors in soil. Proceedings of the
12th International Conference on Soil Mechanics and Foundation Engineering, 2: 1019-1022.
Lutenegger, A.J., 2011a. Historical development of iron screw-pile foundations: 1836–1900.
International Journal for the History of Engineering and Technology, 81(1):108-128.
Lutenegger, A.J., 2011b. Behavior of grouted shaft helical anchors in clay. Journal of the Deep
Foundations Institute, 5(1): 63-74.
Lutenegger, A.J., 2011c. Behavior of multi-helix screw anchors in sand. Proceedings of the 14th
Pan-American Conference on Geotechnical Engineering, Toronto, Canada.
Lutenegger, A.J., 2013. Factors affecting torque correlations for screw-piles and helical anchors.
Proceedings of the 1st International Geotechnical Symposium on Helical Foundations, 211-
224.
Lutenegger, A.J., 2015. Load tests on grouted shaft helical micropiles in some U.K. soils.
Proceedings of the 16th European Conference on Geotechnical Engineering.
Lutenegger, A.J., Erikson, J. and Williams, N., 2014. Evaluating installation disturbance of
helical anchors in clay from field vane tests. Proceedings of the 39th Annual Deep
Foundations Institute Conference.
Mitchell, A., 1849. On submarine foundations: Particularly screw-pile and moorings. Civil
Engineers and Architects Journal, 12: 35-40.
Skempton, A.W., 1950. discussion of The bearing capacity of screw piles and screwcrete
cylinders. Journal of the Institution of Civil Engineers – London, 34(5): 76-81.
Tshua, C.H.C., Aoki, N., Rault, G., Thorel, L. and Garnier, J., 2012. Evaluation of the
efficiencies of helical anchor plates in sand by centrifuge model tests. Canadian Geotechnical
Journal, 49: 1102-1114.
Vickars, R. and Clemence, S.P., Vickars, R.A. & Clemence, S.P., 2000. Performance of helical
piles with grouted shafts. New Technological and Design Developments in Deep
Foundations, ASCE, 327-341.
Wilson, G., 1950. The bearing capacity of screw piles and screwcrete cylinders. Journal of the
Institution of Civil Engineers – London, 34(5): 4-73.
ABSTRACT. Current design of piled rafts often neglects a number of aspects well known in
scientific literature, sometimes leading to increased costs without appreciable increase in
performance. This work makes an attempt to partially fill this lack between State of Art and State
of Practice; to this end, the paper briefly recalls available simple design methods regarding both
geotechnical and seismic issues and proposes a novel simple procedure to estimate the load-
settlement curve of a piled raft as well as the load sharing between piles and raft for vertical
centered loads. The method has been validated through comparison with numerical and
experimental data and applied to a case history; it will be shown that the proposed procedure,
despite its simplicity, is able to account for the non-linearity in the soil behavior, the latter being
responsible of progressive variation of the load sharing with increasing applied top load. Further,
regarding seismic issues, simplified formulae from past works are recalled and discussed.
1. INTRODUCTION
From a conceptual standpoint, a piled raft consists of a group of piles surmounted by a raft in
direct contact with the ground. In order to describe the portion of the total load Qpr taken by the
raft (Fig. 1), it is possible to refer to the load sharing ratio αr defined as:
α r = Qr Qpr (1)
where Qr is the load taken by the raft. A load sharing ratio αr = 1 represents a shallow
foundation with no piles, while αr = 0 represents a pile group with a raft not in contact with
ground; piled raft foundations cover the range 0 < αr < 1.
When designing the structure foundations, engineers typically consider as a first option to
adopt a shallow foundation system such as a raft. If this choice does not guarantee a satisfactory
performance, the adoption of piles underneath the shallow foundation is certainly a common
design alternative. Although this rationale cannot be generalized as the design process is the
outcome of a number of aspects, some of them lying beyond the mere technical considerations, it
is indeed evident that piles are generally required to increase bearing capacity and/or reduce
settlement. However, once the decision of adopting piles has been made, it is common
assumption that the design loads are carried solely by the piles, thus neglecting the contribution
of the raft-soil contact. This traditional design approach is clearly conflicting with the collected
experimental evidence (see for example Mandolini et al. 2005). For pile groups with pile at
relatively small spacing/diameter ratio (say s/d = 3÷4) and covering the entire raft area AR
(AG/AR ∼ 1, with AG the area covered by pile group), the percentage of load taken by the raft is
Fig. 2. Percentage of load carried by the raft: experimental data (Mandolini et al., 2005).
The common simplification of neglecting the raft contribution, which is conservative when
piles are required to increase the foundation bearing capacity, becomes totally erroneous when
the raft alone possesses adequate bearing capacity, often larger than the one provided by the pile
group, and piles have the only role of decreasing and/or regulating settlements.
In the last decades, many specialists focused on this field, either from theoretical or
experimental point of view (for a more comprehensive coverage, reference may be made to
Randolph 1994; Poulos et al. 2001; Mandolini 2003). It can be certainly stated that is nowadays
10,0
CBD or CSBD SBD o DSBD
[2]
[3]
[4]
[5]
[-]
adm [-]
adm
1,0
/w
UR/w
unrealistic [1]
raft
0,1
0 1 2 3 4 5 6
FSUR
FS =Q
UR = Rr,u
UR//V
QPR [-]
pr [-]
Fig. 3. Chart for selection of the design approach (modified from Mandolini 2003).
⎡ 1 α pr ⎤
⎢ ⎥
⎡ wp ⎤ ⎢ K p K r ⎥ ⎡Q p ⎤
⎢ ⎥=⎢ ⋅⎢ ⎥ (2)
⎣ wr ⎦ ⎢ α rp 1 ⎥ ⎣Qr ⎦
⎥
⎢⎣ K p K r ⎥⎦
where Qi and Ki are load and stiffness of pile group (subscript “p”) and raft (subscript “r”)
considered as stand-alone systems, while wi is the settlement in the combined foundation; αpr
and αrp are interaction terms related to the influence of raft settlement on the settlement of the
pile group and vice versa, respectively.
By imposing the compatibility condition (wp = wr) and introducing the further hypothesis that
the terms on the secondary diagonal are equal (αrp/Kp = αpr/Kr), one obtains the expressions for
the stiffness of the piled raft and the load sharing:
Note that the last condition should be considered as a further hypothesis of the method, rather
than a consequence of reciprocity, which presents some difficulty to be applied with the aim of
proving the equality of the interaction terms. In addition, the validity of such hypothesis has been
proven numerically only for the capped pile (i.e. a circular raft equipped with an isolated pile) in
Randolph (1983). However, this hypothesis will be accepted also in the proposed method and its
validity for any piled raft will be not discussed further in this work.
Regarding the value of αrp, Clancy and Randolph (1993) performed a number of numerical
analyses, finding out that for square pile groups made by more than 16 piles (4x4 or bigger) at
moderate-to-large spacing (larger than 4 diameters) such a coefficient remains constant with a
value of about 0.8.
Equations (3) and (4) are valid until piles reach failure (Location A in Fig. 4). Beyond this
point and until raft failure, any load increment is transferred directly to the raft and therefore the
settlement is given by:
Q A Q pr - Q A
w= + (5)
K pr Kr
where Kp and Kr are the tangent stiffnesses of pile group and raft (with Kp0 and Kr0 their initial
value), Qp,u and Qr,u the ultimate loads, np and nr real positive numbers regulating the shape of
the curves. It is easy to verify that np [nr] = 0 corresponds to the assumption of elastic behavior
of the piles [raft] until failure.
Equations (6a,b) may be readily integrated to get the load displacement curves:
ni
⎛ Q ⎞
(Qi ,u − Qi ) − Qi ,u ⎜1 − Q i ⎟
wi = ⎝ i ,u ⎠
with i = p, r (7)
ni
⎛ Qi ⎞
K i ,0 ( ni − 1) ⎜1 − ⎟
⎝ Qi ,u ⎠
The above expression is valid for ni ≠ 1. For ni = 1 the expression of the load-settlement curve
is obtained by calculating the limit, so that:
Qi ,u ⎛ Q ⎞
wi = − ln ⎜1 − i ⎟ with i = p, r (8)
Ki ,0 ⎝ Qi ,u ⎠
Note that for ni < 1 the settlement for Qi approaching Qi,u is finite and equal to:
Qi ,u
wi = with i = p, r (9)
Ki ,0 (1 − ni )
Once the load-settlement curves have been defined for both the pile group and the raft, the
proposed method assumes that each load increment on the piled raft is shared between piles and
raft depending on their current stiffnesses. On the contrary, interaction terms always depend on
the initial stiffnesses. This assumption, verified numerically and experimentally (Caputo and
Viggiani 1984) for pile-to-pile interaction, will be adopted for the pile group-raft interaction.
In matrix form, the method assumes:
⎡ 1 α rp ⎤
⎢ ⎥
⎡ dwp ⎤ ⎢ K p K p ,0 ⎥ ⎡ dQ p ⎤
⎢ ⎥=⎢ ⋅⎢ ⎥ (10)
dw
⎣ r ⎦ ⎢ α rp 1 ⎥ ⎣ dQr ⎦
⎥
⎣⎢ K p ,0 K r ⎦⎥
K pr =
{
K p ,0 ⋅ K p ⋅ ⎡⎣ K p − ( 2 ⋅α rp ⋅ K r )⎤⎦ + ( K p ,0 ⋅ K r ) } (12)
K p2 ,0 − ⎡⎣α rp2 ⋅ ( K p ⋅ K r )⎤⎦
dQr K r ⋅ ⎡⎣ K p ,0 − ( α rp ⋅ K p )⎤⎦
β= = (13)
dQp K p ⋅ ⎡ K p ,0 − ( α rp ⋅ K r )⎤
⎣ ⎦
The method may be readily employed in a spreadsheet. Note that although Eqs. (6)-(9) are
considered in this study, the proposed procedure allows to employ any kind of load-settlement
curve, including a piecewise function.
Applying a constant load increment ΔQpr, at the load increment i the procedure involves the
following steps:
1) Calculation of the tangent stiffness Kpr,i of the piled raft and of βi as:
K pr ,i =
{ }
K p ,0 ⋅ K p ,i −1 ⋅ ⎣⎡ K p ,0 − ( 2 ⋅α rp ⋅ K r ,i −1 )⎦⎤ + ( K p ,0 ⋅ K r ,i −1 )
(14)
K 2 p ,0 − ⎡⎣α 2 rp ⋅ ( K p ,i −1 ⋅ K r ,i −1 )⎤⎦
K r ,i −1 ⋅ ⎡⎣ K p ,0 − ( α rp ⋅ K p ,i −1 )⎤⎦
βi = (15)
K p ,i −1 ⋅ ⎡⎣ K p ,0 − ( α rp ⋅ K r ,i −1 )⎤⎦
2) Evaluation of settlement increment Δwpr,i, current settlement wpr,i and load increment
percentages on pile group and raft as:
ΔQ pr
Δw pr ,i = (16)
K pr ,i
w pr ,i = w pr ,i −1 + Δw pr ,i (17)
3) Evaluation of load increments ΔQp,i, ΔQr,i and total loads Qp,i, Qr,i on pile group and raft as:
1 β
ΔQ p ,i = ⋅ ΔQ pr and ΔQr ,i = ⋅ ΔQ pr (19)
1+ βi 1 + βi
4) Calculation of the total load Qpr,i on piled raft and of the total load percentages αp,i αr,i
carried by pile group and raft as:
Q pr ,i = Q p ,i + Qr ,i (21)
Q p ,i Qr ,i
α p ,i = and α r ,i = (22)
Q pr ,i Q pr ,i
5) Adjournment of current raft and pile group stiffnesses through the functions selected to
describe the evolution of stiffness with load. If these functions obey to Eqs. (6a,b), new
values of piles and raft stiffness are:
np nr
⎛ Qp,i ⎞ ⎛ Qr ,i ⎞
K p,i = K p,0 ⎜1 − ⎟ and K r ,i = K r ,0 ⎜ 1 − ⎟ (23)
⎜ Q ⎟
⎝ p,u ⎠ ⎝ Qr ,u ⎠
The above steps are repeated for a number of load increments equal to Qpr,u/ΔQpr.
Fig. 6. Comparison between results from the proposed method and rigorous FEM analyses
provided by de Sanctis and Mandolini (2006).
Fig. 8. Comparison between results from the proposed method and rigorous FDM analyses
provided by Comodromos et al. (2009).
Fig. 10. Comparison with results from the proposed method and centrifuge test results by
Fioravante and Giretti (2010).
Regarding the pile, a value of shaft coefficient β = 0.25 has been assumed, while for base
resistance a value of Nq = 20 has been considered (Berezantzev, 1961).
To account for group effects the ultimate axial capacity of the pile group has been taken as the
sum of the individual contributions of single piles multiplied by an efficiency factor of 0.8,
following the indications by De Mello (1969). Bearing capacities of raft and pile group were
found to be 540 and 480 MN, respectively.
The initial stiffness of the raft has been calculated by the formula proposed by Carrier and
Christian (1973), employing a low-strain stiffness of the soil varying according to the law G0(z*)
= 3.8 + 1.5z* (with shear modulus in MPa), where z* = 4 m is the depth of the raft base from soil
surface. The resulting initial stiffness of the raft-soil system has been estimated equal to Kr0 =
550 MN/m, thus leading to an estimated settlement of the unpiled raft not lower than 20 cm. The
unpiled raft stiffness has been considered as constant with the load (that is, nr = 0) as for the case
at hand the raft-soil contact is expected to be linear at least until the working loads, due to the
expected low load percentage carried by the raft.
For sake of completeness, it is worth mentioning that the raft bearing capacity has been
reduced by 88% following the indications by de Sanctis and Mandolini (2006); however, for the
hypothesis of linear behavior of the raft-soil system, this assumption has no effect on the
behavior of the piled raft under working loads.
According to Mandolini and Viggiani (1997), pile group stiffness has been assessed by
summing the elastic contribution due to the pile-to-pile interaction to the nonlinear component of
the settlement of the single pile. Initial stiffness of single pile was estimated by the classical
Randolph and Wroth (1978) formula and was calculated as 200 MN/m, therefore obtaining Kp0 =
2387 MN/m (Butterfield and Douglas 1981). Adopting the suggestion from the numerical studies
above for fine-grained soils, a factor np = 1.5 has been utilized to describe the nonlinear behavior
of the single pile. Finally, a raft-pile group interaction factor αrp = 0.7 has been adopted,
according to the graphs shown by Clancy and Randolph (1993).
The comparison between the results obtained by the proposed procedure and the experimental
load-settlement curve is reported in Fig. 11. Looking at the final measured settlement (that is,
after consolidation), the agreement between experimental data and predicted settlement is very
good. In particular, under the working load of 113 MN the estimated settlement is 53 mm as
opposed to the long-term measured settlement of 52 mm. Interestingly, the method predicts that
piles adsorb 88% of the total working load, almost identical to the measured value of 87% after
10 years from construction (see Mandolini et al. 2005; Viggiani et al. 2012).
4. SEISMIC DESIGN
Modern seismic codes (e.g. Eurocode 8) require that piles shall be designed to withstand both
inertial forces (i.e. coming from the superstructure vibrations) and kinematic action (i.e. due to
the soil deformations arising from the passage of seismic waves). Due to seismic soil
deformations, even in an hypothetical absence of a superstructure, piles are subjected to bending
moments resulting from their deflection. Long fixed-head piles in homogeneous soil (i.e., with
stiffness constant with depth) are unable to offer significant opposition to the displacement that
soil tries to impose (at least at low frequencies, which are of major concern), so that pile
curvature at the top is equal to the curvature of soil at surface (Di Laora et al. 2013). Simplified
expressions for the assessment of kinematic bending are available for subsoil conditions other
than homogeneous. A generalized formulation for inhomogeneous soil has been recently
proposed by Di Laora and Rovithis (2015). The authors considered the problem of a flexible pile
embedded in a subsoil with stiffness varying continuously with depth according to the
generalized function (Fig. 12a):
n
⎡ z⎤
G s (z) = G sd ⎢a + (1 − a) ⎥ (24)
⎣ d⎦
where Gs is the shear modulus of the soil, Gsd being the corresponding value at the depth of
one diameter, z is the depth from ground surface while a and n are coefficients regulating the
distribution of stiffness with depth.
Fig. 12. Pile-head kinematic bending moment (Di Laora and Rovithis, 2015).
a s ρs
(1/ R )p ≅ (1/ R )s,eff = (25)
G s ( La / 2 )
where as and ρs are surface acceleration imposed by the earthquake and mass density of the
soil, supposed constant along depth, respectively. The authors also provide an iterative
expression to calculate pile active length in such inhomogeneous soil. Instead of the iterative
procedure, Karatzia and Mylonakis (2016) furnish the closed-form expression:
4
⎧ 1 4+ n ⎫
⎪ ⎡ ⎤ ⎪⎪
1 ⎪⎢ 4 n+4
5 ⎛ π Ep ⎞ ⎥4
La = d ⎨ a + ( n + 4 )(1 − a ) ⎜ ⎟ − a ⎬ (26)
1 − a ⎪⎢ 16 ⎝ 2 Esd ⎠ ⎥ ⎪
⎢ ⎦⎥
⎪⎩ ⎣ ⎪⎭
where Esd refers to soil Young’s modulus at the depth of one diameter. The bending moment at
head is then equal to:
πd4
M = Ep (1/ R )p (27)
64
Fig. 13. Reduction of structural acceleration due to pile-induced filtering effect (Rovithis et al.
2015).
6. REFERENCES
Berezantzev, V.G., Khristoforov, V., and Golubkov, V., 1961. Local bearing capacity and
deformation of piled foundations. Proc. of the V Int. Con. Soil Mech. Found. Eng., Paris,
July 17-22, Vol. 2. pp. 11-15.
Butterfield, R., and Douglas, R. A. 1981. Flexibility coefficients for the design of piles and pile
groups. Construction Industry Research and Information Association, Technical Note, 108.
Caputo, V., and Viggiani, C., 1984. Pile foundation analysis: a simple approach to nonlinearity
effects. Rivista Italiana di Geotecnica, 18(2), 32-51.
Carrier, W. D., and Christian, J. T., 1973. Rigid circular plate resting on a non-homogenous
elastic half-space. Geotechnique, 23(1), 67-84.
Chin, F. K., 1970. Estimation of the ultimate load of piles not carried to failure. In “Proceedings
of the 2nd Southeast Asian Conference on Soil Engineering”, pp. 81-90.
Clancy, P., and Randolph, M. F., 1993. An approximate analysis procedure for piled raft
foundations. International Journal for Numerical and Analytical Methods in Geomechanics,
17(12), 849-869.
Comodromos, E. M., Papadopoulou, M. C., and Rentzeperis, I. K., 2009. Pile foundation
analysis and design using experimental data and 3-D numerical analysis. Computers and
Geotechnics, 36(5), 819-836.
Davis, E.H., and Booker, J.R., 1971. The bearing capacity of strip footings from the standpoint
of plasticity theory. Proc. I Australian-New Zeland Con. in Geomech., Melbourne, pp. 276-
282.
De Mello, V.F.B., 1969. Foundations of buildings on clay. State of the Art Report, Proc. VII
Int. Conf. Soil Mech. Found. Eng., Mexico City, Vol. 1, pp. 49-136.
de Sanctis, L., and Mandolini, A., 2006. Bearing capacity of piled rafts on soft clay soils. Journal
of Geotechnical and Geoenvironmental Engineering, 132 (12), 1600-1610.
Di Laora, R., and de Sanctis, L., 2013. Piles-induced filtering effect on the foundation input
motion. Soil Dynamics and Earthquake Engineering, 46, 52-63.
Di Laora, R., Mylonakis, G., and Mandolini, A., 2013. Pile-head kinematic bending in layered
soil. Earthquake Engineering & Structural Dynamics, 42(3), 319-337.
Di Laora, R., and Rovithis, E., 2015. Kinematic Bending of Fixed-Head Piles in
Nonhomogeneous Soil. Journal of Geotechnical and Geoenvironmental Engineering, 141(4),
04014126.
Massarsch, K.R.(1)
(1)
Geo Risk & Vibration AB, Bromma, Sweden, <rainer.massarsch@georisk.se>
ABSTRACT. During the past two decades, the capacity and performance of hydraulic vibrators
has undergone a rapid development. By the use of vibrators with variable frequency and eccentric
moment, the transmission of the vibration energy to the surroundings can be controlled. Methods
and empirical rules for estimating the drivability of sheet piles are presented. The application of
modern vibrators is not limited to the driving of piles and sheet piles. The efficiency of vibratory
compaction can be enhanced by adapting the vibrator to the resonance frequency of the vibrator-
pile-soil system. The optimization of the compaction process by electronic process control is
highlighted and illustrated. A new concept is proposed where the dynamic penetration resistance
is correlated to the pile penetration speed, measured in terms of the number of vibration cycles
during the driving. The validity of the concept is demonstrated by a case history. Finally, examples
of pile and ground improvement solutions are presented which take advantage of the fundamental
aspects of vibratory driving.
1. INTRODUCTION
Modern, hydraulic vibrators can be used for different construction processes, such as the
installation and/or extraction of piles and sheet piles and compaction . However, in spite of their
frequent use in the foundation industry, many aspects of vibratory driving are still based on overly
simplistic empirical rules. There is a need for a better understanding of the parameters of
importance for vibratory driving of piles and sheet piles. When using vibrators for pile and sheet
pile driving or vibratory compaction of granular soils, the following issues need to be considered:
• drivability: select appropriate vibrator type and capacity to assure efficient installation
• bearing capacity: verify that the required bearing capacity can be achieved
• vibratory compaction: assure that the compaction requirements can be achieved
• environmental impact: minimize emission of ground vibrations and/or noise.
The four issues are closely interrelated and one cannot be addressed without considering the other
three.
2. VIBRATOR PERFORMANCE
2.1 Fundamentals
The performance of modern vibrators can be controlled and adapted for optimal performance.
Several parameters govern the vibratory driving process and it is important to realize how they can
be varied to achieve efficient driving or compaction. Vibratory excitation affects a pile in a
different way than does impact driving. An impact hammer imparts a short duration blow to the
pile head, sending a stress wave down the pile. The initial hammer energy is reduced in passing
through the pile helmet and pile cushion. Each blow must be able to activate the inertia of the pile
(1)
Many modern vibrators also enable the centrifugal force to be adjusted continuously during
operation, resulting in a variable displacement amplitude, s. This feature allows the machine
operator to start up and shut down the vibrator at zero vibration amplitude, thereby reducing the
risk of vibration amplification due to resonance effects. An additional important factor for the
vibrator performance is the displacement amplitude, which together with the centrifugal force, Fv,
determines the driving ability of the vibrator. For a vibrator suspended above the ground surface,
the axial displacement amplitude (single amplitude), s can be determined from the following
relationship.
𝑀𝑀𝑒𝑒
𝑠𝑠 = (2)
𝑀𝑀𝑑𝑑
The “total dynamic mass”, Md, is the sum of all masses, which need to be accelerated by the
vibrator. This includes, e.g., the mass of vibrator (but not the static mass separated by soft springs
from the vibrator), the pile, and the vibrator clamp. Note that most equipment manufacturers
express the displacement amplitude as peak-to-peak ("double") amplitude. From Eq. (2), it can be
appreciated that the free pile displacement amplitude, s, is independent of the vibration frequency,
f. In order to maximize the displacement amplitude, the total dynamic mass, Md, should be kept as
small as possible.
Before 1980, most hydraulic vibrators had fixed eccentric moment, with a typical operating
frequency between 20 and 30 Hz. These vibrators were used for basic construction works, such as
driving and extracting of sheet piles. Table 1 shows a typical range of performance parameters for
vibrators with fixed eccentric moment.
Eccentric mass
Space for
eccentric mass
In the 1990s, vibrators with variable frequency and variable amplitude were introduced with
resonance-free starting-up and shutting-down of vibratory driving. Such vibrators allow the
operating frequency and eccentric moment (and thus amplitude) to be varied according to specific
driving requirements and soil conditions. Typical performance parameters of vibrators with
variable frequency and variable eccentric moment (amplitude) are shown Table 3. It should be
mentioned that vibrator performance data of most modern manufacturers are similar to those
shown in Table 1 through 3, for instance vibrators manufactured in North America (APE, ICE) or
Europe (Dieseko, PTC, RTG).
Vibrator manufacturers have provided empirical guidelines for the selection of vibrators. In Figure
2, the centrifugal force required for driving sheet piles is shown as a function of pile mass and
depth of penetration.
Fig. 2. Empirical chart for estimation of required vibrator centrifugal force for driving of sheet
piles. The example (red lines): sheet pile mass: 3.6 ton maximum penetration depth: 20
m, driving into medium dense sand (III) requires vibrator with centrifugal force of at least
1,600 kN (based on Müller-ThyssenKrupp 2004).
Table 4. Drivability of sheet piles into granular soils, based on cone penetration tests, cf. Fig. 2.
I II III IV V
Test Type Very loose Loose Medium Dense Very Dense
Dense
SPT, N <4 4 - 10 10 - 30 30 - 35 > 50
(blows/0.3m)
CPT, qc <5 5 - 10 10 - 15 15 - 20 > 20
(MPa)
DPH1), N10 <5 5 - 10 10 - 15 15 - 20 > 20
(blows/10 cm)
1)
Heavy dynamic probing: mass 50 kg, height of fall 500 mm
Figure 2 and Table 4 are based on experience from sheet pile driving, compiled by one
particular vibrator manufacturer (Müller-ThyssenKrupp, 2004). Such guidelines must be treated
with caution. The guidelines assume that the user understands the limitations involved. In the
example shown in Fig. 2, a pile with mass 3.6 ton shall be driven to a depth of 20 m into medium
dense sand (III) (qc ≈12 MPa). According to Figure 2, a vibrator with a centrifugal force of about
1,600 kN will be required.
An important aspect of Figure 2 is that the diagram was based primarily on experience from
sheet pile projects, where the shaft resistance dominates. In the case of tubular piles, however, the
toe resistance contributes significantly to the total driving resistance. Westerberg et al. (1995)
suggested that the dynamic toe resistance during vibratory driving of piles is approximately equal
to the CPT cone stress, qc. Figure 3 shows the centrifugal force, required to drive a tubular pile
with closed toe into granular soils of variable density, Massarsch and Fellenius (2017). The soil
conditions are expressed according to the classification given in Table 4.
The application of Figure 3 is illustrated by the following example. The centrifugal force
required to drive a tubular pile with a toe diameter of 350 mm into a granular soil with a cone
resistance qc ≈ 10 MPa is approximately 950 kN. If the soil at the pile toe has a cone stress qc ≈ 15
MPa, the required centrifugal force will be approximately 1450 kN.
An additional parameter to be considered when assessing vibratory driving is the relative
displacement of the pile to the soil. The relative displacement is particularly important for
overcoming the shaft resistance in fine-grained (cohesive) soils. The larger the displacement
amplitude, the more effective the driving process will be. The displacement amplitude depends on
the total dynamic mass to be accelerated by the vibrator (dynamic mass) and the eccentric moment.
In the following example, a sheet pile shall be driven by a vibrator (MS-28) with a 3,250-kg
dynamic mass (vibrator and clamp). The mass of the 15 m long pile is 2,700 kg. The displacement
amplitude can now be estimated from Eq. (2). The double displacement amplitude (2s) of the
suspended pile, which is usually quoted by vibrator manufacturers, is 9 mm (Me = 28 kgm). During
sheet pile penetration, the vibration amplitude decreases, depending on the stiffness of the soil to
be penetrated.
An important advantage of vibratory driving is that many aspects of the installation process can be
monitored, controlled, and documented. With modern electronic measuring equipment, it is
possible to acquire, display, and record information from different sensors, which can be mounted
on the pile, the vibrator, the power pack, and the ground. Figure 4 shows the set-up of a vibratory
monitoring system, which initially was developed for deep vibratory compaction.
When monitoring vibratory driving of piles and sheet piles, it is desirable that a data logger
be used to record the following parameters:
Ground vibrations can be recorded using geophones, Figure 5a, and can be displayed to the
machine operator during vibratory driving, Figure 5b. Monitoring of information obtained during
the vibratory driving process and response of the ground and/or of adjacent structures is an
important aspect of modern vibratory works. For instance, in the case of vibratory driving in the
vicinity of vibration-sensitive buildings or installations, a computer-operated system can be used
to control the maximum vibration intensity in order to ensure that specified limiting values are not
exceeded. Moreover, when vibrators are used for deep vibratory compaction, the vibration
measurements can be used to guide the operation to ensure that maximum transfer of vibration
energy is obtained from the vibrator/probe system to the surrounding soil, for example when using
the resonance compaction system (Massarsch, 1991).
From vibration monitoring, several important parameters can be derived such as: pile and soil
displacement amplitude, resonance frequency of vibrator-pile-soil system, pile/probe penetration
speed. These parameters can be displayed to the machine operator in real time and assist in the
optimization of the driving or compaction system. The displacement amplitude provides important
information regarding the dynamic mass to be accelerated by the vibrator. This information is
particularly important when driving piles or sheet piles into cohesive soils. Soil compaction is
enhanced when the vibrator is operated at the system resonance frequency, at which ground
vibrations are strongly amplified. However, pile or sheet pile penetration will be low at the
resonance frequency, at which the risk of vibration problems in surrounding buildings increases.
The pile penetration speed is another important parameter, which reflects the efficiency of
vibratory driving. However, the penetration speed is relevant only when the pile or sheet pile is
allowed to penetrate under the full weight of the vibrator.
The above described monitoring system has been used on a large number of projects, where
resonance was used to increase the compaction efficiency in coarse-grained soil. In the following
case history, a steel pipe pile (with closed toe) was vibrated into loose to medium dense sandy soil,
using a Müller MS100 variable-frequency vibrator, see Table 2. The dynamic ground was
monitored by a tri-axial geophone, placed 4 m away from the compaction point. The soil profile
consisted of medium to dense sand, which prior to compaction was loose to medium dense. During
the initial probe penetration phase, the compaction probe was vibrated at high frequency (35 to
39 Hz). During the compaction phase, the vibration frequency was gradually lowered until
resonance and, thus, maximum vibration amplification was achieved (about 13 - 16 Hz). The
response at the ground surface during vibratory driving is shown in Figure. 6. The vertical ground
vibration velocity during penetration varied typically between 0 and 4 mm/s (average 2 mm/s),
with peak values around 6 mm/s. At resonance (between 12 and 16 Hz), the vertical vibration
velocity increased significantly, with maximum values reaching 20 mm/s. The vibration
amplification was about 5 to 10. At high frequencies, the probe penetration velocity is usually
significantly higher (by a factor of up to 10) compared to driving at resonance frequency.
Field monitoring of ground vibrations can provide valuable information regarding the transfer
of vibration energy from the pile to the surrounding soil. It is apparent that efficient pile penetration
occurs when the vibration frequency is significantly higher than the resonance frequency of the
vibrator-pile-soil system. On the other hand, transfer of vibration energy and, therefore, the
compaction effect is enhanced when the vibrator is operated at the resonance frequency.
When a pile or sheet pile is to be installed by a vibrator, the selection of the driving equipment and
the installation process should be carefully planned. The vibrator type and capacity as well as the
driving parameters (operating frequency and amplitude) should be chosen, based on geotechnical
information. Using an unsuitable vibrator will not only result in project delays and infer additional
costs but under unfavorable conditions, also produce damaging ground vibrations in buildings or
installations in the ground. The following paragraph proposes a concept which can be used to
estimate vibrator centrifugal force, based on CPT results. It is recommended to determine the
optimal vibratory driving process by field trials. The objective is to develop a correlation between
CPT penetration resistance and pile penetration speed for a given vibrator type and pile size.
Rational design of a vibratory driving project requires site information that includes a well-
established soil profile with soil description obtained from laboratory study of soil specimens. The
best additional direct information is a continuous record of soil layering and density, such as
The driveability of piles or sheet piles can be related to the penetration resistance from dynamic
probing. In the case of heavy dynamic probing (DPH), a mass of 50 kg with a height of fall of 500
mm impacts on a steel rod and the number of blows for a penetration of 10 cm (N10) is determined.
An approximate relationship between different penetration testing methods and strength of
granular soils is given in Table 4. It is possible to correlate the penetration resistance to the number
of vibration cycles required for a pile or sheet pile to penetrate into granular soil. The speed of
probe (pile or sheet pile) penetration can be measured (preferably recorded every second, but
reported as penetration per minute). During driving, it is important that the vibrator rests fully on
the probe and is not held back by the machine operator, as this will affect the probe penetration
speed. It is also important, that the vibration frequency is high (above the resonance frequency)
and kept as constant as possible. At a high vibration frequency, the vibration resistance will be
caused primarily by the toe resistance of the probe, similar to the case at dynamic penetration
testing.
In the following example, a sheet pile is driven by a vibrator operating at 40 Hz into a sand
deposit consisting of several layers with variable DPL penetration resistance, blows/10 cm, shown
in Figure 7.
Fig. 7. Typical dynamic penetration test, Light Dynamic Penetrometer (DPL), from Schönit
(2009). The standard for the DPL (10 kg, 500 mm height of fall) is to record the
penetration as blows per 10 cm penetration).
Fig. 8. Sheet pile penetration speed measured at three different vibration frequencies (two tests
for each frequency), after Schönit (2009).
At field trials, the key vibrator performance parameters to record are the eccentric moment,
the vibration frequency, the centrifugal force, and the displacement amplitude. It is important to
record the displacement amplitude of the vibrator-pile system prior to driving of the pile
penetration into the ground. The intensity of ground vibrations adjacent to the pile should also be
recorded, as this information can be used to determine the risk of vibration amplification should
the vibrator operating frequency be too close to the resonance frequency of the vibrator-pile-soil
system. As the vibrator operating frequency, f (Hz or rpm) is known, it is possible to convert the
measured pile penetration speed, v (cm/min) into an equivalent number of penetration cycles, ce
per depth interval (cycles/cm).
(3)
In the above example, the vibrator operating frequency was kept constant at 40 Hz (2,400
rpm). It is now possible to convert the penetration speed from Figure 8 into an equivalent number
of vibration cycles, as shown in Figure 9. It is important to appreciate that the above shown
correlation is soil-type specific and influenced by the type and capacity of vibrator, and the size
and geometry (toe area) of the sheet pile.
Vibrators can be used for the efficient installation of piles and sheet piles in coarse-grained soil.
However, innovative solutions have been developed to broaden the range of vibrator applications
and to facilitate vibratory driving in difficult soils.
At present, it is difficult to assess the bearing capacity of preformed piles, installed by
vibratory driving. However, recent studies regarding the bearing capacity of cast-in situ piles have
been published by Zeilinger and Hudelmaier (2009). They investigated the bearing capacity of
cast-in-situ piles, installed by different driving methods (impact-driven, vibrated and vibro-jetted).
Static and dynamic loading tests were performed on these piles that suggest that vibratory driving
and water jetting does not have a detrimental effect on the static bearing capacity of in situ cast
piles.
One interesting new concepts in foundation design during the past decades is the use of
settlement-reducing elements (piled rafts). The foundation slab is supported by conical concrete
elements, installed by vibratory driving (cf. Figure 11). This foundation concept is similar to
ground improvement by stone columns or mixed in place columns. Prefabricated elements
(concrete, steel or timber) can be used to increase the stiffness of loose or soft soil layers. However,
vibrated conical elements do not suffer from the uncertainties associated with the installation of
stone columns in soft, compressible soils. Massarsch et al. (1997) described a design concept,
which is based on the load-sharing between the foundation slab foundation and reinforcing
elements.
6. CONCLUSIONS
Vibratory driving of piles and sheet piles is a common method, especially for the installation of
foundation units and sheet piles. The most important vibrator parameters for vibratory driving are
the centrifugal force, the vibration frequency, and thus also of the eccentric moment.
The vibrator-pile-soil interaction is a function of the resonance vibration frequency. Results
of a theoretical analysis are presented which can be used to predict reliably the resonance of the
vibrator-pile-soil system. The most important parameters governing the resonance frequency are
the stiffness (shear wave speed) of the soil and the mass of the vibrator and the sheet pile. The
eccentric moment does not affect the resonance frequency.
Empirical rules can be used to assess the required vibrator capacity (centrifugal force) for
sheet pile driving in primarily coarse-grained soil. However, a more reliable concept can be used
to predict the drivability of sheet piles or piles. A method is described which makes it possible to
correlate the penetration resistance to the number of vibration cycles during sheet pile driving. A
case history is presented which demonstrates the concept and shows that a correlation exists in
coarse-grained soil between penetration resistance (dynamic penetrometer) and number of
vibration cycles per depth interval. Based on this information, which can be obtained either from
field trials or experience from past projects, it is possible to determine the penetration speed during
7. ACKNOWLEDGEMENT
The advise and assistance by Bengt H. Fellenius is gratefully acknowledged. Also, the astute
comments by the reviewers of the manuscript are appreciated.
8. REFERENCES
Massarsch, K.R. 1991. Deep Soil Compaction Using Vibratory Probes, ASTM Symposium on
Design, Construction and Testing of Deep Foundation Improvement: Stone Columns and
Related Techniques, Robert C. Bachus, Ed. ASTM Special Technical Publication STP 1089,
Philadelphia, 1990, pp. 297-319.
Massarsch, K.R., Westerberg, E., and Broms, B.B. 1997. Footings supported on settlement-
reducing vibrated soil nails. XIV, International Conference on Soil Mechanics and Foundation
Engineering, Hamburg 97, Proceedings, Vol. 3, pp 1533-1539.
Massarsch, K.R. and Fellenius, B.H. 2017. Grundlagen des vibrationsrammens von Pfählen und
Spundbohlen (Fundamentals of vibratory driving of piles and sheet piles). Geotechnik. Ernst
und Sohn – Wiley. Accepted for publication.
Müller-ThyssenKrupp. 2004. MÜLLER Vibrator technology. ThyssenKrupp GfT Bautechnik,
Machinery - Product range. 15 p.
Schönit, M. 2009. Online-Abschätzung der Rammguttragfähigkeit beim langsamen Vibrations-
rammen in nichtbindigen Böden (Online-estimation of bearing capacity of driven piles during
slow vibratory driving in non-cohesive soils). Doctoral thesis, Universität Karlsruhe (TH),
Fakultät für Bauingenieur-, Geo- und Umweltwissenschaften. Universitätsverlag Karlsruhe,
Reihe F, Heft 65, 200 p.
Westerberg, E., Massarsch, K. R., and Eriksson, K. 1995. Soil resistance during vibratory pile
driving. Proceedings: International Symposium on Cone Penetration Testing, CPT´95, Vol. 3,
pp. 241 – 250.
Zeilinger, H. and Hudelmaier, K. 2009. The bearing capacity of vibrated cast-in-place-piles. Pile
Driver, Q4, V. 6, p. 44 – 47.
ABSTRACT. For routine site investigations, the use of hybrid tools such as the seismic piezocone
and seismic flat dilatometer offer superior efficiency and economy to provide sufficient profiling
of subsurface conditions and the evaluation of geoparameters, as compared with routine soil
boring, drilling, and sampling. A single sounding provides up to five separate measurements on
soil behavior that can be used to calculate capacity and displacements for axial and lateral pile
foundation capacity, as well as in applications involving quality control for ground modification
projects. Axial pile capacity can be assessed using traditional approaches that utilize limit plasticity
and static equilibrium, or alternatively via direct in-situ testing methods. Load tests from a bridge
project in Minnesota are presented to illustrate the applicability of seismic piezocone testing.
1. INTRODUCTION
Each and every civil & environmental project requires a geotechnical site investigation to
determine the makeup of the underlying ground at that particular location. This can be
accomplished today using a combination and assortment of geophysical methods, soil drilling &
sampling for laboratory testing, and in-situ field testing.
There are currently many different types of deep foundation systems being used commercially
in support of civil engineering works for building loads, bridge piers, ports, and tower structures
both onshore and offshore. Pile types can include driven steel H and pipe (open-end versus closed-
end), precast versus prestressed concrete, timber, composite, tapered, and monotubes, or systems
of bored or augered deep foundations, such as cased or uncased drilled piers, slurry shafts,
augercast pilings, and caissons, as well as specialty types (e.g., Fundex, Omega, Screw, Dewaal).
Consequently, the notion that a single test number such as the SPT-N value can suffice for all the
needs in the analysis and design of modern piling foundations must be replaced with a more
rational belief that multiple measurements are paramount.
Herein, we shall explore the utilization of more modern tests, such as the seismic cone and
seismic dilatometer, for the collection of several measurements concerning soil behavior. In
addition to assessing soil stratigraphy and geoparameters in an efficient and economic manner,
these tests permit an evaluation of the nonlinear axial load-displacement response of deep
foundations.
An illustration showing the importance of this correction is depicted in Figure 1 using SPT
data from the National Geotechnical Experimentation Site (NGES) at Northwestern University
(Finno et al. 2000). Here, the upper soils are comprised of clean fine sands (0.15 mm < D50 < 0.30
mm) that were subjected to two sets of SPTs in soil test borings using a safety hammer and an
automatic hammer. Figure 1a shows the raw uncorrected N-values while Figure 1b presents the
energy-corrected values. The significance and importance of the energy corrections on the N-
values is quite evident.
Fig 1. SPT profiles at the national test site at Northwestern University: (a) uncorrected N-values;
(b) corrected N60 (data from Finno et al. 2000)
using autohammers has now risen to a mean value ± one standard deviation of ER (ave) ≈ 82% ±
7% based on measurements taken in the past five years (Honeycutt et al. 2014).
Fig. 3. Hybrid in-situ geotechnical tests: seismic piezocone and seismic dilatometer
Then the index Ic is recalculated. Iteration converges quickly and generally only 3 cycles are
needed to secure the operational Ic at each depth. The soil zones and associated Ic values are
detailed in Figure 5. The sensitive soils of zone 1 can be screened using the following expression:
In the case of soft to firm intact clays and silty clays, the effective friction angles is determined
from the normalized cone resistance and porewater pressure parameters (Senneset et al. 1989;
Mayne 2016), as shown in Figure 7. The exact solution when the angle of plastification = 0 is
given as:
tan 2 (45 ' / 2) exp( tan ' ) 1
Q
1 6 tan '(1 tan ' ) Bq
This algorithm is specifically applicable for the following ranges of porewater pressure parameter
(0.1 < Bq < 1) and effective stress friction angles (20° < ' < 45°).
The exponent m' has also been calibrated with CPT material index:
0.28
m' 1
1 ( I c / 2.65) 25 (11)
The value of K0 finds applicability in assessing the pile side friction via the beta method, whereby
as a first approximation: = K0 ∙ tan'.
where CM = pile material factor = 1 (rough cast-in-place concrete); 0.9 (prestressed concrete)
0.8 (timber); and 0.7 (steel);
CK = installation factor = 0.9 (bored or augered); 1.0 (low displacement, e.g. H-pile or
open end pipe); and 1.1 (driven solid, e.g. prestressed concrete, closed-end pipe).
The evaluation of toe resistance (rt) is often taken from limit plasticity solutions with associated
shape and depth factors (Brown et al. 2010). For undrained toe response, the full value may be
attained, thus:
where Nq' ≈ 0.77 ∙ exp ('/7.5°) = approximate expression for bearing factor
fx' = strain incompatibility factor = 0.1 (bored piles); 0.2 (jacked); 0.3 (driven)
where Cse is a coefficient for shaft friction based on soil type and Cte = toe resistance coefficent.
Figure 9 provides a quick summary overview of the approach and full details on UniCone are
given elsewhere (Fellenius 2016).
Fig. 9. Overview of UniCone Method (after Eslami and Fellenius 1997; Fellenius 2016)
and for the remaining soil types, the shaft coefficient may be estimated from the following
expression:
where 1 = pile type factor: 0.84 (bored piles), 1.02 (jacked), 1.13 (driven); 2 = load direction
factor: 1.11 (compression) and 0.85 (tension), and 3 = loading rate factor (1.0 for soils with Ic <
2.6 and for Ic > 2.6: 0.97 (stepped load) and 1.09 (constant rate of penetration).
The toe coefficient may be estimated from:
The total axial compression capacity (Qult) is the sum of shaft capacity (Rs) plus toe capacity (Rt):
where As = circumferential area of the pile at depth z and At = toe area of the pile.
where g' = fitted exponent (≈ 0.3 ± 0.1 for uncemented and non-structured soils).
The seismic piezocone test (SCPTu) thus finds special application to deep foundation analysis
because it provides sufficient data (qt, fs, and u2) for axial pile capacity calculations, as well as
supplying the necessary soil stiffness (Emax) for the evaluation of displacements and load transfer.
To illustrate the usefulness of the SCPTu results in axial pile evaluation, a case study from load
tests for the Wakota Bridge are presented.
The ten-lane Wakota Bridge is located southeast of Saint Paul, Minnesota and was completed in
2010. The bridge enables interstate loop 494 to cross the Mississippi River. During the design
phase, load tests on both driven open-end and closed end steel pipe piles were conducted with test
piles having an outer diameter d = 0.457 m, an embedded length L = 32 m, and wall thickness t =
12.5 mm (Dasenbrock 2006). Both piles were loaded in axial compression, then afterwards loaded
in tension, using a static reaction frame arrangement.
Soil conditions at the site can be assessed via the SCPTu sounding presented in Figure 11,
indicating primarily firm sands with a few interbedded clay layers found at depths of 1, 3.5, 7 - 12,
17, and 22 - 27 m. The post-processing of the SCPTu provided direct estimates of the soil type via
material index (Ic), unit weight (t), effective friction angle ('), preconsolidation stress (p'), and
K0 profiles for input into beta method for pile shaft resistance (ave. rs = 65 kPa), as well as by
direct CPT methods using Unicone (ave. rs = 70 kPa) and Modified Unicone (ave. rs = 75 kPa).
Evaluation of the toe resistance determined rte = 3593 kPa by the Modified Unicone expression.
The evaluation of the two tests in axial compression are presented in Figure 12, while Figure
13 shows the two tests in tension. Overall, the agreement between the load-displacement responses
Fig. 12. Axial compression load tests at Wakota site: (a) open-end pile; (b) closed-end pile
are comparable between the measured field load tests and those generated using the elastic
continuum solutions that rely on SCPTu data for input. For reference, the Euro criterion for
"capacity" is taken as that load corresponding to (s/d) = 10% is shown on the measured load test
curve.
5. DEVELOPMENTS IN GEOPHYSICS
The MASW results were combined to form the cross section shown in Figure 14 with a clear
indication of poor ground (Vs < 150 m/s), normal soil conditions (Vs ≈ 200 m/s), and very strong
ground conditions (Vs > 300 m/s). Such information would be valuable prior to selection of
locations of pile load tests, production pile installation, and/or necessary pile length
determinations, as well as beneficial to applications in extent of ground modification and other
construction activities.
6. CONCLUSIONS
Site characterization is an important component for the proper selection of deep foundation
alternatives. Several advances in geotechnical site investigation have been made that quantify
needs in foundation performance, including specialized probes for evaluation of scour potential,
thermal soil properties, pile friction roughness, and soil stiffness, as well as a suite of noninvasive
geophysics techniques. Of particular benefit is the utilization of hybrid tests such as seismic
piezocone and seismic dilatometer that collect information at two ends of the stress-strain-strength
curves in soils, namely the small-strain stiffness (Gmax and Emax) from the shear wave velocity and
the shear strength (max or su, c', and ') that are both needed in the assessment of the axial load-
displacement-capacity of deep foundations.
7. ACKNOWLEDGMENTS
The author gives thanks to Derrick Dasenbrock for sharing the MnDOT SCPTu and load test
results. Appreciation is given to ConeTec of Richmond, BC for support of GT in-situ research
activities.
Akrouch, G.A., Briaud, J-L., Sanchez, M. and Yilmaz, R. 2016. Thermal cone test to determine
soil thermal properties. J. Geotechnical & Geoenvironmental Engrg. 142 (3): 04014085.
Amoroso, S., Monaco, P., Lehane, B.M. and Marchetti, D. 2014. Examination of the potential of
the SDMT to estimate in-situ stiffness decay curves. Soils and Rocks 37 (3): 177-194.
Ali, H., Reiffsteck, P., Baguelin, F., van de Graaf, H., Bacconnet, C. and Gourvès, R. 2010.
Settlement of pile using cone loading test: load settlement curve approach. Proc. 2nd Intl.
Symp. on Cone Penetration Testing (CPT'10), Huntingdon Beach, California: Paper ID 3-24.
Brown, D.A., Turner, J.P. and Castelli, R.J. 2010. Drilled shafts: construction procedures and
LRFD design Methods. Report FHWA-NHI-10-016, National Highway Institute, Federal
Highway Administration, Washington, DC: 972 p.
Dasenbrock, D.D. 2006. Assessment of pile capacity by static and dynamic methods to reconcile
design predictions with observed performance. Proc. 54th Minnesota Geotechnical
Engineering Conference, Edited by Labuz, J., Univ. of Minnesota, St. Paul: 95-108.
Davidson, J.L., Maultsby, J.P. and Spoor, K.B. 1999. Standard penetration test energy calibrations.
Report Number WPI 0510859 by Univ. Florida - Gainesville submitted to Florida Dept.
Transportation, Tallahassee: 123 pages.
Eslami, A. and Fellenius, B.H. 1997. Pile capacity by direct CPT and CPTu methods applied to
102 case histories. Canadian Geotechnical Journal 34 (6): 886-904.
Fellenius, B.H, 2016. Basics of Foundation Design, Electronic Edition, Sidney, British Columbia,
453 pages: www.fellenius.net
Finno, R.J., Gassman, S.L. and Calvello, M. 2000. The NGES at Northwestern University.
National Geotechnical Experimentation Sites (GSP 93), ASCE, Reston, Virginia: 130-159.
Gabr, M., Caruso, C., Key, A. and Kayser, M. 2013. Assessment of in-situ scour profile in sand
using a jet probe. ASTM Geotechnical Testing J. 36 (2): 1-11.
Hebeler, G.L. and Frost, J.D. 2006. A multi-piezo-friction attachment for penetration testing.
Geotechnical Engineering in the Information Technology Age (Proc. GeoCongress, Atlanta),
ISBN 9781604234909; ASCE, Reston, Virginia: 1-6.
Honeycutt, J.N., Kiser, S.E. and Anderson, J.B. 2014. Database evaluation of energy transfer for
cme automatic hammer SPT. J. Geotechnical & Geoenv. Engrg. 140 (1): 194-200.
Jardine, R.J., Gens, A., Hight, D.W. and Coop, M.R. 2004. Developments in understanding soil
behavior. Advances in Geotechnical Engineering (Proc. Skempton Conference), Thomas
Telford, London: 103-206.
Ku, T., Mayne, P.W, and Cargill, E. 2013. Continuous-interval shear wave velocity profiling by
auto-source and seismic piezocone tests. Canadian Geotechnical J. 50 (1): 382–390.
Lunne, T., Robertson, P.K. and Powell, J.J.M. 1997. Cone Penetration Testing in Geotechnical
Practice. EF Spon/Blackie Academic, London, 352 p.
Marchetti, S., Monaco, P., Totani, G., and Marchetti, D. (2008). In-situ tests by seismic dilatometer
(SDMT). From Research to Practice in Geotechnical Engineering (Proc. GeoCongress 2008,
GSP 180, New Orleans), ASCE, Reston, Virginia: 292-311.
Mayne, P.W., Christopher, B., Berg, R., and DeJong, J. 2002. Subsurface Investigations -
Geotechnical Site Characterization. Publication No. FHWA-NHI-01-031, National Highway
Institute, Federal Highway Administration, Washington, D.C., 301 p.
ABSTRACT. The axial capacity of pile foundations in sand is affected by the method of
installation. Pile driving (or jacking) causes significant changes in soil state (the density of the
sand below and in the vicinity of the pile increases). However, at any given depth, the frictional
resistance of displacement piles degrades due to release of lateral stresses acting on the pile shaft.
Therefore, in addition to stress state and density, design methods for displacement piles in sand
have incorporated a shaft resistance degradation factor into the design equations. Accounting for
plug formation and evolution during driving and loading of partial-displacement piles (e.g.,
open-end pipe piles, H piles) are important. Accordingly, design methods have added the
incremental filling ratio and/or plug length ratio as input variables. This paper presents some of
the current pile design methods and discusses recent advances in research that have led to
significant improvement in pile design methods.
1. INTRODUCTION
The response of piles to axial loading depends, to a significant extent, on the installation
methods employed. Accordingly, piles are often classified according to the level of soil
displacement imposed on the in situ soil by the installation method. Non-displacement piles
(e.g., bored piles or drilled shafts), which are constructed by removing soil from the ground
down to the desired depth with partial augers and placing concrete and reinforcement into the
excavated hole, are on one end of the spectrum of pile response to loading. Full-displacement
piles (e.g., closed-end pipe piles or precast reinforced concrete piles) are on the other end of the
spectrum of pile response to loading. Figure 1 shows how the main types of piles are classified
according to the soil displacement induced by their installation processes.
Full-displacement piles are mostly installed by driving or jacking (monotonic or multi-stroke
jacking) them into the ground. The installation of full-displacement piles both displaces laterally
the sand in the vicinity of the shaft of the pile and preloads the soil at the base; both of these
effects lead to changes in void ratio and stress state. Installation of a displacement pile in sand
causes densification of the sand within an annular zone surrounding the pile with fast dissipation
of the excess pore pressure generated during installation. In addition, degradation of frictional
resistance along the pile shaft due to shear loading and unloading cycles (induced by hammer
blows or jacking strokes) is caused by the contractive tendency of the sand immediately adjacent
to the pile shaft. As shown in pile installation simulations, the sand in this zone is subjected to a
shear loading cycle that generates a dilative response, leading to an increase in void ratio and
thus setting up a tendency for contractive response in subsequent shearing cycles. Such
contractive tendency induces a release of the lateral stress acting on the pile shaft and a reduction
in shaft resistance (Figure 2). Friction degradation decreases with the installation depth and
needs to be considered in design. Note that pipe piles may be installed using vibratory hammers
Pile Types
Fig. 1. Effect of pile installation on soil displacement and unit shaft and base resistances.
Initial stage Just after pile tip passes element A (Expansion of a cavity)
σʹh0 A σʹ h1
(a) (b)
σʹ h2
τs Normal Stress σʹ h2 reduces with increasing
number of shear cycles: Friction degradation
Formation of
shear band (c)
Fig. 2. Mechanics of friction degradation along the shaft of a displacement pile: (a) initial state at
point A within the ground (b) pile tip passes point A – a cavity is created within the ground, and
(c) shearing associated with installation and/or loading.
Pile head
The ultimate resistance Qult is associated with an ultimate load criterion typically defined in
many design methods as the load corresponding to a relative settlement w/B at the pile head of
10%. This has been done to account for the fact that a structure supported by piles will
experience the same settlement as the head, not the base, of the piles. Design methods are
developed based on a specific ultimate load criterion, but simulations have the advantage that
any settlement level may be considered (either with respect to the head or the base of the pile), so
different ultimate load criteria or serviceability limit states may also be checked in design.
The ultimate base and limit shaft resistances are given by:
n
QsL = ∑ qsLi Asi (2)
i =1
where the subscript i represents a soil layer (i = 1, 2, 3, …) for which the shaft resistance is to be
calculated, n is the total number of layers in the soil profile that are crossed by the pile, qsLi and
qb,ult are the unit limit shaft and unit ultimate base resistances, Ab is the projected plan area of the
pile base and Asi is the surface area of the pile shaft within the ith layer.
The unit shaft and base resistances can be estimated using indirect and direct design
methods. Indirect design methods require that fundamental soil variables of interest (e.g., relative
density and critical-state friction angle for sands, and undrained shear strength and critical-state
and residual-state friction angles for clays) be estimated first either from laboratory or in situ test
data. Direct design methods correlate field test data (e.g., cone resistance qc or SPT blow count
number N) directly with the unit shaft and base resistances of different piles. The focus of this
paper is mainly on indirect and CPT-based design methods applicable to sand, as discussed next.
where σʹv0 is the initial in situ vertical effective stress in the center of the layer where qsL is
calculated, δ is the sand-pile interface friction angle (it is usually expressed as a function of the
critical-state friction angle ϕc of the sand and pile surface roughness), K is the coefficient of
lateral earth pressure. The effective stress analysis expressed in equation (4) is referred to as the
β method ( β = K tan δ ) in the pile literature as well. The fundamental design equation for the
unit ultimate base resistance of a single pile in sand is:
where Nq,ult is a bearing capacity factor that is linked to the ultimate load criterion.
⎡ K DR ⎡ ⎛ σ v 0 ' ⎞⎤ ⎤
⎢1.5− 0.35ln ⎜ ⎟⎥
100 ⎝ p A ⎠⎦⎥ ⎥
qsL = ⎢ 0.3 K0 −0.4 0.67e ⎣⎢ σ vʹ0 tan φc (6)
⎢e 0 ⎥
⎣ ⎦
and
1.83 0.4
qb ,ult qb ,10% ⎛ D ⎞ ⎛ σ h' 0 ⎞
= = 62 ⎜ R ⎟ ⎜ ⎟ (7)
pA pA ⎝ 100% ⎠ ⎝ pA ⎠
where K0 = coefficient of earth pressure at-rest, pA = reference stress (= 100 kPa) and DR =
relative density of the sand (%) and σʹh0 is the initial in situ horizontal effective stress at the pile
base.
Eq. 6 shows the direct dependence of qsL on the sand relative density and stress state (K
increases with increasing sand density and decreasing initial vertical effective stress) and the
critical-state friction angle, which is an intrinsic variable of the sand (ϕc is used instead of δ in
The cone resistance and relative density of the sand are further related by (Salgado and Prezzi
2007):
⎛q ⎞ ⎛σʹ ⎞
ln ⎜ c ⎟ − 0.4947 − 0.1041φc − 0.841ln ⎜ h 0 ⎟
p ⎝ p A ⎠ ≤ 100%
DR (%) = ⎝ A ⎠ (9)
⎛σʹ ⎞
0.0264 − 0.0002φc − 0.0047 ln ⎜ h 0 ⎟
⎝ pA ⎠
2.3 Displacement Piles
Simulation of the installation of displacement piles is difficult as it requires use of not only an
advanced constitutive model for the sand but also advanced numerical techniques that are able to
handle the large-deformations and rotations induced in the sand during driving or jacking.
Research on developing the capabilities for simulating large-deformation problems (cone
penetration and pile installation) using FEM and the Material Point Method (MPM) is underway.
Many design methods are available in the literature for driven closed-end pipe piles. In the
Purdue method, following the general expression proposed by Randolph (2003), the unit shaft
and base resistances are estimated as (Han et al. 2016; Salgado 2008; Salgado et al. 2011):
h
qsL = 0.2 + (0.02qc / σ vʹ0 − 0.2) exp(−0.05 )σ vʹ0 tan δ c (10)
B
and
where h is the distance from the center of the layer under consideration to the pile base, δ c is the
sand-pile interface friction angle, and qcb ,avg is the average cone resistance within a zone B above
and 1.5B below the base of the pile. Note that the exponential term in Eq. 10 is a frictional
resistance degradation term that is equal to 1 at the base of the pile.
The unit shaft and base resistances are estimated from the cone resistance qc in the ICP
method (Jardine et al. 2005) as:
and
where R is the radius of the pile, Δr is equal to 0.02mm for lightly rusted steel piles, B is the
diameter of the pile and BCPT is the diameter of the cone. The factor Fload is equal to 1 for
compressive loading and 0.8 for tensile loading. In the ICP method, qcb,avg is the qc value
averaged within a zone 1.5B above and 1.5B below the pile base.
The UWA method (Lehane et al. 2005) has a formulation for the unit shaft resistance that is
somewhat similar to that of the ICP method and also includes a factor on the shaft resistance that
accounts for the difference in compressive and tensile loading:
f ⎛ 4GΔr ⎞
qsL = ⎜ σ rcʹ + ⎟ tan δ c
fc ⎝ B ⎠
−0.5
⎡ ⎛ h ⎞⎤
σ rcʹ = 0.03qc ⎢ max ⎜ , 2 ⎟ ⎥ (14)
⎣ ⎝ B ⎠⎦
−0.75
⎡ q /p ⎤
c A
G / qc = 185 ⎢ 0.5 ⎥
⎢⎣ (σ vʹ0 / p A ) ⎥⎦
and
where f / fc = 1 for compressive loading, f / fc = 0.75 for tensile loading and Δr = 0.02mm. In the
UWA method, qcb,avg is calculated following the Dutch averaging technique (Schmertmann 1978;
de Kuiter and Beringen 1979) in which the CPT cone resistance is averaged over a zone
extending from 8B above the pile base to 0.7B to 4B below the pile base, depending on whether
qc increases or decreases below the pile base. Both the ICP and UWA methods as well as the
Purdue method have a degradation factor that is a function of h/B.
For piles jacked in sand, shaft resistance degradation was considered by Basu et al. (2011)
with a term that is a function of the number of jacking strokes N. The following expressions were
obtained from the FEAs performed for monotonic installation of a jacked pile (N = 1):
⎛ K ⎞ ⎡ DR ⎧⎪ ⎛ σ vʹ0 ⎞ ⎫⎪⎤
⎜ ⎟ = 0.93exp ⎢ ⎨2.8 − 0.45ln ⎜ ⎟ ⎬⎥ (17)
⎝ K 0 ⎠ N =1 ⎣⎢100 ⎪⎩ ⎝ p A ⎠ ⎪⎭⎦⎥
With (K/K0)N=1, the ratio K/K0 for a number of jacking strokes N can be calculated from:
⎛ K ⎞ ⎛ K ⎞
⎜ ⎟ = µN ⎜ ⎟ (18)
⎝ K0 ⎠ N ⎝ K 0 ⎠ N =1
{
µ N = A1 + (1 − A1 ) exp − A2 ( N − 1)
A3
} (19)
where A1, A2, and A3 are coefficients that for an initially isotropic sand fabric are given by:
A1 = 0.02
0.3
⎛ p ⎞
A2 = 0.2 ⎜ A ⎟ (20)
⎝ σ vʹ0 ⎠
1.1
⎛D ⎞
A3 = 1.0 − 0.4 ⎜ R ⎟
⎝ 100 ⎠
A1 = 0.02
0.3
⎛ p ⎞
A2 = 0.25 ⎜ A ⎟ (21)
⎝ σ vʹ0 ⎠
2
⎛D ⎞
A3 = 1 − 0.3 ⎜ R ⎟
⎝ 100 ⎠
In the Purdue method (Paik and Salgado 2003, Paik et al. 2003, Paik et al. 2004), which is
based on model pile tests performed in a calibration chamber, the unit limit shaft resistance qsL is
calculated as:
where qb,ultND is the unit ultimate base resistance of a non-displacement pile (Lee and Salgado
1999).
The unit limit shaft resistance in the ICP method for open-end pipe piles is obtained from:
where R* is the equivalent radius of a closed-end pipe pile with the same steel area as the open-
end pipe pile with outer radius Ro and inner radius Ri. Similar to what was proposed for closed-
end pipe piles, the factor Fload is equal to 1 for compressive loading and 0.8 for tensile loading.
However, the ICP method recommends that qsL be further multiplied by 0.9 for open-end piles
loaded in tension. The unit ultimate base resistance depends on whether the open-end pipe pile
behaves as fully plugged or unplugged. For unplugged piles:
*
π ( B02 − Bi2 )
where A = . Open-end pipe piles are considered to be fully plugged when:
4
qc
Bi < 0.02 ( DR − 30 ) or Bi < 0.083 BCPT (26)
Pa
and
f ⎛ 4GΔr ⎞
qsL = ⎜ σ rcʹ + ⎟ tan δ c
fc ⎝ Bo ⎠
−0.5
* 0.3
⎡ ⎛ h ⎞⎤
σ rcʹ = 0.03qc ( A rs) ⎢ max ⎜ , 2 ⎟ ⎥
⎣ ⎝ Bo ⎠ ⎦
0.2
Ars* = 1 − IFR ( Bi2 Bo2 ) , IFRmean ≈ min ⎡1, ( Bi 1.5m ) ⎤
⎣ ⎦
−0.75
⎡ q /p ⎤
c A
G / qc = 185 ⎢ 0.5 ⎥
⎢⎣ (σ vʹ0 / p A ) ⎥⎦ (28)
and
qb,ult = ⎡⎣0.15 + 0.45 Arb* ⎤⎦ qcb,avg
0.2
(29)
Arb* = 1 − FFR ( Bi2 Bo2 ) , FFR ≈ min ⎡1, ( Bi 1.5m ) ⎤
⎣ ⎦
where f / fc = 1 for compression and 0.75 for tension, and FFR is average IFR for the last 3B
depth of pile penetration.
2.4.2 H Piles
Design methods for displacement piles are also used to design H piles but most of them do not
provide specific recommendations for H piles. The limit shaft capacity of H piles depends on
whether under loading a fully plugged mode is operative or not. If the soil in the space between
the flanges forms a plug and becomes an integral part of the pile during loading then the H pile
outside perimeter (the outer boundaries of the H pile cross section are used for calculation of the
shaft surface area) and the gross cross-sectional area (flange width × depth) have been proposed
for use in shaft and base resistance calculations, respectively. When a plug does not form, which
is most likely the case for H piles in sand under static loading condition, shaft resistance may be
mobilized along the entire surface area of the H pile (full H pile-soil interface contact perimeter),
with the actual cross-sectional steel area used in base capacity calculations. In the case of sandy
soils, the use of the total H pile-soil interface area is often recommended for shaft capacity
calculations, and the actual H pile cross-sectional area for base capacity calculations (Tomlinson
1987; Salgado 2008). However, plug formation in H piles and pipe piles needs to be further
studied to improve or develop design methods that are more accurate.
The ICP method provides specific design recommendations for H piles. The H pile cross-
sectional base area Ab is calculated as (see Figure 5):
Ab = As + 2 X p (D − 2t ) (30)
where As = steel area of the H pile, D = depth, t = thickness, Bf = width of flange and
As is the
area of the D
hatched
section
B
f
Fig. 5. Geometry for H pile design problem.
The unit ultimate base resistance is taken as the average cone resistance at the pile base
(qb,ult/qcb,avg = 1). For shaft resistance calculations, Jardine et al. (2005) recommended the use of
the H pile outer perimeter, with qsL calculated as:
0.13 −0.38
⎪⎧ ⎛ σ vʹ ⎞ ⎡ ⎛ h ⎞⎤ Δr ⎪⎫
qsL = ⎨0.029 Fload qc ⎜ ⎟ ⎢ max ⎜ * , 8 ⎟ ⎥ + 2G * ⎬ tan δ (33)
⎩⎪ ⎝ pA ⎠ ⎣ ⎝R ⎠⎦ R ⎪
⎭
where Fload = 0.8 for tension and 1.0 for compression, and R* = [Ab/π]0.5.
Seo et al. (2009) reported the results of a pile load test performed on a H pile installed in
multilayered profile and indicated that the use of reduction factors to the shaft capacity of H piles
using displacement pile methods seems appropriate when full H pile-soil interface contact
perimeter is assumed in calculations [note that reduction factors between 0.5 and 0.8 have been
suggested; e.g., Meyerhof 1976 and Fleming et al. 1992].
uθ
σ′r0 A τ1 τ2 τ3
uz uz
σ′r1 σ′r2 σ′r3 tan η =
η uθ
Fig. 6. Idealization of DD pile installation and loading: (a) initial state, (b) insertion of the
drilling tool into the ground (cavity expansion followed by vertical downward and torsional
shearing), (c) removal of the drilling tool (upward shear unloading along the borehole wall), (d)
axial loading of the pile(simulating an axial load test), and (e) installation parameter η.
The following design equations for the lateral earth pressure coefficient K [which are to be
used for calculation of qsL following Eq. 4] that take into account the initial soil state and the rate
of penetration of the drilling tool into the ground during pile installation resulted from the FEAs:
0.11
K ⎛σʹ ⎞ ⎡⎛ D (%) ⎞ ⎧⎪ ⎛ pA ⎞ ⎫⎪ ⎤
= 0.33 ⎜ v0 ⎟ exp ⎢⎜ R ⎟ ⎨3.59 + 0.53ln ⎜ ʹ ⎟ ⎬ (1 − 0.11tan η )⎥ (34)
K0 ⎝ pA ⎠ ⎢⎣⎝ 100 ⎠ ⎪⎩ ⎝ σ v0 ⎠ ⎪⎭ ⎥⎦
−0.11
K ⎛σʹ ⎞ ⎡⎛ D (%) ⎞ ⎧⎪ ⎛ pA ⎞ ⎫⎪ ⎤
= 1.30 ⎜ v0 ⎟ exp ⎢⎜ R ⎟⎨ 2.91 + 0.38ln ⎜ ⎟⎬ (1 − 0.12 tan η ) ⎥ (35)
K0 ⎝ pA ⎠ ⎢⎣⎝ 100 ⎠ ⎩⎪ ʹ ⎠ ⎭⎪
⎝ σ v0 ⎦⎥
for initially isotropic sand fabric. η is a rate installation parameter expressed as:
where uz and u are the vertical and rotational (torsional) displacements of the drilling tool used
θ
in pile installation (Figure 6e). The installation parameter η represents the ratio of the
advancement of the drilling tool into the ground and the rotation of the drilling tool. Lower
values of η imply slower rate of penetration of the drilling tool into the ground during drilling (as
expected for installation in dense sands), while higher values of η imply easier drilling
conditions (as expected for installation in loose sands). For a drilling condition in which the
auger rotates a single full rotation to penetrate a length equal to the pitch length of the auger, η is
the flight angle of the auger. Note that the value of K/K0 for sand with initially anisotropic sand
fabric is always smaller than of initially isotropic sand fabric.
3. CONCLUSIONS
This paper presents several pile design methods that directly or indirectly account for the effects
of pile installation methods on calculation of axial capacity of piles in sand. Degradation of
frictional resistance along the pile shaft with increasing number of installation-induced shear
cycles is recognized in all design methods for displacement piles discussed in this paper. The
use of cone resistance qc in the calculation of unit shaft and base resistances is widely adopted.
However, guidelines to arrive at a representative qc value at the pile base differ from one method
to the other. Although the capacity calculation methods discussed herein are based on analytical,
numerical and experimental studies, theoretical simulations of installation of displacement piles
and partial displacement piles are still not at a sufficient level of rigor and there is still the need
for field load test data on instrumented piles, with side-by-side comparison of different pile types
installed in the same profiles, in order to further improve the prediction capabilities of pile
capacity calculation methods.
4. REFERENCES
Basu, P., Prezzi, M. and Salgado, R. (2014). Modeling of installation and quantification of shaft
resistance of drilled-displacement piles in sand. International Journal of Geomechanics,
ASCE, 14(2), pp. 214–229.
Basu, P., Loukidis, D., Prezzi, M., and Salgado, R. (2011). Analysis of shaft resistance of jacked
piles in sand. International Journal for Numerical and Analytical Methods in
Geomechanics, John Wiley & Sons, 35(15), pp. 1605-1635.
Basu, P., Prezzi, M., and Basu, D. (2010). Drilled displacement piles – current practice and
design. DFI Journal: The Journal of the Deep Foundations Institute, 4(1), pp. 3-20.
Brown, D.A. and Drew, C. (2000). Axial capacity of augered displacement piles at Auburn
University, New Technological and Design Developments in Deep Foundations.
Proceedings of sessions of Geo- Denver 2000, Geotechnical Special Publication No. 100,
ASCE, pp. 397-403.
Brown, D.A. (2005). Practical considerations in the selection and use of continuous flight auger
and drilled displacement piles. Advances in auger pressure grouted piles: design,
construction and testing. Advances in Designing and Testing Deep Foundations,
Geotechnical Special Publication No. 129, ASCE, pp. 251-261.
ABSTRACT. While observations during pile installation have been used for centuries to produce
a quality deep foundation, using electronic measurements routinely has only become possible with
modern sensors and computers. The beginning of this development took place more than 50 years
ago in the 1960s. While the traditional and one of the most important methods of deep foundation
quality assurance is the static loading test, it is now often partially or completely replaced by the
Dynamic or High Strain Dynamic Loading Test. After a brief review of all available Quality
Assurance and Quality Control methods, the paper describes the High Strain Test procedure, its
theoretical background, equipment, software and standards. Two examples, one for driven and one
for a cast-in-situ pile, demonstrate the benefits and limitations of the available methods.
1. INTRODUCTION
All deep foundation types, whether prefabricated and driven or cast-in-situ after drilling will be
called a “pile” in this paper. Deep foundations are needed where surficial soils have insufficient
strength to support a building, bridge or other structure. Because of their importance for adequate
performance and because of their cost, deep foundation specifications and building codes
increasingly demand thorough testing. As a reward for a thorough testing regime, modern codes
reward the project with reduced factors of safety (AASHTO, 2014). The economic impact from
the savings realized with reduced factors of safety can be very significant (Likins 2015).
Tests were always performed, most notably the static top-down loading test, to verify adequate
foundation performance. The load carrying capacity of both driven piles and bored piles has
traditionally been evaluated by static load tests (SLT). Using either dead weights or reaction piles,
the test pile is loaded by a hydraulic jack pressing against a reaction frame either in compression
(ASTM D1143) or in tension (ASTM D3689). For larger test loads ASTM D1143 requires an
instrumented load cell.
Although the top load static test is the reference test against which all other tests methods are
judged, the test’s shortcomings are numerous: The test is expensive and time consuming, contains
many potential error sources and is non-unique as far as interpretation, as has been widely
documented in the literature (i.e., Fellenius, 1990). Moreover, when it comes to large loads, the
test can be very dangerous due to high “dead” loads (kentledge) or locked-in energies when elastic
reaction systems are in use.
For large cast-in-situ piles with high capacities the safer and usually more economical bi-
directional test has been developed (Elisio Da Silva, 1983 and Osterberg, 1984) and is being
For driven piles, dynamic formulas have been widely accepted because of their simplicity; their
benefits, if any, and limitations have been described by Allin et al. (2015) and Likins et al. (2012a).
Dynamic formulas use generally only an assumed energy and an observed blow count to assess
pile bearing capacity. Their use is, therefore, limited to driven piles. Their simplicity is often given
as a reason for their use, however, that simplicity is also the reason for poor correlations with static
test results which a myriad of versions and factor adjustments could not overcome, thus requiring
uneconomically high factors of safety. In addition these formulas do not allow for stress
calculations.
A numerical solution of the underlying wave equation became practical with the availability of the
electronic computer. Great interest among both construction and academic professionals was
generated by Smith (1960) in the United States. At about the same time in Europe, Fischer (1960),
for example, developed his graphical solution of the wave equation which he applied to a number
of pile driving problems and which was related to software developed by others such as Meunier
(1984). Any of these impact simulation programs proved to be much more accurate than the
traditional dynamic formulas and not only provided a relationship between the number of blows
per unit penetration (blow count) and bearing capacity, but also a reasonably accurate prediction
of dynamic pile stresses. Among the additional options included in these programs was the
driveability analysis which became particularly popular since it simulated not only what happens
under one hammer blow, but rather what can be expected during the complete pile installation
process using diesel, hydraulic, air, steam or even vibratory hammers (e.g., Rausche et al., 2009).
Beginning in the late 1950s measurements which could be routinely performed on construction
sites were developed at Case Western Reserve University (Goble and Rausche, 1970a); Hussein
and Goble (2004) briefly described these developments. Interpretation of the measured force and
velocity signals were based on closed-form solutions of the wave equation (Goble and Rausche,
1970b). This approach was so successful, particularly after being introduced in Sweden and from
there around the world, that the first Stress Wave Conference was held in Gothenburg in 1980 with
repeat conferences every 4 years. A wealth of papers describing the advancement of the dynamic
pile testing methods has been generated by these conferences. Similar development efforts were
also made in Europe, see for example Beringen et al. (1980). The Case Method is further described
below.
The sensors developed during the 12 years of research at Case Western Reserve University
(CWRU) included reusable strain transducers which would be bolted to the side of the pile. The
strain signal was then multiplied with the pile’s elastic modulus and cross sectional area to yield
the pile force. Alternatively a so-called “top transducer” was used (Goble and Rausche, 1970a)
which could be calibrated in a universal testing machine and avoided concrete modulus
uncertainties when cast-in-situ piles were tested. The top transducer was basically an instrumented,
short section of heavy-wall pipe. Velocity was measured using accelerometers whose signals were
integrated to yield velocity. While various changes in sensor technology improved significantly,
the quality of the signals recorded and evaluated, the basic measurement systems are still the same
as developed during the original research. ASTM D4945, “Standard Test Method for High Strain
Dynamic Testing of Deep Foundations”, specifies how the dynamic tests have to be performed.
Similar standards now exist in many other countries or regions such as Australia, Brazil, China,
and Europe.
For the calculation of pile bearing capacity two approaches were chosen; one was a simple
equation that could be solved by computer between hammer blows and one a signal matching
approach which would be more computer time extensive. The simple equation was first based on
a rigid pile model with resistance calculated at the time of zero velocity used (Goble and Rausche,
1970a). Today this approach is also called an unloading point method and is applied to the
interpretation of the Rapid Load Testing Method (Middendorp et al., 1992). However, it soon
became clear that a closed form solution to the wave equation such as has been described by
Timoshenko and Goodier, (1951) would be more accurate (Rausche et al., 1972); it was called the
“Case Method Equation”.
The signal matching procedure developed by the Case team, called CAPWAP®, relied first on
Smith’s numerical analysis approach (Rausche et al., 1972); it was later replaced by the
characteristics solution to the wave equation (Rausche, 1988) called CAPWAPC (today it is again
referred to as CAPWAP). Today a fast automatic signal matching program, iCAP (Likins, et al.,
2012b) is also available for uniform, driven piles. While not taking the place of the more accurate
and powerful CAPWAP program, iCAP can replace the Case Method with the advantage that it
does not require an estimated damping factor. Also, results from iCAP are unique, i.e., independent
of user experience. In the 1970s other formulas were also developed for the assessment of hammer
performance, pile stresses and pile integrity from pile top force and velocity as discussed below.
Two different deep foundation failures have to be distinguished Geotechnical and Structural
failures. The former is best detected or prevented by loading tests and, for driven piles, dynamic
monitoring. The latter is generally only measured and assessed by dedicated integrity tests. While
the static loading test is least suitable for detecting structural deficiencies unless they cause the
pile to catastrophically fail during the loading, the dynamic loading test has a better chance to
detect defects as will be discussed in the Section on dynamic monitoring. The following describes
the most common and recognized integrity test methods and lists additional integrity tests that are
less commonly used after mentioning a few construction monitoring methods which can help
prevent pile integrity problems.
After completion of the drilling and prior to placing concrete in the bored pile, the shape or profile of the
hole can be measured with a so-called caliper. Assuming that no further changes in the hole occur after the
measurement and during concreting, the shape of the finished pile has been established. Some caliper
devices use mechanical arms while more modern devices use an ultrasonic technique.
For bored piles, when end bearing is considered in the design, the condition of the bottom of the drilled
hole is important and must be “clean”, meaning loose sediment removed, so that end bearing is activated at
a relatively small displacement rather than first compressing a weak debris layer. This is particularly
important in rock sockets. While not directly measuring pile capacity, the SID is being used with a thin
measuring rod penetrating the soft sediments. The rod penetration into the sediment is viewed with a remote
camera which also displays the cleanliness of the bottom surface.
A more advanced inspection tool, the SQUID, actually measures the force and distance required to penetrate
any potential debris layer and also the resistance in the bearing layer using one or more instrumented cones
(Fig. 1). The device is conveniently attached to the Kelly bar or drill stem and quickly inserted in the water
or slurry filled or dry hole. The whole process typically takes less than 15 minutes. When the hole is
confirmed as clean and the required penetrometer force is adequately confirmed, the pile can be concreted
and the end bearing included in the design. Such a device can be particularly cost effective to minimize the
depth of a rock socket by determining when the rock is of sufficient strength.
One of the earliest methods, and one which has great similarities with the high strain method, is low strain
integrity testing (Rausche et al, 1988). The method of data collection is specified in ASTM D5882. The
method can be quickly applied and testing 100 piles in one day is not impossible. Higher testing rates may
lead to poor data quality.
Once concrete of the pile has hardened sufficiently, the pile top is struck by a hand-held hammer which
generates a force wave that travels down the pile shaft, reflects off the pile toe or other cross section changes
and then propagates back to the pile top. An accelerometer, attached to the pile top, measures both the input
and reflections. The acceleration is integrated to velocity and various enhancements to the velocity record
are made to facilitate data interpretation which is based on basic wave propagation theory (Likins and
Rausche, 2000). For example, Fig. 2 shows two records of pile top velocity vs. time enhanced by an
exponential amplification function to compensate for signal losses due to soil and pile material damping.
The top signal shows a strong positive reflection at a time corresponding to the designed 25 m length of the
pile. The bottom graph, on the other hand, displays a smaller reflection corresponding to a 15 m depth
suggesting an anomaly of pile size or concrete quality 10 m above the pile toe.
Occasionally a different type of low strain data interpretation, referred to as the Transient Response
Method, (Rausche et al., 1991), is employed. Actually, it requires that, in addition to the velocity, the impact
force of the hand held hammer is measured. It calculates by Fourier Transfer the Mobility of the pile in the
frequency domain thereby yielding, for example, a pile stiffness value.
Crosshole Sonic Logging (Likins et al, 2004), commonly called “CSL”, requires installation of at least two
water filled access tubes in the pile so that concrete quality can be assessed by measuring the wave travel
time in the concrete between a transmitter in one tube and a receiver in the other tube. Typically one tube
for each 300 mm of pile diameter is required. The method is standardized in ASTM D6760. From the “First
Arrival Time” (FAT) of the signal and the spacing between access tubes, the wave speed can be calculated.
Cement produces heat during the curing process. This phenomenon is the basis for Thermal Integrity
Profiling of the entire cross section of bored piles (Piscsalko et al., 2013). During the curing of a bored pile,
the center of the pile has the highest temperature while the perimeter has the lowest temperature since it is
adjacent to the soil and the heat is flowing from the pile into the soil. The more cement content in a concrete
mix, the higher the temperature created and, therefore, the lower the temperature the lower the concrete
quality or pile size. The Thermal Integrity Profiling method procedures are governed by ASTM D7949.
While temperature can be sensed by infrared probes in access tubes, it is more convenient to measure the
temperature by attaching cables with thermal sensors to the reinforcing cage. One such instrumented cable
is installed equidistantly around the cage for each 300 mm of pile diameter. The average temperature of the
shaft can be correlated to the effective average shaft radius. Local deviations from the average shaft
temperature can then be related directly to deviations from the average shaft radius, allowing for a complete
evaluation of the entire cross section including the concrete cover outside the reinforcing cage. Fig. 4 shows
a 3D image calculated from temperature measurements. It should be noted that the test has to be performed
while the concrete cures which makes for a quick turnaround of results (often be completed within 24 hours
of casting concrete), saving valuable construction time.
There are other methods which have been reported on and which are occasionally used depending
on the purpose and or preference of the specifier. The following lists first three of more
successfully used methods and then two additional ones that are less frequently used.
1. The Gamma-Gamma Method helps identify concrete cover defects by measuring
concrete density around the inspection tubes and which is used in conjunction with CSL
testing.
2. The Length Inductive Test Method (LITE) which also requires a borehole next the pile to
be tested for length; in this case the pile has to be made of metal.
3. The Parallel Seismic Method which requires a borehole next to the pile, an accelerometer
in the borehole and a light hammer impact on the pile top for pile length determination.
4. The Bending Wave method which applies impacts to a pile (typically a timber pile under
a structure) perpendicular to its axis (there are similarities with the Low Strain Method).
5. The Vibration Method which applies a variable frequency vibration to the pile top and
which is evaluated in the frequency domain similar to the Transient Response Method.
Fig. 5. Example Bearing Graph Input. Fig. 6. Example Bearing Graph Output
For impact driven piles, an installation log showing hammer energy setting and observed blow
count is still an indispensable quality control tool, however, in addition electronic measurements
are frequently taken for a quantification and more accurate assessment of pile driving equipment
where
F(t1) is the measured force at a chosen time, t1,
V(t1) is the measured pile velocity at the same time t1
t2 = t1 + 2L/c
L is the length of the pile below the sensors
c is the pile material wave speed and
Z = EA/c is the pile impedance
E is the pile elastic modulus,
A is the pile cross sectional area and
J is the damping factor which is related to the soil type, typically ranging from 0.4 for coarse
grained soils to 1.0 for cohesive soils. It is normally determined by CAPWAP analysis of
one of the records obtained.
Equation (1) is evaluated for all times t1 between impact and the end of each of the acquired
records, i.e., of all monitored hammer blows. The maximum value of resistance, Rx, is considered
the best estimate of the ultimate soil resistance. The damping factor, J, has to be estimated based
on soil grain size or determined by correlation with either CAPWAP analyses or static load tests.
The maximum value of E(t) is often referred to as EMX or ENTHRU and is a valuable hammer
performance indicator. The ratio of EMX to rated energy, Er, is called transfer ratio (also transfer
efficiency or global efficiency). Experience has shown that certain types of hammers such as
diesel, air/steam or hydraulically powered ones, have certain acceptable ranges of energy transfer
for certain types of piles.
Stresses at the pile top are directly obtained from the strain measurements. Using one-
dimensional wave propagation theory, the average compression at the pile toe and the maximum
tension at any location along the shaft can also be evaluated from the pile top measurements.
Keeping these stresses below the recommended limits based on structural material properties
reduces the possibility of pile damage (Hannigan et al., 2016).
The extent of damage and its location can be quantified using the so-called Beta Method
(Rausche and Goble (1979), Rausche et al., 1988 and Likins and Rausche (2014)). This method,
applicable to uniform piles, is based on the wave propagation based theory and basically calculates
an integrity indicator less than 100% if a tension wave reaches the pile top prior to the arrival of
the stress wave from the pile toe. Clearly, this method has a great similarity to damage detection
by the Low Strain method however, rather than investigating only the velocity record for
reflections, the force signal now serves as a reference line. Fig. 8 shows two records of a 41 m
long 380 mm square section concrete pile. The records consist of force and velocity times pile
impedance (Z=EA/c, see Eq. 1 above); the product of velocity and impedance has the units of
force. The vertical lines in these records indicate the onset of impact, to, and the onset of the wave
return from the pile toe, i.e., to plus 2 times the pile length below sensors divided by the wave
speed (2L/c).
Fig. 8: Pile top force (solid) and velocity (dashed) records of a pile tested
before (top) and after damage occurred (bottom).
Dynamic monitoring results for each hammer impact are conveniently plotted either vs. pile
penetration depth or vs. blow number. These results are instructive and may be used, for example,
to optimize the driving process. Fig. 9 shows for the 1724x25.4 mm open ended pipe pile,
discussed above, pile top and bottom stresses (left), transferred energy and blows/minute (center)
and blow count plus Case Method resistance for a damping factor J=0.5 (right). This is a
comprehensive summary of results for every hammer blow and would, for example, indicate that
the hammer energy setting should be reduced for areas with high stresses or that a reduced pile
length is acceptable, if capacity is reached earlier than originally anticipated. Also, this information
can be conveniently used to performed so-called refined wave equation analyses, i.e., wave
equation analyses with hammer, driving system and soil parameter input values adjusted for
energy, stress and blow count results matching the field observations.
Each hammer blow applied to a pile loads the pile during a short time period and activates soil
resistance forces. The force and velocity data recorded under an impact loading can, therefore, be
interpreted as a very quick loading test. Both driven piles and cast-in-situ piles can be tested in that
manner (Rausche and Seidel, 1984 and Seidel and Rausche, 1984).
Since changes in the soil occur during pile driving, the end of driving blows may not be
indicating the soil resistance that a static loading test would or what would be present under long
term loading. Both soil resistance increases (soil setup) or decreases (relaxation) with time after
pile installation have been observed. While for monitored and impact driven piles the end-of–
driving (EOD) data is valuable for quality assurance and control purposes, it is, advisable to
perform a restrike test after a certain, soil-type dependent waiting time. The beginning of restrike
(BOR) data are usually closely related to the long term bearing capacity. During restrike testing it
is possible that soil setup is so high that the installation pile hammer may not enough energy to
move the pile under a single blow sufficiently (at least 2 mm) to mobilize the available ultimate
soil resistance. In that case only a lower bound of the available soil resistance, or “proof load”,
would be calculated from the pile top measurements. To remedy this problem a high energy drop
hammer is often mobilized as would also be necessary when cast-in-situ piles are tested. Another
remedy would be the so-called superposition, i.e., combining EOD end bearing with BOR shaft
resistance (Hussein et al., 2002).
The dynamic loading can be performed in two ways; the first often uses a pile driving hammer
whose driving system, in particular, the cushions have been designed for a quick and efficient
loading with cushioning just enough to protect the pile from excessive stresses. If a drop hammer
is used, the pile top cushion is designed so that a high energy transfer is possible. This test
configuration is covered by ASTM D4945.
As an example of a CAPWAP result, the beginning of restrike (BOR) test of the wave equation
and monitoring case discussed earlier is shown in Fig. 11. This example, the same that was used
for demonstrating the wave equation approach, is particularly interesting because of occasional
problems encountered with the dynamic testing of large diameter open ended pipe piles where full
end bearing may not be mobilized in dynamic test due to soil plug slipping caused by inertia effects
(Brown and Thompson, 2015). The CAPWAP summary page of Fig. 11 includes the record
analyzed (upper right), the force signal match (upper left) the calculated resistance distribution
(lower right) and the calculated load-displacement curve (lower left). The numerical values shown
indicate a calculated total capacity of 11,700 kN with 3,400 kN end bearing. The end-of-driving
(EOD) results, also analyzed by CAPWAP, indicated a lower capacity of 6,300 kN with 1,600 kN
end bearing demonstrating an almost doubling of capacity during the 10 day waiting time between
EOD and BOR.
Compared to the wave equation prediction shown above, obviously much lower capacity
values have been calculated by CAPWAP. This difference is primarily due higher dynamic
resistance factors: for the end-of-drive the CAPWAP quakes were 7.5 and 9.6 mm vs. the standard
2.5 mm used in the wave equation analysis. This is probably due to an increased pore water
pressure which made the soil “spongy” or “bouncy” and thus more energy absorbing. For the
restrike it was noticed that primarily the shaft damping was much higher than normally assumed
(1.0 vs 0.16 s/m). Repeating the analyses, as so-called Refined Wave Equation Analyses, with the
CAPWAP calculated quakes and damping for EOD and a strongly increased toe damping for BOR
yielded the bearing graphs shown in Fig. 9, right, with capacities of 6,600 and 11,500 kN vs. 6,300
and 11,800 kN CAPWAP. (Note that adjustments to the CAPWAP soil parameters are usually
needed to achieve blow count agreement, because of differences in the blow count calculation
approaches of the two programs). Refined Wave Equation Analyses are useful when analyzing
different situations at the same site based on a representative dynamic load test. Dynamic
measurements and/or local experience are obviously invaluable for assessing proper dynamic
resistance quantities and therefore better wave equation predictions.
1
250
0 1
250
0
5
000 5
000
1
5 9
5ms 1
5 9
5ms
-2
500 8L
/c -2
500 8L
/c
-1
000
0 -1
000
0
P
ileIm
ped
ance
1
500
S
h aftRe sista
nce L
e ngthb .S enso rs 4
5.4m
L
oad(kN
)
P
ileTop D
istrib
ution E
m be d m e nt 1
9.1m
0 2
000 4
000 6
000 8
000 1
000
0 1
200
0 1
200
Botto
m T
o pA re a 133
4.4cm 2
0
.0 E
n dB e a ringA re a 2
.33m2
9
00 T
o pP e rim eter 5
.42m
kN/m
5
.0 T
o pE -M od ulus 2
06843MP a
R
U= 1
1659k
N 6
00 T
o pS p e c.W eigh t 7
7.3kN/m 3
1
0.0 S
F= 8259k
N T
o pW aveS pd. 5123m/s
E
B= 3400k
N 3
00 Overa llW .S. 5275m/s
Displacement (mm)
D
y= 20
.1mm
1
5.0 0
D
x= 22
.1mm MatchQu a lity 1.39
0
S
ET/B
l= 2
.0mm T
opC o
m p r.S tre ss 1
21.0MP
a
2
0.0 2
000 E
B
MaxC om pr.S tress 1
44.1MP
a
MaxT ensio nS tress -2
0.52MP
a
2
5.0 4
000
Avg.S
h aftQu
a ke 2.9m m
6
000
kN
3
0.0 T
o eQua ke 2.8m m
S
F Avg.S
h aftSmithDpg
. 1
.01s/m
8
000
3
5.0 T
o eSmithDam pin
g 0
.58s/m
1
000
0 P
ileF
orce
a
tR u
4
0.0 1
200
0
Fig. 11. CAPWAP summary for the 2nd restrike blow of the open ended pipe example.
The same open ended pipe pile was also subjected to an instrumented static loading test, 16 days
after the restrike test., allowing for a comparison of statically and dynamically determined pile top
load-displacement curves; the dynamic and static displacements are shown cumulatively in
chronological testing order in Fig. 12, left; obviously the EOD test showed a relative lack of
stiffness compared to the BOR and the static tests, apparently caused by the large quakes
(equivalent to a reduced soil stiffness). The static test was conducted by applying 5 loading cycles.
The envelope enclosing the five cycles indicated a 12,500 kN capacity for the Davisson failure
criterion (Fellenius, 1990) which suggests a small, additional soil setup during the 16 day waiting
period after the restrike test; in any event a good agreement with the BOR test was achieved.
Instrumentation attached to the pile along its length also allowed for a comparison of the forces in
the pile at the corresponding CAPWAP load level and the Davisson failure load. Obviously, the
resistance distribution calculated from the dynamic test is similar but not very close to the statically
determined one (Fig. 12, right).
Continuous Flight Auger (CFA) or Augered-Cast-in-Place piles (ACIP) are produced by drilling
with an auger of the length of the pile and then pressure injecting grout from the bottom up through
the hollow stem of the auger while the auger is extracted. While the monitoring of this installation
process is valuable (see Pile Installation Recorder above) testing for geotechnical and structural
sufficiency is equally important. Statically testing a few piles in the beginning of production piling
may be good to proof the overall suitability of the pile for the project, however, usually only a few
static loading tests are performed even though the number of piles on a site may be very large.
Quick and relatively inexpensive dynamic testing is, therefore, frequently performed too, and
agreement between static and dynamic test results has generally been very good. Also in recent
years the use of embedded sensors has generated valuable comparison records which should, in
the future, lead to even better agreement between dynamically and statically determined resistance
distributions. Of course, for cast-in-situ piles, the question of elastic modulus and cross sectional
area below grade is always a problem and may lead to inaccurate calculation of force from strain.
The following example shows results from both static and dynamic loading tests on a 44.5 m
long CFA pile of 610 mm diameter which was auger-cast through deep soft cohesive soils into a
sedimentary rock. Dynamic acceleration and force were measured on pile and with a top
transducer, respectively. The pile had been instrumented with strain gaged sister bars attached to
the center reinforcement bar of the pile at six locations. After the grout had sufficiently hardened,
the dynamic load test was performed by dropping a 16-ton APPLE ram from various drop heights
onto the lightly cushioned top transducer. The permanent displacement occurring under the
dynamic impact analyzed was 3.2 mm. A few days later a static load test was performed.
For the interpretation of the static test on cast-in-situ piles, AASHTO specifies an offset
criterion of 3.8 mm + D/120 for a 610 mm (or less) diameter (D) pile and 3.8 mm + D/30 for a 914
mm pile (or greater), with linear interpolation. In the present case the offset elastic line would
intersect the slightly extrapolated static load test curve at a little less than 10,000 kN while the
6. SUMMARY
Ideally, the monitoring of pile installation, both by driving and drilling, would avoid the need for
additional quality assurance methods and both pile installation recording and dynamic monitoring
are a great help in these efforts. However, the complexity of soil behavior and construction
procedures require testing of the final deep foundation elements though good construction
monitoring may reduce the number of required tests.
A variety of structural (integrity) and geotechnical (loading) tests have been developed, all
with their specific benefits and limitations. Dynamic methods, based on wave propagation theory,
generally have the advantage of being the least costly and least time consuming.
With proper care and good loading equipment (drop weights), today’s dynamic loading tests
are of high quality and yield good agreement with static tests. However, differences due exist; they
can often be avoided by understanding the effects of the dynamic loading on the soil and potential
error sources in both static and dynamic test procedures and interpretations. Increased use of
instrumented dynamic and static testing will lead to further understanding of the load transfer in
7. ACKNOWLEDGEMENT
The authors wish to thank Berkel and Company Contractors as well as Six Construct for
contributing to this paper by releasing the data of the examples demonstrating the use of
instrumentation on steel piles as well as ACIP piles.
8. REFERENCES
AASHTO, (2014), LRFD Bridge Design Specifications, 7th Ed., Pub. Code LRFDUS-7, American
Assoc. of State Highway and Transportation Officials, Washington DC, 20001
Allin, R., Likins, G., Honeycutt, J., (2015), Pile Driving Formulas Revisited. Proc. of the Intl.
Foundations Congress and Equipment Expo 2015: San Antonio, TX. American Society of
Civil Engineers: Reston, VA
Alm, T., and Hamre, I. (2001), Soil Model for Pile Driveability Predictions Based on CPT
Interpretation, Proc. of the 15th Int. Conf. on Soil Mechanics and Geotechnical Engineering,
2, Istanbul, Turkey, 1297-1302.
ASTM D1143, “Standard Test Methods for Deep Foundations Under Static Axial Compressive
Load”, American Society for Testing and Materials, West Conshohocken, PA.
ASTM D3689, “Standard Test Methods for Deep Foundations Under Static Axial Tension Load”,
American Society for Testing and Materials, West Conshohocken, PA.
ASTM D 4945, “Standard Test Method for High Strain Dynamic Testing of Deep Foundations”,
American Society for Testing and Materials, West Conshohocken, PA.
ASTM D5882, Standard Test Method for Low Strain Integrity Testing of Deep Foundations,
American Society for Testing and Materials, West Conshohocken, PA.
ASTM D6760, Standard Test Method for Integrity Testing of Concrete Deep Foundations by
Ultrasonic Crosshole Testing, American Society for Testing and Materials, West
Conshohocken, PA.
ASTM D7383, Standard test Methods for Axial Compressive Force Pulse (Rapid) Testing of Deep
Foundations, American Society for Testing and Materials, West Conshohocken, PA.
ASTM D7949, Standard Test Method for Thermal Integrity Profiling of Concrete Deep
Foundations, American Society for Testing and Materials, West Conshohocken, PA.
Beringen, F.l., van Hooydonk, W.R., and Schaap L.H.J., (1980) Dynamic Pile Testing: an Aid in
Analyzing Pile Driving Behavior. Proc. of the Intl. Seminar on the Appl. of Stress wave
Theory on Piles, Stockholm, A.A. Balkema, Rotterdam.
Bermingham, P., Janes, M., (1989), An Innovative Approach to Load Testing of High Capacity
Piles, Int’l Conf. on Piling and Deep Foundations, London.
Brown, D., Dapp, S., Thompson, R., Lazarte, C., (2007), Geotechnical Engineering Circular NO.
8 Design and Construction of Continuous Flight Auger (CFA) Piles, Report No. FHWA-HIF-
07-03, Federal Highway Administration, Washington, D.C.
1. INTRODUCTION
Energy is required to sustain our lives. The world’s population could reach 9 billion by the year
2040. Population growth and increasing energy demands from developing nations combine to
cause an anticipated 50% increase in the power demand over the next two decades. Fossil fuels
(petroleum, coal and natural gas) account for about 85 percent of the primary energy consumed
worldwide. Fossil fuel reserves exceed the needs for the next century, however, there are
increased concerns with rising CO2 levels in the atmosphere and climate change implications.
The situation is aggravated by the disparity in time scales between the time required for national-
scale decisions (e.g., the 2-to 4-year political cycle), and the time scales for phenomena that
affect energy infrastructure and the environment (e.g., the 50-year life of energy infrastructure
and the 100,000 year half-life of some radioactive isotopes in high-level nuclear waste).
The geotechnical discipline is central to energy. Energy geo-engineering involves site
characterization, infrastructure development, resource recovery, energy storage,
decommissioning, and waste geo-storage. Facilities and infrastructure often require deep
foundations, including wind turbines, tanks, buildings, and towers. However, unique conditions
develop in the context of energy applications.
This manuscript explores three special cases. The first case centers on deep foundations in
fractured rock in compression and tension. The second study addresses the addition of heat
exchangers for underground heat storage, and the potential for thermo-mechanical ratcheting.
The third study extends soil-structure interaction concepts developed for deep foundation to the
analysis of wells used for dewatering and resource recovery, in particular, the generation of
negative skin friction which may cause the structural collapse of the well (the case of sand drains
used to accelerate the compaction of fine-grained layers has similar hydro-mechanical coupled
features). These problems require alternative analyses, new constitutive models, and adequate
numerical simulation schemes.
Qi = Qi +1 + si ⋅ π ⋅ D ⋅ L0 equilibrium (1)
where the shaft resistance is the side friction acting on the ith-slice si [Pa] times the element
contact area.
Side friction si [Pa]. The pile-soil shear strength siult = σ′vi·K0·µ reflects the vertical effective
stress σ′vi [Pa], the coefficient of horizontal stress K0 [-], and the pile-soil interface friction
coefficient µ [-]. The mobilized side friction has an elasto-plastic displacement response:
⎧ ⎛ δ i + δ i +1 ⎞
⎪ − siult if ⎜ ⎟ ≤ −δ s*
⎪ ⎝ 2 ⎠
⎪ ⎛ δ + δ i +1 ⎞ ⎛ δ + δ i +1 ⎞
−δ * < ⎜ i ⎟ < δs , (2)
*
si = si (δ i , δ i +1 ) = ⎨k i ⋅ ⎜ i ⎟ if
⎪ ⎝ 2 ⎠ ⎝ 2 ⎠
⎪ ⎛ δ + δ i +1 ⎞
⎪ siult if δ s* ≤ ⎜ i ⎟
⎩ ⎝ 2 ⎠
where δi [m] and δi+1 [m] are the relative pile-soil displacements at the element upper and lower
interfaces, ki= siult/δs* [Pa/m] is the shaft stiffness, and δs* [m] is the critical relative displacement
to mobilize the pile-soil shear strength siult [Pa]. This value includes the elastic displacement field
away from the pile and it is pile-size dependent (δs* = 0.005·D to 0.02·D - Hirayama 1990, Reese
1978).
Compatibility. The change in the element length Δi = δi+1 − δi [m] combines thermal ΔTi [m] and
stress-dependent effects Δσi [m]:
Qi + Qi +1 L0
Δi = δ i +1 − δ i = ΔTi − Δσi = α ΔT L0 - (3)
2 A⋅ E
in terms of the thermal expansion coefficient α [°C-1], the temperature change ΔT [°C], the pile
cross-section A = π · D2/4 [m2] and Young’s modulus E [Pa].
Pile tip resistance Qtip [N]. We adopt an elasto-plastic tip response, with a constant critical
relative displacement δtip* [m] required to mobilize the ultimate tip resistance Qtipult [N]:
ult
⎧⎛ Qtip ⎞
⎪⎜ * ⎟⎟δ tip
⎜ if 0 < δ tip < δ tip*
Qtip = Qtip (δ tip ) = ⎨⎝ δ tip ⎠ (4)
⎪ Q ult if *
δ < δ tip
⎩ tip tip
z/L
Ultimate pile capacity Qult [N]. It is the sum of the shaft and the tip resistance Qult = S ult + Qtip
ult
.
head ult head
Then, the factor of safety for an applied load Q is FS = Q /Q .
2006). Furthermore, we assume that the soil temperature at depth remains relatively unaffected
by daily and seasonal weather changes.
0.5 0.5 0.5 Cycle 1
z/L
3.2 Results
The mobilized shaft friction and axial load changes with the number of cycles. Fig. 2 shows the
initial condition and after 50 cycles
1 1 1
0 0 1000 2000 0 0 30 60 0 0 5 10
Normalized Depth z/L
Normalized Depth
1 1 1
0 1000 2000 0 30 60 0 5 10
(a) Axial force
Axial Force Side Friction
(b) Side friction Rel. Displ.
(c) Rel. Displace.
Q [kN]
Q [kN] ss [kPa]
[kPa] δδ [mm]
[mm]
Heating expands the pile in both directions: up in the upper part and down in the lower part.
Thermal contraction upon cooling partially reverses this trend, and plastic displacements
accumulate. The following cases help clarify trends:
• End-bearing pile with no shaft resistance (Sult/Qtult 0). The pile elongates and contracts with
thermal cycles, but the tip load is constant and there is no thermally induced settlement at the
tip.
• No static load applied Qhead =0 (FS = Qult / Qhead ∞). Shaft and tip resistance become
established after the first thermal cycle and there is no accumulation of permanent
displacements. As a corollary, we can conclude that settlement accumulation during thermal
cycles requires the bias effect of an applied pile load Qhead
• With static load applied Qhead>0. Thermo-mechanical ratcheting (continuous settlement) can
only happen if the applied load Qhead and the shaft resistance Sult exceed the tip resistance
Qhead + Sult > Qtipult. Otherwise, the accumulation of permanent displacement will tend
towards an asymptotic value (shakedown condition).
Settlement accumulation reaches 63% of the asymptotic pile settlement when the number of
cycles n equals the characteristic number N*. Most cases analyzed as part of this study show that
the characteristic number of cycles is less than 10-or-20 for common field conditions where the
factor of safety exceeds FS=3-or-4.
Global: small w/h ratio. In this case, depressurization causes an extensive compression of the
reservoir, all layers above the production layer will move downwards and will tend to mobilize
negative skin friction (typically, this is the case of sand column patterns used in field
compaction). The solution is based on standard soil-pile interaction analyses, where equilibrium
conditions relate the change in the casing axial force Pz [N] at depth z [m] to the mobilized shear
resistance against the casing τz [kPa] at the same depth (Poulos and Davis 1980)
∂Qz
dz = − π d well τ z dz (6)
∂z
This analysis of sediment-well interaction assumes that the sediment column above the
production layer is a perfectly rigid body that settles a prescribed amount across the production
horizon. The solution is sensitive to values selected for yield displacement δy (different values
than those assumed in 3D FEM in order to account for sediment deformation), settlement of the
production horizon, and tip stiffness and bearing capacity.
Typical results in Fig. 3 show the mobilization of the negative skin friction above the
production horizon, high peak axial loads near the top of the production horizon, and the critical
role of tip stiffness on the mobilized peak load. This last observation suggests the need for
engineering well completion using soft-end conditions. Note that large axial loads may force the
well to penetrate into the lower layer (i.e., mobilize tip resistance) or cause the longitudinal
failure of the well.
(a) Shaft Resist. fs [kPa] (b) Pile Load Pz [kN] (c)
-500 0 500 -2000 0 2000 4000
0 0 12,000
20 20 10,000
controlled by
40 40 shaft capacity
Max Axial Load [kN]
8,000
60 60
6,000
Depth z [m]
Depth z [m]
80 80
4,000
100 100
2,000
120 controlled by production
120
induced settlement
0
140 140 0.01 0.1 1 10 100 1000 10000 100000
Tip Stiffness k [MN/m]
160 160
Fig. 3. Production well under “global” settlement condition. (a) mobilized shaft
resistance, (b) axial load, (c) maximum axial load as a function of tip stiffness. Case:
155m long well, casing OD=128mm, production horizon between z=140 and z=155m,.
Local: large w/h ratio. The analysis of single wells, or wells with separation greater than the
layer thickness s/h>>1 cannot be conducted with an equivalent 1D formulation, in fact, we used
a fully coupled hydro-mechanical finite element model to explore the consequences of
depressurization. In particular, the analysis must carefully capture the evolution of hydraulic
conductivity k [m/s] as a function of the void ratio e: we used a power equation k/k0=(e/e0)b
where k0 is the hydraulic conductivity at the reference void ratio e0, and the b-exponent captures
the sensitivity of hydraulic conductivity to changes in the void ratio. For fine grained sediments
the exponent can be b=4 and larger (Ren and Santamarina 2017). The coupled hydro-mechanical
FEM model shows the complex nature of sediment-well interaction during depressurization (Fig.
4). Trends from the parametric study show:
-20 b=0
b=1
-40 b=2
b=3
-60
Depth [m] b=4
Depth [m]
-80 b=5
b=4.0
b=4.0 b=6.0
b=6.0
b=6
-100
-120
-140
-160
Fig. 4. Production well under “local” settlement conditions. (left) Pore water pressure
around the production zone for various b-exponents in the constitutive model for
hydraulic conductivity. (right) Axial force distribution along the casing.
5. CONCLUSIONS
Fracture characteristics and orientation greatly affect the stress field, deformation and strength of
deep foundation systems in fractured rock masses. Special caution is required when analyzing
caisson-caisson and caisson-boundary interaction such as in the vicinity of excavations and
tunnels. The situation is aggravated when changes in hydrological conditions are anticipated.
Energy piles subjected to thermal cycles experience thermally-induced pile displacements
that can result in the accumulation of plastic displacements with the number of thermal cycles,
even though, the ultimate pile capacity remains constant. Ratcheting is unlikely in most
applications. Instead, the plastic displacements accumulate towards an asymptotic settlement that
is proportional to the free thermal expansion of the pile and inversely proportional to the factor
of safety. Preliminary results suggest that most of the thermally-induced plastic displacements
take place in the first few cycles (typically less than 10 to 20 cycles for standard applications).
6. ACKNOWLEDGMENTS
Support for this research was provided by the KAUST Endowment at King Abdullah University
of Science and Technology. G. Abelskamp edited the manuscript.
7. REFERENCES
Amadei, B. and Goodman, R.E. (1981). A 3-D constitutive relation for fractured rock masses,
Proceedings of the international symposium on mechanical behaviour of structured media,
pp. 249-268.
Amatya, B. L., Soga, K., Bourne-Webb, P. J., Amis, T., and Laloui, L. (2012). Thermo-
mechanical behaviour of energy piles. Géotechnique, 62(6), pp. 503-519.
Bourne-Webb, P. J., Amatya, B., and Soga, K. (2012). A framework for understanding energy
pile behaviour. Proceedings of the ICE - Geotechnical Engineering. pp.?
Bourne-Webb, P. J., Amatya, B., Soga, K., Amis, T., Davidson, C., and Payne, P. (2009). Energy
pile test at Lambeth College, London: geotechnical and thermodynamic aspects of pile
response to heat cycles. Géotechnique, 59(3), pp. 237-248.
Brandl, H. (2006). Energy foundations and other thermo-active ground structures. Geotechnique,
56(2), pp. 81-122.
Cundall, P.A. (1980). UDEC-a generalized distinct element program for modelling jointed rock.
Report PCAR-1-80, Peter Cundall Associates, European Research Office, US Army Corps
of Engineers- pp.??
Coyle, H. M., and Reese, L. C. (1966). Load transfer for axially loaded piles in clay. Journal of
Soil Mechanics and Foundations Div, 92(2), pp. 1-26.
Desai, C.S., Zamman, M.M., Lightner, J.G., and Siriwardane, H.J. (1984). Thin layer element for
interfaces and joints, Int J Numer Anal Methods Geomech, 8, pp. 19-43.
Gan, Q., and Elsworth, D. (2016). A continuum model for coupled stress and fluid flow in
discrete fracture networks, Geomech Geophys Geo-Energy Geo-Resour, 2(1), pp. 43-61
Goodman, R.E., Taylor, R.L., and Brekke, T.L. (1968). A model for the mechanics of jointed
rock, J Soil Mech Div ASCE 94, SM3. pp. 637-59.
Goodman, R. E. (1989). Introduction to Rock Mechanics, Wiley, New York
ABSTRACT. The paper presents two technologies for deep foundations expansion devices for
increasing stiffness and reducing uncertainty of foundation: the Expander Body (EB) and the Toe
Box (TB). The common characteristic of both devices is their ability to stiffen the toe response by
pre-compressing the soil at the toe level. This is achieved by expanding the devices by pressure
grouting, the EB expands horizontally and the TB vertically. Both devices compress the soil, which
increases its strength and stiffness. During grouting, pressure and volume are measured, providing
important information of the initial and final state of stress of the soil. The measured pressure-
volume relationship can be correlated to geotechnical parameters, similar to a pressuremeter test.
The measured pressure-volume parameters, combined with the soil characterization, allow
estimating the toe response for every pile. The EB and TB devices can be used as quality control
of the performance of each pile, as it becomes possible to identify unexpected soft deposits and to
take remedial measures during the construction phase.
Table 1 shows the size of EB models, prior to and after full expansion. Note that as a result
of inflation, the length of the EB is reduced. As this effect can result in some soils in
decompression below the pile toe, the above-described post grouting device can be used to
increase the pressure below the EB, thereby enhancing the strength and stiffness of the soil
below the EB.
Inspection of the pressure-volume curve provides valuable information regarding the strength
and deformation properties of the soil at the pile toe. An important advantage is that soil properties
are obtained at the pile toe after installation of the pile, thereby taking into account the effects of
pile installation process. This aspect is illustrated by the following example. Figure 3 shows the
effect of pile installation method on the pressure-volume behavior of an EB. Piles TP3 and TP4
were located a few meters apart and founded on a dense sandy soil with SPT value N60>20. TP3
was installed to a depth of 9.6m and TP4 to a depth of 17.5m.
Fig. 3. Effect of the pile installation method on the pressure-volume response. TP3: full
displacement auger pile, 9.6m long, TP4: slurry drilled pile, 17.5m long.
Fig. 5. Pressure vs. injection curves of piles with EB 612, founded in weathered sandstone in
Asunción-Paraguay.
In the case of auger piles, the EB can be grouted shortly after installation and up to a period
of several weeks. The maximum time period of grouting after installation of auger piles was 7
months. As the concrete cover of EB auger and drilled piles under bentonite is relatively thin and
expansion of the EB produces tension stresses, the grout cover breaks easily during expansion of
the EB. In a test performed in a concrete cylinder of 60 cm in diameter, the needed pressure to
break the grout cover was only 200 kPa. Thus, grouting of EBs in a pile group can be performed
in time intervals of one or two weeks.
Post-grouting of the soil can be used to increase the stress below the inflated EB, cf. Figure 8.
Figure 8a illustrates the stress field developed during EB expansion (red arrows) and the
subsequent post-grouting (yellow arrow). A special valve was developed for injecting grout under
controlled conditions below the EB. During the post-grouting process, grout volume and injection
pressure are measured, thus providing additional information about the response of the soil below
the pile toe. Figure 8b shows a typical post-grouting diagram. Note that this diagram was obtained
after EB expansion had been carried out.
Fig. 8a. Illustration of stress field due to grouting Fig. 8b. Post-grouting diagram
TABLE 2. Typical report of EB test results, providing design information for the EB database.
Fig. 9a. Variation of grouting results in a project with EBs embedded into a medium dense sand
layer.
Fig. 9b. Variation of grouting results in a project with EBs embedded into a gravely clay layer.
Fig. 9c. Variation of grouting results in a project with EBs embedded into a very stiff clay layer.
Extensive pile loading tests have been performed on piles with and without EBs. In all cases
the pile toe response with EB shows a remarkable increase of service load. For similar geometry
(pile diameter ranging between 350mm and 450 mm, lengths between 12 m and 18 m) and similar
soils (mostly silty sandy soils, medium dense, with clayey layers), the piles with EB have service
loads 2.5 to 4 times higher than identical piles without EB, for the same deformations measured at
the pile head. This behavior is consistent with the increase of the toe response, (cf. Fellenius et al,
2014, Terceros et al. 2014, Rosales et al. 2015). Table 3 shows statistical information about average
values and standard deviation values of injection pressure from different projects.
TABLE 3. Examples of injection pressures and their variation from projects in different
geotechnical conditions.
STANDARD % of SD REFERRED
AVERAGE GROUT
PROJECT DEVIATION TO AVERAGE
PRESSURE (MPa)
(MPa) VALUE
SILOS SOFIA 1.97 0.400 20.263
TRES CARABELAS 4.54 0.680 14.978
SANTA ELENA 4.06 0.622 15.320
FLORENCIA 3.81 0.721 18.924
ITAGUAZÚ 4.44 0.784 17.659
UAGRM 3.89 0.432 11.112
PARQUEO ALAS 3.63 0.330 9.091
ITATI 4.96 0.612 12.342
Fig. 10a. Toe Box. Fig. 10b. Toe Box with non-woven
geotextile.
About one week after concreting the pile, the toe-box is expanded axially by pressure-grouting
with the shaft resistance serving as reaction. The expansion of the toe-box (mostly the pile toe
displacement) is measured by two telltales. All associated pile compression and pile head
movements are also measured during the process. The grouting process is slow, taking at least an
hour to complete, in order to prevent build-up of pore pressure. The grout pressure and grout
volume is monitored at the pump and the records are saved and graphically displayed.
10
20
30
40
50 Toe, downward movement
60
6
Pressure vs. volume
5
Pressure (MPa)
0
0 100 200 300 400 500 600
Volume (L)
5.1
GRAVELLY SAND
13
(5mm gravel)
GRAVELLY SAND
6 8.3
(5mm gravel)
HIGH PLASTICITY
68
9.5 CLAY
10.2
17
DENSE SAND 12 11.4
13.4
GRAVELLY SAND (8mm
7
gravel) DENSE SAND 17
14.9
16.5
HIGH PLASTICITY
77 GRAVELLY SAND
CLAY 21
(8mm gravel)
18.3
19.9
DENSE SAND 22
20.5
TOE-LEVEL TOE-LEVEL
HIGH PLASTICITY
93
CLAY
DENSE SAND 13
23.5
LOW PLASTICITY
71
CLAY
25.0 25.0
3 CONCLUDING REMARKS
Both of expansion devices, the Expander Body and the Toe Box, are presented in the paper as
stiffness improvement devices and as quality control devices, giving the possibility to verify the
desired pile toe response. Should the pile response show to be less than expected, the verification
gives the possibility to reinforce the foundation if necessary, while the project is still in its
construction phase, avoiding any possible further problem.
Furthermore, both the Expander Base and the Toe Box can be installed with a wide variety of
drilling systems. The main decision whether to use one or the other rests with the pile diameter.
While the EB is recommended to be installed in a pile of up to 600 mm, because its maximum
practical expansion diameter is 800 mm, the Toe Box starts to be practical with pile diameters
ranging from 600 mm and above.
Loading tests have been performed in piles with both of the systems, demonstrating that, when
the devices are used, the supported working loads of the piles can be considerably increased for
the same level of limiting deformation.
4 REFERNCES
1 General
It is the intention that readers have access to the widest possible range of content and
that this content can be shared to enable the most effective research, study, teaching,
and practice of geotechnical and geo-environmental engineering and associated fields of
application. Efficient dissemination of published information is in the interest of the
geo-engineering profession.
The three main players have – sometimes – conflicting interest in the publication and
information dissemination process. The primary interests are briefly outlined below:
The author(s) of a manuscript is usually not aware of the rights and limitations when
submitting manuscripts to publishers.
This document outlines the Publishing Policy endorsed by ISSMGE and intended to
provide a compromise for producing future publications.
b. these rights should not be used in a way that involves duplicate publication
that will compete with the Publisher’s own publications; and
Ownership of the copyright contained in the Manuscript (“the Material”) remains with
the Author(s) who retain the following rights:
3. Authors and any academic institution where they work at the time may
reproduce their Manuscript for the purpose of course teaching;
Permission must be sought for uses other than those defined above. Other than the use
outlined above no publication may be captured or downloaded electronically into any
format without the Publisher's and Author’s specific written permission.
The substantial investment made by the Publisher in protecting and enhancing the
quality of the content is recognised. It is necessary, therefore, to impose the following
conditions on the use of published Material.
The Publisher shall maintain a firm policy against duplicate publication unless there are
exceptional circumstances.
All requests to reproduce or re-use the whole or a substantial part of the Material in
another publication will be conditional upon the Authors’ consent and subject to the
Publisher’s approval.
The party seeking permission will be instructed to write to the Publisher and to obtain
the Authors’ consent.
3 BOLIVIAN
rd
PROCEEDINGS
VOLUME 1
VOLUME 1